1004

farms where there were infected and excreting animals, orthat contamination of the environment was caused by wildbirds or other animals. It is difficult to imagine how wildbirds or animals could selectively contaminate one area ofNewport. It is far more likely that C. burneti was brought outof farms on dust, straw, and manure, either on farm vehiclesor by farm animals using the roads.4 Presumably infectionwould result from inhalation of infected dust particles.The features of this outbreak do not indicate that one farm

premises was the source of infection. Farming contact whichcould be ascertained from the patients involved at least tenestablishments. A survey of farms was attempted but provedto be unsuccessful. There were no reports of Q feveroccurring amongst farm animals during the outbreak but C.burneti is almost always a symptomless infection of cattle andsheep. Vast numbers of organisms may be excreted in faeces,urine, and products of parturition by apparently healthyanimals. The main group of patients in this outbreak hadbeen exposed to infection between the last week of May andthe middle of June. This was well after lambing had finishedbut just before the winter barley was harvested, and so farmsmay have been short of straw. Local reports suggested that

large quantities of straw had been moved through the areaduring this period. Furthermore, when visiting patients inone area of Newport, we saw the roadsides strewn with loosestraw which had fallen from farm vehicles. Contamination ofstraw bales may have occurred during storage or when theywere used for pens during lambing. The difficultiesencountered when pursuing these possibilities within thefarming community were disappointing and precluded a fullinvestigation of the source of this zoonosis.

We thank all patients, employers, and staff who helped us in the survey; andMr R. A. Carson, assistant director of environmental health and Mr J. B.Williams, director of environmental health, Newport, for their invaluablesupport.

Correspondence should be addressed to S. R. P., Communicable DiseaseSurveillance Centre, 61 Colindale Avenue, London NW9 5EQ.

REFERENCES

1. Harman JB. Q fever in Great Britain. Clinical account of eight cases. Lancet 1949; ii;1028-30.

2. Marmion BP, Stoker MGP. The epidemiology of Q fever in Great Britain. An analysis ofthe findings and some conclusions. Br Med J 1958; ii: 809-16.

3. Evans AD, Baird TT. An interim account of an autumnal outbreak of Q fever in Cardiff.Proc Roy Soc Med 1959; 52: 616-20.

4. Lennette EH, Clark WH. Observations on the epidemiology of Q fever in NorthernCalifornia. JAMA 1951; 145: 306-09.

LABORATORY OUTBREAK OF Q FEVERACQUIRED FROM SHEEP

CHRISTOPHER J. HALL SHIRLEY J. RICHMONDE. OWEN CAUL

Public Health Laboratory, Bristol

NORMAN H. PEARCE

Safety Office, University of Bristol

IAN A. SILVER

Department of Pathology, University of Bristol

Summary In April and May, 1981, an outbreak of Qfever occurred in a university department

which used gravid sheep for fetal respiratory studies. Duringthe subsequent investigation 91 people were studied and 28were found to have complement-fixing antibodies to thephase II antigen of Coxiella burneti in their serum. Symptomscompatible with C. burneti infection occurred in 14 of theseropositive patients. The majority of symptomaticinfections could be related to the delivery of twin fetuses inone laboratory, but some people remote from this incidentwho were exposed to sheep in other parts of the building alsoshowed evidence of infection. Those people with serologicalevidence of recent infection were treated with tetracycline for4 weeks. Defects were apparent in the facilities being used forthe research, and the risk of Q fever infection should be bornein mind by groups engaged in work with pregnant sheep.

INTRODUCTION

THERE have been several reports of human infection withCoxiella burneti acquired from sheep in research institutionsin the U.S.A.’-4 We report here a laboratory outbreak of Qfever in England, associated with the use of gravid sheep forfetal respiratory studies.

