determined to be 80% and oral ribavirin had been found asan effective treatment for the patients with CCHF.4 A studyfrom Turkey also had showed the efficacy of oral ribavirinfor reducing case-fatality rate in severe cases with CCHF.2

However, in another study from Turkey, the duration ofhospitalization, the need for blood and blood products andfatality rate had not been found different for the ribavirinand control groups.5

In conclusion, the efficacy of ribavirin in the treat-ment of CCHF is not clear. In this study, i.v. ribavirintherapy was found ineffective for the survival rate of thepatients with severe CCHF. Literature reveals noprospective randomized, controlled clinical trials of oral/i.v. ribavirin against CCHF. We need more clinical trialswith higher number of patients to evaluate the efficacyof i.v. ribavirin on the survival rate of patients affectedby CCHF.

Acknowledgements

There are no funding and conflict of interest. Ethicalapproval is not required.

References

1. Centers for Disease Control and Prevention. Viral hemorrhagicfever: initial management of suspected and confirmed cases.MMWR Morb Mortal Wkly Rep 1983;32(Suppl. 2):27Se38S.

2. Ergonul O, Celikbas A, Dokuzoguz B, Eren S, Baykam N,Esener H. Characteristics of patients with CrimeaneCongohemorrhagic fever in a recent outbreak in Turkey and impactof oral ribavirin therapy. Clin Infect Dis 2004 Jul 15;39(2):284e7.

3. Bakir M, Ugurlu M, Dokuzoguz B, Bodur H, Tasyaran MA,Vahaboglu HTurkish CCHF Study Group. CrimeaneCongo hae-morrhagic fever outbreak in Middle Anatolia: a multicentrestudy of clinical features and outcome measures. J Med Micro-biol 2005 Apr;54(Pt 4):385e9.

4. Mardani M, Jahromi MK, Naieni KH, Zeinali M. The efficacy oforal ribavirin in the treatment of CrimeaneCongo hemorrhagicfever in Iran. Clin Infect Dis 2003 Jun 15;36(12):1613e8.

5. Ozkurt Z, Kiki I, Erol S, Erdem F, Yilmaz N, Parlak M, et al.CrimeaneCongo hemorrhagic fever in Eastern Turkey: clinicalfeatures, risk factors and efficacy of ribavirin therapy. J Infect2006 Mar;52(3):207e15.

6. World Health Organization. Communicable disease profile Iraq,http://www.who.int; 2003. p. 20e3.

7. Watts DM, Ussery MA, Nash D, Peters CJ. Inhibition of Crimeane

Congo hemorrhagic fever viral infectivity yields in vitro byribavirin. Am J Trop Med Hyg 1989 Nov;41(5):581e5.

8. Tignor GH, Hanham CA. Ribavirin efficacy in an in vivo model ofCrimeaneCongo hemorrhagic fever virus (CCHF) infection.Antiviral Res 1993 Dec;22(4):309e25.

Mustafa Aydın CevikAnkara Numune Education and Research Hospital,

Infectious Diseases and Clinical Microbiology Clinic,Samanpazari, 06100 Ankara, Turkey

Nazif ElaldıSivas Cumhuriyet University Hospital, Infectious Diseases

and Clinical Microbiology Clinic, Turkey

Esragul Akıncı*Pınar Onguru

Aysxe ErbayAnkara Numune Education and Research Hospital,

Infectious Diseases and Clinical Microbiology Clinic,Samanpazari, 06100 Ankara, Turkey

*Corresponding author. Tel.: þ90 3122878183;fax: þ90 3123103030.

E-mail address: esragulakinci@yahoo.com (E. Akıncı)

Turan BuzganRamazan Uzun

Ministry of Health, Turkey

Ayhan KubarGulhane Military Medical Academia, Turkey

Hurrem BodurAnkara Numune Education and Research Hospital,

Infectious Diseases and Clinical Microbiology Clinic,Samanpazari, 06100 Ankara, Turkey

Accepted 10 July 2008

Available online 22 August 2008

0163-4453/$34 Crown Copyright ª 2008 Published by Elsevier Ltdon behalf of The British Infection Society. All rights reserved.doi:10.1016/j.jinf.2008.07.007

Letters to the Editor 351

Assessment of EpsteineBarr virus (EBV) serostatus byenzyme immunoassays: Plausibility of the isolatedEBNA-1 IgG positive serological profile

Enzyme immunoassay (EIA) technology is widely used forroutine evaluation of Epstein-Barr virus (EBV) immunologicalstatus, yet the ‘‘gold standard’’ technique is the indirectimmunofluorescence assay (IFA).1 Testing for the presenceof IgG and IgM antibodies to viral capsid antigens (VCA), andIgG antibodies to EBV nuclear antigens (EBNA), particularlyEBNA-1, is generally sufficient to discriminate accuratelybetween primary and past infections, and to demonstratesusceptibility to EBV infection. The interpretation of EBVserology in immunocompetent individuals, as determined byEIA, is not always straightforward. In fact, the occurrence ofthe so called undetermined, unresolved, or implausibleserologic profiles is not uncommon.1,2 Among these, thecombination: VCA IgG negative/VCA IgM negative/EBNA-1IgG positive - classically considered implausible - is observedregularly at our laboratory. The present study was aimed atassessing the frequency of such serological pattern in oursetting, and determining its biological plausibility.

