Study design
Retrospective cohort study
To assess potential risk factors associated with the infection a 33-week retrospective cohort study was conducted. Susceptible subjects were enrolled at the time of their first HD starting on Monday February 7 2005 and exited either on Saturday September 24 2005 (end of follow-up), the last know HD before Saturday September 24 2005 (lost to follow-up) or the day of the first anti-HCV positivity (incident cases). The date of HD and relevant patients' data (i.e.: age, gender, HBsAg status and all available data about HCV testing) were obtained either by the medical director of the unit or directly from patients' medical records. A 33-week time period was considered since it is the time when we believe transmission most likely occurred. It includes a 24-week time period before the onset first anti-HCV (we considered the earliest possible contagion date) and the end of the week before the onset of the last known case (end of outbreak).
Molecular epidemiology
To evaluate the presence of additional cases, to confirm the anti-HCV status of known incident/prevalent cases and to assess whether cases were infected with the same HCV molecular variant(s), all patients who were still undergoing dialysis in the HD unit on 26 September 2005 were sampled to be tested for anti-HCV and HCV-RNA. HCV-RNA positive samples were genotyped and, for samples with the same genotype, a molecular characterization of 2 different genomic regions (i.e. NS5B and HVR-1) was performed. Serum was sampled at the HD unit and sent at the virology unit of "INMI Lazzaro Spallanzani" to be analyzed by the second week of October 2005.
Prospective surveillance
All subjects that tested negative for both anti-HCV and HCV-RNA at the first test (see above) were re-tested at 12 and 24 week to confirm the end of transmission. On those occasions tests were performed locally and the results reported as soon as available by the medical director of the HD unit to the epidemiological team. For patients who were only occasionally dialyzed in the HD unit or who moved to another HD centre, information about anti-HCV status was obtained by contacting patients' HD centres staff.
Auditing procedure
Auditing procedure was conducted to reveal potential inadequacies in the practices of staff at the HD unit and to define possible transmission pathways. All medical records and HD treatment logs were analyzed. Application of standard precautions was assessed by interviewing all HCWs and by reviewing all internal protocols in use. Moreover infected patients were interviewed to evaluate the presence of other risk factors for HCV infection and to obtain additional information on the actual implementation of standard precautions. The HCWs' anti-HCV status was evaluated through the examination of the compulsory tests for viral hepatitis done according to the Italian health regulation [11, 12].
Clinical follow-up of incident cases
Between September 2005 and October 2006 all incident cases were followed-up at the infectious disease unit of the "Azienda Sanitaria di Latina Ospedale Santa Maria Goretti". Patients underwent all the medical treatments according to their needs and independently from the present analysis. Relevant data to define clinical outcome (i.e. ALT level, HCV-RNA and therapy with PegInterferon alfa-2a) were obtained directly from the medical director of the unit.
Participants and clinical setting
The HD unit was staffed with 7 doctors and 6 nurses on duty from Monday to Saturday on 12 HD machines. Patients were allotted to 4 different shifts, i.e.: Monday, Wednesday and Friday mornings (MWF-am) or afternoon (MWF-pm), Tuesday, Thursday and Saturday morning (TTS-am) or afternoon (TTS-pm).
Case definition
Susceptible subject: a subject anti-HCV negative at his/her last test before Monday 7 February 2005*.
Prevalent case: a subject already known to be anti-HCV positive by Monday 7 February 2005*.
Incident case: a susceptible subject who eventually became anti-HCV positive.
Possible case: an incident case without sufficient data to define the HCV molecular variant (i.e. genotype and characterization of both NS5B and HVR1 sequences)
Confirmed case: an incident case infected with the same HCV molecular variant of another incident or prevalent case (i.e. identical genotype, NS5B and HVR1 sequences).
Index case: a prevalent case infected with the same HCV molecular variant of another incident case (i.e. identical genotype, NS5B and HVR1 sequences).