OUTBREAK AND INVESTIGATION

On May 8, 1981, Q fever was diagnosed in a worker at theUniversity of Bristol on the basis ofa complement-fixing (CF)

antibody titre of 1500 against C. burneti phase II antigen in aserum sample taken 12 days after the onset of a "flu-like"illness. He worked in a laboratory in which experimentalprocedures were performed on gravid ewes; the laboratorywas on the first floor of a two-storey building. 6 other workersin that laboratory could recall pyrexial illnesses during April,and sera from these and 2 symptomless colleagues all had CFantibody titres of 64 or more.Since it has been recommended that all acute cases of Q

fever should be treated with an antibiotic to reduce thelikelihood of chronic infection,5 blood samples, clinical

histories, and epidemiological information were obtainedfrom all people who between mid-March and mid-April werein parts of the building that contained sheep. This

investigation was carried out during the latter half of May andearly June. When possible, second serum samples were takenapproximately 2 weeks after the first from all people whowere seropositive, and acute and convalescent sera were takenfrom anyone in whom symptoms compatible with Q feverdeveloped during the investigation. Two gravid ewes housedin a side-room of the implicated laboratory were destroyed onMay 21, and sera and tissue samples were taken for

microbiological investigations. The records of sheepmovements and laboratory procedures were scrutinised. Thelaboratory and associated side-room were fumigated withformaldehyde gas on May 22.

Serological results and reasons for the investigation weregiven by the medical officer for environmental health to thegeneral practitioner of each person tested. Unless

contraindicated, oxytetracycline, 500 mg 8 hourly for 4weeks, was recommended for all people showing evidence ofrecent infection-i.e., seroconversion or a stable CF antibodytitre of more than 32. This level of antibody was chosenbecause seroconversion to a stable convalescent titre of only64 was demonstrated in paired sera from a sheep handler withsymptoms.

SerologyThe complement fixation test was used to measure

antibody to C. burneti phase II antigen; any reactive sera (titre8) were subsequently tested for CF antibodies to the phase Iantigen. All sera were heat inactivated before testing by the

1005

standard four-volume overnight complement fixation test inmicrotitre plates. Titres were expressed as reciprocals of thedilutions giving 50% fixation. Antigens were provided by thedivision of ’microbiological reagents and quality control,Public Health Laboratory Service, Colindale.

RESULTS

Human Data

Five groups of people could have been exposed to C. burnetiinfection, either through contact with the laboratory (F69) onthe first floor or through contact with animal facilities on theground floor. The serological and clinical evidence for

infection in these groups is given in table 1. Of the 91 peopleTABLE I-SEROLOGICAL AND CLINICAL EVIDENCE OF

COXIELLA BURNETI INFECTION AMONG 91 PEOPLE AT RISKr---

*CF antibody titre 8 against C. burneti phase II antigen.f CF antibody titre <8 against C. burneti phase II antigen.

TABLE II-DETAILS OF ILLNESS AND ANTIBODY TITRES IN THE

28 SEROPOSITIVE PEOPLE

A=symptomless.*Symptoms: F=fever; LR=Iower respiratory; UR=upper respiratory;M=myatgia; S=sweats; H=headache; T=tiredness.tPeople involved with experiments with gravid sheep.

-Seronegative (<8) in October, 1980.Human-to-human transmission possible.Infection may have been acquired from subject’s own sheep.

tested 28 were seropositive (titre>8) (tables I & II). 3

symptomless seropositive people had stable titres32, whichwere regarded as indicating previous rather than recentinfection. Serum antibody titres and symptoms in the

seropositive cases are shown in table II. Symptoms varied, butthe usual findings were acute onset of fever and chills withsweating, myalgia, and headache. Most symptoms resolvedwithin 7-10 days, although malaise often persisted forseveral weeks. Symptoms recalled by people who wereseronegative (titre<8) were generally of a mild upperrespiratory nature (colds and sore throats).All sera tested for the presence of CF antibody against

phase I antigen were negative (titre<8).Workers in laboratory F69.-4 of the 5 people directly

involved with the sheep experiments were seropositive, and 3of these 4 people had been ill. A serum sample had been takenduring the acute phase of illness from 1 of these workers andseroconversion was demonstrated, whilst another

seropositive person with symptoms had been seronegative inOctober, 1980. 9 other people used the same laboratory butwere not directly involved with the sheep experiments. 5 ofthese people were seropositive, and 4 of these 5 had been ill.All illness occurred between April 19 and 24, those actuallyinvolved with sheep experiments becoming ill a few daysbefore the other workers.