A total of 1612 sera were submitted to our laboratoryduring 2007 for physician-ordered EBV serological testing.The sera were initially analyzed by the following EIAsmanufactured by Trinity-Biotech (Bray, Ireland): Captia�VCA IgG/IgM, and Captia� EBNA-1 IgG, which use animmunogenic peptide of p18 and a full-length recombinantEBNA-1 protein (p72), respectively as the antigen. Eighty-seven sera (5.3%) from 87 patients (47 males and 40females; mean age of 7 years; range, 0-66 years) had only

352 Letters to the Editor

detectable EBNA-1 IgG antibodies. These sera had beensent either for the diagnosis of primary EBV infection(patients had fever, subclinic hepatitis, febrile exan-thema, or mononucleosis syndrome) (n Z 61), or forroutine pre-transplant (renal and hematopoietic) serolog-ical screening for EBV infection (both from donors andrecipients) (n Z 26). All these sera were subsequentlyanalyzed by a marketed EBNA anticomplement immuno-fluorescence assay-ACIF- (Merifluor ENV-NA ACIF, fromMeridian Biosciences, Cincinatti, Ohio, USA), to determinewhether the patients were truly infected by EBV or not.Thirty-six sera (41.3%) were positive by the ACIF assay.The sera were further analyzed by two other commerciallyavailable EBNA-1 IgG EIAs, which were manufactured byBiotest (Biotest, Dreieich, Germany) and DiaSorin (Dia-Sorin, Salugia, Italy), and by a VCA IgG/IgM assay fromDiaSorin. The Biotest assay uses a recombinant EBNA-1protein lacking the glycine-alanine repeat sequence,known to occasionally cross-react with cellular proteins,as the antigen. The DiaSorin EBNA-1 IgG and VCA IgG/IgMassays employ a synthetic peptide of p72, and a syntheticpeptide of p18, respectively, as the antigen. Data areshown in Table 1. Forty-four sera (50.5%) tested positive inthe Biotest EBNA-1 IgG assay (31 of these sera werepositive in the ACIF assay), and 34 sera (39.0%) werepositive in the DiaSorin EBNA-1 assay (27 of these serawere positive in the ACIF assay). The three EBNA-1 IgGassays yielded concordant results in 32 sera (36.7%), 26 ofwhich were positive in the ACIF assay. Fourteen sera(16.0%) tested positive in the VCA IgG assay from DiaSorin(11 were ACIF positive). Twenty-six out of the 87 seratested with the EIAs from DiaSorin gave the serologicalprofile VCA IgG negative/VCA IgM negative/EBNA-1 IgGpositive.

A number of sera (n Z 13) were tested in a Lineimmunoblot assay (EBV Line, Genzyme Virotech GmbH,Russelsheim, Germany), which detects IgG antibodiesagainst p72, gp125 (VCA), a synthetic peptide of p18, andEA-D (early-antigens diffuse pattern). Six sera reactedexclusively with EBNA-1 (3 of which were negative in theACIF test). The remaining sera reacted with EBNA-1 andgp125.

Commercially available EBV EIAs differ considerably interms of sensitivity and specificity.1e4 Our data fullysupport this assumption. In a previous study,2 a frequencyof unresolved EBV serological patterns of 10.1% was

Table 1 Performance of the EBNA-1 IgG EIAs from Biotestand DiaSorin

ACIF EBNA-1 IgG VCA IgG

Biotest DiaSorin DiaSorin

Pos Neg Pos Neg Pos Neg

Pos (n Z 36) 31 5 27 9 11 25Neg (n Z 51) 13 38 7 44 3 48

ACIF, anticomplement immunofluorescence; EBNA, Epsteine

Barr nuclear antigens; and VCA, Viral capsid antigens.The VCA IgG EIA from DiaSorin, and the ACIF assay on sera yieldingthe serological profile: VCA IgG/IgM negative and EBNA-1 IgGpositive, as determined by EIAs from Trinity Biotech.

