Molecular epidemiology of a hepatitis C virus outbreak in a ...

Nephrol Dial Transplant (1999) 14: 1188–1194

Nephrology Dialysis Transplantation

Original Article

Molecular epidemiology of a hepatitis C virus outbreak in a haemodialysis unit Antigoni Katsoulidou1, Dimitrios Paraskevis1, Victoria Kalapothaki1, Dimitrios Arvanitis2, Peter Karayiannis3, Valsamakis Hadjiconstantinou2 and Angelos Hatzakis1, for the Multicentre Haemodialysis Cohort Study on Viral Hepatitis* 1Department of Hygiene and Epidemiology, Athens University Medical School, Athens, Greece, 2Renal Unit, A. Fleming Hospital, Athens, Greece, and 3Department of Medicine, St Mary’s Hospital, London University Medical School, London, UK

Abstract Background. Haemodialysis patients are at high risk of infection by hepatitis C virus. The aim of this study was to investigate a hepatitis C virus outbreak which occurred in a haemodialysis unit, using epidemiological and molecular methods. Methods. Five seroconversions to hepatitis C virus antibody (anti-HCV ) were observed over a 6 month period and these were added to the four previously recorded anti-HCV-positive patients. All nine patients involved in the outbreak were tested for HCV RNA by reverse transcription-polymerase chain reaction and hepatitis C genotype determination was accomplished by a reverse hybridization assay. Furthermore, part of the NS5 region of hepatitis C genome (nucleotide positions 7904–8304) was amplified and sequenced in all HCV RNA-positive patients. Then, phylogenetic analysis of the nucleotide sequences obtained was carried out in order to investigate any possible epidemiological linkage among patients. Detailed epidemiological records were also available for all haemodialysis patients. Results. Samples from all five incident cases and three out of four prevalent HCV infections were found positive for HCV RNA. HCV genotyping studies revealed that all incident cases were classified as 4c/d, whereas one and two prevalent cases were 1a and 4c/d respectively. Sequence comparisons and phylogenetic Correspondence and offprint requests to: A. Hatzakis, MD, PhD, Department of Hygiene and Epidemiology Athens University Medical School, 75 M. Asias Street, 11527 Athens, Greece. *Investigators and Institutions in the Multicentre Hemodialysis Cohort Study on Viral Hepatitis: A. Hatzakis, A. Katsoulidou, M. Psichogiou, E. Vaindirli (Athens University Medical School, Athens); G. Vosnidis, J. Boletis, D. Stamatiadis (Laiko General Hospital, Athens); S. Voudiclari, Ch. Priftis, M. Koulousiou (Hippokration General Hospital, Athens); S. Moutafis, K. Ifandi, I. Thanou ( Kianous Stavros Clinic, Athens); G. Skoutelis, K. Triandafilou, M. Matzaropoulou (Athinaiki Clinic, Athens); V. Hadjiconstantinou, D. Arvanitis, E. Logothetis (A. Fleming Hospital, Athens).

tree analysis revealed that six of the patients harboured very similar strains and clustered together, including all incident and one prevalent case, which was implicated as index case. Further epidemiological analysis was consistent with patient to patient transmission. Conclusions. Molecular and epidemiological analysis suggested that horizontal nosocomial patient to patient transmission was the most likely explanation for the virus spread within the haemodialysis unit under study. Key words: genotyping, haemodialysis, hepatitis C virus, nosocomial infection, nucleic acid sequencing, phylogenetic analysis

Introduction Haemodialysis patients are at high risk of acquiring hepatitis C infection since the prevalence of anti-HCV in this group is exceptionally high. Initial studies carried out with first generation enzyme-linked immunosorbent assays reported an HCV seroprevalence ranging from 2 to 47% [1–9]. With the introduction of second generation assays an almost two-fold increase in the prevalence of anti-HCV has been demonstrated [10–13]. A number of factors account for the increased risk of HCV infection in haemodialysis patients. These include blood transfusions [4,11,14–15] and duration of haemodialysis. Moreover, the presence of anti-HCV has been often documented in non-transfused haemodialysis patients suggestive of nosocomial transmission of HCV [2,5,7,8,16–20]. However, studies using molecular epidemiology methods provided conflicting results [21–26 ] and it has been postulated that in addition to patient to patient transmission, other recognized (blood products such as albumin, immunoglobulins) or unrecognized routes may be involved [25]. In the present study we describe an outbreak of