Visitors to F69. -The 3 seropositive people in this groupwere regular visitors to the laboratory, whereas the 8

seronegative visitors visited only occasionally. 1 visitor, whobecame ill on May 23 and subsequently seroconverted, wasalso closely associated domestically with a symptomaticseropositive sheep worker who had been ill 1 month earlier,so case-to-case spread may have occurred in this instance.Other workers on the first floor. -The 3 seropositive people

in this group all worked in a room (F71) which adjoined andwas the main route of entry into F69. The 5 workers whowere seronegative worked in rooms further away from theimplicated laboratory.

Ground floor animal facility users. -9 of 2 people who hadhad little or no association with F69 showed serologicalevidence of infection. 3 of these seropositive people wereanimal attendants in contact with both gravid and non-gravidsheep and other animals. 1 person in this group complained ofsymptoms on May 25 and subsequently seroconverted, butshe also kept her own lambing flock, which may have been thesource of her infection.

Ancillary staff.-4 of 35 domestic, maintenance, andportering staff investigated were found to be seropositive,only 1 of whom regularly visited F69. 1 of the seropositivepeople had visited the building only once, on April 4, when hepassed through some ground floor rooms, whilst another hadonly been in the building on April 1, 25, and 26.

Sheep Data

Gravid sheep used during 1981 had been purchased inSeptember, 1980, from a farm in South Wales. They weremixed with the general flock at the University farm, and sentto the university in batches of two or three. They spentseveral days in mixed animal housing on the ground floor ofthe veterinary school, and when required each ewe waswalked on a halter along corridors and up two flights of stairsto F69. During March and April, 1981, eight sheep had beenstudied in F69. After an anaesthetic, the uterus was openedand cannulae and electrical leads were inserted to allow fetal

monitoring. After surgery the sheep were kept in a side-roombut were returned to the laboratory for observations. When

1006

labour began the animals were killed and taken to the

necropsy room on the ground floor where the fetuses wereremoved and returned to F69. However, an unusual

procedure was used on April 3 when, after their intrauterinedeath, twin fetuses were surgically removed from one sheep(no. 993) in F69.Two gravid sheep present in F69 at the onset of this

investigation were killed and examined at the VeterinaryInvestigation Laboratory, Langford. No CF antibody tophase II antigen was detected in the sheeps’ sera but theirplacental tissue contained C. burneti, demonstrated byseroconversion of guinea pigs inoculated with this material.

DISCUSSION

An outbreak of Q fever associated with contamination oflaboratory F69 with C. burneti occurred in late April and earlyMay. The contamination was probably responsible for theinfections in 15 people, of whom 11 (75%) were ill. Surgicaldelivery of twin fetuses from sheep 993 was the most likelyevent to have caused this contamination, although othersurgical procedures performed on sheep in F69 early in Aprilmay also have been relevant.

13 people who had little or no association with F69 alsoshowed serological evidence of infection. These peopleprobably acquired infection from contaminated animal

quarters on the ground floor rather than from F69, but sinceonly 3 of them (23%) were ill, it is impossible to be certain ofthe exact date or source of the majority of these infections. Allinfections may have been acquired from organisms excretedby sheep 993 before transfer to F69; the one infection in lateMay being acquired either from organisms persisting in theenvironment or from the patient’s own sheep. Alternatively,since two sheep from the same source as sheep 993 wereshown post mortem to be infected with C. burneti, it is

possible that these and other experimental sheep had alsobeen excreting organisms whilst on the ground floor and thatthey were the source of some infections.Several defects were apparent in the protocol used for the

laboratory procedures conducted on pregnant sheep;rectification of the defects according to recommendationsmade after similar outbreaksl-4 might have reduced thenumber of infections, although it would probably not haveprevented them totally. The main defects were: the use of ageneral laboratory for surgical procedures rather than a