reported when using a combination of several EIAs fromdifferent manufacturers, and of 7.7% when employinga multiplexed bead assay. In our setting, the mostfrequently found unresolved serological pattern is thecombination of VCA IgG/IgM negative and EBNA-1 IgGpositive, whose plausibility has been traditionally ques-tioned. According to our results, almost half of patientsexhibiting this antibody profile were truly infected (pastinfection with undetectable VCA IgG antibodies) by EBV,as determined by the reference EBNA ACIF assay. In ourexperience, either EIA analysis of follow-up sera frompatients displaying this antibody pattern or performanceof line immunoblot assays often do not contributeto resolve the EBV immunological status, so that theperformance of the EBNA ACIF assay is warranted.The sera subject to analysis were selected on the basisof the results in the EIAs from Trinity Biotech; conse-quently, the overall prevalence of the referred serologicalprofile in EIAs from the different manufacturers could notbe compared. Our data, however, show that EIAs fromDiaSorin frequently yield this antibody pattern as well. Insummary, our study underscores the fact that the findingof an isolated EBNA-1 IgG serological profile must not beregarded as necessarily implausible. Rather, it shouldprompt to carry out the EBNA ACIF assay in order toelucidate whether the patient is latently infected by EBVor not. This extent is of critical relevance in patientsbeing evaluated for solid organ or hematopoietic trans-plantation, as susceptibility to EBV is a major risk factorfor the development of postransplant lymphoproliferativedisorders.

References

1. Hess RD. Routine EpsteineBarr virus diagnostics from thelaboratory perspective: still challenging after 35 years. J ClinMicrobiol 2004;42(8):3381e7.

2. Klutts JS, Liao S, Dunne WM, Gronowski AM. Evaluationof a multipliexed bead assay for assessment of EpsteineBarr virus immunologic status. J Clin Microbiol 2004;42(11):4996e5000.

3. Bruu AL, Hjetland R, Holter E, Mortensen L, Natas O,Petterson W, et al. Evaluation of 12 commercial tests fordetection of EpsteineBarr virus-specific and heterophile anti-bodies. Clin Diagn Lab Immunol 2000;7(3):451e6.

4. Gartner BC, Hess RD, Bandt D, Kruse A, Rethwilm A, Roemer K,et al. Evaluation of four commercially available EpsteineBarrvirus enzyme immunoassays with an immunofluorescence assayas the reference method. Clin Diagn Lab Immunol 2003;10(1):78e82.

Tomas GarcıaNuria Tormo

Microbiology Service, Hospital Clınico Universitario,Valencia, Spain

Concepcion GimenoDavid Navarro*

Microbiology Service, Hospital Clınico Universitario,Valencia, Spain

Department of Microbiology, School of Medicine,University of Valencia, Valencia, Spain

Patient 1

Patient 1 is a 45-year-old Dutch male with diabetes mellituswho presented with general unwell being and pain in hisright shoulder. His complaints had existed for nearly twomonths.

The medical history reported recurrent K. pneumoniaeurinary tract infection complicated by septicemia in 1999and 2005. In 2005 echographic examination of the livershowed two small lesions, probably abscesses.

On physical examination he had no fever and the shouldershowed no abnormalities. Radiographic studies of theshoulder and thorax were normal. Echography of the liver

Letters to the Editor 353

*Corresponding author. Microbiology Service, HospitalClınico Universitario, Av. Blasco Ibanez 17, 46010 Valencia,

Spain. Tel.: þ34 (96)3864657; fax: þ34 (96)3864173.E-mail address: david.navarro@uv.es (D. Navarro)

Accepted 31 July 2008

Available online 11 September 2008

0163-4453/$34 ª 2008 The British Infection Society. Published byElsevier Ltd. All rights reserved.doi:10.1016/j.jinf.2008.07.017

Liver abscess due to Klebsiella pneumoniae showingthe mucoid phenotype

In recent years Klebsiella pneumoniae of the mucoidphenotype has emerged as a cause of invasive infections. InAsia, especially in Taiwan, these often result in liverabscess, whereas in South Africa they mainly result inpneumonia.1e3

In Western Europe and North America only very fewcases of such liver abscesses have been reported.4,5 Herewe report two patients from the Netherlands with a liverabscess from which K. pneumoniae, respectively ofserotype K1 and K2, was cultured. Both patients suffer fromdiabetes mellitus.

however showed a lesion consistent with an abscess. Cultureof pus obtained from this lesion resulted in K. pneumoniae.Blood cultures and a urine culture remained sterile.

The liver abscess was drained and the patient wastreated with oral ciprofloxacin, which was continued forthree months after discharge. The patient showedcomplete recovery. Colonoscopy, MR angiography of thebiliary tract and X-rays of the dental region did not showthe cause of this recurrent infection with K. pneumoniae.

Patient 2

Patient 2 is a 59-year-old Dutch male who was admittedbecause of general unwell being and fever for 10 days.Apart from a fever of 39 �C (102 �F) no abnormalities were

Figure 1 PFGE profiles of Klebsiella serotype K1 and K2 isolates from several countries, among which strains NL1 and NL2 (UK:United Kingdom, TW: Taiwan, AS: Australia, HK: Hong Kong, IS: Israel).