Sexual transmission of hepatitis C virus (HCV) is considered to be rare , but over the past decade, there is an increasing evidence for sexual transmission of HCV among HIV-infected MSM [2,3]. Given the reduced spontaneous clearance  and more rapid progression to liver fibrosis and cirrhosis , the increased incidence of HCV among this population is of great concern. Nevertheless, the infection can be cured, especially during the acute phase, with reported success rates above 60% [6,7]. With new antivirals becoming available, success rates are expected to increase further, although this comes with a substantial cost .
Unfortunately, after viral eradication, patients may become reinfected, as observed among people who inject drugs [9–12]. Whether clearance of a primary infection, either treatment-induced or spontaneously, results in some degree of protection against (persistence of) reinfection is still undecided, given the inconsistent findings in studies that addressed this issue thus far [12–16].
To date, the number of studies on examining the rate of HCV reinfection among HIV-infected MSM remains very limited. We previously reported a very high reinfection rate of 15.2 per 100 person-years among patients successfully treated for acute HCV infection . This finding was confirmed and extended in a recent study from the United Kingdom that investigated the rate of reinfection among both treatment-induced and spontaneous clearers of acute HCV infection .
The true incidence of secondary infections among HIV-infected MSM is likely to be underestimated for several reasons. First, depending on the interval of testing, secondary infections that were rapidly cleared may have been missed. Second, patients with persistent infection were not included in both studies, whereas persistent viremia may have been caused by a superinfection with a new virus, with or without (observed) clearance of the primary virus. Third, without genetic analysis of the ‘relapsing’ virus, recurrence of viremia within 24 weeks following clearance of viremia may have been misclassified as a relapse [9,19].
Increasing our knowledge on the incidence of multiple infections and assessing the genetic relatedness of primary and successive viral strains will provide more insight into correlates of immunity against HCV and is crucial for vaccine development and for targeted preventive strategies. We therefore studied the occurrence of multiple HCV infections in persistently infected patients and in patients who cleared the infection with or without treatment. The presence of new infections with a different genotype (genotype switch), or with a different strain from the same genotype (clade switch), was assessed by systematically sequencing the virus present during the entire viremic period.
The study population consisted of 85 HIV-infected MSM attending the HIV treatment clinic of the Academic Medical Center (AMC) in Amsterdam, the Netherlands, who acquired a primary HCV infection sexually between 1994 and 2011. Acute HCV infection was defined by a positive HCV RNA test preceded by a negative HCV RNA test in patients without evidence of past HCV infection. Patients were either prospectively identified during acute infection or retrospectively by determining HCV RNA in anti-HCV-positive patients in stored sera from earlier time points. The majority of patients were participants of the MSM Observational Study of Acute Infection with hepatitis C (MOSAIC) – a prospective cohort study on acute hepatitis C infection in HIV-infected MSM . Informed consent was obtained, and the study was conducted according to the hospital ethical guidelines and the Dutch code of conduct for responsible use of human tissue and medical research 2011.
The presence of HCV RNA was assessed by either transcription-mediated amplification (VERSANT HCV RNA Qualitative Assay; Siemens Healthcare Diagnostics Inc., Tarrytown, New York, USA) or branched-chain DNA (VERSANT HCV RNA 3.0 Assay, Siemens). For genotyping, a 389-base pair (bp) fragment spanning positions 8616–8275 relative to the H77-strain (AF009606) of the NS5B region was amplified and sequenced as described by Murphy et al.. For the detection of new infections with the same genotype (i.e. clade typing), a 590-bp fragment from the envelope, spanning positions 1295–1885 relative to the H77 strain, which includes the hyper-variable region 1 (designated ‘E2/HVR1’), was amplified and directly sequenced. E2/HVR1 sequences were submitted to GenBank, under accession numbers KP399220–KP399593. All NS5B sequences were submitted to GenBank (accession numbers KP398885–KP399219).
Analysis of multiple infections
For each patient, the first and the most recent RNA-positive samples were selected for NS5B-genotyping. If a genotype switch was observed between the first and the last RNA-positive time points, samples taken between these two time points were genotyped to determine the interval of genotype switch.
If, at the most recent time point, the original genotype was still present, sequence analysis of E2/HVR1 was performed longitudinally on the stored sera, to detect secondary infections with the same genotype (i.e. clade switch).
