c. neoformans

Access EMERGING INFECTIOUS DISEASES at www.cdc.gov or www.cdc.gov/ncidod/EID/eid.htm

EditorsJoseph E. McDade, Editor-in-ChiefAtlanta, Georgia, USAStephen S. Morse, Perspectives EditorNew York, New York, USA

Phillip J. Baker, Synopses EditorBethesda, Maryland, USA

Stephen Ostroff, Dispatches EditorAtlanta, Georgia, USA

Polyxeni Potter, Managing EditorAtlanta, Georgia, USA

International EditorsPatrice CourvalinParis, FranceKeith KlugmanJohannesburg, Republic of South AfricaTakeshi KurataTokyo, JapanS.K. LamKuala Lumpur, MalaysiaJohn S. MacKenzieBrisbane, AustraliaHooman MomenRio de Janeiro, BrazilSergey V. NetesovNovosibirsk Region, Russian FederationV. RamalingaswamiNew Delhi, IndiaDiana WalfordLondon, United Kingdom

Editorial and ComputerSuitorial AssistanceMaria T. Brito

Electronic DistributionCarol Y. CrawfordDavid L. Smith

Editorial and Computer SupportBeatrice T. DivineTeresa M. HoodP. Lynne StocktonMaria T. BritoCarol Y. CrawfordRachna Dube

Retrieve the journal electronically on theWorld Wide Web (WWW), through file transferprotocol (FTP), or by electronic mail (e-mail).WWW and FTP: Access the journal at http://www.cdc.gov/ncidod/EID/eid.htm, from the CDChome page (http://www.cdc.gov), or download itthrough anonymous FTP at ftp.cdc.gov. Thefiles can be found in the directory pub/EID.LISTSERVer (e-mail lists): To subscribe to a list,send an e-mail to [email protected] with thefollowing in the body of your message: subscribelistname (e.g., subscribe EID-TOC). EID-TOCwill send the table of contents automatically toyour e-mail box. Obtain individual articles onWWW or through FTP, or request copies by e-mail. EID-ASCII will send the the entire journal toyou in ASCII format.

Electronic Access

Emerging Infectious DiseasesEmerging Infectious Diseases is published

four times a year by the National Center forInfectious Diseases, Centers for DiseaseControl and Prevention (CDC), 1600 CliftonRoad, Mailstop C-12, Atllanta, GA 30333, USA.Telephone 404-639-3967, fax 404-639-3075,e-mail [email protected].

The opinions expressed by authorscontributing to this journal do not necessarilyreflect the opinions of CDC or the institutionswith which the authors are affiliated.

All material published in Emerging Infec-tious Diseases is in the public domain and maybe used and reprinted without special permis-sion; proper citation, however, is appreciated.

Use of trade names is for identification onlyand does not imply endorsement by the PublicHealth Service or by the U.S. Department ofHealth and Human Services.

Emerging Infectious Diseases is printed on acidfree paper that meets the requirements of ANSI/NISO 239.48-1992 (Permanence of Paper).

∞

The journal is distributed electronically and in hard copy and is available at no charge .

YES, I would like to receive Emerging Infectious Diseases.

Please print your name andbusiness address in the box andreturn by fax to 404-639-3075 ormail to Journal/CDC/MS C12 1600 Clifton Rd. Atlanta, GA 30333

In Index Medicus/Medline, Current Contents, and other databases

Moving? Please give us your new address (in the box) and

print the number of your old mailing label here__________

Editorial BoardAbdu F. Azad, Baltimore, Maryland, USAJohan Bakken, Duluth, Minnesota, USABarry J. Beaty, Ft. Collins, Colorado, USAGus Birkhead, Albany, New York, USAMartin J. Blaser, Nashville, Tennessee, USAS.P. Borriello, London, United KingdomDonald S. Burke, Baltimore, Maryland, USACharles Calisher, Ft. Collins, Colorado, USAArturo Casadevall, Bronx, New York, USAThomas Cleary, Houston, Texas, USABarnett L. Cline, New Orleans, Louisiana, USAJ. Stephen Dumler, Baltimore, Maryland, USADurland Fish, New Haven, Connecticut, USARichard L. Guerrant, Charlottesville,

Virginia, USABrian Gushulak, Geneva, SwitzerlandScott Halstead, Bethesda, Maryland, USASeyed Hasnain, New Delhi, IndiaDavid L. Heymann, Geneva, SwitzerlandWalter Hierholzer, Atlanta, Georgia, USADagmar Hulìnskà, Prague, Czech RepublicPeter B. Jahrling, Frederick, Maryland, USASuzanne Jenkins, Richmond, Virginia, USAMohamed A. Karmali, Toronto, Ontario, CanadaRichard Krause, Bethesda, Maryland, USABruce R. Levin, Atlanta, Georgia, USAMyron Levine, Baltimore, Maryland, USAStuart Levy, Boston, Massachusetts, USAJohn E. McGowan, Jr., Atlanta, Georgia, USAPatrick S. Moore, New York, New York, USAPhilip P. Mortimer, London, United KingdomFred A. Murphy, El Macero, California, USABarbara E. Murray, Houston, Texas, USAJames M. Musser, Houston, Texas, USANeal Nathanson, Philadelphia, Pennsylvania, USARosanna W. Peeling, Winnipeg, Manitoba, CanadaDavid H. Persing, Rochester, Minnesota, USARichard Platt, Boston, Massachusetts, USADidier Raoult, Marseille, FranceDavid Relman, Palo Alto, California, USARebecca Rico-Hesse, San Antonio,

Texas, USAConnie Schmaljohn, Frederick, Maryland, USARobert Shope, Galveston, Texas, USAPeter Small, Stanford, California, USABonnie Smoak, US Army Medical Research

Unit, KenyaRosemary Soave, New York, New York, USAP. Frederick Sparling, Chapel Hill, North

Carolinia, USAG. Thomas Strickland, Baltimore, Maryland, USAJan Svoboda, Prague, Czech RepublicRobert Swanepoel, Sandringham, South AfricaPhillip Tarr, Seattle, Washington, USALucy Tompkins, Stanford, California, USAElaine Tuomanen, New York, New York, USADavid Walker, Galveston, Texas, USABurton W. Wilcke, Jr., Burlington, Vermont, USAMary E. Wilson, Cambridge, Massachusetts, USAWashington C. Winn, Jr., Burlington, Vermont, USA

Liaison RepresentativesAnthony I. Adams, DHSH, AustraliaDavid Brandling-Bennett, WHO, USAGail Cassell, Lilly Research Lab, USARichard A. Goodman, CDC, USAJoseph Losos, Dept. Health, CanadaGerald L. Mandell, U. VA. Sch. Med., USAWilliam J. Martone, NFID, USARoberto Tapia-Conyer, Sec. de Salud, MéxicoKaye Wachsmuth, USDA, USA

Contents

EMERGING INFECTIOUS DISEASESVolume 4 • Number 1 January–March 1998

Cover: Landscape with Metallic Interventions (1991) by Antonio Henrique Amaral (Brazil, b. 1935). Used with permission fromBrenda and Stephen Morse, Atlanta, Georgia (Photo by Troy Hall).

Antonio Amaral is known for successfully combining elements of Pop Art with Brazilian themes. His paintings are viewed bysome as the revival of the tropicalismo of the Tarsila do Amaral generation.

International Editors—UpdateEmerging Infectious Diseases—Brazil 1

H. Momen

PerspectivesRisk for Transfusion-Transmitted Infectious Diseases in Central and South America 5

G.A. Schmunis, F. Zicker, F. Pinheiro, and D. Brandling-Bennett

Calicivirus Emergence from Ocean Reservoirs: Zoonotic and Interspecies Movements 13

A.W. Smith, D.E. Skilling, N. Cherry, J.H. Mead, and D.O. Matson

Outbreak Investigations—A Perspective 21A.L. Reingold

SynopsesGenetic Diversity of Wild-Type Measles Viruses: Implications for Global Measles Elimination Programs 29

W.J. Bellini and P.A. Rota

Diversity among Multidrug-Resistant Enterococci 37B.E. Murray

Proteases of Malaria Parasites: New Targets for Chemotherapy 49

P.J. Rosenthal

Zoonotic Tuberculosis due to Mycobacterium bovis in Developing Countries 59

O. Cosivi, J.M. Grange, C.J. Daborn, M.C. Raviglione, T. Fujikura,D. Cousins, R.A. Robinson, H.F.A.K. Huchzermeyer, I. de Kantor,and F.-X. Meslin

What Makes Cryptococcus neoformans a Pathogen? 71K.L. Buchanan and J.W. Murphy

DispatchesHantavirus Infection in Children in Argentina 85

N.C. Pini, A. Resa, G. del Jesús Laime, G. Lecot, T.G. Ksiazek,S. Levis, and D.A. Enria

Above: La Voz de la Conciencia (1977)by Perez Celis (Argentina) Used withpermission from Brenda and StephenMorse, Atlanta, Georgia (Photo by TroyHall).

Left: Electron photomicrograph ofcetacean calicivirus Tursiops-1. Photocourtesy of A.L. Smith.

Contents

EMERGING INFECTIOUS DISEASESVolume 4 • Number 1 January–March 1998

Reemergence of Dengue in Cuba: A 1997 Epidemic in Santiago de Cuba 89G. Kourí, M.G. Guzmán, L. Valdés, I. Carbonel, D. del Rosario, S. Vazquez, J. Laferté,J. Delgado, and M.V. Cabrera

Hantavirus Pulmonary Syndrome in a Chilean Patient with Recent Travel in Bolivia 93

R. Espinoza, P. Vial, L.M. Noriega, A. Johnson, S.T. Nichol, P.E. Rollin, R. Wells,S. Zaki, E. Reynolds, and T.G. Ksiazek

Prevalence of Tick-Borne Pathogens in Ixodes scapularis in a Rural New Jersey County 97

S. Varde, J. Beckley, and I. Schwartz

Plague, a Reemerging Disease in Madagascar 101S. Chanteau, L. Ratsifasoamanana, B. Rasoamanana, L. Rahalison,J. Randriambelosoa, J. Roux, and D. Rabeson

Bayou Virus-Associated Hantavirus Pulmonary Syndrome in Eastern Texas: Identification of the Rice Rat, Oryzomys palustris, as Reservoir Host 105

N. Torrez-Martinez, M. Bharadwaj, D. Goade, J. Delury, P. Moran, B. Hicks,B. Nix, J. L. Davis, and B. Hjelle

Laboratory Survey of Drug-Resistant Streptococcus pneumoniae in New York City, 1993–1995 113

R. Heffernan, K. Henning, A. Labowitz, A. Hjelte, and M. Layton

B-virus from Pet Macaque Monkeys: An Emerging Threat in the United States? 117S.R. Ostrowski, M.J. Leslie, T. Parrott, S. Abelt, and P.E. Piercy

LettersInfectious Diseases and Mental Illness: Is There a Link? E. McSweegan 123Mycobacterium nonchromogenicum Bacteremia in an AIDS Patient, 124 J. Mayo, J. Collazos, and E. Martínez

Escherichia coli O157:H7 Infection in Colombia, S. Mattar and E. Vásquez 126Autofluorescence and the Detection of Cyclospora Oocysts, O.G.W. Berlin, 127 J.B. Peter, C. Gagne, C.N. Conteas, and L.R. Ash

Partnerships for Detecting Emerging Infectious Diseases: Nepal and Global Influenza Surveillance, J.M. Gambel, D.R. Shlim, L.C. Canas, N.J. Cox, 128 H.L. Regnery, R.M. Scott, D.W. Vaughn, C.H. Hoke Jr., and P.W. Kelley

HIV-2 Infection and HIV-1/HIV-2 Dual Reactivity in Patients With and Without AIDS-Related Symptoms in Gabon, C. Tevi-Benissan, M. Okome, 130 M. Makuwa, M.N. Nkoume, J. Lansoud-Soukate, A. Georges, M.-C. Georges-Courbot, and L. Belec

Q Fever in French Guiana: New Trends, F. Pfaff, A. François, D. Hommel, I. Jeanne, 131 J. Margery, G. Guillot, Y. Couratte-Arnaude, A. Hulin, and A. Talarmin

Ixodes dammini: A Junior Synonym for Ixodes scapularis, M. Sanders 132The Name Ixodes dammini Epidemiologically Justified, S.R. Telford III 132Ebola/Athens Revisited, P.E. Olson, A.S. Benenson, and E.N. Genovese 134

News and NotesThe Hot Zone–1997: Conference on Emerging Infectious Diseases 135International Meeting on Borreliosis, Prague, Czech Republic 140Workshop on Climate Change and Vector-Borne and Other Infectious Diseases 140The Fourth International Conference on HFRS and Hantaviruses 141Third International Congress on Tropical Neurology 141International Conference on Emerging Infectious Diseases 142

Above: Macaquegrabbing a snackfrom a passerby inthe downtown areaof Lop Buri, Thai-land. Photo by CyrilRuoso, reprintedwith permission ofBIOS.

1Vol. 4, No. 1, January–March 1998 Emerging Infectious Diseases

Update

In this new section, we welcome commentaryand updates from our newly formed Board ofInternational Editors. (See inside front cover fora list of names.)

Emerging Infectious Diseases—Brazil

Hooman MomenInstituto Oswaldo Cruz, FIOCRUZ, Rio deJaneiro, Brazil

Dr. Momen is a senior researchscientist; head of the Departmentof Biochemistry and MolecularBiology, Oswaldo Cruz Institute,FIOCRUZ; and Editor ofMemorias do Instituto OswaldoCruz. His research interests in-clude biochemical systematics ofpathogenic microorganisms. Dr.Momen is temporary adviser toboth WHO and PAHO at severaltechnical consultations and scien-tific working groups.

Brazil’s large size (more than 8.5 million km2

and 150 million population) and inadequatepublic health infrastructure pose a considerablechallenge in assessing the status of emerginginfectious diseases. Under the current system,most infectious diseases are not notifiable. Indiseases for which notification is required,underreporting is common and varies widely byregion and disease, and notification is oftendelayed, which causes the data to be revisedfrequently. Moreover, in hospital and clinicalsettings, the etiologic agent of an infectiousdisease is often not identified. For example, morethan a million hospital admissions are recordedper year by the public health system underparasitic and infectious diseases (the categoryexcludes AIDS and respiratory illnesses); of thesediseases, more than 70% are diagnosed as ill-defined intestinal infections and a further 10% asfood poisoning and septicemia, both withoutidentification of the etiologic agents. For thesereasons, the numbers reported here (1996 data,

unless otherwise stated) may not reflect the truenumbers of cases, and only diseases whoseprevalence has changed markedly in recent yearswill be included in the review.

Parasitic Diseases

MalariaAmong parasitic diseases, malaria causes the

most illness, approximately half a million casesannually, nearly all from the northern (Amazon)region. The disease has been controlled in theremainder of the country for years. Wheresystematic and sustained efforts have been madein the Amazon, control has also been successful.Due to the different epidemiologic situations inthe Amazon, no single strategy is effective.Worrying developments in this region include theestablishment of the disease in the periphery ofthe principal cities and the increase in drug-resistant parasites.

American TrypanosomiasisAmerican trypanosomiasis (Chagas disease)

is the most lethal parasitic disease in Brazil,causing more than 5,000 deaths per year. Severalmillion people are chronically infected carriers ofthis disease, for which there is still no effectivetreatment. A very effective control program againstthe principal insect vector, Triatoma infestans, andimproved control of the blood supply have reducedthe incidence of new cases to a very low level.Factors that continue to make the disease a re-emerging threat include secondary vectors, forestclearance, and congenital transmission.

LeishmaniasisLeishmaniasis (cutaneous and visceral) has

increased in incidence in recent years. Thisincrease is modified by apparent cyclic variationsof undetermined cause (cyclic duration for thevisceral form has been estimated at 10 years).The disease has increased not only in the usualareas of forest and recent clearance, but also inareas of traditional colonization as well as in theurban areas of the northeast. One factor involvedin the spread of this disease, uncontrolled urbangrowth due principally to rural migration, has ledto pockets (normally on the periphery of the

Address for correspondence: Hooman Momen, InstitutoOswaldo Cruz, FIOCRUZ, Av. Brazil 4365, Rio de Janeiro21045-900, Brazil; fax: 55-21-590-3545; e-mail:[email protected].

updateInternational Editors

2Emerging Infectious Diseases Vol. 4, No. 1, January–March 1998

Update

cities) of extreme poverty that favor thespread of many diseases.

Viral Diseases

DengueAmong the notifiable viral diseases, dengue

causes the most illness. Its mosquito vector,Aedes aegypti, was reintroduced into Brazil in the1970s; large outbreaks followed in the nextdecade. Since the vector’s reintroduction, morethan 700,000 cases have been reported (180,000cases in 1996). The presence of dengue-1 and -2indicates that cases of hemorrhagic dengue arealso occurring. Ae. aegypti is now present in allstates, and a plan has been formulated by thefederal ministry of health to eradicate the vector,although plan implementation has met withoperational difficulties. The widespread distribu-tion of Ae. aegypti, also the vector of yellow fever,in areas where this disease was consideredendemic poses a serious risk for the reemergenceof the epidemic form of yellow fever. Vaccinationand sanitation campaigns earlier in the centuryconsiderably reduced the incidence of the disease(fewer than 100 cases per year—the last urbancases were reported in 1942).

MeaslesMeasles—a viral disease that was controlled

and nearly eradicated—reemerged with avengeance in 1997. The outbreak began in thestate of São Paulo, where only 15 cases of measleswere confirmed in 1996, and has spread to otherstates. As of October 1997, 61,000 cases (48,000 inSão Paulo) have been reported, and 17,000(13,000 in São Paulo) have been confirmed. Mostcases have occurred in young adults under 30years of age. The vaccination campaign has beenreformulated in response to the outbreak, and anational campaign has been launched. GivenBrazil’s history of success in organizing suchcampaigns, it is likely that the outbreak will berapidly contained and that the effort for measleseradication will be resumed.

Other Viral DiseasesOther viral diseases include AIDS, with more

than 17,000 new cases registered in the last year.The number of deaths due to AIDS was 30% lowerin the first half of 1997 than in the first half of 1996.This decrease is attributed to the introduction ofnew antiviral drug combinations. The spread of

the AIDS epidemic, particularly into rural areas,has resulted in coinfections with differentinfectious agents, producing a variety of novelpathologies. Several subtypes of HIV, as well asnovel recombinations, have been reported.

Enteric transmission of hepatitis continues;hepatitis A and D comprise most of the cases,although the peak incidence is now in youngadults rather than in young children. Theincidence of hepatitis B is declining as a result ofvaccination, while that of hepatitis C isincreasing. A high prevalence of hepatitis Dexists in some regions of the Amazon, where inconjunction with hepatitis B, it is believed to bethe principal cause of Labrea black fever.

Influenza is a major concern. Importedvaccine is expensive and not generally available,and the proportion of elderly people (and thethreat for a large-scale epidemic) is increasing.

The vast forests that still cover large areas ofBrazil are home to many known and unknownviruses. The Evandro Chagas Institute in Belemalone has isolated more than 183 newarboviruses. The increasing exploitation ofsylvatic resources put humans in direct contactwith these viruses, but because of inadequate ornonexistent medical facilities (entire municipali-ties still lack a single medical doctor), fevers andeven deaths often are not diagnosed or aremisattributed. Among new viruses causing fatalinfections recently discovered in Brazil are Rociovirus, which causes encephalitis, and Sabiá virus,which causes hemorrhagic fever. Several cases ofhantavirus infection have also been reported, buton these and other occasions, the lack of abiosafety level 4 laboratory in the countryimpeded further work.

Bacterial DiseasesEnteric bacterial infections are an important

cause of illness in Brazil. National figures onprevalence are not available, except for cholera.However, an alarming increase has beenreported in antibiotic-resistant strains.

CholeraThe cholera epidemic in Brazil started with

the reemergence of the disease in Latin Americain 1991, reached a peak of more than 60,000confirmed cases and 670 deaths in 1993, and thendeclined to 1,000 cases in 1996. In 1997, thedisease resurged with more than 5,000 reportedcases and 2,600 confirmed cases. The reawakening


Update

of epidemiologic research in cholera caused by itsresurgence led to the discovery of a new biotype ofVibrio cholerae in the Amazon region. This biotypeis of the O1 serotype; it has distinct multilocusenzyme electrophoresis and RAPD profiles fromother pathogenic O1 V. cholerae. About 50isolates have been made from cases of diarrhea inthe upper Amazon (Solimoes) River. The microbeapparently lacks the principal known virulencefactors (e.g., the toxin gene cassette and themajor colonization factor, TCP); however, someisolates present a cytotoxic effect for Y-1 cells.

Mycotic DiseasesFungal infections, including histoplasmosis,

paracoccidioidomycosis, and cryptococcosis, occur;however, their cause is often unknown and even if adiagnosis is made, the diseases are not reported.

Public Health InfrastructureThe emerging infectious disease picture in

Brazil will not change markedly without asustained and determined effort to improve thecountry’s public health infrastructure. Theexisting, generally passive epidemiologic surveil-lance system produces information that arrivestoo late to be effective; however, a number of

measures, if implemented immediately, canmitigate the impact of any future epidemic: acontainment laboratory (biosafety level 4) thatcan handle known and unknown microbes of highvirulence; at least one infirmary, properly designedand fully equipped, to treat highly contagious andvirulent diseases. The current lack of this facilityposes a great danger to the population should anoutbreak of such a disease occur.

The financial, technical, and human re-sources needed to activate these measuresalready exist. In infectious diseases, Brazil has along tradition of fruitful international collabora-tion that can be tapped for additional support. Thesuccess of national vaccination and sanitationcampaigns in the eradication and control of someinfectious diseases at the beginning of this centuryand in more recent times demonstrates that muchcan be accomplished. Brazil can become asuccessful model for other developing countries.

AcknowledgmentsI thank the Fundação Nacional de Saude for

providing data on different diseases; Drs. H. Schatzmayr,J.R. Coura, and P. Mason for interesting discussions andcriticisms of the manuscript; and Drs. Ana Gaspar, ClaraYoshida, Ana Carolina Vicente, and Marilda Siqueira foruseful information.


Perspectives

Preventing the transmission of infectiousdiseases through blood transfusion in developingcountries is difficult given that the resourcesneeded are not always available, even whenpolicies and strategies are in place (1). Avoidingpaid donors, selecting blood donors throughquestionnaires, and limiting the number oftransfusions can prevent the transmission ofinfections. Testing for specific antibodies is thefinal measure for eliminating unsafe blood.

The risk for transfusion-transmitted infec-tious diseases can be estimated on the basis of

screening level for each infectious agent and theprevalence rate of the infection in the donorpopulation. Estimates may also take into accountthe sensitivity, specificity, and window period ofthe testing assays. We report here an estimate ofsuch a risk in 12 Central and South Americancountries and the cost of reagents required for thescreening of these infectious diseases as a proxyof resources needed to reduce the risk.

Source of InformationThis report analyzes data from 1993 on

screening of blood donors from five countries(Costa Rica, El Salvador, Guatemala, Honduras,and Nicaragua) of Central America. Data werealso analyzed for 1993 from five countries

Risk for Transfusion-TransmittedInfectious Diseases in

Central and South America

Gabriel A. Schmunis, Fabio Zicker,Francisco Pinheiro, and David Brandling-Bennett

Pan American Health Organization, Washington, D.C., USA

Address for correspondence: Gabriel A. Schmunis, PanAmerican Health Organization—PAHO, 525 23th StreetNW, Washington, D.C. 20037, USA; fax: 202-974-3688; e-mail: [email protected].

We report the potential risk for an infectious disease through tainted transfusion in10 countries of South and Central America in 1993 and in two countries of SouthAmerica in 1994, as well as the cost of reagents as partial estimation of screeningcosts. Of the 12 countries included in the study, nine screened all donors for HIV; threescreened all donors for hepatitis B virus (HBV); two screened all donors forTrypanosoma cruzi; none screened all donors for hepatitis C virus (HCV); and sixscreened some donors for syphilis. Estimates of the risk of acquiring HIV through bloodtransfusion were much lower than for acquiring HBV, HCV, or T. cruzi because ofsignificantly higher screening and lower prevalence rates for HIV. An index of infectiousdisease spread through blood transfusion was calculated for each country. The highestvalue was obtained for Bolivia (233 infections per 10,000 transfusions); in five othercountries, it was 68 to 103 infections per 10,000. The risks were lower in Honduras(nine per 10,000), Ecuador (16 per 10,000), and Paraguay (19 per 10,000). While thereal number of potentially infected units or infected persons is probably lower than ourestimates because of false positives and already infected recipients, the datareinforce the need for an information system to assess the level of screening forinfectious diseases in the blood supply. Since this information was collected, Chile,Colombia, Costa Rica, and Venezuela have made HCV screening mandatory;serologic testing for HCV has increased in those countries, as well as in El Salvadorand Honduras. T. cruzi screening is now mandatory in Colombia, and thepercentage of screened donors increased not only in Colombia, but also in Ecuador,El Salvador, and Paraguay. Laws to regulate blood transfusion practices have beenenacted in Bolivia, Guatemala, and Peru. However, donor screening still needs toimprove for one or more diseases in most countries.


Perspectives

(Bolivia, Chile, Colombia, Peru, and Venezu-ela) and for 1994 from two countries (Ecuadorand Paraguay) of South America. Informationwas obtained from Ministry of Health reportsduring technical meetings in which thesituation of each country was reviewed (2-5) orfrom an official report (6).

In addition, data are presented on the leastexpensive reagents for detecting antibodies forHIV, hepatitis C virus (HCV), Trypanosomacruzi, and Treponema pallidum, and for detectinghepatitis B virus (HBV) antigen (HBsAg) (2,3).All data are national except for Peru, where theinformation was for the city of Lima only (3).Population data were from the Pan AmericanHealth Organization’s publication Health Condi-tions in the Americas (7). Estimates are based onreported results of donor screening activities (2-6).

AssumptionsFor the best possible scenario, the following

assumptions were made: 1) Because thelaboratory procedures and brands of reagentsused in the 12 countries may differ in sensitivityand specificity, comparisons between them arenot straightforward. In addition, results of thescreening are influenced by the existence of anorganized system of quality control and profi-ciency testing for the serology and for theevaluation of the diagnostic kits, which mostcountries lacked from 1993 to 1994. Mostcountries reported the use of different brands ofsecond, third, and fourth generation immuno-logic assays for screenings of HCV, HIV, andHBV, respectively. Therefore, we assumed thatthe specificity of the tests for viral diagnosis was100%, but the sensitivity was 90.00% for HCV,99.99% for HIV, and 99.90% for HBV. Thesespecificity and sensitivity estimates fit well withthose reported for second, third, and fourthgenerations of assays for HCV, HIV, and HBV,respectively, as mentioned in the package insertby two of the manufacturers of reagents used inthe countries. Average window periods for thoseassays were 20 to 25 days (8,9), 82 to 84 days(9,10), and 51 days (9) for HIV, HBV, and HCV,respectively. In the case of T. cruzi serology, weselected the upper range of reported sensitivityand specificity (90% and 95%, respectively)(11,12). For T. cruzi, the probability that a personmay become a donor during the window period islow because infection is usually acquired inchildhood and in rural areas. 2) We assumed that

prevalence of infection in unscreened donors wasthe same as the national average prevalence foreach infectious disease. 3) Chile (6) and Peru (3)were the only countries that reported a fraction-ation index, 1.85 and 1.5, respectively. As no othercountry provided data on the fractionation index ordata allowing one to be calculated, to put thecountries in the same category, it was assumed thatevery blood donation corresponded to a singletransfusion to one recipient.

Screening Coverage and Prevalence RatesTable 1 shows coverage of screening and

prevalence rates of seropositive tests for specificinfectious agents among blood donors reported byeach of the 12 countries. For HIV, 100% of thedonors were screened in all countries, exceptBolivia (36.20%), Ecuador (89.50%), and Colom-bia (98.80%). Prevalence rates for HIV variedfrom 3.90 per 1,000 in Honduras to 0.04 per 1,000in Nicaragua. For HBV, only Costa Rica, Peru,and Venezuela screened 100% of donors. Thehighest values of HBV prevalence estimated were14.40 per 1,000 for Venezuela and 13.00 per 1,000for Paraguay. Bolivia, Costa Rica, and Paraguaydid not screen for HCV at all, and all othercountries screened fewer than 58% of donors;prevalence rates varied from 0.50 to 9.40 per1,000. Screening for syphilis was not complete inBolivia, Chile, Colombia, Ecuador, Nicaragua,and Paraguay; prevalence rates were 5.00 to28.00 per 1,000. For T. cruzi infection, onlyVenezuela and Honduras screened 100% ofdonors; prevalence rates were 2.00 per 1,000 inEcuador to 147.90 per 1,000 in Bolivia. In 1993,Peru and Costa Rica had not yet introducedscreening for T. cruzi.

Estimating Potential Infectivity of theBlood Supply

The probability of receiving an infectedtransfusion unit P(R) in each country wasestimated by multiplying the prevalence of aspecific infection by 1-level of screening (Table 1).For those estimates, the sensitivity andspecificity of the different tests were taken intoaccount. As the overall assumed sensitivity ofHIV screening was 99.99%, the adjustment ofprevalence rates makes no material difference tothe precision of the figures in Table 1. Theprobability of getting a transfusion-transmittedinfection P(I) was calculated as the result of theprobability of receiving an infected transfusion


Perspectives

P(R) multiplied by the infectivity risk. Forcountries reporting 100% of screening coveragefor a specific disease, a residual P(R) wasestimated as prevalence x 1-screening sensitivity(Table 2). Infectivity risk (defined as the likelihoodof being infected when receiving an infectedtransfusion unit) was assumed to be 90% for HIV(13), 75% for HBV (14), 90% forHCV (15), and 20% for T. cruzi(16) (Table 2). Estimates fortransfusion-acquired syphilisare not presented because theinfectivity risk depends onlength of refrigeration (17).

Considering the low preva-lence rates and the incomplete-ness of HIV screening, onlyBolivia, Colombia, and Ecua-dor could have missed detect-ing an HIV-infected transfu-sion unit; the probability ofgetting an infection in thesecountries was estimated at0.57, 0.22, and 0.95 per 10,000transfusions, respectively. ForHBV and HCV this risk ishigher. Up to 14.21 HBVinfections (Nicaragua) and 67.09HCV infections (Colombia) per10,000 transfusions may haveoccurred. The highest risk for

transfusion-transmitted infection was estimatedfor T. cruzi: 219.28 per 10,000 and 49.56 per10,000 for Bolivia and Peru, respectively, andapproximately 2 to 24 per 10,000 for the otherseven countries (Table 2).

Table 3 shows estimates of the absolutenumber of infections that may have been induced

Table 1. Coverage of screeninga of blood donors and seroprevalence rates (per 1,000) of infectious diseases, bycountryb

HIV HBVc HCV Syphilis T. cruziCountry Cov.d Prev.e Cov. Prev. Cov. Prev. Cov. Prev. Cov. Prev.

(%) (/103) (%) (/103) (%) (/103) (%) (/103) (%) (/103)Bolivia 36.2 0.10 14.5 2.00 0 ?f 37.9 18.10 29.4 147.90Chile 100 3.40 98.7 2.00 34.0 6.40 95.2 11.40 76.7 12.00Colombia 98.8 2.00 98.3 7.00 24.7 9.00 87.3 13.00 1.4 12.00Costa Rica 100 0.34 100 4.50 0 ? 100 5.00 0 ?Ecuador 89.5 1.00 88.2 3.80 32.9 1.40 86.7 11.50 51.0 2.00El Salvador 100 1.30 96.0 8.00 31.4 2.50 100 19.00 42.5 14.70Guatemala 100 3.00 79.8 7.00 37.2 8.00 100 19.00 75.0 14.00Honduras 100 3.90 83.5 2.70 27.8 0.50 100 7.00 100 12.40Nicaragua 100 0.04 53.1 4.00 53.1 4.40 88.4 16.00 58.4 2.40Paraguay 100 0.70 92.9 13.00 0 ? 66.9 28.00 86.8 45.00Peru 100 2.80 100 8.60 57.4 4.40 100 9.60 0 23.60g

Venezuela 100 2.10 100 14.40 31.0 9.40 100 10.70 100 13.20aCoverage of screening = (number of screened donors ÷ total number of donors) x 100.

bData as reported by the countries from 1993, except for Ecuador and Paraguay, which were for 1994.

cHBsAg only.

dCoverage.

ePrevalence.

fScreening not performed and prevalence not known.

gData from a survey of 2,237 samples.

Table 2. Probability of receiving an infected transfusion P(R)a and probabilityof getting a transfusion-transmitted infection P(I)b, by countryc

HIV HBV HCV T. cruzi(x104) (x104) (x104) (x104)

Country P(R) P(I) P(R) P(I) P(R) P(I) P(R) P(I)Bolivia 0.64 0.57 17.27 12.95 NSPd NSP 1096.38 219.28Chile 0.00 0.00 0.26 0.20 46.46 41.82 29.36 5.87Colombia 0.24 0.22 1.20 0.90 74.55 67.09 124.24 24.85Costa Rica 0.00 0.00 0.45x 0.34 NSP NSP NSP NSPEcuador 1.05 0.95 4.52 3.39 10.33 9.38 10.29 2.06El Salvador 0.00 0.00 3.23 2.42 18.87 16.97 88.75 17.75Guatemala 0.00 0.00 14.28 10.71 55.26 49.74 36.75 7.35Honduras 0.00 0.00 4.49 3.37 3.97 3.57 13.02x 2.60Nicaragua 0.00 0.00 18.95 14.21 22.70 20.43 10.48 2.10Paraguay 0.00 0.00 9.32 6.99 NSP NSP 62.37 12.47Peru 0.00 0.00 0.87x 0.65 20.62 18.56 247.80 49.56Venezuela 0.00 0.00 1.45x 1.09 71.35 64.21 13.86x 2.77aP(R) = probability of receiving an infected transfusion = prevalence of infection x1- level of screening; xfor countries in which reported screening level was 100%, aresidual P(R) was estimated as prevalence x 1- screening sensitivity rate x 10,000.bP(I) = probability of getting a transfusion-transmitted infection = P(R) xinfectivity index (infectivity indexes used were HIV=90%; HBV=75%; HCV=90%;T.cruzi=20%). For calculations of P(R) and P(I) the prevalence was correctedtaking into account the sensitivity of the screening.cData from 1993, except for Ecuador and Paraguay, which were for 1994.dNo Screening performed, so P(R) and P(I) not known.


Perspectives

by transfusion, calculated as [no. of donors x P(I)],for each country. Because Chile (6) and Peru (3)reported fractionation of blood units by 1.85 and 1.5,respectively, the estimated number of infectedunits transfused in those countries was multipliedby these factors. For the remaining countries, it wasassumed that each donated unit was given to onlyone recipient. An index of infectious disease spreadthrough blood transfusion was calculated bydividing the estimated total number of transfusion-related infections (for any one of the infectiousagents considered) by the total number of donors.This index indicates the health risks associatedwith blood transfusion and can be used as anoutcome indicator to assess the cost-effectiveness ofscreening programs.

The highest value for the infection spreadingindex was obtained for Bolivia, where 233transfusion-related infections may have occurredper 10,000 donations. This was a result of a veryhigh prevalence rate of antibodies to T. cruzi anda lower level of screening. For most othercountries considered, the index was 68 to 103infections per 10,000 donations. Due to lowseroprevalence rates and good screening levels insome cases, the risk for transfusion-relatedinfections was relatively low in Honduras (nineper 10,000), Ecuador (16 per 10,000), andParaguay (19 per 10,000).

Table 3 also shows the ratio of number ofinfections per donation by country. One infection(HIV, HBV, HCV, or T. cruzi) might have beentransmitted in every 43 (Bolivia) to 1,072(Honduras) donated units.

Screening CostsThe unitary cost for serologic screening,

estimated solely from expenditures on the leastexpensive laboratory reagents in each countryand considering the prevalence rates reported bythe countries was US$0.9 to US$2.4 for an HIVenzyme-linked immunosorbent assay (ELISA),US$0.5 to US$3.5 for HBV screening (enzymeimmunoassay, radioimmunoassay, or passivereverse hemagglutination), US$3.5 to US$10.0for HCV ELISA, US$0.25 to US$1.0 for a T. cruzitest (ELISA, radioimmunoassay, or indirecthemagglutination) (Table 4), and US$0.09 toUS$0.60 for syphilis serology (RPR or VDRL).Using other tests might have increased the costssignificantly for some of the infections. Forexample, the rapid agglutination test for HIV isusually more expensive than ELISA.

The cost of preventing the transfusion of oneinfected unit was estimated as [(no. of donors xcost of each test)/ total number of positive donors]as reported by each country. This valuerepresents the cost of detecting one unit positive

for any one of the infectionsstudied in each country byusing one diagnostic test foreach infectious disease. Forexample, using two tests, onefor antibody detection and onefor antigen detection of HIV,increases costs. Detection of T.cruzi was the least expensive(US$11-$209 per positive unit),followed by HBV (US$90-USS$599 per unit), HCV(US$438-$7,136 per unit), andHIV (US$232-$23,000 per unit)(Table 4). The wide variation ofcost primarily reflects differ-ences in the prevalence of eachinfection and in the cost of eachtest in the countries.

The costs per capita tocarry out a complete screeningof blood donors in each countrywas US$0.008 to US$0.04 forHIV, US$0.008 to US$0.02 for

Table 3. Estimates of transfusion-transmitted infectious diseases, by countrya

InfectionAbsolute no. of transfusion- spreading Ratio of

No. of transmitted infectious diseasesb indexc infections:Country donors HIV HBV HCV T. cruzi Total /104 donationsBolivia 37,948 2 49 NAe 832 883 233 1:43Chiled 217,312 0 8 1681 236 1925 88 1:113Colombia 352,316 8 32 2364 875 3279 93 1:107Costa Rica 50,692 0 2f NA NA NA NA NAEcuador 98,473 9 33 92 20 154 16 1:639El Salvador 48,048 0 12 82 85 179 37 1:268Guatemala 45,426 0 49 226 33 308 68 1:147Honduras 27,885 0 9 10 7f 26 9 1:1072Nicaragua 46,001 0 65 94 10 169 37 1:272Paraguay 32,893 0 23 NA 41 64 19 1:514Perug 52,909 0 4f 147 393 544 103 1:97Venezuela 204,316 0 22f 1312 57f 1391 68 1:147aData from 1993 except for Ecuador and Paraguay, which were for 1994.bNumber of cases transmitted by blood transfusion = [number of donors x P(I) ]. Forcalculations of number of infections, the prevalence was corrected taking into accountthe sensitivity of the screening.cInfection spreading index = (total number of infections transmitted ÷ number of donors)x 10,000.dFractionation index = 1.85.eData not available.fResidual infectivity considering that sensitivity of diagnostic tests is not 100%.gFractionation index = 1.5.


Perspectives

HBV, US$0.01 to US$0.08 for HCV, US$0.0008 toUS$0.003 for syphilis, and US$0.0025 toUS$0.009 for T. cruzi.

Condition of the Blood SupplyThese estimates indicate that the condition of

the blood supply in Central and South America isfar from ideal. Roughly, one case of transfusion-related infection occurs every 43 to 1,072donations, varying with the infectious agent andthe country.

In three of the 12 countries, transfusionrecipients might become infected with HIV; innine countries, with HCV; in all countries, withHBV; and in ten countries, with T. cruzi.However, it was not possible to establish thepotential number of tainted units/infections fromcountries in which there was no information ondonor prevalence for HCV (e.g., Bolivia, CostaRica, and Paraguay).

No serologic tests for T. cruzi were done inCosta Rica and Peru. Data on the prevalence of T.cruzi serology in blood donors from Costa Ricafrom 1983 to 1985 (18,19) suggest a risk. Datafrom a recent report of a survey among donors in

Lima indicate a prevalence of 2.36% (3). If this isthe real prevalence in the city, the number oftainted units transfused would have been 1,872in 1993, while the number of persons infectedthrough blood transfusion could have been 393.

On the other hand, considering the number ofdonors and the prevalence of the infection in the12 countries, if blood had not been screened at all,more than 35,000 infected units would have beentransfused. However, infections have differentpatterns of evolution. HIV-infected persons areexpected to get AIDS at some time during theirlives (20), while only 50% and 38% of persons,respectively, will get posttransfusion hepatitisafter infection with HBV or HCV (21). On theother hand, 20% to 30% of those infected withT. cruzi will get Chagas disease (16,18,19).

Limitations of the DataDifficulties and limitations of the use of

public health data for policy decisions, even inindustralized countries, are well recognized.Figures presented here were generated toestablish an approximation of the problem byproviding an overview of the risk of receiving

Table 4. Estimated unitary cost of preventing transfusion-transmitted infections, by country, 1993a

Cost (US$)HIV HBVb HCV T. cruzi

Preventing Preventing Preventing PreventingSingle one infected Single one infected Single one infected Single one infected

Country test unit test unit test unit test unitChile 2.3 676 1.2c 599 3.5d 547 NAe NA

Costa Rica 1.1 3,280 0.5c 111 NA NA NA NAEcuadorf 1.7 1,708 1.0c 263 10.0 7,136 0.35g 175El Salvador 2.0 1,550 1.9c 238 4.5 1,802 1.0g,h 68

1.0Guatemala 1.8 601 1.7c 243 3.5 438 0.9g 65Honduras 0.9 232 0.9c 334 3.5 6,971 0.45h 36

0.5h 186Nicaragua 1.0 23,000 0.5h 125 3.5 797 0.5h 209Peru (Lima) 2.4 858 3.5c 407 8.2 1,862 0.25g 11

2.4i 279Venezuela 1.3 619 1.3c 90 4.5 479 0.5 38

0.3g 23aCost of preventing (=detecting) one infected unit was calculated as [(number of donors x test cost) ÷ (total number of positivedonors detected)]. All costs refer to enzyme-linked immunosorbent assay, unless otherwise indicated.bHBsAg onlycEnzyme immunoassay.dEstimated cost based on cost of test in other countries.eData not available.fDonors and prevalence for 1994, costs for 1993.gIndirect hemagglutination.hRadioimmunoassay.iPassive reverse hemagglutination.


Perspectives

tainted blood in different countries in Centraland South America. One potential cause ofunderestimation of viral infections transmittedby blood is the residual risk because of thewindow period, even when 100% of donors arescreened by serology (8,9). However, this residualrisk would be difficult to ascertain in mostcountries of Central and South America.Investigation of clinically identified cases after atransfusion, follow-up of recipients forseroconversion, and special laboratory studiesdetecting seronegative donors for missed infec-tions are laborious and expensive and couldseldom be undertaken in those countries.Another possibility would be studies thatcombine estimates of incidence rates of infectionamong repeated and first-time donors (whoseroconvert) with estimates for the duration ofthe preseroconversion period for a specificinfectious agent (22). Excellent results wereobtained by this method in the United States,where more than 80% of donations come fromrepeat donors. Those studies involved hundredsof thousands of donors and millions of donations.To have incidence data on repeat donors, it isnecessary to have a significant number ofvoluntary donors who will repeat donations.Therefore, it is unlikely that studies of that sortcould be carried out in the countries mentionedhere. First, the population of the countries ismuch smaller; therefore, the number of donationsis smaller. For example, in all Central Americathe number of donations is approximately210,000 per year. Second, the number of repeatdonations from voluntary donors is small.Voluntary donors accounted for 30% and 40% ofall donations in Colombia and Costa Rica and 4%to 10% of all donors in Chile, Bolivia, Peru, andVenezuela (2-6). The number of voluntary donorswas also small in the remaining countries. In allcountries of Central and South America, mostdonations come from directed donors, relatives orfriends of patients. In addition, there is nonational registry of donors to allow for follow-up.Using incidence rates for first-time donorsinstead of repeat donors is not a solution becauseofficial incidence rates for HIV or other viruseswere not available at the time of this study.

The risk for transfusion-related infectioncould also be overestimated. Recipients mayalready be infected. This is especially likely for T.cruzi infection in Bolivia, where theseroprevalence in the general population could be

higher than 20% (18,19). Another source ofoverestimation is that only some of the casesdetected by screening would be confirmed. Inseveral countries, a confirmatory test ismandatory for HIV, syphilis, and HCV. However,as the primary function of blood banks is donorscreening, seropositive donors for any of thediseases mentioned here are usually referred tospecialized services or reference laboratories forconfirmation of the results of the screening, and ifresults are confirmed, for treatment andcounseling. Results of this confirmatory serologyare not often sent back to the blood bank, evenwhen privacy concerns allow for it. Chile was theonly country that reported results of confirma-tory tests for HIV: results indicated that only 9%of those found positive by the screening wereconfirmed positive (6). With T. cruzi, as there isno confirmatory test, it is assumed that a truepositive is a unit that is positive on more than onetest. By these criteria, a recent study in Brazilsuggested that only one out of five donors positivefor T. cruzi could be considered a true positive(23). Those facts, however, do not reduce thepublic health relevance of the problem presentedhere, although the real numbers of potentiallyinfected units/infected persons may be still lowerthan our estimates.

Establishment of a screening process in everycountry will depend on balancing the benefits andcosts. Although costs for preventing transfusionof one tainted unit or preventing one infectionseem high for some etiologic agents, they are notso. Even in the case of Nicaragua, the countrywith the lowest HIV prevalence, the cost toprevent the transfusion of one potentially HIV-infected unit (by testing all donors with ELISA)was estimated at US$23,000, while treatmentcosts (drugs only) for an AIDS patient would beapproximately US$12,000 per year.

In general, the risk for an infectious diseasethrough tainted transfusion is not as high as thatreported from some countries of Africa (24). Since1993, donor screening has improved in severalcountries. Chile, Colombia, Costa Rica, andVenezuela, for example, have made screening forHCV mandatory, and coverage for serology forthat infection has increased in those countries, aswell as in El Salvador and Honduras. T. cruziscreening is now mandatory in Colombia, and thepercentage of screened donors not only increasedin Colombia but also in Ecuador, El Salvador, andParaguay. Laws to regulate blood transfusion


Perspectives

practices have been enacted in Bolivia, Guatemala,and Peru. The figures presented, however,underline the need for improvement and stress theimportance of an information system that allowsassessing the level of screening for infectiousdiseases in the blood supply. Universal screening ofdonors for HCV is still a priority in most countries,and increased donor screening for T. cruzi is apriority for Bolivia and possibly for Peru.

Continuous collection of the type of informa-tion shown here, which has only been partiallyavailable (1), provides a baseline against whichfuture achievements can be measured and isessential for obtaining the support needed tomaintain or expand the screening of blood donors.

References 1. Linares J, Vinelli E, editors. Taller Latinoamericano de

servicios de transfusión sanguínea y óptimo uso de losrecursos. Cruz Roja Finlandesa; 1994. p. 167.

2. Organización Panamericana de la Salud. Taller para elcontrol de calidad de sangre en transfusiones: serologíapara la detección de Chagas, hepatitis B y C, sífilis yHIV/SIDA. Document OPS/HPC/HCT/94.42.

3. Organización Panamericana de la Salud. Paísesandinos. Taller sobre control de calidad de sangre entransfusiones: serología para la detección de hepatitisB y C, sífilis, tripanosomiasis americana y VIH/SIDA.Document OPS/HCP/HCT/95-61.

4. Organización Panamericana de la Salud. Simposiointernacional sobre control de calidad en bancos desangre del Cono Sur y de Brasil. Informe final OPS/HCP/HCT/95.55.

5. Organización Panamericana de la Salud. Taller sobrecontrol de calidad en serología de bancos de sangre.OPS/HCP/HCT/96/79.

6. Ministerio de Salud, Chile. Diagnóstico de la situaciónde los bancos de sangre y medicina transfusionál enChile 1993. Santiago, Chile: La Ministerio; 1995. SerInf Téc No.14.

7. Pan American Health Organization. Health conditionsin the Americas. Washington (DC): The Organization;1994. PAHO Sci Pub No. 549.

8. Lackritz EM, Satten GA, Aberle-Grasse J, Dodd RY,Raimondi VP, Janssen RS, et al. Estimated risk oftransmission of the human immunodeficiency virus byscreened blood in the United States. N Eng J Med1995;333:1721-5.

9. Schreiber GB, Busch MP, Kleinman SH, Korelitz JJ.The risk of transfusion-transmitted viral infections. NEng J Med 1996;334:1685-90.

10. Van de Poel CL, Cuypers HT, Reesink HW. Hepatitis Cvirus six years on. Lancet 1994;344:1475-9.

11. Andrade ANS, Martelli CMT, Luquetti AO, Oliveira OS,Almeida e Silva S, Zicker F. Triagem sorologica ra oTrypanosoma cruzi entre doadores de sangue do Brasilcentral. Bol Oficina Sanit Panam 1992;113:19-27.

12. Lorca MH, Child RB, Garcia AC, Silva MG, Osorio JS,Atias M. Evaluación de reactivos comercialesempleados en el diagnóstico de la enfermedad deChagas en bancos de sangre de Chile. I Selección dereactivos. Rev Med Chile. 1992;120:420-6.

13. Donegan E, Stuart M, Niland JC, Sacks HS, Azen SP,Dietrich SL, et al. Infection with humanimmunodeficiency virus type 1 (HIV-1) amongrecipients of antibody-positive blood donations. AnnIntern Med 1990;113:733-9.

14. Gocke DJ. A prospective study of post-transfusionalhepatitis. JAMA 1972;219:1165-70.

15. Aach RD, Stevens CE, Hollinger FB, Mosley JW,Peterson DA, Taylor PE, et al. Hepatitis C virusinfection in post-transfusion hepatitis. An analysiswith first- and second-generation assays. N Engl J Med1991;325:1325-9.

16. World Health Organization. The control of Chagasdisease. Geneva: WHO Tech Rep Ser No.811:32;990.

17. National Institutes of Health. Infectious diseasetesting for blood transfusion. NIH Consens Statement1995;13:13-4.

18. Schmunis GA. Trypanosoma cruzi, the etiologic agentof Chagas disease: status in the blood supply inendemic and non endemic countries. Transfusion1991;31:547-55.

19. Schmunis GA. American trypanosomiasis as a publichealth problem. Chagas disease and the nervoussystem. PAHO Sci Pub 994;547:3-29.

20. Pedersen C, Lindhardt BO, Jensen BL, Lavritzen E,Gerstoft J, Dickmeiss E, et al. Clinical course ofprimary HIV infection: consequences for subsequentcourse of infection. BMJ 1989;299:154-7.

21. Alter M. Residual risk of transfusion associatedhepatitis. In: Program and Abstracts of the NationalInstitutes of Health Development Conference onInfectious Diseases Testing for Blood Transfusions.1995 Jan 9-11; Bethesda (MD): National Institutes ofHealth;1995. p. 23-7.

22. Busch MP. Incidence of infectious disease markers inblood donors, implications for residual risk of viraltransmission by transfusion. In: Program andAbstracts of the National Institutes of HealthDevelopment Conference; 1995. Jan 9-11; Bethesda(MD): National Institutes of Health; 1995. P. 29-30.

23. Hamerschlak N, Pasternak J, Amato Neto V, CarvalhoMB, Guerra CS, Coscina AL, et al. Chagas’ disease, analgorithm for donor screening and positive donorcounseling. Rev Soc Bras Med Trop 1997;30:205-9.

24. McFarland W, Mvere D, Shandera W, Reingold A.Epidemiology and prevention of transfusion-associatedhuman immunodeficiency virus transmission in sub-Saharan Africa. Vox Sang 1997;72:85-92.


Perspectives

In vitro cultivation of caliciviruses indicatesthat these pathogens have been emergingperiodically from ocean sources for 65 years (1).The best-documented example of oceancaliciviruses causing disease in terrestrialspecies is the animal disease vesicular exan-thema of swine (VES) (1). Feline calicivirus (theonly member of the group with a seeminglyubiquitous and continuous terrestrial presence)also appears to have ocean reservoirs (2). Thesource of caliciviruses causing gastroenteritis inhumans is frequently shellfish, which do notalways come from beds contaminated withhuman waste (3,4). The origins of hepatitis E areoften obscure, but water is one suspected source(5). The most recent emerging calicivirus isassociated with rabbit hemorrhagic disease(RHD), and although an ocean association has notbeen reported, the agent readily moves betweencontinents and crosses ocean channels (6).

Finally, the only reported in vitro isolation andsequential propagation of a calicivirus patho-genic for humans is a virus residing in the sea (7).

The Caliciviridae are divided into fivegroups, tentatively designated distinct genera,on the basis of sequence relatedness and genomicorganization (8). Four are known humanpathogens—Sapporo, Norwalk-like small roundstructured viruses, hepatitis E, and the marine(animal) caliciviruses—while the fifth group,which includes RHD virus, is not yet proven to bea human pathogen. The human Sapporo virusesare more closely related to the marinecaliciviruses than to the other human groupcausing gastroenteritis, the Norwalk-like vi-ruses. On the basis of homology and genomicorganization, RHD virus falls between these twogroups. In addition, the genomic organization ofhepatitis E is most closely related to that of theonly other hepatotropic calicivirus currentlydescribed, RHD virus (8).

Many marine calicivirus strains in thetentative genus of VES virus-like caliciviruseshave been passaged in vitro; their characterization

Calicivirus Emergence from OceanReservoirs: Zoonotic andInterspecies Movements

Alvin W. Smith,* Douglas E. Skilling,* Neil Cherry,†Jay H. Mead,‡ and David O. Matson§

*Oregon State University, Corvallis, Oregon, USA; †Lincoln University,New Zealand; ‡Red Cross, Portland, Oregon, USA; and §Children’s

Hospital of the King’s Daughters, Eastern Virginia MedicalSchool, Norfolk, Virginia, USA

Address for correspondence: Alvin W. Smith, College ofVeterinary Medicine, Laboratory for Calicivirus Studies,Oregon State University, Corvallis, OR 97331, USA; fax: 541-737-2730; e-mail: [email protected].

Caliciviral infections in humans, among the most common causes of viral-inducedvomiting and diarrhea, are caused by the Norwalk group of small round structuredviruses, the Sapporo caliciviruses, and the hepatitis E agent. Human caliciviruses havebeen resistant to in vitro cultivation, and direct study of their origins and reservoirsoutside infected humans or water and foods (such as shellfish contaminated with humansewage) has been difficult. Modes of transmission, other than direct fecal-oral routes,are not well understood. In contrast, animal viruses found in ocean reservoirs, whichmake up a second calicivirus group, can be cultivated in vitro. These viruses can emergeand infect terrestrial hosts, including humans. This article reviews the history of animalcaliciviruses, their eventual recognition as zoonotic agents, and their potentialusefulness as a predictive model for noncultivatable human and other animalcaliciviruses (e.g., those seen in association with rabbit hemorrhagic disease).


Perspectives

has facilitated understanding of calicivirusgeographic distribution and host versatility (1).Dozens of serotypes were described on the basis ofserum neutralization tests; this antigeniccomplexity complicated serodiagnosis and ham-pered studies of effects on host species.

The illnesses associated with two recentlydiscovered viruses classified as Caliciviridae,hepatitis E virus and RHD virus, have altered thenotion that caliciviruses produce only transientclinical disease but not death (1,9,10). Hepatitis Evirus is fatal for 25% of the pregnant women indeveloping countries who contract hepatitis E(11); RHD can kill 95% of infected rabbits within24 to 48 hours of exposure (6).

The oceans are reservoirs in whichcaliciviruses are exposed in a water substrate tolife forms from zooplankton to whales and inwhich they, like other RNA viruses, can amplifyto very high numbers with variants occurring inevery replicative cycle (12). Such a variedreplicative setting has served this parasite well.With the right tools, evidence of previousinfection with caliciviruses can often be shown infish, avian, and many mammalian species,including humans (1). It is not known why somecaliciviruses have become potential hemorrhagicagents associated with purpura hemorrhagica inaborted piglets (13), neonatal hemorrhagicsyndrome in pinnipeds (A.W. Smith, D.E.Skilling, unpub. data), hemorrhagic disease infatal hepatitis E in humans (5), and RHD (6).However, it is known that caliciviral diseases canbe difficult or impossible to contain and eradicate.Pathogenic caliciviruses can be expected tocontinue emerging from the sea in unexpectedforms at unexpected times in unexpected places.Studying those that have emerged and arecompliant to in vitro propagation can provideinsights into those that cannot be cell-cultureadapted and those yet to be discovered.

History and Its LessonsThe 66-year history of the caliciviruses with

ocean reservoirs can be divided into threeperiods: 1932 to 1972, the species-specific era(14); 1972 to 1982, the new era of virology, duringwhich oceans were first found to be reservoirs ofviral disease infecting domestic animals (7,9);and 1976 to the present.

The first evidence of infection withcaliciviruses of marine origin can be traced to1932. A large herd of swine in Orange County,

California, was being fed raw garbage collectedfrom restaurants and institutions in the LosAngeles area. When some animals became sickwith vesicular lesions on the feet and nose,regulatory veterinarians were notified becausevesicular diseases of livestock were reportable.The farm with sick swine and adjacent farmswere quarantined for foot-and-mouth disease,and more than 19,000 head of exposed cattle andswine were destroyed and buried in quicklime(14). The outbreak was contained. One year laterand 100 miles to the south in San Diego,California, the second known outbreak occurredand was contained (14). This time the disease wasfound not to be foot-and-mouth disease, becausethe virus would not infect cattle, but instead wasdescribed as a new disease of swine and wascalled VES (14). In 1934, a third outbreakoccurred in San Francisco, California, and VESwas again contained (14). In 1935, the eventsrepeated themselves, but from 1936 through mid-December 1939, the disease disappeared andthen abruptly reappeared, at times involving 40%of California swine herds. All of these outbreaksin the 1930s and 1940s were shown by cross-infectivity studies to be caused by many distinctbut related VES virus strains.

The embargoes placed on raw California porkwere successful in containing VES withinCalifornia until 1952. That year, a passengertrain between San Francisco and Chicago servedCalifornia pork and discarded the raw porktrimmings into the garbage in Cheyenne,Wyoming; the garbage was fed to swinesubsequently redistributed by auction sale yard.Within 14 months, all major swine-growing areasin the United States (41 states) had reportedVES. For the first time, the federal government,rather than just the California state government,activated eradication and quarantine measuresagainst VES, including enforcement of federallaws requiring garbage to be cooked before it wasfed to swine. By 1956, the last reportedoutbreak of VES had been contained, and thedisease was said to have been eradicated. In1959, VES was declared a foreign animaldisease, even though it had never been reportedoutside the United States (14).

Forty years later, the natural history of thiscalicivirus still contained few details. Its originswere not known but were said to have been denovo or from some unknown wild animalreservoir, which was extensively sought but not


Perspectives

found (14,15). Swine were the only naturallyinfected host species; no evidence of humaninfection had been observed (14). Control of VEShad been a notable success story for regulatoryveterinary medicine in the United States; within24 years of its discovery as an entirely newdisease, it was said to have been eradicated (14).Internationally accepted animal disease diag-nostic tests using swine, horse, and bovineinfectivity profiles for vesicular stomatitis virus,foot-and-mouth virus, and VES virus were usedroutinely; VES virus did not infect cattle,whereas the other viruses did (14,15).

The discovery of a new paradigm in “viraltraffic” (16) began the second period of calicivirushistory. The movement of a member of theCaliciviridae from ocean reservoirs to terrestrialhosts changed the understanding of the naturalhistory of a virus thought to be host specific anderadicated (1,15,17).

The first virus isolate from a pinnipedoccurred in 1972. The agent, named San Miguelsea lion virus type 1 (SMSV-1), was a calicivirusthat caused classic VES in swine (17). Thus begana series of isolating and characterizing viruses inocean species that were officially designated as“viruses indistinguishable from VES virus”because they were additional VES virus types.They could not be called VES virus (18) since VEShad been officially eradicated. Should the VESvirus reappear, its status as a foreign animaldisease would mandate immediate implementa-tion of eradication measures; eradication wasviewed as an impossibility because of the widerange of reservoirs for VES virus, both migratoryand ocean species. The 13 marine calicivirusesserotypes (100 TCID50 vs. 20 units of neutralizingantibody) isolated from swine before 1956 werecalled VES viruses; those isolated after 1972 havebeen designated SMSV-1 through 17 or givenmore proper nomenclature, e.g., the pygmychimpanzee isolate (primate calicivirus Panpaniscus type 1) (1,19).

By 1982, 11 species of pinnipeds andcetaceans of the North Pacific Ocean and BeringSea (monk seals, California sea lions, northernsea lions, northern elephant seals, northern furseals, walrus, gray whales, sei whales, spermwhales, bowhead whales, and Pacific bottlenoseddolphins) were known to be susceptible tocalicivirus infection, as was an ocean fish, theopaleye perch (Girella nigricans) (20). Further-more, in many instances, the virus had crossed

the intertidal zone to infect terrestrial species(18). On the basis of these data and theestablished ocean ranges of known calicivirushost species, the shores of Mexico, the UnitedStates, Canada, Russia, Korea, Japan, China,and perhaps others bordering the North PacificOcean had been regularly exposed to largenumbers of marine caliciviruses with unknownhost ranges and tissue trophisms (21). By thistime, type-specific neutralizing antibodies to twoof four serotypes tested were reported in humanpatients in the United States (22). Cumulatively,these findings lead to the conclusion that fish andperhaps other ocean products provide a vehiclefor transmission of these marine caliciviruses toterrestrial animals.

The magnitude of potential exposure tomarine caliciviruses from the sea is substantial.For example, a 35-ton gray whale, shown byelectron microscopy to have more than 106

caliciviruses per gram of feces, can eat 5% or moreof its body weight per day and eliminate anequivalent quantity of feces containing anestimated 1013 caliciviruses daily. Marinecaliciviruses remain viable more than 14 days in15°C seawater (20).

Although marine mammals were ofteninfected, fish and fish products were more likelyto transport the virus from sea to land (23). Incontrast to the 1932 to 1936 introductions of VESfrom raw fish, the rapid and uncontrolled spreadof VES virus throughout California after 1939and then across the United States in 1952 wasa pig-to-pig cycle through raw garbage feed.However, new virus serotypes were alsointroduced through feeding raw fish scraps toswine (1,15). That VES was not a species-specific disease became accepted, but thepossibility of human infection, althoughsuspected (22), was largely untested.

During the third historical period (1976 topresent), which overlaps with the second, viraltraffic across the land/sea interface has beenobserved repeatedly, as the following examplesshow. A calicivirus isolated from an opaleye perchand designated SMSV-7 produced fulminatingVES in exposed swine and spread from pig to pigby contact transmission (23). A reptiliancalicivirus Crotalus-1 was isolated from threespecies of snakes and one species of amphibian(24) and from three species of marine mammalswhose population distributions spanned theNorth Pacific from Mexico to the Bering Sea (1).


Perspectives

Mink fed a diet of ground-up calicivirus-infectedseal meat became infected with VES virus (25).Parasitic nematode larvae from California sealions in San Diego, California, were used to infectopaleye fish with calicivirus (SMSV-5); when thefish were killed and fed to Northern fur seals onthe Pribilof Islands in the Bering Sea 30 dayslater, the seals developed vesicular disease, andthe virus was recovered from the lesions (1).Shellfish resident to the tidal zone were exposedto marine caliciviruses and held at less than 10°Cin a continuous flow of sterile seawater. Thecaliciviruses were reisolated 60 days later inmammalian cell lines (26). When tested with acDNA calicivirus group-specific hybridizationprobe from a marine calicivirus (SMSV-5), someshellfish beds on U.S. coasts were positive forcaliciviruses of unknown type (26). Felinecalicivirus was shown to cause disease not only indogs, but also in seals (on the basis of 17 of 20adult sea lions having neutralizing antibody toFCV-F9 with titers of 1:15 to 1:220). Only 11 of 20of these sea lions had neutralizing antibody to asea lion isolate, SMSV-13 (titer 1:10). Thisdemonstrates a probable feline calicivirus oceanpresence in California sea lions (2). In swine, a so-called mystery pig disease (porcine respiratoryreproductive syndrome) was reproduced inpregnant sows in 1992 with a three-plaquepassage purified cytolytic calicivirus isolate fromstillborn piglets with mystery pig disease (13). Asecond calicivirus serotype isolated from thesame piglets was the same as that isolated fromwalruses in 1976 (1). A white tern (Gygis albarothschildi), a migratory sea bird sampled in themid-Pacific (French Frigate Shoals), had ablistering disease caused by a calicivirus (27). Inthe first fully documented human case of clinicaldisease caused by a marine calicivirus, SMSV-5was isolated from blisters on the hands and feet ofa patient (7). A second, less well-documented,case involved a field biologist who was handlingsea lions and developed severe facial blistering.An untypable calicivirus was isolated in tissueculture (Vero cells) from throat washings (7).

Extent of ExposureThe extent of human disease is not known

because test reagents are not readily availableand diagnosticians are not alerted to caliciviralcauses of human disease, except for diarrhea andoccasionally hepatitis. However, evidence ofhuman exposure was shown when 150 serum

specimens from normal blood destined for donoruse were tested. The samples were antibody-negative for hepatitis B surface and core antigen,HIV-1 and -2, HIV P-24 antigen, human T-celllymphotropic virus Type 2, and hepatitis C virus.Approximately 19% had antibodies reactive to apolyvalent antigen made up of equal quantities ofcesium chloride-banded SMSV-5, 13, and 17.(Figure 1A). To demonstrate that these reactionswere not cross-reactions to the human Norwalkcalicivirus antibody, serum samples from eightpersons with Norwalk virus-induced diarrheawere also tested. Both acute- and convalescent-phase serum specimens were tested by enzyme-linked immunosorbent assay with the sameSMSV antigens and Norwalk capsid protein(Figure 1B, C, and D). Although cross-reactivitywas not detected, the serum samples may stillhave been able to cross-react with the caliciviruscausing hepatitis E or Sapporo calicivirus, whichwere not tested. However, sera typed to all 40marine caliciviruses reacted negatively whentested against Sapporo antigen (data not shown).These results suggest possible human exposureand antigenic response to marine caliciviruses. Ifthat is not the case, such results present aconfusing diagnostic picture of calicivirusexposure and diagnosis of human disease.