© 1999 European Renal Association–European Dialysis and Transplant Association

Molecular study of a hepatitis C virus outbreak

1189

HCV infection in a haemodialysis unit. Molecular analysis of viral strains in association with epidemiological data reconstructed the plausible routes of transmission and provided evidence for a patient to patient nosocomial infection.

Subjects and methods Haemodialysis procedure At the time of the outbreak, 40 patients were regularly dialysed, on one of the two shifts per day (either morning or evening shift), 3 times per week on either a MondayWednesday-Friday or a Tuesday-Thursday-Saturday schedule. Thus, every dialysis machine was used by two persons per day. Patients were normally dialysed on the same shift and on the same console although exceptions could occur and these were recorded. The dialysis unit employed one large room with 10 dialysis machines (consoles no. 1–10) and a small room with one dialysis machine (console no. 11). Consoles no. 1–9 were used by all anti-HCV-positive and -negative patients, who were not separated, and shared the same dialysis machines. One dialysis machine was dedicated to patients in critical condition (console no. 10) and another machine in the small room (console no. 11) to HBsAg-positive patients (Figure 1). Nurses worked in two shifts per day and had to take care of three patients per shift, but they could also move from patient to patient if needed. Haemodialysis was carried out using Gambro dialysis machines (Ab, Sweden, AK-10) and bicarbonate as dialysate. All patients were receiving heparin and dialysers were never reused. Machines were heat disinfected between treatments and chemically every month. AK-10 machines do not allow for the dialysis procedure to start if the disinfection cycle has not been completed.

Data collection The renal dialysis unit started its function in the beginning of 1989. Screening of anti-HCV antibodies and monitoring of alanine aminotransferase (ALT ) was part of the haemodialysis unit routine since December 1989. By May 1993 an intensive serological follow-up study was initiated in all haemodialysed patients as part of the Multicenter Haemodialysis Cohort Study on Viral Hepatitis. Blood samples were collected every 2 weeks from anti-HCVnegative patients and every 4 weeks from anti-HCV-positive ones. Blood was centrifuged within 2 h from collection to achieve optimal conditions for HCV RNA detection. ALT levels were recorded and multiple serum aliquots from each bleed were kept at −70°C until further use. Medical records were kept constantly for all patients as part of the haemodialysis unit routine. These records included data such as symptoms of hepatitis, history of liver enzyme abnormalities, medication history, past medical history and history of transfusions. Furthermore, the dialysis schedule (day and shift), seating arrangement (Figure 1), bleeding episodes, staff-patient assignments, work schedules, dialysis equipment maintenance as well as infection control practices were all recorded. Demographic and risk factors data including information on intravenous drug use, household contact with any hepatitis case, occupational exposure to human blood and temporary dialysis in other centres, were also available.

Fig. 1. Schematic diagram of the haemodialysis unit: dialysis machines (1–11), seating arrangements of study patients, shifts (A: Monday-Wednesday-Friday morning, B:Monday-WednesdayFriday evening, C: Tuesday-Thursday-Saturday morning, D: Tuesday-Thursday-Saturday evening).

Hepatitis C serology Routine detection of anti-HCV begun in December 1989, when the first generation enzyme immunoassay (ELISA) became available. A second generation immunoassay was adopted in June 1991 (Abbott’s Labs, Wiesbaden, Germany). All anti-HCV positive patients included in the study were confirmed by a second or third generation recombinant immunoblot assay (Riba-2; HCV-2.0 SIA or Riba-3; HCV3.0 SIA, Chiron Corp., Emeryville, CA, USA).