Phylogenetic analysis was used to determine whether the evolving primary virus was still present or replaced by a new viral strain from the same genotype as outlined in detail below. First, sequences were analyzed using Codoncode version 3.7.1, aligned using Clustal X version 2.0.11, and edited using GeneDoc version 2.7 software. For visual inspection of sequences, maximum likelihood trees were constructed for each genotype under a Hasegawa–Kishino–Yano evolutionary model with invariant sites and a gamma distribution of among-site rate heterogeneity (HKY + I + G) as implemented in Molecular Evolution Genetic Analysis version 5 . This substitution model exhibited the best-fit evolutionary substitution model for this fragment, using the Akaike information criterion in model test version 3.7 , as implemented in Paup*4.0 . Trees were unrooted and bootstrap values were determined from 1000 bootstrap resamplings of the original data.
To define an objective criterion for distinguishing intrahost evolution from a clade switch, nucleotide substitution rates were estimated for the E2/HVR1 genomic region for all patient-specific internal branches using a Bayesian Markov Chain Monte Carlo (MCMC) approach as implemented in the program BEAST, version 1.7.4 . For each genotype, the model was run separately by applying an uncorrelated relaxed lognormal molecular clock  and a coalescent exponential population size with a random starting tree [27,28]. The MCMC chains were run for 10e7 states to obtain a good convergence and effective sample size (ESS) above 200. The model was run while enforcing monophyly for each patient, thereby forcing the model to impose an unrealistically high nucleotide substitution rate for branches with sequences derived from a heterologous virus. Substitution rates for all branches were extracted from the maximum clade credibility tree as created by TreeAnnotator version 1.7.4 and constructed using FigTree version 1.4.0. Mean substitution rates from all internal branches from patient-specific monophyletic clades were calculated, and a rate exceeding the mean + three times the SD was considered as evidence for the occurrence of a clade switch between the two time points.
Cleared infections were defined as absent HCV RNA for at least 60 days in untreated patients or when a sustained viral response (SVR) was achieved 24 weeks after treatment discontinuation. If a reinfection occurred within the 24-week period following the end of treatment, the response was regarded as SVR.
A reinfection was defined as recurrence of HCV RNA with a viral strain other than the primary one after a cleared infection. A superinfection was defined as the detection of a new viral strain other than the primary virus during follow-up without documented HCV RNA-negative time points in between. Throughout the study, the term new infection refers to either reinfection or superinfection.
The incidence rate of new infections following primary infections and its confidence interval (CI) were calculated using person-time methods and are given per 100 person-years. The mid-P test was used to compare incidence rates. The date of primary HCV infection was estimated as the mid-point between the date of last RNA-negative sample and the date of the first RNA-positive sample. Follow-up time was calculated from estimated time of infection until the date when either a genotype or clade switch occurred, or the last date of HCV genotyping if no viral switch had occurred. Cumulative incidence curves were estimated within a competing-risks framework to determine the incidence of the first genotype switch compared to the first clade switch during follow-up. In addition, using Cox proportional-hazards analysis, the effect of genotype (1a versus non1a) at primary infection on the risk of a secondary infection with genotype 1a infection was determined. During treatment, patients were considered not to be at risk for a new infection. A P value less than 0.05 was considered to be statistically significant. The R language and environment for statistical computing, version 2.8 and SPSS statistical software (version 19.0; SPSS Inc., Chicago, Illinois, USA) were used for the data analysis.
At the time of HCV infection, the median age was 41.6 [interquartile range (IQR) 36.2–46.8] years and the median CD4+ cell count was 500 (IQR 393–638) cells/μl. HIV load was available for 83 out of 85 patients: in 48 (57.8%) no HIV RNA was detectable at the time of HCV infection. Median interval between HCV RNA testing was 2.2 (IQR 0.9–4.8) months. Fifty-six out of 85 patients (65.9%) were treated with peg-interferon and ribavirin during the acute stage of infection, and 46 of 56 (82.1%) achieved SVR. Five out of 29 untreated patients (17.2%) cleared the primary infection spontaneously. Baseline and follow-up patient demographics and infection data stratified by HCV treatment and outcome are summarized in Table 1. Patients were predominantly infected with genotype 1a (60.0%) and genotype 4d (24.7%). Supplement Fig. 1 (https://links.lww.com/QAD/A761) depicts the date of the first RNA-positive sample per genotype for all primary infections observed during follow-up.