Figure 1. A) 150 blood donor sera tested against apolyvalent antigen containing San Miguel sea lionviruses (SMSVs) 5, 13, and 17 purified by CsCl; B) Eightacute- and eight convalescent-phase sera from aconfirmed outbreak of Norwalk gastroenteritis testedagainst the polyvalent SMSVs 5, 13, 17 antigen; C) Theeight acute-phase sera from the same outbreak ofNorwalk gastroenteritis tested in B also tested againstthe baculovirus expressed Norwalk virus capsid protein;D) The eight convalescent-phase sera paired with theacute-phase sera (See C) tested as in C.

Opt

ical

den

sity

Individual sera


Perspectives

Organism CharacteristicsThe history of marine caliciviruses demon-

strates that their biologic properties have greatplasticity. The VES virus-like caliciviruses canreplicate at temperatures of 15°C to 39°C, havediverse tissue trophisms, and travel by land(terrestrial reptiles, amphibians, and mammals),sea (pinnipeds, cetaceans, teleosts, and perhapsfilter-feeding mollusks), and air (pelagic birds,e.g., the white tern). They can persist in nonlyticcycles in many reservoir hosts, and they have awide diversity of successful antigenic types (1)(more than 40 serotypes on the basis of virusneutralization, e.g., no cross-protection be-tween types). Their cup-like surface morphol-ogy is characteristic (Figure 2). Finally, theyare zoonotic: this is a paradigm shift (7). Noother virus has been shown to have its originsand primary reservoirs in the sea yet emerge tocause disease in humans.

To measure calicivirus adaptivity andpreclude strong presumptions of host specificityon the basis of calicivirus type or species of origin,the following list of 16 hosts is given for a singlevirus serotype, SMSV-5: known natural hosts—five genera of seals, cattle, three genera ofwhales, donkeys, fox, and humans—and suscep-tible hosts—opaleye fish, horses, domestic swine,and primates (1). The lists are still growing.SMSV-5 can also persist for 60 days in shellfish,but infectivity has not been measured (26). Felinecalicivirus (FCV-F9) has an apparent oceanpresence among California sea lions and is nothost-specific (2). All members of the familyFelidae are susceptible to infection, not just

domestic cats. In addition, cheetahs aresusceptible; the agent has naturally infected andcaused disease in dogs and experimentallyinfected coyotes (28,29). Reports of humanantibody against the feline virus suggest zoonoticpotential for the feline calicivirus (30).

Tissue TrophismsThe broad host range and diverse mecha-

nisms of transmission and survival of marinecaliciviruses are expected of an RNA virusquasispecies (12). If structural simplicity associ-ated with a capsid made up of a single proteinspecies and replicative strategies conservedacross rather broad tissue and phylogeneticdistances is a measure, caliciviruses areprimitive RNA viruses. Caliciviral RNA replica-tive mechanisms are thus expected to generatenumerous mutants (perhaps as high as one to 10per template copy (12), which will come in contactwith many pelagic and terrestrial biota.Opportunity exists to form clusters of virusadapted across a diversity of life forms. Actualmutation rates have not been demonstrated forthe Caliciviridae, but plaque-size reversionstudies have found that the mutation rate for thisphenomenon is one per 106 replicates (14,15). Inaddition, the expected versatility from RNA virusreplicative infidelity and the resulting successfuladaptive mechanisms are manifested in the widespectrum of calicivirus tissue trophisms.

Disease conditions involving calicivirustissue trophisms include blistering of the skin(particularly on the appendages and around themouth), pneumonia, abortion, encephalitis,myocarditis, myositis, hepatitis, diarrhea, andcoagulation/hemorrhage (1,3,5,7; Table 1).Caliciviruses have the inherent potential andadaptive mechanisms to successfully parasitizeessentially all organ systems of the many animalspecies that have been examined in detail.

The Future Calicivirus disease manifestations in ani-

mals will likely continue but will only becomewell defined with improved diagnostic means.With cultivatable marine caliciviruses as models,the role of disease-causing caliciviruses can befurther defined. Now caliciviruses infectinghumans can only be visualized by electronmicroscopy or histochemistry but cannot bepropagated in vitro. Thus, miscarriage and birthdefects in human patients, hepatitis other than

Figure 2. Electron photomicrograph of CetaceanCalicivirus Tursiops - 1 (CCVTur-1). Negative stainingusing phosphotungstic acid on a carbon-coated gridshowing typical surface cup morphologic features ascommonly seen by electron microscopy. Bar = 100 nm.


Perspectives

types A through G, hand-foot-and-mouth–likediseases, viral myocarditis, viral encephalitis ofunknown etiology, and joint and muscle disease,for example, should be examined for caliciviruseswhen other causes of disease are not found.

In the absence of data, extrapolating fromcultivatable caliciviruses to predict future effectsof poorly characterized caliciviruses should beuseful, particularly when there is an urgent needto assess possible human risk. The calicivirusimplicated in RHD is a case in point for it mightbe expected to infect humans.

Additional evidence exists. An anecdotalaccount mentions a Mexican worker whodeveloped antibodies to RHD while eradicatingthe disease in Mexico (31). An Australian studydesigned to assess the risk for illness after RHDescaped from Wardang Island (32) examined agroup of 269 persons (153 reporting exposure torabbits or samples infected with RHD virus and116 reporting no known RHD virus contact) fromtwo Australian states with the greatest amount ofRHD virus activity in rabbits. Exposure wascategorized by degree of skin exposure to infectedmaterials. Date of first exposure was noted, butno cumulative exposure index was developed. A“high” exposure category could derive from oneexposure, and “low” exposure categories could

include multiple exposures, each with lowexposure. Symptoms were assessed by recall ofillness over the previous 13 months. Because theRHD agent was in high security containmentfacilities for the first 3 months of the recall periodand geographically confined for the following 3months, that period was considered a low exposureperiod. Because of the rapid spread of the virus inthe two states, the last 6 months of the recall periodwere considered the high exposure period.

The data (Table 2) show the rate ratios for theoccurrence of different illness in the two periods.All rate ratios were considerably greater than1.00, and the rate ratios for any illness, diarrhea/gastroenteritis, flu/fever, and neurologic illnessare significant (p < 0.005). Because each groupcontained health histories for 3 spring months or3 autumn months, 1 summer month, and 2 wintermonths, the data are seasonally adjusted; hence,winter illness does not explain the excesssymptoms observed in the high exposure group,and RHD virus exposure remains a plausibleexplanation for increased disease incidence.

It is difficult to produce pure cultures ofnoncultivatable caliciviruses to carry out Koch’spostulates and establish cause and effect for asingle pathogen strain or species. For RHD, botha calicivirus and a parvovirus have beenidentified in ill rabbits, and a parvovirus hasbeen isolated in vitro and reported to fulfillKoch’s postulates (33-35). Yet, caliciviruses havebeen purified from infected organs to the limits ofpurity by physical means, and those preparationsalso cause RHD (35). The caliciviruses purified byphysical means cannot be proven to be free of

Table 1. Calicivirus tissue trophisms

Disease Species Calicivirusconditions affected groupa

Skin Cattle, cats, dogs, VESV, SMSV, blistering humans, primates, FCV, CCV

seals, swinePneumonia Cats, cattle, swine FCV, SMSVAbortion Seals, swine VESV, SMSVEncephalitis Cats, primates, VESV, SMSV

seals, swineMyocarditis Seals, swine VESV, SMSVHepatitis Humans, rabbits, VESV, RHDV,

swine HEVDiarrhea Cattle, dogs, VESV, SMSV,

humans, reptiles, CCV, SRSV, swine Sapporo

Coagulation/ Humans, rabbits, RHDV, VESV, hemorrhage seals, swine HEVaThe family Caliciviridae has been tentatively divided intofive groups, each proposed to be a genus. Group 1: Vesicularexanthema of swine (VESV), San Miguel sea lion virus(SMSV), Feline calicivirus (FCV), Canine calicivirus (CCV);Group 2: Sapporo calicivirus (Sapporo); Group 3: Rabbithemorrhagic disease virus (RHDV); Group 4: Hepatitis Evirus (HEV); Group 5: Small round structured virus (SRSV),which includes Norwalk virus.

Table 2. Population incidence of rabbit hemorrhagicdisease (RHD) virus for seasonally equalized periods(July–December and February–July), derived fromMead et al. (32)

95%Jul–Dec Feb–Jul Rate Confidence

1995 1996 Ratio IntervalExposure to Low High RHD virusAny illness 112 210 1.88 1.49–2.36Flu/fever 94 189 2.01 1.57–2.57Diarrhea/ 41 73 1.78 1.21–2.61 gastroenteritisNeurologic 18 49 2.72 1.58–4.67 symptomsRashes/skin 3 10 3.33 0.92–12.1Bleeding/ 2 4 2.00 0.18–22.1 hepatitis


Perspectives

contaminating agents, such as parvovirus (35). IfRHD is parvovirus-driven, extrapolation fromwhat is known of other small DNA virusessuggests a rather stable genome and a reducedhost range with less likelihood of new hostrelationships (12). On the other hand, ifcalicivirus is the primary pathogen, thegenomic infidelity that occurs during smallRNA virus replication and the documentedcross-species transmission of the cultivatablecaliciviruses suggest that RHD might alsomove across species barriers (1,12).

Adequate diagnostic reagents for epidemio-logic studies need to be made available; theyinclude antigens, monoclonal antibodies, poly-merase chain reaction primer sets, and cDNAprobes based on group epitopes. In addition,biotype- or pathotype-specific reagents areneeded to differentiate pathogenic from non-pathogenic infections.

The future also holds the confoundingproblem of vaccines. Although vaccines can beproduced, because of calicivirus antigenicdiversity, their efficacy would be predictablyshort-lived and marginal. Other approaches willneed to be sought. Conserved traits that renderthe Caliciviridae viable as a virus with certainpredictable genomic expressions must be sought,and if they exist, targeted for immune attack.

ConclusionsOnly one of the five known calicivirus groups

can be grown in vitro and subjected to the fullrange of host-parasite tests and conditionsnecessary to more fully define a virus in nature.Therefore, extrapolations developed from thisgroup, the cultivatable marine caliciviruses,should provide insights as a predictive model tohelp answer questions for the noncultivatablecaliciviruses such as small round structuredvirus, Sapporo virus, hepatitis E virus, and rabbitcaliciviruses. From the replicative strategy of theCaliciviridae (as RNA viruses), one would predictconsiderable diversity. In vitro cultivation hasshown that caliciviruses exhibit survivability andplasticity in nature. Many of the factorsregarding host spectrum, zoonotic potential,disease conditions, transport, intermediate hosts,and abrupt appearance or disappearance, whichmay be unknown in newly emerging calicivirusdiseases (e.g., RHD), may be more reliablypredicted with an established model such as thecultivatable marine caliciviruses. New and better

biologic tools for diagnostic and epidemiologicassessments must be developed. This should beaugmented by recognizing the zoonotic potentialof the cultivatable caliciviruses of ocean originand then examining them as possible models tohelp solve many unanswered questions forpathogenic Caliciviridae.

AcknowledgmentsWe thank Ms. Christine Robinette for manuscript

preparation. This study was funded by the OregonAgricultural Experiment Station through the College ofVeterinary Medicine.

Alvin W. Smith is professor of veterinary virologyand head, Laboratory for Calicivirus Studies at theCollege of Veterinary Medicine, Oregon StateUniversity, Corvallis, Oregon. Dr. Smith is interestedin mechanisms for the preservation and movement ofpathogenic viruses in nature, particularly thosecontained in ocean reservoirs. His research hasfocused on the marine caliciviruses.

References 1. Smith AW, Boyt PM. Caliciviruses of ocean origin: a

review. The Journal of Zoo Wildlife Medicine1990;21:3-23.

2. Smith AW, Skilling DE. Comparison of antibody titers for13 calicivirus serotypes between aborting and full-termadult Zalophus californianus. Special report submitted tothe National Marine Mammal Laboratory, NationalMarine Fisheries Service. Corvallis (OR): Laboratory forCalicivirus Studies, Oregon State University; 1994.

3. Gunn RA, Janowski HT, Lieb S, Prather EC, GreenbergHB. Norwalk virus gastroenteritis following raw oysterconsumption. Am J Epidemiol 1982;115:348-51.

4. Morse DL, Grabau JJ, Hanrahan R, Stricof M,Shaegani M, Diebal R, et al. Widespread outbreaks ofclam- and oyster-associated gastroenteritis: role ofNorwalk virus. N Engl J Med 1986;314:678-81.

5. Bradley DW. Hepatitis E virus: a brief review of thebiology, molecular virology, and immunology of a novelvirus. J Hepatol 1995;22:140-5.

6. Chasey D. Possible origins of rabbit haemorrhagic diseasein the United Kingdom. Vet Rec 1994;135:496-9.

7. Smith AW, Berry ES, Skilling DE, Barlough JE, Poet SE,Berke T, et al. In vitro isolation and characterization of acalicivirus causing a vesicular disease of the hands andfeet. Clin Infect Dis. In press 1998.

8. Berke T, Golding B, Jiang K, Cubitt DW, Wolfaardt M,Smith AW, Matson DO. A phylogenetic analysis of thecaliciviruses. J Med Virol. In press 1997.

9. Barlough JE, Berry ES, Skilling DE, Smith AW. Themarine calicivirus story-part I. Compendium onContinuing Education for the Practicing Veterinarian1986;8:F5-F14.

10. Barlough JE, Berry ES, Skilling DE, Smith AW. Themarine calicivirus story-part II. Compendium onContinuing Education for the Practicing Veterinarian1986;8:F75-F82.


Perspectives

11. Bradley DW. Enterically transmitted non A non Bhepatitis. Br Med Bull 1990;46:442-61.

12. Holland J. Replication error, quasispecies populations,and extreme evolution rates of RNA viruses. In: MorseSS, editor. Emerging Viruses. Oxford: OxfordUniversity Press; 1993. p. 203-17.

13. Smith AW, Skilling DE, Applegate GL, Trayor TR, LolaTJ, Poet SE. Calicivirus isolation from stillborn pigletsin two outbreaks of swine infertility and respiratorysyndrome (SIRS). Proceedings of the World Associationof Veterinary Microbiologists, Immunologists, andSpecialists in Infectious Disease. Davis (CA):University of California, Davis; 1992.

14. Smith AW, Madin SH. Vesicular exanthema of swine.In: Leman AD, Straw B, Glock RD, Mengeling WL,Penny RHC, Scholl E, editors. Diseases of swine. 6thedition. Ames (IA): Iowa State Press; 1986. p. 358-68.

15. Bankowski RA. Vesicular exanthema. In: BankowskiRA, editor. Advances in veterinary science. AcademicPress; 1965. p. 23-64.

16. Morse SS. Examining the origins of emerging viruses.In: Morse SS, editor. Emerging viruses. Oxford: OxfordUniversity Press; 1993. p. 10-28.

17. Smith AW, Akers TG, Madin SD, Vedros NA. SanMiguel sea lion virus isolation, preliminarycharacterization and relationship to vesicularexanthema of swine virus. Nature 1973;244:108-10.

18. Smith AW, Akers TG. Vesicular exanthema of swine. JAm Vet Med Assoc 1976;169:700-3.

19. Smith AW, Skilling DE, Ensley PK, Benirschke K, LesterTL. Calicivirus isolation and persistence in a pygmychimpanzee (Pan paniscus). Science 1983;221:79-81.

20. Smith AW, Skilling DE, Prato CM, Bray HL.Calicivirus (SMSV-5) infection in experimentallyinoculated opaleye fish (Girella nigricans). Arch Virol1981;67:165-8.

21. Smith AW, Skilling DE, Barlough JE, Berry ES.Distribution in the North Pacific Ocean, Bering Sea,and Arctic Ocean of animal populations known to carrypathogenic caliciviruses. Diseases of Aquatic Organisms1986;2:73-80.

22. Smith AW, Prato CM, Skilling DE. Calicivirusesinfecting monkeys and possibly man. Am J Vet Res1978;39:287-9.

23. Smith AW, Skilling DE, Dardiri AH, Latham AB.Calicivirus pathogenic for swine: a new serotypeisolated from opaleye Girella nigricans, an ocean fish.Science 1980;209:940-1.

24. Smith AW, Anderson MP, Skilling DE, Barlough JE,Ensley PK. First isolation of calicivirus from reptilesand amphibians. J Am Vet Med Assoc 1986;47:1718-21.

25. Sawyer JC. Vesicular exanthema of swine and SanMiguel sea lion virus. J Am Vet Med Assoc1976;169:707-9.

26. Smith AW, Reno P, Poet SE, Skilling DE, Stafford C.Retention of ocean-origin caliciviruses in bivalvemollusks maintained under experimental depurationconditions. In: Fenwick B, editor. Proceedings of the25th Annual Meeting of the International Associationof Aquatic Animals; 1994; Vallejo, California.Baltimore: International Association of AquaticAnimal Medicine; 1994.

27. Poet SE, Skilling DE, Megyesi JL, Gilmartin WG,Smith AW. Detection of a non-cultivatable calicivirusfrom the white tern (Gygis alba rothschildi). J WildlDis 1996;32:461-7.

28. Evermann JF, Bryan CM, McKiernan AJ. Isolation of acalicivirus from a case of canine glossitis. CaninePractice 1981;8:36-9.

29. Evermann JF, McKiernan AJ, Smith AW, Skilling DE,Ott RL. Isolation of a calicivirus from dogs with entericinfections. Am J Vet Res 1985;46:218-20.

30. Cubitt DW. Proceedings of the European Society ofVeterinary Virology. Readings, United Kingdom:Reading University; 1996.

31. Bureau of Resource Studies. Rabbit calicivirus disease.Canberra, Australia: Australian Government PrintingOffice; August 1996. p. 20-56.

32. Mead C, Kaldor J, Canton M, Gamer G, Crerar S,Thomas S. Rabbit calicivirus and human health.Canberra, Australia: Department of Primary Industriesand Energy, Australian Government (Released underthe Official Information Act). Report of the RabbitCalicivirus Human Health Study Group; 1996.

33. Xu WY. Viral haemorrhagic disease of rabbits in thePeople’s Republic of China: epidemiology and viruscharacterization. Revue Scientifique et Technique,Office International des Epizooties 1991;10:2393-408.

34. Gregg DA, House C, Meyer R, Berninger M. Viralhaemorrhagic disease of rabbits in Mexico: epidemiologyand viral characterization. Revue Scientifique etTechnique, Office International des Epizooties1991;10:2435-51.

35. Ohlinger VF, Thiel HJ. Identification of the viralhaemorrhagic disease in rabbits as a calicivirus. RevueScientifique et Technique, Office International desEpizooties 1991;10:2311-23.


Perspectives

Investigations of acute infectious diseaseoutbreaks are very common, and the results ofsuch investigations are often published;however, surprisingly little has been writtenabout the actual procedures followed duringsuch investigations (1,2). Most epidemiologistsand public health officials learn the proceduresby conducting investigations with the initialassistance of more experienced colleagues. Thisarticle outlines the general approach toconducting an outbreak investigation. Theapproach applies not only to infectious diseaseoutbreaks but also to outbreaks due tononinfectious causes (e.g., toxic exposure).

How Outbreaks Are RecognizedPossible outbreaks of disease come to the

attention of public health officials in variousways. Often, an astute clinician, infection controlnurse, or clinical laboratory worker first noticesan unusual disease or an unusual number ofcases of a disease and alerts public healthofficials. For example, staphylococcal toxic shocksyndrome and eosinophilia myalgia syndromewere first noted by clinicians (3,4). Frequently, itis the patient (or someone close to the patient)who first suspects a problem, as is often the casein foodborne outbreaks after a shared meal andas was the case in the investigation of a cluster ofcases of apparent juvenile rheumatoid arthritisnear Lyme, Connecticut, which led to thediscovery of Lyme disease (5). Review of routinelycollected surveillance data can also detect

outbreaks of known diseases, as in the case ofhepatitis B infection among the patients of an oralsurgeon in Connecticut and patients at a weightreduction clinic (6,7). The former outbreak wasfirst suspected when routinely submittedcommunicable disease report forms for severalpatients from one small town indicated that allof the patients had recently had oral surgery.However, it is relatively uncommon foroutbreaks to be detected in this way and evenmore uncommon for them to be detected in thisway while they are still in progress. Finally,sometimes public health officials learn aboutoutbreaks of disease from the local newspaperor television news.

Reasons for Investigating OutbreaksThe most compelling reason to investigate a

recognized outbreak of disease is that exposure tothe source(s) of infection may be continuing; byidentifying and eliminating the source ofinfection, we can prevent additional cases. Forexample, if cans of mushrooms containingbotulinum toxin are still on store shelves or inhomes or restaurants, their recall and destruc-tion can prevent further cases of botulism.

However, even if an outbreak is essentiallyover by the time the epidemiologic investigationbegins—that is, if no one is being further exposedto the source of infection—investigating theoutbreak may still be indicated for many reasons.Foremost is that the results of the investigationmay lead to recommendations or strategies forpreventing similar future outbreaks. Forexample, a Legionnaires’ disease outbreakinvestigation may produce recommendationsfor grocery store misting machine use that mayprevent other outbreaks (8). Other reasons for

Outbreak Investigations—A Perspective

Arthur L. ReingoldUniversity of California, Berkeley, California, USA

Address for correspondence: Arthur L. Reingold, Division ofPublic Health Biology and Epidemiology, School of PublicHealth, University of California, Berkeley, 140 Warren Hall,Berkeley, CA 94720-7360, USA; fax: 510-643-5163; e-mail:[email protected].

Outbreak investigations, an important and challenging component of epidemiologyand public health, can help identify the source of ongoing outbreaks and preventadditional cases. Even when an outbreak is over, a thorough epidemiologic andenvironmental investigation often can increase our knowledge of a given disease andprevent future outbreaks. Finally, outbreak investigations provide epidemiologic trainingand foster cooperation between the clinical and public health communities.


Perspectives

investigating outbreaks are the opportunity to 1)describe new diseases and learn more about knowndiseases; 2) evaluate existing prevention strategies,e.g., vaccines; 3) teach (and learn) epidemiology;and 4) address public concern about the outbreak.

Once a decision is made to investigate anoutbreak, three types of activities are generallyinvolved—the epidemiologic investigation; theenvironmental investigation; and the interactionwith the public, the press, and, in manyinstances, the legal system. While these activitiesoften occur simultaneously throughout theinvestigation, it is conceptually easier to considereach of them separately.

Epidemiologic InvestigationOutbreak investigations are, in theory,

indistinguishable from other epidemiologic in-vestigations; however, outbreak investigationsencounter more constraints. 1) If the outbreak isongoing at the time of the investigation, there isgreat urgency to find the source and preventadditional cases. 2) Because outbreak investiga-tions frequently are public, there is substantialpressure to conclude them rapidly, particularly ifthe outbreak is ongoing. 3) In many outbreaks,the number of cases available for study is limited;therefore, the statistical power of the investiga-tion is limited. 4) Early media reports concerningthe outbreak may bias the responses of personssubsequently interviewed. 5) Because of legalliability and the financial interests of persons andinstitutions involved, there is pressure toconclude the investigation quickly, which maylead to hasty decisions regarding the source of theoutbreak. 6) If detection of the outbreak is delayed,useful clinical and environmental samples may bevery difficult or impossible to obtain.

Outbreak investigations have essential com-ponents as follows: 1) establish case definition(s);2) confirm that cases are “real”; 3) establish thebackground rate of disease; 4) find cases, decide ifthere is an outbreak, define scope of the outbreak;5) examine the descriptive epidemiologic featuresof the cases; 6) generate hypotheses; 7) testhypotheses; 8) collect and test environmentalsamples; 9) implement control measures; and 10)interact with the press, inform the public. Whilethe first seven components are listed in logicalorder, in most outbreak investigations, manyoccur more or less simultaneously. The impor-tance of these components may vary dependingon the circumstances of a specific outbreak.

Case DefinitionIn some outbreaks, formulating the case

definition(s) and exclusion criteria is straightfor-ward; for example, in an outbreak of gastroenteri-tis caused by Salmonella infection, a laboratory-confirmed case would be defined as a culture-confirmed infection with Salmonella or perhapswith Salmonella of the particular serotypecausing the outbreak, while a clinical casedefinition might be new onset of diarrhea. Inother outbreaks, the case definition and exclusioncriteria are complex, particularly if the disease isnew and the range of clinical manifestations isunknown (e.g., in a putative outbreak of chronicfatigue syndrome). In many outbreak investiga-tions, multiple case definitions are used (e.g.,laboratory-confirmed case vs. clinical case; definitevs. probable vs. possible case; outbreak-associatedcase vs. nonoutbreak-associated case, primary casevs. secondary case) and the resulting data areanalyzed by using different case definitions.When the number of cases available for study isnot a limiting factor and a case-control study isbeing used to examine risk factors for becominga case, a strict case definition is oftenpreferable to increase specificity and reducemisclassification of disease status (i.e., reducethe chance of including cases of unrelatedillness or no illness as outbreak-related cases).

Case ConfirmationIn certain outbreaks, clinical findings in

reported cases should be reviewed closely, eitherdirectly, by examining the patients, or indirectly,by detailed review of the medical records anddiscussion with the attending health-careprovider(s), especially when a new diseaseappears to be emerging (e.g., in the earlyinvestigations of Legionnaires’ disease, AIDS,eosinophilia myalgia syndrome, and hantaviruspulmonary syndrome) (4,9-11). Clinical findingsshould also be examined closely when some or allof the observed cases may be factitious, perhapsbecause of laboratory error (12); a discrepancybetween the clinical and laboratory findingsgenerally exists, which may be discernible onlyby a detailed review of the clinical findings.

Establishing the Background Rate of Diseaseand Finding Cases

Once it is clear that a suspected outbreak isnot the result of laboratory error, a set ofactivities should be undertaken to establish the


Perspectives

background rate of the disease in the affectedpopulation and to find all the cases in a givenpopulation in a certain period. This set ofactivities should prove that the observednumber of cases truly is in excess of the “usual”number (i.e., that an outbreak has occurred),define the scope of the outbreak geographicallyand temporally, find cases to describe theepidemiologic features of those affected and toinclude them in analytic epidemiologic studies(see below) or, most often, accomplish acombination of these goals.

When hundreds of acute onset diarrhea casesare suddenly seen daily in a single outpatientsetting (10), an outbreak is clearly occurring. Onthe other hand, when too many hospitalizedpatients are dying unexpectedly of cardiac arrest(13) or the number of cases of listeriosis in a givencounty in recent months is moderately elevated,it may be necessary to establish the backgroundrates in the population to determine whether anoutbreak is occurring. In such situations, the periodand geographic areas involved would provide themost useful baseline data, keeping in mind that thelabor and time required to collect such informationis often directly proportional to the length of theperiod and the size of the geographic area selected.Because disease incidence normally fluctuates byseason, data from comparable seasons in earlieryears should be included.

Establishing the background rate of a diseaseis generally more straightforward if confirmatorytests are available than if laboratory tests areunavailable or infrequently used. The rate ofcertain invasive bacterial infections (e.g., listerio-sis and meningococcal infections) in a given areacan be easily documented by reviewing therecords of hospital clinical microbiology laborato-ries; however, cases for which specimens were notsubmitted to these laboratories for testing will goundetected. When a disease is less frequentlylaboratory-confirmed because health-care pro-viders may not have considered the diagnosis orordered the appropriate laboratory tests (e.g., forLegionnaires’ disease), establishing the back-ground rate of disease in a community or ahospital suspected of having an outbreakgenerally requires alternative case-finding strat-egies and is almost invariably more laborintensive. In an outbreak of a new disease,substantial effort is often necessary to determinewhether or not cases of that disease had beenoccurring but had gone unrecognized.

Once data concerning the background rate ofa disease (including case-finding for the currentperiod) have been collected, it is generallypossible to determine whether or not an outbreakis occurring or has occurred, although in somesituations it may remain unclear whether or notthe number of cases observed exceeds thebackground rate. In part, the problem may relateto how an outbreak is defined. To paraphrase aU.S. Supreme Court justice speaking aboutpornography, “I can’t define an outbreak, but Iknow one when I see one.” Thus, it may bedifficult to detect and prove the existence of smalloutbreaks, but large ones are self-evident.

An outbreak can also be difficult to identifywhen during the period under study changesoccur in the care-seeking behavior and access tocare of patients; the level of suspicion, referralpatterns, and test-ordering practices of health-care providers; the diagnostic tests and otherprocedures used by laboratories; and theprevalence of underlying immunosuppressiveconditions or other host factors in the population.All these factors, which can affect the apparentincidence of a disease and produce artifactualchanges perceived as increases (or decreases) inthe actual incidence, need to be considered wheninterpreting the findings.

Descriptive EpidemiologyBy collecting patient data, the case-finding

activities provide extremely important informa-tion concerning the descriptive epidemiologicfeatures of the outbreak. By reviewing andplotting on an “epidemic curve” the times of onsetof the cases and by examining the characteristics(e.g., age, sex, race/ethnicity, residence, occupa-tion, recent travel, or attendance at events) of theill persons, investigators can often generatehypotheses concerning the cause(s)/source(s) ofthe outbreak. While linking the sudden onset ofgastroenteritis among scores of persons whoattended a church supper to the single commonmeal they shared is generally not a challenge, anotherwise cryptic source can be at least hinted atby the descriptive epidemiologic features of thecases involved. For example, in a particularlyperplexing outbreak of Salmonella Muencheninfections ultimately traced to contaminatedmarijuana, the age distribution of the affectedpersons and of their households was markedlydifferent from that typically seen for salmonello-sis (14). Or, similarly, in the outbreak of


Perspectives

legionellosis due to contaminated mistingmachines in the produce section of a grocerystore, before the link to this exposure was evensuspected, it was noted that women constituted asubstantially higher proportion of the casesusually seen with this disease (5). The shape ofthe epidemic curve can also be very instructive,suggesting a point-source epidemic, ongoingtransmission, or a combination of the two.

Generating a HypothesisThe source(s) and route(s) of exposure must

be determined to understand why an outbreakoccurred, how to prevent similar outbreaks in thefuture, and, if the outbreak is ongoing, how toprevent others from being exposed to thesource(s) of infection. In some outbreaks, thesource and route are obvious to those involved inthe outbreak and to the investigators. However,even when the source of exposure appearsobvious at the outset, a modicum of skepticismshould be retained because the obvious answer isnot invariably correct. For example, in anoutbreak of nosocomial legionellosis in RhodeIsland, the results of an earlier investigation intoa small number of hospital-acquired cases at thesame hospital had demonstrated that Legionellapneumophila was in the hospital potable watersupply, and a sudden increase in new cases wasstrongly believed to be related to the potablewater (15). However, a detailed epidemiologicinvestigation implicated a new cooling tower atthe hospital as the source of the second outbreak.

While the true source of exposure, or at leasta relatively short list of possibilities, is apparentin many outbreaks, this is not the case in the morechallenging outbreaks. In these instances, hypoth-eses concerning the source/route of exposure can begenerated in a number of ways beyond a detailedreview of the descriptive epidemiologic findings. Areview of existing epidemiologic, microbiologic, andveterinary data is very useful for learning aboutknown and suspected sources of previous outbreaksor sporadic cases of a given infection or disease, aswell as the ecologic niche of an infectious agent.Thus, in an outbreak of invasive Streptococcuszooepidemicus infections in New Mexico due toconsumption of soft cheese made from contami-nated raw milk, the investigation focused onexposure to dairy products and animals because ofprevious microbiologic and veterinary studies (16).

A review of existing data generally only helpsconfirm what is already known about a particular

disease and is far less helpful in identifyingtotally new and unsuspected sources or routes ofinfection (i.e., marijuana as a source ofSalmonella). When neither review of thedescriptive epidemiologic features of the casesnor review of existing scientific informationyields the correct hypothesis, other methods canbe used to generate hypotheses about what thepatients have in common. Open-ended interviewsof those infected (or their surrogates) are onesuch method in which investigators try to identifyall possibly relevant exposures (e.g., a list of allfoods consumed) during a given period. Forexample, in an investigation of Yersiniaenterocolitica infections in young children inBelgium, open-ended interviews of the mothersof some of the ill children showed that many gavetheir children raw pork sausage as a weaningfood, providing the first clue as to the source ofthese infections (17). Similarly, in two outbreaksof foodborne listeriosis, a variant of this processled to the identification of the source of theoutbreak. In one of these outbreaks, a search ofthe refrigerator of one of the case-patients who,as a visitor to the area, had had very limitedexposure to foods there, suggested cole slaw as apossible vehicle of infection (18). In the otheroutbreak, an initial case-control study found nodifferences between cases and controls regardingexposure to a number of specific food items butshowed that case households were more likelythan control households to buy their food at aparticular foodstore chain. To generate a list ofother possible food sources of infection, investiga-tors shopped with persons who did the shoppingfor case households and compiled a list of foodspurchased at that foodstore chain that had notbeen reported in the previous study. Thisapproach implicated pasteurized milk from thatchain as the source of the outbreak (19).

In some particularly perplexing outbreaks,bringing together a subset of the patients todiscuss their experiences and exposures in a waythat may reveal unidentified links can be useful.

Testing the HypothesisWhether a hypothesis explaining the occur-

rence of an outbreak is easy or difficult to generate,an analytic epidemiologic study to test the proposedhypothesis should be considered. While in manyinstances a case-control study is used, otherdesigns, including retrospective cohort and cross-sectional studies, can be equally or more


Perspectives

appropriate. The goal of all these studies is to assessthe relationship between a given exposure and thedisease under study. Thus, each exposure ofinterest (e.g., each of the meals eaten together bypassengers on a cruise ship and each of the foodsand beverages served at those meals) constitutes aseparate hypothesis to be tested in the analyticstudy. In outbreaks where generating the correcthypothesis is difficult, multiple analytic studies,with additional hypothesis-generating activities inbetween, are sometimes needed before the correcthypothesis is formed and tested (19).

In interpreting the results of such analyticstudies, one must consider the possibility that“statistically significant” associations betweenone or more exposures and the disease may bechance findings, not indicative of a truerelationship. By definition, any “statisticallysignificant” association may have occurred bychance. (When the standard cut point of p < 0.05is used, this occurs 5% of the time.) Because manyanalytic epidemiologic studies of outbreaksinvolve testing many hypotheses, the problem of“multiple comparisons” arises often.

While there are statistical methods foradjusting for multiple comparisons, when andeven whether to use them is controversial. At aminimum, it is important to go beyond thestatistical tests and examine the magnitude ofthe effect observed between exposure and disease(e.g., the odds ratio, relative risk) and the 95%confidence intervals, as well as biologic plausibil-ity in deciding whether or not a given“statistically significant” relationship is likely tobe biologically meaningful. Evidence of a dose-response effect between a given exposure andillness (i.e., the greater the exposure, the greaterthe risk for illness) makes a causal relationshipbetween exposure and disease more likely.Whether the time interval between a givenexposure and onset of illness is consistent withwhat is known about the incubation period of thedisease under study must also be assessed. Whenillness is “statistically significantly” related tomore than one exposure (e.g., to eating each ofseveral foods at a common meal), it is importantto determine whether multiple sources ofinfection (perhaps due to cross-contamination)are plausible and whether some of the notedassociations are due to confounding (e.g.,exposure to one potential source is linked toexposure to other sources) or to chance.

When trying to decide if a “statisticallysignificant” exposure is the source of an outbreak,it is important to consider what proportion of thecases can be accounted for by that exposure. Oneor more of the patients may be classified as“nonexposed” for various reasons: incorrectinformation concerning exposure status (due topoor memory, language barriers); multiplesources of exposure or routes of transmission(perhaps due to cross-contamination); secondaryperson-to-person transmission that followed acommon source exposure; or patients withoutthe suspected exposure, representing back-ground cases of the disease unrelated to theoutbreak. The plausibility of each of theseexplanations varies by outbreak. While there isno cutoff point above or below which theproportion of exposed case-patients should fallbefore an exposure is thought to account for anoutbreak, the lower this proportion, the lesslikely the exposure is, by itself, the source.

Other possibilities need to be consideredwhen the analytic epidemiologic study finds noassociation between the hypothesized exposuresand risk for disease. The most obvious possibilityis that the real exposure was not among thoseexamined, and additional hypotheses should begenerated. However, other possibilities shouldalso be considered, particularly when the settingof the outbreak makes this first explanationunlikely (e.g., when it is known that thoseinvolved in the outbreak shared only a singleexposure or set of exposures, such as eating asingle common meal). Two other explanations forfailing to find a “statistically significant” linkbetween one or more exposures and risk forillness also need to be considered—the number ofpersons available for study and the accuracy ofthe available information concerning the expo-sures. Thus, if the outbreak involves only a smallnumber of cases (and non-ill persons), thestatistical power of the analytic study to find atrue difference in exposure between the ill andthe non-ill (or a difference in the rate of diseaseamong the exposed and the unexposed) is verylimited. If the persons involved in the outbreak donot provide accurate information about theirexposure to suspected sources or vehicles ofinfection because of lack of knowledge, poormemory, language difficulty, mental impair-ment, or other reasons, the resultingmisclassification of exposure status also can


Perspectives

prevent the epidemiologic study from implicatingthe source of infection. Studies have documentedthat even under ideal circumstances, memoryconcerning such exposures is faulty (20). However,given the usually enormous differences in rates ofdisease between those exposed and those notexposed to the source of the outbreak, even smallstudies or studies with substantial misclassificationof exposure can still correctly identify the source.

Environmental InvestigationSamples of foods and beverages served at a

common meal believed to be the source of anoutbreak of gastroenteritis or samples of thewater or drift from a cooling tower believed to bethe source of an outbreak of Legionnaires’ diseasecan support epidemiologic findings. In the bestscenario, the findings of the epidemiologicinvestigation would guide the collection andtesting of environmental samples. However,environmental specimens often need to beobtained as soon as possible, either before theyare no longer available, as in the case of residualfood from a common meal, or before environmen-tal interventions are implemented, as in the caseof treating a cooling tower to eradicateLegionella. Because laboratory testing of envi-ronmental samples is often expensive and labor-intensive, it is sometimes reasonable to collectand store many samples but test only a limitednumber. Collaborating with a sanitarian, envi-ronmental engineer, or other professional duringan environmental inspection or collection ofspecimens is always beneficial.

While finding or not finding the causativeorganism in environmental samples is oftenperceived by the public, the media, and the courtsas powerful evidence implicating or exoneratingan environmental source, either positive ornegative findings can be misleading for severalreasons. For example, finding Legionella in ahospital potable water system does not prove thatthe potable water (rather than a cooling tower orsome other source) is responsible for an outbreakof Legionnaires’ disease (21). Similarly, notfinding the causative organism in an environ-mental sample does not conclusively rule out asource as the cause of the problem, in partbecause the samples obtained and tested maynot represent the source (e.g., because of errorin collecting the specimens, interveningchanges in the environmental source) and inpart because the samples may have been

mishandled. Furthermore, in some outbreakscaused by well-characterized etiologic agents,laboratory methods of detecting the agent inenvironmental samples are insensitive, techni-cally difficult, or not available, as in the case ofrecent outbreaks of Cyclospora infections associ-ated with eating imported berries (22,23).

Control MeasuresCentral to any outbreak investigation is the

timely implementation of appropriate controlmeasures to minimize further illness and death. Atbest, the implementation of control measureswould be guided by the results of the epidemiologicinvestigation and possibly (when appropriate) thetesting of environmental specimens. However, thisapproach may delay prevention of further exposureto a suspected source of the outbreak and is,therefore, unacceptable from a public healthperspective. Because the recall of a food product,the closing of a restaurant, or similar interventionscan have profound economic and legal implicationsfor an institution, a manufacturer or owner, and theemployees of the establishments involved, actingprecipitously can also have substantial negativeeffects. The recent attribution of an outbreak ofCyclospora infections to strawberries from Califor-nia demonstrates the economic impact that canresult from releasing and acting on incorrectinformation (22,23). Thus, the timing and nature ofcontrol measures are difficult. Balancing theresponsibility to prevent further disease with theneed to protect the credibility and reputation of aninstitution is very challenging.

Interactions with the Public and PressWhile the public and the press are not aware

of most outbreak investigations, media attentionand public concern become part of someinvestigations. Throughout the course of anoutbreak investigation, the need to shareinformation with public officials, the press, thepublic, and the population affected by theoutbreak must be assessed. While press, radio,and television reports can at times be inaccurate,overall the media can be a powerful means ofsharing information about an investigation withthe public and disseminating timely informationabout product recalls.

Dr. Reingold worked as an epidemiologist at theCenters for Disease Control and Prevention for 8 yearsbefore joining the faculty of the School of Public


Perspectives

Health at the University of California, Berkeley. He iscurrently professor of epidemiology and head of theDivision of Public Health Biology and Epidemiology.

References 1. Goodman RA, Buehler JW, Koplan JP. The

epidemiologic field investigation: science and judgmentin public health practice. Am J Epidemiol 1990;132:9-16.

2. MacKenzie WR, Goodman RA. The public healthresponse to an outbreak. Current Issues in PublicHealth1996;2:1-4.

3. Chesney PJ, Chesney RW, Purdy W, Nelson D,McPherson T, Wand P, et al. Epidemiologic notes andreports: toxic-shock syndrome—United States. MMWRMorb Mortal Wkly Rep 1980;29:229-30.

4. Hertzman PA, Blevins WL, Mayer J, Greenfield B,Ting M, Gleich GJ, et al. Association of eosinophilia-myalgia syndrome with the ingestion of tryptophan. NEngl J Med 1980;322:871.

5. Steere AC, Malawista SE, Syndman DR, Shope RF,Andman WA, Ross MR, Steele FM. Lyme arthritis: anepidemic of oligoarticular arthritis in children andadults in three Connecticut communities. ArthritisRheum 1977;20:7.

6. Reingold AL, Kane MA, Murphy BL, Checko P, FrancisDP, Maynard JE. Transmission of Hepatitis B by anoral surgeon. J Infect Dis 1982;145:262.

7. Canter J, Mackey K, Good LS, Roberto RR, Chin J,Bond WW, et al. An outbreak of hepatitis B associatedwith jet injections in a weight reduction clinic. ArchIntern Med 1990;150:1923-7.

8. Mahoney FJ, Hoge CW, Farley TA, Barbaree JM,Breiman RF, Benson RF, McFarland LM.Communitywide outbreak of Legionnaires’ diseaseassociated with a grocery store mist machine. J InfectDis 1992;165:736.

9. Fraser DW, Tsai TR, Orenstein W, Parkin WE,Beecham HJ, Sharrar RG, et al. Legionnaires’ disease:description of an epidemic of pneumonia. N Engl J Med1977;297:1189-97.

10. Kriedman-Kien A, Laubenstein L, Marmor M, HymesK, Green J, Ragaz A, et al. Kaposi’s sarcoma andPneumocystis pneumonia among homosexual men—New York City and California. MMWR Morb MortalWkly Rep 1981;30:305-8.

11. Koster F, Levy H, Mertz G, Young S, Foucar K,McLaughlin J, et al. Outbreak of acute illness—southwestern United States. MMWR Morb MortalWkly Rep 1993;42:421-4.

12. Weinstein RA, Bauer FW, Hoffman RD, Tyler PG,Anderson RL, Stamm WE. Factitious meningitis:diagnostic error due to nonviable bacteria incommercial lumbar puncture trays. JAMA1975;233:878.

13. Buehler JW, Smith LF, Wallace EM, Heath CW,Rusiak R, Herndon JL. Unexplained deaths in achildren’s hospital: an epidemiologic assessment. NEngl J Med 1985;313:211.

14. Taylor DN, Wachsmuth IK, Shangkuan Y-H, SchmidtEV, Barrett TJ, Schrader JS, et al. Salmonellosisassociated with marijuana: a multistate outbreaktraced by plasmid fingerprinting. N Engl J Med1982;306:1249.

15. Garbe PL, Davis BJ, Weisfeld JS, Markowitz L, MinerP, Garrity F, et al. Nosocomial Legionnaires’ disease:epidemiologic demonstration of cooling towers as asource. JAMA 1985;254:521.

16. Espinosa FH, Ryan WM, Vigil PL, Gregory DF, HilleyRB, Romig DA, et al. Group C streptococcal infectionsassociated with eating homemade cheese: New Mexico.MMWR Morb Mortal Wkly Rep 1983;32:514.

17. Tauxe RV, Walters G, Goossen V, VanNoyer R,Vandepitte J, Martin SM, et al. Yersinia enterocoliticainfections and pork: the missing link. Lancet1987;5:1129.

18. Schlech WF, Lavigne PM, Bortolussi RA, Allen AC,Haldane EV, Wort AJ, et al. Epidemic listeriosis:evidence for transmission by food. N Engl J Med1983;308:203.

19. Fleming DW, Cochi SL, MacDonald KL, Brondum J,Hayes PS, Plikaytis BD, et al. Pasteurized milk as avehicle of infection in an outbreak of listeriosis. N EnglJ Med 1985;312:404.

20. Decker MD, Booth AL, Dewey MJ, Fricker RS,Hutcheson RH, Schaffner W. Validity of foodconsumption histories in a foodborne outbreakinvestigation. Am J Epidemiol 1986;124:859.

21. Hayes EB, Matte TD, O’Brien TR, McKinley TW,Logsdon GS, Rose JB, et al. Large community outbreakof cryptosporidiosis due to contamination of a filteredpublic water supply. N Engl J Med 1989;390:1372.

22. Chambers J, Somerfieldt S, Mackey L, Nichols S, BallR, Roberts D, et al. Outbreaks of Cyclosporacayetanensis infection—United States, 1996. MMWRMorb Mortal Wkly Rep 1996;45:549-51.

23. Hofman J, Liu Z, Genese C, Wolf G, Manley W, Pilot K,et al. Update: outbreaks of Cyclospora cayetanensisinfection—United States and Canada, 1996. MMWRMorb Mortal Wkly Rep 1996;45:611-2.


Synopses

Until the advent of a live-attenuated vaccinein the early 1960s, measles was an epidemicdisease worldwide. Today many countries havecontrolled measles, but the disease remainsendemic on most continents. Development of alive-attenuated measles vaccine and implemen-tation of laws that required proof of vaccinationupon school entry dramatically reduced theincidence of measles in the United States. Thenumber of reported cases plummeted fromapproximately 500,000 before vaccine intro-duction in 1963 to fewer than 1,500 in 1983.Despite these measures, a reemergence orresurgence of measles in the United Statesfrom 1989 to 1991 resulted in more than 55,000cases of measles and approximately 120measles-associated deaths (Figure 1; 1). Inexploring the reasons for the resurgence, ourlaboratory genetically characterized measles vi-ruses isolated from wild-type virus-infectedpersons from the same outbreak or temporally andgeographically distinct outbreaks in the UnitedStates; in the regions examined, all measles virusesisolated during the period of resurgence were

almost identical in nucleotide sequence andgenetically distinct from vaccine strains (2).

Measles isolates from many regions of theworld have been characterized in parallel studiesby our laboratory and by others. In conjunctionwith classic epidemiologic investigations, thesewell-characterized viruses have formed a pictureof the distribution of wild-type measles viruses inmost areas of the world. Eight distinct genotypeshave been identified, and undoubtedly more willbe added. Some are localized to specific regions,while most are widely distributed. The assembly

Genetic Diversity of Wild-Type MeaslesViruses: Implications for GlobalMeasles Elimination Programs

William J. Bellini and Paul A. RotaCenters for Disease Control and Prevention, Atlanta, Georgia, USA

Address for correspondence: William J. Bellini, MS C-22,Centers for Disease Control and Prevention, 1600 Clifton Rd.,Atlanta, GA, 30333 USA; fax: 404-639-4187; e-mail:[email protected].

Wild-type measles viruses have been divided into distinct genetic groupsaccording to the nucleotide sequences of their hemagglutinin and nucleoprotein genes.Most genetic groups have worldwide distribution; however, at least two of the groupsappear to have a more limited circulation. To monitor the transmission pathways ofmeasles virus, we observed the geographic distribution of genetic groups, as well aschanges in them in a particular region over time. We found evidence of interruption ofindigenous transmission of measles in the United States after 1993 and identified thesources of imported virus associated with cases and outbreaks after 1993. The patternof measles genetic groups provided a means to describe measles outbreaks andassess the extent of virus circulation in a given area. We expect that molecularepidemiologic studies will become a powerful tool for evaluating strategies to control,eliminate, and eventually eradicate measles.

Figure 1. Incidence of U.S. measles cases and measlesdeaths between 1980 and 1997. Total number ofmeasles cases for years 1993-1997 (week 35) areindicated above each bar.

YEARS

Rep

orte

d D

eath

s

Mea

sles

Cas

es

80 81 82 83 84 85 86 87 88 89 90 91 92 93 94

20

40

60

80

100

00

5,000

10,000

15,000

20,000

25,000

30,000

95

312 958 301

96

488

Interruption in transmission

97 (wk 35)

102


Synopses

of these sequences into a large database, whichincludes their geographic distribution, has becomea new means by which measles transmissionpathways can be traced and control measures canbe assessed. This “molecular epidemiology” hasaffected the U.S. measles elimination program, andif used appropriately with standard epidemio-logic methods, it will affect global measleselimination and eradication. This article summa-rizes the status of the known measles genotypedistribution throughout the world and describeshow molecular epidemiologic information hasbeen used to assess the effectiveness of measleselimination in the United States.

Global Distribution of Measles GenotypesIn most cases, genetic characterization of

wild-type measles viruses has been conducted bysequencing the genes coding for the hemaggluti-nin (H) protein or the nucleoprotein (N). Of thesix genes on the viral genome, the H and N genesare the most variable. Over their protein codingregions, the H and N genes vary by up to 7% at thenucleotide level. The single most variable part ofthe measles genome is the 450 nucleotides thatcode for the COOH-terminus of the N protein,where nucleotide variability between variouswild-type viruses can approach 12%. Severallaboratories have analyzed the sequences of wild-type measles viruses and assigned the viruses tovarious genetic groups (2-14). At present, nowidely accepted standard is available fordescribing genetic groups of measles virus.However, with a few exceptions, the assignmentof viruses to a particular group has beenconsistent between laboratories (10).

Many of our studies have focused on thegenetic characterization of measles virusesassociated with cases and outbreaks in theUnited States during the last 10 years (2,11).These viruses can be separated into at least eightdistinct genetic groups (Table; Figure 2).Phylogenetic analyses using various computerprograms (15-17) indicated good statisticalsupport for each of the groups described below.Actually, more than eight genetic groups arelisted when viruses from groups not yet found inthe United States are included (e.g., Zambia:1993, Germany: 1992) (Figure 2). The number ofgenetic groups is likely to increase since the trueextent of genetic heterogeneity among wild-typemeasles viruses is still unknown, and virologic

surveillance has not been conducted or has onlyjust begun in many areas of the world.

Group 1 contains the prototype, Edmonstonstrain, which was isolated in 1954. This groupalso contains all vaccine viruses sequencedregardless of whether they were derived fromEdmonston (Attenuvax, Edmonston-Zagreb,AIKC, Schwarz) or from temporally andgeographically independent wild-type isolates(Shanghai-191: China, Changchun-47: China,CAM-70: Japan, Leningrad-16: Russia) (18).Relatively few wild-type viruses from theprevaccine era are available for molecularcharacterization. These viruses, which wereisolated in Japan, Russia, Finland, Romania, andthe United States during the 1950s and 1960s,are in group 1 (9). Therefore, while virusesbelonging to the other genetic groups may havebeen present, group 1 viruses must have hadwidespread distribution during the prevaccineera. Group 1 viruses continue to circulate, andviruses from group 1 were isolated from patientswith clinical measles in the United States, UnitedKingdom, Russia, China, and Argentina duringthe last 7 years (5,11,19,20, and unpub.observations). These recent group 1 wild-type

Table. Sources of genotypes isolated in the UnitedStates, 1995–1997

No. of Imports AlsoGroup isolates from circulating in4 13 Germany, Spain, Central Europe,

United Kingdom, Canada, Brazil, Austria, Brazil, United Italy, Greece, Kingdom Ukraine

3 9 Japan Japan, Thailand5 8 Italy, Germany Central Europe,

United King- dom, Brazil

7 3 Pakistan, Kenya South Africa, Canada

8 2 China, Vietnam China2 1 Philippines United States

1989–1992, Micronesia

1 1 Unknown United King- dom, Russia, China, Argentina

6 1 Kenya The Gambia, Cameroon, Gabon, Zambia


Synopses

viruses have several nucleotide substitutionsthat distinguish them from vaccine viruses. Incontrast, measles vaccine viruses reisolated fromimmunosuppressed patients with giant cellpneumonia had nucleotide sequences nearlyidentical to those of the vaccine virus found in thevaccine vial (unpub. observations). This suggeststhat vaccine viruses are very stable even afterprolonged replication in a human host. There-fore, it is unlikely that the group 1 wild-typeviruses represent laboratory contamination ofcultures with vaccine virus or reisolation of

vaccine virus from recently vaccinated persons.Sequence studies have failed to identify a distinctset of genetic markers that consistently differenti-ate wild-type and presumably virulent viruses fromattenuated viruses. Current studies are focusing onthe analysis of the noncoding regions of the viralgenome. More studies are needed to compareattenuated strains with their more virulent orreactogenic precursors. The recent development ofan infectious clone for measles (21) will, no doubt,contribute to those studies.

Group 2 viruses were associated with theresurgence of measles in the United Statesbetween 1989 to 1991, an epidemic that had anunusually high incidence of deaths and hospital-izations (Figure 1). The circulation of group 2viruses within the United States was interruptedin 1993, and this will be described in moredetail below. Among these viruses, the Illinois-1(Chicago-1) strain has become a representative ofrecent wild-type viruses, and almost the entiregenome has been sequenced. Group 2 viruses werefirst isolated in Japan during the early 1980s; morerecently, they were isolated in Japan, thePhilippines, and Micronesia (2,11-13,22).

The group referred to as group 3 viruses canactually be divided into two distinct groups with acommon geographic distribution. These viruseshave been isolated from outbreaks in Japan andThailand and from sporadic cases followingimportation into North America and Europe(2,11). Although virus from groups 2 and 3cocirculated in Japan during the late 1980s andearly 1990s, the group 3 viruses have recentlybecome the predominant genotype (23).

Groups 4 and 5 appear to be circulatingwidely in many countries in western Europe,particularly Germany, Spain, and the UnitedKingdom, where virologic surveillance has beenconducted (6,7,24). Viruses from this group arealso circulating in France, Italy, Austria, andGreece since they have been associated withmultiple importations from these areas into theUnited States (2,11; Table).

All representatives of the group 6 viruseshave been isolated in the African countries of TheGambia (4), Cameroon, Gabon, and Zambia orassociated with importations into the UnitedStates from Kenya. There is more geneticvariability within the group 6 viruses than amongmost of the other genetic groups, yet all group 6viruses contain a subset of nucleotide substitu-tions that places them on this African lineage.

Figure 2. Phylogenetic tree showing genetic relation-ships between the eight genetic groups of measlesvirus associated with U.S. outbreaks and cases since1988. The location and year of isolation is given foreach virus. Viruses not assigned to one of the eightgroups are labeled in brown. The unrooted tree isbased on the sequence of the protein coding region ofthe H gene (1854 nt). Wt-Edmonston = low passageseed of the original Edmonston isolate. SSPE =sequences obtained from cases of subacute sclerosingpanencephalitis.


Synopses

Relatively few viruses from central Africa, wheremost measles infections are occurring, have beenisolated for genetic analysis, and it will beinteresting to determine if other genotypes arealso present in this area.

Group 7 viruses were first isolated during anepidemic in Montreal, Canada, in 1988 (2,11).Group 7 was the predominant group among anumber of recently isolated viruses fromJohannesburg, South Africa (9,25,26). The identifi-cation of a group 7 virus in association with animportation to the United States from Pakistansuggests (11) that the viruses in this group may becirculating widely in Africa and Asia.

The group 8 viruses form a highly distinctgroup isolated in four provinces within thePeople’s Republic of China during the early 1990s(19). Like group 6, group 8 viruses have morenucleotide variability within the group (up to 3%)than the other groups. Recent evidence alsosuggests that group 8 viruses are circulating inother parts of China (Hong Kong) and in Vietnam.

Several recently isolated viruses do not fitinto the eight genetic groups that, so far, containmost recent isolates (Figure 2). Some outliersrepresent single isolations of a unique genetictype. However, preliminary analysis of a numberof wild-type viruses from Zambia indicates thatthese viruses belong to a genetic group that isdistinct from the eight groups described thus far(unpub. observations).

If viruses isolated during the early to mid-1980s were included into the genetic analysis (notshown), it would be apparent that more geneticgroups exist. However, viruses representingthese groups have not been isolated in the last 10to 15 years, and it must be assumed that thesegroups are circulating in restricted geographicregions, are circulating at such a low frequency asto escape surveillance, or are extinct.

A summary of genetic groups (Figure 3)represents a static picture that simply identifieswhere particular genetic groups have beenisolated, with no accounting for frequency ofisolation or source of the virus. Certain regions ofthe world (including much of Africa and most ofsouthern Asia) are still vastly underrepresented. Asurvey of recent Australian isolates is in progress.The pattern of genotypes in the United Kingdomand the United States is very complex because ofrelatively good strain surveillance and thefrequency of international travel to these locations.

Molecular EpidemiologyMeasles has long been considered one of the

most communicable of diseases. The resurgenceof disease from 1989 to 1991 in the United States(Figure 1) provides a good example of the rapidtransmissibility of the virus. During thisresurgence only group 2 viruses were isolated,and the sequences from these viruses were highlyrelated (2). With continued molecular surveil-lance, we were able to document the interruptionof transmission of the group 2 viruses andmonitor the change in measles genotypesassociated with outbreaks and sporadic casesfrom 1994 to the present. Molecular surveillanceof measles viruses was most useful when thechange in genotypes was observed over time.Without that information, it would not have beenpossible to describe the transition from anapparently “indigenous” lineage to importation ofmultiple lineages (Figure 4). This is in contrast tothe situation in South and Central America. Inthese areas, viral surveillance was not conductedbefore mass vaccination campaigns were initiated,so the identity of the prevailing genotype could notbe determined. Therefore, it is difficult to interpretthe genetic data obtained from viruses currentlycausing outbreaks in these regions.

The molecular surveillance of wild-typeviruses in the United States between 1989 and

Key to genotypes:

Group 3 Group 4Group 2

Group 5 Group 6

Group 1

Group 7 Group 8

Americas, 1988-1997 Europe and Asia, 1988-1997

Figure 3. Global distribution of measles geneticgroups. Colored circles indicate areas where measlesviruses from various genetic groups have beenisolated. Viruses not assigned to one of the eightgroups are labeled in brown.


Synopses

1997 provides the best example of dynamicmolecular surveillance (Figure 4). Virusesisolated over a 4-year period from majoroutbreaks in New York, Philadelphia, Chicago,Los Angeles, Houston, and southern Texas variedby less than 0.4% at the nucleotide level in the Hand N genes (2). Analysis of the few wild-typemeasles viruses isolated in the United Statesbefore 1988 indicates lineages other than group 2.This suggests that the group 2 viruses wereprobably imported during the late 1980s andwere rapidly transmitted to the entire country.While numerous importations occurred duringthe resurgence, apparently these viruses did notcirculate widely enough to be detected bymolecular surveillance. Perhaps the number ofmeasles-susceptible persons in the U.S. popu-lation during the resurgence was high enoughto sustain continuous transmission withoutaccumulation of variants or displacement byother imported viruses.

More aggressive childhood vaccination pro-grams, the introduction of a two-dose schedule,

and successful mass vaccination campaignsconducted by the Pan American Health Organi-zation in South and Central America greatlyreduced the number of reported measles cases inthe United States in 1993 (Figure 1). During a 6-week period at the end of 1993, no indigenouscases of measles were reported (27). Molecularsurveillance of measles viruses associated withcases and outbreaks in the United States during1994, 1995, 1996, and 1997 documented thisinterruption of transmission of what had been theindigenous genotype (2,11). Only one group 2virus was detected in the United States after1993, and this was directly linked to importationfrom the Philippines (11).

Molecular surveillance data allow us to drawseveral conclusions about the transmission ofmeasles virus in the United States. The first isthat increasing the level of population immunityby vaccination can interrupt the transmission ofmeasles virus. This is hardly new information,and interruption of transmission was describedfor The Gambia in 1983 and more recently in

Figure 4. Change in genetic groups of measles viruses associated with U.S. cases and outbreaks between 1988 andSeptember 1997. Arrows indicate sources of virus, if known.


Synopses

Finland. However, our studies are the first inwhich genetic analysis of measles strains hasbeen used to document interruption oftransmission. Secondly, long-term asymptom-atic transmission of virus is unlikely since nogroup 2 viruses were detected in the UnitedStates after 1993 that were not directly linkedto importation. Finally, measles will not befully controlled anywhere until it is controlledglobally. Virus introduced by importation willcontinue to fuel sporadic outbreaks andepidemics even in areas with relatively goodcontrol measures. These observations shouldstrengthen our resolve to accelerate measlescontrol activities on a global level.

The molecular data imply that underconditions of continuous indigenous transmissionof measles virus, the number of circulatinggenotypes is limited. As population immunityincreases, the pattern of genotypes becomes morecomplex to reflect the multiple sources ofimported virus. We hope to test this modelfurther by conducting molecular surveillance ofwild-type measles viruses circulating in areasthat still have endemic measles.

ConclusionsGenetic characterization of wild-type measles

viruses provides a valuable means to measure thelevel of virus circulation in areas just beginningto implement measles control plans. In areas thatalready achieved good measles control, molecularepidemiologic studies provide a means todescribe outbreaks and cases. Identifying thesource of the virus can lead to improved controlmeasures. To be maximally effective, molecularepidemiologic studies must include surveys ofviral genetic groups from all areas of the world.Specimens for viral isolations should be obtainedfrom as many chains of transmission as possible.Obtaining specimens must become an integralpart of measles surveillance and be included inthe standard operating procedures for investigat-ing measles cases. If we can establish a largedatabase to describe the indigenous geneticgroups before large-scale control measures areenacted, we can closely monitor the ability ofthese control measures to reduce or interrupttransmission of measles. Molecular epidemiologywill greatly enhance measles elimination anderadication efforts.

Dr. Bellini is chief of the Measles Section,Respiratory and Enteric Viruses Branch, Division ofViral and Rickettsial Diseases, National Center forInfectious Diseases, CDC. Dr. Rota is a microbiologistin the Measles Section, CDC. The authors work closelywith national and international public healthagencies to support global measles eradicationactivities. They provide expertise and training on thegenetic characterization of measles viruses andlaboratory diagnosis of measles infections. Researchinterests include molecular virology, viralpathogenesis, and immune response to viral infections.

References 1. Atkinson WL, Orenstein WA, Krugman S. The

resurgence of measles in the United States, 1989-1990.Annu Rev Med 1992;43:451-63

2. Rota JS, Heath JL, Rota PA, King GE, Celma ML,Carabaña J, et al. Molecular epidemiology of measlesvirus: identification of pathways of transmission andimplications for measles elimination. J Infect Dis1996;173:32-7.

3. Kreiss S, Whistler T. Rapid identification of measlesvirus strains by the heteroduplex mobility assay. VirusRes 1997;47:197-203.

4. Outlaw MC, Jaye A, Whittle H, Pringle C. Clustering ofhemagglutinin sequences of measles viruses isolated inthe Gambia. Virus Res 197;48:125-31.

5. Outlaw MC, Pringle C. Sequence variation within anoutbreak of measles virus in the Coventry area duringspring/summer 1993. Virus Res 1995;39:3-11.

6. Rima BK, Earle JAP, Yeo RP, Herlihy L, Baczko K, terMeulen V, et al. Temporal and geographicaldistribution of measles virus genotypes. J Gen Virol1995;76:1173-80.

7. Rima BK, Earle JAP, Baczko K, ter Meulen V, LiebertU, Carstens C, et al. Sequence divergence of measlesvirus haemagglutinin during natural evolution andadaptation to cell culture. J Gen Virol 1997;78:97-106.

8. Rima B, Earle JAP, Baczko K, Rota PA, Bellini WJ.Measles virus strain variations. Current TopicsMicrobiol Immunol 1995;191:65-84.

9. Rota PA, Bloom AE, Vanchiere JA, Bellini WJ.Evolution of the nucleoprotein and matrix genes ofwild-type strains of measles virus isolated from recentepidemics. Virology 1994;198:724-30.

10. Rota PA, Rota JS, Bellini WJ. Molecular epidemiologyof measles virus. Seminars in Virology 1995;6:379-86.

11. Rota JS, Rota PA, Redd SC, Pattamadilok S, BelliniWJ. Phylogenetic analysis of measles viruses isolatedin the United States 1995-1996. J Infect Dis. In press1997.

12. Saito H, Sato H, Abe M, Harata S, Amano K, Suto T, etal. Cloning and characterization of the cDNA encodingthe HA protein of a hemagglutination-defectivemeasles virus strain. Virus Genes 1994;8;2:107-13.

13. Sakata H, Kobune F, Sato TA, Tanabayashi K, YamadaA, Sugiura A. Variation in field isolates of measlesvirus during an 8-year period in Japan. MicrobiolImmunol 1993;37:233-7.


Synopses

14. Taylor MJ, Godfrey E, Baczko K, ter Meulen V, WildTF, Rima BK. Identification of several differentlineages of measles virus. J Gen Virol 1991;72:83-8

15. Devereaux J, Haeberli P, Smithies O. A comprehensiveset of sequence analysis programs for the VAX. NucleicAcids Res 1984;12:387-95.

16. Felsenstein J. Phylogenies from molecular sequences;inferences and reliability. American Review ofGenetics 1988;22:512-65.

17. Swofford DL. PAUP: phylogenetic analysis usingparsimony [computer program]. Version 3.1.1.Champaign (IL): Illinois Natural History Survey; 1991.

18. Rota JS, Wang ZD, Rota PA, Bellini WJ. Comparison ofsequences of the H, F, and N coding genes of measlesvirus vaccine strains. Virus Res 1994;31:317-30.