Hepatitis C virus RNA quantification, extraction, amplification and genotyping The detection of HCV RNA was carried out with the Amplicor HCV Test (Roche Diagnostic Systems Inc., Branchburg, NJ, USA) [27]. For the quantification of HCV RNA QuantiplexTM HCV RNA 2.0 Assay (bDNA), (Chiron Dignostics), was used. The determination of HCV genotypes was done on amplicons (generated by the Amplicor HCV Test), using a commercially available assay (Inno-Lipa HCV II, Innogenetics N.V, Zwijndrecht, Belgium). The last is a reverse amplification assay based on the hybridization of the

1190

A. Katsouldou et al.

Fig. 2. Rooted phylogenetic tree analysis based on the NS5 sequence alignments. For the reference sequences used in phylogenetic analysis, Genbank accession numbers, their sequence names as well as the genotype they belong to, are the following: genotype 1a: M67463 (HPCCGAA), genotype 1b: D10750 (HPCJ491), genotype 2a: D00944 (HPCPOLP), genotype 2b: D10988 (HPCJ8G), genotype 3a: D28917 (HPCHK6), genotype 3b: D10080 (HPCT9), genotype 4: L36437 (HPCNS5BA), genotype 4c: L29605 (HPCGB215N ), L29614 (HPCGB48NS), L29607 (HPCGB358N ), L29602 (HPCGB116N ), L38376 (HPCFRAU ), L44603 (HCNL81I ) genotype 4e: L29590 (HPCCAMNS5), L29626, (HPCGB8093), L29627 (HPCGB8091), genotype 4f: L38370 (HPCFRAO), L29596 (HPCG22NS1), L29593 (HPCG22NS5), genotype 4g: L29618 (HPCGB549N ), L29621 (HPCGB5491), genotype 4h: L29613 (HPCGB4381), L29611 (HPCGB438N ).

amplified sequence from the 5∞-non coding region of HCV with oligonucleotide probes immobilized as parallel lines on membrane strips [28]. The probes correspond to the sequence of the six more common HCV genotypes and their subtypes. The nomenclature system proposed by Simmonds et al. was used [29]. To obtain sequences from the NS5 region of the viral genome, HCV RNA extraction, reverse transcription and amplification were done as follows: RNA was extracted from

100 ml of serum or plasma using a mono-phasic solution of phenol and guanidine isothiocyanate (TrizolTM LS Reagent, Gibco BRL, Grand Island, NY, USA), according to manufacturers instructions. The final product was dissolved in 20 ml of RNase-free water. Reverse transcription was done using the Gene Amp RNA PCR Kit (Perkin Elmer Cetus, Branchburg, NJ, USA), in a total volume of 20 ml with 50 units of MuLV reverse transcriptase and random hexamers. The PCR reaction was carried out under standard conditions,


using 2.5 units of Taq polymerase (AmpliTaq, Perkin Elmer Cetus, Branchburg, NJ, USA), in a total volume of 100 ml on a DNA Thermal Cycler (Perkin Elmer Cetus 2400). The thermal profile of the reaction was as follows: initial denaturation at 94°C for 2 min; 40 cycles with denaturation at 94°C for 15 s, annealing at 45°C for 30 s and polymerization at 72°C for 30 s. In the PCR reaction and for the specific amplification of NS5 sequences, primers thought to be highly conserved amongst different variants of HCV were used [30]. The sequence of PCR primers were 243: 5∞-TGGG GATCCCGTATGATACCCGCTGCTTTGA-3∞ (sense) and 242: 5∞-GGCGGAA TTCCTGGTCATAGCCTCCGTGAA3∞ (anti-sense).The PCR product is a 400 bp fragment between HCV RNA nucleotide positions 7904 and 8304 based on the numbering system of Choo et al. [31,32]. The amplified products were electrophoresed through a 1.7% agarose gel and visualized by ethidium bromide staining. It should be noted that PCR reactions were carried out under strict conditions [33], in order to avoid contamination and false positive results. Furthermore, in every PCR run two reagent controls, two positive and two negative samples were also included.