NS5B genotyping identified 18 genotype switches, with a switch from HCV-4d to HCV-1a being the most common (n = 9). Other observed genotype switches were 1a to 4d (n = 3), 1a to 2b (n = 2), 1a to 3a (n = 1), 1b to 4d (n = 1), 1a to 1b (n = 1), and 3a to 1a (n = 1). Genotype switches occurred in both treated and untreated patients following both cleared and persistent infections. Sixteen of these 18 genotype switches were secondary infections, and two were tertiary infections in two patients.
Given the conserved character of NS5B , for the remaining 69 patients without a genotype switch, serial sequencing of the more variable E2/HVR1 region was performed, to determine whether changes in E2/HVR1 over time were compatible with intrahost evolution or with a new infection with the same genotype (clade switch). In total, 380 sequences were generated. Median interval between sequences was 0.57 year (IQR 0.21–0.99).
Visual inspection of maximum likelihood (ML)-E2/HVR1 trees indicated that in general, sequences clustered per patient, with some intermingling of sequences from different patients in the beginning of the infection. However, in six HCV-1a-infected patients, variants were detected during follow-up that formed distinct phylogenetic clusters, suggesting replacement of the primary strain with another HCV-1a strain (Supplement Fig. 2, https://links.lww.com/QAD/A761). These six clade switches were also observed in the NS5B phylogenetic tree (data not shown).
Intermingling in the beginning of infection was also observed in the HCV-4d phylogenetic tree. Here, sequences derived from one patient occasionally formed separate clusters during follow-up (Supplement Fig. 3, https://links.lww.com/QAD/A761, e.g. patient 013 with patient 038, and patients 004, 0053, 0061, and 0037), which impeded distinguishing intrahost evolution from a clade switch by visual inspection only.
Therefore, to define a more objective criterion for the presence of clade switch, we estimated nucleotide substitution rates across internal branches of the tree using the program BEAST . The mean substitution rate across all branches was 9.44e−03 substitutions/site per year (SD 1.56e−02). Using a cut-off of 5.62e−02 substitutions/site per year for clade-switch designation (mean +3*SD) in the BEAST run with forced monophyly, seven outliers in substitution rates from six patients were identified (Fig. 1), confirming that the clusters with large intrapatient genetic distances in the HCV-1a phylogenetic tree indeed represented clade switches. Nucleotide substitution rates above the threshold were not present among patients infected with genotypes other than 1a. Clade switches occurred as a second infection (n = 3), third infection (n = 2), and even fourth infection (n = 1).
Incidence of new infections: association with clearance and primary genotype
In total, 24 new infections, both super and reinfections, were observed in 19 patients. Figure 2 summarizes all new infections in a flow chart in relation to treatment and treatment outcome. In four out of 19 patients, multiple viral switches were observed. Most new infections were reinfections, following either spontaneous or treatment-induced clearance of the primary infection, with a total of 19 reinfections among 59 previous cleared infections, and five superinfections in 47 persistent infections. Interestingly, in the five spontaneously cleared infections, three reinfections occurred. Figure 3 illustrates the individual infection history of patients with multiple infections.
Overall incidence rate of a secondary infection was 5.39 cases per 100 person-years (95% CI 3.34–8.26), based on 19 secondary infections among 85 patients with a median follow-up time of 2.87 years (IQR 1.51–5.87). The incidence rate of reinfection among those with a resolved primary infection (n = 51) was 14.5 per 100 person-years (95% CI 8.41–23.34). In contrast, incidence rate of superinfections was significantly lower, with a rate of 1.6 cases per 100 person-years (IQR 0.5–3.9, P < 0.01.)
Cause-specific cumulative incidence curves are shown in Fig. 4a. At 5 years following primary infection, the cumulative incidence of a clade switch was 4.8% (95% CI 0.0–10.1), whereas the cumulative incidence of a genotype switch was 26.7% (95% CI 13.3–38.1). In addition, in a Cox proportional-hazards analysis, patients with HCV-1a during primary infection had a decreased risk for acquiring an HCV-1a again as secondary infection (hazard ratio 0.25, 95% CI 0.07–0.93) compared to patients with a non-HCV-1a at primary infection (Fig. 4b; P = 0.03, log-rank test).