_ properly equipped and ventilated operating theatre; the largenumber of people in the vicinity of the experiments yet notinvolved with them; and the general lack of awareness ofinfections which can be acquired from sheep. The use of onlysheep free of C. burneti may be difficult to achieve since

investigation showed that CF antibodies cannot always bedemonstrated in infected animals.We had difficulty in deciding whether the presence of

antibodies in symptomless people represented recent orprevious infection. Great variation in primary CF antibodyresponses occur in acute Q fever,6 and these antibodies maybe detectable for 2 years or more after infection.6 Thereforetitres we regarded as indicating recent infection may haverepresented past infection and vice versa. A method of

detecting specific IgM has recently been described,’ and weare assessing the value of this technique in resolving thesedifficulties.The decision to treat recently infected people with

oxytetracycline was made upon the recommendation ofBrownS and, although such action has not been shown toprevent chronic Q fever, we thought it a sensible precaution

to take, bearing in mind the difficulties experienced intreating this rare but serious complication. Except forresidual malaise no symptoms lasted longer than 2 weeks, andthe only person admitted to hospital was a diabetic requiringstabilisation.Other groups engaged in research with pregnant sheep

should be reminded of the risk of acquiring Q fever.We thank Dr S. K. R. Clarke for her advice during the investigation and her

help in the preparation of this manuscript; Dr R. Buttery of Avon Area HealthAuthority for his assistance in the investigation; Dr A. Goodship for hisassistance, particularly in the tracing of sheep movements; the VeterinaryInvestigation Centre, Langford, and the Central Veterinary Laboratory,Weybridge, for providing details and results of their sheep investigations; andall members of the University of Bristol who cooperated in this investigation.

Correspondence should be addressed to C. J. H., Public Health Laboratory,Royal United Hospital, Combe Park, Bath BA1 3NG.

REFERENCES

1. Schachter J, Sung M, Meyer KF. Potential danger of Q fever in a university hospitalenvironment. J Infect Dis 1971; 123: 301-04.

2. Curet LB, Paust JC. Transmission of Q fever from experimental sheep to laboratorypersonnel. Am J Obstet Gynecol 1972; 114: 566-68.

3. Center for Disease Control. Q fever at a university research center-California. MorbidMortal Wkly Rep 1979; 28: 333-34.

4. Meiklejohn G, Reimer LG, Graves PS, Helmick C. Cryptic epidemic of Q fever in amedical school. J Infect Dis 1981; 144: 107-13.

5. Brown GL. Q fever. Br Med J 1973; ii: 41-43.6. Murphy AM, Field PR. The persistence of complement fixing antibodies to Q fever

(Coxiella burneti) after infection. Med J Aust 1970; 1: 1148-50.7. Murphy AM, Magro L. IgM globulin response in Q fever (Coxiella burneti) infections.

Pathology 1980; 12: 391-96.

Personal Paper

CHEMOTHERAPY FROM AN INSIDER’SPERSPECTIVE

KENNETH H. COHN

Harvard Surgical Service/New England Deaconess Hospital,Boston, Massachusetts, U.S.A.

THIS paper is based on my experience of receivingchemotherapy from November, 1980, to July, 1981. I waswell until Oct. 15, 1980, when I discovered a 3 cm mass in myleft mid-neck. Apart from intermittent pruritus and anintertriginous fungus infection at the end of August, 1980, Ihad had no symptoms. I had not noticed any masses beforeOct. 15, and I had not lost weight or experienced nightsweats. A sonogram showed the mass to be cystic. With thepreoperative diagnosis of cystic hygroma, I underwent acervical exploration on Oct. 20, 1980, and was informed inthe recovery room that the frozen section was malignant.Because of a normal bone-marrow aspiration, as well as anegative computerised tomographic scan of the abdomen andpelvis, I was designated stage IA, diffuse undifferentiatedlymphoma. After sperm-banking had been completed I

began, on Nov. 17, 1980, a 37-week course of chemotherapy(including bleomycin, adriamycin, cyclophosphamide,oncovin, dexamethasone, methotrexate, and citrovorum).

PROBLEMS CONFRONTING PHYSICIANS WHO BECOME

PATIENTS -

The abrupt change from being a surgical resident to

becoming a patient gave rise to six problems:1. Reaction of Health-care Workers to their own MortalityThe first difficulty stemmed from the realisation by the

health-care team that people of their own age and profession