Sequencing and phylogenetic analysis of viral strains Amplified DNA was purified by the GeniePrep DNA isolation kit and directly subjected to double strand sequencing with dye-labelled dideoxy terminators. The ABI PRISM 377 automated sequencer (Applied Biosystems) was used for electrophoresis and data collection. DNA sequences were edited and assembled using Sequencer software (Macintosh release 3.0). Clustal V was used for multiple sequence alignment of DNA sequences under default conditions [34]. Reference sequences for the various subtypes were obtained from the GenBank database. Phylogenetic analysis of NS5 sequences was carried out in order to confirm the genotyping performed by Inno-Lipa and in addition, to investigate any possible epidemiological linkage among HCV strains from the haemodialysis patients. All the NS5 sequences were compared with each other and with selected HCV sequences of genotypes 1, 2, 3 and 4 from GenBank. Phylogenetic tree analysis was done by the neighbour-joining method using Kimura’s two parameter correction, as implemented in the PHYLIP package [35]. To further confirm the reliability of the phylogenetic tree, BOOTSTRAPPING (1000 replicates) was accomplished.

Results Six patients who had been on haemodialysis were found positive for anti-HCV with 2nd generation HCV assays (June 1991). During 1992 three of these died, and one other patient seroconverted to anti-HCV (July 1992). No further seroconversions were observed until an outbreak of HCV infection was recorded between May and October 1993, when five new seroconversions to anti-HCV were noted ( Table 1). At that time, the prevalence of anti-HCV in the haemodialysis unit was found to be 22.5%, which is much higher than the one reported for the general population (1.0% in our region). After October 1993, no more seroconversions occurred till the end of follow-up (October 1997). A more intensive follow-up started in May 1993 with weekly or biweekly collection of samples when all

1191 Table 1. Date of the first anti-HCV positive sample, HCV RNA and genotype status Patient

anti-HCV (+)

HCV RNA

HCV genotype

HD HD HD HD HD HD HD HD HD

12/89 12/89 6/91 7/92 5/93 6/93 7/93 8/93 10/93

+ − + + + + + + +

1a − 4c/d 4c/d 4c/d 4c/d 4c/d 4c/d 4c/d

808 824 819 836 807 806 804 813 816

patients were tested for HCV RNA and ALT level. At that time, eight patients of the haemodialysis unit were positive for HCV RNA. Three of them were prevalent cases of hepatitis C documented by anti-HCV antibodies, whereas the five were incident cases that seroconverted to anti-HCV between May and October 1993. All the other patients of the unit were negative for HCV RNA and did not develop anti-HCV antibodies until the end of the study. The five new cases of hepatitis C concerned patients who had never been transfused or had no transfusions of blood or blood products in the 14 months preceding seroconversion. Moreover, none of them had any other identifiable risk factor for HCV infection. It should be also noted that all members of the staff were regularly checked for HCV antibodies and were repeatedly found negative until the end of the study. The seroconverters had no obvious clinical symptoms and in two of them (HD 813 and HD 816) no ALT elevation was observed during the whole study period. Genotyping using primers from the 5∞-non-coding region (Inno-Lipa) revealed that seven out of the eight HCV RNA-positive patients belonged to subtype 4c/d (discrimination between subtypes 4c and 4d with InnoLipa method is not possible), including all the new seroconversions, whereas one patient was genotyped 1a ( Table 1). Sequences in the NS5 region of the viral genome were obtained with primers 242, 243 for all the HCV RNA positive patients of the study. In all cases, the first available serum sample with documented viraemia for hepatitis C was selected for sequence analysis (no serum samples were available before May 1993). Two other HCV RNA-positive patients, completely unrelated to the haemodialysis unit (Co1 and Co2), which were classified as 1b and 4 respectively by Inno-Lipa, were used as controls. Phylogenetic tree analysis confirmed the HCV genotyping results obtained by Inno-Lipa for all the patients and controls, except for patient HD 819 who did not fall within the 4c clade, although it clearly clusters with type 4 isolates, as indicated by high bootstrap values (99%). In addition, genotyping by sequencing the NS5 region for the second control patient (Co2) was more precise than Inno-Lipa since it was clustered with the 4c group.