19. Xu WB, Tamin A, Rota JS, LiBi J, Bellini WJ, Rota PA.New genetic group of measles virus isolated in thePeople’s Republic of China. Virus Res. In press 1998.

20. Heider A, Santibanez S, Tischer A, Gerike E,Tikhonova N, Ignatyev G, et al. Comparativeinvestigation of the long non-coding M-F genomeregion of wild-type and vaccine measles viruses. ArchVirol. In press 1998.

21. Radecke F, Speilhofer P, Schneider H, Kaelin K, HuberM, Dotsch C, et al. Rescue of measles virus from clonedDNA. EMBO J 1995;14:5773-84.

22. Guris D, Auerbach S, Vitek C, Maes E, McCready J,Durand M, et al. Measles outbreaks in Micronesia,1991-1994. J Pediatr. In press 1997.

23. Katayama Y, Shibahara K, Kohama T, Homma M,Hotta H. Molecular epidemiology and changingdistribution of genotypes of measles virus field strainsin Japan. J Clin Microbiol 1997;35:2651-3.

24. Jin L, Brown DWG, Ramsay MEB, Rota PA, BelliniWJ. The diversity of measles virus in the UK, 1992-1995. J Gen Virol 1997;78:1287-94.

25. Kreis S, Whistler T. Rapid identification of measlesvirus strains by the heteroduplex mobility assay. VirusRes 1997;47:197-203.

26. Kreis S, Vardas E, Whistler T. Sequence analysis of thenucleocapsid gene of measles virus isolates from SouthAfrica identifies a new genotype. J Gen Virol1997;78:1581-7.

27. Centers for Disease Control and Prevention. Absenceof reported measles—United States, November 1993.MMWR Morb Mortal Wkly Rep 1993;42(48):925-6.


Synopses

Enterococci, which have been known as acause of infective endocarditis for close to acentury, more recently have been recognized as acause of nosocomial infection and “superinfec-tion” in patients receiving antimicrobial agents(1). The enterococcus is now receiving increasedattention because of its resistance to multipleantimicrobial drugs, which probably explains alarge part of its prominence in nosocomialinfections. The most common enterococci-associated nosocomial infections are infections ofthe urinary tract, followed by surgical woundinfections and bacteremia (1-3). Enterococci areoften present in intraabdominal and pelvicinfections, although not all patients with suchinfections require specific antienterococcaltherapy. Other enterococcal infections includeinfections (including meningitis and bacteremia)in very ill neonates; central nervous systeminfections in adults, typically with a history ofcentral nervous system surgery or intrathecalchemotherapy; and rarely, osteomyelitis andpulmonary infections. Enterococci frequentlyarise from colonization of indwelling T tubes,

causing liver or biliary infection in livertransplant patients (1).

Antimicrobial ResistanceMost enterococci have naturally occurring or

inherent resistance to various drugs, includingcephalosporins and the semisynthetic penicillinase-resistant penicillins (e.g., oxacillin) and clinicallyachievable concentrations of clindamycin andaminoglycosides. Compared with streptococci,most enterococci are relatively resistant topenicillin, ampicillin, and the ureidopenicillins,with MICs of 1 µg/ml to 8 µg/ml for mostEnterococcus faecalis and even higher for mostE. faecium. Many enterococci are also tolerantto the killing effects of cell-wall active agents,including ampicillin and vancomycin; recent datasuggest that this property may not be inherent,but rather acquired after exposure to antibiotics(4). Inherent in vivo resistance of E. faecalis totrimethoprim-sulfamethoxazole may explain thelack of efficacy in animal models. In vitro,trimethoprim-sulfamethoxazole readily inhibitsmost enterococci at low concentrations, but thisactivity is lessened by exogenous folates (5).Moreover, bactericidal activity against E. faecalisseems unreliable and very method dependent (6).In animal models, this combination has notshown good activity and is not generally accepted

Diversity amongMultidrug-Resistant Enterococci

Barbara E. Murray, M.D.University of Texas Houston-Medical School, Houston, Texas, USA

Address for correspondence: Barbara E. Murray, Division ofInfectious Diseases, University of Texas Houston-MedicalSchool, 6431 Fannin, 1.728 JFB, Houston, TX 77030, USA;fax: 713-500-6761; e-mail: [email protected].

Enterococci are associated with both community- and hospital-acquired infections.Even though they do not cause severe systemic inflammatory responses, such asseptic shock, enterococci present a therapeutic challenge because of their resistanceto a vast array of antimicrobial drugs, including cell-wall active agents, all commerciallyavailable aminoglycosides, penicillin and ampicillin, and vancomycin. The combinationof the latter two occurs disproportionately in strains resistant to many otherantimicrobial drugs. The propensity of enterococci to acquire resistance may relate totheir ability to participate in various forms of conjugation, which can result in the spreadof genes as part of conjugative transposons, pheromone-responsive plasmids, orbroad host-range plasmids. Enterococcal hardiness likely adds to resistance byfacilitating survival in the environment of a multidrug-resistant clone, thus enhancingpotential spread from person to person. The combination of these attributes within thegenus Enterococcus suggests that these bacteria and their resistance to antimicrobialdrugs will continue to pose a challenge.


Synopses

as an effective antienterococcal therapy, espe-cially for systemic infections (7,8).

In addition to natural resistance to manyagents, enterococci have also developed plasmid-and transposon-mediated resistance to tetracy-cline (as well as minocycline and doxycycline),erythromycin (plus the newer compoundsazithromycin and clarithromycin), chloram-phenicol, high levels of trimethoprim, and highlevels of clindamycin.

The propensity of E. faecalis to acquiremultiple antibiotic-resistance traits may resultfrom a variety of distinctly different mechanismsfor conjugation, i.e., bacterial mating. The beststudied system of conjugation involvesoligopeptides called pheromones and pheromone-responsive plasmids (9; Figure 1). Briefly, strainsof E. faecalis typically secrete into the culturemedium a number of different small peptide sexpheromones specific for different types ofplasmids. When a cell containing a pheromone-responsive plasmid (the potential donor cell)comes into contact with its correspondingpheromone, transcription of a gene on theplasmid is turned on, resulting in the synthesis ofa sticky substance (called aggregation substance)on its surface. When the donor cell bumps intoanother E. faecalis, aggregation substance,which contains two Arg-Gly-Asp motifs, sticks tothe binding substance on the surface of mostE. faecalis cells, causing them to clump together.In the test tube, clumps of cells actually fall to the

bottom of the tube, resulting in a visibleaggregate. By a process not yet well understood,the pheromone-responsive plasmid can thentransfer from the donor bacterium to the other(recipient) bacterium. Once the recipient cell hasacquired this particular plasmid, the synthesis ofthe corresponding sex pheromone is shut off toprevent self-clumping. This system of conjuga-tion, which occurs primarily in E. faecalis, ishighly efficient and results in transfer ofplasmids in both filter and broth matings.

Another system of conjugation, also not wellunderstood, involves broad host-range plasmidsthat can transfer among species of enterococciand other gram-positive organisms such asstreptococci and staphylococci (10). The transferfrequency is generally much lower than with thepheromone system and is much more efficientwith filter than with broth matings. Sincestaphylococci, streptococci, and enterococci sharea number of resistance genes, these broad host-range plasmids may be a mechanism by whichsome of these resistance genes have spreadamong different genera.

A third type of conjugation, which involvesconjugative transposons, may also explain thespread of resistance genes to many differentspecies (11). As opposed to ordinary transposons,which can jump within a cell from one DNAlocation to another, conjugative transposons alsoencode the ability to bring about conjugationbetween different bacterial cells. Since plasmidstypically require rather complex machinery forreplication (often depending on successfulinteractions with host proteins) and must faceadditional problems of surface exclusion andincompatibility, conjugative transposons (whichdo not replicate, but instead insert into thechromosome or into a plasmid of the new host)appear to be an even more efficient and far-reaching way of disseminating a resistance gene.This may explain why the tetM gene of theconjugative transposon Tn916 has spread beyondthe gram-positive species into gram-negativeorganisms, including gonococci, meningococci,and Haemophilus ducreyi, as well as intomycoplasma and ureaplasma, among others(12,13). Other resistance genes, including thoseencoding resistance to erythromycin and kana-mycin, are also found on conjugative transposons;these frequently contain or are related to Tn916.Such transposons may have evolved from aTn916 ancestor; their emergence suggests the

Figure 1. Enterococcus faecalis pheromone-responsiveconjugative system.Pheromone A released from the potential recipient cell(right) interacts with plasmid A in the potential donorcell (left) to induce synthesis of aggregation substance.Attachment of aggregation substance to bindingsubstance causes the cells to clump into visibleaggregates. Once the pheromone-responsive plasmidA has transferred from donor to recipient cell,synthesis of pheromone A is shut off.


Synopses

possibility of further dissemination of resistanceamong gram-positive organisms. Particularlyominous are reports of the vanB gene clusterwithin large conjugative chromosomal elementsthat appear similar, at least in function, toconjugative transposons (14).

High-Level Aminoglycoside ResistanceAlthough some acquired resistance of

enterococci is not clinically important becausethe agents involved are not commonly used, otherresistance greatly affects enterococcal therapy;high-level resistance (HLR) to aminoglycosides isan example. This resistance is added onto thenormal low-level resistance of enterococci toaminoglycosides and typically results in MICs of> 2,000 µg/ml. This degree of resistance predicts,without exception, resistance to synergism betweencell-wall active agents and the aminoglycoside towhich the organism is highly resistant (1). High-level aminoglycoside resistance is most often dueto aminoglycoside-modifying enzymes; HLR tostreptomycin can also be ribosomal, that is, due toa mutation that results in ribosomes resistant tostreptomycin inhibition. HLR to kanamycin(without gentamicin) is a fairly common trait andis due to the production of a 3'-phosphotransferase,APH(3')-III. This enzyme is important because italso eliminates synergism between cell-wallactive agents and amikacin (through phosphory-lation of the 3'-hydroxyl group), although it doesnot necessarily confer HLR to amikacin. HLR togentamicin results from the bifunctional protein(AAC(6')-I/APH(2")-I), encoded by a single genewith two active sites, one with 6'-acetyltransferaseactivity and the other, 2"-phosphotransferaseactivity (15). The combination of these activitiesresults in HLR or resistance to synergism for allcommercially available aminoglycosides exceptstreptomycin, which is not modified by thisenzyme. However, HLR to streptomycin (due toeither ribosomal resistance or a streptomycinadenylyltransferase) is also common and cancoexist with the gene(s) for HLR to otheraminoglycosides. Spectinomycin is also notmodified by the bifunctional enzyme, but thisagent, which is not a true aminoglycoside, is notgenerally bactericidal against enterococci anddoes not appear to show synergism with cell-wall active agents.

Strains of enterococci from patients withendocarditis and other serious infections forwhom combination therapy is desired should be

screened for HLR to streptomycin and gentami-cin. HLR screening for tobramycin is notgenerally performed or advisable It could, inprinciple, be used for E. faecalis, but E. faeciumisolates have a chromosomally encoded, natu-rally occurring gene for a 6'-acetyltransferasethat eliminates synergism with tobramycin,although it does not cause HLR (MICs aretypically 128-500 µg/ml); a probe for this gene hasbeen used to confirm the identification of E.faecium isolates to species. In addition, HLR of anE. faecium isolate (without HLR to gentamicin) totobramycin, due to an adenylyltransferase, wasrecently described (16). Therefore, the use oftobramycin for possible synergism in seriousenterococcal infections would need to be precededby screening for HLR to tobramycin (a test that isnot commonly available), as well as for identifica-tion to species, neither of which is practical.

More recently, veterinary and human isolatesresistant to moderate levels of gentamicin (256 µg/ml) were found to have a new gentamicin-modifyingenzyme encoded by a gene designated aph(2")-Ic(17). This gene conferred resistance to synergismbetween gentamicin and cell-wall active agents andmay be less easily detected than strains producingthe bifunctional enzyme.

Beta-Lactamase– and non-Beta-Lactamase–Associated Penicillin Resistance

The first known penicillinase-producingisolate of enterococcus was an isolate of E.faecalis recovered from a patient in Houston,Texas, in 1981 (18). Although rare, these isolateshave been reported from the United States(Texas, Florida, North Carolina, Delaware,Pennsylvania, New York, Massachusetts, andConnecticut), Lebanon, Canada, and Argentina(19). Like other enterococci, beta-lactamase-producing strains have been found ascolonizers, as in the large “outbreak” ofcolonization in a Boston children’s hospital, butthey have also been associated with trueinfections, as was demonstrated by cases at aVirginia Veterans Administration hospital, byisolates from Argentina, and in other reports(20,21). The enterococcal penicillinase gene,identical to the gene encoding staphylococcaltype A penicillinase, almost always occurs instrains with HLR to gentamicin and is oftenfound on a transferable plasmid that alsocontains aph(2")-Ia/aac(6')-Ie. The relatively lowlevels of beta-lactamase produced by enterococci


Synopses

result in a marked inoculum effect with thesestrains so that at low and even moderate inocula(103-105 CFU/ml), penicillinase-producing en-terococci usually appear no more resistant thanother enterococci, while at high inocula (> 107

CFU/ml), these organisms are usually highlyresistant to penicillin, ampicillin, andureidopenicillins. The activity of the penicillinaseis reversed by the beta-lactamase inhibitorsclavulanate, sulbactam, and tazobactam; inanimal models of endocarditis, beta-lactamaseinhibitors have been shown to markedly enhancethe therapeutic efficacy of ampicillin or penicil-lin. In the clinical laboratory, penicillinase-producing enterococci are generally not detectedby routine laboratory susceptibility testing,such as MICs or disk diffusion. For this reason,if a penicillin is to be used for therapy,enterococcal isolates from patients withendocarditis or other serious infections shouldbe tested for penicillinase production by using aspecific beta-lactamase test such as thechromogenic cephalosporin nitrocefin.

Nonpenicillinase-producing, penicillin-resistantenterococci have been reported for decades andusually are E. faecium. Until recently, MICs ofpenicillin typically ranged from 8 µg/ml to 64 µg/ml,with an occasional isolate having higher levels ofresistance. However, increasingly, strains withmuch higher levels of penicillin resistance havebeen reported (22). Whether a large number ofstrains have converted from low-level to high-levelresistance or a more limited number of strains havebeen disseminated is unclear. The mechanismsinvolved in this resistance are overproduction of alow-affinity penicillin-binding protein (a cell-wallsynthesis enzyme) and a further decrease in theaffinity of one of these enzymes for penicillin (23).As a possible explanation for why manyvancomycin-resistant E. faecium also have veryhigh levels of resistance to ampicillin, Rice andcolleagues (24) showed that transfer of vancomycinresistance from one strain to another was linked totransfer of ampicillin resistance.

Vancomycin ResistanceMost surprising in recent years has been the

emergence among enterococci of acquired resis-tance to vancomycin. Vancomycin had been inclinical use since the 1950s, although it was notheavily used until the late 1970s and particularlythe 1980s. Because multiple genes are involved ingenerating vancomycin resistance, the development

of resistance was neither easy nor recent. Threephenotypes of vancomycin resistance (types A, B,and C) are now well described; a fourth, type D,has been recently reported (25). VanA-typestrains are typically highly resistant to vancomy-cin and moderately to highly resistant toteicoplanin. This phenotype is often plasmid ortransposon mediated and is inducible (i.e.,exposure of bacteria to vancomycin results in theinduction of the synthesis of several proteins thattogether confer resistance) (26).

In vancomycin-susceptible enterococci, D-alanyl-D-alanine (formed by an endogenous D-alanine-D-alanine ligase) is added to a tripeptideprecursor to form a pentapeptide precursor. TheD-Ala-D-Ala terminus is the target of vancomy-cin; once vancomycin has bound, the use of thispentapeptide precursor for further cell-wallsynthesis is prevented. In the VanA phenotype,one of the proteins whose synthesis is induced byexposure of bacterial cells to vancomycin is calledVanA; VanA is a ligase and resembles the D-alanine-D-alanine ligase from Escherichia coliand other organisms, including vancomycin-susceptible enterococci (27). VanA generates D-Ala-D-X, where X is usually lactate; theformation of D-lactate is due to the presence ofVanH, a dehydrogenase encoded by vanH. Thedepsipeptide moiety, D-Ala-D-Lac, is then addedto a tripeptide precursor, resulting in adepsipentapeptide precursor. Vancomycin doesnot bind to the D-Ala-D-Lac terminus, so thisdepsipentapeptide can be used in the remainingsteps of cell-wall synthesis. However, when thenormal pentapeptide precursor ending in D-Ala-D-Ala is also present, cells are not fullyvancomycin resistant, despite the presence of D-Ala-D-Lac containing precursors. This apparentproblem is taken care of in large part by vanX,which encodes a dipeptidase, VanX, that cleavesD-Ala-D-Ala, preventing its addition to thetripeptide precursor. Should any D-Ala-D-Alaescape cleavage and result in a normalpentapeptide precursor, vanY encodes an ancil-lary or back-up function. That is, it codes for acarboxypeptidase, VanY, which cleaves D-alanine and D-lactate from D-Ala-D-Ala and D-Ala-D-Lac termini, respectively, resulting intetrapeptide precursors, to which vancomycindoes not bind. The other genes involved in theVanA resistance complex include vanR and vanS,whose encoded proteins are involved in somehowsensing the presence of extracellular vancomycin or


Synopses

its effect and signaling intracellularly to activatetranscription of vanH, vanA, and vanX (27). A finalgene in the vanA cluster is vanZ, which encodesVanZ, the role of which is not known.

VanB, encoded by vanB in the vanB genecluster, is also a ligase that stimulates theformation of D-Ala-D-Lac. The VanB phenotypeis typically associated with moderate to highlevels of vancomycin resistance but is withoutresistance to teicoplanin. This is explained by theobservation that vancomycin, but not teicoplanin,can induce the synthesis of VanB and of VanHBand VanXB. However, because mutants resistantto teicoplanin can readily be selected from VanBstrains on teicoplanin-containing agar, clinicalresistance would likely occur among VanB strains ifteicoplanin were widely used. Most of the proteinsencoded by the vanA gene cluster have homologuesencoded by the vanB gene cluster, except for VanZ.The vanB gene cluster has an additional gene,vanW, of unknown function.

The VanC phenotype (low-level resistance tovancomycin, susceptible to teicoplanin) is aninherent (naturally occurring) property of E.gallinarum and E. casseliflavus. This property isnot transferable and is related to the presence ofspecies-specific genes vanC-1 and vanC-2,respectively (28); a third possible species, E.flavescens and its gene vanC-3, are so closelyrelated to E. casseliflavus and vanC-2 thatdifferent names are probably not warranted (29).These species appear to have two ligases; the cell-wall pentapeptide, at least in E. gallinarum, endsin a mix of D-Ala-D-Ala and D-Ala-D-Ser (29,30).The genes vanC-1 and vanC-2 apparently lead tothe formation of D-Ala-D-Ser containing cell-wallprecursors, while D-Ala-D-Ala ligases, alsopresent in these organisms, result in D-Ala-D-Ala. The presence of both D-Ala-D-Ala and D-Ala-D-Ser precursors may explain why many isolatesof these species test susceptible to vancomycinand why even those isolates with decreasedsusceptibility display only low-level resistance.

VanD-type glycopeptide resistance has beenrecently described in an E. faecium isolate fromthe United States (25). The organism wasconstitutively resistant to vancomycin (MIC ≥≥≥≥≥ 64µg/ml) and to low levels (4 µg/ml) of teicoplanin.Following polymerase chain reaction amplifica-tion with primers that amplify many D-Ala-D-Alaligases, a 605-bp fragment was identified whosededuced amino acid sequence showed 69% identityto VanA and VanB and 43% identify to VanC.

Molecular Epidemiology of NewerResistance Traits

High-Level Gentamicin ResistanceThe DNA sequence of the gene encoding HLR

to gentamicin in E. faecalis is the same as thesequence of the gentamicin resistance gene ofstaphylococci (15). Since this gene was wellestablished in staphylococci by the 1970s butHLR to gentamicin was not reported inenterococci until 1979, the seemingly obviousconclusion is that this gene spread fromstaphylococci to enterococci rather than viceversa or, at least, staphylococci acquired it first.However, the disk diffusion method used in the1970s and microtiter dilution MICs done later arecapable of detecting gentamicin resistance instaphylococci but do not distinguish enterococciwith high-level gentamicin resistance from thosewith low-level, inherent resistance. Therefore,since laboratories were not screening enterococciby special techniques for high-level gentamicinresistance, it cannot be definitively stated thatthis resistance did not appear in enterococciearlier or coincident with its emergence instaphylococci. However, several observationssupport the likelihood that gentamicin resistanceappeared and disseminated in staphylococcibefore it did in enterococci. In 1971, Moellering etal. reported the lack of high-level gentamicinresistance among enterococci (31).Watanakunakorn reported the absence of high-level gentamicin resistance among 126 entero-cocci from 1980 to 1984, with HLR subsequentlyappearing in 1985 (32). Phillips et al. from theUnited Kingdom reported no highly gentamicin-resistant strains in 1969 to 1979 or 1980 to 1985and appearance of strains in 1986 (33). Zervos etal. reported that only one (0.04%) of 269 isolatesof E. faecalis had high-level gentamicin resis-tance in 1981; this figure gradually increased to7.7% in 1984 (34). High-level gentamicinresistance in E. faecium appears to have occurredafter its appearance in E. faecalis, with the firstreport occurring in 1988 (35).

Delineation of the molecular epidemiology ofstrains of enterococci was limited in the past bythe lack of an easy, reliable, and widely accessiblemethod for subspecies strain differentiation.Zervos and Schaberg reported the use of plasmidpatterns in enterococci to suggest the intrahospitalspread of strains with high-level gentamicinresistance. Pulsed-field gel electrophoresis (PFGE)


Synopses

of E. faecalis with HLR to gentamicin found thatdifferent isolates from both the same anddifferent locations had markedly differentrestriction endonuclease digestion patterns. Thatis, it found no evidence of a common strain orstrains that predominated among gentamicin-resistant organisms (36). Strains isolated be-tween 1981 and 1984 at the University ofMichigan demonstrated that plasmids encodinghigh-level gentamicin resistance were heteroge-neous, which again argues against clonaldissemination of a limited number of strains orplasmids to account for spread of this property(34,37). We have subsequently shown thatgentamicin resistance in enterococci can beencoded on a transposon identical to that instaphylococci (38). In addition, the enterococcalgentamicin-resistance gene has been found inother genetic settings, one of which has also beenfound in North American Staphylococcus aureuswith gentamicin resistance (39). Since allenterococci with HLR to gentamicin (MIC > 2,000µg/ml) that have been tested have hybridizedwith the same gene probe, this property could betermed a “gene epidemic.” However, by the timegentamicin resistance was discovered in entero-cocci, this gene was already widespread with noevidence of either a common plasmid or acommon or predominant strain.

Other genetic elements encoding HLR togentamicin have also been described. Thal et al.have described a 27-kb element designatedTn924 that encodes HLR to gentamicin and couldbe mobilized from the chromosome of an E.faecalis by a coresident plasmid (40). Rice et al.have described a large (ca. 60 kb) transferableelement, tentatively named Tn5385, whichappears to contain within it an 18-kb conjugativetransposon (Tn5381) encoding tetracycline resis-tance and a 26-kb IS256-based compositetransposon (Tn5384) encoding resistance togentamicin and to erythromycin (41).

Penicillin-Resistant Penicillinase-ProducingEnterococci

PFGE analyses of penicillinase-producingenterococci have shown that a common penicilli-nase-producing strain (or “clone”), defined ashaving an identical or related chromosomaldigestion pattern, was present in Texas, Florida,North Carolina (unpub. observation), Delaware,Pennsylvania, and Virginia, which had a largeoutbreak with numerous infections (42,43).

Moreover, at each of these locations, all isolates ofpenicillinase-producing E. faecalis examinedwere derivatives of this strain; in the hospital inwhich this strain has been endemic for manyyears, a single penicillinase-producing isolate ofE. faecium, the only such isolate ever reported,was also found (44). Among numerous non-Bla+E. faecalis, a PFGE pattern similar to that of Bla+E. faecalis has been found only once from anisolate from a Philadelphia hospital that had oneof the Bla+ isolates (43 and unpub. obs.).Penicillinase-producing isolates from Connecti-cut, Boston, Canada (unpub. observations),Lebanon, and Argentina represent differentstrains (19). However, clonal relatedness of threeisolates with an almost identical pattern wasdemonstrated in Connecticut (despite one ofthese lacking HLR to gentamicin) (45), and all sixisolates in Argentina appeared to be a singlestrain. The Boston isolates have not been studiedby PFGE, but a number of isolates from the samehospital had a single shared plasmid, suggestingthat these also represent clonal dissemination ofa single strain. While all of these penicillinase-producing E. faecalis could be detected bynitrocefin, a vancomycin-resistant E. faecalisthat tested negative for penicillinase bynitrocefin but showed a marked inoculum effectwith penicillin and destroyed penicillin in abioassay has been reported (46). Thus, in contrastto high-level gentamicin resistance in entero-cocci, which appeared to be widely disseminatedon different plasmids and in different strains bythe time this phenotype was studied, penicilli-nase production in enterococci is still largelyassociated with a limited number of strains;moreover, in locations known to have morethan one isolate, oligoclonal spread within eachsetting remains the rule.

Nonpenicillinase-Producing Penicillin-Resistant Strains

PFGE analyses of highly penicillin-resistantE. faecium from the Medical Colleges of Virginiaand Pennsylvania have shown that within eachlocation, most highly resistant strains repre-sented a single clone (47). Analysis of the PFGEpatterns also raised the possibility, on the basis ofsimilarities between patterns of isolates in thesetwo different locations, that a strain may havespread from one institution to the other (47); thisconclusion was supported by the finding thatthese isolates belonged to the same multilocus


Synopses

enzyme cluster (unpub. observations). Circum-stantial evidence of intrahospital spread of highlypenicillin-resistant E. faecium also comes fromthe study of Grayson et al. (22), who reported asudden increase in more highly penicillin-resistant isolates of E. faecium at the Massachu-setts General Hospital in 1988. Although nogenetic analysis was done, the fact thatgentamicin resistance also simultaneously ap-peared and most of the E. faecium highlyresistant to penicillin were also highly resistantto gentamicin suggests clonal dissemination ofone or a few strains within that hospital (22).

Vancomycin ResistanceVancomycin resistance in enterococci is

heterogeneous on many levels. For example,three different, well-described types of vancomy-cin resistance are known, each associated withdifferent ligase genes, (vanA, vanB, vanC1, andvanC2), and a fourth type, VanD, has beenreported recently (25). VanA and VanB typeresistance is encoded by gene clusters that areacquired (i.e., not part of the normal genome ofenterococci) and are often transferable. In contrast,vanC1 and vanC2 are normally occurring genesthat are endogenous species characteristics of E.gallinarum and E. casseliflavus, respectively, andare not transferable. The acquired gene clustersassociated with vanA and vanB are found indifferent genetic surroundings. The vanA genecluster has been found in a small Tn3-liketransposon, Tn1546, and in elements that appear tobe closely related (e.g., Tn5488, which has aninsertion sequence [IS1251] within Tn1546[48,49]) or lacking vanZ (50). These elementshave in turn been found on both transferable andnontransferable plasmids, as well as on thechromosome of the host strain. VanB typeresistance was initially not found to be transferable,but at least in some instances, the vanB gene clusterhas been found on large (90 kb to 250 kb)chromosomally located transferable elements, oneof which contains within it a 64-kb compositetransposon (Tn1547) (Figure 2; 14). More recently,vanB has been found as part of plasmids.

In addition to being found in different geneticsurroundings, the vanA and vanB gene clustershave also been found in a number of differentbacterial species. vanA has been found in multipleenterococcal species as well as in lactococci,Orskovia, and Arcanobacteria (51). The distributionof the vanB gene cluster seems somewhat more

restricted, having been found primarily in E.faecium and E. faecalis, although it has recentlybeen found in Streptococcus bovis (52).

When vancomycin-resistant enterococci (VRE)from patients in a given hospital have beenexamined, particularly after the first recovery ofVRE, evidence is often found of a single orpredominant strain (53-58). Finding isolates withidentical or highly related PFGE patterns indifferent hospitals indicates interhospital clonaltransmission (59-61). Some reports do not find asingle or a predominant strain, especially whenVRE have been present in a hospital or area forsome time (62,63). This was also true of tworeports from France in which all of 16 and all of 24vancomycin-resistant E. faecium were different(50,64,65). The reports from France likely reflectanother observation regarding the diversity of

Figure 2. Potential modes of spread of vancomycin-resistant genes. Adapted in part from Quintiliani andCourvalin (14).The vanB gene cluster (shown on the left) on a 64-kbtransposon is part of a 250-kb mobile element shownto move from the chromosome of one enterococcus andinsert into the chromosome of another. Although notdemonstrated, circularization of the vanB containinglarge mobile elements resembles the mechanismdescribed for conjugative transposons that can excisefrom the chromosome of one strain, circularize,transfer from one enterococcus to another, andreinsert into the chromosome of the recipient (such asthe one on the right). The 64-kb transposon shown onthe left can also jump to another plasmid within thehost enterococcus. If it is a conjugative (Tra+) plasmid,that plasmid can then transfer by conjugation to otherbacteria, taking the van resistance genes with it. Inone instance, the vancomycin resistance transposonwas shown to transpose to a plasmid encoding thevirulence factor hemolysin (Hly).


Synopses

vancomycin resistance. vanA has also beenshown to be present in normal fecal enterococci ofhealthy, nonhospitalized persons in differentparts of Europe (66-68); in one study, 20 differentstrains (identified by PFGE) of vancomycin-resistant E. faecium were found in the fecal floraof 17 persons in two areas in Belgium (68). vanAcontaining E. faecium have also been found in thefeces of healthy animals as well as from animalproducts in Europe (50,67,69-71); in one study,VRE were found in the feces of healthy meat-eaters but not vegetarians (72). VRE have not,however, been found as part of the normal fecalflora in the United States (73) possibly becauseglycopeptides, often used in animal feed inEurope, are not used in the United States. Forexample, 24,000 kg annually of the glycopeptideavoparcin were reportedly used in recent years inDenmark (74). Reports of VRE in the feces ofanimals on farms using glycopeptides, but not inthose without such use, support this hypothesis(71,75). While oral glycopeptide use markedlyincreases the numbers of VRE per gram of stool inhumans (68) and by analogy, presumably does soin animals, glycopeptide use does not explain theorigin of these gene clusters.

The problem of multidrug-resistant entero-cocci promises to be with us for the foreseeablefuture. The enterococcus has likely emerged as amajor nosocomial pathogen in part because of itsresistance to multiple antibiotics, which allows itto survive and subsequently infect patients. Withits propensity to acquire new traits, such as high-level gentamicin, penicillin, and vancomycinresistance, the enterococcus continues to createnew therapeutic problems and dilemmas; itsability to transfer some of its plasmids tostreptococci and staphylococci and the implica-tions of a possible spread of penicillin andvancomycin resistance to these and other gram-positive species are also of concern.

Dr. Murray is professor and director, Division ofInfectious Diseases, and codirector, Center for theStudy of Emerging and Re-emerging Pathogens,University of Texas Houston-Medical School. Her maininterests are in the areas of antibiotic resistance,molecular epidemiology, and bacterial pathogenesis.Recent work has been focused on possible mechanismsof virulence of enterococci.

References 1. Murray BE. The life and times of the enterococcus. Clin

Microbiol Rev 1990;3:46-65. 2. Moellering RC Jr. Emergence of enterococcus as a

significant pathogen. Clin Infect Dis 1992;14:1173-8. 3. Schaberg DR, Culver DH, Gaynes RP. Major trends in

the microbial etiology of nosocomial infection. Am JMed 1991;91(Suppl 3B):72S-75S.

4. Hodges TL, Zighelboim-Daum S, Eliopoulos GM,Wennersten C, Moellering RC Jr. Antimicrobialsusceptibility changes in Enterococcus faecalisfollowing various penicillin exposure regimens.Antimicrob Agents Chemother 1992;36:121-5.

5. Zervos MJ, Schaberg DS. Reversal of the in vitrosusceptibility of enterococci to trimethoprim-sulfamethoxazole by folinic acid. Antimicrob AgentsChemother 1985;28:446-8.

6. Najjar A, Murray BE. Failure to demonstrate aconsistent in vitro bactericidal effect of trimethoprim-sulfamethoxazole against enterococci. AntimicrobAgents Chemother 1987;31:808-10.

7. Chenoweth CE, Robinson KA, Schaberg DR. Efficacy ofampicillin versus trimethprim-sulfamethoxazole in amouse model of lethal enterococcal peritonitis.Antimicrob Agents Chemother 1990;34:1800-2.

8. Grayson ML, Thauvin-Eliopoulos C, Eliopoulos GM,Yao JDC, DeAngelis DV, Walton L, et al. Failure oftrimethoprim-sulfamethoxazole therapy inexperimental enterococcal endocarditis. AntimicrobAgents Chemother 1990;34:1792-4.

9. Clewell DB, Keith EW. Sex pheromones and plasmidtransfer in Enterococcus faecalis. Plasmid 1989;21:175-84.

10. Clewell DB. Plasmids, drug resistance, and genetransfer in the genus Streptococcus. Microbiol Rev1981;45:409-36.

11. Clewell DB. Conjugative transposons and thedissemination of antibiotic resistance in streptococci.Annu Rev Microbiol 1986;40:635-59.

12. Roberts MC. Characterization of the Tet Mdeterminants in urogenital and respiratory bacteria.Antimicrob Agents Chemother 1990;34:476-8.

13. Roberts MC, Hillier SL. Genetic basis of tetracyclineresistance in urogenital bacteria. Antimicrob AgentsChemother 1990;34:261-4.

14. Quintiliani R Jr, Courvalin P. Characterization of Tn1547,a composite transposon flanked by the IS16 and IS256-like elements, that confers vancomycin resistance inEnterococcus faecalis BM4281. Gene 1996;172:1-8.

15. Ferretti JJ, Gilmore KS, Courvalin P. Nucleotidesequence analysis of the gene specifying the bifunctional6'-aminoglycoside acetyltransferase 2"-aminoglycosidephosphotransferase enzyme in Streptococcus faecalis andidentification and cloning of gene regions specifying thetwo activities. J Bacteriol 1986;167:631-8.

16. Carlier C, Courvalin P. Emergence of 4',4"-aminoglycoside nucleotidyltransferase in enterococci.Antimicrob Agents Chemother 1990;34:1565-9.

17. Chow JW, Zervos MJ, Lerner SA, Thal LA, DonabedianSM, Jaworski DD, et al. A novel gentamicin resistancegene in Enterococcus. Antimicrob Agents Chemother1997;41:511-4.


Synopses

18. Murray BE. ß-lactamase-producing enterococci.Antimicrob Agents Chemother 1992;36:2355-9.

19. Murray BE, Singh KV, Markowitz SM, Lopardo HA,Patterson JE, Zervos MJ, et al. Evidence for clonalspread of a single strain of ß-lactamase-producingEnterococcus faecalis to six hospitals in five states. JInfect Dis 1991;163:780-5.

20. Rhinehart E, Smith NE, Wennersten C, Gorss E,Freeman J, Eliopoulos GM, et al. Rapid dissemination ofß-lactamase-producing, aminoglycoside-resistant Entero-coccus faecalis among patients and staff on an infant-toddler surgical ward. N Engl J Med 1990;323:1814-8.

21. Wells VD, Wong ES, Murray BE, Coudron PE, WilliamsDS, Markowitz DM. Infections due to beta-lactamase-producing, high-level gentamicin-resistant Enterococcusfaecalis. Ann Intern Med 1992;116:285-92.

22. Grayson ML, Eliopoulos GM, Wennersten CB, RuoffKL, de Girolami PC, Ferraro M-J, Moellering, RC Jr.Increasing resistance to ß-lactam antibiotics amongclinical isolates of Enterococcus faecium: a 22-yearreview at one institution. Antimicrob AgentsChemother 1991;35:2180-4.

23. Fontana R, Ligozzi M, Pittaluga F, Satta G. Intrinsicpenicillin resistance in enterococci. Microbial DrugResist 1996;2:209-13.

24. Rice LB, Carias LL, Rudin SB, Donskey CJ. Abstract#LB-17 Transferable ampicillin resistance inEnterococcus faecium and linkage of ampicillin andvancomycin-resistance determinants [abstract]. Inprogram of the 37th ICAAC Program Addendum; 1997Sep 28–Oct1; Toronto, Ontario, Canada.

25. Perichon B, Reynolds P, Courvalin P. VanD-typeglycopeptide-resistant Enterococcus faecium BM4339.Antimicrob Agents Chemother 1997;41:2016-8.

26. Derlot E, Courvalin P. Mechanisms and implications ofglycopeptide resistance in enterococci. Am J Med1991;91(Suppl 3B):82S-5S.

27. Arthur M, Molinas C, Bugg TDH, Wright GD, WalshCT, Courvalin P. Evidence for in vivo incorporation ofD-lactate into peptidoglycan precursors of vancomycin-resistant enterococci. Antimicrob Agents Chemother1992;36:867-9.

28. Dutka-Malen S, Molinas C, Arthur M, Courvalin P.Sequence of the vanC gene of Enterococcus gallinarumBM4174 encoding a D-alanine:D-alanine ligase-related protein necessary for vancomycin resistance.Gene 1992;112:53-8.

29. Navarro F, Courvalin P. Analysis of genes encoding D-alanine-D-alanine ligase-related enzymes inEnterococcus casseliflavus and Enterococcus flavescens.Antimicrob Agents Chemother 1994;38:1788-93.

30. Billot-Klein D, Gutmann L, Sable S, Guittet E, vanHeijenoort J. Modification of peptidoglycan precursors is acommon feature of the low-level vancomycin-resistantVANB-type enterococcus D366 and of the naturallyglycopeptide-resistant species Lactobacillus casei,Pediococcus pentosaceus, Leuconostoc mesenteroides, andEnterococcus gallinarum. J Bacteriol 1994;176:2398-405.

31. Moellering RC Jr, Wennersten C, Medrek T,Weinberg AN. 1971. Prevalence of high-levelresistance to aminoglycosides in clinical isolates ofenterococci. Antimicrobial agents and chemotherapy—1970. Washington: American Society for Microbiology;1971. p. 335-40.

32. Watanakunakorn C. The prevalence of high-levelaminoglycoside resistance among enterococci isolatedfrom blood cultures during 1980-1988. J AntimicrobChemother 1989;24:63-8.

33. Phillips I, King A, Gransden WR, Eykyn SJ. Theantibiotic sensitivity of bacteria isolated from the bloodof patients in St. Thomas’ Hospital, 1969-1988. JAntimicrob Chemother 1990;25:59-80.

34. Zervos MJ, Dembinski S, Mikesell T, Schaberg DR.High-level resistance to gentamicin in Streptococcusfaecalis: risk factors and evidence for exogenousacquisition of infection. J Infect Dis 1986;153:1075-83.

35. Eliopoulos GM, Wennersten C, Zighelboim-Daum S,Reiszner E, Goldmann D, Moellering RC Jr. High-levelresistance to gentamicin in clinical isolates ofStreptococcus (Enterococcus) faecium. AntimicrobAgents Chemother 1988;32:1528-32.

36. Murray BE, Singh KV, Heath JD, Sharma BR,Weinstock GM. Comparison of genomic DNAs ofdifferent enterococcal isolates using restrictionendonucleases with infrequent recognition sites. J ClinMicrobiol 1990;28:2059-63.

37. Zervos MJ, Mikesell TS, Schaberg DR. Heterogeneityof plasmids determining high-level resistance togentamicin in clinical isolates of Streptococcus faecalis.Antimicrob Agents Chemother 1986;30:78-81.

38. Hodel-Christian SL, Murray BE. Characterization ofthe gentamicin resistance transposon Tn5281 fromEnterococcus faecalis and comparison to staphylococcaltransposons Tn4001 and Tn4031. Antimicrob AgentsChemother 1991;35:1147-52.

39. Hodel-Christian SL, Smith M, Zscheck KZ, MurrayBE. 1991. Comparison of a gentamicin resistancetransposon and a beta-lactamase gene from enterococcito those from staphylococci. In: Dunny GM, Cleary PP,McKay LL, editors. Genetics and molecular biology ofstreptococci, lactococci, and enterococci. Washington:American Society for Microbiology; 1991. p. 54-8.

40. Thal LA, Chow JW, Clewell DB, Zervos MJ. Tn924, achromosome-borne transposon encoding high-levelgentamicin resistance in Enterococcus faecalis.Antimicrob Agents Chemother 1994;38:1152-6.

41. Rice LB, Carias LL, Marshall SH, Bonafede ME. 1996.Sequences found on staphylococcal ß-lactamaseplasmids integrated into the chromosome ofEnterococcus faecalis CH116. Plasmid 1996;35:81-90.

42. Seetulsingh PS, Tomayko JF, Coudron PE, MarkowitzSM, Skinner C, Singh KV, Murray BE. ChromosomalDNA restriction endonuclease digestion patterns of ß-lactamase-producing Enterococcus faecalis isolatescollected from a single hospital over a 7-year period. JClin Microbiol 1996;34:1892-6.

43. Tomayko JF, Murray BE. Analysis of Enterococcusfaecalis isolates from intercontinental sources bymultilocus enzyme electrophoresis and pulsed-field gelelectrophoresis. J Clin Microbiol 1995;33:2903-7.

44. Coudron PE, Markowitz SM, Wong ES. Isolation of a ß-lactamase-producing, aminoglycoside-resistant strainof Enterococcus faecium. Antimicrob Agents Chemother1992;36:1125-6.

45. Patterson JE, Singh KV, Murray BE. Epidemiology ofan endemic strain of ß-lactamase-producingEnterococcus faecalis. J Clin Microbiol 1991;29:2513-6.


Synopses

46. Handwerger S, Perlman DC, Altarac D, McAuliffe V.Concomitant high-level vancomycin and penicillinresistance in clinical isolates of enterococci. Clin InfectDis 1992;14:655-61.

47. Miranda AG, Singh KV, Murray BE. DNA fingerprintingof Enterococcus faecium by pulsed-field gelelectrophoresis may be a useful epidemiologic tool. JClin Microbiol 1991;29:2752-7.

48. Handwerger S, Skoble J. Identification of chromosomalmobile element conferring high-level vancomycinresistance in Enterococcus faecium. Antimicrob AgentsChemother 1995;39:2446-53.

49. Handwerger S, Skoble J, Discotto LF, Pucci MJ.Heterogeneity of the vanA gene cluster in clinicalisolates of enterococci from the Northeastern UnitedStates. Antimicrob Agents Chemother 1995;39:362-8.

50. van den Bogaard AE, Jensen LB, Stobberingh EE.Vancomycin-resistant enterococci in turkeys andfarmers. N Engl J Med 1997;337:1558-9.

51. Power EGM, Abdulla YH, Talsania HG, Spice W,Aathithan S, French GL. vanA genes in vancomycin-resistant clinical isolates of Oerskovia turbata andArcanobacterium (Corynebacterium) haemolyticum. JAntimicrob Chemother 1995;36:595-606.

52. Poyart C, Pierre C, Quesne G, Pron B, Berche P, Trieu-Cuot P. Emergence of vancomycin resistance in thegenus Streptococcus: characterization of a vanBtransferable determinant in Streptococcus bovis.Antimicrob Agents Chemother 1997;41:24-9.

53. Biavasco F, Miele A, Vignaroli C, Manso E, Lupid R,Varaldo PE. Genotypic characterization of a nosocomialoutbreak of VanA Enterococcus faecalis. Microb DrugResist 1996;2:231-7.

54. Boyce JM, Opal SM, Chow JW, Zervos MJ, Potter-BynoeG, Sherman CB, et al. Outbreak of multidrug-resistantEnterococcus faecium with transferable vanB classvancomycin resistance. J Clin Microbiol 1994;32:1148-53.

55. Handwerger S, Raucher B, Altarac D, Monka J,Marchione S, Singh KV, et al. Nosocomial outbreakdue to Enterococcus faecium highly resistant tovancomycin, penicillin and gentamicin. Clin InfectDis 1993;16:750-5.

56. Livornese LL Jr, Dias S, Samel C, Romanowski B,Taylor S, May P, et al. Hospital-acquired infection withvancomycin-resistant Enterococcus faecium transmittedby electronic thermometers. Ann Intern Med1992;117:112-6.

57. Moreno F, Grota P, Crisp C, Magnon K, Melcher GP,Jorgensen JH, Patterson JE. Clinical and molecularepidemiology of vancomycin-resistant Enterococcusfaecium during its emergence in a city in southernTexas. Clin Infect Dis 1995;21:1234-7.

58. Perlada DE, Smulian AG, Cushion MT. Molecularepidemiology and antibiotic susceptibility of enterococciin Cincinnati, Ohio: a prospective citywide survey. JClin Microbiol 1997;35:2342-7.

59. Chow JW, Kuritza A, Shlaes DM, Green M, Sahm DF,Zervos MJ. Clonal spread of vancomycin-resistantEnterococcus faecium between patients in three hosptialsin two states. J Clin Microbiol 1993;31:1609-11.

60. Clark NC, Cooksey RC, Hill BC, Swenson JM, TenoverFC. Characterization of glycopeptide-resistantenterococci from U.S. hospitals. Antimicrob AgentsChemother 1993;37:2311-7.

61. Sader HS, Pfaller MA, Tenover FC, Hollis RJ, JonesRN. Evaluation and characterization of multiresistantEnterococcus faecium from 12 U.S. medical centers. JClin Microbiol 1994;32:2840-2.

62. Boyle JF, Soumakis SA, Rendo A, Herrington JA,Gianarkis DG, Thurberg BE, Painter BG. Epidemiologicanalysis and genotypic characterization of a nosocomialoutbreak of vancomycin-resistant enterococci. J ClinMicrobiol 1993;31:1280-5.

63. Morris JG Jr, Shay DK, Hebden JN, McCarter RJ Jr,Perdue BE, Jarvis W, et al. Enterococci resistant tomultiple antimicrobial agents, including vancomycin.Ann Intern Med 1995;123:250-9.

64. Bingen EH, Denamur E, Lambert-Zechovsky NY,Elion J. Evidence for the genetic unrelatedness ofnosocomial vancomycin-resistant Enterococcus faeciumstrains in a pediatric hospital. J Clin Microbiol1991;29:1888-92.

65. Plessis P, Lamy T, Donnio PY, Autuly F, Grulois I,LePrise PY, Avril JL. Epidemiologic analysis ofglycopeptide-resistant Enterococcus strains inneutropenic patients receiving prolonged vancomycinadministration. Eur J Clin Microbiol Infect Dis1995;14:959-63.

66. Jordens JZ, Bates J, Griffiths DT. Faecal carriage andnosocomial spread of vancomycin-resistant Enterococcusfaecium. J Antimicrob Chemother 1994;34:515-28.

67. Klare I, Heier H, Claus H, Bohme G, Marin S,Seltmann G, et al. Enterococcus faecium strains withvanA-mediated high-level glycopeptide resistanceisolated from animal foodstuffs and fecal samples ofhumans in the community. Microbial Drug Resist1995;3:265-72.

68. Van der Auwera P, Pensart N, Korten V, Murray BE,Leclercq R. Influence of oral glycopeptides on the fecalflora of human volunteers: selection of highly-glycopeptide-resistant enterococci. J Infect Dis1995;173:1129-36.

69. Aarestrup FM, Ahrens P, Madsen M, Pallesen LV,Poulsen RL, Westh H. Glycopeptide susceptibilityamong Danish Enterococcus faecium and Enterococcusfaecalis isolates of animal and human origin and PCRidentification of genes within the VanA cluster.Antimicrob Agents Chemother 1996;40:1938-40.

70. Bates J, Jordens Z, Selkon JB. Evidence for an animalorigin of vancomycin-resistant enterococci [letter].Lancet 1993;342:490.

71. Klare I, Heier H, Claus H, Reissbrodt R, Witte W.vanA-mediated high-level glycopeptide resistance inEnterococcus faecium from animal husbandry. FEMSMicrobiol Lett 1995;125:165-72.

72. Schouten MA, Voss A, Hoogkamp-Korstanje JAA. VREand meat. Lancet 1997;349:1258.

73. Coque TM, Tomayko JF, Ricke SC, Okhuysen PO,Murray BE. Vancomycin-resistant enterococci fromnosocomial, community, and animal sources in theUnited States. Antimicrob Agents Chemother1996;40:2605-9.


Synopses

74. Kjerulf A, Pallesen L, Westh H. Vancomycin-resistantenterococci at a large university hospital in Denmark.APMIS 1996;104:475-9.

75. Klare I, Heier H, Claus H, Witte W. Environmentalstrains of Enterococcus faecium with inducible high-level resistance to glycopeptides. FEMS Microbiol Lett1993;106:23-30.


Synopses

Hundreds of millions of cases of malaria occurannually, and infections with Plasmodiumfalciparum, the most virulent human malariaparasite, cause more than one million deaths peryear (1). Despite extensive control efforts, theincidence of the disease is not decreasing in mostmalaria-endemic areas of the world, and in someit is clearly increasing (2). Malaria also remains amajor risk to travelers from industrialized todeveloping countries. Because malaria parasitesare increasingly resistant to antimalarial drugs,appropriately counseled travelers to malaria-endemic regions are more likely to contractmalaria now than they were 40 years ago.

Malaria control efforts include attempts todevelop an effective vaccine, eradicate mosquitovectors, and develop new drugs (2,3). However,the development of a vaccine has proven verydifficult, and a highly effective vaccine willprobably not be available in the near future (4).Efforts to control Anopheles mosquitoes have hadlimited success, although the use of insecticide-impregnated bed nets does appear to reducemalaria-related death rates (5). In addition,methods to replace natural vector populationswith mosquitoes unable to support parasitedevelopment are under study and may contributeto malaria control in the long term (6). However,

the current limitations of vaccine and vectorcontrol, as well as the increasing resistance ofmalaria parasites to existing drugs, highlight thecontinued need for new antimalarial agents.

Established Antimalarial DrugsAntimalarial drugs have been used for

centuries. Early natural products, including thebark of the cinchona tree in South America andextracts of the wormwood plant in China, wereamong the first effective antimicrobial agents tobe used. Cinchona bark was used in Europebeginning in the 17th century, and upon itsisolation from bark in 1820, quinine becamewidely used. In the last 50 years, extensive efforts,including the screening of hundreds of thousandsof compounds, have led to the development of anumber of effective synthetic antimalarial drugs.The most important of these, chloroquine, hasbeen the mainstay of antimalarial chemotherapyfor the last 50 years. The compound eradicatesparasites rapidly, has minimal toxicity, is widelyavailable at low cost throughout the world, andneeds to be taken only once a week forchemoprophylaxis. However, resistance to chloro-quine has been steadily increasing since thedrug’s initial use in South America and SoutheastAsia in the late 1950s. Chloroquine resistance isnow widespread in most P. falciparum-endemicareas of the world (3). Thus, the use of chloroquinefor presumptive treatment of falciparum malaria

Proteases of Malaria Parasites: NewTargets for Chemotherapy

Philip J. RosenthalSan Francisco General Hospital and University of California,

San Francisco, California, USA

Address for correspondence: Philip Rosenthal, Box 0811,University of California, San Francisco, CA 94143-0811,USA; fax: 415-206-6015; e-mail: [email protected].

The increasing resistance of malaria parasites to antimalarial drugs is a majorcontributor to the reemergence of the disease as a major public health problem and itsspread in new locations and populations. Among potential targets for new modes ofchemotherapy are malarial proteases, which appear to mediate processes within theerythrocytic malarial life cycle, including the rupture and invasion of infected erythrocytesand the degradation of hemoglobin by trophozoites. Cysteine and aspartic proteaseinhibitors are now under study as potential antimalarials. Lead compounds have blocked invitro parasite development at nanomolar concentrations and cured malaria-infected mice.This review discusses available antimalarial agents and summarizes experimental resultsthat support development of protease inhibitors as antimalarial drugs.


Synopses

or for chemoprophylaxis is usually no longerappropriate (7). Moreover, resistance to chloro-quine of P. vivax, the second most lethal humanmalaria parasite, is increasing in South Asia (8).

No other antimalarial drug (9-12) is asefficacious and safe as chloroquine (Table 1). Thebest antimalarial drug for treating chloroquine-resistant falciparum malaria remains quinine (orintravenous quinidine), which is fairly toxic;quinine resistance is increasing in SoutheastAsia, particularly in the border areas of Thailand(9). Amodiaquine, used to treat chloroquine-resistant malaria in developing countries, is alsoquite toxic, and resistance to it is also common(13). Mefloquine (14) is widely used forchemoprophylaxis against chloroquine-resistantP. falciparum, but its use is limited by toxicity (15)and (in the developing world) high cost. Mefloquineis not approved for treatment of malaria in theUnited States because of the neurotoxicity ofdoses required for the treatment. Fansidar, acombination of sulfadoxine and pyrimethamine,is no longer recommended for chemoprophylaxisbecause of its dermatologic toxicity (15). Fansidaris also not an ideal drug for treatment because it isslow acting, but it is increasingly important intreating chloroquine-resistant malaria in developing

countries because economic constraints limit theuse of other agents (16). The use of bothmefloquine and Fansidar will increasingly belimited by drug resistance, already widespread inparts of Southeast Asia (9,17).

Other antimalarial drugs have specializeduses. Tetracyclines and some other antibiotics(clindamycin, sulfas) are slow acting andgenerally best used as an adjunct to quininetherapy in treating falciparum malaria (9).Doxycycline is also used for chemoprophylaxis inregions with high levels of drug resistance,especially Southeast Asia (10,17). Other drugs forchemoprophylaxis include proguanil, whichremains effective in combination with chloro-quine in many areas other than Southeast Asia,and Maloprim, a combination of dapsone andpyrimethamine (10,17). Resistance to these drugsis fairly common, however. Primaquine has awell-defined specific role: eradicating chronicliver stages of P. vivax and P. ovale after treatingthe acute blood infection with chloroquine.

New Antimalarial DrugsRelatively few new antimalarial drugs are

undergoing clinical testing (Table 2). Halofantrine,identified in the 1940s, was not developed until

Table 1. Established antimalarial drugsa

Drug Role Best Feature(s) LimitationsChloroquine TX of and CP against non- Very safe; low Widespread R

Pf and sensitive Pf parasites cost; long half-life

Quinine/quinidine Best TX for Pf malaria; Limited R; Fairly toxic (cinchonism, low cost rapidly acting cardiac)

Amodiaquineb TX of R Pf malaria Low cost Toxicity (bone marrow, liver); R common

Mefloquine CP against R Pf malaria; Relatively little R, Moderately toxic not approved for TX though increasing; (mostly CNS); high in United States long half-life cost; R in SE Asia

Fansidar TX of Pf malaria; no longer Relatively low cost; Skin toxicity (can be recommended for CP long half-life fatal); increasing R

Primaquine Eradication of chronic Only drug for this Hemolysis with G6PD liver stage Pv, Po malaria indication deficiency; increasing R

Proguanilb CP only (often with Low cost; nontoxic R common chloroquine)

Maloprimb CP only (often with Low cost R common; skin rashes chloroquine)

Tetracyclines CP; TX of Pf malaria in Low cost Skin and gastrointes- combination with quinine tinal toxicity

aTX, therapy; CP, chemoprophylaxis; R, resistance/resistant; Pf, Plasmodium falciparum; Pv, P. vivax; Po, P. ovale; CNS, centralnervous system; G6PD, glucose 6-phosphate dehydrogenase.bNot available in the United States.


Synopses

the 1980s; its use has been limited by variable oralabsorption and cardiac toxicity (12,18). The drugis approved in the United States for treatment ofchloroquine-resistant P. falciparum infection,although in most cases quinine (or intravenousquinidine) is preferable. The most effective newdrugs are artemisinin and related compounds.Artemisinin was isolated in 1972 from Artemisiaannua, a plant used in China for centuries to treatfever (19). Artemisinin derivatives (artesunate,artelinate, artemether, arteether, dihydroartemisinin)have been synthesized and are undergoingextensive clinical testing. These compounds, whichare already widely used in some areas, are potent,rapidly acting antimalarials that are effectiveagainst chloroquine-resistant P. falciparum (20).Because recrudescences of infection after treat-ment are common, however, artemisinin andrelated compounds might best be used incombination with another drug.

Other compounds are under evaluation.Atovaqone (21), which is approved for treatingpatients with Pneumocystis infections, appearsto be effective against malaria in combinationwith proguanil (22), but its use has been limitedby recrudescence after treatment. Pyronaridine,an acridine derivative used to treat malaria inChina, has shown efficacy against falciparummalaria (23). The iron chelator desferrioxamineenhances the clearance of parasites in mildmalaria (24) and, in conjunction with quinineand Fansidar, hastens recovery from deep comain severe falciparum malaria (25). Azithromycin,a quinolone antibiotic, appears efficacious inmalaria chemoprophylaxis (26).

Malarial Proteases: New Targets forChemotherapy

The limitations of antimalarial chemotherapyunderscore the need for new drugs, ideallydirected against new targets. Potential targets forchemotherapy include malarial proteases (27).The erythrocytic life cycle, which is responsible forall clinical manifestations of malaria, begins whenfree merozoites invade erythrocytes. Theintraerythrocytic parasites develop from smallring-stage organisms to larger, more metaboli-cally active trophozoites and then to multi-nucleated schizonts. The erythrocytic cycle iscompleted when mature schizonts ruptureerythrocytes, releasing numerous invasive mero-zoites. Proteases appear to be required for therupture and subsequent reinvasion of erythro-cytes by merozoite-stage parasites and for thedegradation of hemoglobin by intraerythrocytictrophozoites (Figure).

Proteases and Erythrocyte Rupture andInvasion

The rupture of erythrocytes by matureschizonts and the subsequent invasion oferythrocytes by free merozoites appear to requiremalarial protease activity, possibly to breach theerythrocyte cytoskeleton, a complex network ofproteins. In addition, a number of malarialproteins are proteolytically processed during thelate schizont and merozoite life-cycle stages; forexample, merozoite surface protein-1 is processedin a manner inhibited by serine proteaseinhibitors (28), presumably to facilitate thecomplex series of events involved in erythrocyte

Table 2. New antimalarial drugsDrug Role Best Feature(s) LimitationsHalofantrine TX of Pf malaria; not Usually effective Variable bioavailability,

approved for CP against R Pf malaria cardiac toxicity

Artemisinin and TX of Pf malaria Rapidly acting; effec- Recurrence after TX fairly related compoundsa tive against multidrug-R common

strains

Atovaquone ? TX of Pf malaria; ? Limited toxicity Limited studies so far show CP (probably in combi- frequent recurrence after TX nation with proguanil)

Pyronaridinea ? TX of Pf malaria Effective against R strains Studies limited to date

Desferrioxamine ? TX of severe Pf Well tolerated when Studies limited to date malaria used for iron overload

Azithromycin ? CP Limited toxicity Studies limited to dateFor abbreviations, see Table 1, footnote a.aNot available in the United States.


Synopses

rupture and invasion (29). Although the specificroles of different classes of proteases are notcompletely clear, inhibitors of cysteine and serineproteases have consistently blocked erythrocyterupture and invasion (27).

Candidate P. falciparum rupture/invasionproteases have been identified, but none has beenfully characterized biochemically or molecularly:1) a 68 kD cysteine protease was identified inschizonts and merozoites and localized to themerozoite apex, suggesting that it may bereleased from the rhoptry organelle duringinvasion (30); 2) a cysteine protease of matureschizonts and a serine protease of merozoiteswere identified in highly synchronized parasites(31); 3) a serine protease was shown to be bound tothe schizont/merozoite membrane by a glycosyl-phosphatidylinositol anchor, to be activated byphosphatidylinositol-specific phospholipase Cduring the merozoite stage, and to be capable ofcleaving the erythrocyte cytoskeletal proteinband 3 (32,33); 4) another protease, inhibited byboth cysteine and serine protease inhibitors,hydrolyzed the erythrocyte cytoskeletal proteinsspectrin and band 4.1 (34); and 5) the serinerepeat antigen (35,36) and the related proteinSERP H (37), both expressed in matureschizonts, have important similarities in theirsequences with cysteine proteases. Furtherresearch should identify the specific biologicroles of the proteases mentioned and better

characterize these enzymes, thus fostering thedevelopment of specific inhibitors.

Host proteases may also play a role inerythrocyte rupture by P. falciparum. In recentstudies, host urokinase was shown to bind to thesurface of P. falciparum-infected erythrocytes,and the depletion of urokinase from parasiteculture medium inhibited erythrocyte rupture bymature schizonts (38). This inhibition wasreversed by exogenous urokinase.

Drug Development EffortsSynthetic peptide inhibitors of the P.

falciparum schizont cysteine protease Pf 68inhibited erythrocyte invasion by culturedparasites (39,40). The most effective peptide,GlcA-Val-Leu-Gly-Lys-NHC2H5, inhibited theprotease and blocked parasite development athigh micromolar concentrations (40; Table 3).Although these results do not demonstrate levelsof inhibition expected to be therapeuticallyrelevant, they suggest that a specific proteaseactivity is required for erythrocyte invasion bymalaria parasites and thus is a potential target forantimalarial drugs.

Proteases and Malarial HemoglobinDegradation

Extensive evidence suggests that the degra-dation of hemoglobin is necessary for the growthof erythrocytic malaria parasites, apparently toprovide free amino acids for parasite proteinsynthesis (27,50). In P. falciparum, hemoglobindegradation occurs predominantly in trophozoitesand early schizonts, the stages at which theparasites are most metabolically active. Tropho-zoites ingest erythrocyte cytoplasm and transportit to a large central food vacuole. In the foodvacuole, hemoglobin is broken down into heme, amajor component of malarial pigment (51), andglobin, which is hydrolyzed to its constituent aminoacids. The food vacuole is an acidic organelleanalogous to lysosomes. Several lysosomal pro-teases are well characterized, including cysteine(cathepsins B, H, and L) and aspartic (cathepsin D)proteases (52), and malaria parasites containanalogous food vacuole proteases that degradehemoglobin. At least two aspartic proteases andone cysteine protease have been isolated frompurified P. falciparum food vacuoles (53).

Malarial aspartic protease activities havebeen identified (54-60). Two recently character-ized aspartic proteases (plasmepsin I and

Figure. Protease targets in erythrocytic malariaparasites. The Plasmodium falciparum erythrocyticlife cycle is shown schematically, and data supportingcysteine (CP), serine (SP), and aspartic (AP) proteasesof the different parasite stages as chemotherapeutictargets are provided in italics.


Synopses

plasmepsin II) are located in the food vacuole, haveacid pH optima, and share sequence homology withother aspartic proteases (41,53,61,62). Further-more, the aspartic proteases can cleave hemoglo-bin. One of the enzymes, plasmepsin I, cleavesnative hemoglobin (53,59). Plasmepsin II appearsto prefer denatured globin as a substrate (53). Onthe basis of these data, plasmepsin I is thought to beresponsible for initial cleavages of hemoglobin afterthe molecule is transported to the food vacuole (53).

Incubation of cultured P. falciparum para-sites with the protease inhibitor leupeptin causedtrophozoite food vacuoles to fill with apparentlyundegraded erythrocyte cytoplasm (63-65). Analy-sis of the leupeptin-treated parasites showed thatthey contained large quantities of undegradedglobin, while minimal globin was detectable incontrol parasites (64,66). Leupeptin inhibits bothcysteine and some serine proteases, but the highlyspecific cysteine protease inhibitor E-64 alsocaused undegraded globin to accumulate. Afterparasites were incubated with inhibitors of otherclasses of proteases including the asparticprotease inhibitor pepstatin (63-67), globin didnot accumulate. More recent studies that usednondenaturing electrophoretic methods demon-strated that cysteine protease inhibitors not onlyblocked malarial globin hydrolysis, but alsoinhibited earlier steps in hemoglobin degrada-tion, including denaturation of the hemoglobintetramer and the release of heme from globin (68).Another study showed that E-64, but notpepstatin, inhibited the production of hemozoin

(the malarial end product of heme) by culturedparasites (69). These results suggest that acysteine protease is required for initial steps inhemoglobin degradation by P. falciparum.

A P. falciparum trophozoite cysteine proteasewith biochemical features expected for a foodvacuole hemoglobinase has been identified (31)and biochemically (70-72) and molecularly (73)characterized. This protease, called falcipain,degraded denatured and native hemoglobin invitro; its acid pH optimum, substrate specificity,and inhibitor sensitivity indicated that it was apapain family cysteine protease (64,70,71).Specific inhibitors of falcipain blocked hemoglo-bin degradation and prevented parasite develop-ment. The degree of inhibition of falcipain byfluoromethyl ketones (44) and vinyl sulfones (46)correlated with their inhibition of hemoglobindegradation and parasite development, support-ing the hypothesis that falcipain is the cysteineprotease required for hemoglobin degradation.

The specific mechanism for hemoglobindegradation in the malarial food vacuole remainsunclear. As noted above, both the asparticprotease plasmepsin I and the cysteine proteasefalcipain have been identified in parasite foodvacuoles and shown to cleave denatured andnative hemoglobin in vitro (53,71). Resultsshowing that only cysteine protease inhibitorsblock hemoglobin processing and globin hydroly-sis in cultured parasites suggest that falcipain isrequired for initial steps of hemoglobin degrada-tion (66-68,74). However, other studies have

Table 3. Protease targets for chemotherapyEffective inhibitorsa

In vitrob In vivoc

Protease Biologic role Compound (Reference) (IC50; µM) (mg/kg/day)Pf68 Erythrocyte invasion GlcA-Val-Leu-Gly-Lys-NHC2H5 (40) 900Plasmepsin I Hemoglobin degradation SC-50083 (41) 2-5

Ro 40-4388 (42) 0.25Plasmepsin II Hemoglobin degradation Compound 7 (43) 20Falcipain Hemoglobin degradation Z-Phe-Arg-CH2F (44) 0.064

Mu-Phe-HPh-CH2F (45) ~0.03 400Mu-Leu-HPh-VSPh (46) 0.01Oxalic bis ((2-hydroxy-1-naph- 7 thylmethylene)hydrazide) (47)1-(2,5-dichlorophenyl)-3- 0.23 (4-quinolinyl)-2-propen-1-one (48)7-chloro-1,2-dihydro-2-(2,3-di- 2 methoxy-phenyl)-5,5-dioxide- 4-(1H,10H)-phenothiazinone (49)

aThe structures of these compounds and details of the described studies are in the references noted.bAssays compared the development of new ring-form parasites or the uptake of [3H]hypoxanthine by treated and controlparasites.cCure of Plasmodium vinckei-infected mice.