Given the high similarity of circulating viruses in an emerging epidemic, identification of new infections with the same genotype as the original one can be challenging, in particular, in patients with persistent viremia or a relapse following treatment. Apart from sequencing the rather conserved NS5B region to identify genotype switches, we also sequenced a genetically highly diverse fragment of the second envelope gene (E2). The intrahost nucleotide substitution rate of this region was determined using serial samples, and a threshold for genetic divergence between two sequences in a certain timeframe was defined, enabling the distinction between new infections and intrahost evolution. This allowed us to precisely estimate the incidence of new infections with both the original and a different genotype in both persistent and cleared acute HCV infections.
The overall incidence rate of secondary infections was 5.39 cases per 100 person-years. However, among patients with cleared infections, incidence of secondary infections was as high as 14.5 per 100 person-years. A novel and important finding is that, for genotype 1a – the most common genotype – the risk of acquiring a genotype 1a infection again appeared to be reduced. This suggests that, even in HIV-infected individuals, partial genotype-specific immunity is generated at the time of the primary infection.
The high incidence of reinfection following spontaneous or treatment-induced clearance of the primary infection observed in this study is in line with the high reinfection rate of 15.2 cases/100 person-years observed in an earlier smaller study on the incidence of reinfection following treatment-induced clearance in HIV-infected MSM with acute HCV from two HIV clinics . The incidence rate is somewhat higher than that observed in a recent study from the United Kingdom , which showed a reinfection rate of 8.0/100 person-years among HIV-infected MSM who cleared their primary HCV infection. The higher estimate in our study might be explained by our extensive sequence analysis, which enabled us to identify clade switches in patients who ‘relapsed’ within 24 weeks after the end of the treatment. This results in the identification of additional reinfections, whereas such patients were considered not to have cleared their primary infection and were thus excluded in the UK study. In addition, the smaller testing interval in our study might have resulted in the identification of reinfections that would have been missed otherwise .
Martin et al. reported a nonsignificant trend towards a lower incidence of reinfection among spontaneous clearers versus treatment-induced clearance. Our study population comprised only five spontaneous clearers among untreated patients, and three of them became reinfected, suggesting that spontaneous clearance of a previous infection does not reduce the risk of reinfection.
In this study, the replacement of the primary virus by other viruses in persistently infected patients was also investigated. We cannot prove that such replacements are indeed superinfections since these could also have been caused by dynamic changes in dominance when different viral strains are present at the same time. However, we believe that the newly identified viruses are true superinfections, given the overall high incidence of new infections , although another possibility is that such superinfections are in fact reinfections, where aviremic time points were missed. The likelihood of this scenario is small, given the small RNA-testing interval in this study. In any case, the incidence rate of such superinfections was relatively low with 1.6 cases per 100 person-years. There are several explanations for a lower incidence of superinfection in persistently infected patients as compared to the incidence of reinfection in patients with cleared infection. First, patients who are aware of their chronic HCV infection may be less likely to engage in high-risk sexual behaviour. Second, in the HCV cell culture system, it has been shown that superinfection is excluded at a postentry step [30,31]. Third, cross-neutralizing antibodies are present in most patients during the chronic phase of infection . These circulating antibodies may immediately neutralize any new virus, before a new infection can be established.
Perhaps, the most remarkable finding of this study is the observed strikingly reduced incidence of new infections with the original genotype as compared to new infections with a different genotype. Whether the incidence of such new infections is truly reduced, or whether they are cleared more rapidly and therefore may be missed as a consequence of our testing interval of 2.2 months remains unknown. Indeed, reinfections are characterized by lower viral load and a shorter duration of viremia, suggesting the existence of acquired immunity [15,33]. However, our study suggests that there is a strong genotype-specific component to this acquired immunity, resulting in partial protection against the genotype present at the primary infection. Whether such genotype-specific immunity is driven by B or T cells, or even natural killer (NK) cells , needs to be further explored. It is also unclear whether the new antiviral regimens, without pegylated interferon, may have an effect on acquired partly genotype-specific immunity. For vaccine development, however, our findings suggest that a strategy directed at generating genotype-specific responses may be more successful than pursuing the holy grail of a pan-genotype vaccine for a virus as variable as HCV.