1192

The very close relation between the six haemodialysis patients (HD 836, HD 807, HD 806, HD 804, HD 813, HD 816) which formed a distinct cluster, as well as the high bootstrap values (1000 of 1000), clearly points towards a common source of infection. Four out of six patients (HD 804, HD 816, HD 836, HD 813) displayed identical nucleotide sequences, and the two remaining (HD 806, HD 807) had one nucleotide substitution each. Phylogenetic analysis also revealed that (i) patient HD 819 was not associated with the outbreak and (ii) the index case was HD 836 who seroconverted to anti-HCV in July 1992. It should be also noted that in May 1993 HD 836 presented HCV RNA levels 1.13×106 Equivalents/ml and had ALT levels 104 IU/ml.

Discussion There is much evidence supporting the nosocomial transmission of HCV infection among the haemodialysis patients of this study: (i) the existence of five seroconversions, which took place over a 6 month period and was observed in non-transfused patients, (ii) the finding of a single hepatitis C virus genotype (type 4) in all the patients of the outbreak, (iii) the extremely high prevalence of genotype 4 among the anti-HCV positive patients of the haemodialysis unit (77.8%), a genotype which is relatively rare in Greece (5%, in chronic hepatitis C patients, unpublished data) and (iv) the close relationship of the NS5 amplified sequences between the six haemodialysis patients as shown by phylogenetic tree analysis. That nosocomial transmission was implicated, became apparent when molecular epidemiological studies were undertaken which included sequencing of the NS5 region of the HCV genome. The NS5 region has been used before for analogous studies [25,36 ]. Its suitability for phylogenetic analysis relies on the fact that it is relatively well conserved and therefore sequence changes in that part of the HCV genome, are slow enough to preserve evidence of relatedness over a considerable period of time [36 ]. Although the hypervariable region (HVR) of the E2 gene of hepatitis C has been used by many investigators for molecular epidemiology studies [37–43], it is believed that a higher rate of sequence changes might hinder attempts to link contact to the source of infection, as has been shown with the analogous V3 region of the envelope gene of HIV [44,45] and might lead to underrecognition of clusters infected from a common source, especially when samples are taken several years after infection [46 ]. Moreover our efforts to amplify HVR in genotype 4 samples, from various transmission groups was unsuccessful (data not shown). A similar epidemiological problem, the high prevalence of hepatitis C genotype 4 virus in patients treated in the same haemodialysis unit, was studied by Sampietro et al. [22] using single strand confirmation polymorphism (SSCP) analysis of polymerase chain reaction products

A. Katsouldou et al.

obtained from the 5∞-untranslated region of the viral genome. Based on genotyping by Inno-Lipa, the suspected index cases were HD 819 or HD 836 or both. Both were of genotype 4 and known carriers before the present outbreak (June 1991 and July 1992 for HD 819 and HD 836, respectively). Phylogenetic tree analysis of the NS5 region ruled out that HD 819 was the index case. On the contrary, the same phylogenetic tree analysis placed the six remaining patients (HD 836, HD 807, HD 806, HD 804, HD 813, HD 816) in the same cluster. This cluster included the five new cases and patient HD 836 who was anti-HCV positive since July 1992, thus strongly suggesting that the latter was the index case of the outbreak. Looking back to the epidemiological data it was found out that (i) patient HD 836 was regularly dialysed in the same shift with patient HD 813 (every Tuesday-Thursday-Saturday evening) from September 1992 and during the whole of 1993, but never shared the same dialysis machine suggesting the horizontal mode of transmission as the most probable explanation and (ii) four out of the five patients who seroconverted between May and October 1993 (HD 804, HD 806, HD 807, HD 816) were dialysed in the same morning shift (every Monday, Wednesday and Friday) when patient HD 836 was not regularly served. Nevertheless, according to the phylogenetic analysis, the infection presumably passed from patient HD 836 to the morning shift. Careful review of epidemiological records revealed that although patient HD 836 was regularly treated in the evening shift D, he exceptionally attended the morning shift A (console no. 10) once on Monday February 1, 1993. While anti-HCV seroconversions started in May 1993, late seroconversion to anti-HCV is well documented among haemodialysis patients [47–49]. These patients can be viraemic and therefore able to transmit hepatitis C, but anti-HCV is not detected for a relatively longer period of time. Thus, it is possible that one or more patients of the morning shift were accidentally infected in February 1993 and, being viraemic, they passed the transmission to the rest of the patients of the morning shift during the following months (horizontal mode of transmission). Patients HD 836 and HD 816 shared the same dialysis machine. However, a vertical route of transmission through the dialysis machines is a very remote probability since patient HD 836 used to be in the Tuesday-Thursday-Saturday evening shift, after which the dialysis machines were routinely carefully disinfected and dried. The unit was ending its operation around 20.00 h starting again next morning (MondayWednesday-Friday) around 08.00 h and it was in this shift that patient HD 816 was dialysed. It is also worth mentioning that AK-10 machines do not allow for the dialysis procedure to start if the disinfection cycle has not been completed; they drain utterly following disinfection and no dialysate recirculation takes place in any part of the hydraulic circuit. Retrospective analysis of probable risk practices suggested that the vehicle of transmission was a nurse