Synopses

shown that native hemoglobin is cleaved byplasmepsin I, but not falcipain, in nonreducingconditions that may be present in the foodvacuole (53,59,72). In any event, regardless of theexact sequence of hemoglobin processing,multiple enzymes, including at least the threeproteases already identified, appear to partici-pate in the degradation of hemoglobin. Theseproteases are thus logical targets for antimalarialdrug development.

Aminopeptidase activity has also beendescribed in malaria parasites (75-77). Thisactivity, with a neutral pH optimum, was notfound in food vacuole lysates (77). When theselysates were incubated with hemoglobin, discretepeptide fragments, but not free amino acids, wereidentified (77). These results suggest thathemoglobin is degraded to small peptides in thefood vacuole, that these peptides are transportedto the parasite cytosol, and that additionalprocessing of hemoglobin peptides is mediated bycytosolic aminopeptidase activity (77).

Drug Development EffortsBoth the cysteine protease inhibitor E-64 and

the aspartic protease inhibitor pepstatin blockedP. falciparum development (63-67). Administeredtogether, the two inhibitors acted synergistically(67). However, only E-64 blocked globin hydroly-sis (64-67). Numerous peptide-based cysteineprotease inhibitors, including fluoromethyl ke-tones (44,70,78) and vinyl sulfones (46), inhibitedfalcipain at low nanomolar concentrations andinhibited P. falciparum development and hemo-globin degradation at concentrations below 100nanomolar (Table 3). In a malaria animal model, afluoromethyl ketone that inhibited falcipain atlow nanomolar concentrations blocked P. vinckeiprotease activity in vivo after a single subcutane-ous dose, and, when administered for 4 days,cured 80% of murine malaria infections (45).Thus, despite the theoretical limitations ofpotentially rapid degradation in vivo and inhibitionof host proteases, peptide protease inhibitors showpromise as candidate antimalarial drugs.Fluoromethyl ketones have subsequently showntoxicity in animal studies, but evaluations ofrelated, apparently nontoxic inhibitors of falcipainas antimalarial drugs are under way.

A computer model for the structure offalcipain was used to identify nonpeptideinhibitors (47). Screening of potential nonpeptide

inhibitors identified a low micromolar leadcompound (47; Table 3). Subsequent synthesis andtesting of small molecules based on the structure ofthe lead compound have identified biologicallyactive falcipain inhibitors, including chalcones thatblock parasite metabolism at submicromolarconcentrations (48) and phenothiazines that blockparasite metabolism and development at lowmicromolar concentrations (49).

Peptidelike aspartic protease inhibitors arepotent inhibitors of plasmepsins I and II. Inindependent studies SC-50083 (41), Ro 40-4388(42), and “compound 7” (43) inhibited plasmepsinI or II at nanomolar concentrations and blockedparasite development at high nanomolar tomicromolar concentrations (Table 3). Drug develop-ment efforts should be assisted by the recentdetermination of the structure of plasmepsin II(43). Inhibitors of aspartic and cysteine proteaseshave synergistic effects in inhibiting the growth ofcultured malaria parasites (67), and theseproteases also act synergistically to degradehemoglobin in vitro (41). Therefore, the combina-tion of inhibitors of malarial cysteine and asparticproteases may provide the most effectivechemotherapeutic regimen and best limit thedevelopment of parasite resistance to proteaseinhibitors. Ultimately, a better understanding ofthe biochemical properties and biologic roles ofmalarial proteases will foster the development ofprotease inhibitors that specifically inhibitparasite enzymes and thus are the most suitablecandidates for chemotherapy.

AcknowledgmentsWork in the author’s laboratory was supported by grants

from the National Institutes of Health, the UNDP/WorldBank/WHO Special Programme for Research and Training inTropical Diseases, and the American Heart Association.

Dr. Rosenthal is associate professor, Departmentof Medicine, Division of Infectious Diseases, SanFrancisco General Hospital and the University ofCalifornia, San Francisco. His research interests includethe evaluation of proteases of malaria parasites aschemotherapeutic targets.

References 1. Walsh JA. Disease problems in the Third World. Ann N

Y Acad Sci 1989;569:1-16. 2. Oaks SC, Mitchell VS, Pearson GW, Carpenter CCJ,

editors. Malaria: obstacles and opportunities.Washington: National Academy Press; 1991.

3. Olliaro P, Cattani J, Wirth D. Malaria, the submergeddisease. JAMA 1996;275:230-3.


Synopses

4. Hoffman SL, Miller LH. Perspectives on malariavaccine development. In: Hoffman SL, editor. Malariavaccine development. Washington: ASM Press; 1996.p. 1-13.

5. Alonso PL, Lindsay SW, Armstrong JRM, Conteh M,Hill AG, David PH, et al. The effect of insecticide-treated bed nets on mortality of Gambian children.Lancet 1991;337:1499-502.

6. Collins FH, Besansky NJ. Vector biology and thecontrol of malaria in Africa. Science 1994;264:1874-5.

7. Centers for Disease Control. Recommendations for theprevention of malaria among travelers. JAMA1990;263:2729-40.

8. Murphy GS, Basri H, Purnomo, Andersen EM, BangsMJ, Mount DL, et al. Vivax malaria resistant totreatment and prophylaxis with chloroquine. Lancet1993;341:96-100.

9. Panisko DM, Keystone JS. Treatment of malaria—1990. Drugs 1990;39:160-89.

10. Keystone JS. Prevention of malaria. Drugs 1990;39:337-54.

11. Wyler DJ. Malaria chemoprophylaxis for the traveler.N Engl J Med 1993;329:31-7.

12. White JJ. The treatment of malaria. N Engl J Med1996;335:800-6.

13. Olliaro P, Nevill C, LeBras J, Ringwald P, Mussano P,Garner P, et al. Systematic review of amodiaquinetreatment in uncomplicated malaria. Lancet1996;348:1196-201.

14. Palmer KJ, Holliday SM, Brogden RN. Mefloquine: areview of its antimalarial activity, pharmacokineticproperties and therapeutic efficacy. Drugs 1993;45:430-75.

15. Luzzi GA, Peto TEA. Adverse effects of antimalarials:an update. Drug Saf 1993;8:295-311.

16. Hellgren U, Kihamia CM, Bergqvist Y, Lebbad M,Premji Z, Rombo L. Standard and reduced doses ofsulfadoxine-pyrimethamine for treatment ofPlasmodium falciparum in Tanzania, withdetermination of drug concentrations and susceptibilityin vitro. Trans R Soc Trop Med Hyg 1990;84:469-73.

17. Bradley DJ, Warhurst DC. Malaria prophylaxis:guidelines for travellers from Britain. BMJ1995;310:709-14.

18. Bryson HM, Goa KL. Halofantrine: a review of itsantimalarial activity, pharmacokinetic propertiesand therapeutic potential. Drugs 1992;43:236-58.

19. Meshnick SR, Taylor TE, Kamchonwongpaisan S.Artemisinin and the antimalarial endoperoxides: fromherbal remedy to targeted chemotherapy. MicrobiolRev 1996;60:301-15.

20. de Vries PJ, Dien TK. Clinical pharmacology andtherapeutic potential of artemisinin and its derivativesin the treatment of malaria. Drugs 1996;52:818-36.

21. Haile LG, Flaherty JF. Atovaquone: a review. AnnPharmacother 1993;27:1488-94.

22. Radloff PD, Philipps J, Nkeyi M, Hutchinson D, KremsnerPG. Atovaquone and proguanil for Plasmodiumfalciparum malaria. Lancet 1996;347:1511-4.

23. Ringwald P, Bickii J, Basco L. Randomised trial ofpyronaridine versus chloroquine for acuteuncomplicated falciparum malaria in Africa. Lancet1996;347:24-8.

24. Gordeuk VR, Thuma PE, Brittenham GM, Biemba G,Zulu S, Simwanza G, et al. Iron chelation as achemotherapeutic strategy for falciparum malaria. AmJ Trop Med Hyg 1993;48:193-7.

25. Gordeuk V, Thuma P, Brittenham G, McLaren C,Parry D, Backenstose A, et al. Effect of iron chelationtherapy on recovery from deep coma in children withcerebral malaria. N Engl J Med 1992;327:1473-7.

26. Anderson SL, Berman J, Kuschner R, Wesche D,Magill A, Wellde B, et al. Prophylaxis of Plasmodiumfalciparum malaria with azithromycin administered tovolunteers. Ann Intern Med 1995;123:771-3.

27. McKerrow JH, Sun E, Rosenthal PJ, Bouvier J. Theproteases and pathogenicity of parasitic protozoa.Annu Rev Microbiol 1993;47:821-53.

28. Blackman MJ, Holder AA. Secondary processing of thePlasmodium falciparum merozoite surface protein-1(MSP1) by a calcium-dependent membrane-boundserine protease: shedding of MSP133 as a noncovalentlyassociated complex with other fragments of the MSP1.Mol Biochem Parasitol 1992;50:307-16.

29. Perkins ME. Erythrocyte invasion by the malarialmerozoite: recent advances. Exp Parasitol 1989;69:94-9.

30. Bernard F, Schrevel J. Purification of a Plasmodiumberghei neutral endopeptidase and its localization inmerozoite. Mol Biochem Parasitol 1987;26:167-74.

31. Rosenthal PJ, Kim K, McKerrow JH, Leech JH.Identification of three stage-specific proteinases ofPlasmodium falciparum. J Exp Med 1987;166:816-21.

32. Braun-Breton C, Rosenberry TL, Pereira da Silva L.Induction of the proteolytic activity of a membrane proteinin Plasmodium falciparum by phosphatidyl inositol-specific phospholipase C. Nature 1988;332:457-9.

33. Roggwiller E, Bétoulle MEM, Blisnick T, Braun BretonC. A role for erythrocyte band 3 degradation by theparasite gp76 serine protease in the formation of theparasitophorous vacuole during invasion of erythrocytesby Plasmodium falciparum. Mol Biochem Parasitol1996;82:13-24.

34. Deguercy A, Hommel M, Schrevel J. Purification andcharacterization of 37-kilodalton proteases fromPlasmodium falciparum and Plasmodium bergheiwhich cleave erythrocyte cytoskeletal components. MolBiochem Parasitol 1990;38:233-44.

35. Knapp B, Hundt E, Nau U, Küpper HA. Molecularcloning, genomic structure and localization in a bloodstage antigen of Plasmodium falciparum characterized bya serine stretch. Mol Biochem Parasitol 1989;32:73-84.

36. Li W-B, Bzik DJ, Horii T, Inselburg J. Structure andexpression of the Plasmodium falciparum SERA gene.Mol Biochem Parasitol 1989;33:13-26.

37. Knapp B, Nau U, Hundt E, Küpper HA. A new bloodstage antigen of Plasmodium falciparum highlyhomologous to the serine-stretch protein SERP. MolBiochem Parasitol 1991;44:1-14.

38. Roggwiller E, Fricaud A-C, Blisnick T, Braun-BretonC. Host urokinase-type plasminogen activatorparticipates in the release of malaria merozoites frominfected erythrocytes. Mol Biochem Parasitol1997;86:49-59.

39. Schrevel J, Grellier P, Mayer R, Monsigny M. Neutralproteases involved in the reinvasion of erythrocytes byPlasmodium merozoites. Biol Cell 1988;64:233-44.


Synopses

40. Mayer R, Picard I, Lawton P, Grellier P, Barrault C,Monsigny M, et al. Peptide derivatives specific for aPlasmodium falciparum proteinase inhibit the humanerythrocyte invasion by merozoites. J Med Chem1991;34:3029-35.

41. Francis SE, Gluzman IY, Oksman A, Knickerbocker A,Mueller R, Bryant ML, et al. Molecular characterizationand inhibition of a Plasmodium falciparum aspartichemoglobinase. EMBO J 1994;13:306-17.

42. Moon RP, Tyas L, Certa U, Rupp K, Bur D, Jacquet C,et al. Expression and characterisation of plasmepsin Ifrom Plasmodium falciparum. Eur J Biochem1997;244:552-60.

43. Silva AM, Lee AY, Gulnik SV, Majer P, Collins J, BhatTN, et al. Structure and inhibition of plasmepsin II, ahemoglobin-degrading enzyme from Plasmodiumfalciparum. Proc Natl Acad Sci U S A 1996;93:10034-9.

44. Rosenthal PJ, Wollish WS, Palmer JT, Rasnick D.Antimalarial effects of peptide inhibitors of aPlasmodium falciparum cysteine proteinase. J ClinInvest 1991;88:1467-72.

45. Rosenthal PJ, Lee GK, Smith RE. Inhibition of aPlasmodium vinckei cysteine proteinase cures murinemalaria. J Clin Invest 1993;91:1052-6.

46. Rosenthal PJ, Olson JE, Lee GK, Palmer JT, Klaus JL,Rasnick D. Antimalarial effects of vinyl sulfonecysteine proteinase inhibitors. Antimicrob AgentsChemother 1996;40:1600-3.

47. Ring CS, Sun E, McKerrow JH, Lee GK, Rosenthal PJ,Kuntz ID, et al. Structure-based inhibitor design byusing protein models for the development ofantiparasitic agents. Proc Natl Acad Sci U S A1993;90:3583-7.

48. Li R, Kenyon GL, Cohen FE, Chen X, Gong B, DominguezJN, et al. In vitro antimalarial activity of chalcones andtheir derivatives. J Med Chem 1995;38:5031-7.

49. Domínguez JN, López S, Charris J, Iarruso L, Lobo G,Semenov A, et al. Synthesis and antimalarial effects ofphenothiazine inhibitors of a Plasmodium falciparumcysteine protease. J Med Chem 1997;40:2726-32.

50. Rosenthal PJ, Meshnick SR. Hemoglobin catabolismand iron utilization by malaria parasites. Mol BiochemParasitol 1996;83:131-9.

51. Slater AFG. Malaria pigment. Exp Parasitol1992;74:362-5.

52. Bond JS, Butler PE. Intracellular proteases. Annu RevBiochem 1987;56:333-64.

53. Gluzman IY, Francis SE, Oksman A, Smith CE, DuffinKL, Goldberg DE. Order and specificity of thePlasmodium falciparum hemoglobin degradationpathway. J Clin Invest 1994;93:1602-8.

54. Gyang FN, Poole B, Trager W. Peptidases fromPlasmodium falciparum cultured in vitro. MolBiochem Parasitol 1982;5:263-73.

55. Aissi E, Charet P, Bouquelet S, Biguet J. Endoproteasein Plasmodium yoelii nigeriensis. Comp BiochemPhysiol 1983;74B:559-66.

56. Sherman IW, Tanigoshi L. Purification of Plasmodiumlophurae cathepsin D and its effects on erythrocytemembrane proteins. Mol Biochem Parasitol 1983;8:207-26.

57. Sato K, Fukabori Y, Suzuki M. Plasmodium berghei: astudy of globinolytic enzyme in erythrocytic parasite.Zbl Bakt Hyg A 1987;264:487-95.

58. Bailly E, Savel J, Mahouy G, Jaureguiberry G.Plasmodium falciparum: isolation and characterizationof a 55-kDa protease with a cathepsin D-like activityfrom P. falciparum. Exp Parasitol 1991;72:278-84.

59. Goldberg DE, Slater AFG, Beavis R, Chait B, Cerami A,Henderson GB. Hemoglobin degradation in the humanmalaria pathogen Plasmodium falciparum: a catabolicpathway initiated by a specific aspartic protease. J ExpMed 1991;173:961-9.

60. Vander Jagt DL, Hunsaker LA, Campos NM, ScalettiJV. Localization and characterization of hemoglobin-degrading aspartic proteinases from the malarialparasite Plasmodium falciparum. Biochim BiophysActa 1992;1122:256-64.

61. Dame JB, Reddy GR, Yowell CA, Dunn BM, Kay J,Berry C. Sequence, expression and modeled structureof an aspartic proteinase from the human malariaparasite Plasmodium falciparum. Mol BiochemParasitol 1994;64:177-90.

62. Hill J, Tyas L, Phylip LH, Kay J, Dunn BM, Berry C.High level expression and characterisation ofplasmepsin II, an aspartic proteinase from Plasmodiumfalciparum. FEBS Lett 1994;352:155-8.

63. Dluzewski AR, Rangachari K, Wilson RJM, GratzerWB. Plasmodium falciparum: protease inhibitors andinhibition of erythrocyte invasion. Exp Parasitol1986;62:416-22.

64. Rosenthal PJ, McKerrow JH, Aikawa M, Nagasawa H,Leech JH. A malarial cysteine proteinase is necessaryfor hemoglobin degradation by Plasmodium falciparum.J Clin Invest 1988;82:1560-6.

65. Vander Jagt DL, Caughey WS, Campos NM, HunsakerLA, Zanner MA. Parasite proteases and antimalarialactivities of protease inhibitors. Prog Clin Biol Res1989;313:105-18.

66. Rosenthal PJ. Plasmodium falciparum: effects ofproteinase inhibitors on globin hydrolysis by culturedmalaria parasites. Exp Parasitol 1995;80:272-81.

67. Bailly E, Jambou R, Savel J, Jaureguiberry G.Plasmodium falciparum: differential sensitivity invitro to E-64 (cysteine protease inhibitor) andpepstatin A (aspartyl protease inhibitor). Journal ofProtozoology 1992;39:593-9.

68. Gamboa de Domínguez ND, Rosenthal PJ. Cysteineproteinase inhibitors block early steps in hemoglobindegradation by cultured malaria parasites. Blood1996;87:4448-54.

69. Asawamahasakda W, Ittarat I, Chang C-C, McElroy P,Meshnick SR. Effects of antimalarials and proteaseinhibitors on plasmodial hemozoin production. MolBiochem Parasitol 1994;67:183-91.

70. Rosenthal PJ, McKerrow JH, Rasnick D, Leech JH.Plasmodium falciparum: inhibitors of lysosomalcysteine proteinases inhibit a trophozoite proteinaseand block parasite development. Mol BiochemParasitol 1989;35:177-84.


Synopses

71. Salas F, Fichmann J, Lee GK, Scott MD, Rosenthal PJ.Functional expression of falcipain, a Plasmodiumfalciparum cysteine proteinase, supports its role as amalarial hemoglobinase. Infect Immun 1995;63:2120-5.

72. Francis SE, Gluzman IY, Oksman A, Banerjee D,Goldberg DE. Characterization of native falcipain, anenzyme involved in Plasmodium falciparum hemoglobindegradation. Mol Biochem Parasitol 1996;83:189-200.

73. Rosenthal PJ, Nelson RG. Isolation and characterizationof a cysteine proteinase gene of Plasmodiumfalciparum. Mol Biochem Parasitol 1992;51:143-52.

74. Kamchonwongpaisan S, Samoff E, Meshnick SR.Identification of hemoglobin degradation products inPlasmodium falciparum. Mol Biochem Parasitol1997;86:179-86.

75. Vander Jagt DL, Baack BR, Hunsaker LA. Purificationand characterization of an aminopeptidase fromPlasmodium falciparum. Mol Biochem Parasitol1984;10:45-54.

76. Curley GP, O’Donovan SM, McNally J, Mullally M,O’Hara H, Troy A, et al. Aminopeptidases fromPlasmodium falciparum, Plasmodium chabaudi, andPlasmodium berghei. J Eukaryot Microbiol 1994;41:119-23.

77. Kolakovich KA, Gluzman IY, Duffin KL, Goldberg DE.Generation of hemoglobin peptides in the acidicdigestive vacuole of Plasmodium falciparum implicatespeptide transport in amino acid production. MolBiochem Parasitol 1997;87:123-35.

78. Rockett KA, Playfair JHL, Ashall F, Targett GAT,Angliker H, Shaw E. Inhibition of intraerythrocyticdevelopment of Plasmodium falciparum by proteinaseinhibitors. FEBS Lett 1990;259:257-9.


Synopses

Tuberculosis (TB), one of the most wide-spread infectious diseases, is the leading cause ofdeath due to a single infectious agent amongadults in the world. Mycobacterium tuberculosisis the most common cause of human TB, but anunknown proportion of cases are due to M. bovis(1). In industrialized countries, animal TBcontrol and elimination programs, together withmilk pasteurization, have drastically reduced theincidence of disease caused by M. bovis in bothcattle and humans. In developing countries,however, animal TB is widely distributed, controlmeasures are not applied or are appliedsporadically, and pasteurization is rarely prac-ticed. The direct correlation between M. bovisinfection in cattle and disease in the humanpopulation has been well documented in

industrialized countries. Whereas little informa-tion is available from developing countries (2,3),risk factors for M. bovis in both animals andhumans are present in the tropics.

TB is a major opportunistic infection in HIV-infected persons (4). The vast majority of peoplecarrying this dual infection live in developingcountries; however, dual HIV and M. bovisinfection has been reported in industrializedcountries (5-11). The epidemic of HIV infection indeveloping countries, particularly countries inwhich M. bovis infection is present in animals andthe conditions favor zoonotic transmission, couldmake zoonotic TB a serious public health threatto persons at risk (3,12-14).

We summarize available epidemiologic infor-mation on TB and zoonotic TB, examine riskfactors that can influence the occurrence ofzoonotic TB in developing countries, and describethe most recent TB activities of the World HealthOrganization (WHO) (15-18).

Zoonotic Tuberculosis due toMycobacterium bovis inDeveloping Countries

O. Cosivi,* J.M. Grange,† C.J. Daborn,‡ M.C. Raviglione,* T. Fujikura,¶D. Cousins,§ R.A. Robinson,** H.F.A.K. Huchzermeyer,††

I. de Kantor,‡‡ and F.-X. Meslin**World Health Organization, Geneva, Switzerland; †Imperial College Schoolof Medicine, National Heart and Lung Institute, London, United Kingdom;‡Edinburgh University, Edinburgh, Scotland; ¶The University of Zambia,

Lusaka, Zambia; §Department of Agriculture, South Perth, Australia;**College of Veterinary Medicine, Veterinary Teaching Hospital,St. Paul, Minnesota, USA; ††Onderstepoort Veterinary Institute,

Onderstepoort, South Africa; and ‡‡Pan American Health Organization/World Health Organization, Buenos Aires, Argentina

The World Health Organization (WHO) estimates that human tuberculosis (TB)incidence and deaths for 1990 to 1999 will be 88 million and 30 million, respectively, withmost cases in developing countries. Zoonotic TB (caused by Mycobacterium bovis) ispresent in animals in most developing countries where surveillance and control activitiesare often inadequate or unavailable; therefore, many epidemiologic and public healthaspects of infection remain largely unknown. We review available information onzoonotic TB in developing countries, analyze risk factors that may play a role in thedisease, review recent WHO activities, and recommend actions to assess themagnitude of the problem and control the disease in humans and animals.

Address for correspondence: Ottorino Cosivi, Division ofEmerging and Other Communicable Diseases, Surveillanceand Control, CH-1211 Geneva 27, Switzerland; fax: 41-22-791-4893; e-mail: [email protected].


Synopses

Human TB: Global Situation and TrendsThe global incidence of TB is greatly

underestimated. In 1995, 3.3 million cases werereported to the Global Tuberculosis Programmeof WHO, whereas a more likely number is 8.8million. Of the reported cases, 62% occurred inthe Southeast Asian and Western Pacific regions,16% in sub-Saharan Africa, and 7% to 8% in eachof the regions of the Americas, EasternMediterranean, and Europe. Many countries,especially those with few resources, are unable toreport all TB cases because of difficulties inidentifying suspected cases, establishing adiagnosis, and recording and reporting cases.

In 1995, an estimated 8.8 million new TBcases occurred—5.5 million (62%) in theSoutheast Asian and Western Pacific regions and1.5 million (17%) in sub-Saharan Africa. Theannual global incidence is predicted to increase to10.2 million by the year 2000, an increase of36% from 1990. Southeast Asia, WesternPacific regions, and sub-Saharan Africa willaccount for 81% of these new cases (Table 1).For 1990 to 1999, in the absence of effectivecontrol, global TB incidence and deaths willreach 88 million and 30 million, respectively(19); 70% of the new cases will occur in patients15 to 59 years of age, the most economicallyproductive segment of the population.

As a result of the HIV epidemic, the crudeincidence rate of TB is expected to increase insub-Saharan Africa from 191 cases per 100,000 in

1990 to 293 in 2000. However, the total number ofnew cases will double by the year 2000. Becauseof the HIV epidemic, the decline of the crudeincidence rate in the Southeast Asian andCentral and South American regions is expectedto be slower than in previous years. Inindustrialized countries, a small increase incrude incidence rate and total cases is expected asthe result of immigration from countries with ahigh prevalence of dual HIV and TB infection.

The worldwide incidence of HIV-attributableTB cases is estimated to increase from 315,000(4% of the total TB cases) in 1990 to 1.4 million(14% of the total TB cases) by the year 2000. In2000, approximately 40% of these HIV-attribut-able cases will occur in sub-Saharan Africa and40% in Southeast Asia. Ten percent of the totalnumber of TB cases expected during 1990 to 1999are estimated to be attributable to HIV infection.

While demographic factors, such as popula-tion growth and changes in population structure,will largely account for the expected increase inTB incidence worldwide, the HIV epidemic insub-Saharan Africa will have a greater role thandemographic factors.

By the year 2000, 3.5 million persons will bedying of TB annually, an increase of 39% from1990. In Southeast Asia alone, 1.4 million deathswill occur annually. During 1990 to 1999, anestimated 30 million will die of TB, with 9.7% of thecases attributable to HIV infection. M. tuberculo-sis will be largely responsible for the new TB

Table 1. Estimated human tuberculosis and HIV-attributable tuberculosis cases in 1990, 1995, and 2000, by region (19)1990 1995 2000

HIV- HIV- HIV-Region TB cases Ratea attributed TB cases Rate attributed TB cases Rate attributedSoutheast Asia 3,106,000 237 66,000 3,499,000 241 251,000 3,952,000 247 571,000Western Pacificb 1,839,000 136 19,000 2,045,000 140 31,000 2,255,000 144 68,000Africa 992,000 191 194,000 1,467,000 242 380,000 2,079,000 293 604,000Eastern 641,000 165 9,000 745,000 168 16,000 870,000 168 38,000 MediterraneanAmericasc 569,000 127 20,000 606,000 123 45,000 645,000 120 97,000Eastern Europed 194,000 47 1,000 202,000 47 2,000 210,000 48 6,000Industrialized 196,000 23 6,000 204,000 23 13,000 211,000 24 26,000 countriese

Total TB cases 7,537,000 143 315,000 8,768,000 152 738,000 10,222,000 163 1,410,000 Attributed to HIV 4.2% 8.4% 13.8% Increase since 1990 16.3% 35.6%aRate: incidence of new cases per 100,000 population.bWestern Pacific Region of WHO except Japan, Australia, and New Zealand.cAmerican Region of WHO, except USA and Canada.dEastern European countries and independent states of the former USSR.eWestern Europe, USA, Canada, Japan, Australia, and New Zealand.


Synopses

cases and deaths, but an unknown, and potentiallyimportant, proportion will be caused by M. bovis.

Bovine TB in Developing CountriesAlthough prevalence data on animal TB in

developing countries are generally scarce,information on bovine TB occurrence and controlmeasures exists (20,21).

AfricaOf 55 African countries, 25 reported sporadic/

low occurrence of bovine TB; six reported enzooticdisease; two, Malawi and Mali, were described ashaving a high occurrence; four did not report thedisease; and the remaining 18 countries did nothave data (Figure 1).

Of all nations in Africa, only seven applydisease control measures as part of a test-and-slaughter policy and consider bovine TB anotifiable disease; the remaining 48 control thedisease inadequately or not at all (Figure 2).Almost 15% of the cattle population are found incountries where bovine TB is notifiable and atest-and-slaughter policy is used. Thus, approxi-mately 85% of the cattle and 82% of the humanpopulation of Africa are in areas where bovine TBis either partly controlled or not controlled at all.

AsiaOf 36 Asian nations, 16 reported a sporadic/

low occurrence of bovine TB, and one (Bahrain)described the disease as enzootic; ten did notreport bovine TB; and the remaining nine did nothave data (Figure 3). Within the Asian region,seven countries apply disease control measuresas part of a test-and-slaughter policy andconsider bovine TB notifiable. In the remaining29 countries, bovine TB is partly controlled or notcontrolled at all (Figure 4).

Of the total Asian cattle and buffalopopulations, 6% and less than 1%, respectively,are found in countries where bovine TB isnotifiable and a test-and-slaughter policy is used;94% of the cattle and more than 99% of thebuffalo populations in Asia are either only partlycontrolled for bovine TB or not controlled at all.Thus, 94% of the human population lives incountries where cattle and buffaloes undergo nocontrol or only limited control for bovine TB.

Latin American and Caribbean CountriesOf 34 Latin American and Caribbean

countries, 12 reported bovine TB as sporadic/lowoccurrence, seven reported it as enzootic, and one(Dominican Republic) described occurrence as

No data

Not reported

Sporadic

Enzootic

Figure 1. Bovine tuberculosis occurrence, Africa (21).

Not applied

Applied

Figure 2. Control measures for bovine tuberculosisbased on test-and-slaughter policy and diseasenotification, Africa (21).


Synopses

high. Twelve countries did not report bovine TB.No data were available for the remaining twocountries (Figure 5).

In the entire region, 12 countries apply diseasecontrol measures as part of a test-and-slaughterpolicy and consider bovine TB a notifiable disease.In the remaining 22 nations, the disease is partlycontrolled or not controlled at all (Figure 6). Theregional prevalence of bovine TB has beenestimated at 1% and higher in 67% of the total cattlepopulation and 0.1% to 0.9% in a further 7%; theremaining 26% are free of the disease or areapproaching the point of elimination (22).

Of the total Latin American and Caribbeancattle population, almost 76% is in countries wherebovine TB is notifiable and a test-and-slaughter

Figure 3. Bovine tuberculosis occurrence, Asia (21).

Figure 4. Control measures for bovine tuberculosisbased on test-and-slaughter policy and disease notifi-cation, Asia (21).

Figure 5. Bovine tuberculosis occurrence, Latin Americaand the Caribbean (21).

Figure 6. Control measures for bovine tuberculosisbased on test-and-slaughter policy and diseasenotification, Latin America and the Caribbean (21).


Synopses

policy is used. Thus, approximately 24% of thecattle population in this region is either only partlycontrolled for bovine TB or not controlled at all. It isalso estimated that 60% of the human populationlive in countries where cattle undergo no control oronly limited control for bovine TB.

Zoonotic TB in HumansTB caused by M. bovis is clinically

indistinguishable from TB caused by M.tuberculosis. In countries where bovine TB isuncontrolled, most human cases occur in youngpersons and result from drinking or handlingcontaminated milk; cervical lymphadenopathy,intestinal lesions, chronic skin TB (lupusvulgaris), and other nonpulmonary forms areparticularly common. Such cases may, however,also be caused by M. tuberculosis. Little is knownof the relative frequency with which M. boviscauses nonpulmonary TB in developing nationsbecause of limited laboratory facilities for theculture and typing of tubercle bacilli.

Agricultural workers may acquire the diseaseby inhaling cough spray from infected cattle; theydevelop typical pulmonary TB. Such patientsmay infect cattle, but evidence for human-to-human transmission is limited and anecdotal.

In regions where bovine TB has been largelyeliminated, a few residual cases occur amongelderly persons as a result of the reactivation ofdormant lesions. These are fewer than 1% of allTB cases. Surveys in the United States,Scandinavia, and South England have shownthat approximately half of these postprimarycases are pulmonary, a quarter involve thegenitourinary tract (a rare occurrence in primarydisease), and the remainder involve othernonpulmonary sites, notably cervical lymphnodes (23). In the same regions, approximately10% of cases caused by M. tuberculosis arenonpulmonary, although, for reasons that are notclear, the incidence is higher, approximately20%, in ethnic minority populations.

Information on human disease due to M. bovisin developed and developing countries is scarce.From a review of a number of zoonotic tuberculosisstudies, published between 1954 and 1970 andcarried out in various countries around the world, itwas estimated that the proportion of human casesdue to M. bovis acccounted for 3.1% of all forms oftuberculosis: 2.1% of pulmonary forms and 9.4% ofextrapulmonary forms (24). Table 2 summarizes

the findings of more recent reports of TB caused byM. bovis in industrialized countries.

Human disease caused by M. bovis has beenconfirmed in African countries. In an investiga-tion by two Egyptian health centers, theproportions of sputum-positive TB patientsinfected with M. bovis, recorded during threeobservations, were 0.4%, 6.4%, and 5.4% (33). Inanother study in Egypt, nine of 20 randomlyselected patients with TB peritonitis wereinfected with M. bovis, and the remaining withM. tuberculosis (34).

Isolation of M. bovis from sputum samples ofpatients with pulmonary TB has also beenreported from Nigeria. Of 102 M. tuberculosiscomplex isolates, 4 (3.9%) were M. bovis (35).Another study in Nigeria reported that one of 10mycobacteria isolated from sputum-positivecultures was M. bovis (36).

In a Zaire study, M. bovis was isolated fromgastric secretions in two of five patients withpulmonary TB (37). In the same study, theprevalence of the disease in local cattle wasapproximately 8% by tuberculin testing andisolation of M. bovis.

In a recent investigation in Tanzania, sevenof 19 lymph node biopsies from suspectedextrapulmonary TB patients were infected withM. tuberculosis and four with M. bovis (14). Nomycobacteria were cultured from the remaining

Table 2. Human tuberculosis due to Mycobacteriumbovis, industrialized countries

Cases% of Pulmonarytotal (% of total

Country (ref.) Years No. TB M. bovis)Australia (25) 1970-94 240 0.43-3.1 71.6a

England (23) 1977-90 232 1.2 40.0Germany (26) 1975-80 236 4.5 73.7Ireland Rural (27) 1986-90 17 6.4 70.6 Urban (28) 1982-85 9 0.9 88.8New Zealand (29) 1983-90 22 7.2 31.8Spain (30) 1986-90 10 0.9 50.0Sweden (17) 1983-92 96 2.0 -Switzerland (31) 1994 18 2.6 -U.S. (32) 1954-68 6 0.3 33.3U.S. (9) 1980-91 73 3.0 52.0b

12.0c

a Overall percentage includes 80.6 % males and 51.2%females.bAdults.cChildren.


Synopses

eight (Table 3). Although the number of sampleswas low, the high proportion (36%) of M. bovisisolates is of serious concern.

In an epidemiologic study in Zambia (38), anassociation between tuberculin-positive cattleand human TB was found. Households thatreported a TB case within the previous 12 monthswere approximately seven times more likely toown herds containing tuberculin-positive cattle(odds ratio = 7.6; p = 0.004). Although this couldbe explained by zoonotic TB transmission, otherfactors such as transient sensitivity to tuberculinof cattle exposed to TB patients and coincidentalenvironmental factors favoring both humanclinical TB and sensitivity to bovine tuberculinshould also be considered.

In Latin America, a conservative estimatewould be that 2% of the total pulmonary TB casesand 8% of extrapulmonary TB cases are caused byM. bovis. These cases would therefore account for7,000 new TB cases per year, a rate of nearly 2 per100,000 inhabitants. From a nationwide study inArgentina during 1982 to 1984, 36 (0.47%) of 7,672mycobacteria cultured from sputum samples wereM. bovis (39). However, in another study in SantaFe province (where most of the dairy cattle industryis concentrated) during 1984 to 1989, M. boviscaused 0.7% to 6.2% of TB cases (40).

Very limited data on the zoonotic aspects ofM. bovis are available from Asian countries.However, cases of TB caused by M. bovis were notreported in early investigations in India (41).

EpidemiologyMuch information on the epidemiologic

patterns of zoonotic TB has been obtained in thiscentury from industrialized countries. However,some striking epidemiologic differences related toboth animal and human populations in developingcountries require particular attention.

Risk Factors: Animal Population

Animal reservoirs . The widespread distribu-tion of M. bovis in farm and wild animalpopulations represents a large reservoir of thismicroorganism. The spread of the infectionfrom affected to susceptible animals in bothindustrialized and developing countries is mostlikely to occur when wild and domesticatedanimals share pasture or territory (42). Well-documented examples of such spread includeinfection in badgers (Meles meles) in the UnitedKingdom and possums (Trichosurus vulpecula)in New Zealand. Wild animal TB represents apermanent reservoir of infection and poses aserious threat to control and elimination programs.

Milk production and animal husbandry. Milkproduction has increased in most developingcountries as a consequence of greater demand formilk for human consumption (43; Figure 7). Thisincreased demand for milk—estimated at 2.5%per year for 1970 to 1988 for sub-Saharan Africa(44)—led to increases in the number of productiveanimals and milk imports and intensification ofanimal production through the introduction ofmore productive exotic breeds.

Although the prevalence of the disease withina country varies from area to area, the highestincidence of bovine TB is generally observed

Table 3. Isolates from suspected extrapulmonarytuberculosis patients, Tanzania, 1994 (14)

No. of M. tuber- M.Occupation samples culosis bovis Neg.Livestock 4 0 2 2 keeperFarmers 6 2 1 3Children 3 2 1 0Unknown 6 3 0 3Total 19 7 4 8

Figure 7. Cow milk production by region (43).

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

1979-81 1992 1993 1994

Years

Ton

nes

Africa

Latin American andCaribbean countries

Asia

0


Synopses

where intensive dairy production is mostcommon, notably in the milksheds of larger cities(1). This problem is exacerbated where there isinadequate veterinary supervision, as is thecase in most developing countries. In addition,in some industrialized countries such as theUnited States, where bovine TB is close toelimination, large dairy herds (i.e., 5,000 ormore cows) that are crowded together repre-sent the main source of infection (45).

In developing countries, bovine TB infects ahigher proportion of exotic dairy breeds (Bostaurus) than indigenous zebu cattle (Bos indicus)and crossbred beef cattle (1). However, underintensive feedlot conditions, a death rate of 60%and depression of growth have been found intuberculous zebu cattle (46). In those areas whereextensive management is more common, animalcrowding (e.g., near watering ponds, dip tanks,markets, and corrals) still plays a major role inthe spread of the disease.

Control measures and programs. The basicstrategies required for control and elimination ofbovine TB are well known and well defined (47).However, because of financial constraints,scarcity of trained professionals, lack of politicalwill, as well as the underestimation of theimportance of zoonotic TB in both the animal andpublic health sectors by national governmentsand donor agencies, control measures are notapplied or are applied inadequately in mostdeveloping countries.

Successful conduct of a test-and-slaughterpolicy requires sustained cooperation of nationaland private veterinary services, meat inspectors,and farmers, as well as adequate compensationfor services rendered. Only a few developingcountries can adhere to these requirements.

In addition, bovine TB does not often justifythe emergency measures required for otherzoonotic diseases (e.g., Rinderpest, East Coastfever, and foot and mouth disease). The fulleconomic implications of zoonotic TB are,however, overlooked in many developing nationswhere the overall impact of the disease on humanhealth and animal production needs to be assessed.According to recent estimates, annual economicloss to bovine TB in Argentina is approximately 63million US dollars (48). In a study recentlyconducted in Turkey, the estimated socioeconomicimpact of bovine TB to both the agriculture and

health sectors was approximately 15 to 59 millionUS dollars per year (49).

Several Latin American countries, throughagreements between governments and cattleowners associations, have made the decision tocontrol and eliminate bovine TB. Where foot andmouth disease has been eliminated, bovine TBand other existing infections such as brucellosisbecome important because of their impact on themeat and live animal export trade. Bovine TB andbrucellosis also limit the development of the dairyindustry and its expansion at the regional level.

Risk Factors: Human Population

Close physical contact. Close physicalcontact between humans and potentially infectedanimals is present in some communities,especially in developing regions. For example, inmany African countries cattle are an integralpart of human social life; they represent wealthand are at the center of many events and,therefore, gatherings. In addition, with 65% ofAfrican, 70% of Asian, and 26% of Latin Americanand Caribbean populations working in agricul-ture, a significant proportion of the population ofthese regions may be at risk for bovine TB.

Food hygiene practices. Consumption ofmilk contaminated by M. bovis has long beenregarded as the principal mode of TB transmis-sion from animals to humans (1). In regionswhere bovine TB is common and uncontrolled,milkborne infection is the principal cause ofcervical lymphadenopathy (scrofula) and ab-dominal and other forms of nonpulmonary TB.Although proper food hygiene practices couldplay a major role in controlling these forms of TB,such practices are often difficult to institute indeveloping countries.

In all countries of sub-Saharan Africa, thereis active competition between large-scale, oftenstate-run, processing and marketing enterprisesand the informal sector. The informal sector canignore standards of hygiene and quality, andproducers often sell directly to the finalconsumers. In addition, an estimated 90% of thetotal milk produced is consumed fresh or soured(44). Although it has been stated that Africansgenerally boil milk and that the souring processdestroys M. bovis (44), other sources stronglycontradict these statements (39). M. bovis was


Synopses

isolated from seven (2.9%) of 241 samples of rawmilk in Ethiopia (17). Both M. bovis andM. tuberculosis have also been found in milksamples in Nigeria (36) and Egypt (34). Thus,serious public health implications of poten-tially contaminated milk and milk productsshould not be underestimated.

HIV/AIDS. According to recent WHO globalestimates, of the 9.4 million people infected withboth HIV and TB in mid-1996, 6.6 million (70%)live in sub-Saharan Africa (4). The greatestimpact of HIV infection on TB is in populationswith a high prevalence of TB infection amongyoung adults. The occurrence of both infections inone person makes TB infection very likely toprogress to active disease.

In many developing countries, TB is the mostfrequent opportunistic disease associated withHIV infection. HIV seroprevalence rates greaterthan 60% have been found in TB patients invarious African countries (4). Persons infectedwith both pathogens have an annual risk ofprogression to active TB of 5% to 15%, depending ontheir level of immunosuppression; approximately10% of non-HIV infected persons newly infectedwith TB become ill at some time during their lives.In the remaining 90%, effective host defensesprevent progression from infection to disease.

TB cases due to M. bovis in HIV-positivepersons also resemble disease caused byM. tuberculosis. Thus, they manifest aspulmonary disease, lymphadenopathy, or, inthe more profoundly immunosuppressed, dis-seminated disease.

M. bovis has been isolated from HIV-infectedpersons in industrialized countries. In France, M.bovis infection accounted for 1.6% of TB cases inHIV-positive patients. All isolated strains wereresistant to isoniazid (7). Taking into consider-ation the intrinsic resistance of M. bovis topyrazinamide, two of the first-line anti-TB drugswere not effective. WHO-recommended standardtreatment for new TB cases includes, in the initialphase, isoniazid, rifampicin, pyrazinamide, andstreptomycin or ethambutol. In situations of highprimary resistance to isoniazid and streptomy-cin, the intrinsic resistance of M. bovis topyrazinamide may severely limit the efficacy oftreatment of TB caused by M. bovis.

In a Paris hospital, a source patient withpulmonary TB due to a multidrug-resistantstrain of M. bovis led to active disease in five

patients. Disease occurred 3 to 10 months afterinfection (10). This observation led to threeconcerns: 1) human-to-human M. bovis transmis-sion leading to overt disease, 2) a short intervalbetween infection and overt disease, and 3)disseminated multidrug-resistant M. bovis.

In another study, conducted in San Diego,California, one of 24 adults with pulmonary TB and11 of 24 adults with nonpulmonary TB due to M.bovis had AIDS. One of 25 children, a 16-year-oldboy with abdominal TB, was also HIV-positive (9).

It is commonly believed that M. bovis is lessvirulent than M. tuberculosis in humans andtherefore less likely to lead to overt postprimarydisease and that human-to-human transmissionleading to infectious disease is rare. However, ifthe apparent difference in virulence is the resultof differences in responsiveness of the hostdefense mechanisms, HIV-induced immunosup-pression could well lower host defenses leading toovert disease after infection.

Surveillance of TB due to M. bovisThe use of direct smear microscopy as the

only method for diagnosis of suspected TB,although an essential requirement of anynational TB program, could partly explain therelatively low notification rate of disease causedby M. bovis in developing countries. Direct smearmicroscopy does not permit differentiationbetween species of the M. tuberculosis complex;in addition, culture and speciation are often notcarried out, and even when culture facilities areavailable, M. bovis grows poorly in standardLöwenstein-Jensen medium, one of the mostwidely used culture media (50). In somecountries, human disease caused by M. bovis ismerely reported as TB to avoid inquiries fromdisease control agencies, which might generateproblems of patient confidentiality (2).

The collection of representative data on theincidence of TB due to M. bovis from mostlaboratories in developing countries has addi-tional problems. For example, the location andcoverage of laboratories are often biased towardscity populations; sputum specimens may pre-dominate, with relatively few specimens fromextrapulmonary lesions, particularly amongchildren. Specimens from children with TB arefrequently negative on culture, and biopsies aredifficult to take from lesions.

Recent outbreaks of multidrug-resistant TBin some parts of the world underscore the need for


Synopses

surveillance through wider application of reliableculture and drug susceptibility tests.

Control Measures and Programs inDeveloping Countries

Bovine TB can be eliminated from a countryor region by implementing a test-and-slaughterpolicy, if no other reservoir host of infectionexists. While the test-and-slaughter policy islikely to remain the backbone of nationalelimination bovine TB programs, the policy hasnumerous constraints in developing countries.Alternative strategies (e.g., programs based onslaughterhouse surveillance and traceback oftuberculous animals to herds of origin) may betechnically and economically more appropriate inthese countries.

Measures to prevent transmission of infectionshould be the primary objective to be achieved withtrained public health personnel, public education,and proper hygienic practices. Test-and-slaughterprograms may be feasible and appropriate in areaswith low bovine TB prevalence and effective controlof animal movement.

Animal Vaccination and ResearchDevelopments

Although not usually considered relevant toelimination programs in livestock (47), vaccinationof animals against TB would be a viable strategy intwo disease control situations: in domesticatedanimals in developing countries and in wildlife andferal reservoirs of disease in industrializedcountries where test-and-slaughter programs havefailed to achieve elimination of the disease.

Many issues need to be addressed beforevaccination becomes a realistic option for controlof disease in cattle and other animals. First, ahighly effective vaccine needs to be developed.The results obtained globally with bacillusCalmette-Guérin (BCG) have been suboptimal,and efficacy has varied considerably from regionto region (42,51). Secondly, the delivery of thevaccine poses few problems in domesticatedanimals, but it is fraught with difficulties in wildanimals. Thirdly, vaccination may compromisediagnostic tests. A vaccine that induces tubercu-lin reactivity would invalidate the key diagnostictool used in control programs. Fourthly, short ofperforming lengthy and expansive field studies,evaluation of the protective efficacy of a newvaccine will pose serious difficulties. Tradition-ally, the guinea pig and mouse have been used for

this purpose, but the information gained has beenof little value. Recent work has, however,indicated that deer may well prove a suitablemammal for evaluating new vaccines andoptimum delivery systems (52).

Enzyme-linked immunosorbent assay andgamma-interferon tests may prove to be moresensitive and specific than the tuberculin test andmay facilitate diagnostic procedures. Nucleicacid-based technology, notably polymerase chainreaction and related methods, may provide morerapid, sensitive, and specific diagnostic tools.Multicenter studies of the applicability of thesetechniques to the diagnosis of human TB have,however, shown that their sensitivity andspecificity are not as high as originally expectedand that many problems need to be solved beforethe techniques are introduced into routinelaboratory practice (53). Restriction fragmentlength polymorphism analysis (DNA fingerprint-ing) could be useful in epidemiologic studies thattrace the spread of disease between cattle, otheranimals, and humans (54) or in the rapiddifferentiation of M. bovis within the M.tuberculosis complex (55). The use of thesetechniques is limited by resources in mostdeveloping countries.

WHO and Zoonotic TBThe public health importance of animal TB

was recognized early by WHO, which in its 1950report of the Expert Committee on Tuberculosis(56) stated: “The committee recognizes theseriousness of human infection with bovinetuberculosis in countries where the disease incattle is prevalent. There is the danger oftransmission of infection by direct contactbetween diseased cattle and farm workers andtheir families, as well as from infected foodproducts.” Since then, TB in animals has beencontrolled and almost eliminated in severalindustrialized countries but in very fewdeveloping countries.

More recently, WHO has been involved inzoonotic TB through the activities of the Divisionof Emerging and other Communicable DiseasesSurveillance and Control at WHO in Geneva(WHO/EMC) and the Veterinary Public Healthprogram of the WHO Regional Office for theAmericas, Pan American Health Organization(PAHO/HCV).

WHO/EMC has organized and coordinated aworking group of experts from countries


Synopses

worldwide (15-17). Their subjects are epidemiol-ogy, public health aspects, control, and researchon zoonotic TB. In addition, a joint WHO, Foodand Agriculture Organization of the UnitedNations (FAO), and Office International desEpizooties (OIE) Consultation on Animal Tuber-culosis Vaccines was held to review currentknowledge on TB vaccine development forhumans and animals and make recommenda-tions for animal TB vaccine research anddevelopment (57). Promising results of cattlevaccination with low doses of BCG were reported.It is also planned for field trial cattle vaccinationto commence early in 1998 in Madagascar incollaboration with national and internationalresearch institutions, OIE and WHO. In theframework of the working group activities, theguidelines for speciation within the Mycobacte-rium tuberculosis complex (50) have beenprepared to respond to the growing need forreliable differentiation between M. tuberculosis,M. africanum, and M. bovis and to promote andstrengthen surveillance.

A Plan of Action for the Eradication of BovineTuberculosis in the Americas (18) has beendeveloped by PAHO in collaboration withmember countries of the region. PAHO/HCV, incooperation with the Pan American Institute forFood Protection and Zoonosis (INPPAZ), BuenosAires, Argentina, and other technical institutions(e.g., FAO), provides technical support to theregional plan. PAHO/HCV activities trainspecialists in diagnosis, reporting, surveillancesystems, and quality control of reagents, as wellas supporting the planning and implementationof national programs. INPPAZ acts as a referencecenter for these activities. The first phase of theregional plan is expected to lead, in the next 10years, to the elimination of bovine TB fromcountries with more advanced national pro-grams. In the remaining countries, the objectiveswill be to strengthen epidemiologic surveillance,defining areas at risk and setting up control andelimination programs.

ConclusionsAlthough the epidemiology of bovine TB is

well understood and effective control andelimination strategies have been known for a longtime, the disease is still widely distributed andoften neglected in most developing countries. Itspublic health consequences, although well

documented from the past experiences ofindustrialized countries, have scarcely beeninvestigated and are still largely ignored in theseregions. Because of the animal and public healthconsequences of M. bovis, disease surveillanceprograms in humans should be considered apriority, especially in areas where risk factorsare present. The increase of TB in such areascalls for stronger intersectoral collaborationbetween the medical and veterinary profes-sions to assess and evaluate the scale of theproblem, mostly when zoonotic TB couldrepresent a significant risk, for example, inrural communities and in the workplace.

Industrialized countries, where the test-and-slaughter policies have not completely elimi-nated infection in cattle because of wild animalreservoirs, are now reconsidering wild animalvaccination. Any vaccination research anddevelopment program should therefore also takeinto account the possible application of vaccinesto cattle, particularly in developing countries.

In developing countries, where HIV andbovine TB are likely to be common, particularly inyoung persons, the ability of HIV infection toabrogate any host factors that prevent theprogression of infection by M. bovis to overtdisease may lead to higher incidence and case-fatality rates for human TB caused by this speciesand increased human-to-human transmission ofthis disease. This should be of great concern inthose developing countries where bovine TB ispresent and measures to control spread ofinfection are not applied or are appliedinadequately. Research is needed to determinewhen M. bovis is of zoonotic importance and whatthe underlying mechanisms of transmission are.Locally operative risk factors for zoonotic TBshould therefore be identified to determinepersons at risk and develop appropriate controlmeasures. International cooperation in allaspects of zoonotic TB remains essential in thefight against this disease.

References 1. Acha PN, Szyfres B. Zoonotic tuberculosis. In:

Zoonoses and communicable diseases common to manand animals. 2nd edition. Washington: Pan AmericanHealth Organization/World Health Organization;1987: Scientific Publication No. 503.

2. Collins CH, Grange JM. A review. The bovine tuberclebacillus. J Appl Bacteriol 1983;55:13-29.


Synopses

3. Cosivi O, Meslin F-X, Daborn CJ, Grange JM. Theepidemiology of Mycobacterium bovis infection in animalsand humans, with particular reference to Africa. Scientificand Technical Review 1995;14:733-46.

4. Raviglione MC, Snider DE, Kochi A. Globalepidemiology of tuberculosis. JAMA 1995;273:220-6.

5. Houde C, Dery P. Mycobacterium bovis sepsis in aninfant with human immunodeficiency virus infection.Pediatr Infect Dis J 1988;7:810-2.

6. Cornuz J, Fitting JW, Beer V, Chave JP. Mycobacteriumbovis and AIDS. AIDS 1991;5:1038-9.

7. Dupon M, Ragnaud JM. Tuberculosis in patientsinfected with human immunodeficiency virus 1. Aretrospective multicentre study of 123 cases in France.Quarterly Journal of Medicine [New Series 85]1992;306:719-30.

8. Yates MD, Pozniak A, Grange JM. Isolation ofmycobacteria from patients seropositive for the humanimmunodeficiency virus (HIV) in south east England:1984-92. Thorax 1993;48:990-5.

9. Dankner WM, Waecker NJ, Essey MA, Moser K,Thompson M, Davis CH. Mycobacterium bovisinfections in San Diego: a clinicoepidemiologic study of73 patients and a historical review of a forgottenpathogen. Medicine (Baltimore) 1993;72:11-37.

10. Bouvet E, Casalino E, Mendoza-Sassi G, Lariven S,Vallee E, Pernet M, et al. A nosocomial outbreak ofmultidrug-resistant Mycobacterium bovis among HIV-infected patients. A case-control study. AIDS1993;7:1453-60.

11. Boletín Epidemiológico Semanal. Vigilancia de latuberculosis zoonótica por M. bovis: sistema deenfermedades de declaración obligatoria 1982-1995.Red Nacional de Vigilancia Epidemiológica de España,Centro Nacional de Epidemiología 1995;3:183-4.

12. Grange JM, Daborn C, Cosivi O. HIV-relatedtuberculosis due to Mycobacterium bovis. Eur Respir J1994;7:1564-6.

13. Moda G, Daborn CJ, Grange JM, Cosivi O. The zoonoticimportance of Mycobacterium bovis. Tubercle andLung Disease 1996;77:103-8.

14. Daborn CJ, Grange JM, Kazwala RR. The bovinetuberculosis cycle—an African perspective. J ApplBacteriol [Symposium Supplement] 1997;81:27s-32s.

15. World Health Organization. Report of the WHOmeeting on animal tuberculosis; 1992 Apr 27; Cairo,Egypt. Geneva: The Organization; 1992. Unpub.document WHO/CDS/VPH/92.112.

16. World Health Organization. Report of the WHOmeeting on zoonotic tuberculosis (Mycobacteriumbovis), with the participation of the FAO; 1993 Nov 15.Geneva, Switzerland. Geneva: The Organization; 1993.Unpub. document WHO/CDS/VPH/93.130.

17. World Health Organization. Report of WHO WorkingGroup on zoonotic tuberculosis (Mycobacterium bovis),with the participation of FAO; 1994 June 14. Mainz,Germany. Geneva: The Organization; 1994. Unpub.document WHO/CDS/VPH/94.137.

18. World Health Organization/Pan American HealthOrganization. VIII Inter-American meeting, at theministerial level on animal health; 1993 Apr 27-29.Washington, DC. Washington: The Organizations;1993. Document RIMSA 8/26.

19. Dolin PJ, Raviglione MC, Kochi A. Global tuberculosisincidence and mortality during 1990-2000. Bull WorldHealth Organ 1994;72:213-20.

20. Thoen CO, Steele JH, editors. Regional and CountryStatus Reports. Part 2. Mycobacterium bovis infectionin animals and humans. Ames (IA): Iowa StateUniversity Press; 1995. p. 167-345.

21. Food and Agriculture Organization. In: Welte WR,editor. FAO/OIE/WHO Animal Health Yearbook 1993.Washington: The Organization; 1994. FAO AnimalProduction and Health Series No. 33.

22. de Kantor IN, Ritacco V. Bovine tuberculosis in LatinAmerica and the Caribbean: current status, controland eradication programs. Vet Microbiol 1994;40:5-14.

23. Grange JM, Yates MD. Zoonotic aspects ofMycobacterium bovis infection. Vet Microbiol1994;40:137-51.

24. Gervois M, Vaillant JM, Fontaine JF, Laroche G,Dubois G. Epidemiologie de l’infection humaine parMycobacterium bovis. Archivio Monaldi per latisiologia e le malattie dell’apparato respiratorio1972;27:294-317.

25. Cousins DV, Williams SN. A study of Mycobacteriumbovis infection in Australian patients 1970-1994. In:Griffit F, de Lisle G, editors. Tuberculosis in wildlifeand domestic animals. Otago, New Zealand: Universityof Otago Press: 1995. p. 260-3.

26. Krebs A, Käppler W. Die Bedeutung von Mycobacteriumbovis in der Tuberkuloseepidemiologie. Zeitschift fürErkrankungen der Atmungs Organe 1982;158:101-9.

27. Cormican MG, Flynn J. Tuberculosis in the West ofIreland 1986-1990. Ir J Med Sci 1992;161:70-2.

28. Collins C, Kelly P, Bryne C, Denham F, Clancy L. Isbovine, atypical or resistant tuberculosis a problem? IrMed J 1987;80:66-7.

29. Brett JL, Humble MW. Incidence of humantuberculosis caused by Mycobacterium bovis. N Z MedJ 1991;104:13-4.

30. Sauret J, Jolis R, Ausina V, Castro E, Cornudella R.Human tuberculosis due to Mycobacterium bovis: reportof 10 cases. Tubercle and Lung Disease 1992;73:388-91.

31. La tuberculose en Suisse en 1994. Bern,Switzerland: Bulletin 37, Office fédéral de la santépublique;1994. p.10-1.

32. Karlson AG, Carr DT. Tuberculosis caused byMycobacterium bovis. Report of six cases: 1954-1968.Ann Intern Med 1970;73:979-83.

33. Elsabban MS, Lofty O, Awad WM, Soufi HS, MikhailDG, Hammam HM, et al. Bovine tuberculosis and itsextent of spread as a source of infection to man andanimals in Arab Republic of Egypt. In: Proceedings ofthe International Union Against Tuberculosis andLung Disease Conference on Animal Tuberculosis inAfrica and the Middle East; 1992 Apr 28-30; Cairo,Egypt. Paris: The Union; 1992. p.198-211.

34. Nafeh MA, Medhat A, Abdul-Hameed A-G, Ahmad YA,Rashwan NM, Strickland GT. Tuberculous peritonitisin Egypt: the value of laparoscopy in diagnosis. Am JTrop Med Hyg 1992;47:470-7.

35. Idigbe EO, Anyiwo CE, Onwujekwe DI. Humanpulmonary infections with bovine and atypicalmycobacteria in Lagos, Nigeria. Journal of TropicalMedicine and Hygiene 1986;89:143-8.


Synopses

36. Idrisu A, Schnurrenberger P. Public health significanceof bovine tuberculosis in four northern states ofNigeria: a mycobacteriologic study. Nigerian MedicalJournal 1977;7:384-7.

37. Mposhy M, Binemo-Madi C, Mudakikwa B. Incidencede la tuberculose bovine sur la santé des populations duNord-Kivu (Zaïre). Rev Elev Med Vet Pays Trop1983;36:15-8.

38. Cook AJC, Tuchili LM, Buve A, Foster SD, Godfrey-Faussett P, Pandey GS, McAdam KPWJ. Human andbovine tuberculosis in the Monze district of Zambia—across-sectional study. Br Vet J 1996;152:37-46.

39. Barrera L, de Kantor IN. Nontuberculous mycobacteriaand Mycobacterium bovis as a cause of human diseasein Argentina. Tropical and Geographical Medicine1987;39:222-7.

40. Sequeira de Latini MD, Latini OA, Lopez ML, Cecconi JO.Tuberculosis bovina en seres humanos. 2a. parte: Periodo1977-1989. Revista Argentina del Torax 1990;51:13-1.

41. Lall JM. Tuberculosis among animals in India. TheVeterinary Bulletin 1969;39:385-90.

42. O’Reilly LM, Daborn CJ. The epidemiology ofMycobacterium bovis infections in animals and man: areview. Tubercle and Lung Disease 1995;76(Suppl 1):1-46.

43. Food and Agriculture Organization. Production. Vol.47. Rome: Food and Agriculture Organization of theUnited Nations; 1993. FAO Statistic Series No. 117.

44. Walshe MJ, Grindle J, Nell A, Bachmann M. Dairydevelopment in sub-Saharian Africa. WashingtonDC: World Bank, 1991. African Technical DepartmentSeries; World Bank Technical Paper No. 135.

45. National Research Council. Livestock diseaseeradication—evaluation of the cooperative State-Federal Bovine Tuberculosis Eradication Program.Washington: National Academy Press; 1994.

46. Blancou J, Cheneau Y. Influence de la tuberculose surle gain de poids de zébus à l’engrais. Rev Elev Med VetPays Trop 1974;27:75-80.

47. World Health Organization. Third Report of the JointFAO/WHO Expert Committee on Zoonoses. Geneva: TheOrganization; 1967. Technical Report Series No. 378.

48. Maggi C, Alvarez E, de Kantor I, Nader A.Methodology for estimating the losses caused bybovine tuberculosis in Argentina. Scientific andTechnical Review. In press 1998.

49. Barwinek F, Taylor NM. Assessment of the socio-economic importance of bovine tuberculosis in Turkeyand possible strategies for control or eradication.Bakanliklar, Ankara, Turkey: Turkish-German AnimalHealth Information Project, General Directorate ofProtection and Control; 1996.

50. Grange JM, Yates MD, de Kantor I. Guidelines forspeciation within the Mycobacterium tuberculosiscomplex. 2nd ed. Geneva: World Health Organization;1996. Unpub. document WHO/EMC/ZOO/96.4.

51. Malin AS, Young DB. Designing a vaccine fortuberculosis. Unraveling the tuberculosis genome—can we build a better BCG? Br Med J 1996;312:1495.

52. Griffin JFT, Mackintosh CG, Buchan GS. Animal modelsof protective immunity in tuberculosis to evaluatecandidate vaccines. Trends Microbiol 1995;3:418-24.

53. Doucet-Populaire F, Lalande V, Carpentier E,Bourgoin A, Dailloux M, Bollet A, et al. A blind study ofthe polymerase chain reaction for the detection ofMycobacterium tuberculosis DNA. Tubercle and LungDisease 1996;77:358-62.

54. van Embden JDA, Schouls LM, van Soolingen D.Molecular techniques: application in epidemiologicstudies. In: Thoen CO, Steele JH, editors.Mycobacterium bovis infection in animals and humans.Ames (IA): Iowa State University Press; 1995. p. 15-27.

55. Liebana E, Aranaz A, Francis B, Cousins D.Assessment of genetic markers for speciesdifferentiation within the Mycobacterium tuberculosiscomplex. J Clin Microbiol 1996;34:933-8.

56. World Health Organization. Expert Committee onTuberculosis report on the Fifth Session; 1950 Sep 11-16; Geneva, Switzerland. Geneva: The Organization;1951. Technical Report Series 32.

57. World Health Organization. Report of a WHO/FAO/OIE consultation on animal tuberculosis vaccines;1994 Aug 3-5; Geneva, Switzerland. Geneva: TheOrganization; 1994. Unpub. document WHO/CDS/VPH/94.138.


Synopses

Cryptococcus neoformans is an encapsulatedfungal organism (Figure 1) that can cause diseasein apparently immunocompetent, as well asimmunocompromised, hosts (1,2). Most suscep-tible to infection are patients with T-celldeficiencies (1,2). C. neoformans var. neoformanscauses most cryptococcal infections in humans, sothis review will focus on information from theneoformans variety of this basidiomycetousfungus. C. neoformans var. neoformans is foundworldwide; its main habitats are debris aroundpigeon roosts and soil contaminated with decayingpigeon or chicken droppings (1,3). Not part of thenormal microbial flora of humans, C. neoformansis only transiently isolated from persons with nopathologic features (2,4). It is generally acceptedthat the organism enters the host by therespiratory route in the form of a dehydratedhaploid yeast or as basidiospores. After sometime in the lungs, the organism hematogenouslyspreads to extrapulmonary tissues; since it has apredilection for the brain, infected personsusually contract meningoencephalitis (1). Ifuntreated, cryptococcal meningoencephalitis is100% fatal, and even when treated with the mosteffective antifungal drugs, cryptococcal infec-tions can be fatal if the host does not haveadequate T-cell-dependent immune function (2).

What Makes Cryptococcus neoformansa Pathogen?

Kent L. Buchanan and Juneann W. MurphyUniversity of Oklahoma Health Sciences Center,

Oklahoma City, Oklahoma, USA

Address for correspondence: Juneann W. Murphy, Universityof Oklahoma Health Sciences Center, Oklahoma City, OK,USA; fax: 405-271-3117; e-mail: [email protected] [email protected].

Life-threatening infections caused by the encapsulated fungal pathogenCryptococcus neoformans have been increasing steadily over the past 10 yearsbecause of the onset of AIDS and the expanded use of immunosuppressive drugs.Intricate host-organism interactions make the full understanding of pathogenicity andvirulence of C. neoformans difficult. We discuss the current knowledge of thecharacteristics C. neoformans must possess to enter the host and establish progressivedisease: basic growth requirements and virulence factors, such as the polysaccharidecapsule; shed products of the organism; melanin production; mannitol secretion;superoxide dismutase; proteases; and phospholipases.