Our study is limited by small numbers and a relatively short follow-up. In addition, the ongoing epidemic in Amsterdam may convey immunity only against locally circulating variants of the same genotype with saturation of infection with these variants among the population at risk. Therefore, the observed partial protection against reinfection with the same genotype needs to be confirmed by other studies, before firm conclusions can be drawn.
In conclusion, this study confirms the high rate of HCV reinfections following primary infection among HIV-positive MSM, highlighting the need for public health interventions in this high-risk group. In addition, this study demonstrates that observational cohort studies with frequent sampling of individuals with acute HCV infection are important to better understand the correlates of immunity against HCV . Such studies ultimately will contribute to the development of a protective vaccine as the most powerful public health intervention.
All other authors contributed significantly to the intellectual content of the manuscript. X.V.T. performed laboratory work, executed data analysis, and drafted the manuscript. S.P.R. performed laboratory work. B.P.X.G., C.K.H., and J.W.V. contributed to data analysis. J.T.M. M. and M.V. are physicians treating HIV-infected MSM, and together with M.D.J. and R.M., they also contributed to the intellectual content of the manuscript. M.P. is MOSAIC principal investigator. J.S. was the principal investigator of this study.
We kindly thank all MOSAIC study participants and other members of the MOSAIC study team: J. Arends; D. van Baarle; K. Brinkman; M. van den Ende; L. Gras; D. Kwa; T. van de Laar; F. Lambers; J. Mulder; C. Smit; W. van der Veldt. In addition, the authors would like to thank J. Karlas for coordinating HCV testing; M. Bakker for assistance with sample retrieval, G.R. Visser for data management, M. Mutschelknauss for data collection and patient recruitment and R. Rose for assistance with data analysis.
Financial support: This work was supported by a grant from ‘AIDS fonds’ Netherlands (2008026) and from the Virgo consortium, funded by the Dutch government (project number FES0908), and the Netherlands Genomics Initiative (project number 050-060-452).
Conflicts of interest
There are no conflicts of interest.
1. Tohme RA, Holmberg SD. Is sexual contact a major mode of hepatitis C virus transmission?
2. Urbanus AT, van de Laar TJ, Stolte IG, Schinkel J, Heijman T, Coutinho RA, et al. Hepatitis C virus infections among HIV-infected men who have sex with men: an expanding epidemic
3. Van de Laar T, Pybus O, Bruisten S, Brown D, Nelson M, Bhagani S, et al. Evidence of a large, international network of HCV transmission in HIV-positive men who have sex with men
4. Van den Berg CH, Grady BP, Schinkel J, van de Laar T, Molenkamp R, Van HR, et al. Female sex and IL28B, a synergism for spontaneous viral clearance in hepatitis C virus (HCV) seroconverters from a community-based cohort
5. Thein HH, Yi Q, Dore GJ, Krahn MD. Natural history of hepatitis C virus infection in HIV-infected individuals and the impact of HIV in the era of highly active antiretroviral therapy: a meta-analysis
6. Lambers FA, Brinkman K, Schinkel J, Spijkerman IJ, Molenkamp R, Coutinho RA, et al. Treatment of acute hepatitis C virus infection in HIV-infected MSM: the effect of treatment duration
7. Vogel M, Dominguez S, Bhagani S, Azwa A, Page E, Guiguet M, et al. Treatment of acute HCV infection in HIV-positive patients: experience from a multicentre European cohort
8. Puoti M, Rossotti R, Travi G, Panzeri C, Morreale M, Chiari E, et al. Optimizing treatment in HIV/HCV coinfection
. Dig Liver Dis
2013; 45 (Suppl 5):S355–S362.