or a group of nurses who did not change gloves regularly when moving from patient to patient. By December 1993, additional measures were adopted to prevent further spread of the virus: (i) all anti-HCV-positive patients were moved into one dialysis shift and (ii) frequent reminders to staff and patients for strict adherence to the universal precautions for the prevention of spreading of viral diseases in the dialysis environment, as recommended by CDC, were applied. During the following 4 year period (January 1994–October 1997) no new seroconversion has been observed. Patient HD 836 was successfully transplanted (May 1994), HD 819 has been transferred to another unit (July 1994), patient HD 816 died from septicaemia (February 1996) and HD 806 from cerebrovascular accident (June 1997). HCV prevalence in the unit at the end of the study period was 11.4% (5/44). The molecular and epidemiological investigation of this outbreak, in addition to the majority of HCV outbreaks studied by molecular epidemiology methods [21–24,26 ], suggest that patient to patient transmission is the most frequently involved transmission mechanism and other recognized or unrecognized routes are extremely uncommon. Acknowledgements. The authors would like to acknowledge Adrienne Varelas and Zissis Moschidis for technical assistance.

References 1. Esteban JI, Esteban R, Viladomiu L et al. Hepatitis C virus antibodies among risk groups in Spain. Lancet 1989; 334: 294–296 2. Schlipko¨ter U, Roggendorf M, Ernst G et al. Hepatitis C virus antibodies in hemodialysis patients. Lancet 1990; 335: 1409 3. Gilli P, Moretti M, Soffritti S et al. Non-A, non-B hepatitis and anti-HCV antibodies in dialysis patients. Int J Artif Organs 1990; 13: 737–741 4. Jeffers LJ, Perez GO, De Medina MD et al. Hepatitis C infection in two urban hemodialysis units. Kidney Int 1990; 38: 320–322 5. Hayashi J, Nakashima K, Kajiyama W et al. Prevalence of antibody to hepatitis C virus in hemodialysis patients Am J Epidemiol 1991; 134: 651–657 6. Ponz E, Campistol JM, Barrera J et al. Hepatitis C virus antibodies in patients on hemodialysis and after kidney transplantation. Transplant Proc 1991; 23: 1371–1372 7. Lin HH, Huang CC, Sheen IS, Lin DY, Liaw YF. Prevalence of antibodies to hepatitis C virus in the hemodialysis unit. Am J Nephrol 1991; 11: 192–194 8. Mondelli MU, Smedile V, Piazza V et al. Abnormal alanine aminotransferase activity reflects exposure to hepatitis C in haemodialysis patients. Nephrol Dial Transplant 1991; 6: 480–483 9. Chan TM, Lok ASF, Cheng IKP. Hepatitis C infection among dialysis patients: A comparison between patients on maintenance haemodialysis and continuous ambulatory peritoneal dialysis. Nephrol Dial Transplant 1991; 6: 944–947 10. Oguchi H, Miyasaka M, Tokunaga S et al. Hepatitis virus infection (HBV and HCV ) in eleven Japanese hemodialysis units. Clin Nephrol 1992; 38: 36–43 11. Muller GY, Zabaleta ME, Arminio A et al. Risk factors for dialysis-associated hepatitis C in Venezuela. Kidney Int 1992; 41: 1055–1058 12. Mondelli MU, Cristina G, Pazza V, Cerino A, Villa G, Salvadeo A. High prevalence of antibodies to hepatitis C virus in hemodialysis units using a second generation assay. Nephron 1992; 61: 350–351