Figure 1. Transmission electron micrograph ofbudding C. neoformans showing the characteristicpolysaccharide capsule.


Synopses

To be classified as a pathogen, an organismmust be able to cause infection under certainconditions. By this definition, C. neoformans cancertainly be classified as a pathogen. Because theimmunodeficient are more susceptible than theimmunocompetent to infection with this yeast-like organism, C. neoformans is frequentlyreferred to as an opportunistic pathogen. Thefactors that make C. neoformans a pathogen canbe divided into two major groups. The firstcomprises the basic characteristics needed toestablish an infection and survive in the humanhost; the second comprises the virulence factorsthat affect the degree of pathogenicity.

Basic Characteristics for Pathogenicity

The Infectious ParticleTo enter the alveolar spaces of the lungs and

establish pulmonary infection, an organism mustproduce viable forms smaller than 4 µm indiameter. The typical vegetative form of C.neoformans is the yeast form with a cell diameterof 2.5 µm to 10 µm. The organism can alsoundergo sexual reproduction, and since it is abasidiomycete (Filobasidiella neoformans), itforms basidiospores. Sexual reproduction ap-pears to occur much less frequently in naturethan asexual or vegetative reproduction. Thesexual spores (basidiospores) are approximately1.8 µm to 3 µm in diameter and result from crossesof the α- and a-mating types on an appropriatemedium (1). Although the exact nature of theinfectious particles of C. neoformans is notknown, they are presumed to be the dehydratedyeast cells or basidiospores of the appropriatesize range to get into the lungs. Once inside thelungs, the yeast cells become rehydrated andacquire the characteristic polysaccharide capsule(Figure 2). In the case of basidiospores, theywould convert to encapsulated blastoconidia.

Recently, Wickes et al. (5) reported that theα-mating type of C. neoformans can producemonokaryotic hyphae on a solid medium withouta nitrogen source or water. The monokaryotichyphae contain clamp connections and basidiawith short chains of basidiospores and are similarin all respects except in nuclei number to thedikaryotic hyphae of the sexual forms (5). Thehaploid fruiting bodies formed by nitrogen-starvedα-mating type isolates also produce abundantamounts of blastoconidia, which on the appropri-ate medium can produce haploid hyphae (5).

None of the a-mating type strains testedproduced the haploid fruiting bodies. Backcrossstudies indicated that the ability to undergohaploid fruiting lies within (or is tightly linked to)the MATα locus, which is responsible for the α-mating type phenotype (5). These results suggestbasidiospores from the monokaryotic hyphaemay be another source of the infectious particle innature. Regardless of the nature of the infectiousparticle (yeast cell or basidiospore), if a cryptococcalisolate is not able to produce small infectiousparticles, it cannot be pathogenic under the usualconditions for establishing the disease.

Mating TypesKwon-Chung and Bennett (6) surveyed the

mating types of C. neoformans isolates fromenvironmental and clinical sources and found a30- to 40-fold higher frequency of α-mating typethan a-mating type. The skewed ratio of α-matingtype to a-mating type was not thought to be due toa genetic preponderance towards α-mating typeprogeny because crosses between two test strainsresulted in a 1:1 ratio of α-mating and a-matingtype progeny (7). To examine the reason for thedisparate ratio, Kwon-Chung and colleagues (8)constructed a pair of congenic strains of C.neoformans that differed only in their matingtype. Survival studies with these congenic strainsdemonstrated that the α-mating type (B-4500)was significantly more virulent than the a-mating strain (B-4476) when injected intrave-nously into mice (8). Eighty percent of miceinfected with 106 B-4500 (α-mating type) were

Figure 2. Proposed means of infection by C.neoformans.


Synopses

dead within 36 days; whereas, it took 93 days for80% of mice infected with 106 B-4476 (a-matingtype) to die (8). Although the mating type locusof C. neoformans has been cloned and partiallycharacterized, the genes or gene products thatcontribute to the increased virulence ofα-mating type isolates are not known.

The mating type locus (MATα) cloned from anα-mating type C. neoformans isolate is 35 kb to 45kb long (9). Moore and Edman (9) demonstratedthat introducing a 2.1-kb fragment of the MATαlocus into an a-mating strain isolate stimulatedfilament formation on starvation medium.Basidia and basidiospores were not produced onthe filaments. In contrast, electroporation of anα-mating strain with the same MATα DNAfragment was ineffective in stimulating filamentformation under the same growth conditions (9).Sequence analysis of the 2.1-kb fragment fromthe MATα locus identified a 114-bp open readingframe that encodes a 38-amino acid propheromonepeptide (9). The nature and the effects of theactive pheromone peptide on mating and possiblyvirulence of C. neoformans are unknown;consequently, how the mating type relates tovirulence at the genetic level is also unknown.However, the observation of haploid fruiting byα-mating type isolates of C. neoformans onnitrogen- and water-depleted medium (5) mayhelp explain why the α-mating type isolates fromclinical specimens are more frequent than thea-mating type strains (6).

Growth In VivoTo cause infection in humans, a C.

neoformans isolate must grow at 37oC in anatmosphere of approximately 5% CO2 and at a pHof 7.3 to 7.4. To survive at 37oC, the organismmust have an intact gene that encodes the C.neoformans calcineurin A catalytic subunit (10).Calcineurin is a serine-threonine specific phos-phatase that is activated by Ca2+-calmodulin andis involved in stress responses in yeasts (10).Although calcineurin A mutant strains of C.neoformans can grow at 24oC, they cannot survivein vitro at 37oC, in 5% CO2, or at alkaline pH (10).Since these are similar to conditions in the host,one would predict that the calcineurin A mutantwould not survive in the human host. In supportof that prediction, Odom et al. have shown thatsuch mutants are not pathogenic for immunosup-pressed rabbits (10). Calcineurin A appears to be

a basic requirement for C. neoformans survival inthe host and consequently is a necessary factorfor the pathogenicity of the organism.

Virulence FactorsVirulence factors increase the degree of

pathogenicity of a microbe. C. neoformans has anumber of virulence factors; generally, thevirulence of an isolate cannot be attributed to anysingle factor, but rather it is attributed to manyworking in unison to cause progressive disease.As virulence factors go, those of C. neoformanswould be considered low-grade. We will discusseach virulence factor separately; however, theseverity of the host’s disease results from acombination of several virulence factors superim-posed on the host’s innate and immune resistancestatus. The virulence factors that will be discussedare capsule, cryptococcal products, melaninproduction, mannitol production, and potentialfactors such as superoxide dismutase, proteases,phospholipase B, and lysophospholipase. Thepolysaccharide capsule and the soluble extracellu-lar constituents of C. neoformans (referred to hereas cryptococcal products) are probably thedominant virulence factors.

CapsuleC. neoformans has a capsule composed

primarily of a high molecular weight polysaccha-ride that has a backbone of α-1,3-D-mannopyranoseunits with single residues of ß-D-xylopyranosyl andß-D-glucuronopyranosyl attached (11-15). This polysac-charide is referred to as glucuronoxylomannan (GXM)(11) and has four serotypes: A and D, produced by C.neoformans var. neoformans, and B and C,produced by C. neoformans var. gattii. The evidenceindicates that the capsule is a key virulence factorfor C. neoformans; acapsular mutants are typicallyavirulent, whereas encapsulated isolates havevarying degrees of virulence (16-19). Two capsulargenes, CAP59 and CAP64, have been described (20-22). Complementation of an acapsular, avirulentmutant with CAP59 resulted in a virulenttransformant with a capsular phenotype, anddeletion of CAP59 by homologous integrationcaused the encapsulated organism to becomeacapsular and avirulent (20). Similar observationshave been made with the CAP64 gene (21,22).Although their biochemical functions have not beendefined, it appears that the two genes are essentialfor capsule formation and virulence.


Synopses

Chemotactic Effects on LeukocytesSome properties of the C. neoformans capsule

enable the host to more effectively clear theorganism from tissues; however, others protectthe organism from host defenses. The capsules ofserotype A and D isolates are chemotactic forneutrophils (23). Moreover, complement is fixedby cryptococcal capsules by the alternativepathway (24), and this process produceschemotactic peptides such as C5a (23). Chemot-axis of leukocytes induced by either of thesemechanisms would be advantageous to the host.

Effects of Complement InteractionsComplement fixing by C. neoformans in

tissue would result in chemotactic factorproduction and attraction of leukocytes into theinfection site. Once in the infected tissue, theleukocytes would interact with and kill theorganism. As complement is fixed, C3b and C3biare deposited on the surface of the cryptococci(24). The capsule can mask the C3b and C3bideposits (24-26); however, if they are not completelymasked, the deposited complement componentsfacilitate binding of the cryptococci to CR3receptors on leukocytes (27). Such bindinginteractions are advantageous to the host; theyenhance the opportunity for the leukocytes to killthe cryptococci either extracellularly or afterphagocytosis. The organism can also be opsonizedby antibodies to GXM, but the capsule may blockthe Fc portion of the antibody and prevent it frombinding to Fc receptors on the phagocytic host cells(28). Some of these functions of the capsule favor thehost; however, if the capsule is very large, theorganism is protected (24-26). The cryptococci coulddeplete complement in the host, creating anenvironment that favors the cryptococci (29).

Effects on PhagocytosisAll considered, the capsule is more beneficial

to the organism than to the host. Encapsulated C.neoformans cells are not phagocytized or killed byneutrophils, monocytes, or macrophages to thesame degree as acapsular mutants (25,30-33).Encapsulated C. neoformans have a strongernegatively charged surface than acapsular cells orSaccharomyces cerevisiae cells (34). The highnegative charge could cause electrostatic repul-sion between the organism and the negativelycharged host effector cells and reduce intimatecell-cell interactions required for clearance ofthe cryptococci (34).

Altered Antigen PresentationThe inability of macrophages to ingest the

encapsulated organisms could also diminishantigen presentation to T cells and consequentlyreduce immune responses. This speculation hasbeen experimentally demonstrated by Collinsand Bancroft (35). Other studies with humanalveolar macrophages have confirmed that antigenpresentation is not as effective with encapsulatedcryptococci as with acapsular strains (33,36).Unlike acapsular cells, encapsulated isolatescannot stimulate proliferative responses in T cellsbecause of the reduced secretion of interleukin-1(IL-1) by the antigen-presenting cells stimulatedwith the encapsulated yeasts (36).

Effects on Cytokine ProductionIn addition to being antiphagocytic, more

resistant to killing, and less able to stimulate T-cell proliferation (25,30-33,35,36), highly encap-sulated isolates of C. neoformans opsonized witha heat-labile serum component, presumablycomplement, do not stimulate monocytes andmacrophages to produce proinflammatorycytokines such as TNFα, IL-1ß, and IL-6 aseffectively as similarly opsonized nonencapsulatedor weakly encapsulated isolates (37-41). Thesignal for induction of cytokine production can bea direct result of the attachment of the monocyteor macrophage to the acapsular cryptococci or canbe an outcome of the phagocytic process, inducedby acapsular cryptococci. Since the capsule blocksphagocytosis, any cytokines induced by thephagocytic process would not be induced by theencapsulated C. neoformans cells. If the cryptococ-cal cell wall materials must be exposed to inducecytokine production, the capsule would block thedirect induction of cytokine production. Regard-less of the mechanism of stimulation, the lack ofproduction of proinflammatory cytokines couldhave a bearing on protection. TNFα is necessaryfor the induction of the protective immuneresponse against C. neoformans (42). Conse-quently, the lack of or reduced production ofTNFα in infections with highly encapsulatedisolates of C. neoformans would prevent theinduction of protective immunity, resulting inprogressive disease. The roles of IL-1ß and IL-6 inprotecting against C. neoformans have not beendefined, but it is highly probable that the lack ofthese two cytokines could compromise theprotective responses of the host and givecryptococci the advantage.


Synopses

In contrast to the reduced TNFα, IL-1ß, andIL-6 produced by stimulating monocytes andmacrophages with highly encapsulated crypto-cocci, IL-10 produced by these host cells increasesafter interaction with encapsulated strains (43).Neutralization of IL-10 with anti-IL-10 incocultures of human peripheral blood mono-nuclear cells and encapsulated cryptococciincreased the amounts of TNFα and IL-1ßproduced, which indicates that the induction ofIL-10 production by stimulating macrophageswith encapsulated C. neoformans downregulatesthe production of the proinflammatory cytokinesTNFα and IL-1ß (43). One might predict that theinduction of high levels of IL-10 would alsopreferentially stimulate the induction of a T-helper 2 (Th2) response rather than a Th1response in the T cells (44). Since the Th1response is associated with protection againstC. neoformans (45), increased levels of IL-10would dampen induction of the protectiveimmune response.

Encapsulated cryptococcal cells do not affectthe different types of leukocytes in the samemanner. Although encapsulated isolates do notstimulate macrophages to produce proinflammatorycytokines, they do stimulate neutrophils toproduce proinflammatory cytokines and thechemotactic factor IL-8 more effectively thanweakly encapsulated or acapsular organisms(46). As with the stimulation of macrophages byacapsular cryptococci to produce proinflammatorycytokines, serum is required for the encapsulatedorganisms to induce neutrophils to producecytokines (46). In the case of cytokine productionby neutrophils in response to encapsulated C.neoformans, it appears that the opsonizationprocess releases a factor into the supernatantthat induces the neutrophils to produce thecytokines (46). How these opposing in vitroobservations with macrophages and neutro-phils relate to the in vivo system orpathogenicity is yet to be determined.

Cryptococcal ProductsIn disseminated cryptococcosis, measurable

levels of cryptococcal products are present in thebody fluids of the patients (47). Although GXM isthe major cryptococcal component in body fluids,it is highly probable that the organism also shedsgalactoxylomannan (GalXM) and mannoproteins(MP) in vivo. This speculation is based on indirectevidence from in vivo studies and on the fact that

GalXM and MP are in culture medium when theorganism is grown in vitro (15,48). Cryptococcalantigens in serum or spinal fluid are diagnosticfor cryptococcosis (47). Furthermore, if dissemi-nated cryptococcosis patients have high crypto-coccal antigen titers at the onset of therapy, theyare less likely to respond to therapy or more likelyto die before therapy is completed than patientswith low cryptococcal antigen titers (49). Thedirect relationship of cryptococcal antigen levelsin body fluids and the severity of disease suggeststhat the cryptococcal antigens in the host’scirculatory system or spinal fluid may haveadverse effects on host defenses.

Effects on Leukocyte MigrationIt has been recently demonstrated in the

mouse model that intravascular cryptococcalantigens inhibit the migration of leukocytes fromthe bloodstream into an inflammatory site (50).Intravascular cryptococcal antigens significantlyreduce leukocyte infiltration into a site of acuteinflammation induced by such stimuli ascryptococcal culture filtrate antigen, TNFa, orthe chemotactic peptide FMLP (formylmethionylleucyl phenylalanine) or into a delayed-typehypersensitivity reaction site induced by purifiedprotein derivative of Mycobacterium tuberculosisor by C. neoformans antigen(s) (50). Each of theidentified cryptococcal products, GXM, GalXM,and MP, when given intravenously to mice, caninhibit leukocyte migration (50). Considering thatGXM is the dominant cryptococcal component inthe host’s circulation, one might predict thatGXM is mainly responsible for the reducedleukocyte migration into inflammatory sites (50).These observations imply that pulmonarycryptococcosis patients, who have low to nocryptococcal antigen concentrations in theirserum, would have a normal influx of leukocytesinto pulmonary sites of infections. On the otherhand, in severe pulmonary infections or indisseminated cryptococcosis patients who havehigh levels of circulating cryptococcal antigens,minimal inflammation would be expected in theinfected tissues. In fact, for years investigatorshave commented on the minimal host tissueresponses observed in patients with disseminatedcryptococcosis (3). These recent data demonstratethat the circulating cryptococcal antigens areresponsible, at least in part, for the lack of hosttissue response.


Synopses

Cryptococcal antigen(s) can potentially pre-vent leukocytes from migrating into an inflam-matory site in two nonexclusive ways (51,52).First, GXM can stimulate neutrophils to shedL-selectin, a surface molecule necessary for thefirst step in neutrophil movement into tissues(51). Without L-selectin the neutrophils do notslow down and roll along the inflamed endothelialcells lining the blood vessels. With this first stepin extravasation blocked, the numbers ofneutrophils in infected tissues would be greatlyreduced. Second, GXM and GalXM can bind toCD18, the beta chain of the adhesion moleculeLFA-1, and prevent the binding of anti-CD18 toLFA-1 (52). Consequently, this binding of GXMand GalXM could prevent LFA-1 on the neutrophilsfrom binding to the LFA-1 ligand, ICAM-1 on theinflamed endothelial cell surface. If leukocytes areinhibited from entering tissues by either or both ofthese mechanisms, the organism is not effectivelyeliminated, and the disease is more severe.

Induction of Immunomodulatory CellsCryptococcal antigens injected into the

bloodstream of experimental animals can induceregulatory T cells, which dampen or ablate theanticryptococcal humoral as well as cell-mediated immune responses (53-71). Someclinical correlates support the concept that theantigenemia seen in disseminated cryptococcosisdownmodulates the immune responses (72,73).

A long-lasting, specific immunologic unre-sponsiveness has been reported in persons curedof cryptococcal meningitis (72-74). Although theymade antipneumococcal polysaccharide antibod-ies to the same degree as control volunteers aftervaccination with pneumococcal polysaccharide,cured patients could not make antibodies tocryptococcal polysaccharides after vaccinationwith cryptococcal antigens (72,73). Henderson etal. (72,73) speculated that the intense, prolongedantigenemia associated with the cryptococcosismay account for the observed unresponsiveness.It is not unusual for patients with disseminatedcryptococcosis to have depressed cell-mediatedimmune (CMI) responses to cryptococcal anti-gens (75,76). However, sufficient information isnot available to determine whether the patientshad a generalized defect in CMI function or had aspecifically depressed CMI response because ofthe cryptococcal antigenemia.

Experimental animal models demonstrateconvincingly that cryptococcal antigens given

intravenously can induce immunoregulatory Tcells that downmodulate the anticryptococcalCMI response (56-71). Serum from C. neoformans-infected mice with a cryptococcal antigen titer of10-4 when transferred intravenously to naivemice induces regulatory T cells that specificallydepress the anticryptococcal CMI response (67).Similar immunoregulatory T cells are induced bysimulating the antigenemia with intravenousinjection of cryptococcal culture filtrate antigen(56,57,60-63,66-69). The immunoregulatory CD4+

T cells, which appear in the lymph nodes of themice 7 days after antigen injection, diminish theinduction of T cells responsible for theanticryptococcal delayed-type hypersensitivityresponse and reduce the ability of the mice toclear cryptococci from tissues (56,57,60,68). Asecond immunoregulatory T cell, a CD8+ cell, hasbeen found in spleens of mice (57). The CD4+

immunoregulatory cells do not alter the numbersor types of leukocytes that infiltrate ananticryptococcal delayed-type hypersensitivityreaction; however, they do have a downregulatingeffect on IL-2 and IFNγ production at the site (77).In contrast, the CD8+ immunoregulatory cellsappear to affect the numbers of neutrophils thatinfiltrate the delayed-type hypersensitivityreaction site (Murphy, unpub. data). The detailsof the characteristics and functions of theseimmunoregulatory T cells have been reviewed(45). The available data strongly support thenotion that cryptococcal products in thecirculation induce an array of immunoregulatoryT cells, which depress the anticryptococcalimmune responses and protection.

More consideration should be given to theimpact of high levels of cryptococcal antigen inthe cerebrospinal fluid (CSF) on progression ofdisease. As suggested by Denning et al. (78), highcryptococcal antigen concentrations could changethe osmolality of the CSF, thereby affecting itsoutflow and adsorption and increasing intracra-nial pressure. The increased pressure may causeheadaches, visual loss, and early death (78). It isalso possible that release of mannitol by C.neoformans contributes to increased intracranialpressure in cryptococcal meningitis patients (78).

Melanin SynthesisOne characteristic that differentiates patho-

genic isolates of C. neoformans from nonpatho-genic isolates and other Cryptococcus species isthe organism’s ability to form a brown to black


Synopses

pigment on a medium (such as birdseed or caffeicacid agar) that contains diphenolic compounds(1). This pigment, first described by Staib (79), isa melaninlike compound produced by C. neoformansisolates with phenoloxidase activity (80). Theimportance of melanin production in C. neoformansvirulence was first demonstrated in the early1980s. Rhodes, Polacheck, and Kwon-Chung (81)reported that naturally occurring C. neoformansmutants lacking melanin (Mel-) were lessvirulent in mice than melanin-producing strains.Others (82-84), using chemically induced mu-tants or isogenic pairs of C. neoformans, haveconfirmed and extended this observation.

Biochemical analyses suggest that in C.neoformans melanogenesis is accomplished byconversion of dihydroxyphenols such as 3,4-dihydroxyphenylalanine (DOPA) to dopaquinone(Figure 3). This conversion is catalyzed by aphenoloxidase and is the rate-limiting step,presumably because subsequent steps in thepathway, such as dopaquinone rearranging todopachrome and ultimately autoxidation tomelanin, are spontaneous (85). C. neoformanslacks a tyrosinase enzyme required for endog-enous production of dihydroxyphenols (83); thusto produce melanin, a C. neoformans isolate mustbe able to acquire diphenolic compounds from itsgrowth environment, and it must have theenzyme phenoloxidase to catalyze conversion ofthese compounds into the subsequent melaninintermediates. The brain is a tissue rich incatecholamines such as DOPA and is a favoritetarget for infection by C. neoformans. However,the organism cannot use catecholamines as a solecarbon source, which suggests that the brain isnot a preferred nutritional niche for growth of

C. neoformans (86); rather, it may serve as asurvival niche as described below.

Polacheck et al. (86) reported that melanin-producing isolates of C. neoformans wereresistant to damage by an in vitro epinephrineoxidative system, whereas mutants lackingphenoloxidase activity were highly susceptible,as evidenced by decreased survival. Jacobsonand Tinnell (87) found that melanin protectedC. neoformans from damage by hypochlorite andpermanganate but not by hydrogen peroxide. Inthese experiments, hypochlorite was 100 timesmore fungicidal than hydrogen peroxide, andC. neoformans could produce sufficient levels ofmelanin to effectively protect the organism fromoxidative compounds produced by macrophages(87). Wang and Casadevall (88) extended thesefindings by examining survival of C. neoformansin the presence of nitric oxide and reactive nitrogenintermediates as well as in the epinephrineoxidative system described by Polacheck et al. (86).Wang and Casadevall cultured C. neoformans cellswith L-DOPA to melanize the yeast cells (88).Melanized cryptococci survived damage by bothnitrogen- and oxygen-derived oxidants signifi-cantly better than nonmelanized organisms ofthe same strain. These results support thehypothesis that C. neoformans uses catechola-mines in the brain to make melanin, therebyprotecting the organism from oxidative damageby scavenging free radicals (86).

Recently, cryptococcal diphenoloxidase hasbeen purified, and the gene CNLAC1 encodingthis enzyme was cloned (89). The glycosylatedprotein has a molecular weight of 75 kDa andcontains copper. The substrate specificity of theenzyme indicates that it is a laccase. CNLAC1contains 14 introns, a 624 amino acid openreading frame; transcriptional activity is dere-pressed in the absence of glucose (89), whichconfirms an earlier report of low glucoserequirements for melanin formation (90). Disrup-tion of CNLAC1 resulted in loss of virulence ofC. neoformans, whereas complementation withCNLAC1 increased virulence of Mel- mutants inmice (84); these results suggest that the laccase(phenoloxidase) encoded by CNLAC1 is a potentialvirulence factor of C. neoformans. Furthermore,transcripts of the CNLAC1 gene have beendetected by reverse transcription-polymerasechain reaction (RT-PCR) in C. neoformans yeastcells isolated from CSF in a rabbit model ofcryptococcal meningitis (84).

Figure 3. Proposed pathway for melanin synthesis byC. neoformans (85).


Synopses

Besides acting as an antioxidant, melaninproduction may help C. neoformans survive inthe host in other ways. Melanized yeast cells areless susceptible to amphotericin B thannonmelanized yeast cells, and this may contrib-ute to the inability to effectively treat infectionsin immunocompromised hosts (91). Furthermore,phagocytosis of melanized C. neoformans by amacrophage cell line in the presence of an anti-GXM antibody was decreased, which suggeststhat melanin deposition in the cell wall mayinhibit opsonization by specific antibodies (92).Recently, Huffnagle and coinvestigators (93)reported that melanized heat-killed C. neoformansstrain 145 yeast cells stimulated less TNFαproduction by alveolar macrophages and lessantigen-specific T-cell proliferation thannonmelanized heat-killed 145 strain cells. Theauthors suggest that melanin “cloaks” C. neoformansfrom recognition by host effector cells andinhibits induction of a protective T-cell-mediatedimmune response (93); however, it is possiblethat scavenging of host-produced oxidants andinhibition of phagocytosis by melanin maycontribute to the decreased TNFα production andlymphoproliferation observed.

Although melanin production is important inthe virulence of C. neoformans, there is very littleevidence demonstrating the presence of melaninin vivo. One report indicates that phenoloxidaseactivity in C. neoformans is greatly diminished at37oC compared with activity at 25oC, whichsuggests that melanin production may be limitedin vivo (94). Detection of melanin in vivo ishampered by the lack of specific antibodies orstains. A modified Fontana-Masson stain hasbeen used to detect a brown pigment in the cellwall of C. neoformans cells in histologic brainsections; however, the stain is not specificbecause Cryptococcus laurentii, which is Mel-,also stains with Fontana-Masson (95).

In summary, accumulating evidence indi-cates that melanin production is an effectivesurvival (virulence) factor of C. neoformans, andmelanin may serve multiple roles in protectingthis medically important fungus from hostdefenses. Two caveats should be noted in labelingmelanin as a major virulence factor for C.neoformans. The C. neoformans isolate must beable to internalize the melanin precursors, andthe phenoloxidase enzyme must make asufficient amount of melanin precursor at 37oC.Consequently, much more work is needed to

delineate the role of melanin production invirulence of C. neoformans, especially in light ofthe apparent temperature sensitivity of thecryptococcal phenoloxidase enzyme.

Mannitol ProductionAccumulating evidence suggests that produc-

tion of the hexitol D-mannitol may contribute tosurvival of C. neoformans in the host. Wong etal. (96) reported that of the 12 human isolatesof C. neoformans examined, all produced andsecreted D-mannitol into growth medium.Further, by using a rabbit cryptococcal meningi-tis model, they showed that C. neoformans canproduce D-mannitol in vivo. CSF from rabbitstreated with cortisone and infected intracister-nally with C. neoformans contained more D-mannitol than CSF from controls, cortisone-treated uninfected rabbits, or rabbits infectedwith C. neoformans but not treated with cortisone(96). In the latter group, cryptococcal infectionwas limited. The levels of D-mannitol in infectedrabbit CSF correlated well with both the numbersof culturable C. neoformans and the cryptococcalantigen titer of the CSF, which suggests thatlevels of D-mannitol in CSF may be prognostic(96); however, it is not known whether differentisolates of C. neoformans vary in mannitolproduction. These authors suggested two meansby which mannitol production may contribute toC. neoformans pathogenesis. First, high concen-trations of D-mannitol in the CNS may contributeto brain edema. Second, mannitol is a potentscavenger of hydroxyl radicals, and cryptococcal-produced D-mannitol may help protect theorganism from oxidative damage (96).

To investigate the role of mannitol incryptococcosis, Chaturvedi and co-workers (97)isolated a low mannitol producing mutant afterUV irradiation of C. neoformans strain H99. Themutant, mannitol low producer (MLP), wassimilar to the parent strain H99 in many growthand morphology characteristics and in knownvirulence factors such as capsule production andphenoloxidase activity (97). However, the mutantMLP strain was more susceptible to growthinhibition and killing by heat and high salt thanthe parent strain. In addition, the mutant MLPstrain was significantly less virulent than theparent strain (MLP LD50 = 3.7 X 106 CFU, H99LD50 = 6.9 X 102) (97). Further comparisons of H99and MLP strains showed that polymorphonuclearleukocytes (PMNL) killed MLP significantly


Synopses

better than H99 at several effector-to-targetratios (98). Addition of superoxide dismutase,mannitol, or dimethyl sulfoxide inhibited PMNLkilling of both strains, but addition of catalase didnot alter killing, which suggests that reactiveoxygen intermediates such as the hydroxylradical and hypochlorous acid are potent effectormolecules against C. neoformans and mannitolmay protect against oxidative killing byscavenging such compounds (98).

The results above indicate that mannitolproduction by C. neoformans correlates withincreased resistance to heat stress, osmoticstress, and damage by reactive oxygen intermedi-ates, as well as increased pathogenicity of thisfungal agent. Additional studies are required todetermine the role of mannitol production invirulence of C. neoformans. Isogenic strainslacking the enzyme required for mannitolproduction, namely mannitol dehydrogenase, arenot available because the gene for this enzymehas not been found. However, a gene (Mtl) hasbeen isolated from C. neoformans that can induceexpression of the S. cerevisiae mannitol dehydro-genase gene (99). The 346 amino acid productencoded by the Mtl gene is thought to be involvedin regulating mannitol production in C. neoformansand may be beneficial in future studies ofmannitol production (99).

Other Potential Virulence Factors

Superoxide DismutaseJacobson and coinvestigators (100) exam-

ined superoxide dismutase (SOD) productionby C. neoformans to determine whether SODlevels increased at 37oC to compensate for possibledecreases in melanin production at this tempera-ture. These investigators observed an increase inSOD levels at 37oC, which suggests that SOD mayparticipate in free radical scavenging at this highertemperature in vivo (100); however, there is noevidence that SOD production serves as a virulencefactor for C. neoformans.

ProteasesMuller and Sethi (101) reported that C.

neoformans grown in culture produced a proteasethat could digest human plasma proteins, andBrueske (102) reported that C. neoformansculture supernatants contained a protease capableof digesting casein. However, neither of thesestudies determined whether the proteases were

manufactured in vivo. Limited evidence indicat-ing in vivo production of proteases has beenpresented by Salkowski and Balish (103). Theseinvestigators observed skin lesions on T-cell–deficient mice after intravenous infections withC. neoformans strain SLHA (103). Microscopicexamination of the lesions suggested that thecryptococcal yeast cells were degrading collagenbundles in the dermis (103). Supernatants ofSLHA strain cultures liquefied gelatin in vitroindicating that this cryptococcal strain secreted acollagenaselike protein (103). Thus, C. neoformansproteases possibly serve as virulence mecha-nisms by initiating invasion of host tissues;however, more studies with isogenic strains ofC. neoformans are required before proteasescan be listed as virulence factors.

PhospholipasesRecently, Chen and co-workers (104,105)

reported phospholipase, lysophospholipase,and lysophospholipase-transacylase activity ofC. neoformans grown on egg yolk agar. Of 50strains tested, 49 had phospholipase activity(104), due to phospholipase B secreted intocultures (105). Analysis of four strains withvarying levels of phospholipase activity indicateda correlation between phospholipase activity andvirulence in BALB/c mice (104). The authorssuggest that extracellular phospholipase activityproduced by C. neoformans may disrupt mamma-lian cell membranes and allow the yeast cells topenetrate into host tissues (104,105); however,further investigations are necessary to establishthe role, if any, of these types of products in thevirulence of C. neoformans.

Regulation of VirulenceRegulation of virulence factors such as

capsule production and melanin formation is notwell understood. However, the gene GPA1, whichencodes a G-protein α-subunit, is involved in theregulation of these virulence factors as well as inC. neoformans mating (106). Disruption of GPA1resulted in defects in mating in response to nitrogenstarvation, capsule production in response to ironlimitation, and melanin synthesis in response toglucose starvation (106). Furthermore, gpa1mutants were much less virulent in a rabbitmodel of cryptococcal meningitis (106). Reconsti-tution of the gpa1 mutant with wild-type GPA1restored mating, capsule production, melaninsynthesis, and virulence. Addition of cyclic AMP


Synopses

also restored these phenotypes, which suggeststhat C. neoformans GPA1 regulates these factorsby sensing the nutritional signals of theenvironment and regulating cyclic AMP metabo-lism in the organism (106). These findings alongwith results of other molecular studies areintriguing and represent the initial steps indefining at the molecular level the factorscontrolling virulence in C. neoformans.

ConclusionsVirulent isolates of C. neoformans must be

able to produce small particles that can get intothe alveolar spaces, must be able to grow at 37oCat a pH of 7.3 to 7.4 in an atmosphere ofapproximately 5% CO2, must have an intactcalcineurin pathway, and (possibly) must be anα-mating type. The ability to produce a largecapsule and shed great amounts of capsularmaterial into the body fluids makes theorganism highly virulent. Other factors, suchas melanin, mannitol, superoxide dismutase,protease, and phospholipase production, mayenhance the pathogenicity of C. neoformans.The effectiveness of many of these cryptococcalvirulence factors depends on the status of thehost’s defensive mechanisms. Although wehave learned much about the pathogenicity andvirulence of C. neoformans, many gaps stillremain in our knowledge.

AcknowledgmentsWe thank Drs. J.A. Alspaugh, J.R. Perfect, and J.

Heitman for making their manuscript available beforepublication.

Dr. Buchanan is a research assistant professor inthe Department of Microbiology and Immunology,University of Oklahoma Health Sciences Center,Oklahoma City, Oklahoma, USA. Dr. Murphy is aGeorge Lynn Cross Research Professor of Microbiologyand Immunology in the same department.

References 1. Kwon-Chung KJ. Cryptococcosis. In: Kwon-Chung KJ,

Bennett JE, editors. Medical mycology. Philadelphia:Lea & Febiger, 1992:397-446.

2. Mitchell TG, Perfect JR. Cryptococcosis in the era ofAIDS—100 years after the discovery of Cryptococcusneoformans. Clin Microbiol Rev 1995;8:515-48.

3. Rippon JW. Cryptococcosis. In: Rippon JW, editor.Medical mycology. Philadelphia: W.B. SaundersCompany; 1988, p. 582-609.

4. Duperval B, Hermans PE, Brewer NS, Roberts GD.Cryptococcosis, with emphasis on the significance ofisolation of Cryptococcus neoformans from therespiratory tract. Chest 1977;72:13-9.

5. Wickes BL, Mayorga ME, Edman U, Edman JC.Dimorphism and haploid fruiting in Cryptococcusneoformans: association with the a-mating type. ProcNatl Acad Sci U S A 1996;93:7327-31.

6. Kwon-Chung KJ, Bennett JE. Distribution of andmating types of Cryptococcus neoformans amongnatural and clinical isolates. Am J Epidemiol1978;108:337-40.

7. Kwon-Chung KJ, Hill WB. Sexuality and pathogenicityof Filobasidiella neoformans (Cryptococcus neoformans).In: Vanbreuseghem R, DeVroey C, Vanbreuseghem R,editors. Sexuality and pathogenicity of fungi. Paris:Masson; 1981. p. 243-50.

8. Kwon-Chung KJ, Edman JC, Wickes BL. Geneticassociation of mating types and virulence in Cryptococcusneoformans. Infect Immun 1992;60:602-5.

9. Moore TDE, Edman JC. The α-mating type locus ofCryptococcus neoformans contains a pheromone gene.Mol Cell Biol 1993;13:1962-70.

10. Odom A, Muir S, Lim E, Toffaletti DL, Perfect J,Heitman J. Calcineurin is required for virulence ofCryptococcus neoformans. EMBO J 1997;16:2576-89.

11. Cherniak R, Reiss E, Slodki ME, Plattner RD, BlumerSO. Structure and antigenic activity of the capsularpolysaccharide of Cryptococcus neoformans. MolImmunol 1980;17:1025-32.

12. Bhattacharjee AK, Kwon-Chung KJ, Glaudemans CP.Capsular polysaccharides from a parent strain and from apossible mutant strain of Cryptococcus neoformansserotype A. Carbohydr Res 1981;95:237-48.

13. Bhattacharjee AK, Bennett JE, Glaudemans CP.Capsular polysaccharides of Cryptococcus neoformans.Reviews of Infectious Diseases 1984;6:619-24.

14. Cherniak R. Soluble polysaccharides of Cryptococcusneoformans. Cur Top Med Mycol 1988;2:40-54.

15. Cherniak R, Sundstrom JB. Polysaccharide antigens ofthe capsule of Cryptococcus neoformans. Infect Immun1994;62:1507-12.

16. Bulmer GS, Sans MD, Gunn CM. Cryptococcusneoformans. I. Nonencapsulated mutants. J Bacteriol1967;94:1475-9.

17. Fromtling RA, Shadomy HJ, Jacobson ES. Decreasedvirulence in stable, acapsular mutants of Cryptococcusneoformans. Mycopathologia 1982;79:23-9.

18. Kozel TR, Cazin J Jr. Non-encapsulated variant ofCryptococcus neoformans. I. Virulence studies andcharacterization of soluble polysaccharide. InfectImmun 1971;3:287-94.

19. Kwon-Chung KJ, Rhodes JC. Encapsulation and melaninformation as indicators of virulence in Cryptococcusneoformans. Infect Immun 1986;51:218-23.

20. Chang YC, Kwon-Chung KJ. Complementation of acapsule-deficient mutation of Cryptococcus neoformansrestores its virulence. Mol Cell Biol 1994;14:4912-9.

21. Chang YC, Penoyer LA, Kwon-Chung KJ. The secondcapsule gene of Cryptococcus neoformans, CAP64, isessential for virulence. Infect Immun 1996;64:1977-83.


Synopses

22. Chang YC, Cherniak R, Kozel TR, Granger LC, MorrisLC, Weinhold LC, Kwon-Chung KJ. Structure andbiological activities of acapsular Cryptococcusneoformans 602 complemented with the CAP64 gene.Infect Immun 1997;65:1584-92.

23. Dong ZM, Murphy JW. Mobility of human neutrophilsin response to Cryptococcus neoformans cells, culturefiltrate antigen, and individual components of theantigen. Infect Immun 1993;61:5067-77.

24. Kozel TR, Pfrommer GS. Activation of the complementsystem by Cryptococcus neoformans leads to binding ofiC3b to the yeast. Infect Immun 1986;52:1-5.

25. Kozel TR, Mastroianni RP. Inhibition of phagocytosisby cryptococcal polysaccharide: dissociation of theattachment and ingestion phases of phagocytosis.Infect Immun 1976;14:62-7.

26. Kozel TR, Highison B, Stratton CJ. Localization onencapsulated Cryptococcus neoformans of serumcomponents opsonic for phagocytosis by macrophagesand neutrophils. Infect Immun 1984;43:574-9.

27. Griffin FMJ. Roles of macrophage Fc and C3b receptorsin phagocytosis of immunologically coated Cryptococcusneoformans. Proc Natl Acad Sci U S A 1981;78:3853-7.

28. Kozel TR, Pfrommer GS, Guerlain AS, Highison BA,Highison GJ. Role of the capsule in phagocytosis ofCryptococcus neoformans. Reviews of InfectiousDiseases 1988;10:S436-9.

29. Macher AM, Bennett JE, Gadek JE, Frank MM.Complement depletion in cryptococcal sepsis. JImmunol 1978;120:1686-90.

30. Bulmer GS, Sans MD. Cryptococcus neoformans II.Phagocytosis by human leukocytes. J Bacteriol1967;94:1480-3.

31. Bulmer GS, Sans MD. Cryptococcus neoformans. III.Inhibition of phagocytosis. J Bacteriol 1968;95:5-8.

32. Kozel TR, Gotschlich EC. The capsule of Cryptococcusneoformans passively inhibits phagocytosis of the yeastby macrophages. J Immunol 1982;129:1675-80.

33. Vecchiarelli A, Pietrella D, Dottorini M, Monari C, RetiniC, Todisco T, Bistoni F. Encapsulation of Cryptococcusneoformans regulates fungicidal activity and the antigenpresentation process in human alveolar macrophages.Clin Exp Immunol 1994;98:217-23.

34. Nosanchuk JD, Casadevall A. Cellular charge ofCryptococcus neoformans: Contributions from thecapsular polysaccharide, melanin, and monoclonalantibody binding. Infect Immun 1997;65:1836-41.

35. Collins HL, Bancroft GJ. Encapsulation of Cryptococcusneoformans impairs antigen-specific T-cell responses.Infect Immun 1991;59:3883-8.

36. Vecchiarelli A, Dottorini M, Pietrella D, Vecchiarelli A,Dottorini M, Pietrella D, et al. Role of human alveolarmacrophages as antigen-presenting cells inCryptococcus neoformans infection. Am J Respir CellMol Biol 1994;11:130-7.

37. Levitz SM, Tabuni A, Kornfeld H, Reardon CC,Golenbock DT. Production of tumor necrosis factoralpha in human leukocytes stimulated by Cryptococcusneoformans. Infect Immun 1994;62:1975-81.

38. Naslund PK, Miller WC, Granger DL. Cryptococcusneoformans fails to induce nitric oxide synthase inprimed murine macrophage-like cells. Infect Immun1995;63:1298-1304.

39. Vecchiarelli A, Retini C, Pietrella D, Monari C, TasciniT, Beccari T, Kozel TR. Downregulation bycryptococcal polysaccharide of tumor necrosis factoralpha and interleukin-1 beta secretion from humanmonocytes. Infect Immun 1995;63:2919-23.

40. Delfino D, Cianci L, Migliardo M, Mancuso G,Cusumano V, Corradini C, Teti G. Tumor necrosisfactor-inducing activities of Cryptococcus neoformanscomponents. Infect Immun 1996;64:5199-204.

41. Delfino D, Cianci L, Lupis E, Celese A, Petrelli ML,Curro F, et al. Interleukin-6 production by humanmonocytes stimulated with Cryptococcus neoformanscomponents. Infect Immun 1997;65:2454-6.

42. Huffnagle GB, Toews GB, Burdick MD, Boyd MB,McAllister KS, McDonald RA, et al. Afferent phaseproduction of TNF-α required for the development ofprotective T cell immunity to Cryptococcus neoformans.J Immunol 1996;157:4529-36.

43. Vecchiarelli A, Retini C, Monari C, Tascini C, Bistoni F,Kozel TR. Purified capsular polysaccharide of Cryptococcusneoformans induces interleukin-10 secretion by humanmonocytes. Infect Immun 1996;64:2846-9.

44. Moore KW, O’Garra A, de Waal Malefyt R, Vieira P,Mosmann TR. Interleukin-10. Annu Rev Immunol1993;11:165-90.

45. Murphy JW. Cell-mediated immunity. In: Howard DH,Miller JD, editors. The Mycota VI. Human and animalrelationships. Berlin: Springer-Verlag; 1996. p. 67-97.

46. Retini C, Vecchiarelli A, Monari C, Tascini C, Bistoni F,Kozel TR. Capsular polysaccharide of Cryptococcusneoformans induces proinflammatory cytokine release byhuman neutrophils. Infect Immun 1996;64:2897-903.

47. Gordon MA, Vedder DK. Serologic tests in diagnosisand prognosis of cryptococcosis. JAMA 1966;197:961-7.

48. Reiss E, Cherniak R, Eby E, Kaufman L. Enzymeimmunoassay detection of IgM to galactoxylomannanof Cryptococcus neoformans. Diagnostic Immunology1984;2:109-15.

49. Diamond RD, Bennett JE. Prognostic factors incryptococcal meningitis. A study in 111 cases. AnnIntern Med 1974;80:176-81.

50. Dong ZM, Murphy JW. Intravascular cryptococcalculture filtrate (CneF) and its major componentglucuronoxylomannan (GXM) are potent inhibitors ofleukocyte accumulation. Infect Immun 1995;63:770-8.

51. Dong ZM, Murphy JW. Cryptococcal polysaccharidesinduce L-selectin shedding and tumor necrosis factorreceptor loss from the surface of human neutrophils. JClin Invest 1996;97:689-98.

52. Dong ZM, Murphy JW. Cryptococcal polysaccharidesbind to CD18 on human neutrophils. Infect Immun1997;65:557-63.

53. Murphy JW, Cozad GC. Immunologicalunresponsiveness induced by cryptococcal capsularpolysaccharide assayed by the hemolytic plaquetechnique. Infect Immun 1972;5:896-901.

54. Kozel TR, Gulley WF, Cazin J Jr. Immune response toCryptococcus neoformans soluble polysaccharide:immunological unresponsiveness. Infect Immun1977;18:701-7.

55. Breen JF, Lee IC, Vogel FR, Friedman H. Cryptococcalcapsular polysaccharide-induced modulation of murineimmune responses. Infect Immun 1982;36:47-51.


Synopses

56. Murphy JW, Moorhead JW. Regulation of cell-mediated immunity in cryptococcosis. I. Induction ofspecific afferent T suppressor cells by cryptococcalantigen. J Immunol 1982;128:276-83.

57. Murphy JW, Mosley RL, Moorhead JW. Regulation ofcell-mediated immunity in cryptococcosis. II.Characterization of first-order T suppressor cells (Ts1)and induction of second-order suppressor cell. JImmunol 1983;130:2876-81.

58. Morgan MA, Blackstock R, Bulmer GS, Hall NK.Modification of macrophage phagocytosis in murinecryptococcosis. Infect Immun 1983;40:493-500.

59. Blackstock R, Hall NK. Nonspecific immune suppressionby Cryptococcus neoformans infection. Mycopathologia1984;86:35-43.

60. Murphy JW. Effects of first-order Cryptococcus-specific T-suppressor cells on induction of cellsresponsible for delayed-type hypersensitivity. InfectImmun 1985;48:439-45.

61. Murphy JW, Mosley RL. Regulation of cell-mediatedimmunity in cryptococcosis. III. Characterization ofsecond-order T suppressor cells (Ts2). J Immunol1985;134:577-84.

62. Mosley RL, Murphy JW, Cox RA. Immunoadsorption ofCryptococcus-specific suppressor T-cell factors. InfectImmun 1986;51:844-50.

63. Khakpour FR, Murphy JW. Characterization of a third-order suppressor T cell (Ts3) induced by cryptococcalantigen(s). Infect Immun 1987;55:1657-62.

64. Blackstock R, McCormack JM, Hall NK. Induction of amacrophage-suppressive lymphokine by solublecryptococcal antigens and its association with models ofimmunological tolerance. Infect Immun 1987;55:233-9.

65. Blackstock R, Hernandez NC. Inhibition of phagocytosisin cryptococcosis: phenotypic analysis of the suppressorcell. Cell Immunol 1988;114:174-87.

66. Murphy JW, Cox RA. Induction of antigen-specificsuppression by circulating Cryptococcus neoformansantigen. Clin Exp Immunol 1988;73:174-80.

67. Murphy JW. Influence of cryptococcal antigens on cell-mediated immunity (CMI). Reviews of InfectiousDiseases 1988;10:5432-5.

68. Murphy JW. Clearance of Cryptococcus neoformansfrom immunologically suppressed mice. Infect Immun1989;57:1946-52.

69. Murphy JW. Immunoregulation in cryptococcosis. In:Kurstak E, editor. Immunology of fungal diseases. NewYork: Marcel Dekker; 1989. p. 319-45.

70. Blackstock R, Hall NK, Hernandez NC.Characterization of a suppressor factor that regulatesmacrophage phagocytosis in murine cryptococcosis.Infect Immun 1989;57:1773-9.

71. Blackstock R, Zembala M, Asherson GL. Functionalequivalence of cryptococcal and haptene-specific Tsuppressor factor (TsF). I. Picryl and oxazolone-specific TsF, which inhibit transfer of contactsensitivity also inhibit phagocytosis by a subset ofmacrophages. Cell Immunol 1991;136:435-47.

72. Henderson DK, Bennett JE, Huber MA. Long-lasting,specific immunologic unresponsiveness associatedwith cryptococcal meningitis. J Clin Invest1982;69:1185-90.

73. Henderson DK, Kan VL, Bennett JE. Tolerance tocryptococcal polysaccharide in cured cryptococcosispatients: failure of antibody secretion in vitro. Clin ExpImmunol 1986;65:639-46.

74. Salvin SB, Smith RF. An antigen for detection ofhypersensitivity to Cryptococcus neoformans. Proc SocExp Biol Med 1961;108:498-501.

75. Diamond RD, Bennett JE. Disseminated cryptococcosisin man: decreased lymphocyte transformation inresponse to Cryptococcus neoformans. J Infect Dis1973;127:694-7.

76. Graybill JR, Alford RH. Cell-mediated immunity incryptococcosis. Cell Immunol 1974;14:12-21.

77. Buchanan KL, Murphy JW. Regulation of cytokineproduction during the expression phase of theanticryptococcal delayed-type hypersensitivityresponse. Infect Immun 1994;62:2930-9.

78. Denning DW, Armstrong RW, Lewis BH, Stevens DA.Elevated cerebrospinal fluid pressures in patients withcryptococcal meningitis and acquired immunodeficiencysyndrome. Am J Med 1991;91:267-72.

79. Staib F. Cryptococcus neoformans und Guizotiaabyssinica. Zeitschrift für Hygiene 1962;148:466-75.

80. Shaw CE, Kapica L. Production of diagnostic pigmentby phenoloxidase activity of Cryptococcus neoformans.Applied Microbiology 1972;24:824-30.

81. Rhodes JC, Polacheck I, Kwon-Chung KJ. Phenoloxidaseactivity and virulence in isogenic strains of Cryptococcusneoformans. Infect Immun 1982;36:1175-84.

82. Kwon-Chung KJ, Polacheck I, Popkin TJ. Melanin-lacking mutants of Cryptococcus neoformans and theirvirulence for mice. J Bacteriol 1982;150:1414-21.

83. Torres-Guerrero H, Edman JC. Melanin-deficientmutants of Cryptococcus neoformans. J Med Vet Mycol1994;32:303-13.

84. Salas SD, Bennett JE, Kwon-Chung KJ, Perfect JR,Williamson PR. Effect of the laccase gene, CNLAC1, onvirulence of Cryptococcus neoformans. J Exp Med1996;184:377-86.

85. Polacheck I, Kwon-Chung KJ. Melanogenesis inCryptococcus neoformans. Journal of GeneralMicrobiology 1988;134:1037-41.

86. Polacheck I, Platt Y, Aronovitch J. Catecholamines andvirulence of Cryptococcus neoformans. Infect Immun1990;58:2919-22.

87. Jacobson ES, Tinnell SB. Antioxidant function offungal melanin. J Bacteriol 1993;175:7102-4.

88. Wang Y, Casadevall A. Susceptibility of melanized andnonmelanized Cryptococcus neoformans to nitrogen-and oxygen-derived oxidants. Infect Immun1994;62:3004-7.

89. Williamson PR. Biochemical and molecularcharacterization of the diphenol oxidase of Cryptococcusneoformans: identification as a laccase. J Bacteriol1994;176:656-64.

90. Polacheck I, Hearing VJ, Kwon-Chung KJ. Biochemicalstudies of phenoloxidase and utilization ofcatecholamines in Cryptococcus neoformans. J Bacteriol1982;150:1212-20.

91. Wang Y, Casadevall A. Growth of Cryptococcusneoformans in presence of L-dopa decreases itssusceptibility to amphotericin B. Antimicrob AgentsChemother 1994;38:2648-50.


Synopses

92. Wang Y, Aisen P, Casadevall A. Cryptococcusneoformans melanin and virulence: mechanism ofaction. Infect Immun 1995;63:3131-6.

93. Huffnagle GB, Chen GH, Curtis JL, McDonald RA,Strieter RM, Toews GB. Down-regulation of theafferent phase of T cell-mediated pulmonaryinflammation and immunity by a high melanin-producing strain of Cryptococcus neoformans. JImmunol 1995;155:3507-16.

94. Jacobson ES, Emery HS. Temperature regulation ofthe cryptococcal phenoloxidase. J Med Vet Mycol1991;29:121-4.

95. Kwon-Chung KJ, Hill WB, Bennett JE. New, specialstain for histopathological diagnosis of cryptococcosis.J Clin Microbiol 1981;13:383-7.

96. Wong B, Perfect JR, Beggs S, Wright KA. Production ofthe hexitol D-mannitol by Cryptococcus neoformans invitro and in rabbits with experimental meningitis.Infect Immun 1990;58:1664-70.

97. Chaturvedi V, Flynn T, Niehaus WG, Wong B. Stresstolerance and pathogenic potential of a mannitolmutant of Cryptococcus neoformans. Microbiology1996;142:937-43.

98. Chaturvedi V, Wong B, Newman SL. Oxidative killingof Cryptococcus neoformans by human neutrophils.Evidence that fungal mannitol protects by scavengingreactive oxygen intermediates. J Immunol1996;156:3836-40.

99. Perfect JR, Rude TH, Wong B, Flynn T, Chaturvedi V,Niehaus W. Identification of a Cryptococcus neoformansgene that directs expression of the crypticSaccharomyces cerevisiae mannitol dehydrogenasegene. J Bacteriol 1996;178:5257-62.

100. Jacobson ES, Jenkins ND, Todd JM. Relationshipbetween superoxide dismutase and melanin in apathogenic fungus. Infect Immun 1994;62:4085-6.

101. Muller HE, Sethi KK. Proteolytic activity of Cryptococcusneoformans against human plasma proteins. MedMicrobiol Immunol (Berl) 1972;158:129-34.

102. Brueske CH. Proteolytic activity of a clinical isolate ofCryptococcus neoformans. J Clin Microbiol 1986;23:631-3.

103. Salkowski CA, Balish E. Cutaneous cryptococcosis inathymic and beige-athymic mice. Infect Immun1991;59:1785-9.

104. Chen SCA, Muller M, Zhou JZ, Wright LC, Sorrell TC.Phospholipase activity in Cryptococcus neoformans: anew virulence factor? J Infect Dis 1997;175:414-20.

105. Chen SCA, Wright LC, Santangelo RT, Muller M,Moran VR, Kuchel PW, Sorrell TC. Identification ofextracellular phospholipase B, lysophospholipase, andacyltransferase produced by Cryptococcus neoformans.Infect Immun 1997;65:405-11.

106. Alspaugh JA, Perfect JR, Heitman J. Cryptococcusneoformans mating and virulence are regulated by theG-protein α subunit GPA1 and cAMP. Genes Dev1997;11:3206-17.


Dispatches

In May 1993, a new hantaviral illness,hantavirus pulmonary syndrome (HPS), wasrecognized in the southwestern region of the UnitedStates (1). HPS is a viral zoonosis characterized bya febrile prodrome in young, healthy adults; thedisease progresses to respiratory failure with theclinical picture of adult respiratory distresssyndrome (ARDS). The striking pulmonaryinvolvement differentiates HPS from a previ-ously described hantaviral disease known ashemorrhagic fever with renal syndrome.

In the first 100 HPS cases in the UnitedStates, the average age was 34.9 years (range 11to 69); eight cases were in children or adolescentsunder 16 years of age (2). In Argentina, from 1987to July 1997, 114 cases were diagnosed in threeareas of the country where several strains of newworld hantaviruses are known to cause HPSdiseases (3,4). Before 1995, no cases weredetected in Argentine children under 12 years ofage. Ten cases were reported among adolescents(13 to 19 years) with a case-fatality rate of 30%(Instituto Nacional de Enfermedades ViralesHumanas, [INEVH], unpub. data).

The initial case definition referred to ARDSand included adults and young adults (5) as theaffected population. The lack of HPS cases amongchildren in the original outbreaks led to acirculating hypothesis that children were not atrisk or were at a very low risk for HPS. Anotherhypothesis was that children were protected frompulmonary involvement, perhaps by immunesystem immaturity or a lack of other risk factors(such as cigarette smoking) for lung injury.

In this report we describe five cases inchildren; in all of them the etiologic diagnosis wasestablished by the presence of immunoglobulin M(IgM) antibody to Sin Nombre virus (SNV)

antigens. Serologic results for two of the childrenwere also positive for SNV IgG antibody. Serumsamples were tested for IgM and IgG antibodiesto SNV by enzyme-linked immunosorbent assay(ELISA) (6). An ELISA titer greater than or equalto 1:400 was considered positive (Table 1).

Patient 1 was identified during the study of thefirst outbreak in southern Argentina in 1995 (5).Four patients in this outbreak were from the samefamily. During interviews of the family members,we found that a 9-year-old boy had a febrile diseasewithout respiratory involvement, beginning onApril 19. Serology performed on May 3, 14 daysafter the onset of symptoms, demonstrated IgM andIgG antibodies to SNV antigens.

Hantavirus Infection inChildren in Argentina

Clinical hantavirus infection was diagnosed in five Argentine children ages 5 to 11years by immunoglobulin M (IgM)- capture enzyme-linked immunosorbent assay usingSin Nombre virus (SNV) antigens. Death in three of the children was associated withabsence of detectable IgG to SNV antigens. An additional two cases in healthy childrenwere studied: one, a breast-fed 15-month-old whose mother died of suspectedhantavirus pulmonary syndrome (HPS) 8 months previously, had hantavirusIgG (≥ 1:6400); a second, whose mother survived HPS during month three ofpregnancy, apparently had maternal antibodies no longer detectable 1 year after birth.

Table 1. Hantavirus infection in children, Argentina, 1995–1997

Serologya

DateAge Date of IgM

Case Sex (yrs) onset IgG Areab Outcome 1 M 9 4-19-95 5-3-95 South Alive

>6400 1600

2 F 5 3-21-97 3-23-97 North Dead>6400 Neg

3 M 9 3-30-97 4-3-97 North Alive 1600 400

4 F 11 4-14-97 4-16-97 North Dead 1600 Neg

5 M 5 4-27-97 4-28-97 Central Dead 1600 Neg

aTiter expressed as the reciprocal of the serum dilutionreactive in enzyme-linked immunosorbent assay.bArea of origin in Argentina.


Dispatches

The other four cases in children wereidentified during routine surveillance. From1995 to 1997, samples from 25 children (ages 3months to 12 years; mean = 5.8 years) were sentto INEVH for diagnosis. Table 2 summarizesthe main clinical and laboratory findings. Noneof the children had renal failure; patient 2 haduremia of 0.30 g/l, and patient 4 had a serumcreatinine level of 1.40 g/l. All patients wholater died received supplemental oxygen aspart of their treatment.

Two special situations involving childrenarose during the study of the first cases of HPS inEl Bolsón in 1995. 1) A woman belonging to thefamily of patient 1 contracted HPS during thefirst quarter of pregnancy. She had a febrilesyndrome, without respiratory failure; chest X-rays showed bilateral interstitial infiltrates.Serologic tests showed both SNV IgM (≥ 1:6400)and IgG (≥ 1:6400) on April 22, 1995, 8 days afterthe onset of symptoms. She delivered a healthyinfant in October 1995. A sample of the newborn’scord blood was positive for SNV IgG (≥ 1:6400)and negative for SNV IgM. A serum sampledrawn from the mother at the same time had aSNV IgG titer ≥ 1:6400 and was negative for SNVIgM. A second serum sample, taken from the baby ayear later during November 1996, had nodetectable SNV IgG or IgM. 2) During aretrospective search for cases fulfilling the HPScase definition, a woman who died of ARDS inSeptember 1994 was considered to have a possible

case. No serum samples or autopsy tissues wereavailable to make an etiologic diagnosis. Beforedying, this woman breast-fed a 7-month-old baby;when tested for antibodies 8 months later in May1995 at 15 months of age, the baby had SNV IgG(≥ 1:6400) and no detectable SNV IgM antibodies.A second serologic sample, collected 18 monthslater in November 1996, still had SNV IgGantibodies, with a similar titer. Both babies havecontinued to develop normally as of October 1997.

A case similar to that of patient 1 wasdetected in New Mexico in June 1993 (7) in thecourse of the investigation of a fatal HPS case. Inthis case, the patient also had a mild clinicalcourse that did not meet the surveillance casedefinition for HPS. This case definition (revised9/96) is as follows: “a febrile illness characterizedby bilateral diffuse interstitial edema that mayradiographically resemble ARDS, with respira-tory compromise requiring supplemental oxygen,developing within 72 hours of hospitalization,and occurring in a previously healthy person; oran unexplained respiratory illness resulting indeath, with an autopsy examination demon-strating noncardiogenic pulmonary edemawithout an identifiable cause” (8).

Our remaining four cases were sporadic, inpersons without previous contact with other HPSpatients, and were suspected because theirclinical symptoms were typical of HPS. Results ofserologic testing with SNV antigens of thehousehold contacts in cases 2 and 5 (five persons

Table 2. Laboratory results and clinical features of children with hantavirus infection, Argentina, 1995–1997Tests and Casefeatures 1 2 3 4 5Leukocytes 9,200 27,000 12,800 10,600 69,200 (/mm3)Hematocrit (%) 44 66 43 55 53Thrombocytes NDa 266,000 200,000 97,000 ND (/mm3)Sedimentation rate ND 4 28 8 1 (mm/hr)GOT/GPTb ND Increased Increased Increased Increased (mild)Chest X-ray HIc DIId DII DII DIIRespiratory None Distress Slight dyspnea Tachypnea, Acute respiratory symptoms clinical and insufficiency

X-raydisassociation,hypoventilation

aND: Not done.bGOT/GPT: Glutamic oxalacetic transaminase/Glutamic pyruvic transaminase.cHI: Hilar indistinctness.dDII: Diffuse interstitial infiltrate.


Dispatches

each) were negative. The clinical, radiologic, andlaboratory findings were similar in children andin adults; severely ill patients had greatervariation in laboratory values than mild cases,and in fatal cases, only SNV IgM was present.

The case-fatality rate in this series was60%, but the small number of cases does notpermit conclusions. In previously reportedcases in adolescents 13 to 19 years of age, thecase-fatality rate was 30%.

These cases originated in the three areaswhere the illness is endemic in Argentina. This isan important point because an unusual case ofHPS in southern Argentina, with the possibilityof person-to-person transmission, had beenreported (9,10). Patient 1 and the baby that wasbreast-feeding when the mother died of suspectedHPS could be further instances of person-to-person transmission.

A case of hemorrhagic fever with renalsyndrome and pregnancy was reported in 1992(11); the dynamics of serum antibody persistencewere similar to those found in the one instancewhere we believe antibody was passivelytransferred from mother to baby. These resultsindicate that HPS should be considered in thedifferential diagnosis of respiratory distress oratypical bilateral pneumonia in children, at leastin areas where these diseases have beenconfirmed. Mild disease should be considered too,especially in contacts of HPS patients and inyounger age groups.

Our findings also suggest the transfer ofpassive antibodies from mother to fetus (withoutfetal infection) and the possibility of transmissionof infection by maternal breast feeding.

Noemí C. Pini,* Amanda Resa,† Gladys delJesús Laime,‡ Gustavo Lecot,¶ Thomas G.

Ksiazek,§ Silvana Levis,* and Delia A. Enria**Instituto Nacional de Enfermedades ViralesHumanas “Dr. Julio I Maiztegui,” Pergamino,Argentina; †Hospital de El Bolsón, Río Negro,

Argentina; ‡Hospital de Orán, Salta, Argentina;¶Hospital de Olavarría, Buenos Aires, Argentina;

and §Centers for Disease Controland Prevention, Atlanta, Georgia, USA

References 1. Centers for Disease Control and Prevention.

Outbreak of acute illness: southwestern United States,1993. MMWR Morb Mortal Wkly Rep 1993;42:421-4.

2. Khan AS, Khabbaz RF, Armstrong LR, Holman RC,Bauer SP, Graber J, et al. Hantavirus pulmonarysyndrome: the first 100 US Cases. J Infect Dis1996;173:1297-303.

3. Parisi MN, Enria DA, Pini NC, Sabattini M. Detecciónretrospectiva de infecciones clínicas por hantavirus enla Argentina. Medicina (B Aires) 1996;56:1-13.

4. Levis SC, Briggiler AM, Cacase M, Peters CJ, KsiazekTG, Cortés J, et al. Emergence of hantaviruspulmonary syndrome in Argentina [abstract]. Am JTrop Med Hyg 1995;53:233.

5. Hughes JM, Peters CJ, Cohen ML, Mahy BWJ.Hantavirus pulmonary syndrome: an emerginginfectious disease. Science 1993;262:850-1.

6. Ksiazek TG, Peters CJ, Rollin PE, Zaki S, Nichol S,Spiropoulou C, et al. Identification of a new NorthAmerican hantavirus that causes acute pulmonaryinsufficiency. Am J Trop Med Hyg 1995;52:117-23.

7. Armstrong LR, Bryan RT, Sarisky J, Khan AS, Rowe T,Ettestad PJ, et al. Mild hantavirus disease caused bySin Nombre virus in a four-year-old-child. PediatrInfect Dis J 1995;12:1108-10.

8. Centers for Disease Control and Prevention. Casedefinitions for infectious conditions under publichealth surveillance. MMWR Morb Mortal Wkly Rep1997;46:16.

9. Enria D, Padula P, Segura EL, Pini N, Edelstein A,Riva Posse C, Weissenbacher MC. Hantaviruspulmonary syndrome in Argentina. Possibility ofperson-to-person transmission. Medicina (B Aires)1996;56:709-11.

10. Wells RM, Sosa Estani S, Yadon ZE, Enria DA, PadulaP, Pini N, et al. An unusual hantavirus outbreak insouthern Argentina: person-to-person transmission?Emerg Infect Dis 1997;2:171-4.

11. Silberberg L, Rollin PE, Keourani G, Courdrier D.Haemorrhagic fever with renal syndrome andpregnancy: a case report. Trans R Soc Trop Med Hyg1993;87:65.


Dispatches

Cuba had its first dengue epidemic of moderntimes in 1977; transmission continued probablyuntil 1981, and more than 500,000 mild cases werereported. A 1978 serologic survey for flavivirusantibody indicated that 44.6% of the Cubanpopulation had been infected with dengue-1 virus,whereas before 1977 only 2.6% had antibodies (1,2).