9. Grady BP, Schinkel J, Thomas XV, Dalgard O. Hepatitis C virus reinfection following treatment among people who use drugs
. Clin Infect Dis
10. Grebely J, Knight E, Ngai T, Genoway KA, Raffa JD, Storms M, et al. Reinfection with hepatitis C virus following sustained virological response in injection drug users
. J Gastroenterol Hepatol
11. Pham ST, Ra Bull, Bennett JM, Rawlinson WD, Dore GJ, Lloyd AR, et al. Frequent multiple hepatitis C virus infections among injection drug users in a prison setting
12. Van de Laar TJW, Molenkamp R, van den Berg C, Schinkel J, Beld MGHM, Prins M, et al. Frequent HCV reinfection and superinfection in a cohort of injecting drug users in Amsterdam
. J Hepatol
13. Grebely J, Pham ST, Matthews GV, Petoumenos K, Bull RA, Yeung B, et al. Hepatitis C virus reinfection and superinfection among treated and untreated participants with recent infection
14. Mehta SH, Cox A, Hoover DR, Wang XH, Mao Q, Ray S, et al. Protection against persistence of hepatitis C
15. Osburn WO, Fisher BE, Dowd KA, Urban G, Liu L, Ray SC, et al. Spontaneous control of primary hepatitis C virus infection and immunity against persistent reinfection
16. Page K, Hahn Ja, Evans J, Shiboski S, Lum P, Delwart E, et al. Acute hepatitis C virus infection in young adult injection drug users: a prospective study of incident infection, resolution, and reinfection
. J Infect Dis
17. Lambers Fa E, Prins M, Thomas X, Molenkamp R, Kwa D, Brinkman K, et al. Alarming incidence of hepatitis C virus re-infection after treatment of sexually acquired acute hepatitis C virus infection in HIV-infected MSM
18. Martin TCS, Martin NK, Hickman M, Vickerman P, Page EE, Everett R, et al. HCV reinfection incidence and treatment outcome among HIV-positive MSM in London
19. Hara K, Rivera MM, Koh C, Demino M, Page S, Nagabhyru PR, et al. Sequence analysis of the hepatitis C virus in relapse after sustained virological response: relapse or reinfection?
. he J Infect Dis
20. FEDERA. Human tissue and medical research: code of conduct for responsible use. Rotterdam: Stichting FMWV; 2011.
21. Murphy DG, Willems B, Deschênes M, Hilzenrat N, Mousseau R, Sabbah S, et al. Use of sequence analysis of the NS5B region for routine genotyping of hepatitis C virus with reference to C/E1 and 5’ untranslated region sequences
. J Clin Microbiol
22. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods
. Mol Biol Evol
23. Posada D, Crandall KA. MODELTEST: testing the model of DNA substitution
24. Swofford DL. PAUP*. Phylogenetic analysis using parsimony (*and other methods),
version 4. Sunderland, MA: Sinauer Associates; 2012.
25. Drummond AJ, Suchard MA, Xie D, Rambaut A. Bayesian phylogenetics with BEAUti and the BEAST 1.7
. Mol Biol Evol
26. Drummond AJ, Ho SY, Phillips MJ, Rambaut A. Relaxed phylogenetics and dating with confidence
27. Drummond AJ, Nicholls GK, Rodrigo AG, Solomon W. Estimating mutation parameters, population history and genealogy simultaneously from temporally spaced sequence data
28. Kingman JF. Origins of the coalescent
29. Vickerman P, Grebely J, Dore GJ, Sacks-Davis R, Page K, Thomas DL, et al. The more you look, the more you find: effects of hepatitis C virus testing interval on reinfection incidence and clearance and implications for future vaccine study design
. J Infect Dis
30. Van der Helm JJ, Prins M, del Amo J, Bucher HC, Chêne G, Dorrucci M, et al. The hepatitis C epidemic among HIV-positive MSM: incidence estimates from 1990 to 2007
31. Tscherne DM, Evans MJ, Von HT, Jones CT, Stamataki Z, McKeating JA, et al. Superinfection exclusion in cells infected with hepatitis C virus
. J Virol
32. Pestka JM, Zeisel MB, Bläser E, Schürmann P, Bartosch B, Cosset F-L, et al. Rapid induction of virus-neutralizing antibodies and viral clearance in a single-source outbreak of hepatitis C
. Proc Natl Acad Sci U S A
33. Sacks-Davis R, Aitken CK, Higgs P, Spelman T, Pedrana AE, Bowden S, et al. High rates of hepatitis C virus reinfection and spontaneous clearance of reinfection in people who inject drugs: a prospective cohort study
. PLoS One
34. Sun JC, Beilke JN, Lanier LL. Adaptive immune features of natural killer cells
35. Cox AL, Page K, Bruneau J, Shoukry NH, Lauer GM, Kim AY, et al. Rare birds in North America: acute hepatitis C cohorts