1193 13. Chauveau P, Courouce AM, Lemarec N, et al. Antibodies to hepatitis C virus by second generation test in hemodialyzed patients. Kidney Int 1993; 43 [Suppl 41]: S149–S152 14. Knudsen F, Wantzin P, Rasmussen K et al. Hepatitis C in dialysis patients: relationship to blood transfusions, dialysis and liver disease. Kidney Int 1993; 43: 1353–1356 15. Dussol B, Berthe`zene P, Brunet P, Roubicek C, Berland Y. Hepatitis C virus infection among chronic dialysis patients in the south of France: A collaborative study. Am J Kidney Dis 1995; 25: 399–404 16. Yamaguchi K, Nishimura Y, Fukuoka N et al. Hepatitis C virus antibodies in haemodialysis patients. Lancet 1990; 335: 1409–1410 17. Calabrese G, Vagelli G, Guaschino R, Gonella M. Transmission of anti-HCV within the household of haemodialysis patients. Lancet 1991; 338: 1466–1467 18. Giammaria U, De Meo F, Acitelli S et al. HCV infection in hemodialyzed patients: incidence and correlation with dialysis age. Nephron 1992; 61: 335–336 19. Cantu` P, Mangano S, Masini M, Limido A, Crovetti G, De Fillipo C. Prevalence of antibodies against hepatitis C virus in a dialysis unit. Nephron 1992; 61: 337–338 20. Jadoul M, Cornu C, Van Ypersele De Strihou C. Incidence and risk factors for hepatitis C seroconversion in hemodialysis: A prospective study. Kidney Int 1993; 44: 1322–1326 21. Allander T, Medin C, Jacobson SH, Grillner L, Persson MAA. Hepatitis C transmission in a haemodialysis unit: Molecular evidence for spread of virus among patients not sharing equipment. J Med Virol 1994; 43: 415–419 22. Sampietro M, Badalamenti S, Salvadori S et al. High prevalence of a rare hepatitis C virus in patients treated in the same hemodialysis unit: Evidence for nosocomial transmission of HCV. Kidney Int 1995; 47: 911–917 23. Stuyver L, Claeys H, Wyseur A et al. Hepatitis C in a hemodialysis unit: Molecular evidence for nosocomial transmission. Kidney Int 1996; 49: 889–895 24. Olmer M, Bouchouareb D, Zandotti C, De Micco P, De Lamballerie X. Transmission of the hepatitis C virus in an hemodialysis unit: evidence for nosocomial infection. Clin Nephrol 1996; 47: 263–270 25. Zeuzem S, Scheuermann EH, Waschk D et al. Phylogenetic analysis of hepatitis C virus isolates from hemodialysis patients. Kidney Int 1996; 49: 896–902 26. De Lamballerie X, Olmer M, Bouchouareb D, Zandotti C, De Micco Ph. Nosocomial transmission of hepatitis C in hemodialysis patients. J Med Virol 1996; 49: 296–302 27. Zeuzem S, Ru¨ster B, Roth WK. Clinical evaluation of a new polymerase chain reaction assay (AmplicorTM HCV ) for detection of hepatitis C virus. Z Gastroenterol 1994; 32: 342–347 28. Stuyver L, Rossau R, Wyseur A et al. Typing of hepatitis C virus isolates and characterization of new subtypes using a line probe assay. J Gen Virol 1993; 74: 1093–1102 29. Simmonds P, Holmes EC, Cha TA et al. Classification of hepatitis C virus into six major genotypes and a series of subtypes by phylogenetic analysis of the NS-5 region. J Gen Virol 1993; 74: 2391–2399 30. Enomoto N, Takada A, Nakao T, Date T. There are two major types of hepatitis C virus in Japan. Biochem Biophys Res Commun 1990; 170: 1021–1025 31. Choo QL, Kuo G, Weiner AJ, Orerby LR, Bradly DW, Houghton M. Isolation of a cDNA clone derived from a blood borne non-A, non-B viral hepatitis genome. Science 1989; 244: 359–362 32. Choo QL, Richman KH, Han JH et al. Genetic organization and diversity of the hepatitis C virus. Proc Natl Acad Sci USA 1991; 88: 2451–2455 33. Kwok S, Higuchi R. Avoiding false positives with PCR. Nature 1989; 339: 237–238 34. Higgins DG. and Sharp PM. Fast and sensitive multiple sequence alignments on a microcomputer. Comput Applic Biosci 1989; 5: 151–153 35. Felsenstein J. PHYLIP (Phylogeny Inference Package version 3.5c edn). Department of Genetics, University of Washington, Seattle, USA, 1993