A second dengue epidemic in 1981, caused bydengue-2 virus (2), was unusually severe andwidespread. Of 344,203 cases, 10,312 wereclinically classified as dengue hemorrhagic fever/dengue shock syndrome (DHF/DSS), and 158persons (101 children and 57 adults) died (3).Before 1981, only 60 suspected or confirmed DHFsporadic cases had been reported in the region(4). Dengue-2 virus isolated during the 1981epidemic was classified in the same genotype asNew Guinea 1944 (5). Not previously known tocirculate in the Americas, this genotype was notisolated again in the region until 1994 inVenezuela and in 1995 in Mexico (6). Retrospec-tive studies show that although the 1981epidemic was detected in May, the first casesoccurred in December 1980. After the epidemicended on October 10, 1981, a campaign toimprove mosquito control and eradicate Aedesaegypti was immediately launched. Eradicationwas not achieved, but most of the 169 Cubanmunicipalities were free of the vector.

Passive Surveillance—1981A passive dengue surveillance system was

established at the end of the 1981 epidemic. Of9,543 paired sera (acute- and convalescent-phase) from all suspected dengue patients, only14 showed seroconversion to immunoglobulin G(IgG) by enzyme-linked immunosorbent assay(ELISA) (7); none developed IgM antibodies todengue virus by capture IgM ELISA (8). Dengue

virus infection was excluded on the basis ofclinical and epidemiologic investigation. No Ae.aegypti mosquitoes were found in the residencelocalities of these patients. The surveillancesystem detected cases, imported from other LatinAmerican countries, that had no evidence ofindigenous transmission. Since 1987, 4,983samples received through the surveillance systemfor measles and rubella, as well as paired sera ofpatients with rash, were studied for dengueantibodies [María Guzmán, World Health Organi-zation (WHO)/Pan American Health Organization(PAHO) Collaborating Center for the Study of ViralDiseases, unpub. info.]. No dengue cases wereidentified. The low Ae. aegypti premise indexes andthe results of the passive surveillance systemindicate no dengue transmission in Cuba between1981 and the end of 1996. However, reinfestationhas occurred in some areas; the municipality ofSantiago de Cuba was reinfested in 1992 byAe. aegypti transported in imported tires (9).

Active Surveillance—1997In January 1997, the Institute of Tropical

Medicine “Pedro Kourí” of the Cuban Ministry ofHealth (a WHO/PAHO Collaborating Center forthe Study of Viral Diseases) established an activesurveillance system for dengue in Santiago deCuba municipality. The municipality is located inSantiago de Cuba province, in the eastern part ofthe country, and has several risk factors for thereemergence of dengue: limited water supply,inadequate eradication efforts, high vectorinfestation, and increasing migration of peoplefrom Latin American and Caribbean disease-endemic countries to the municipality. Followingthe Guidelines for the Prevention and Control ofDengue and Dengue Hemorrhagic Fever in theAmericas (4), this surveillance system actively

Reemergence of Dengue in Cuba: A 1997Epidemic in Santiago de Cuba

After 15 years of absence, dengue reemerged in the municipality of Santiagode Cuba because of increasing migration to the area by people from disease-endemic regions, a high level of vector infestation, and the breakdown oferadication measures. The 1997 epidemic was detected early through an activesurveillance system. Of 2,946 laboratory-confirmed cases, 205 were denguehemorrhagic fever, and 12 were fatal. No deaths were reported in persons under 16years of age. Now the epidemic is fully controlled.


Dispatches

searched for febrile patients in the primaryhealth-care subsystem whose clinical picture wascompatible with dengue fever and whose seracollected 5 to 6 days after onset of the diseasecontained dengue IgM antibodies. As a result ofthis system, dengue cases were detected onJanuary 28, 1997, in one area of the municipality.In three of the first seven cases, dengue-2 viruswas detected by polymerase chain reaction (10)and was confirmed by viral isolation andidentification using C6/36 cell line and mono-clonal antibodies to the four dengue serotypes.

Although retrospective seroepidemiologicstudies indicated that the initial transmissionoccurred during the second half of December1996, it is highly probable that the cases detectedon January 28 were the first. Of 60,000 casesreported from the emergency rooms of Santiagode Cuba hospitals from November 1 to January28, 592 were clinically compatible with denguefever. Home interviews of these 592 patientsreduced the figure to 154. Blood samples from 143of 154 patients were examined for IgMantibodies, but no positive cases were detected.

The breakdown of the vector controlcampaign in this municipality interfered with ourefforts to abort the epidemic, despite the earlydetection of the first dengue cases; however, thepartial vector control measures implementedonce the outbreak was detected prevented itsextension to the other 30 Cuban municipalitiesinfested with the Ae. aegypti mosquito. Active surveillance continued from January toJuly 1997. Serologic confirmation of cases wascarried out by IgM capture ELISA, confirmingrecent infection. The serologic diagnosis wasdecentralized to the Provincial Laboratory inSantiago de Cuba, which used an ultramicro-ELISA for dengue IgM detection (11). The Instituteof Tropical Medicine served as the nationalreference laboratory for serology, viral isolation,and strain identification and characterization.

During the epidemic, 17,114 febrile patientswere initially considered to have dengue, but serologictesting of 10,024 of these patients confirmed denguein only 2,946; 46 dengue-2 isolates from 160 serumsamples were obtained. The nucleotide sequence ofthe E\NS1 gene junction of the first isolated strain(12) indicated that it belonged to the Jamaicagenotype, which during recent years is beingtransmitted extensively throughout Latin Ameri-can and Caribbean countries and is associated withDHF/DSS in some countries (6,13).

EpidemiologyAfter the end of the 1981 Cuban DHF

epidemic, seroepidemiologic studies in Palmira,Cienfuegos, and Cerro municipalities examineddengue-1 and dengue-2 seroprevalence in thesepopulations (14,15). Taking into considerationthese data and the total population of theSantiago de Cuba municipality, we estimated theprevalence of dengue-1 and dengue-2 antibodies.The estimated total population at risk fordengue-2 infection was 301,986 adults andchildren susceptible to a primary infection by anydengue virus serotype (63.5% of the population)and 88,108 adults with antibodies to dengue-1virus acquired during the epidemic of 1977 to1980, now susceptible to a secondary infectionwith dengue-2 and at increased risk for DHF/DSS(18.5% of the population).

The earlier Cuban experience (3) confirmsother reports of secondary infection (dengue-1 anddengue-2) as the main risk factor for DHF/DSS.During the 1997 dengue outbreak, secondaryinfection was again confirmed as a risk factor forDHF/DSS. Of the 2,946 confirmed cases, 205(including 12 fatal adult cases) were classified asDHF/DSS cases according to the criteria estab-lished by PAHO (4). DHF/DSS was observed mostlyin adults, the only age group in whom secondaryinfection was possible. DHF/DSS-compatible symp-toms were seen only in one child with primaryinfection. Preliminary studies indicated thatsecondary infection was present in 100 (98%) of 102DHF/DSS cases. In fatal cases, secondary infectioncould be documented in 11 (92%) of 12 cases. InThailand the greatest risk appeared when thesecondary infection occurred 6 months to 5 yearsafter the primary one (16). For that reason, anepidemic of DHF/DSS was not expected inSantiago de Cuba, perhaps only sporadic cases.However, DHF/DSS in adults who contracted asecondary infection at least 16 years after theprimary infection was not previously reported.

Because in Cuba dengue-1 circulated from1977 to 1980-81, the youngest patients expectedto contract secondary infection should be olderthan 16 years of age; the youngest DHF/DSSpatient with confirmed secondary infection was a17-year-old, which indicates that the “enhancing”antibodies can circulate and be effective for atleast 16 years and maybe for life.

A significant number of febrile patients withsuspected dengue had respiratory signs andsymptoms; therefore, simultaneous circulation of


Dispatches

respiratory or other pathogens was considered.Serologic screening for respiratory viruses usinghemagglutination-inhibition and ELISA con-firmed that 29.3% of 41 nonconfirmed denguecases were influenza A, influenza B, oradenovirus infections. Additionally, some chil-dren had fever and rash clinically compatiblewith herpangina, and some had diarrheal diseasewith fever, as is common in Cuba during thesummer. These febrile syndromes contributed tothe high number of patients whose infectionswere provisionally considered suspect denguecases. Suspect dengue cases were broadly definedto maximize sensitivity of detection and retain allpossible dengue cases. This active surveillanceexcluded other febrile syndromes but recordedthem as suspected cases. In practice, the riskperception by the population was very high,especially when the epidemic was officially declaredand deaths were noted. Both the patients and thehealth providers appeared to think of dengue asthe first diagnostic possibility. For this reason,the figure of 17,114 cases was considered themagnitude of the epidemic from the clinicalmanagement perspective. Since most cases weretested serologically, the incidence of clinical caseswas probably close to the 2,946 serologically orvirologically confirmed cases. Because asymp-tomatic and subclinical dengue cases are frequent,especially in children, the true rate of infectionmay be higher. In a separate and limited study onasymptomatic contacts of dengue cases, for everyclinical case, 13.9 asymptomatic or subclinicalcases were produced. Serologic studies of contactsin Santiago de Cuba are planned for a more in-depth study of this question.

Clinical ManagementThe health authorities established a liberal

policy of hospitalization that varied with theavailability of beds. Hospitalization permittedvector control of the human reservoir, moreprecise case classification, and close clinicalsurveillance.

When beds were available, all patients withsuspected cases were hospitalized. When thenumbers of patients surpassed the availability ofbeds, patients were treated at home under thesupervision of the family doctor. The familydoctor transferred the patient to the hospital ifany medical complication appeared. Wards withspecialized personnel were established where thepatients were protected from vectors, and

observation wards were organized for patientswith complications. Intensive and intermediatecare units, as well as an emergency subsystem forthe transfer of patients from one unit to another,were available. As in 1981, some patients rapidlydeveloped hypovolemic shock and died withinhours of admission to the hospital (17).

An ad hoc task force followed the casedefinitions for dengue and DHF/DSS establishedby PAHO (4) for classifying the cases at theclosure of the medical record. The accumulatedexperience of the Cuban scientists and doctorsand the increased international knowledge aboutdengue and DHF/DSS in the last 15 yearspermitted a much deeper and more comprehen-sive study of this outbreak with more accurateclassification and management of cases than in1981. Nevertheless, the case-fatality rate wasthree times higher, mainly because of a muchbetter classification of DHF/DSS cases. Othercountries in the region with a very accurate caseclassification, such as Puerto Rico (13), also havea high case-fatality rate.

Vector ControlThe campaign to control the vector started

before the beginning of the 1997 dengue outbreakand is well established. Although the campaignrequired the mobilization of scarce financialresources and experts from all over the country,early intervention prevented spread of theoutbreak to other potentially vulnerable mu-nicipalities. Of 169 municipalities in Cuba, 30had Ae. aegypti mosquitoes. The epidemic waslimited to the municipality of Santiago de Cuba;no autochthonous transmission to other munici-palities of the province or country was detected.

An active search for cases detected transmis-sion very early, before “fever alert” signaled anoutbreak. In the Provincial Center for Hygiene,Epidemiology, and Microbiology of Santiago deCuba, a special Unit for Analysis and Trendsmaintains a permanent fever alert system. Forseveral years, this system has provided a weeklytabulation of febrile patients for every popula-tion. The tabulation allows us to evaluate feveralert (4) as applied to an active surveillancesystem. Because the fever alert did not appear inthe epidemic area until May 1997, after theepidemic was already occurring, we considerfever alert an indicator with low sensitivity for theearly and timely detection of dengue transmission,at least under the conditions of this study.


Dispatches

As a result of the 1997 epidemic, anepidemiologic alert was established, andantivector intervention, as well as activeseroepidemiologic surveillance, was reinforced inthe entire country. The epidemiologic character-ization of the outbreak (now fully controlled) is inthe final phase. Although mosquitoes persisted ata low level after the 1981 DHF/DSS epidemic, thecampaign was successful in eradicating denguefrom Cuba for more than 15 years, precisely whenthe disease was reemerging in nearly all the othertropical regions of the Americas. According toPAHO, 250,707 cases of dengue fever and 4,440cases of DHF/DSS were reported in 1996 alone;29 countries reported dengue in 1996, and 10 ofthese reported DHF/DSS. Overall, from 1981 to1996, 25 countries reported 41,000 cases of DHF/DSS (F. Pinheiro, pers. comm.).

The 1997 Cuban dengue outbreak demon-strated once again that dengue reappears whereAe. aegypti control is relaxed. Taking into accountthese facts, Cuba maintains its policy of vectoreradication and recommends an exerted effort inthe American region to prevent a recurrence ofdengue similar to the one in Southeast Asia,where DHF/DSS is the leading cause ofhospitalization and death among children (18).

Gustavo Kourí,* María Guadalupe Guzmán,†Luis Valdés,‡ Isabel Carbonel,‡

Delfina del Rosario,* Susana Vazquez,*José Laferté,* Jorge Delgado,§ and

María V. Cabrera‡*Instituto de Medicina Tropical “Pedro Kourí” (IPK),

Marianao, Havana, Cuba; †WHO/PAHOCollaborating Center for the Study of Viral Diseases,

Havana, Cuba; ‡Provincial Center for Hygiene,Epidemiology and Microbiology, Santiago de Cuba,

Cuba; and §Ministry of Public Health, Cuba

References 1. Cantelar N, Fernández A, Albert L, Pérez E.

Circulación de dengue en Cuba 1978-1979. Rev CubanaMed Trop 1981;33:72-8.

2. Kourí G, Mas P, Guzmán MG, Soler M, Goyenechea A,Morier L. Dengue hemorrhagic fever in Cuba, 1981:rapid diagnosis of the etiologic agent. Bull Pan AmHealth Org 1983;17:126-32.

3. Kourí G, Guzmán MG, Bravo J, Triana C. Denguehemorrhagic fever/dengue shock syndrome: lessonsfrom the Cuban epidemic. Bull World Health Organ1989;67:375-80.

4. Dengue and dengue hemorraghic fever in theAmericas: guidelines for prevention and control.Washington: Pan American Health Organization;1994. Scientific publication No. 548.

5. Guzmán MG, Deubel V, Pelegrino JL, Rosario D, SariolC, Kourí G. Partial nucleotide and amino-acidsequences of the envelope and the envelope/nonstructural protein-1 gene junction of four dengue 2virus strains isolated during the 1981 Cuban epidemic.Am J Trop Med Hyg 1995:52:241-6.

6. Ricco-Hesse R, Harrison LM, Salas RA, Tovar D,Nisalak A, Ramos C, et al. Origins of dengue type 2viruses associated with increased pathogenicity in theAmericas. Virology 1997;230:244-51.

7. Fernández R, Vázquez S. Serological diagnosis ofdengue by an ELISA inhibition method (EIM). MemInst Oswaldo Cruz 1990;85:347-51.

8. Vázquez S, Saenz E, Huelva G, González A, Kourí G,Guzmán MG. Detección de IgM contra el virus deldengue en sangre entera absorbida en papel de filtro.Rev Panamericana de Salud Pública. In press 1998.

9. Ministerio de Salud Pública de Cuba. Dengue en Cuba.Julio 1997. Boletín Epidemiológico OrganizaciónPanamericana de la Salud 1997;18:7.

10. Lanciotti RS, Calisher CH, Gubler DG, Chang G,Vordam V. Rapid detection and typing of dengueviruses from clinical samples by using reversetranscriptase-polymerase chain reaction. J ClinMicrobiol 1992;30:545-51.

11. Laferte J, Pelegrino JL, Guzmán MG, González G,Vázquez S, Hermida C. Rapid diagnosis of dengue virusinfection using a novel 10µl IgM antibody captureultramicroELISA assay (MAC UMELISA Dengue).Advances in Modern Biotechnology 1992;1:19.4.

12. Rico-Hesse R. Molecular evolution and distribution ofdengue viruses type 1 and 2 in nature. Virology1990;174:479-93.

13. División de Prevención y Control de Enfermedades,Programa de Enfermedades, Programa deEnfermedades Transmisibles, HCP/HCT, OPS.Resurgimiento del dengue en las Américas. BoletínEpidemiológico. Organización Panamericana de laSalud 1997;18:1-6.

14. Guzmán MG, Kourí G, Bravo J, Hoz de la F, Soler M,Hernández B. Encuesta seroepidemiológicaretrospectiva a virus dengue en los municipiosCienfuegos y Palmira. Rev Cubana Med Trop1989;41:321-32.

15. Guzmán MG, Kourí G, Bravo J, Soler M, Vázquez S,Morier L. Dengue hemorrhagic fever in Cuba, 1981: aretrospective seroepidemiologic study. Am J Trop MedHyg 1990;42:179-84.

16. Halstead SB. Observations related to pathogenesis ofdengue hemorrhagic fever. Yale J Biol Med1970;42:350-60.

17. Díaz A, Kourí G, Guzmán MG, Lobaina L, Bravo J, RuizA, et al. Description of the clinical picture of denguehemorrhagic fever/dengue shock syndrome (DHF/DSS) in adults. Bull Pan Am Health Organ1988;22:133-44.

18. Gubler DJ, Clark GG. Dengue/dengue hemorrhagicfever: the emergence of a global health problem. EmergInfect Dis 1995;1:55-7.


Dispatches

Since its discovery in 1993 (1) and itsassociation with Sin Nombre virus in NorthAmerica (2), hantavirus pulmonary syndrome(HPS) has been identified in several countries inSouth America and is associated with Juquitibavirus in Brazil (3), Andes virus in Argentina andChile (4-6), and Laguna Negra virus in Paraguay(7,8). Hantaviruses are rodent-borne, and each isassociated with a specific primary rodentreservoir. Sigmodontine rodents are the vectorsof hantaviruses associated with HPS (3).

Case ReportA 20-year-old male resident of Santiago,

Chile, who had no prior history of medicalproblems, became ill after he backpacked (fromFebruary 4 to March 9, 1997) as a tourist inBolivia. His travel itinerary included Barrio,Oruro, La Paz, Cochabamba, Villa Tunari, SantaCruz, Vallegrande, Higueras, and Sucre. Thetourist and a fellow traveler stayed in hotel inSanta Cruz where they saw black rats runningthrough the bathroom they used. While inHigueras, they stayed in a rustic adobe housewith no floor and joined local villagers inagricultural jobs, primarily harvesting hay. Adiagnosis of HPS was considered.

On March 26, 3 weeks after returning toChile, the young man became ill with fever andcough. On March 28, he was admitted to theemergency room of a private hospital in Santiagowith high fever (39°C), cough, and chest pain.Chest X-rays showed interstitial infiltrates in theleft lower lobe. Laboratory results were asfollows: hemoglobin 15.3 gm/dl, white blood cells5,900, platelets 130,000, erythrocyte sedimenta-

tion rate 6, and a C-reactive protein 2.8 mg/dl. Hewas sent home on clarithromycin 250 mg, twice aday. Three days later (March 31), he returned tothe emergency room with persistent high fever,myalgia, and shortness of breath; distal cyanosiswas noted. Vital signs were as follows: bloodpressure 115/80, pulse 120, temperature 38.5°C,and respiratory rate 32. Physical examinationfound few petechiae on the forearms, diffusebilateral rales, regular cardiac rhythm, nomurmurs, and no hepatomegaly or splenomegaly.Chest X-rays showed diffuse bilateral alveolarinfiltrates. Oxygen saturation was 90%. Labora-tory values included white blood cells 11,700 witha left shift, platelets 150,000, hemoglobin 19.4,erythrocyte sedimentation rate 2, C-reactiveprotein 8.18, INR 1.6, activated partial thrombo-plastin 43s, blood urea nitrogen 38.9 mg/dl, andliver function tests normal limits.

The patient was transferred to the intensivecare unit with a diagnosis of bilateral pneumoniaof unknown etiology and secondary respiratoryfailure. In 3 to 4 hours, acute respiratory distressdeveloped; arterial blood gases (on 50% oxygen)were PaO2 61, PaCO2 22.7, pH 7.49, and HCO317.1. The patient was connected to a ventilatorand was started on imipenen, erythromycin,amantadine, dopamine, dobutamine, fluids, andnitric oxide. Two 500-mg doses of methylpred-nisolone were administered, and a Swan-Ganzcatheter was installed. The pulse wedge pressurewas 7, and the cardiac output 6.1. During the first24 hours, acute renal failure developed, andhemodialysis was started.

Hypotension was quite refractory to vasoac-tive drugs (mean BP 30 to 40), and critical

Hantavirus Pulmonary Syndromein a Chilean Patient

with Recent Travel in Bolivia

A case of hantavirus pulmonary syndrome (HPS) was serologically confirmed in acritically ill patient in Santiago, Chile. The patient’s clinical course had many similaritiesto that of other HPS patients in North and South America but was complicated by acutesevere renal failure. The patient’s history included self-reported urban and probablerural rodent exposure during travel in Bolivia. Comparison of a viral sequence from anacute-phase serum sample with other known hantaviruses showed that the hantavirusnucleic acid sequence from the patient was very similar to a virus recently isolated fromrodents associated with HPS cases in Paraguay.


Dispatches

hypoxemia (O2 saturation 70% to 80%) andhypotension persisted for 48 hours.Echocardiography showed a mild pericardialeffusion. Blood cultures were negative forbacteria, as well as for influenza, adenovirus,respiratory syncytial virus, and parainfluenzavirus. Serologic results for Mycoplasma,Legionella, and HIV were also negative. Bloodgases started to improve on day 3, and thepatient’s clinical condition also slowly improved.On day 10, hemolytic anemia (Coombs negative)developed, and a bone marrow aspirate showedhemophagocytosis. On day 11, an episode ofbleeding from the respiratory tract occurred.However, the patient’s ventilatory functioncontinued to improve. On day 15 after admission,his chest X-ray was clearing, but he was still on amechanical ventilator (O2 saturation 98% on 40%O2). However, he was anuric (BUN 100). On day17, he had a second episode of bronchial bleeding,and bronchoscopy showed only a 4-mm trachealulcer. Platelets were 72,000, and hematocrit fellto 24%. Hantavirus serology was reportedpositive [Centers for Disease Control andPrevention (CDC), April 18], and intravenousribavirin was started on April 22. A centralvenous catheter-related infection with Pseudomo-nas and Staphylococcus aureus was documented.The patient’s condition deteriorated progres-sively, with further bronchial bleeding andmarkedly unstable ventilatory function, andrequired constant administration of vasoactivedrugs; he died of massive pulmonary hemorrhageand shock on April 28. A postmortem lung samplewas taken with a needle.

The patient’s serum, collected on April 1, wastested for immunoglobulin G (IgG) and IgMantibodies to Sin Nombre virus at CDC. Both IgGand IgM antibodies were found, which suggestedrecent infection with a hantavirus associatedwith Sigmodontine rodents. Subsequent testingof sera collected on April 21 and April 25confirmed the initial findings. Hematoxylin andeosin stained sections of the lung sample showeddiffuse alveolar damage with extensive hyalinemembrane formation, proliferation of type IIpneumocytes, and fibroblastic and edematousthickening of alveolar walls. Abundant fibrin,necrotic debris, and acute inflammatory cellularinfiltrates were also observed in the alveolarspaces. Rare endothelial cells and macrophageswere hantavirus-antigen positive by previously

described immunohistochemical procedures (9).The destructive changes seen by histopathologyand a small amount of antigen found in thepatient’s tissues are compatible with the longcourse of the patient’s illness (9).

Viral RNA was extracted from the earliestserum sample, and reverse transcriptase-polymerase chain reaction amplification withprimers designed specifically for hantavirusesassociated with Sigmodontine rodents (8) yieldedpolymerase chain reaction fragments for the Sand M segments. These cDNA fragments weresequenced and compared with those of otherAmerican hantaviruses. These comparisons showthat the virus is closely related to other SouthAmerican hantaviruses, and most closely relatedto Laguna Negra virus detected in patients andCalomys laucha rodents (vesper mice) in theChaco region of Paraguay (7,8). The nucleotidesequence identity was 84% for the G1 proteinencoding fragment and 87% for the N proteinencoding fragment. However, the deduced aminoacid sequences for both fragments were identicalto Laguna Negra virus. This shows that the virusassociated with this HPS case is a Laguna Negravirus variant and suggests that the virus isprobably associated with the same or a veryclosely related species of rodent host.

C. laucha, the apparent primary rodentreservoir for Laguna Negra virus, is common insavanna and grassland areas as far north assouthern Brazil and Bolivia, throughout much ofParaguay and Uruguay, and in Argentina as farsouth as Rio Negro province, but not in Chile (10).A number of the lowland rural locations that thepatient visited in Bolivia are within the range ofC. laucha.

The clinical course and travel history of thepatient and the laboratory serology andmolecular characterization of the viral RNA arecompatible with infection in Bolivia with aLaguna Negra virus variant. This report andevolving information concerning hantavirusesassociated with clinical HPS in Argentina andParaguay strongly suggest that a diagnosis ofHPS should be considered in patients with febrilerespiratory distress syndrome throughout LatinAmerica.

Ricardo Espinoza,* Pablo Vial,*† Luis M.Noriega,*† Angela Johnson,‡ Stuart T. Nichol,‡Pierre E. Rollin,‡ Rachel Wells,‡ Sherif Zaki,‡Enrique Reynolds,* and Thomas G. Ksiazek‡


Dispatches

*Clínica Alemana de Santiago, Santiago, Chile;†Escuela de Medicina, Universidad Católica de

Chile, Santiago, Chile; ‡Centers for Disease Controland Prevention, Atlanta, Georgia, USA.

References 1. Duchin JS, Koster F, Peters CJ, Simpson G, Tempest

B, Zaki S, et al. Hantavirus pulmonary syndrome: aclinical description of 17 patients with a newlyrecognized disease. N Engl J Med 1994;330:949-55.

2. Nichol ST, Spiropoulou CF, Morzunov S, Rollin PE,Ksiazek TG, Feldmann H, et al. Genetic identificationof a hantavirus associated with an outbreak of acuterespiratory illness. Science 1993;262:914-7.

3. Nichol ST, Ksiazek TG, Rollin PE, Peters CJ.Hantavirus pulmonary syndrome and newly describedhantaviruses in the United States. In: Elliott RM,editor. The Bunyaviridae. New York: Plenum Press,1996:269-80.

4. Lopez N, Padula P, Rossi C, Lazaro ME, Franze-Fernandez MT. Genetic identification of a newhantavirus causing severe pulmonary syndrome inArgentina. Virology 1996;220:223-6.

5. Centers for Disease Control and Prevention. Hantaviruspulmonary syndrome in Chile—1997. MMWR MorbMortal Wkly Rep 1997;46:949-51.

6. Lopez N, Padula P, Rossi C, Miguel S, Edelstein A,Ramirez E, et al. Genetic characterization andphylogeny of Andes virus and variants from Argentinaand Chile. Virus Research 1997;50:77-84.

7. Williams RJ, Bryan RT, Mills JN, Palma RE, Vera I, deVelasquez F, et al. An outbreak of hantaviruspulmonary syndrome in Western Paraguay. Am J TropMed Hyg 1997;57:274-82.

8. Johnson AM, Bowen MD, Ksiazek TG, Williams RJ,Bryan RT, Mills JN, et al. Laguna Negra virusassociated with HPS in western Paraguay and Bolivia.Virology. In press 1997.

9. Zaki S, Greer PW, Coffield LM, Goldsmith CS, NolteKB, Foucar K, et al. Hantavirus pulmonary syndrome:pathogenesis of an emerging infectious disease.American Journal of Pathology 1995;146:552-79.

10. Redford KH, Eisenberg JF. Mammals of theneotropics, the southern cone. Chicago: The Universityof Chicago Press, 1992:279-80.


Dispatches

The vector-borne diseases Lyme disease,human babesiosis, and human granulocyticehrlichiosis (HGE) are emerging in the Northeastand upper Midwest regions of the United States(1,2). The etiologic agents for these diseases(Borrelia burgdorferi, Babesia microti, and theHGE agent, respectively) appear to share thesame vertebrate reservoir host (Peromyscusleucopus) and tick vector (Ixodes scapularis) (3-6).Immunologic evidence of human coinfectionwith these pathogens has been reported (7-10),and a culture-confirmed case of coinfection withB. burgdorferi and the HGE agent has recentlybeen described (11). Thus, in Lyme disease–endemic areas, there may be a substantial riskfor coinfection with B. burgdorferi and either B.microti or the HGE-causing rickettsia.

Hunterdon County (population 118,000) is arural, but rapidly developing, county in westernNew Jersey (Figure), with many homes in woodedsettings. In 1996, the county had the third highestcase rate of Lyme disease (524 per 100,000) of allcounties in the United States (Centers for DiseaseControl and Prevention case reports). The countyhas also had many suspected, but no serologicallyconfirmed, cases of HGE.

The risk of acquiring a tick-borne pathogen ina specific geographic location depends largely ontick density and the prevalence rate of the agentin the tick population. Although cultivation of amicrobial agent is considered optimal, moleculardetection is more practical since it permits therapid assay of large numbers of individualspecimens. We report here the prevalence rate ofB. burgdorferi, B. microti, and the agent of HGE

in I. scapularis in Hunterdon County by species-specific polymerase chain reaction (PCR).

One hundred adult I. scapularis werecollected by drag cloth or from personal clothingat 10 sites in the county during the fall of 1996(Figure). Ticks were stored at room temperaturein 70% ethanol until analysis. DNA was isolatedfrom individual ticks with the Isoquick DNAextraction kit (ORCA Research, Bothell, WA)

Prevalence of Tick-Borne Pathogensin Ixodes scapularis in a Rural

New Jersey County

To assess the potential risk for other tick-borne diseases, we collected 100adult Ixodes scapularis in Hunterdon County, a rapidly developing rural county inLyme disease–endemic western New Jersey. We tested the ticks by polymerasechain reaction for Borrelia burgdorferi, Babesia microti, and the rickettsial agent ofhuman granulocytic ehrlichiosis (HGE). Fifty-five ticks were infected with at leastone of the three pathogens: 43 with B. burgdorferi, five with B. microti, and 17 withthe HGE agent. Ten ticks were coinfected with two of the pathogens. The resultssuggest that county residents are at considerable risk for infection by a tick-bornepathogen after an I. scapularis bite.

Figure. Map of New Jersey showing HunterdonCounty. Black dots indicate tick collection sites.


Dispatches

(12). The final DNA pellets were resuspended in50 µl of sterile water, and a 10-µl aliquot wasused for each PCR test. B. burgdorferi–specificPCR used primers IS1 and IS2 (13). The HGEagent was detected by PCR using primers GER3and GER4, as described by Munderloh et al. (14),and B. microti DNA was detected according tomethods used by Persing et al. (15). PCR productswere electrophoresed on 2% agarose gels stainedwith ethidium bromide. For B. burgdorferi and B.microti, hybridization with specific probes wascarried out after transfer to nylon membranes(13). In each PCR experiment appropriatenegative controls were used; these includedmaster mix controls lacking template andreactions containing B. burgdorferi DNA, HGEagent DNA, or Escherichia coli DNA as additionalcontrols for HGE agent, B. burgdorferi, or B. microtiPCR, respectively. To prevent cross-contamina-tion, specimen preparation, PCR amplification,and post-PCR analysis were performed inseparate laboratories.

Results of PCR analysis for each of thepathogens on individual ticks are presented inthe Table. At least one of the three pathogens waspresent in 55% of the ticks. The highestprevalence rate was that for B. burgdorferi (43/100), followed by the agent of HGE (17/100) andB. microti (5/100). Furthermore, these pathogenswere widely distributed. At least one tick at eachsite was infected with B. burgdorferi, nine of 10sites had HGE agent–positive ticks, and B.microti–positive ticks were found at fivedifferent collection sites. Ten ticks werecoinfected with two of the three agents; no tickswere infected with all three pathogens. The B.microti PCR amplification products were gelpurified and subjected to automated DNAsequencing. The 238 bp products yielded

sequences identical to that reported for B.microti 18S ribosomal DNA (15).

The prevalence rates for the three pathogensreported here are consistent with earlier studiesat selected sites in the northeastern UnitedStates. Telford et al. reported prevalence rates of36%, 11%, and 9% for B. burgdorferi, the HGEagent, and B. microti, respectively, in adult I.scapularis on Nantucket Island, Massachusetts;coprevalence of B. burgdorferi and the HGEagent was 4%, and no simultaneous infectionwith B. burgdorferi and B. microti was observed(4). We have reported prevalence rates of 52%and 53% for B. burgdorferi and the HGEpathogen, and coprevalence of 26% at a site inWestchester County, New York (5). Given thesimultaneous infection of I. scapularis with thesepathogens, it is not surprising that numerousserosurveys of Lyme disease patients indicate thecoincident presence of antibodies to B. burgdorferiand B. microti and human granulocytic ehrlichia(7-10). These findings do not establish concur-rent, active infection, nor do they demonstratesimultaneous transmission by a single tick bite;however, they suggest that persons living indisease-endemic areas are exposed to theseagents, presumably as a result of tick bites.

Human infection by any of the three tick-borne agents alone generally results in similaracute manifestations, including fever, headache,and myalgia (1,10,16). Reported cases of Lymedisease greatly outnumber those of babesiosis orHGE. This, in conjunction with seroepidemiologicdata, suggests that many infections with B. microtiand granulocytic ehrlichia are subclinical. Dualinfection with B. burgdorferi and B. microti oftenresults in more severe illness (10); whether this isalso true for concurrent infection by the agents ofLyme disease and HGE remains to be established.Human coinfection by multiple tick-borne agentsmay account for the variable nature of the clinicalmanifestations of Lyme disease.

This study demonstrates that the agents ofLyme disease, human babesiosis, and HGEcoexist in the tick population of a previouslyidentified Lyme disease–endemic region of NewJersey. The study provides further evidence thatthese three pathogens may be prevalentthroughout the range of I. scapularis. Infectionwith any of these three tick-borne pathogensshould be considered for residents or visitors of adisease-endemic area who have flulike symptomsand a history of a tick bite.

Table. Prevalence rate of Borrelia burgdorferi , humangranulocytic ehrlichiosis (HGE) agent, and Babesiamicroti in 100 adult Ixodes scapularis

Pathogen No. infected ticksB. burgdorferi 43HGE agent 17B. microti 5B. burgdorferi and HGE agent 6B. burgdorferi and B. microti 2HGE agent and B. microti 2B. burgdorferi, HGE agent, and B. microti 0


Dispatches

AcknowledgmentsWe thank Gary Wormser and Dennis White for helpful

comments, Chris Kolbert and David Persing for providingadvice and positive controls for B. microti PCR, and theHunterdon County Tickborne Disease Research Group andHunterdon Medical Center for their support. This work wassupported in part by NIH grant AR41511.

Shobha Varde,* John Beckley,†and Ira Schwartz*

*New York Medical College, Valhalla, New York,USA; and †Department of Health, County ofHunterdon, Flemington, New Jersey, USA

References 1. Walker DH, Barbour AG, Oliver JH, Lane RS, Dumler

JS, Dennis DT, et al. Emerging bacterial zoonotic andvector-borne diseases: ecological and epidemiologicalfactors. JAMA 1996;275:463-9.

2. Persing DH, Conrad PA. Babesiosis: new insightsfrom phylogenetic analysis. Infect Agents Dis1995;4:182-95.

3. Anderson J, Mintz E, Gadbaw J, Magnarelli L. Babesiamicroti, human babesiosis and B. burgdorferi inConnecticut. J Clin Microbiol 1991;29:2779-83.

4. Telford SR III, Dawson JE, Katavolos P, Warner CK,Kolbert CP, Persing DH. Perpetuation of the agent ofhuman granulocytic ehrlichiosis in a deer tick-rodentcycle. Proc Natl Acad Sci U S A 1996;93:6209-14.

5. Schwartz I, Fish D, Daniels TJ. Prevalence of therickettsial agent of human granulocytic ehrlichiosis inticks from a hyperendemic focus of Lyme disease[letter]. N Engl J Med 1997;337:49-50.

6. Piesman J, Mather TN, Telford SR III, Spielman A.Concurrent Borrelia burgdorferi and Babesia microtiinfection in nymphal Ixodes dammini. J Clin Microbiol1986;24:446-7.

7. Benach JL, Coleman JL, Habicht GS, MacDonald A,Grunwaldt E, Giron JA. Serological evidence forsimultaneous occurrences of Lyme disease andbabesiosis. J Infect Dis 1985;152:473-7.

8. Magnarelli LA, Dumler JS, Anderson JF, Johnson RC,Fikrig E. Coexistence of antibodies to tick-bornepathogens of babesiosis, ehrlichiosis, and Lyme borreliosisin human sera. J Clin Microbiol 1995;33:3054-7.

9. Mitchell PD, Reed KD, Hofkes JM. Immunoserologicevidence of coinfection with Borrelia burgdorferi,Babesia microti, and human granulocytic Ehrlichiaspecies in residents of Wisconsin and Minnesota. J ClinMicrobiol 1996;34:724-7.

10. Krause PJ, Telford SR III, Spielman A, Sikand V, RyanR, Christianson D, et al. Concurrent Lyme disease andbabesiosis. Evidence for increased severity andduration of illness. JAMA 1996;275:1657-60.

11. Nadelman RB, Horowitz HW, Hsieh TC, Wu JM,Aguero-Rosenfeld ME, Schwartz I, et al. Simultaneoushuman granulocytic ehrlichiosis and Lyme borreliosis.N Engl J Med 1997;337:27-30.

12. Schwartz I, Varde S, Nadelman RB, Wormser GP, FishD. Inhibition of efficient polymerase chain reactionamplification of Borrelia burgdorferi DNA in blood-fedticks. Am J Trop Med Hyg 1997;56:339-42.

13. Schwartz I, Wormser GP, Schwartz JJ, Cooper D,Weissensee P, Gazumyan A, et al. Diagnosis of early Lymedisease by polymerase chain reaction amplification andculture of skin biopsies from erythema migrans lesions. JClin Microbiol 1992;30:3082-8.

14. Munderloh UG, Madigan JE, Dumler JS, Goodman JL,Hayes SF, Barlough JE, et al. Isolation of the equinegranulocytic ehrlichiosis agent, Ehrlichia equi, in tickcell culture. J Clin Microbiol 1996;34:664-70.

15. Persing D, Mathiesen D, Marshall W, Telford S,Spielman A, Thomford J, Conrad P. Detection ofBabesia microti by polymerase chain reaction. J ClinMicrobiol 1992;30:2097-103.

16. Boustani MR, Gelfand JA. Babesiosis. Clin Infect Dis1996;22:611-5.


Dispatches

In the last 15 years, Madagascar (population13 million) has accounted for 45% of the cases ofplague in Africa (1).

EpidemiologyPlague was brought to the island of

Madagascar in 1898 by steamboats from Indiaand has never disappeared. As a result ofvaccination campaigns, improved housing andpublic hygiene, and the discovery of streptomycinand insecticides, plague was controlled in the1950s. During the next 30 years, only 20 to 50cases per year were reported in the entirecountry. However, since 1989, the number ofsuspected cases has increased steadily (Figure 1).Since 1990, 800 to 1,500 cases of suspected plaguehave been reported, of which 150 to 230 weresmear-positive (presumptive cases) or confirmedby the isolation of Y. pestis (2,3). The populationexposed to plague in the plateaus is approxi-mately 5 million; the mean annual rate of known

human cases (except in Mahajanga) is 3 to 4 per100,000 inhabitants. The mean annual death rateis 20% of the confirmed or presumptive cases.Bubonic plague is the main clinical form of thedisease (approximately 95% of cases).

With the exception of the west coast port ofMahajanga, plague is endemic in areas more than800 m high. The major focus area is a centraltriangle, and the minor focus is a northerndiamond (Figure 2). The main plague foci (inorder of importance) are Mahajanga, the districtof Ambositra, the town of Antananarivo, and thedistricts of Fianarantsoa II, Miarinarivo, Betafo,and Soavinandriana. In the plateaus, the humanplague season is September to April, while inMahajanga it is July to November.

In 1996, 1,644 suspected plague cases werereported; biologic samples were available for1,316. A total of 173 confirmed and 56presumptive cases were officially reported. In1997, 2,127 suspected cases were reported fromJanuary to October, and more than 2,500 totalcases were expected by the end of 1997. So far,260 are confirmed and 33 are presumptive cases.The number of confirmed or presumptiveplague cases is underestimated because ofunderreporting in remote areas and the lack ofsensitivity of the bacteriologic techniques usedfor diagnosis. Our preliminary results, ob-tained by immunodiagnostic tests, such as anti-F1 antibody detection (4) and F1 antigen tests(tests supplied by the Naval Medical ResearchInstitute, Bethesda, MD), suggest a number ofplague cases two to three times higher than thatobtained by conventional methods (S. Chanteauand J. Burans, unpub. data). Whether thesignificant increase between 1996 and 1997

Plague, a Reemerging Diseasein Madagascar

Human cases of plague, which had virtually disappeared in Madagascar after the1930s, reappeared in 1990 with more than 200 confirmed or presumptive cases reportedeach year since. In the port of Mahajanga, plague has been reintroduced, and epidemicsoccur every year. In Antananarivo, the capital, the number of new cases has increased,and many rodents are infected with Yersinia pestis. Despite surveillance for thesensitivity of Y. pestis and fleas to drugs and insecticides and control measures toprevent the spread of sporadic cases, the elimination of plague has been difficultbecause the host and reservoir of the bacillus, Rattus rattus, is both a domestic and asylvatic rat.

Figure 1. Human plague, Madagascar, 1982–1996.


Dispatches

results from better notification of the casesfollowing educational campaigns or a realincrease in disease incidence is unknown.

In Madagascar, the reservoirs of Y. pestis aretwo species of rats, Rattus norvegicus and R. rattus.The sylvatic reservoir of Y. pestis is not welldocumented. The urban domestic rat R. norvegicus,introduced to Madagascar during the 20thcentury, can only be found (in larger numbersthan R. rattus) in coastal cities and Antananarivo.In contrast, R. rattus, probably imported intoMadagascar with the first human migrations, hasinvaded every ecosystem of the island. Thus, it isboth a domestic and a wild rodent. It is generallythe only type of rodent found in small towns,villages, rice fields, and grasslands. The reservoirsof Y. pestis during the interseason period are the R.rattus and R. norvegicus that have resistedinfection; therefore, it is almost impossible toeliminate plague from the island (2).

Two species of rat fleas are vectors of plaguein Madagascar. The most effective are the classic

oriental rat flea, Xenopsylla cheopis, and theendemic flea, Synopsyllus fonquerniei (2); bothare found on R. rattus, although X. cheopis isalmost exclusively on urban and indoor rodents,while S. fonquerniei is found on outdoor and wildrodents (S. Laventure and J.M. Duplantier, unpub.data). Ongoing research is examining theinterrelationships between the two species of ratsand the two species of fleas in the epidemiologiccycle of plague in Madagascar.

The reemergence of plague in Madagascarprobably reflects the general breakdown oftraditional measures of plague control. Prolongedmaintenance of the domestic rat-to-flea cycle thataccounted for the historic plague pandemics willalmost surely provide an opportunity for selectionof new traits that could make the organism evenmore virulent. In fact, three new variants ofY. pestis have recently emerged in the region ofAmbositra and Ambohimahasoa, one of the mostactive plague foci of the island. These new ribotypestend to spread to new geographic areas; whetheror not they have acquired selective advantages isbeing explored (5). Furthermore, the opportunityfor gene exchange with enteric bacteria is greatlyenhanced. It may be important that the firstnaturally occurring antibiotic-resistant strain ofY. pestis was recently isolated in Madagascar (6,7).

Plague in MahajangaIn Mahajanga (population 150,000), after two

epidemics in the beginning of the 20th century,plague remained under control between 1928 and1990 (2). In July 1991, the disease suddenlyreappeared in a shantytown near the market-place of Marolaka; of the 170 suspected cases, 41were confirmed or presumptive. After 3 yearswithout reported cases, three successive out-breaks occurred: July 1995 (8,9), July 1996, andJuly 1997. During the 1995 and 1996 outbreaks, atotal of 1,058 suspected cases were reported; only109 cases were confirmed, and 30 cases weresmear-positive. However, serologic testing, mainlyF1 antibody detection by enzyme-linkedimmunosorbent assay (ELISA) (4), allowed thediagnosis of 93 additional cases (10). Thus, themean annual rate of known human cases is 77 per100,000 inhabitants. During the ongoing epidemic,from July to October 1997, 376 suspected cases (120bacteriologically confirmed) have been reportedand tended to spread in new quarters of the town.

Every human outbreak has been preceded bya high number of rat deaths. During and after

Figure 2. Plague foci in Madagascar. The plague-endemic zones are in the gray areas and in the port ofMahajanga.


Dispatches

epidemics, all shrews captured were infectedwith many X. cheopis. Ongoing studies areexamining the role of the shrew Suncus murinusin the maintenance of plague between epidemics.

Plague in AntananarivoSince plague was introduced in the highlands

in 1921, it has never disappeared from remotevillages; however, in Antananarivo, it has beencontrolled; no human cases were reportedbetween 1953 and 1978. In 1979, the firstconfirmed case was found in one of the ancientfoci of the town (11). In the 1990s, a growingnumber of cases have been reported; 10 to 25 eachyear have been confirmed or are presumptive(Figure 3). In 1996, confirmed plague cases werefound in 17 quarters of the capital. The meanannual rate of human cases is 1.4 per 100,000inhabitants in 1995 and 1996.

In 1995, 10% of 625 rats trapped near amarketplace were infected with Y. pestis; 80%were anti-F1 seropositive (S. Chanteau, J.A.Drominy, and B. Rasoamanana, unpub. data).The monthly flea index (mean number of X. cheopisper rat) in this quarter was more than 4 year-round.Such a high index represents a serious threat,especially since most of these fleas are resistant todeltamethrin (12). Other families of insecticides,such as the carbamates, can be proposed. Since1996, the Ministry of Health and local governmenthave made serious efforts to educate the populationand improve public hygiene.

A preliminary study of rats trapped in eightother quarters of the capital in 1997 showed amean seroprevalence of 14% by ELISA (J.A.

Drominy and S. Chanteau, unpub. data). Thecirculation of Y. pestis in the rat population wassignificantly larger than it was in 1965 (4.5%) (2).

National Plague Control ProgramThe national plague control program is

financially supported by the World Bank and theFrench Ministry of Cooperation. The surveillancesystem used is based on immediate compulsorynotification of every suspected case of plague andits biologic confirmation by the Central Labora-tory. All patients with suspected cases aretreated with streptomycin, and their contacts aretreated with sulfonamides to prevent diseasespread. Insecticides are used for flea control.Every strain of Y. pestis isolated from humans,rats, and fleas is tested for drug sensitivity, andthe susceptibility of fleas to insecticides isdetermined. The isolation of the first multidrug-resistant strain of Y. pestis in 1995 (6,7) and theincreasing resistance of fleas to insecticides havecaused much concern (12).

This national program, implemented inMadagascar for several decades, has beenhampered by economic and operational difficul-ties and urgently needs to be strengthened.

Suzanne Chanteau,* Lala Ratsifasoamanana,†Bruno Rasoamanana,*† Lila Rahalison,* Jean

Randriambelosoa,† Jean Roux,* andDieudonné Rabeson†

*Institut Pasteur, Antananarivo, Madagascar; and†Ministère Santé, Antananarivo, Madagascar

References 1. World Health Organization. Human plague in 1995.

Wkly Epidemiol Rec 1997;46:344-8. 2. Brygoo ER. Epidemiologie de la peste à Madagascar.

Arch Inst Pasteur Madagascar 1966;35:7-147. 3. Blanchy S, Ranaivoson G, Rakotonjanabelo A.

Epidémiologie clinique de la peste à Madagascar. ArchInst Pasteur Madagascar 1993;60:27-34.

4. Rasoamanana B, Leroy F, Boisier P, Rasolomaharo M,Buchy P, Carniel E, Chanteau S. Field evaluation of anIgG anti-F1 ELISA test for the serodiagnosis of humanplague in Madagascar. Clin Diagn Lab Immunol1997;4:587-91.

5. Guiyoule A, Rasoamanana B, Buchreiser C, Michel P,Chanteau S, Carniel E. Recent emergence of newvariants of Yersinia pestis in Madagascar. J ClinMicrobiol 1997;35:2826-33.

6. Rasoamanana B, Leroy F, Raharimanana C, ChanteauS. Surveillance de la sensibilité aux antibiotiques dessouches de Y. pestis à Madagascar de 1989 à 1995. ArchInst Pasteur Madagascar 1995;62:108-10.

Figure 3. Human plague, Antananarivo, 1976–1996.


Dispatches

7. Galimand M, Guiyoule A, Gerbaud G, RasoamananaB, Chanteau S, Carniel E, Courvalin P. Multipleantibiotic resistance in Yersinia pestis mediated by aself-transferable plasmid. N Engl J Med1997;337:677-80.

8. Rasolomaharo M, Rasoamanana B, Andrianirina Z,Buchy P, Rakotoarimanana N, Chanteau S. Plague inMahajanga, Madagascar. Lancet 1995;346:1234.

9. Boisier P, Rasolomaharo M, Ranaivoson G,Rasoamanana B, Rakoto L, Andriamahefazafy B,Chanteau S. Urban epidemic of bubonic plague inMahajanga, Madagascar. Epidemiological aspects.Trop Med Int Health 1997;5:422-7.

10. Chanteau S, Rasoamanana B, Rasolomaharo M, LeroyF, Rahalison L, Buchy P, et al. Apport de la détectiondes anticorps anti-F1 et de l’antigénémie F1 dansl’analyse des épidémies de peste de la ville deMahajanga. Proceedings of Colloque Scientifique de laRéunion du Conseil des Directeurs des InstitutsPasteurs et Instituts Associés; 1997; Paris.

11. Coulanges P. Situation de la peste à Tananarive, de sonapparition en 1921 à sa résurgence en 1979. Arch InstPasteur Madagascar 1989;56:9-35.

12. Laventure S, Ratovonjato J, Rajaonarivelo E, RasoamananaB, Rabarison P, Chanteau S, Roux J. Résistance auxinsecticides des puces pestigènes à Madagascar etimplications pour la lutte antivectorielle. Proceedings ofthe 5ème Congrès International de Médecine Tropicale deLangue Française; 1996; Mauritius.


Dispatches

The 1993 discovery of a clinically distinctform of hantavirus disease now known ashantavirus pulmonary syndrome (HPS) high-lighted the existence of a previously unrecog-nized complex of New World hantaviruses, eachof which is associated with a specific rodent ofthe subfamily Sigmodontinae, family Muridae.The prototype of this complex is the Sin Nombre(SN) virus of the deer mouse, Peromyscusmaniculatus. SN virus, which occurs mostfrequently in the western United States andCanada, is responsible for more than 95% ofcases of HPS in North America. The case-fatality rate of HPS is nearly 50% (1).

Three viruses other than SN have beenassociated with HPS in North America. All casesassociated with viruses other than SN haveoccurred outside the range of the deer mouse.Two cases have been linked to New York virus inthe northeastern United States; the white-footedmouse, P. leucopus, was the only hantaviruscarrier in the area where the infections werecontracted. The cotton rat (Sigmodon hispidus)–associated Black Creek Canal virus is believed tohave caused a single case in southern Florida,albeit without molecular confirmation. Bayou(BAY) virus has been identified from cDNAsequences amplified from the necropsiedtissues or the blood of patients from Louisianaand eastern Texas (1-6).

Evidence implicates the rice rat, Oryzomyspalustris, as the carrier of BAY virus. In a surveyof archived rodent tissue samples obtained fromspecies indigenous to Louisiana, only O. palustrissamples were positive for hantavirus antibodies,and cDNAs of a BAY-like virus were amplified

from two of those samples (7). A few additionalBAY virus–RNA-positive rice rat samples haveappeared in subsequent studies (6). However, nosubstantial rodent collections have been con-ducted in association with BAY virus HPS cases,and no direct molecular association has beenfound between human BAY virus genomicsequences and those obtained from samples fromrodents trapped at possible sites of humaninfection. The purpose of this study was toidentify the carrier rodents associated with twohuman cases of BAY virus infections and tofurther characterize the clinical consequences ofhuman infection with BAY virus.

Diagnostic StudiesInitial serologic investigations to detect

antibodies to hantavirus used a sandwich µ-capture enzyme-linked immunosorbent assay(ELISA) as well as an immunoglobulin (Ig) GELISA, with recombinant SN virus nucleocapsid(N) protein as the target antigen (8). Confirma-tory testing with a larger array of recombinant-expressed viral N antigens was conducted byusing strip immunoblot and Western blot formats(9-11). The strip immunoblot assay incorporatesfive membrane-bound antigens (recombinant-expressed SN virus N and G1, recombinant-expressed Seoul virus N, and syntheticpeptides of SN N and G1)(11). Antibodiesreactive to these antigens are detected with ananti-human immunoglobulin heavy+light chainconjugate, and the assay thus has somesensitivity for IgM but greater sensitivity forIgG. The Western blot uses both anti-IgG andanti-IgM conjugates in separate assays (9).

Bayou Virus-Associated HantavirusPulmonary Syndrome in Eastern Texas:

Identification of the Rice Rat,Oryzomys palustris , as Reservoir Host

We describe the third known case of hantavirus pulmonary syndrome (HPS) due toBayou virus, from Jefferson County, Texas. By using molecular epidemiologic methods,we show that rice rats (Oryzomys palustris) are frequently infected with Bayou virus andthat viral RNA sequences from HPS patients are similar to those from nearby rice rats.Bayou virus is associated with O. palustris; this rodent appears to be its predominantreservoir host.


Dispatches

Western blot studies used a panel of affinity-purified, full-length N antigens produced inEscherichia coli in fusion with the phage T7 gene10 protein in the pET23b vector (10). Recombi-nant N antigens were purified over metalchelation affinity columns through a polyhistidinetag on the C-terminus of each protein. Therecombinant proteins were produced in anisogenic background and loaded at 500ng/lanebefore sodium dodecyl sulfate-polyacrylamide gelelectrophoresis separation and electrophoretictransfer. The antigens were derived from thefollowing hantaviruses: SN (3H226); BAY (OP-LA-475); Rio Mamoré (OM-556), Muleshoe (SH-Tx-339), Puumala (P360), and Seoul (80/39).

Rodent Collection and ProcessingSince the two BAY virus-HPS cases from

Texas occurred in neighboring cities, this studyincludes rodent samples collected in theinvestigation of both case P/Tx (6) and the casedescribed here, T/Tx. Sherman live-traps wereused to collect rodents in Jefferson County andneighboring Orange County. Heart bloodsamples were screened by a recombinant SNvirus N antigen ELISA (8).

Viral Sequencing and PhylogeneticAnalyses

Human peripheral blood mononuclear cells (2x 106) and rodent lung, kidney, and spleensamples (~200 mg total) were used to make RNA(6,10). A standard set of partially nested primersin the viral S segment was used in reversetranscription-polymerase chain reaction (RT-PCR) analyses of the RNA (6,10). This procedureproduces a 442-nucleotide (nt) amplificationproduct; 397 nt of that sequence is internal to theprimers. The PCR products were subjected todirect DNA sequencing with an ABI 377automatic sequencer. Phylogenetic trees wereconstructed from the 397 nt of informativesequence by using maximum parsimony withPAUP 3.1 software (12).

A Third Case of HPS Due to BAY VirusPatient T/Tx is a 54-year-old African-

American man from Jefferson County, Texas. Aheavy tobacco user, he has a more than 100 pack-years* smoking history and chronic shortness ofbreath. Approximately 2 weeks before admission,his spouse noted that he had increased fatigue

and somnolence. By August 20, 1996, hissymptoms worsened, and he sought medical care.He visited a local emergency room withcomplaints of increasing shortness of breath andlow-grade fever for 2 days. Hospital personnelbelieved that he had probable chronic obstructivepulmonary disease and chronic bronchitis.Results of a physical examination of the patientat that time were unremarkable, and a chestradiograph was interpreted as hyperinflatedbut without infiltrates (Figure 1A). He wasafebrile (37°C) in the emergency room.Laboratory analysis showed a white blood cellcount of 4,080/µl and a platelet count of122,000/µl. He was given amoxicillin andbronchodilators and was discharged.

On August 21, the patient returned to theemergency room with persistent and worseningdyspnea. He also had flulike symptoms, includingmyalgias, and had had a temperature of 39.4°C.Physical examination revealed bilateral rhonchi.He had a blood pressure of 90/60, pulse of 96/min,respiratory rate of 16/min, and an oral temperatureof 37°C. Laboratory analysis noted an increase inhis white blood cell count to 7,700/µl with a leftshift and a decrease in platelet count to 65,000/µl.The patient was admitted to the hospital with apreliminary diagnosis of chronic obstructivepulmonary disease exacerbation and possiblepneumonia. He was given broad-spectrumantibiotics, methylprednisolone, aminophylline,and ipatronium nebulizer treatments.

The patient’s condition deteriorated rapidlyover the next 24 hours, and he was severely shortof breath by the second hospital day. He was alsofebrile and was transferred to the intensive careunit to manage his worsening shortness ofbreath. His oxygen demand increased, and on thethird hospital day, the patient required endotra-cheal intubation. A physical examinationfound bilateral expiratory wheezing and bibasilarrales. A chest radiograph showed increasingbilateral interstitial and alveolar infiltrates. Bythe fourth hospital day, the patient required100% oxygen and positive end-expiratory pres-sure support. A chest radiograph again showedinterstitial and alveolar infiltrates with peribron-chial cuffing and pleural effusion. His fluid intakewas restricted to manage his pulmonary edema;he briefly required dopamine for hypotension.Blood and urine cultures remained negative.Sputum Gram stain showed a moderate number

*Pack-year = no. packs/day x no. years


Dispatches

of white blood cells and few budding yeast;culture showed no pathogens. Serologic tests formycoplasma and Legionella were negative.

By the fifth hospital day, the patient’s conditionhad begun to improve. Over the next 5 days thefever diminished, the pulmonary edema im-proved, and requirement for supplemental oxygendecreased. The endotracheal tube was removed onthe 10th day of hospitalization; by the 14th day,the patient was discharged—ambulatory, afebrile,and no longer requiring supplemental oxygen.

The hematologic laboratory findings weregenerally most abnormal when the patient wasmost acutely ill clinically; the peak serumchemistry abnormalities trailed the clinical illnessby several days. The maximum white blood cellcount was 31,700/µl on August 23, with adifferential count of 7% lymphocytes (includingthe characteristic plasmacytoid forms; [13]), 44%mature neutrophils, 35% band neutrophils, 3%metamyelocytes, and 2% myelocytes. The plateletcount reached a nadir of 19,000/µl on that day,and the hemoglobin reached a maximum of 17.2g/dL on August 22. The serum creatinine was 0.6mg/dL on August 24 but peaked at 1.9 mg/dL onAugust 25. This mild azotemia was associated

with proteinuria (300 mg/dL in a spot sampleobtained on August 22). The serum enzymescreatine kinase and lactic dehydrogenase peakedon August 27 at 917 units/L and 933 units/L,respectively. Fractionation of the creatine kinaseand lactic dehydrogenase isoenzyme forms werenot supportive of myocardial injury as a cause fortheir elevation. Serum levels of the livertransaminases were modestly elevated.

Environmental Assessment and RodentCollection

The patient worked as a laborer at a railroadconstruction company. For 2 months before hisillness, he replaced old rail lines at threeindustrial plants (Beaumont, Neches River, andOrange) in Orange and Jefferson Counties andcleaned the yard and garage at work headquar-ters. He manually removed and replaced tracksand ties serving warehouses, transport pipes,and scrap metal piles.

The Beaumont and Orange work sites werechemical plants. All rail work occurred within theconfines of the mowed plant complex. Some of thetrack sites were next to drainage ditches andcanals; others coursed by warehouses reported by

A B

Figure 1. Chest radiographs of patient T/Tx at hospital admission on August 20, 1996 (Panel A), and during aperiod of increasing respiratory distress on August 22, 1996 (Panel B). Note the diffuse interstitial infiltrate andperibronchial cuffing that developed over that interval.


Dispatches

plant employees to have rats. The patient had notseen any rodents in the months before becomingill, but he remembered rodent droppings at awooden crew trailer at the Beaumont plant. Allrodents trapped at these plants tested negativefor hantavirus antibodies.

The Neches River work site was a recyclingplant situated for commerce by rail and ship. Therails were within several meters of permanentbayous and swamps, piles of scrap metal, and theNeches River. The vegetation near the water wasdense with trees and undergrowth. Truck andtrain traffic at the plant created a substantialdust problem. The site contained stray dogs andcats as well as a variety of wildlife, includingsnakes and alligators. Although trapped heavily,few rodents were collected from near the tracksand adjacent swamps. All rodents trapped at thisfacility were negative for hantavirus antibodies.

Although the patient lived in an older pier-and-beam home, he and his wife saw no evidenceof rodent infestation. A Texas Department ofHealth investigator verified the absence ofrodents and rodent excreta. Although traps wereset, no rodents were collected at the residence.

During the 2 months before he became ill, thepatient visited a casino boat in Louisiana andfished three times from the Pleasure Island jettyin southern Jefferson County. The jetty had largerock boulders at the water’s edge and an asphaltroad along its length. Sand and dense 6-foot highgrasses covered the remainder of the jetty.Household trash andfishing debris werescattered throughoutthe area. Most ro-dents collected by theTexas Department ofHealth investigationteam during this in-vestigation were fromthis jetty; and mostwere O. palustris, in-cluding three serop-ositive specimens(Area 1, Figure 2).

Other sites werechosen between thepatient’s work sitesand home because theirhabitats were likely tosupport rodents. Rice

Figure 2. Location of rodents trapped during investigations related to patient P/Tx(March 1996; [6]) and patient T/Tx (October 1996; this report). The locations in whichrodents were collected are broadly classified into four areas, designated 1 through 4.

rat habitats (permanent water and grasses)were among those targeted, because the firstJefferson County hantavirus investigation docu-mented the Bayou strain of hantavirus in O.palustris (6). The Table and Figure 2 show thelocation and species of each rodent collected andthe number of seropositive specimens in each offour targeted areas. Four hantavirus antibody-positive rodents were trapped at these miscella-neous locations during the investigation of thecurrent HPS case. Two of the seropositivesamples were from Area 1 of Figure 2 (not thejetty), and two were in Area 2. Area 2 yielded twoseropositive rodents during the first Texas BAYvirus case investigation (6).

Diagnostic StudiesThe IgG ELISA assay for SN virus antibodies

in patient T/Tx’s serum was negative, but the IgMtest was positive at a 1:6400 dilution (8). The stripimmunoblot assay showed antibody reactivitiestypically seen with HPS caused by a hantavirusother than SN virus, with 2+ reactivity to theimmunodominant SN virus N peptide and 4+reactivity to the full-length recombinant N protein(11). There was no reactivity to the SN virus G1antigen, either in peptide or recombinant form, or toSeoul virus recombinant N protein (data notshown). Antibody reactivity to the SN virus G1antigen is specific for infection with SN virus.

Western blot analysis of patient T/Tx’s serumshowed IgG antibodies to hantavirus N proteins


Dispatches

(Figure 3, Panel A). Although the most intensestaining was observed with the homologous (BAYvirus) antigen, varying cross-reactivity waspresent against the N antigens of Muleshoe, SN,Rio Mamoré, and Puumala viruses but not ofSeoul virus. A similar pattern was observed whenanother blot membrane was probed with an anti-human IgM conjugate, except that reactivitieswere somewhat more intense (data not shown).Serum samples of seropositive rice rats and deermice showed similar cross-reactivities, althoughthe intensity of reactivity to a particular Nantigen was related to the sequence similaritybetween the membrane-bound antigen and thevirus against which the antibodies were directed(Figure 3, Panels B and C).

RT-PCR analyses and sequencing of thecDNA product of the patient’s blood sampleidentified definitively BAY virus as the etiologicagent. When the 442 nt viral S segmentamplification product from patient T/Tx wascompared with the homologous sequence of otherhantaviruses, it was closely similar to previouslydescribed BAY viruses from Louisiana and Texas(Figure 4). Nine viral sequences from seropositiveO. palustris and one from a seropositive S. hispidus

Table. Rodents collected in association with two cases ofBayou virus-hantavirus pulmonary syndrome in Jeffersonand Orange Counties, Texas, March and October 1996

No. specimens collected(no. seropositive) No. PCRb

Area Area Area Area testedSpecies 1a 2 3 4 Tot. (no. pos.)Baiomys 0 0 1 3 4 0 taylori (0) (0) (0)Mus 30 21 12 26 88 1 musculus (0) (0) (1) (0) (1) (0)Ochrotomys 0 0 0 1 1 0 nuttali (0) (0)Oryzomys 45 20 3 8 76 15 palustris (7) (5) (0) (3) (15) (14)Peromyscus 0 3 0 0 3 0 leucopus (0) (0)Rattus 1 1 0 1 3 0 norvegicus (0) (0) (0) (0)Rattus 21 8 0 1 30 0 rattus (0) (0) (0) (0)Reithrodontomys 0 7 0 7 14 0 fulvescens (0) (0) (0)Reithrodontomys 1 0 0 2 3 0 humulis (0) (0) (0)Sigmodon 95 5 11 4 114 1 hispidus (0) (0) (0) (1) (1) (1)aRefer to Figure 2 for map areas.bPCR=polymerase chain reaction.

Figure 3. Western blot assay for detecting IgGantibodies in patient and rodent blood samples. A1:500 dilution of serum was used to probe Westernblots containing equimolar amounts of recombinant-expressed N antigens of various hantaviruses: 1)Bayou; 2) Muleshoe; 3) Puumala; 4) Rio Mamoré; 5)Seoul; and 6) Sin Nombre. Serum samples are directedagainst specific hantaviruses as verified by reversetranscription-PCR and sequence analyses: (A) patientT/Tx, Bayou virus; (B) Bayou virus-seropositiveOryzomys palustris specimen #505, collected inJefferson County, Texas; (C) a Sin Nombre–seropositive deer mouse, collected in California; (D), anegative control (hantavirus-seronegative) deer mouse.Antibodies bound to the solid-phase antigens weredetected with alkaline phosphatase-conjugated anti-human IgG (A) or with anti-Peromyscus leucopus IgG(B-D). The arrow indicates the migration of the full-length T7-viral N antigen, ~55kDa.


Dispatches

from Jefferson County, Texas, and Cameron andTerrebonne Parishes, Louisiana, were comparedwith those of patients T/Tx and P/Tx and withthat of a 1993 case-patient from northernLouisiana (Case-LA-93, Figure 4).