Excluded case: prevalent or incident cases for whom molecular analysis (either genotype, NS5B or HVR1 sequences) excluded relation with at least another incident or prevalent case.
*According to the internal infection control protocols, all patients underwent anti-HCV test at the time of first HD and then every 3 months (see Results "Audit procedure").
Epidemiological and clinical outcomes definition
Standard precautions (which include hand hygiene, proper use of gloves, and safe injection practices) are defined according to Health Care Infection Control Practices Advisory Committee 2007 [8].
Attack rate was calculated as the proportion of incident cases over susceptible subjects at the end of outbreak.
Clinical attack rate was calculated as the proportion of incident cases with ALT > 80 UI/ml at least once after the first positive anti-HCV test out of the exposed susceptible subjects at the end of the outbreak.
Clinical outcome: incident cases were classified as either recovered, if HCV-RNA was undetectable, or chronically infected, if HCV-RNA was detectable, by the end of October 2006.
Virological methods
Standard virology (HCV serology and PCR for HCV-RNA)
A third-generation immunoenzymatic assay (Axsym HCV version 3.0, Abbott Laboratories) was used for anti-HCV antibody testing. A commercially available quantitative polymerase chain reaction (PCR) assay was used to measure HCV-RNA (Amplicor HCV Monitor 2.0, Roche Diagnostics).
HCV genotype analysis and molecular characterization of NS5B and HVR-1
A line probe assay (InnoLipa HCV II, Bayer Diagnostics) was used to genotype HCV in all HCV-RNA positive subjects. Samples from patients with the same HCV genotype were eligible for molecular analysis of NS5B and HVR1. HCV-RNA was extracted from serum samples using the QIAamp Viral RNA kit (QIAGEN, Hilden, Germany), underwent retrotranscription by random hexamer extension and was used to perform molecular analysis. Two-strand direct sequencing was carried out on nested PCR products obtained from the NS5B region and from the hypervariable region 1 (HVR1) encompassing in the E2 gene, as previously reported [13]. Sequencing was performed on ABI Prism 3100, using the BigDye Terminator cycle sequencing kit (Applied Biosystems, Warrington, UK). The sequences, aligned with CLUSTAL W software (version 1.5), were confronted by BLAST with sequences from the National Center for Biotechnology Information (NCBI) database (U.S. National Library of Medicine, Bethesda, MD, http://www.ncbi.nlm.nih.gov), to confirm the genotype assignment. Phylogenetic trees were constructed using the neighbor-joining method, including NS5B and HVR-1 reference sequences from GenBank, as well as sequences of genotype 2c obtained from routine Laboratory diagnostics or from previously analyzed clusters of transmission, all epidemiologically unrelated to the described episode (see figure legend). All the algorithms used are included in the Mega package (version 2.1). The results of this analysis were used to confirm a monophyletic cluster of infection.
Statistical methods
Overall and stratum specific rates with 95% confidence interval (95% CI) were calculated as cases per 100 dialysis-week person. Crude and adjusted rate ratios (RR), 95% CI interval and p-value of dichotomous variables were calculated according to Poisson regression with a robust error variance for binary data as described by Zou G. [14]. Discrete variables were assessed with a conventional Poisson regression. The following were considered as potential risk factors: age >60, gender, having been dialyzed at least once either after or along a HCV genotype 2c positive subject, overall number of HD while anti-HCV negative (as the only discrete variable), being usually dialysed either on MWF-am or MWF-pm or TTS-am or TTS-pm or other (each patient was allocated in one of the 5 above group according to the shift he/she had 75% or more HD or "other" if this criterion was not met for any shift). A multivariate model was built up including only the risk factors with a p-value less than 0.10 from the univariate analysis. Statistical significance was assumed if p < 0.05. The statistical analysis was performed by using STATA 11 statistical package (Stata-Corp, 4905 Lakeway Drive College Station, Texas 77845 USA).