1194 36. Power JP, Lawlor E, Davidson F, Holmes EC, Yap PL, Simmonds P. Molecular epidemiology of an outbreak of infection with hepatitis C virus in recipients of anti-D immunoglobulin. Lancet 1995; 345: 1211–1213 37. Fe´ray C, Samuel D, Thiers V et al. Reinfection of liver graft by a hepatitis C virus after liver tranplantation. J Clin Invest 1992; 89: 1361–1365 38. Weiner AJ, Thaler MM, Crawford K et al. A unique, predominent hepatitis C virus variant found in an infant born to a mother with multiple variants. J Virol 1993; 67: 4365–4368 39. Suzuki K, Mizokami M, Lau JYN et al. Confirmation of hepatitis C virus transmission through needlestick accidents by molecular evolutionary analysis. J Infect Dis 1994; 170: 1575–1578 40. Ohto H, Terazawa S, Sasaki Nobuh et al. Transmission of hepatitis C virus from mothers to infants. N Engl J Med 1994; 330: 744–750 41. Allander T, Gruber A, Naghavi M et al. Frequent patient-topatient transmission of hepatitis C virus in a hematology ward. Lancet 1995; 345: 603–607 42. Esteban JI, Go´mez J, Martell M et al. Transmission of hepatitis C virus by a cardiac surgeon. N Engl J Med 1996; 334: 555–560

A. Katsouldou et al. 43. Aizaki H, Saito A, Kusakawa I et al. Mother-to-child transmission of a hepatitis C virus variant with an insertional mutation in its hypervariable region. J Hepatol 1996; 25: 608–613 44. Ou CY, Ciesielsky CA, Myers G et al. Molecular epidemiology of HIV transmission in a dental practice. Science 1992; 256: 1165–1171 45. Holmes EC, Leigh Brown AJ, Simmonds P. Sequence data as evidence. Nature 1993; 364: 766 46. McCruden EAB, Munro J, King B, Fitzpatrick S, Briggs JD. Patient-to-patient transmission of hepatitis C virus. Lancet 1995; 345: 1443–1444 47. Medin C, Allander T, Roll M, Jacobson SH, Grillner L. Seroconversion to hepatitis C virus in dialysis patients: a retrospective and prospective study. Nephron 1993; 65: 40–45 48. Oliva JA, Maymo RM, Carrio J, Delgado O, Mallafre JM. Late seroconversion of C virus markers in hemodialysis patients. Kidney Int 1993; 43 [Suppl 41]: S153–S156 49. Psichogiou M, Katsoulidou A, Vaindirli E, Francis B, Lee SR, Hatzakis A. Immunologic events during the incubation period of hepatitis C virus infection: the role of antibodies to E2 glycoprotein. Transfusion 1997; 37: 858–862 Received for publication: 7.10.98 Accepted in revised form: 13.01.99