The sequences of all the Jefferson Countyrodents and patient T/Tx were closely related toeach other, differing by 1 to 10 nt (0.25% to 2.5%of 397 evaluable residues) in pairwise comparisons.

Four rodent-derived viral sequences differed fromthat of the patient T/Tx virus by 4 nt (1%), and allwere from rodents collected in Jefferson County.

Except for a brief visit to a Louisiana casino,patient T/Tx reported that he had not traveledoutside east Texas in the 2 months before hisillness. The generally close genetic similarityamong the viral sequences from east Texassupports the hypothesis that patient T/Txbecame infected in that region (14). Potentialexposure sites, as defined by his travel history,were too numerous to allow us to identify a moreprecise site of infection. No rodent sequence wascompletely identical to that of patient T/Tx.

The sequence obtained from patient P/Tx (6),who was thought to have been infected inJefferson County, was not closely aligned withany eastern Texas rodent sequence or with that ofpatient T/Tx. The closest of the 11 eastern Texasviral sequences differed from that of patient P/Txat 13 residues (3.3%). One sequence available fromprevious studies was far closer to that of patientP/Tx, differing at only four residues (1%). Thissequence was obtained from a rice rat (OP-LA-5)collected in Cameron Parish, Louisiana, approxi-mately 30 km from the former residence of patientP/Tx. Although patient P/Tx initially reported notravel back to his former residence (where hismother still lived) in the 6 weeks before his illness(6), he later was not certain about his travelhistory in the weeks before his illness, statingthat he might have visited his mother inLouisiana in the weeks before his illness.

In previous studies, O. palustris wastentatively implicated as the predominant rodentreservoir for BAY virus, but the sample sizeswere small, and specific cases of HPS were notlinked to the occurrence of BAY virus-infectedrodents at the presumed sites of infection (6,7).For this study, analysis of a substantial collectionof rodents collected in the vicinity of sites ofhuman infection showed that O. palustris is thespecies with the highest hantavirus seroprevalence.Furthermore, the hantavirus genotypes incirculation in the eastern Texas/western Louisi-ana area were related to those of two human case-patients from the area, and all but one of thoserodent-derived sequences came from O. palustris.

Rice rats are found in wetlands and marshesfrom Texas throughout the southeastern UnitedStates, extending north as far as New Jersey onthe eastern seaboard. Like all carriers of virusesassociated with HPS, O. palustris is a member of

Figure 4. Unweighted maximum parsimony treeproduced by PAUP 3.1 software comparing a 397 ntportion of the hantavirus S genomic segments(residues 207 to 603) of patients and rodents infectedwith Bayou virus. A selection of other prototypicalhantavirus sequences, each of which was comparedantigenically in Figure 2, was included for comparison.Other hantaviruses are abbreviated as follows:RM=Rio Mamoré; MULE=Muleshoe; SN= Sin Nombre;PUU=Puumala; and SEO=Seoul. Human-derivedBayou virus sequences are indicated by Case, androdent sequences by OP (Oryzomys palustris) or SH(Sigmodon hispidus). TX=Texas, LA=Louisiana.


Dispatches

the subfamily Sigmodontinae, family Muridae.Although only a single species exists in the UnitedStates, the tribe Oryzomyini has scores of species inLatin America, where HPS is increasinglyrecognized as an important zoonotic disease.Viruses similar to BAY virus have beenrecognized recently in members of the Oryzomyinifrom Bolivia and Argentina, and in some casesthese rodent-borne viruses have been linked toHPS in humans (1,15).

The possibility that the genetically distinctclade of viruses associated with oryzomine andSigmodon rodents produces a disease that isqualitatively different from SN virus-associatedHPS (16) has been raised repeatedly as new caseshave been identified (5,6,17). Although patientT/Tx had a relatively mild course of HPS, hisserum creatinine, urine protein concentration, andcreatine kinase were nevertheless slightlyelevated. Both the elevation of creatine kinaseand chemical evidence of myositis have beenobserved in HPS caused by viruses of the oryzomineand Sigmodon clade, but much less commonly inSN virus infection (1). Further studies are neededto determine the basis for the variant clinicalsymptoms of HPS in these patients.

AcknowledgmentsWe thank G. Haigh, McNeese State College, for donating therice rat specimen OP-LA-5; J. Griffith for technicalassistance with sequencing; and W. George, St. ElizabethHospital, for first recognizing HPS as a possible cause ofpatient T/Tx’s illness. This study was supported by DHHSgrant RO1 AI36336 (to B.H.).

Norah Torrez-Martinez,* MausumiBharadwaj,* Diane Goade,* John Delury,†

Peggy Moran,* Bradley Hicks,† Beverlee Nix,‡James L. Davis,¶ and Brian Hjelle*

*University of New Mexico School of Medicine,Albuquerque, New Mexico, USA; †Texas Departmentof Health, Austin, Texas, USA; ‡Texas Department

of Health, Houston, Texas, USA; and ¶St.Elizabeth’s Hospital, Beaumont, Texas, USA.

References 1. Schmaljohn C, Hjelle B. Hantaviruses: a global disease

problem. Emerg Infect Dis 1997;3:95-104. 2. Song J-W, Baek L-J, Gajdusek DC, Yanagihara R,

Gavrilovskaya I, Luft BJ, et al. Isolation of pathogenichantavirus from white footed mouse (Peromyscusleucopus). Lancet 1994;344:1637.

3. Rollin PE, Ksiazek TG, Elliott LH, Ravkov EV, MartinML, Morzunov S, et al. Isolation of Black Creek Canalvirus, a new hantavirus from Sigmodon hispidus inFlorida. J Med Virol 1995;46:35-9.

4. Morzunov SP, Feldmann H, Spiropoulou CF,Semenova VA, Rollin PE, Ksiazek TG, et al. A newlyrecognized virus associated with a fatal case ofhantavirus pulmonary syndrome in Louisiana. J Virol1995;69:1980-3.

5. Khan AS, Spiropoulou CF, Morzunov S, Zaki SR, KohnMA, Nawas SR, et al. Fatal illness associated with a newhantavirus in Louisiana. J Med Virol 1995;46:281-6.

6. Hjelle B, Goade D, Torrez-Martinez N, Lang-WilliamsM, Kim J, Harris RL, Rawlings JA. Hantaviruspulmonary syndrome, renal insufficiency and myositisassociated with infection by Bayou hantavirus. ClinInfect Dis 1996;23:495-500.

7. Torrez-Martinez N, Hjelle B. Enzootic of Bayouhantavirus in rice rats (Oryzomys palustris) in 1983.Lancet 1995;346:780-1.

8. Ksiazek TG, Peters CJ, Rollin PE, Zaki S, Nichol S,Spiropoulou C, et al. Identification of a new NorthAmerican hantavirus that causes acute pulmonaryinsufficiency. Am J Trop Med Hyg 1995;52:117-23.

9. Jenison S, Yamada T, Morris C, Anderson B, Torrez-Martinez N, Keller N, Hjelle B. Characterization ofhuman antibody responses to Four Corners hantavirusinfections among patients with hantavirus pulmonarysyndrome. J Virol 1994;68:3000-6.

10. Rawlings JA, Torrez-Martinez N, Neill SU, Moore GM,Hicks BN, Pichuantes S, et al. Cocirculation of multiplehantaviruses in Texas, with characterization of the Sgenome of a previously-undescribed virus of cotton rats(Sigmodon hispidus). Am J Trop Med Hyg 1996;55:672-9.

11. Hjelle B, Jenison S, Torrez-Martinez N, Herring B,Quan S, Polito A, et al. Rapid and specific detection ofSin Nombre virus antibodies in patients withhantavirus pulmonary syndrome by a strip immunoblotassay suitable for field diagnosis. J Clin Microbiol1997;35:600-8.

12. Swofford DL. PAUP: phylogenetic analysis usingparsimony [computer program]. Version 3.1.1.Champaign (IL): Illinois Natural History Survey;1991.

13. Nolte KB, Feddersen RM, Foucar K, Zaki SR, KosterFT, Madar D, et al. Hantavirus pulmonary syndromein the United States: a pathological description of adisease caused by a new agent. Hum Pathol1995;26:110-20.

14. Hjelle B, Torrez-Martinez N, Koster FT, Jay M, AscherMS, Brown T, et al. Epidemiologic linkage of rodentand human hantavirus genomic sequences in caseinvestigations of hantavirus pulmonary syndrome. JInfect Dis 1996;173:781-6.

15. Bharadwaj M, Botten J, Torrez-Martinez N, Hjelle B.Rio Mamoré virus: genetic characteristics of a newly-recognized hantavirus of the pygmy rice ratOligoryzomys microtis from Bolivia. Am J Trop MedHyg. In press 1997.

16. Duchin JS, Koster FT, Peters CJ, Simpson GL, TempestB, Zaki SR, et al. Hantavirus pulmonary syndrome: aclinical description of 17 patients with a newly recognizeddisease. N Engl J Med 1994;330:949-55.

17. Khan AS, Gaviria JM, Rollin PE, Hlady WG, KsiazekTG, Armstrong LR, et al. Hantavirus pulmonarysyndrome in Florida: association with the newlyidentified Black Creek Canal virus. Am J Med1996;100:46-8.


Dispatches

The emergence of drug-resistant Streptococ-cus pneumoniae underscores the need for timely,local, population-based surveillance of antimicro-bial resistance. The prevalence of resistance inU.S. communities varies widely, with 2% to 53%of S. pneumoniae isolates found to have reducedsusceptibility to penicillin (1-4). The Centers forDisease Control and Prevention recommendsthat empiric antibiotic therapy for pneumococcalinfections be based upon local susceptibilitypatterns (2,5). However, few communities trackdrug-resistant S. pneumoniae.

The SurveyTo estimate the prevalence of drug-resistant

S. pneumoniae in New York City, we surveyedhospital-based clinical microbiology laboratoriesfrom 1993 to 1995. A standardized questionnairewas mailed annually to each laboratory, andthose that did not respond were contacted bytelephone or were visited. To evaluatecompliance with S. pneumoniae penicillinsusceptibility testing guidelines established bythe National Committee for Clinical Labora-tory Standards (NCCLS) (6), we asked aboutcriteria for selecting specimens and techniquesfor oxacillin disk diffusion screening anddetermination of penicillin MICs.

To determine the prevalence of penicillinresistance, we asked for the number of S.pneumoniae isolates identified during the year,the number tested for susceptibility to penicillin,and the number found to be possibly resistant bythe oxacillin disk diffusion test and penicillin-intermediate or -resistant by MIC testing. Wealso asked that information be providedseparately for isolates from normally sterile sites(e.g., blood, cerebrospinal fluid) and from

nonsterile sites (e.g., sputum, nasopharyngealswab). In 1995, we added questions regarding theMIC test results for extended-spectrum cepha-losporins (ESCs), including the number ofpenicillin-nonsusceptible isolates that were alsononsusceptible to an ESC. No individual patientinformation was obtained. A report summariz-ing the results of the survey and describingNCCLS guidelines was mailed annually tomicrobiology laboratories, hospital infectioncontrol departments, and local infectiousdisease physicians and pediatricians.

AnalysisOf 67 hospital-based clinical microbiology

laboratories in New York City, 100% completedthe survey in 1993, 98% in 1994, and 100% in1995. Overall, more than 5,000 S. pneumoniaeisolates were reported annually.

Data were analyzed by using EpiInfo Version6.0 (CDC, Atlanta, GA, USA). Drug-susceptibilityresults are presented for laboratories thatconformed with NCCLS guidelines and providedcomplete data on all S. pneumoniae isolatesidentified (Table 1).

Susceptibility CriteriaThe NCCLS recommends routine screening

by the oxacillin disk diffusion test of clinicallyimportant S. pneumoniae isolates for susceptibil-ity to penicillin. Isolates with a zone size ≤ 19 mm,or any isolate from the blood or cerebrospinalfluid, should be tested with an approved MICmethod such as broth dilution or antibiotic gradientstrips (e.g., E-test). Isolates whose penicillinMICs are either intermediate (MIC ≥ 0.1 and ≤ 1µg/ml) or resistant (MIC ≥ 2 µg/ml) should alsohave MICs determined for susceptibility to an

Laboratory Survey of Drug-ResistantStreptococcus pneumoniae in

New York City, 1993–1995

Wide geographic variation in the prevalence of drug-resistant Streptococcuspneumoniae demonstrates the importance of tracking antimicrobial resistance locally. Thissurvey of hospital microbiology laboratories in New York City found that penicillinresistance (MIC ≥ 2.0 µg/ml) increased from 1.5% of S. pneumoniae isolates in 1993 to6.3% in 1995 and that in 1995, one-third of isolates nonsusceptible to penicillin (MIC ≥ 0.1µg/ml) were also nonsusceptible to an extended-spectrum cephalosporin (MIC ≥ 1 µg/ml).


Dispatches

ESC such as cefotaxime or ceftriaxone (ESC-intermediate MIC = 1 µg/ml; ESC-resistantMIC ≥ 2 µg/ml) (6). We will use the term“nonsusceptible” to refer to both intermediateand resistant isolates.

FindingsThe proportion of laboratories conforming

with NCCLS guidelines for penicillin susceptibil-ity testing of S. pneumoniae increased from 22%in 1993 to 69% in 1995. This was due to anincrease in the number of laboratories thatscreened all isolates, a sharp decrease in the useof automated MIC tests, and a fourfold rise in theuse of antibiotic gradient strips for determiningMICs (Table 2). Overall, the proportion of isolateswith oxacillin disk diffusion test zone size ≤ 19mm, suggesting possible resistance, increasedfrom 8.5% of isolates in 1993 to 20.2% in 1995

(Table 1). MIC test results showed that 5.7% ofisolates were penicillin-intermediate and 1.5%penicillin-resistant in 1993, compared with 8.8%and 6.3%, respectively, in 1995. The prevalence ofresistant organisms increased among isolatesfrom both sterile and nonsterile sites and wassomewhat higher among nonsterile-site isolatesthan among sterile-site isolates in 1995 (6.9% vs.4.9%, Chi-square p = 0.03).

Seven of the laboratories followed NCCLS-recommended methods and provided completepenicillin susceptibility results for both 1993 and1995. These seven reported that among 779isolates in 1993, 25 (3.2%) were intermediate andsix (0.8%) were resistant to penicillin. Among 896isolates in 1995, 83 (9.2%) were intermediate, and47 (5.2%) were resistant.

The prevalence of penicillin resistance variedbetween laboratories and by geographic area. At

Table 1. Penicillin resistance among Streptococcuspneumoniae isolates at New York City hospital laboratories

1993 1994 1995No. of isolates No. (%) No. (%) No. (%)Screened with oxacillin diska 3,227 4,133 4,912

zone size ≤ 19 mmb 273 (9) 549 (13) 995 (20)Screened and confirmed with approved MICc 1,229 2,491 3,535

zone size ≤ 19 mm 154 (13) 350 (14) 704 (20)Id 70 (6) 209 (8) 310 (9)Rd 19 (2) 115 (5) 222 (6)

Sterile-site isolates screened and confirmed with MICe 462 649 1,321

I 16 (3) 66 (10) 83 (6)R 5 (1) 20 (3) 65 (5)

Nonsterile isolates screened and confirmed with MICf 456 662 1,596

I 12 (3) 83 (13) 200 (13)R 5 (1) 31 (5) 110 (7)

aNumber of laboratories reporting oxacillin disk diffusion testresults in 1993, 1994, 1995 was 33, 40, 51, respectively.bOxacillin disk diffusion test zone size.cApproved penicillin MIC tests include antibiotic gradientstrips and broth dilution or micro-dilution using Mueller-Hinton broth supplemented with 2-5% lysed horse blood.Number of laboratories reporting MIC test results in 1993,1994, 1995 was 10, 22, 35, respectively.dI=Penicillin-intermediate (MIC ≥ 0.1 and ≤ 1.0 µg/ml);R=Penicillin-resistant (MIC ≥ 2.0 µg/ml).eNumber of laboratories reporting sterile site isolate resultsin 1993, 1994, 1995 was 8, 12, 32, respectively.fNumber of laboratories reporting nonsterile site isolate resultsin 1993, 1994, 1995 was 5, 9, 28, respectively.

Table 2. Penicillin susceptibility testing protocols forStrepococcus pneumoniae in New York City hospitallaboratories

1993 1994 1995No. of laboratoriesa No. (%) No. (%) No. (%)Screened with oxacillin disk 55 (82) 63 (97) 65 (97)

Screened all isolates 38 (57) 56 (86) 59 (88)Screened only sterile 9 (13) 6 (9) 6 (9)Other 8 (12) 1 (2) 0

Performed penicillin 39 (58) 51 (78) 56 (84) MIC testMIC-tested

All isolates 10 (15) 9 (14) 7 (10)Oxacillin- resistant only 18 (27) 39 (60) 42 (63)Sterile isolates only 5 (7) 1 (2) 0 (0)By request only 4 (6) 1 (2) 5 (7)Other 2 (3) 1 (2) 2 (3)

Penicillin MIC techniqueb

Antibiotic gradient stripc 11(16) 41 (63) 50 (75)Automated testsc 21 (31) 9 (14) 2 (3)Broth dilution 8 (12) 11 (17) 5 (7)Unknown 2 (3) 3 (5) 0

Conformed with NCCLS guidelinesd 15 (22) 38 (58) 46 (69)aTotal number of laboratories responding to survey in 1993,1994, 1995 was 67, 65, 67, respectively.b13 laboratories used more than one MIC technique.cAntibiotic gradient strip: e.g., E-test. Automated tests: e.g.,Microscan, Vitek.dNCCLS=National Committee for Clinical LaboratoryStandards. Screened all S. pneumoniae isolates with oxacillindisk diffusion test and confirmed resistance with approvedpenicillin MIC test.


Dispatches

35 laboratories in 1995, 0% to 80% (median13.1%) of isolates were nonsusceptible topenicillin. Laboratories in the New York Cityborough of Manhattan reported that 18% ofisolates were nonsusceptible to penicillin,compared with 16% in Queens, 14% in the Bronx,and 10% in Brooklyn (MIC data from StatenIsland were not available).

In 1995, 18 laboratories reported that of 275 S.pneumoniae isolates nonsusceptible to penicillin,90 (33%) were also nonsusceptible to an ESC.

LimitationsThe survey design had several limitations.

First, levels of expertise varied among laborato-ries regarding antibiotic susceptibility testing ofpneumococci. Laboratory variation may partlyaccount for the wide range in the proportion ofisolates nonsusceptible to penicillin observed atindividual laboratories, which could in turninfluence citywide estimates. We did not collectand test isolates to confirm laboratory results.Second, the number of laboratories included inour results increased each year as morelaboratories adopted NCCLS methods andprovided complete data. This improved theaccuracy of the survey but made interpretingtrends difficult. Third, citywide estimatesinevitably mask important differences in the riskfor drug-resistant S. pneumoniae infections inspecific subpopulations; for example, day-careattendance and prior antibiotic use have beenassociated with drug-resistant pneumococcalinfections in children (1,7). Finally, sinceinformation was collected on isolates, rather thanindividual patients or infections, actual diseaseincidence could not be calculated.

Conclusions and RecommendationsOur results document a marked improve-

ment in penicillin susceptibility testing protocolsfor S. pneumoniae during this period. Because ofan increase in the proportion of isolates testedand widespread adoption of antibiotic gradientstrips for MIC testing, more than two-thirds oflaboratories conformed with NCCLS guidelines in1995, compared with fewer than one-fourth in 1993.

The survey demonstrates that drug-resistantpneumococci are prevalent and may be increasingin New York City. Penicillin resistance increasedfourfold between 1993 and 1995, and data from1996 indicate that it has remained at this level or

continued to increase. Nine percent of bloodisolates in 1996 were intermediate and 6%resistant to penicillin (New York City Depart-ment of Health, unpub. data), compared with 6%and 5%, respectively, among sterile-site isolatesin our survey for 1995. The proportion of S.pneumoniae isolates resistant to penicillin inNew York City was similar to U.S. nationalaverages during this period (2-4). Our findingthat 33% of penicillin-nonsusceptible S.pneumoniae isolates were resistant to ESCs isalso similar to the 29% (4) observed nationally.

On the basis of these results, we recom-mended that New York City clinicians considercarefully the possibility of penicillin and ESCresistance when treating suspected S. pneumoniaeinfections. Although the response to therapy mayvary by route of administration, site of infection,and age and immune status of the patient,treatment failure is increasingly common andcan have serious consequences (8,9). Educationalefforts are needed to promote appropriate use ofantibiotics and encourage use of the 23-valentpneumococcal vaccine in populations at in-creased risk for pneumococcal disease (10).

This laboratory survey was a relatively low-cost method of estimating the prevalence of drug-resistant pneumococci in New York City. It mayserve as a model in areas where clinicallaboratories routinely perform drug-susceptibil-ity testing but the resources to collect or testisolates centrally for surveillance are limited.Despite the survey’s limitations, the estimatesprovided may be sufficient to guide clinicians inselecting appropriate empiric therapy for sus-pected pneumococcal infections. Considerableeffort was spent each year disseminating thesedata to laboratories, hospital infection controldepartments, infectious disease physicians,pediatricians, and other primary care physicians.

Our ability to expand this survey approachbeyond S. pneumoniae, penicillin, and ESCs islimited by constraints on hospital laboratory staff.As laboratories computerize and standard formatsare developed for the electronic transmission oflaboratory test results, collection and dissemina-tion of susceptibility data will be possible on a widerrange of pathogens and antimicrobial drugs.

AcknowledgmentsThe authors thank Alexandra Dow for her contributions

to the 1993 survey and laboratory staff throughout NewYork City, whose time and effort made this project possible.


Dispatches

Richard Heffernan, Kelly Henning, AnneLabowitz, Annette Hjelte, and Marcelle Layton

New York City Department of Health,New York, New York, USA

References 1. Hoffman J, Cetron MS, Farley MM, Baugham WS,

Facklam RR, Elliott JA, et al. The prevalence of drug-resistant Streptococcus pneumoniae in Atlanta. N EnglJ Med 1995;333:481-6.

2. Breiman RF, Butler JC, Tenover FC, Elliott JA, FacklamRR. Emergence of drug-resistant pneumococcal infectionsin the United States. JAMA 1994;271:1831-5.

3. Doern GV, Brueggemann A, Holley HP, Rauch AM.Antimicrobial resistance of Streptococcus pneumoniaerecovered from outpatients in the United States duringthe winter months of 1994 to 1995: results of a 30-center national surveillance study. Antimicrob AgentsChemother 1996;40:1208-13.

4. Butler JC, Hoffmann J, Cetron MS, Elliott JA, FacklamRR, Breiman RF, et al. The continued emergence of drug-resistant Streptococcus pneumoniae in the United States:an update from the Centers for Disease Control andPrevention’s pneumococcal sentinel surveillance system.J Infect Dis 1996:174:986-93.

5. Cetron MS, Jernigan DB, Breiman RF, The DRSPWorking Group. Action plan for drug-resistantStreptococcus pneumoniae. Emerg Infect Dis 1995;1:64-5.

6. National Committee for Clinical Laboratory Standards.Methods for dilution antimicrobial susceptibility testsfor bacteria that grow aerobically. Approved standard.M7-A4. Villanova (PA): The Committee; 1997.

7. Arnold KE, Leggiadro RJ, Breiman RF, Lipman HB,Schwartz B, Appleton MA, et al. Risk factors forcarriage of drug-resistant Streptococcus pneumoniaeamong children in Memphis, Tennessee. J Pediatr1996;128:757-64.

8. Appelbaum PC. Epidemiology and in vitro susceptibilityof drug-resistant Streptococcus pneumoniae. PediatrInfect Dis J 1996;15:932-9.

9. Chesney PJ, Wilimas JA, Presbury G, Abbasi S,Leggiadro RJ, Davis Y, et al. Penicillin- andcephalosporin-resistant strains of Streptococcuspneumoniae causing sepsis and meningitis in childrenwith sickle cell disease. J Pediatr 1995;127:526-32.

10. Munford RS, Murphy TB. Antimicrobial resistance inStreptococcus pneumoniae: can immunization preventits spread? J Investig Med 1994;42:613-21.


Dispatches

B-Virus in Nonhuman PrimatesCercopithecine herpesvirus 1 (Herpesvirus

simiae or B-virus) frequently infects Old Worldprimates of the genus Macaca. Of at least 19species of macaques, rhesus, Japanese, cynomol-gus, pig-tailed, and stump-tailed macaques arethe species most commonly used in biomedicalresearch (1). Seroprevalence of neutralizingantibodies to B-virus in captive adult macaquepopulations is 73% to 100% (1-3). LikeHerpesvirus simplex virus infection in humans,B-virus infection in monkeys is characterized bylifelong infection with intermittent reactivationand shedding of the virus in saliva or genitalsecretions, particularly during periods of stressor immunosuppression (4). B-virus infection istransmitted among free-ranging or group-housedanimals, primarily through sexual activity andbites. In captivity, as well as in the wild, maturemacaques are more likely than immatureanimals to have been infected with, and shed, thevirus. Antibody titer to B-virus indicatesinfection but can neither confirm nor eliminateactual viral shedding at the time of the bite (4).

B-Virus in HumansB-virus disease in humans usually results

from macaque bites or scratches (4). Incubationperiods may be as short as 2 days, but morecommonly are 2 to 5 weeks (1,3,5-7; Centers forDisease Control and Prevention [CDC], unpub.data). Most documented infections have occurredamong biomedical research employees who hadoccupational exposure to macaques, althoughtransmission has also been documented amonglaboratory workers handling infected centralnervous system and kidney tissues (1,5).

From 1990 to 1992, 28 U.S. residentsreported nonoccupational macaque bites to CDC(L. Chapman, pers. comm.). Since 1993,additional nonoccupational exposure cases havebeen reported, seven of which (involving 24persons and eight macaques) are listed in Table 1.Of the six macaques for which herpes B serologicresults were available, four (67%) were positive.Two owners refused requests for testing. Four(44%) of nine exposed children were bitten,versus only three (20%) of 12 adults. Childrenwere 3.2 times more likely to be bitten thanadults; although a common observation, thisassociation is not statistically significant for thiscase series.

Most free-ranging monkey populations arethought to be part of the exotic fauna of distanttourist destinations and wild animal parks;however, macaque species have established free-ranging feral populations in Texas and Florida.In such settings, contact between humans andmacaques cannot be safely controlled (8-10), andworkers and visitors are at risk. Guidelines for B-virus prevention and diagnosis have recentlybeen published (9-12).

Symptomatic human infection with B-virus israre; fewer than 40 cases were reported from1933 to 1994 (1,4-7,13-15; CDC, unpub. data).However, the consequences of symptomaticinfection may be severe. Viral infection rapidlyprogresses to central loci in the spinal cord and,eventually, the brain. Of 24 known symptomaticpatients whose cases were reviewed in 1992, 19(79%) died (CDC; unpub. data).

Before 1987, most surviving human patientshad moderate to severe neurologic impairment,sometimes requiring lifelong institutionaliza-

B-virus from Pet Macaque Monkeys: AnEmerging Threat in the United States?

Of primary concern when evaluating macaque bites are bacterial and B-virusinfections. B-virus infection is highly prevalent (80% to 90%) in adult macaques and maycause a potentially fatal meningoencephalitis in humans. We examined sevennonoccupational exposure incidents involving 24 persons and eight macaques. Sixmacaques were tested for herpes B; four (67%) were seropositive. A commonobservation was that children were more than three times as likely to be bitten thanadults. The virus must be assumed to be a potential health hazard in macaque bitewounds; this risk makes macaques unsuitable as pets.


Dispatches

tion (1). Recently, acyclovir has preventedprogression of the disease in a limited number ofpatients. In at least three patients, thistreatment reversed the neurologic symptoms andwas life-saving (7,14,15). Rapid diagnosis andinitiation of therapy are of paramount impor-tance in preventing death or permanentdisability in surviving patients.

Human and Macaque InteractionsMost owners form an emotional bond with

infant primates. This bond is probably strength-ened by the neonatal monkey’s physical andbehavioral resemblance to a human infant.Although physically and emotionally dependenton their mothers (or human substitutes) for up to 2years of age, most macaques exhibit unpredictablebehavior as they mature. Males tend to becomeaggressive, and both male and female macaquesbite to defend themselves and to establishdominance. Dominance within the social hierarchyof macaques is established by aggression towardother monkeys, generally the younger and smallermembers of the group. Both veterinary specialists

and breeders of nonhuman primates agree that asa rule, all these animals bite (16,17). Bitingincidents eventually bring the animals to theattention of animal control authorities. Most statehealth departments can require that any bitingnondomestic animal be euthanized and the brain besubmitted for rabies testing.

Regulations, Guidelines, and PoliciesRegarding Nonhuman Primates

Table 2 lists the principal federal regulationsaffecting the possession, distribution, and uses ofnonhuman primates. The United States isobligated under the Convention in InternationalTrade in Endangered Species (CITES) to restrictand control trafficking in exotic and endangeredspecies.

Since October 10, 1975, U.S. Public Healthregulation 42 CFR 71.53(c) has prohibited theimportation of nonhuman primates into theUnited States as pets, and neither nonhumanprimates imported since that date nor theiroffspring may be legally bred or distributed forany uses other than bona fide science, university-

Table 1. Selected pet macaque bite casesa,b

Primate species,Location age, B-virus status Nature of exposure CommentsIllinois Rhesus, 20+ yrs, Household contact Bought at auction, wife bitten

B-virus positive (2 adults/3 children), multiple sites, childrenCynomolgus, 2-4 yrs, bites, scratches (2 adults) hand-fed monkey B-virus negative

Florida Cynomolgus, 2 yrs, Household contact (1 adult), Kissed on lips, ate off owner’s B-virus positive bite (1 child) plate, shared bed

Arizona Cynomolgus, 2 yrs, Bites on toe and buttock Unprovoked attack on neighbor, B-virus negative (child) declared vicious animal by

judge, no. of household contacts (owner) unreported

Cynomolgus, 7 weeks, Household contact Diapered, shared chewed B-virus positive (6 adults), gum, oral ulcers noted by

bite on face (1 adult) veterinarian, bite incident at neighborhood bar

Macaque, (species Bite on thigh (1 child) Unprovoked attack (climbed undetermined), 2 yrs, fence to bite child) B-virus status unknownMacaque, (species Severe bite (1 child) Injured child attended an and age undetermined), unlicensed day-care facility B-virus status unknown run by monkey owner, 7

other monkeys on premisesMinnesota Rhesus, 2 yrs, Household contact, Acquired as “child-substitute”

B-virus positive owners’ friend bitten (full-time baby-sitters hired)aCases referred to Centers for Disease Control and Prevention since 1993.bAs of November 1997, no confirmed transmission of B-virus in these persons has been documented.


Dispatches

level educational programs, or full-time zoologicexhibition. Furthermore, the regulation states,“the maintenance of nonhuman primates as pets,hobby, or an avocation with occasional display toothers is not a permissible use” (18).

All states require their citizens to complywith applicable federal regulations. Many stateofficials, however, may be unaware of regulatoryrestrictions on the uses and distribution ofnonhuman primates and may be confused by thedistinctions among federal agencies regardingregulatory restrictions on captive-bred animals.State wildlife authorities may not know that afederal public health regulation prohibits thekeeping (“maintenance”) of nonhuman primatesimported after October 10, 1975, as pets, for ahobby, or as an avocation; likewise, many do notknow the compelling public health and safetyreasons for enforcement.

Captive-bred offspring of animals purported tohave been imported before October 10, 1975, arefrequently offered for sale. Without documentationit is very difficult to determine whether this is thecase. Depending on the specific circumstances, it ispossible for undocumented animals to be consid-ered deliberately misclassified (i.e., intentionallymislabeled), a violation under the Lacey Act (18USC 42) and under 16 USC 3373 (19).

In 1987 and 1988, occupational safetyguidelines were published based on evidence thatall macaque species are inherently dangerous tohumans because of the risk for B-virustransmission, as well as the likelihood of seriousphysical injury from bite wounds (9-12,14,15).Several recent reviews of monkey-bite injuriesworldwide indicate that severe lacerations,wound infections, and permanent sequelae (e.g.,flexure contractures, osteomyelitis) were present

in 33% of cases (20,21).In 1990, the American Veterinary Medical

Association issued a general policy statementopposing the keeping of wild animals (especiallythose inherently dangerous to humans) as petsand advising veterinarians to exert theirinfluence to discourage this practice (22). In 1995,updated guidelines for the prevention andtreatment of B-virus infections in exposedpersons were published (12). Despite thesecontinuing public health educational efforts,nonhuman primates (including macaques) con-tinue to be marketed and kept as pets in manystates (16,17,23).

The Frequency of Exposure Resulting inInfection

Much remains to be learned about thepathogenesis of B-virus infection in humans. Inthis very limited case series (Table 1), one family(two adults and two of three children) exposed toa B-virus positive macaque had flulike symp-toms. One of the adults had additional symptomsrelated to the injury site, which suggested B-virus infection. In the other six cases, no suspectclinical symptoms were noted, and disease-specific antiviral postexposure prophylaxis wasnot given. B-virus is still rare, and diagnosticevaluation of clinical cases of aseptic meningitisdoes not routinely include B-virus testing.

Owners of pet macaques are often reluctantto report bite injuries from their pets, even totheir medical care providers, and may fail toappreciate that the premonitory headache andflulike symptoms (which may lead them to seekmedical attention) could be associated withhealed, often minor, bite wounds dating back morethan a month (23). The Southwest Foundation for

Table 2. Federal regulations regarding nonhuman primates

Agency Statute Regulation SubjectsDepartments of Health, Public Health Services 42 CFR 71.53 Importation, distribution, bona fide Centers for Disease Act, 42 USCS 201 uses in the U.S., breeding colony Control and Prevention requirements, pet/avocationist usesDepartment of Animal Welfare Act, 9 CFR Licenses (breeders, dealers, laboratories, Agriculture 7 USCS 2131-2159 Subchapter A exhibitors, auctions), interstate

health certificates, humane care and transport

Department of the Endangered Species 50 CFR 10, 11, Endangered species, smuggling, Interior, U.S. Fish Act, 16 USCS 1540 13, 14, 16 interstate sales and Wildlife Service Lacey Act, 18

USCS 42


Dispatches

Biomedical Research, which is the designatedNational Institutes of Health B-virus resourcelaboratory, reports processing 2,000 to 3,000human diagnostic specimens per year between1990 and 1994, or approximately 200 per month,most of which reflect occupational exposure (8).

Some Public Health Consequences of theNonhuman Primate Pet Trade

The pet trade in a variety of nonhumanprimate species, and particularly the apparentincrease in macaque species as part of this trade,may constitute an emerging infectious diseasethreat in the United States. Although the U.S.Fish and Wildlife Service indicates that illegaltraffic in nonhuman primates is a significantaspect of the estimated $3 billion worth ofwildlife illegally traded in the United Statesannually, more data are needed on the actualnumber of macaques in the private sector andon trends in the population (24; U.S. Fish andWildlife special agents, pers. comm.).

The public resources deployed when amonkey-bite case is referred to public healthauthorities are similar to those required forrabies investigations (M. Leslie and T. Parrott,unpub. obs.). Persons bitten by pet and feralmacaques are more likely than persons bitten inthe workplace to require public resources, delayseeking medical care, and have an initial medicalevaluation by care givers who are largelyunfamiliar with the potentially serious conse-quences of B-virus exposure (23). In contrast,occupational exposure generally occurs withinhighly structured workplace settings, where healthprofessionals are prepared to provide prompt,appropriate, and specific care at no public cost.

Ongoing efforts to establish B-virus–freemacaque colonies illustrate the difficulties ofascertaining B-virus–negative status, even witha battery of sophisticated laboratory tests andextended longitudinal follow-up of individualmacaques (25). The high percentage of death inknown cases of human B-virus disease under-scores the potential seriousness of all bite orscratch exposures from macaques.

The extremely high prevalence of B-virusalong with their behavioral characteristics makethe macaque species unsuitable as pets.

AcknowledgmentsThe authors thank Drs. D. Manning, S. O’ Marro, R.

Montrey, T. Burke, D. Morton, and J. Thulin; Mr. J.

Francis, Mr. C. Langkop; Drs. R. Martin, J. Hilliard, D.Watkins; Capt. J. Thompson; Special Agents P. Bosco, D.Burleson, J. English, S. Hamilton, T. Karabinoff, D.Kirkby, D. Manera, G. Phillips, G. Phocas, T. Santelle, J.Sommers, C. Tabor; Dr. J. Cheek; and the Division of Viraland Rickettsial Diseases, Animal Resources Branch, and theDivision of Quarantine, CDC.

Stephanie R. Ostrowski,* Mira J. Leslie,†Terri Parrott,‡ Susan Abelt,¶ and

Patrick E. Piercy§*Centers for Disease Control and Prevention,

Atlanta, Georgia, USA; †Arizona Department ofHealth Services, Phoenix, Arizona, USA; ‡Pembroke

Park Animal Hospital, Pembroke Park, Florida, USA;¶Lake Superior Zoological Gardens, Duluth, Minne-

sota, USA; and §Illinois Department of PublicHealth, Springfield, Illinois, USA

References 1. Palmer AE. B virus, Herpesvirus simiae: historical

perspective. J Med Primatol 1987;16:99-130. 2. Shah KV, Morrison JA. Comparison of three rhesus

groups for antibody patterns to some viruses: absenceof active simian virus 40 transmission in the free-ranging rhesus of Cayo Santiago. Am J Epidemiol1972;99:308-15.

3. Orcutt RP, Pucak GJ, Foster HL, Kilcourse JY.Multiple testing for the detection of B virus antibody inspecially handled rhesus monkeys after capture fromvirgin trapping grounds. Lab Anim Sci 1976;26:70-4.

4. Weigler BJ. Biology of B Virus in macaque and humanhosts: review. Clin Infect Dis 1992;14:555-67.

5. Hummeler K, Davidson WL, Henle W, LaBoccetta AC,Ruch HG. Encephalomyelitis due to infection withHerpesvirus simiae (Herpes B virus): report of twofatal laboratory acquired cases. N Engl J Med1959;261:64-8.

6. Centers for Disease Control. B-virus infections inhumans—Michigan. MMWR Morb Mortal Wkly Rep1989;38:453-4.

7. Holmes GP, Hilliard JK, Klontz KC, Rupert AH,Schindler CM, Parrish E, et al. B-virus (Herpesvirussimiae) infection in humans: epidemiologicinvestigations of a cluster. Ann Intern Med1990;112:833-9.

8. Centers for Disease Control. Guidelines for preventionof Herpesvirus simiae (B-virus) infection in monkeyhandlers. MMWR Morb Mortal Wkly Rep 1987;36:681-2, 687-9.

9. The B Virus working group. Guidelines fromprevention of Herpesvirus simiae (B virus) infection inmonkey handlers. J Med Primatol 1988;17:77-83.

10. Hilliard JK. 1990-1994 yearly comparisons; B-virusResource Laboratory. San Antonio (TX): SouthwestFoundation for Biomedical Research; 1995.

11. Wells DL, Lipper SL, Hilliard JK, Stewart JA, HolmesGP, Herrmann KL, et al. Herpesvirus simiaecontamination of primary rhesus monkey kidney cellcultures. CDC recommendations to minimize risks tolaboratory personnel. Diagn Microbiol Infect Dis1989;12:333-6.


Dispatches

12. Holmes GP, Chapman LE, Stewart JA, Straus SE,Hilliard JK, Davenport DS, et al. Guidelines for theprevention and treatment of B-virus infections inexposed persons. Clin Infect Dis 1995;20:421-39.

13. Artenstein AW, Hicks CB, Goodwin BS, Hilliard JK.Human infection with B virus following a needlestickinjury. Reviews of Infectious Diseases 1991;13:288-91.

14. Centers for Disease Control. B-virus infection inhumans—Pensacola, Florida. MMWR Morb MortalWkly Rep 1987;36:289-90, 295-6.

15. Centers for Disease Control. B-virus infection inhumans—Michigan. MMWR Morb Mortal Wkly Rep1989;38:453-4.

16. Hoctor P. Short hints on raising exotics. AnimalFinders Guide 1995 Sep 1;12:44-5.

17. Johnson-Delaney CA. The pet monkey: health care andhusbandry guidelines. Journal of Small Exotic AnimalMedicine 1991;1:32-7.

18. Centers for Disease Control. Import restrictions onnonhuman primates. Atlanta (GA): Centers forDisease Control; 1982 Nov 17. DQ-CPS AdvisoryMemorandum No. 65.

19. Anderson RS. Lacey Act. Case law update and currentissues in federal wildlife prosecution. Washington:U.S. Department of Justice, Environment and NaturalResources Division; 1993.

20. Goldstein EJC, Pryor EP, Citron DM. Simian bites andbacterial infection. Clin Infect Dis 1995;20:1551-2.

21. Janda DH, Ringler DH, Hilliard JK, Hankin RC,Hankin FM. Nonhuman primate bites. J Orthop Res1990;8:146-50.

22. American Veterinary Medical Association. PolicyStatement on Wild Animals as Pets. In: 1995 AVMADirectory. Schaumburg (IL): The Association; 1995.

23. Paulette J. “Yes, monkeys will bite!” The Simian 1996Feb;6-8.

24. Webster D. The looting and smuggling and fencing andhoarding of impossibly precious, feathered and scalywild things: inside the $10 billion black market inendangered animals. The New York Times Magazine1997 Feb 16: 27-33, 48-50.

25. Hilliard J. Managing macaques and herpes B. Presentedat 4th National Symposium on Biosafety: Working Safelywith Research Animals; 1997 Jan 27-31; Atlanta, Georgia.


Letters

Infectious Diseases and Mental Illness:Is There a Link?

To the Editor : The report by Hatalski et al. (1) onBorna virus as a probable human pathogenprovides yet another example of an infectious agentbeing tentatively associated with neuropsychiatricdisorders. Earlier this year, researchers atRockefeller University and the National Instituteof Mental Health suggested that after streptococ-cal infection, some children may be at increasedrisk for obsessive-compulsive disorders andTourette syndrome (2). The human B-cell antigenD8/17, believed to be a marker for increasedsusceptibility to poststreptococcal rheumaticheart disease, has been tentatively linked to thisincreased risk for psychiatric illness in children.Other reports of patients with complicated Lymeborreliosis, including some whose infections haveprogressed to encephalopathies, describe persis-tent verbal and memory deficits among thesepatients (3). In a few Lyme disease patients, theonly overt symptoms of disease at the time ofinitial diagnosis and treatment were classified asmental confusion (4). Two newly emergentinfectious diseases in the United States, leptospiro-sis and neurocysticercosis, have been found amonginner city residents and poor immigrants,respectively. Occasionally leptospirosis has beenassociated with a variety of postinfectiouspsychiatric symptoms, including depression,dementia, and psychosis (5). Neurocysticercosis,a tropical parasitic infection, is increasinglyassociated with emergency room admissions forseizures and epilepsy (6). Still other infectiousdiseases are being examined for links withcognitive symptoms and emotional disorders.

The primary cause of many commonpsychiatric disorders, including depression,manic depression, anxiety, and schizophrenia,remains a mystery. The World Health Organiza-tion estimates that 1.5 billion people worldwidesuffer from a neuropsychiatric disorder. Of the 10leading causes of disability in 1990, four werepsychiatric disorders: unipolar depression, manicdepression, schizophrenia, and obsessive-com-pulsive disorders (7). The National Institute ofMental Health recently estimated that as manyas 20% of young Americans ages 7 to 14—approximately 10 million children—have mentalhealth problems severe enough to compromisetheir ability to function (8). Infectious agents may

play a role in some of these diseases to someunknown degree. A better understanding of the roleof infection may speed treatment and preventionefforts and reduce the degree of disability andstigma associated with mental illness.

Vaccines and antimicrobial agents mightenhance current therapeutic options for mentalillnesses. Even if infectious diseases were aprimary factor in only 1% of neuropsychiatricillnesses, some 10 million persons might benefitfrom antimicrobial therapies. Identifying thosesusceptible to neuropsychiatric illnesses (be-cause of environmental factors or geneticpredisposition) may also permit vaccination orantimicrobial prophylaxis and a subsequentlowering of disease incidence.

Physicians and federal agencies addressingthe problems of emerging infectious diseasesshould examine the possibility of infection as acause of mental illness. Better communicationamong infectious disease and mental healthexperts, as well as additional training, will beneeded to shed light on the growing phenomenonof infectious diseases manifesting themselves asneuropsychiatric disorders.

Edward McSweeganNational Institute of Allergy and Infectious

Diseases, Bethesda, Maryland, USA

References 1. Hatalski CG, Lewis AJ, Lipkin WI. Borna disease.

Emerg Infect Dis 1997;3:129-35. 2. Swedo SE, Leonard HL, Mittleman BB, Allen AJ,

Rapoport JL, Dow SP, et al. Identification of childrenwith pediatric autoimmune neuropsychiatric disordersassociated with streptococcal infections by a markerassociated with rheumatic fever. Am J Psychiatry1997;154:110-2.

3. Benke T, Gasse T, Hittmair-Delazer M, SchmutzhardE. Lyme encephalopathy: long-term neuropsychologicaldeficits years after acute neuroborreliosis. Acta NeurolScand 1995;91:353-7.

4. Fallon BA, Nields JA. Lyme disease: a neuropsychiatricillness. Am J Psychiatry 1994;151:1571-83.

5. Mumford C, Dudley N, Terry H. Leptospirosispresenting as a flaccid paraplegia. Postgrad Med J1990;66:218-20.

6. Garcia H, Tsang V, Gonzalez A, Gilman R.Cysticercosis in the U.S. and Peru. Presented at the 6thAnnual National Institutes of Health InternationalCenters for Tropical Disease Research Meeting; 1997May 5-7; Bethesda, Maryland.

7. World Health Organization. The global burden ofdisease [cited 1998 Jan 5]; [2 screens]. Available fromURL: http://www.who.ch/programmes/mnh/msamain.htm


Letters

8. Children’s mental health research severely neglected,says National Institute of Mental Health Director.Washington Fax [serial online] 1997 Apr 8 [cited 1998 Jan5]. Available from URL: http://www.washington-fax.com

Mycobacterium nonchromogenicumBacteremia in an AIDS Patient

To the Editor: Mycobacterium avium complex isthe most common nontuberculous mycobacteriumthat causes disseminated infection in HIV-positive patients (1). Other less commonnontuberculous mycobacteria responsible fordisseminated disease in these patients are M.fortuitum (2), M. genavense (3), M. gordonae (4),M. haemophilum (5), M. kansasii (6), M.malmoense (7), M. marinum (8), M. scrofulaceum(9), M. simiae (10), M. szulgai (11), M. terrae (9),and M. xenopi (2,9). Although only M. genavense,M. kansasii, and M. xenopi are significantly morefrequent in these patients (2,3,9), HIV infection islikely a predisposing condition for allnontuberculous mycobacterial infections. Wereport the first case of disseminated infectioncaused by M. nonchromogenicum in an HIV-infected patient.

A 28-year-old man with HIV infectionacquired by sharing injection tools was seen inour outpatient clinic because of intermittentfever, drenching nocturnal sweats, and coughwith purulent sputum of 4 months’ duration. Healso reported a weight loss of 10 kg in theprevious 2 months. He had been diagnosed withbronchial infection in another hospital and hadbeen treated with an unknown antibiotic. Afterthis treatment, respiratory symptoms hadimproved somewhat, but fever and constitutionalsymptoms continued. His only previous opportu-nistic infection had been recurrent oral andesophageal candidiasis. The last CD4-cell counthad been 16/µL 1 year earlier, and he wasreceiving didanosine and prophylactic therapywith cotrimoxazole and fluconazole. On physicalexamination the patient appeared ill; he wasfebrile, cachectic, and had thrush and oral hairyleukoplakia. Neither lymphadenopathy norabnormal cardiopulmonary symptoms werefound. The liver, which had enlarged since thelast examination, was palpated 6 cm below theright costal margin. Abnormal laboratory valuesincluded aspartate aminotransferase 61 U/L,gamma-glutamyl transferase 209 U/L, lactate

dehydrogenase 516 U/L, hemoglobin 12.3 g/dL,leukocyte count 4,300/µL (66% neutrophils, 19%band forms, 1% metamyelocytes, 3% lympho-cytes, 11% monocytes), platelet count 130,000/µL,and erythrocyte sedimentation rate 72 mm/h.Chest X-rays were unremarkable, and a set ofblood cultures was sterile. A sputum cultureyielded Haemophilus influenzae sensitive toampicillin, and smears and cultures for mycobac-teria in one stool and three sputum samples werenegative.

The patient was treated with oral amoxicillinfor 2 weeks without improvement. Empiricaltherapy against M. avium complex withclarithromycin, ciprofloxacin, and ethambutolwas started; the patient’s condition improveddramatically within the next few days, and thefever and diaphoresis disappeared, althoughcough and sputum production remained un-changed. Three weeks later, a slow-growingnonphotochromogenic mycobacterium, identifiedas M. nonchromogenicum in a referencelaboratory (Centro Nacional de Microbiología,Majadahonda, Madrid, Spain) by biochemicaltests and confirmed by polymerase chainreaction-restriction enzyme pattern analysis,was isolated from a blood sample obtained onadmission. This microorganism was sensitive tothe three drugs administered, and the treatmentwas continued. Two months later the patient hadgained 10 kg, hemoglobin had increased to 13.8g/dL, the erythrocyte sedimentation rate haddecreased to 52 mm/h, and the differentialleukocyte count had returned to normal.Antimycobacterial drugs were withheld after 1year of treatment. Twenty-two months after thediagnosis, the patient is doing well. He isreceiving combination antiretroviral therapy,and his CD4-cell count is 128/µL.

M. nonchromogenicum, a slow-growing non-pigmented (Runyon’s group III) mycobacterium,belongs to the M. terrae complex, together withM. triviale; it is traditionally considerednonpathogenic. However, it has been involved ina few cases of pulmonary infection (12) andchronic tenosynovitis secondary to puncturewounds (13), like the related organism M. terrae.In fact, some authors think that M.nonchromogenicum is the true pathogen in theM. terrae complex (13), and it is possible thatsome reports have misidentified this organism.This complex was first isolated in soil washings


Letters

from radishes, but it has been found to beubiquitous in the aquatic environment, includinga hospital potable water supply (14).

Unlike osteoarticular infections, whichcommonly occur in previously healthy people,the scanty reports on pulmonary and dissemi-nated infection by M. terrae complex suggestthat either immunosuppression or local predis-posing conditions (e.g., tuberculous cavities)are necessary pathogenetic cofactors (15). Toour knowledge, M. nonchromogenicum bacter-emia has never been reported before.

No specific DNA probes exist for M. terraecomplex, but false-positive reactions with M.tuberculosis complex DNA probes have beendescribed (16). Isolates are usually resistant to mostantituberculosis drugs, with the exception ofethambutol and streptomycin, and susceptible toerythromycin, ciprofloxacin, and sulfonamides.

Only one case of disseminated infection by M.terrae has been described in a patient withadvanced HIV infection and positive cultures inblood and bronchoalveolar lavage fluid, but noadditional data were provided (9). Although theisolate we recovered might represent a labora-tory contaminant, several pieces of evidencemake this possibility very unlikely: lack ofalternative explanation for a persistent andprogressive clinical picture of 4 months’ duration,absence of response to standard antibiotic therapy,negative results in the search for other pathogens,rapid and sustained clinical and laboratoryresponse to drugs active against this strain, clearimprovement despite the lack of treatment forother conditions, and absence of other isolates ofthis pathogen in our hospital despite the largenumber of samples examined for mycobacteria.

AcknowledgmentWe are indebted to María Soledad Jiménez,

Mycobacterial Laboratory of the Centro Nacional deMicrobiología, Majadahonda, Madrid, for providing themicrobiologic data.

José Mayo, Julio Collazos,and Eduardo Martínez

Hospital de Galdakao, Vizcaya, Spain

References 1. Nightingale SD, Byrd LT, Southern PM, Jockusch JD,

Cal SX, Wynne BA. Incidence of Mycobacteriumavium-intracellulare complex bacteremia in humanimmunodeficiency virus-positive patients. J Infect Dis1992;165:1082-5.

2. Raszka WV Jr, Skillman LP, McEvoy PL, Robb ML.Isolation of nontuberculous, non-avium mycobacteriafrom patients infected with human immunodeficiencyvirus. Clin Infect Dis 1995;20:73-6.

3. Bessesen MT, Shlay J, Stone-Venohr B, Cohn DL,Reves RR. Disseminated Mycobacterium genavenseinfection: clinical and microbiological features andresponse to therapy. AIDS 1993;7:1357-61.

4. Lessnau KD, Milanese S, Talavera W. Mycobacteriumgordonae: a treatable disease in HIV-positive patients.Chest 1993;104:1779-85.

5. Straus WL, Ostroff SM, Jernigan DB, Kiehn TE,Sordillo EM, Armstrong D, et al. Clinical andepidemiologic characteristics of Mycobacteriumhaemophilum, an emerging pathogen inimmunocompromised patients. Ann Intern Med1994;120:118-25.

6. Parenti DM, Symington JS, Keiser J, Simon GL.Mycobacterium kansasii bacteremia in patientsinfected with human immunodeficiency virus. ClinInfect Dis 1995;21:1001-3.

7. Chocarro A, González-López A, Breznes MF, Canut A,Rodríguez J, Diego JM. Disseminated infection due toMycobacterium malmoense in a patient infected withhuman immunodeficiency virus. Clin Infect Dis1994;19:203-4.

8. Tchornobay AM, Claudy AL, Perrot JL, Levigne V,Denis M. Fatal disseminated Mycobacterium marinuminfection. Int J Dermatol 1992;31:286-7.

9. Schafer RW, Sierra MF. Mycobacterium xenopi,Mycobacterium fortuitum, Mycobacterium kansasii,and other nontuberculous mycobacteria in an area ofendemicity for AIDS. Clin Infect Dis 1992;15:161-2.

10. Valero G, Peters J, Jorgensen JH, Graybill JR. Clinicalisolates of Mycobacterium simiae in San Antonio,Texas: an 11-yr review. Am J Respir Crit Care Med1995;152:1555-7.

11. Roig P, Nieto A, Navarro V, Bernacer B, Borras R.Micobacteriosis por Mycobacterium szulgai en pacientecon infección por el virus de la inmunodeficienciahumana. An Med Interna 1993;10:182-4.

12. Tsukamura M, Kita N, Otsuka W, Shimoide H. A study ofthe taxonomy of the Mycobacterium nonchromogenicumcomplex and report of six cases of lung infection due toMycobacterium nonchromogenicum. Microbiol Immunol1983;27:219-36.

13. Ridderhof JC, Wallace RJ Jr, Kilburn JO, Butler WR,Warren NG, Tsukamura M, et al. Chronic tenosynovitis ofthe hand due to Mycobacterium nonchromogenicum: useof high-performance liquid chromatography foridentification of isolates. Rev Infect Dis 1991;13:857-64.

14. Lockwood WW, Friedman C, Bus N, Pierson C, GaynesR. An outbreak of Mycobacterium terrae in clinicalspecimens associated with a hospital potable watersupply. American Review of Respiratory Diseases1989;140:1614-7.

15. Peters EJ, Morice R. Miliary pulmonary infectioncaused by Mycobacterium terrae in an autologous bonemarrow transplant patient. Chest 1991;100:1449-50.

16. Ford EG, Snead SJ, Todd J, Warren NG. Strains ofMycobacterium terrae complex which react with DNAprobes for M. tuberculosis complex. J Clin Microbiol1993;31:2805-6.


Letters

Escherichia coli O157:H7 Infection inColombia

To the Editor: The prevalence of Escherichiacoli O157:H7 in Colombia is not known; weconducted a study to determine its prevalencein children with infectious diarrhea, in adultcattle, and in ground beef.

Between March 1996 and March 1997, weexamined 538 children under 5 years of age withinfectious diarrhea who had been admitted toeither of the two children’s hospitals in Bogotá.Diarrhea was defined as three or more loosestools within the previous 24 hours. One hundredand sixty-one children under 10 years of ageadmitted to the hospital for a medical reasonother than infectious diarrhea served as controls.

Stool samples from children with and withoutdiarrhea were placed in Stuart transport mediumand sent to the laboratory within 24 hours; thesamples were injected into sorbitol MacConkeyagar (Oxoid Basingstoke, United Kingdom). After24 hours of incubation at 37°C, sorbitolnonfermenting colonies were tested with 4-methylumbelliferyl-ß-glucuronide (MUG); alltypical colonies of E. coli O157:H7 that weresorbitol-negative were confirmed as E. coli bybiochemical tests (1,2) and were tested foragglutination with a latex test kit (OxoidBasingstoke, United Kingdom) for detecting E.coli O157 and E. coli H antiserum H7 (Difco,Detroit, MI, USA). All human isolates wereconfirmed as Shiga-toxin producers by latexagglutination (Oxoid Basingstoke, United King-dom). We used as a control strain E. coli O157:H7provided by M. Karmali.

Rectal swabs from 307 healthy adult cattlefrom farms in Cundinamarca and MetaDepartments were placed in Stuart transportmedium, stored at 4°C, and transported to thelaboratory within 6 hours. Swabs were injectedinto sorbitol MacConkey agar, and colonies thatdid not ferment sorbitol were characterized bystandard techniques (3).

One hundred and fifty beef patties (31cooked, 119 raw), collected in Bogotá, wereexamined by direct plating and enrichmentculturing. Samples (1.0 g each) were seriallydiluted (1:10) in 0.85% NaCl solution, and 0.1 mlportions were plated in duplicate onto sorbitolMacConkey agar. Serologic and biochemicalconfirmation was done as mentioned above.

E. coli O157:H7 prevalence among childrenwas 7.2%, with an age range of 0–60 months(average 21 months); diarrheal illness lasted anaverage of 2.5 days. Of 39 patients, eight were 6months old or younger, five were 6 to 12 monthsold, 17 were 12 to 24 months old, and nine wereolder than 24 months. Renal failure associatedwith hemolytic uremic syndrome (HUS) devel-oped in three (7.7%). Epidemiologic data were notcollected regarding contaminated foods as apossible source of E. coli O157:H7 infection in thepatients. E. coli O157:H7 was isolated from five(3.1%) of the 161 controls; the prevalence of E. coliO157:H7 was substantially higher in patientswith infantile diarrhea than in controls.

All 39 strains from human cases weresorbitol-negative; five did not display MUGactivity. Overall, 39 strains agglutinated stronglywith antiserum O157:H7. Antimicrobial suscep-tibility tests were performed by the Bauermethod (4). All E. coli O157:H7 isolated weresusceptible to ciprofloxacin; 92% were resistantto ampicillin; 76% were resistant to furazolidone;and 76% were resistant to trimethoprim-sulfamethoxazole (TMP-SMZ).

E. coli O157:H7 was isolated from 20 (6.5%) of307 rectal swabs from cattle. The strains isolatedwere sorbitol-negative and agglutinated stronglywith antisera; five did not present activity. Allstrains were susceptible to ciprofloxacin; 90%were resistant to ampicillin; and 26% wereresistant to TMP-SMZ. E. coli O157:H7 wasisolated from 13 (87%) of 150 beef patties, sixfrom raw beef, and seven from cooked beef.

Stool cultures of all patients with acutebloody diarrhea should be tested for E. coliO157:H7 to identify those at risk for HUS (5);however, serotyping, cytotoxicity assays, or DNAprobing for E. coli O157:H7 are not routinelyperformed in Colombia.

This preliminary report suggests that E. coliO157:H7 is emerging as an important cause ofendemic childhood diarrhea in Colombia and thatthe chain of contamination is present. Theincidence is greatly underestimated because oflimited surveillance and reporting. Furtherstudies are needed to identify the pathogenicmechanisms of these E. coli O157:H7 strains andto determine the fecal carriage rate in healthychildren. Data obtained will help elucidate therole of E. coli O157:H7 in childhood diarrhea. Inaddition, molecular analysis should be performed


Letters

to establish the connection between the strainsisolated from different sources in Colombia.

Our findings suggest that the risk for E. coliO157:H7 infection in Colombia is high; therefore,more active screening and surveillance wouldenhance case detection, epidemiologic understand-ing of E. coli O157:H7 infection and HUS, and couldlead to more specific therapeutic interventions.

Salim Mattar and Elizabeth VásquezFacultad de Ciencias, Universidad Javeriana,

Bogotá, Colombia

References 1. Bockemuhl J, Aleksic S, Karch H. Serological and

biochemical properties of Shiga-like toxin(verocytotoxin)-producing strains of Escherichia coli,other than O-group 157, from patients in Germany.Zentralbl Bakteriol 1992;276:189-95.

2. Gunzer F, Bohm H, Rossmann H, Bitzan M, Aleksic S,Karch H. Molecular detection of sorbitol-fermentingEscherichia coli O157 in patients with hemolytic-uremic syndrome. J Clin Microbiol 1992;30:1807-10.

3. Chapman PA, Siddons CA, Wright DJ, Norman P, FoxJ, Crick E. Cattle as a possible source of verocytotoxin-producing Escherichia coli O157 infections in man.Epidemiol Infect 1993;439:447.

4. Bauer AW, Kirby WMM, Sherris JC, Turk M.Antibiotic susceptibility testing by a standardizedsingle disc method. Am J Clin Pathol 1966;45:493-6.

5. Wells JG, Davis BR, Wachsmuth IK, Riley LW, RemisRS, Sokolow R, Morris GK. Laboratory investigation ofhemorrhagic colitis outbreaks associated with a rareEscherichia serotype. J Clin Microbiol 1983;18:512-20.

Autofluorescence and the Detection ofCyclospora Oocysts

To the Editor: From May through July 1997, wesearched for the seasonally occurring Cyclosporacayetanensis, along with other coccidia andmicrosporidia, in fecal samples from 385patients. The samples, in 10% formalin forevaluation of coccidia and microsporidia, wereinitially processed by a routine formalin-ethylacetate concentration method; the parasite wasdetected in 18 patients (1,2). The resultingsediment was examined as follows. A drop ofsediment was placed on a slide, cover-slipped,and examined microscopically as a wet mount at200x and 400x magnification and subsequently at200x magnification by epifluorescence with a 330to 380 nm UV filter. Four smears were alsoprepared and stained by routine trichrome (2),modified trichrome (3), auramine-rhodamine (4),

and Kinyoun acid-fast (5) procedures. All wetmount and stained preparations were evaluatedby at least two trained persons.

Of the 385 fecal samples examined, 18 werepositive for C. cayetanensis. The positive sampleswere from eight states, which encompassednortheastern (Rhode Island, New York, Massa-chusetts, Pennsylvania), midwestern (Wiscon-sin), western (Oregon, California), and southern(Florida) sections of the United States.

In 12 of 18 patients, the organisms weredetected without much difficulty in wet mountsas round or partially collapsed nonrefractilebodies; however, in the other six, repeated wetpreparations were needed to detect the organ-isms. When the same wet mounts were examinedwith epifluorescence microscopy, oocysts wereeasily discerned in all samples, even the six inwhich repeated wet preparations and stains wereneeded. While the trichrome procedures wereineffective, the auramine-rhodamine and Kinyounstains gave varied results. The autofluorescencetechnique, however, was distinctly superior tothe wet mount and staining procedures.

Extensive outbreaks of diarrhea caused byC. cayetanensis were reported in 1997 fromdifferent parts of the United States (6-8), andseveral procedures have been used to confirm thediagnosis in clinical samples. While the organ-isms are large enough to be seen in direct wetmounts, they are frequently caught up in mucusor covered by debris, so they are difficult todetect. Autofluorescence in C. cayetanensisoocysts makes them easily visible in clinicalsamples (1,9) with the use of a 330 to 380 nmUV filter; this feature enhanced their detectionat least twofold over the direct wet mount,especially when the wet mount and stainedslides contained few oocysts. (The same wetmount preparation can be used for theepifluorescence procedure.)

The 18 patients with cyclosporiasis were ages2 to 71 years, which indicates that the infectionwas not specific to any age group. Twelve of the 18cases were in women. Massachusetts had 11, thelargest number of C. cayetanenis-positive pa-tients. Of the 18, 16 were adults; the other twowere children with a coexisting parasite(Dientamoeba fragilis). In one instance, threemembers of the same family were infected, theparents with only C. cayetanensis, the son with D.fragilis and Blastocystis hominis.


Letters

Because C. cayetanensis is a seasonal diarrhealagent, fecal samples from persons with persistentunexplained explosive diarrhea during thesummer should be carefully evaluated for thisinfection. Stool specimens should be fixed in 10%formalin and examined with autofluorescencemicroscopy for enhanced detection.

O.G.W. Berlin,*† J.B. Peter,† C. Gagne,† C.N.Conteas,‡ and L.R. Ash*

*School of Public Health, University of California,Los Angeles, California, USA; †Specialty

Laboratories, Santa Monica, California, USA;and ‡Kaiser Permanente Medical Center,

Los Angeles, California, USA

References 1. Berlin OGW, Novak SM, Porschen RK, Long EG,

Stelma GN, Schaeffer FW. Recovery of Cyclosporaorganisms from patients with prolonged diarrhea. ClinInfect Dis 1994;18:606-9.

2. Ash LR, Orihel TC. Collection and preservation offeces. Parasites: a guide to laboratory procedures andidentification. Chicago: ASCP Press; 1991. p. 3-53.

3. Weber R, Bryan RT, Owen RL, Wilcox CM, Gorelkin L,Visvesvara GS. Improved light microscopical detectionof microsporidia spores in stool and duodenal samples.N Engl J Med 1992;326:161-6.

4. Berlin OGW. Mycobacteria. In: Baron EJ, FinegoldSM, editors. Diagnostic microbiology. 8th ed. St. Louis:C.V. Mosby; 1990. p. 597-640.

5. Ash LR, Orihel TC. Atlas of human parasitology. 4thed. Chicago: ASCP Press; 1997.

6. Centers for Disease Control and Prevention. Update:outbreaks of cyclosporiasis—United States, 1997.MMWR Morb Mortal Wkly Rep 1997;46:451-2.

7. Centers for Disease Control and Prevention. Update:outbreaks of cyclosporiasis—United States and Canada,1997. MMWR Morb Mortal Wkly Rep 1997;46:521-3.

8. Colley DG. Widespread food-borne cyclosporiasisoutbreaks present major challenges. Emerg Infect Dis1996;2:354-6.

9. Eberhard ML, Pienazek NJ, Arrowood MJ. Laboratorydiagnosis of Cyclospora infections. Arch Pathol LabMed 1997;121:792-7.

Partnerships for Detecting EmergingInfectious Diseases: Nepal and GlobalInfluenza Surveillance

To the Editor: With new influenza strains emergingeach year, identification of circulating strains bycoordinated global surveillance is crucial to vaccinedevelopment for the coming year (1-3). Approxi-mately 110 laboratories in 80 countriesvoluntarily participate in the World HealthOrganization (WHO) influenza surveillance

network (4). Comprehensive surveillance isespecially important in Asia, since new influenzastrains often originate there. To participate ininfluenza global surveillance, countries need notrely on their own laboratory capability. Clinicalspecimens from patients thought to haveinfluenza can be sent to designated laboratoriesaround the world for analysis. A uniquepartnership has led to the expansion of the WHOglobal influenza surveillance network to Nepal.

The U.S. Army Medical Component - ArmedForces Research Institute for Medical Sciences(AFRIMS) (5) in Bangkok, Thailand, is wellsituated to assist with surveillance in Asia.Scientists at AFRIMS have conducted medicalresearch in collaboration with Nepali colleaguesfor more than 20 years. Several studies have beenconducted in collaboration with the CIWECClinic Travel Medicine Center (a travel medicineclinic that serves the diplomatic, aid, and touristcommunities in Nepal). The clinic has approxi-mately 5,000 patient visits per year, of which halfare drawn from the 2,500 expatriates in Nepaland half from the 200,000 non-Indian touristswho visit Nepal annually.

A protocol was developed for a pilot influenzasurveillance program. The staff of the CIWECClinic was responsible for volunteer recruitment,clinical evaluation, and specimen collection.Febrile upper respiratory infections weredefined as temperature ≥ 100°F (37.8°C, oral orequivalent) and cough or sore throat of ≤ 72 hoursduration. Other symptoms, such as streptococcalpharyngitis, were excluded. No age or genderrestrictions were included. Volunteers had to havebeen in Nepal for the 5 days preceding illness. Onlythe first patient in any single household withsimilar symptoms within days of other householdmembers was asked to participate.

The AFRIMS field station in Kathmandu(locally known as the Walter Reed/AFRIMSResearch Unit - Nepal or WARUN) wasresponsible for shipping specimens collected bythe CIWEC Clinic to AFRIMS, Thailand. Sincedry ice was not available in Kathmandu, dry iceand shipping containers were sent by AFRIMS,Thailand for use by WARUN. Shipments fromWARUN were then sent back to AFRIMS, wherespecimens were repacked in dry ice and sent fortesting at the central laboratory of the U.S. AirForce’s Project Gargle (6) in San Antonio, Texas.Project Gargle has been testing viral respiratory


Letters

specimens from distant Air Force installations formore than 20 years. Each specimen was tested forinfluenza A and B; parainfluenza virus 1, 2, and3; adenovirus; enterovirus; and herpesvirus.Characterization of selected influenza A and Bisolates by hemagglutination-inhibition testingwas performed by the Centers for Disease Controland Prevention (CDC).

Between December 1996 and February 1997,the CIWEC staff collected specimens from 18patients. Samples were collected from 11 (61%)residents and seven (39%) tourists, who wereevenly distributed by gender and had a median ageof 35 years. Influenza B/Beijing/184/93-like viruseswere isolated from five (28%) of the 18 specimens.All patients from whom influenza viruses wereobtained had mild illnesses with fever and upperrespiratory syndromes. Herpes virus type 1 andadenovirus type 6 were each identified in one otherspecimen. No respiratory viruses were identified inthe remaining 11 specimens.

Because of the importance of China in theemergence of new strains of influenza, CDC’sWHO Collaborating Center for Surveillance,Epidemiology, and Control of Influenza hasworked with colleagues in China to establish anational Chinese network of influenza surveil-lance sites. Analysis of viruses isolated in Chinabetween 1988 and 1997 in comparison with otherviruses obtained through WHO’s global influenzasurveillance network has shown that influenzavariants are frequently identified in China beforebecoming prevalent in other regions of the world.Nepal is another especially valuable surveillancesite, given its location between China and India(at the crossroads between northern andsouthern Asia) and its historic importance as atrans-Himalayan trade route.

Especially relevant are data from Chinademonstrating that the two antigenically andgenetically distinct lineages of influenza B virusesrepresented by B/Victoria/02/87 and B/Yamagata/16/88 (7) have continued to circulate and evolve inChina, while only viruses related to B/Yamagatahave been detected elsewhere in the world andare represented in the current trivalent vaccineby the B/Bejing/184/93-like component. Virologicsurveillance in surrounding countries (8) such asNepal is necessary to detect geographic spread ofB/Victoria-like virus in the region. Our datasuggest that these viruses have not yet spread toKathmandu.

Our unique international partnership be-tween several civilian and military organizations(e.g., CIWEC Clinic, CDC, U.S. Air Force, andU.S. Army) demonstrates the feasibility of suchpartnerships as well as the usefulness ofinfluenza surveillance data at both the local andglobal levels. Despite the small number ofisolates obtained during this study, we were ableto determine that the influenza B component ofthe trivalent vaccine prepared for the 1996-1997influenza season would likely have offeredprotection for travelers and the local populationagainst the influenza B strains isolated inKathmandu. Ongoing surveillance data willestablish geographic and temporal patterns ofcirculation of influenza viruses and thusprovide valuable information for guiding publichealth policies for influenza vaccination. On aglobal level, these data are useful for annualvaccine strain selection.

Advances in communication, laboratory, andspecimen transport technologies contributedgreatly to the identification of viral pathogensfrom a new sentinel surveillance site in Nepal. Inevaluating future collaborative sites, priorsurveillance experience and reliable specimenshipping should be prime considerations. Ap-proaches that use existing resources might fostergreater international cooperation toward im-proved global detection and reporting ofinfectious diseases.

AcknowledgmentFinancial support was given by the United States

Department of Defense Emerging Infectious Disease Program.

Jeffrey M. Gambel,* David R. Shlim,†Linda C. Canas,‡ Nancy J. Cox,¶

Helen L. Regnery,§ Robert M. Scott,§David W. Vaughn,§ Charles H. Hoke Jr.,* and

Patrick W. Kelley**Walter Reed Army Institute of Research,Washington, D.C., USA; †CIWEC Clinic,

Kathmandu, Nepal; ‡ArmstrongLaboratory - Diagnostic Virology, Brooks

Air Force Base, Texas, USA; ¶Centers for DiseaseControl & Prevention, Atlanta, Georgia, USA; and

§Armed Forces Research Institute of MedicalSciences, Bangkok, Thailand

References 1. Gross PA. Preparing for the next influenza pandemic: a

reemerging infection. Ann Internal Med 1996;124:682-5.


Letters

2. Glezen WP. Emerging infections: pandemic influenza.Epidemiol Rev 1996;18:64-76.

3. Patriarca PA, Cox NJ. Influenza pandemicpreparedness plan for the United States. J Infect Dis1997;176(Suppl 1):S4-S7.

4. Hampson AW. Surveillance for pandemic influenza. JInfect Dis 1997;176(Suppl 1):S8-S13.

5. Gambel JM, Hibbs RG. U.S. military overseas medicalresearch laboratories. Mil Med 1996;161:638-45.

6. Williams RJ, Cox NJ, Regnery HL, Noah DL, Khan AS,Miller JM, ET AL. Meeting the challenge of emergingpathogens: the role of the United States Air Force inglobal influenza surveillance. Mil Med 1997;162:82-6.

7. Rota PA, Hemphill ML, Whistler T, Regnery HL,Kendal AP. Antigenic and genetic characterization ofthe hemagglutinins of recent circulating strains oninfluenza type B virus. J Gen Virol 1992:73:2737-42.

8. Centers for Disease Control and Prevention. Updateinfluenza activity—United States and worldwide,1996-97 season, and composition of the 1997-98influenza vaccine. MMWR Morb Mortal Wkly Rep1997;46:325-30.

HIV-2 Infection and HIV-1/HIV-2Dual Reactivity in Patients Withand Without AIDS-RelatedSymptoms in Gabon

To the Editor : Between 1996 and 1997, weevaluated the incidence of HIV-2 infection at theFondation Jeanne Ebori, the second largesthospital in Libreville, capital of Gabon; we foundan unexpected high prevalence of HIV-2–infectedor HIV-1/HIV-2–dually reactive patients.

During a 10-month period, 147 (14.3%) of1,029 sera from inpatients and outpatients werefound HIV-positive by the type III methodrecommended by the World Health Organization(two enzyme-linked immunosorbent assays areused to screen anti-HIV antibodies) (1). Furtherdiscrimination between HIV-1 and HIV-2infections was assessed by using syntheticpeptides specific for the gp41 and the gp120 ofHIV-1 and the gp36 of HIV-2 (ImmunoComb II,PBS Orgenics, Illkirch, France). Of the 147 HIV-positive sera, 141 (96.0%) were exclusively HIV-1–positive; four were exclusively HIV-2–positive;and two were both HIV-1– and HIV-2–positive.Of the six sera with anti-gp36/HIV-2 reactivities,two (from patients A and B) were positive on HIV-2 Western blot, with marked anti-gag HIV-1cross-reactivity and a discrimination assaypositive only for HIV-2; two (from patients D andE) were positive on HIV-2 Western blot, withanti-gag and pol reactivities markedly lower than

anti-env reactivities and a discrimination assaypositive only for HIV-2; the two remaining sera(from patients C and F) showed typical dualreactivities for HIV-1 and HIV-2 infections, withpositive patterns of HIV-1 and HIV-2 Westernblots and a discrimination test positive for bothviruses. As a whole, six (4.1%) of 147 HIV-positivesera showed either HIV-2 infection alone (n = 4)or dual reactivity. Of those, four were fromGabonese patients B, C, D, and E, and two werefrom immigrants from West Africa (patient Afrom Mali and patient F from Nigeria); two werefemale patients B and E. Among Gabonesepatients, only one (patient E) had traveled toWest Africa; the remaining three had nevervisited any neighboring country. However, oneGabonese man (patient C) lived in Port-Gentil,which has many West African immigrants. Forall patients, the most likely risk factor for HIVwas a heterosexual relationship with anunknown HIV-infected person. In three asymp-tomatic patients (A, B, and C) the HIV-2–serostatus was unexpected; in contrast, the threeother patients had AIDS-related symptoms.Patients D and E had an HIV-2 Western blotpattern showing a marked decrease in anti-gagand pol reactivities compatible with theiradvanced stage of HIV-2 disease.

The case of a 55-year-old exclusivelyheterosexual asymptomatic woman (patient B)suggests the possibility of a specific variant ofHIV-2 in Central Africa (2). The high frequency inprimates in Gabon of natural infection with simianimmunodeficiency retroviruses, which show ahigh degree of genetic relatedness to HIV-2 (3),could support such a hypothesis.

Two patients had typical dual reactivities toHIV-1 and HIV-2 antigens. To our knowledge,such dual reactivities have never been reportedin Gabon (4). In the patient from Nigeria (patientF), the serologic pattern was typical of thatusually observed in West Africa (5). Dualreactivity can result from genuine mixedinfections and from serologic cross-reactivity inHIV-1 and HIV-2 infection alone; theoretically, itcould also represent infection with a different,cross-reacting recombinant strain (5).

HIV-2 infection in Gabon is epidemiologicallyrelated to West Africa, because of cultural and,above all, economic ties. However, HIV-2 is notlimited to immigrant populations from WestAfrica or to Gabonese citizens traveling in thisarea; it has also reached the indigenous Gabonese


Letters

population. The possibility of rare cases of HIV-1and HIV-2 coinfections, recombinant HIV-1 andHIV-2 strains, and also peculiar HIV-2 variantsfrom Central Africa, should be considered inGabon. A possible entry of HIV-2 infection intoCentral Africa from Gabon in the near futurecould have major public health implications.

AcknowledgmentsWe thank Dr. Jean Wickings for reviewing the

manuscript. CIRMF is funded by the Gabonese government,ELF Gabon, and the French Ministry of Cooperation.

Carol Tevi-Benissan,* Madeleine Okome,†Maria Makuwa,* Moise Ndong Nkoume,†

Joseph Lansoud-Soukate,* Alain Georges,*Marie-Claude Georges-Courbot,*

and Laurent Belec‡*Centre International de Recherches Medicales,Franceville, Gabon; †Fondation Jeanne Ebori,

Libreville, Gabon; and ‡Centre Hospitalo-Universitaire Broussais, Paris, France.

References 1. World Health Organization. Global programme on AIDS.

Recommendations for the selection and use of HIVantibody tests. Wkly Epidemiol Rec 1992;67:145-9.

2. Belec L, Martin PMV, Georges-Courbot MC, Brogan T,Gresenguet G, Mathiot CC, et al. Dementia as the primarymanifestation of HIV-2 infection in a Central Africanpatient. Ann Inst Pasteur Virol 1988;139:291-4.

3. Georges-Courbot MC, Moisson P, Leroy E, PingardAM, Nerrienet E, Dubreuil G, et al. Occurrence andfrequency of transmission of naturally occurringsimian retroviral infections (SIV, STLV, and SRV) atthe CIRMF Primate Center, Gabon. J Med Primatol1996;25:313-26.

4. Delaporte E, Janssens W, Peeters M, Buve A, DibangaG, Perret J-L, et al. Epidemiological and molecularcharacteristics of HIV infection in Gabon, 1986-1994.AIDS 1996;10:903-10.

5. Peeters M, Gershy-Damet G-M, Fransen K, Koffi K,Coulibaly M, Delaporte E, et al. Virological andpolymerase chain reactions studies of HIV-1/HIV-2dual infection in Côte d’Ivoire. Lancet 1992;340:339-40.

Q Fever in French Guiana: New Trends

To the Editor: Q fever, the endemic diseasecaused by the rickettsial organism Coxiellaburnetii, was first described in French Guiana in1955 (1). Only sporadic cases were reported until1996 when three patients were hospitalized inthe intensive care unit of the Cayenne Hospitalfor acute respiratory distress syndrome. One ofthe patients died. Many cases of Q fever werediagnosed in the general population at the same

time. A seroepidemiologic study was performedto determine whether the increase in cases wasdue to an increase in incidence or to animprovement in diagnosis. All paired samples ofsera (acute-phase and convalescent-phase) frompatients sent to the arbovirus laboratory fordiagnosis of dengue infection from January 1,1992, to December 31, 1996, were tested forantibodies to C. burnetii by immunofluorescence.All positive samples were also tested forimmunoglobulin (IgM) by the same method; theIgG and IgM titers were determined by using aserial twofold dilution. A diagnosis of Q feverwas made when there was a seroconversionfrom negative to positive or a twofold increasein IgG titer associated with the presence of IgMin the second sample.

One hundred and fifty-one of 426 paired seracollected between 1992 and 1996 were frompatients recently infected with dengue fever.Twenty-five (9.1%) of 275 remaining sera were fromQ fever patients. Significant differences wereobserved in the rates of Q fever in different years (p< 0.01); one (1.9%) of 53 was positive in 1992, five(9.1%) of 55 in 1993, five (8.6%) of 58 in 1994, three(4.8%) of 63 in 1995; a large increase was observedin 1996 (11 [23.9%] of 46). Differences by residencewere also assessed. Rates of infection were higher inCayenne (21 [13.0%] of 161) than in rural areas (4[3.5%] of 114) (p < 0.01).

This study shows that cases of Q fever haveoccurred in French Guiana in recent years andthat a significant increase in the incidence rateoccurred in 1996. The reasons for this increaseare unclear, and further studies of theepidemiology of Q fever in French Guiana arenecessary. The epidemiology of Q fever is unusualin French Guiana because the rates of infectionare much higher in Cayenne, the capital city,than in rural areas. No link with classical sourcesof infection (cattle, sheep, or goat birth products,or work in a slaughterhouse) was found. Indeed,Cayenne, with 80,000 inhabitants, is located nearthe Atlantic Ocean, and the prevailing windsblow from the sea. Airborne contamination fromrural areas is therefore impossible. Furthermore,no large farm is in the immediate vicinity of thecity. For identical reasons, contamination fromthe abattoirs is not likely; they are located onthe west side of the city, near the CayenneRiver, and the winds blow from the east. In ourstudy, cases were almost equally distributed


Letters

throughout the city, although many patientscame from the same area.

A seroepidemiologic study to determinepossible new sources of infection (e.g., dogs, cats)and estimate rates of seropositivity in cattle andsheep and a case-control study on new cases arebeing conducted.

F. Pfaff,* A. François,* D. Hommel,† I. Jeanne,*J. Margery,‡ G. Guillot,† Y. Couratte-Arnaude,‡

A. Hulin,† and A. Talarmin**Pasteur Institute of French Guiana, Cayenne,French Guiana; †Cayenne Hospital, Cayenne,

French Guiana; and ‡Army Health Care Service,Cayenne, French Guiana

Reference 1. Floch H. La pathologie vétérinaire en Guyane

française (les affections des porcins, des caprins et desovins). Revue Elev Méd Vét Pays Trop 1955;8:11-3.

Ixodes dammini: A Junior Synonym forIxodes scapularis

To the Editor: The authors of “A new tick-borneencephalitis-like virus infecting New Englanddeer ticks, Ixodes dammini” (1) provide usefulinformation regarding a possibly new tick-borneencephalitis-like virus. However, the use of thename Ixodes dammini is not accurate fordescribing this species. I. dammini (Spielman,Clifford, Piesman, and Corwin) was synonymizedwith Ixodes scapularis (Say) in 1993 by Oliver etal. (2) and was redescribed in 1996 (3) to reduceconfusion regarding identification. Keirans andcolleagues summarize a wide array of rigorousstudies involving hybridization, assortativemating, isozymes, and morphometrics, all ofwhich provide evidence supporting thesynonymization of the two tick species (3).

The synonymization of I. dammini with I.scapularis has been widely accepted. “I.scapularis (= I. dammini)” is still often used, butthe use of I. scapularis as the sole nomen for thisspecies is becoming more common (4). Oliver etal. (2) have established I. dammini as a juniorsubjective synonym of I. scapularis. If scientificallyrigorous evidence exists justifying the reestablish-ment of the species name I. dammini, it must bepublished according to proper procedure. Theproper nomenclature of any species, let alone one ofsuch widespread notoriety and public healthimportance, is too important to be relegated to a

footnote. Until such evidence is presented, thecontinued misuse of I. dammini serves only toconfuse health-care providers, public healthprofessionals, and lay persons.

On a secondary matter, on page 167 of thedispatch, the authors state that “I. (Pholeoixodes)cookei is a one-host tick that is only distantlyrelated to I. dammini and only rarely feeds onhumans or mice” (1). I. cookei is a three-host tick(D.E. Sonenshine, pers. comm.), as are all themembers of the genus Ixodes.

Martin SandersMaryland Department of Health and Mental

Hygiene, Baltimore, Maryland, USA

References 1. Telford SR III, Armstrong PM, Katavolos P, Foppa I,

Garcia ASO, Wilson ML, Spielman A. A new tick-borne encephalitis-like virus infecting New Englanddeer ticks, Ixodes dammini. Emerg Infect Dis1997;3:165-70.

2. Oliver JH, Owsley MR, Hutcheson AM, James C,Chen W, Irby S, et al. Conspecificity of the ticks Ixodesscapularis and Ixodes dammini (Acari: Ixodidae). JMed Entomol 1993;30:54-63.

3. Keirans JE, Hutcheson HJ, Durden LA, KlompenJSH. Ixodes scapularis (Acari: Ixodidae): redescriptionof all active stages, distribution, hosts, geographicalvariation, and medical and veterinary importance. JMed Entomol 1996;33:297-318.

4. Centers for Disease Control and Prevention. Lymedisease—United States, 1996. MMWR Morb MortalWkly Rep 1997;46:531-5.

The Name Ixodes damminiEpidemiologically Justified

To the Editor: Although a large body of evidencehas been interpreted as supporting conspecificityof the deer tick (Ixodes dammini) and theblacklegged tick (Ixodes scapularis), according toChapter VI, Article 23 L of the International Codeof Zoological Nomenclature (1), “A name that hasbeen treated as a junior synonym may be used asthe valid name of a taxon by an author whoconsiders the synonymy to be erroneous....”

Current use of I. scapularis to refer to thevector of Lyme disease obscures importantepidemiologic issues. One of the reasons for“sinking” I. dammini was to make it easier todiagnose Lyme disease in areas where thedisease was thought to be nonendemic: “Thebelief that I. dammini does not occur south ofMaryland and that I. scapularis is a separate and


Letters

distinct species yet unproven as a natural vectorof Lyme disease has caused delays in Lymedisease surveillance in the South. The generalattitude among physicians and veterinarians hasbeen that Lyme disease is not a problem in thatarea, although patients present clinical symp-toms of it” (2). Recognizing and reporting Lymedisease in southern and southcentral statesshould not, however, depend on whether the twoticks are conspecific. Only peer-reviewed descrip-tions of human cases of Lyme disease, withappropriate documentation of the diagnoses,should be accepted as evidence. Few such reportsexist, and the evidence does not convincinglysupport a conclusion that Lyme disease exists asan epidemic zoonosis in southern states (3). Thisis not to say that residents outside the well-established eastern United States zoonotic sites(the Northeast and upper Midwest) do not havesymptoms that fit one or more aspects of thecurrent Centers for Disease Control andPrevention/Council of State and TerritorialEpidemiologists/Association of State and Territo-rial Public Health Laboratory Directors surveil-lance definition for Lyme disease. Lyme disease-like infections, mainly manifesting as erythemamigrans and strongly associated with Lone Startick (Amblyomma americanum) bites are com-monly seen in southern and southcentral states,but Borrelia burgdorferi does not seem to be theetiologic agent (4). Enzootic transmission of Lyme disease spiro-chetes among rodents and ticks had beendocumented in southern and southcentral statesby the late 1980s (5-7). The question, however, iswhether there is frequent zoonotic transmission.There are widespread southern U.S. enzooticcycles of Trypanosoma cruzi, but few autochtho-nous human Chagas disease cases seem to occurbecause the vectors (such as Triatoma sanguisuga)have behavioral traits that reduce their capacityto serve as zoonotic vectors (8). Nymphal I.scapularis apparently do not frequently bitehumans (7,9), although adult ticks do. The majorfeature of Lyme disease epidemiology in theNortheast and in the upper Midwest, however, istransmission by nymphal I. dammini (10).

Whether the predilection of nymphal I.dammini to feed on humans is environmentallydetermined or is a heritable trait withundescribed genetic markers remains unex-plored. Particular mitochondrial DNA haplotypesseem to be more characteristic of I. dammini

(11,12), and the use of such typing methods mayenhance future analyses of the vectorial capacityof these ticks. For example, one might test thehypothesis that nymphal ticks removed fromresidents of sites in coastal North Carolinathrough Georgia, where both kinds of ticks havebeen collected, represent only I. dammini. But, ifit is “widely accepted” that no differences existbetween the two ticks, such studies may never bedone. Similarly, many may wrongly assume thatLyme disease, human babesiosis, and humangranulocytic ehrlichiosis are, or will become,epidemic throughout virtually all of the easternUnited States. An equally likely scenario is thatthese zoonoses may never become public healthproblems for more southerly states. For themoment, then, distinguishing tick populations thatfrequently bite humans from those that rarely doseems to be a rational use of nomenclature,particularly for public health officials.

Dr. Sanders is correct in pointing out that allIxodes spp. are “three-host” ticks, although myintent in using the term “one-host” was toindicate that all stages of I. cookei tend to feed onthe same kind of animal (sometimes a singleanimal, within burrows), usually woodchucks,skunks, or raccoons. I regret the confusion from myuse of the acarologic term in a descriptive context.

Sam R. Telford IIIHarvard School of Public Health,

Boston, Massachusetts, USA

References 1. International code of zoological nomenclature. 3rd ed.

London: International Trust for ZoologicalNomenclature; 1985.

2. Oliver JH Jr, Owsley MR, Hutcheson HJ, James AM,Chen C, Irby WS, et al. Conspecificity of the ticksIxodes scapularis and I. dammini (Acari: Ixodidae). JMed Entomol 1993;30:54-63.

3. Barbour AG. Does Lyme disease occur in the South? Asurvey of emerging tickborne infections in the region.Am J Med Sci 1996;311:34-40.

4. Campbell GL, Paul WS, Schriefer ME, Craven RB, Rob-bins KE, Dennis DT, et al. Epidemiologic and diagnosticstudies of patients with suspected early Lyme disease,Missouri, 1990-1993. J Infect Dis 1995;172:470-80.

5. Levine JF, Apperson CS, Nicholson WL. The occurrence ofspirochetes in ixodid ticks in North Carolina. Journal ofEntomological Science1989;24:594-602.

6. Kocan AA, Mukolwe SW, Murphy GL, Barker RW,Kocan KM. Isolation of Borrelia burgdorferi(Spirochaetales: Spirochaetaceae) from Ixodesscapularis and Dermacentor albopictus ticks (Acari:Ixodidae) in Oklahoma. J Med Entomol 1992;29:630-3.


Letters

7. Luckhart S, Mullen GR, Wright JC. Etiologic agent ofLyme disease, Borrelia burgdorferi, detected in ticks(Acari: Ixodidae) collected at a focus in Alabama. J MedEntomol 1991;28:652-7.

8. Pung OJ, Banks CW, Jones DN, Krissinger MW.Trypanosoma cruzi in wild raccoons, opossums, andtriatomine bugs in southeast Georgia, USA. J Parasitol1995;81:324-6.

9. Lavender DR, Oliver JH Jr. Ticks (Acari:Ixodidae) inBulloch County, Georgia. J Med Entomol 1996;33:224-31.

10. Spielman A, Wilson ML, Levine JF, Piesman J. Ecologyof Ixodes dammini borne human babesiosis and Lymedisease. Annu Rev Entomol 1985;30:439-60.

11. Rich SM, Caporale DA, Telford SR III, Kocher TD,Hartl DL, Spielman A, et al. Distribution of the Ixodesricinus-like ticks of eastern North America. Proc NatlAcad Sci U S A 1995;92:6284-8.

12. Norris DE, Klompen JSH, Keirans JE, Black WC IV.Population genetics of Ixodes scapularis (Acari:Ixodidae)based on mitochondrial 16S and 12S genes. J MedEntomol 1996;33:78-89.

Ebola/Athens Revisited

To the Editor: After our hypothesis that theplague of Athens (430 B.C.–425 B.C.) could havebeen caused by Ebola virus was published in this

journal (1996;2:155-6), it was brought to ourattention that this hypothesis had been previ-ously entertained.

Gayle D. Scarrow had published a paperentitled “The Athenian Plague: A PossibleDiagnosis” in The Ancient History Bulletin 2.1(1988). Unfortunately, this had not come to ourattention in our literature search, and thereforewe assumed that we were the first to recognizethe possibility. Clearly, Ms. Scarrow deservescredit for suggesting this first. Her argumentsare compelling, even without the support of morerecently available information and the observa-tions advanced in our publication.

We believe an evolving knowledge base (e.g.,the information about the Côte d’Ivoire outbreakwhere a protracted epidemic has been meticu-lously documented) will serve to enhance thecredibility of the Ebola/Athens hypothesis.

P.E. Olson,* A.S. Benenson,† and E.N. Genovese†*U.S. Navy Balboa Hospital, San Diego,

California, USA; †San Diego State University,San Diego, California, USA


News and Notes

Meeting Summaries

The Hot Zone–1997: Conference onEmerging Infectious Diseases

On June 27-28, 1997, the University ofKentucky College of Medicine, University ofCincinnati College of Medicine, and KentuckyAIDS Consortium held a conference for cliniciansand researchers in Lexington, Kentucky. Partici-pants presented the latest findings fromworldwide epidemiologic studies, basic science,and clinical research on emerging infectiousdiseases. The findings indicate that the waragainst infectious diseases is far from over. In theUnited States, between 1980 and 1992, deathsfrom infectious diseases increased by 58%,making serious infections the third leading causeof death; HIV infection is the leading cause ofdeath among 25- to 44-year-olds; and antibioticresistance costs the health-care system anestimated $100 million to $30 billion each year.

With the global population growing from 2.5to 5.8 billion over the last 25 years, large urbancenters throughout the developing world areovercrowded and have inadequate sanitation,ideal for the emergence of infectious diseases. By2025, the global population will reach 8.6 billion.In developing countries, this represents an 84%increase, which will intensify overcrowding inthese areas. In industrialized countries, an agingpopulation base, the advent of immunosuppres-sive medications, and the emergence of HIV arecombining to increase the risk for opportunisticinfection. Moreover, with increased travel,clinicians see increasing numbers of patientswith exotic diseases acquired abroad. Recentmigration of epidemic diphtheria from the formerSoviet Union to Europe and the emergence ofmultidrug-resistant tuberculosis (TB) in theUnited States and elsewhere are but twoexamples of infections resulting from internationaltravel; in addition, nearly 70% of the fruits andvegetables consumed in the United States originatein developing countries; disease outbreaks relatedto imported food frequently go unreported.

Extensive cross-species contact among hu-mans and certain domestic animals can dictateantigenic shifts in influenza viruses. Thelikelihood of the emergence of a new influenzavirus in the near future increases with the growthof the hog population in China. The emergence of

new viruses, such as HIV and filoviruses, indicatesthe virtually unlimited capacity of pathogenicorganisms to mutate and rapidly adapt toenvironmental changes and selective pressures.

HIV/AIDS ResearchRecent data from long-term survivors

support the concept that HIV replication occurswhen the number of CD4+ cells drops below theminimum level required to maintain CD8+ cellcontrol of HIV. CD4+ cell production of IL-2 isneeded for strong cell-mediated immunity.Without CD8+ cell responses, a more virulent,highly cytotoxic viral strain emerges, killinggreater numbers of CD4+ cells and leading toAIDS. Because CD8+ cell loss appears to berelated to a shift from a TH-1- to a TH-2-typecytokine response, therapeutic approaches thatmaintain TH-1 cell response, or enhance CD8+cell anti-HIV activity through factors yet to befully defined, are being actively investigated.Vaccines relying exclusively on antibody re-sponses will almost certainly prove to be oflimited value, while those using CD8+ cell antiviralactivities hold substantially greater promise.

A basic science research forum described theuse of a murine model of immunodeficiency inducedby a type C retrovirus. Because models of this typereproduce a number of the clinical pathophysiologicmanifestations associated with human AIDS, theycan significantly enhance our understanding ofretrovirus-induced immunodeficiency.

The results of recent research involving hostimmune responses to Pneumocystis cariniiinfection indicated that, compared with healthycontrols, HIV patients with less than 200 CD4+cells have similar IL-4 levels but significantlylower peripheral blood mononuclear cell prolif-erative responses and IFN-γ levels to P. cariniimajor surface class glycoprotein (MSG). Centersfor Disease Control and Prevention (CDC) Class 3patients with previous P. carinii pneumonia(PcP) have significantly higher IL-4 (but not IFN-γ) levels than Class 3 patients with no history ofPcP. HIV+ patients who have recovered from PcPhave sufficient memory cells to recognize MSG,but demonstrate a shift from a TH-1- to a TH-2-type antigen recall response.

HIV TherapyThe current state of CD4 and viral load

testing and the 11 drugs available for HIVtreatment were reviewed. Although combination


News and Notes

drug therapy has consistently proved moreeffective than monotherapy in maintainingreduced viral load, the results of a recentlycompleted follow-up study confirmed the effec-tiveness of zidovudine (AZT) in preventingneonatal HIV transmission; in a large cohort ofpregnant women, AZT plus high titer immuno-globulin G was no more effective than AZT plusplacebo, with both regimens producing resultscomparable with those achieved with AZT alone.A number of issues in HIV treatment remainunresolved, including when treatment shouldbegin, how to manage inadequate response totherapy, whether to use HIV resistance genotypingto direct therapy, and how to best deal withprophylaxis for opportunistic infections in patientsshowing dramatic reductions in viral load.

Dengue and Dengue Hemorrhagic Fever (DHF)Tens of millions of cases of dengue and

hundreds of thousands of cases of DHF arereported annually, and more than 2.5 billionpeople are at risk for infection. Factors contributingto the emergence of dengue include unplannedand uncontrolled population growth associatedwith urbanization in tropical regions, lack ofeffective mosquito control, deteriorating watersystems that increase densities of Aedes aegypti,and viral migration among tropical urban centersdue to increasing international air travel. Morethan 50% of all air travel from the United Statesis to tropical destinations, and from 1977 to 1994,2,248 suspected cases of imported dengue werereported in the United States. Because theirclinical symptoms are initially nonspecific,dengue and other arboviral infections can bedifficult to distinguish from other viral, bacterial,and parasitic infections. Correct diagnosisrequires a detailed clinical summary, thoroughepidemiologic information (including recenttravel history), and a diagnostic laboratory test.

The severe hemorrhagic form, DHF/dengueshock syndrome (DSS), has an average incuba-tion period of 4 to 6 days before sudden onset offever and nonspecific signs and symptoms.Because the major pathophysiologic abnormalityobserved in DHF/DSS is increased vascularpermeability and leakage of plasma from thevascular compartment, early fluid replacement iseffective. The geographic distribution of DHF/DSS has been expanding and now can be found intropical areas of Asia, the Pacific, and theAmericas, including Central America and

Mexico. This expansion is associated withincreased movement of dengue viruses byairplane travelers and the development ofhyperendemicity in the Pacific Region and theAmericas. A similar scenario, generated largelyfrom human encroachment into new environ-ments, may also emerge for other Aedes-transmitted illnesses such as yellow fever. Themost cost-effective approach to control dengueand DHF is larval source reduction in disease-endemic areas. Programs should use bothgovernment and community resources to inte-grate environmental sanitation with the use ofinsecticides and biologic controls, targeted tobreeding grounds such as tire dumps.

Other Viral DiseasesThe exponential increase in ecologic change,

both environmental and behavioral, was cited asthe major driving force for the increasing humanrisk for viral infection. Microbial variability canplay a causal role in disease emergence, but itmore often enables viruses to adapt to newcircumstances. Travel of infected humans andinternational transport of microbes and vectorshelp provide the maximum possible microbialevolutionary opportunities in the minimumamount of time.

Viruses have emerged in the past, withmeasles providing a good example of theworrisome potential for future emerging RNAviruses. The emergence of cities in theMesopotamian basin, resulting largely from theadvent of irrigated agriculture, provided apopulated substrate for interhuman transmis-sion of short-incubation, nonlatent viruses.Domestication of livestock likely brought measlesprogenitors into close proximity with humans; aprecursor of rinderpest/peste de petit ruminantsthen made an interspecies leap; much latermeasles spread to Europe, and, in the post-Columbian interchange, to the Americas.

Viral hemorrhagic fevers include thosecaused by Ebola filovirus and hantaviruses.Ebola hemorrhagic fever is characterized byextensive and disseminated infection andnecrosis in major organs, and lymphoid deple-tion. Aerosol transmission of Ebola virus hasoccurred between nonhuman primates andguinea pigs, but no evidence exists forinterhuman transmission by airborne infection.Barrier nursing precautions generally preventthe spread to humans, but in areas having


News and Notes

inadequate medical care facilities, the virus canamplify in humans and cause epidemics.Hantaviruses belong to a single genus in thefamily Bunyaviridae, and each virus infects alimited or unique rodent species with no apparentdisease. Although hemorrhagic fever with renalsyndrome has rarely been diagnosed in theAmericas, hantaviruses from sigmodontine ro-dents cause hantavirus pulmonary syndrome,characterized by large bilateral pleural effusionsand heavy, edematous lungs, interstitial pneu-monitis, and extensive infection of endothelialcells in the pulmonary microvasculature. The firstdocumented interhuman transmission of ahantavirus was an outbreak of 20 cases inPatagonia, with evidence overwhelmingly indicat-ing spread between patients and physicians. Thereasons for, and mechanism of, the spread areunknown, but the registry of U.S. cases was revisedto ensure that this phenomenon was adequatelymonitored. Although early diagnosis and support-ive care are potentially lifesaving in cases ofhantavirus pulmonary syndrome, such efforts areof limited value in Ebola hemorrhagic fever.

Prion IllnessesIn prion illnesses such as Creutzfeldt-Jakob

disease (CJD), risk factors (family history of CJDor dementia, history of poliomyelitis, exposure tosheep or cows) and iatrogenic factors (dura orcorneal grafts and exposure to pooled humangrowth hormone) are important. At onset, CJD iscommonly misdiagnosed as psychotic illness,Alzheimer’s dementia, paraneoplastic syndrome,vascular brain disease, parkinsonism, or evendrug-induced delirium. A variant of CJD, causedby a prion with an altered protein configuration,is bovine spongiform encephalopathy (BSE ormad cow disease). Although the precise mecha-nism of infection originally reported in U.K.cattle is unclear, BSE has been exported to othercountries by feeding cattle inadequately pro-cessed bone meal. The potential for theemergence of BSE in the United States existsbecause similar reservoirs of infection arepresent. Casual handling of beef and beefproducts in processing plants and uncontrolleddisposal of sick cattle may increase that risk.Several cases of CJD have been reported inKentucky patients who consumed squirrelbrains; however, a causal link has not beenestablished. The most recently identified prionillness is a new variant of CJD reported in

England that, unlike sporadic CJD, occurs inmuch younger patients (16 to 50 years of age), canlast longer than 1 year, and is characterized bythe presence of psychiatric and sensory symp-toms and the absence of kuru plaques andelectroencephalogram periodic complexes. Theillness is thought to be linked to BSE.Mathematical modeling suggests that 75,000 to80,000 cases will occur in the foreseeable future.

Tick-Borne DiseasesAs with many other emerging infectious

diseases, concern about tick-borne zoonoticdiseases in the United States has increasedbecause of environmental changes brought on byhuman settlement and socioeconomic factors thatplace humans at greater risk for tick exposure,such as the development of suburban housing indisturbed natural settings. Current conditionsappear favorable for continued increases invector tick populations and their geographicexpansion and for increasing interaction betweenticks and humans. In the United States, theprevalence of Rocky Mountain spotted fever maycontinue to increase as the urbanization of thewestern and southern regions expands opportuni-ties for human exposure to tick-borne pathogens.

The epidemiology of Lyme disease andtularemia was reviewed, and the concept of asouthern (U.S.) tick-associated rash illness(STARI), the epidemiology of which is still beingdefined, was introduced. STARI is characterizedby an expanding erythematous rash resemblingthat of Lyme disease, mild or absent constitu-tional symptoms, and no well-described sequelae.The rash responds well to antibiotic treatment.STARI is seen in the range of the human-bitingLone Star tick (Amblyomma americanum) and isfrequently associated with a Lone Star tick bite.Studies indicate that this infection is not causedby Borrelia burgdorferi or other known tick-borne agents. Amplified segments of the genomeof this spirochete have recently been described, andsome investigators have proposed that it representsa new species, B. lonestari. It has been suggestedthat STARI may be closely related or identical to acommensal spirochete of deer, B. theileri.

The need for early diagnosis and treatment ofRocky Mountain spotted fever was stressed bypresenting data from 94 infected patients inNorth Carolina. Death rates were significantlylower (6.5% vs. 22.9%, p < 0.03) in patientsreceiving appropriate treatment within 5 days of


News and Notes

onset of illness than in patients who receiveddelayed or no treatment. Predictors of deathincluded renal failure; elevated serum creatinine,aspartate transaminase, and bilirubin; decreasedserum sodium; thrombocytopenia; neurologicinvolvement; and male gender.

Parasitic DiseasesThe role of the cell surface glycoconjugate

lipophosphoglycan (LPG) in the survival ofLeishmania parasites whose life cycle alternatesbetween intracellular parasitism and extracellu-lar life in sand fly vectors was explored. Dataindicating that LPG is a multifactorial moleculemay lead to a biochemical rationale for LPG-targeted chemotherapeutic regimens.

Isolation of a full-length genomic clone of theacid α-mannosidase from an epimastigote genomiclibrary of Trypanosoma cruzi was reported.Sequence analysis showed a single open readingframe encoding the α-mannosidase gene. Be-cause the size of this frame is consistent with thatof lysosomal mannosidases in humans, theseresults could lead to the exploration of newchemotherapeutic options in Chagas disease.

A bank of 5,200 insertion mutants has beenused to characterize the parasitic mechanismsused by Legionella pneumophila. Results fromtransmission electronmicroscopy indicate that L.pneumophila has acquired genetic loci specific forsurvival and replication within mammalian cells,allowing evolution from a protozoan parasite intoits present disease-causing agent. Alternatively,ecologic coevolution of L. pneumophila toparasite protozoa has led to the development ofmultiple redundant mechanisms, some of whichdo not function within mammalian macrophages.

TBPoverty, changing immigration patterns, and

the emergence of HIV disease were cited asfactors contributing to an attenuated rate ofdecrease of TB incidence in the United Statessince the late 1980s and its increase as a majorglobal problem. Poorly managed TB controlprograms, suboptimal access to health care, aninadequate physician knowledge base, and poorpatient compliance have combined to increase theincidence of TB and, especially, multidrug-resistant TB; sensitivity testing is critical in themanagement of resistant TB. Updated CDCguidelines for interpreting the purified proteinderivative skin test call for responses to be

considered positive if induration is greater than 5mm in those with HIV or who have recent TBcontact, and greater than 10 mm in foreign-bornpatients, intravenous drug users, the homeless,and immunocompromised patients. Because ofthe “boosting” phenomenon, two-step purifiedprotein derivative skin testing is now recom-mended for health-care workers and nursinghome patients who are retested periodically.Treatment of active TB should use multipledrugs, avoid adding single agents, and includecompliance monitoring, preferably directly ob-served therapy. Workplace prevention measuresshould be driven by a consistently high index ofsuspicion and should include appropriate isolationof suspected cases, work-ups following exposure,rigorous reporting to health departments, andlocating recalcitrant patients. Effective control ofTB will require social, political, and culturalchanges, as well as medical innovation.

PlagueThe epidemiology, pathophysiology, and

treatment of plague, as well as the improvementof diagnostic techniques for infections caused byYersinia pestis, were reviewed. Such new testswill permit the public health laboratory toquickly identify a plague outbreak and apply theappropriate control measures to limit its spread.

The pathogenesis of plague was explored, andtwo separate, but essential, iron transportsystems were identified in Y. pestis. The first, theyersinabactin (Ybt) system, enables the organismto proliferate from the site of an open wound orbite, while the second, identified as Yfe, is used byYbt mutants to obtain iron during infection ofinternal organs. The Ybt iron transport systemappears to be essential for growth in the earlystages of bubonic plague infection, while the Yfesystem functions to allow growth during, or after,infection of internal organs. Both systems arerequired for Y. pestis to be fully virulent.

A novel mechanism by which a Y. pestisvirulence protein is sequestered in, andtransported within, host cells was described.When yersiniae contact a eukaryotic cell, asignaling event activates the expression andsecretion of a set of four toxins (Yops) and causestheir vectorial translocation into the cell cytoplasm.Three of the Yops derange cell signaling andcytoskeletal functions by kinase, tyrosine phos-phatase, and actin depolymerization activities.Although no activity or intracellular target has


News and Notes

been identified for the fourth known translocatedYop, YopM, immunoblot analysis and laserscanning confocal microscopy demonstrated thatmost YopM is vectorially translocated into HeLacells or the macrophagelike cell line J774 byadherent Y. pestis and travels to the eukaryoticcell nucleus. Because a growing number ofimportant human pathogens (e.g., Salmonella,Shigella, and Pseudomonas) have similar, butless well-studied, secretion/translocation mecha-nisms and putative secreted toxins, these findingswill facilitate studies that ultimately could lead tonovel therapies for these agents.

Antibiotic ResistanceThe current development of staphylococci

resistant to methicillin or fluoroquinolones andgram-negative bacilli resistant to extended-spectrum beta-lactams are but the most recentlyrecognized patterns of antibiotic resistance.General approaches for modifying these trendsinclude 1) source control, particularlyhandwashing, and the need to wear gloves duringcontacts with all patients, 2) improved antibioticuse and control, 3) improved infection controldevices, and 4) better use of pathophysiology andimmunologic modulation.

The growing problem of vancomycin-resis-tant enterococci (VRE), with an incidence of 20%to 40% in some groups of U.S. hospitalizedpatients, necessitates maximal use of all theseapproaches. Skin proliferation of VRE producesextensive environmental contamination that mayrequire universal use of gloves to control outbreaksor hyperendemic disease. In addition, vancomycinshould be limited to treatment of beta-lactamresistant gram-positive bacteria (such as methicil-lin-resistant Staphylococcus aureus and gram-positive bacteria in beta-lactam allergic patients)and Clostridium difficile (only after metronidazolefailure) and to endocarditis prophylaxis. Vancomy-cin should be avoided in routine surgicalprophylaxis, empiric treatment of febrile neutrope-nia with negative cultures, and pneumoniaprophylaxis in the intensive care unit. A number ofexperimental peptides and other agents (such asquinupristin-dalfopristin) under investigation astreatments for VRE infections were identified.

Future trends in resistance may includefurther spread of vancomycin-resistant staphy-lococci (already reported in Japan), quinolone- or

carbapenem-resistant gram-negative bacilli, andtreatment-resistant viruses. Seventeen isolatesof methicillin-resistant S. aureus with uniquegenotypes were studied to determine rates ofresistance to the fluoroquinolones ciprofloxacinand levofloxacin. The mean single-step resis-tance to 4 x MIC ciprofloxacin was 1.05 x 10-5 andto levofloxacin was 4.03 x 10-6. When seriallypassaged in increasing antibiotic concentrations,the geometric mean MICs for ciprofloxacin andlevofloxacin increased 3.0 ± 1.5 times and 1.8 ±1.4 times, respectively (p < 0.0005). Only fourstrains became resistant to levofloxacin, but eightbecame resistant to ciprofloxacin, indicating thatciprofloxacin selects methicillin-resistant S. aureusmore frequently than levofloxacin.

Other TopicsThe laboratory evidence supporting the role

of Chlamydia pneumoniae in the development ofatherosclerosis was reviewed. Recent dataindicate that infection of vascular endothelialcells with C. pneumoniae is associated with theproduction of chemokines and adhesion mol-ecules that promote transendothelial migrationof neutrophils and monocytes. These findingssuggest that immunopathogenic responses to C.pneumoniae infection may contribute to thedevelopment of clogging deposits.

Data were presented demonstrating theproliferation of human CD4+ T cells fromunexposed persons in response to in vitro exposureto Toxoplasma gondii. Further studies showed thatthis proliferative response depends on HLA-DRmolecules and requires processing of Tgantigens. In contrast to typical exogenoussuperantigens, analysis of TCR Vß expressionafter stimulation with Tg did not show apattern of preferential increase of a specificTCR Vß-bearing subpopulation. αßT cellssecreted significant amounts of IFN-γ afterincubation with Tg-infected monocytes. Thisprocess may play an important role in the earlyevents of the immune response to T. gondii.

Richard N. Greenberg,* Judith E. Feinberg,†and Claire Pomeroy*

*University of Kentucky Medical Center, Lexington,Kentucky, USA; †University of Cincinnati College of

Medicine, Cincinnati, Ohio, USA


News and Notes

International Meeting on Borreliosis,Prague, Czech Republic

Approximately 150 participants from 10countries gathered in Prague, the Czech Republic,August 27-29, 1997, to discuss research topicsrelated to the theme of the meeting, “LymeBorreliosis—Basic Science and Clinical Ap-proaches.” The meeting was organized by theNational Institute of Public Health (Centre ofEpidemiology and Microbiology); the World HealthOrganization Collaborating Center for Referenceand Research on Borreliosis; the Second School ofMedicine, Charles University (Prague); and theCzech Medical Association J.E. Purkyn�. Meetingsessions focused on topics including epidemiology,clinical treatment, dermatology, diagnosis andtreatment, neurology, and laboratory diagnosis.

The session on epidemiology presentedsurveillance data on the incidence of Lymeborreliosis (LB) in the Czech Republic (incidencerates were 61.8/100,000 in 1995 and 41.2/100,000in 1996) and in Slovakia, Austria, and Slovenia.Data underscored the high risk for transmission ofLB in central and eastern Europe. The results ofvaccine trials using the recombinant outersurface protein (Osp)A antigen of Borreliaburgdorferi were also presented; more detailedstudies are needed to examine intraspeciesvariability of OspA antigens in Europe.

The session on clinical approaches andtreatment reviewed research conducted in theUnited States and discussed the diagnosticimportance of organism-specific biologic mark-ers, e.g., Borrelia-specific antigens or DNA, aswell as pleocytosis in cerebrospinal or synovialfluid. Experience with the diagnosis and treatmentof LB in the hyperendemic-disease regions of westBohemia underscored the importance of accuratediagnosis in avoiding overtreatment.

The use of nonhuman primates as models forstudying neuroborreliosis was examined in thesession on neurology. Problems related to thediagnosis and treatment of chronic disease, andtheir economic consequences, were identified.Several methods to assist clinicians in making acorrect diagnosis were presented and discussed.The persistence of B. burgdorferi DNA in patientswith Lyme arthritis was considered in therheumatology session. Ultrastructural evidencefor the intracellular location of B. burgdorferi insynovium also was presented.

The session on laboratory diagnosis focusedon the genomic sequence of the linear chromo-some of B. burgdorferi (B31 strain) and thecrystal structure of OspA; both apply to thelaboratory diagnosis of LB. Other studiesaffirmed the importance of standardizing diag-nostic methods to ensure reproducibility anduniformity of the results from different laborato-ries. The influence of certain in vivo–expressedantigens (virulence antigens) on invasivenessand the ability of B. burgdorferi to adapt to thehost environment were noted. Other topics werethe sensitivity and reproducibility of polymerasechain reaction and the importance of the primersselected for the assay.

The studies presented in the poster sessionaddressed a wide array of themes: among them,epidemiology and population awareness, reactiv-ity of B. burgdorferi antigens in immunoblotprocedures when specimens derived fromhumans or animals are used, and incidence ofticks and their association with disease indifferent regions.

The importance of apoptosis in the morphol-ogy of LB, the role of Langerhans cells in the skinreactions, and the role of integrin CR3 in theinteraction of B. burgdorferi with host cells werediscussed. The sensitivity and the selection of theprimers used for polymerase chain reaction todetect B. burgdorferi in ticks were considered.Aspects of vector biology and ecology wereinvestigated (e.g., habitats, the tick as LB’s majorvector, vector capacity). Other diseases transmit-ted by Ixodes ricinus ticks in Europe (e.g., tick-borne encephalitis, babesiosis, ehrlichiosis) as wellas human ehrlichiosis in Europe were reviewed.

Dagmar Hulínská and JiÍí BaštaNational Institute of Public Health,

Prague, Czech Republic

Workshop on Climate Change and Vector-Borne and Other Infectious Diseases

Climate changes may affect human healththrough a myriad of pathways; of particularinterest are pathways affecting the geographicranges and incidence of vector- and water-bornediseases. As society chooses how to deal withprojections of long-term climate change, decisionsmust be based on scientific knowledge. A 2-day


News and Notes

workshop1 was convened in September 1997 todiscuss what is known about the relationshipbetween projected climate changes and theincidence of water-borne diseases (e.g., cholera)and vector-borne diseases, including those typicallyconsidered tropical (malaria, dengue fever, yellowfever, and schistosomiasis), plus subtropical ortemperate-zone diseases whose vectors are likely tobe affected by projected climate changes.

The workshop participants discussed thesystems involved in potential climate changes,from the global ocean-atmosphere-landmass sys-tem that drives climate to the regional ecologic andhuman socioeconomic systems where diseasedynamics occur. These systems are extremelycomplex, as are the interactions among them, whichunderscores the need for more research beforeaccurate projections can be made. Major researchgaps were identified, and an agenda was framedfor a sound scientific basis for public policydebates and decisions. The proposed agendaincluded the following items: climate modeling;ecosystem and habitat dynamics; disease surveil-lance; technologies for disease prevention andmitigation; disease transmission dynamics; datasets for empirical studies; integrated assessments;and detecting, understanding, and responding tounexpected events. Further discussion and imple-mentation of this research agenda is encouraged. Asummary of the workshop is available from theElectric Power Research Institute, TR-109516,EPRI Distribution Center, 207 Coggins Drive,P.O. Box 23205, Pleasant Hill, CA 94523;Telephone: 510-934-4212.

1The workshop was commissioned by the Electric PowerResearch Institute, with additional sponsorship from theDepartment of Energy, the National Institute of Allergy andInfectious Disease, the National Institute of EnvironmentalHealth Sciences, and the National Aeronautics and SpaceAdministration. The workshop was organized and conductedby the Washington Advisory Group. The 28 participantsincluded representatives from agencies and institutions thatconduct or fund research and experts in the fields ofclimatology and global climate modeling, public health, andthe biology and ecology of vectors, pathogens, and theecosystems they inhabit.

The Fourth International Conferenceon HFRS and Hantaviruses

Atlanta, Georgia, USA, March 5–7, 1998

The Centers for Disease Control andPrevention and other cosponsors will host theFourth International Conference on HFRSand Hantaviruses to facilitate the exchangeof scientific information in the followingareas: 1) clinical aspects, 2) laboratorydiagnostics, 3) pathogenesis and immuneresponse, 4) hantavirus ecology, 5) hantavirusepidemiology, 6) molecular biology and cellinteractions, 7) health education and preven-tion, and 8) antiviral and vaccine develop-ment. The meeting will offer plenary sessionswith invited speakers, as well as oral andposter sessions based on accepted abstracts.

For further information, contact AmyCorneli, Centers for Disease Control andPrevention, 1600 Clifton Road, MS A26, Atlanta,GA 30333, USA; fax: 404-639-1509; e-mail:[email protected]; URL: http://www.cdc.gov/ncidod/diseases/hanta/hantconf.htm.

Third InternationalCongress onTropical Neurology

November 30–December 2,1998

Organized by the Groupe Francophone d’Etudeet de Recherche en Neurologie Tropicale, the ThirdInternational Congress on Tropical Neurologywill convene in Fort de France, Martinique, fromNovember 30 to December 2, 1998. The four mainthemes of the congress are central nervoussystem inflammatory, neurodegenerative, epi-leptic, and cerebrovascular disorders in tropicalenvironments; however, presentations on otherthemes are welcome.

A symposium on epilepsy in tropical zones willbe held during the congress.

For additional information, contact ProfessorM. Dumas (phone: 33-5-55-43-58-20, fax: 33-5-55-43-58-21) or Professor J.C. Vernant (phone: 33-5-96-55-22-61, fax: 33-5-96-75-45-90).


News and Notes

International Conference on EmergingInfectious DiseasesMarch 8–11, 1998Atlanta Marriott Marquis Hotel

Late-breaker abstract submission deadline:January 30, 1998

Information on abstract submission, confer-ence registration, and exhibits can be obtained atwww.asmusa.org, by sending an e-mail messageto [email protected], or by calling202-942-9248. Proceedings of the conference willbe published in the journal Emerging InfectiousDiseases.

Registration: limited to 2,500 - register NOW!Preliminary program information is available

at http://www.cdc.gov/ncidod/EID/98conf.htm.


List of Reviewers (1995-1997)

R. AhmedJ. AlexanderJ. AllanB. AndersonP. AndersonF. AnguloR. ArbeiD. ArcherL. AshA. AzadJ. BakkenW.R. BallouT. BarrettC. Barry, IIIM. BarryJ. BartlettJ. BasemanJ. BassL. BeatiB. BeatyE. BelayD. BellW. BelliniJ. BenachJ. BennettE. BergerR. BerkelmanG. BirkheadA. L. BisnoP. BlakeM. BlaserU. BleckerJ. BoaseS. BonhoefferP. BoyleP. Brachman, Jr.M. BrandtE. BreitschwerdtD. BrilesR. BrodersonW. BrogdonP. BrownR. BrubakerJ. BullT. BurkotJ. ButlerR. ByersC. CalisherR. CampbellM. CartterA. CasadevallG. CassellM. ChamberlandL. ChapmanP. CharacheJ. ChesneyJ. ChildsM. ChristensenB. ChomelJ. ClarridgeT. ClearyB. ClineB. ColeM. CoyleA. Cravioto

J. CrawfordS. CzinnG. DaschJ. DawsonS. DebanneK. DeCockD. DennisE. DidierB. DietzscholdD. DohmM. DolanT. DonderoH.D. DonnellJ. DuchinA. DufourJ.S. DumlerM. EberhartM. EberhardI. EckardtD. EnriaP. EpsteinS. EwingJ. EzzelA. FanningM. FarleyP. FarnhamC.P. FarringtonD. FishD. FishbeinI. FisherT. FolksD. FranzF. FrostE. GangarosaL. GarciaR. GaynesM. GerberF. GherardiniM. GilmoreH. GinzburgW. GlezenR. GoeringB. GoldD. GoldbergT. GomezK. GreenD. GriffinD. GublerR. GuerrantG. Gurri GlassB. GushulakT. HackstadtJ. HadlerR. HajjehL. HallS. HalsteadM. HamburgD. HancockC. HanlonL. HarrisonL. HedstromD. HeymannD.A. HendersonL. HerwaldtJ. Hilliard

D. HillisB. HjelleD. HollisA. HolmesR. HolzmanD. HooperJ. HoranR. Horsburgh, Jr.T. HrabaW. HuestonM. Hugh-JonesD. HuxsollP. JahrlingW. JakubowskiR. JamisonJ.M. JandaE. JanoffM. JenkinsP. JensenW. JohnsonD. JuranekK. KafadarJ. KaperM. KarmaliD. KaslowL. KatzA. KaufmannT. KiehnD. KirschnerU. KitronD. KlauckeH. KoprowskiB. KorberT. KozelR. KrauseN. KumarC.C. KuoG. LandM. LandryJ. LaryM. LaytonJ. LeDucD. LeibyB. LevinM. LevineR. LevinsH. LevyJ. LevyJ. LewisP. LindenH. LipmanM.D. LittleH. LobelL. MagnarelliJ. MaguireB. MahyT. MarrieR. MassungC. McDonaldJ. McDougalJ. McGowan, Jr.M. McNeilJ. McNichollG. McQuillanA. Meijer

J. MelnickM. MillerD. MindellB. MishuC. MitchellJ. MittlerC. MoeJ. ModlinR. MohlerT. MonathS. MonroeM. MontecalvoC. MooreP. MooreJ.A. MorrisS. MorrisD. MorseS. MorseM. MufsonU. MunderlohL. MundyF. MurphyB. MurrayC. MurrayG. MyersI. NachamkinT. NashJ. NataroN. NathansonT. NavinR. NelsonS. NicholC. NicholsR. NicolaD. NoahP. NuoriT. NutmanS. OaksS. O’BrienJ. OlsonE. OttesenM. ParmentierJ. ParryA. PaviaP. PellettJ. PennerD. PerrottaD. PersingC. J. PetersM.J. PickettN. PieniazekS. PittsB. PlikaytisJ. PlouffeM. RayfieldR. RegneryL. ReichmanE. ReissD. RelmanL. RiceR. Rico-HesseB. RimaH. RobinsonJ. RoseR. Rosenberg

D. RossC. RupprechtM. SalvatoA. SanchezS. SavarinoC. SchableJ. SchachterG. SchadP. SchantzC. SchmaljohnM. SchwartzD. ShahA. ShepherdR. ShopeD. ShroyerP. SimoneE. SkameneJ. SmithJ. SommerfeldD. SonenshineR. SpiegelJ. SpikaA. SpielmanK. SpitalnyL. StammP. StrebelG.T. StricklandN. StrockbineD. StroupSummersgillB. SwaminathanP. TarrF. TenoverR. TeshS. TeutschS. ThackerD. ThayerM. TippleJ. TokarsL. TompkinsM. TonerT. TsaiV. TsangJ. van EmbdenS. WatermanH. WatsonD. WalkerA. WandelerS. WeaverR. WelchJ. WengerR. WeyantD. WhippleC. WhitneyG. WidmerR. WirtzK. WiseC. Wisseman, Jr.M. WongG. WormserM. Zervos

Emerging Infectious Diseases thanks the following reviewers for their support through thoughtful, thorough,and timely reviews from the journal’s inception to the present. We apologize for any names we may have omittedinadvertently.

Instructions to Authors

Manuscript PreparationThe journal complies with the “Uniform Requirements for

Manuscripts Submitted to Biomedical Journals” (Ann Int Med1997:126[1]36-47) (http://www.acponline.org/journals/resource/unifreqr.htm).

Begin each of the following sections on a new page and in thisorder: title page, abstract, text, acknowledgments, references,tables, figure legends, and figures.

Title page. Give complete information about each author (i.e.,full name, graduate degree(s), affiliation, and the name of theinstitution in which the work was done). Also provide address forcorrespondence (include fax number and e-mail address).

Abstract and key words. Avoid citing references in the abstract.Include up to 10 key words; use terms listed in the Medical SubjectHeadings from Index Medicus (http://www.nlm.nih.gov/tsd/serials/lji.html).

Text. Double-space everything, including the title page,abstract, references, tables, and figure legends. Type only on oneside of the paper and number all pages, beginning with the title page.Indent paragraphs 5 spaces; leave no extra space betweenparagraphs. After a period, leave only one space before beginning thenext sentence. Use Courier font size 10 and ragged right margins.Italicize (rather than underline) scientific names when needed.

Electronic formats. For word processing, use WordPerfect or MSWord. For graphics, use Corel Draw, Harvard Graphics, Freelance,PowerPoint, .TIF (TIFF), .GIF (CompuServe), .WMF (WindowsMetafile), .EPS (Encapsulated Postscript), or .CGM (ComputerGraphics Metafile). The preferred font for graphics files is Helvetica.Convert Macintosh files into one of the suggested formats. Submitslides or photographs in glossy, camera-ready photographic prints.

References. Follow the Uniform Requirements style. Placereference numbers in parentheses, not in superscripts. Numbercitations in order of appearance (including in text, figures, and tables).Cite personal communications, unpublished data, and manuscriptsin preparation or submitted for publication in parentheses in text.Consult List of Journals Indexed in Index Medicus for acceptedjournal abbreviations; if a journal is not listed, spell out the journaltitle in full. List the first six authors followed by “et al.”

Tables and figures. Create tables within the word processingprogram’s table feature (not columns and tabs within the wordprocessing program). For figures, use color as needed; send slides,photographs, or prints. Figures, symbols, lettering, and numberingshould be clear and large enough to remain legible when reduced.Place figure keys within the figure.

Units of measurement. Consult Uniform Requirements.Abbreviations. Use abbreviations sparingly. Spell out a term the

first time it is used.

Manuscript SubmissionInclude a cover letter verifying that the final manuscript has

been seen and approved by all authors.Submit three copies of the original manuscript with three sets

of original figures and an electronic copy (on diskette or by e-mail) tothe Editor, Emerging Infectious Diseases, Centers for DiseaseControl and Prevention, 1600 Clifton Rd., MS C-12, Atlanta, GA30333, USA; e-mail [email protected].

Types of Articles

Perspectives: Articles in this section should provide insightfulanalysis and commentary about new and reemerging infectiousdiseases or related issues. Perspectives may also address factorsknown to influence the emergence of diseases, including microbialadaption and change; human demographics and behavior;technology and industry; economic development and land use;international travel and commerce; and the breakdown of publichealth measures. Articles should be approximately 3,500 words andshould include references, not to exceed 40. Use of additionalsubheadings in the main body of the text is recommended. If detailedmethods are included, a separate section on experimentalprocedures should immediately follow the body of the text.Photographs and illustrations are encouraged. Provide a shortabstract (150 words) and a brief biographical sketch.

Synopses: This section comprises concise reviews of infectiousdiseases or closely related topics. Preference is given to reviews ofnew and emerging diseases; however, timely updates of otherdiseases or topics are also welcome. Synopses should beapproximately 3,500 words and should include references, not toexceed 40. Use of subheadings in the main body of the text isrecommended. If detailed methods are included, a separate sectionon experimental procedures should immediately follow the body ofthe text. Photographs and illustrations are encouraged. Provide ashort abstract (150 words) and a brief biographical sketch.

Dispatches: These brief articles are updates on infectious diseasetrends and research. The articles include descriptions of newmethods for detecting, characterizing, or subtyping new orreemerging pathogens. Developments in antimicrobial drugs,vaccines, or infectious disease prevention or elimination programsare appropriate. Case reports are also welcome. Dispatches (1,000 to1,500 words of text) need not be divided into sections. Provide ashort abstract (50 words); references, not to exceed 10; figures orillustrations, not to exceed two; and a brief biographical sketch.

Letters: This section includes letters that give preliminary data orcomment on published articles. Letters (500 to 1,000 words of text)should not be divided into sections, nor should they contain figuresor tables. References (not more than 10) may be included.

Editorial Policy and Call for Articles

Emerging Infectious Diseases is a peer-reviewed journal established expressly to promote the recognition of new andreemerging infectious diseases around the world and improve the understanding of factors involved in disease emergence,prevention, and elimination.

The journal has an international scope and is intended for professionals in infectious diseases and related sciences. Wewelcome contributions from infectious disease specialists in academia, industry, clinical practice, and public health, as wellas from specialists in economics, demography, sociology, and other disciplines. Inquiries about the suitability of proposedarticles may be directed to the Editor at 404-639-3967 (tel), 404-639-3075 (fax), or [email protected] (e-mail).

Emerging Infectious Diseases is published in English and features three types of articles: Perspectives, Synopses, andDispatches. The purpose and requirements of each type of article are described in detail below. To expedite publication ofinformation, we post journal articles on the Internet as soon as they are cleared and edited.

Spanish translations of some articles can be accessed at ftp://fcv.medvet.unlp.edu.ar/pub/EID. Articles by authors fromnon-English-speaking countries can be made simultaneously available in English and in the author’s native language(electronic version of the journal only).