Re-emergent Hepatitis C Viremia After Apparent Clearance in HIV-Positive Men Who Have Sex With Men: Reinfection or Late Recurrence?

To the Editors: The increased number of cases of acute hepatitis C virus (HCV) infection in HIV-positive men who have sex with men has been well documented.1 The majority are within a network of individuals engaging in higher-risk sexual activity with accompanying recreational drug use.2 Immunologic responses evoked after HCV infection are not sufficient to protect against reinfection or superinfection.3,4 Most of the data regarding reinfection are derived from the intravenous drug-using population or organ transplant or polytransfusion recipients.5-7 There are no data detailing HCV reinfection after sexual exposure. We report data from two London teaching hospitals describing HIV-infected individuals acutely infected with HCV who developed re-emergent hepatitis C viremia after apparent clearance of the initial infection. Apparent clearance occurred either spontaneously or after treatment. The objectives of the study were to: 1) describe the characteristics of the individuals with re-emergent viremia; 2) describe the features of the subsequent episodes of viremia; and 3) attempt to differentiate between relapse and reinfection using viral phylogenetic analyses. Individuals previously diagnosed with acute HCV infection and included in the recent paper2 were examined further by database interrogation. The cohort consisted of all HIV-positive individuals diagnosed with acute infection within two clinics between 1999 and 2008. Those subjects who experienced a second episode of hepatitis C viremia after apparent clearance of the primary viremic episode were included in this study. Individuals with re-emergent viremia were categorized according to the following: 1) undetectable HCV RNA by Roche assay (COBAS AmpliPrep/Cobas TaqMan; Roche, CA) for more than 6 months after acute infection without specific treatment. This was termed spontaneous virologic clearance; 2) undetectable HCV polymerase chain reaction for more than 24 weeks after completion of treatment with pegylated interferon-2a and ribavirin (pIFN/RBV). This was termed sustained virologic response; and 3) a second episode of viremia after an end of treatment response (ETR) but before the elapse of 6 months, thus not meeting the definition of a true sustained virologic response. This subset had sufficient ongoing risk to suspect HCV reinfection had occurred. The dates of each period of hepatitis C viremia were noted. A subset of subjects with paired stored frozen serum samples, one from each period of viremia, was identified. For each of these cases, informed consent was obtained. Samples were anonymized and processed to obtain sequences for phylogenetic analysis. In all subjects, case notes and electronic records were reviewed and variables, including demographics, HIV parameters, antiretroviral treatment, HCV parameters including alanine transferase (ALT), HCV polymerase chain reaction, and genotype were extracted. Genotypes were identified by direct sequencing where stored samples were available and amplifiable. Where this was not the case, genotypes were detected by sequencing within our local service, where distinction between 1a and 1b genotypes was not always possible at the early stages of the epidemic. Concurrent diagnoses of sexually transmitted infections (STIs) were recorded as a surrogate marker of ongoing high-risk activity. HCV RNA was extracted (QIAGEN, West Sussex, UK) following the manufacturer's instructions and reverse-transcribed using Superscript First Strand Synthesis System (Invitrogen, Cambridge, UK) to produce single-stranded cDNA. Nested polymerase chain reaction (PCR) was performed using primers to amplify parts of the E1/E2 region of the genome incorporating HVR1. A terminal extension was carried out to ensure all products were fully double-stranded. Two microliters of the first-round PCR product was amplified with E2F2/E2R2 primers. The PCR products and appropriate size markers were subjected to 1.5% agarose gel electrophoresis and visualized by ultraviolet light transillumination. Bands were excised and the DNA purified using QIAquick Gel Purification Kit (QIAGEN) following the manufacturer's instructions and eluted in buffer (10 mM Tris-HCl, pH 7.6). The amplicon was directly sequenced using forward and reverse primers by The Sequencing Service (School of Life Sciences, University of Dundee, UK) using Applied Biosystems Big Dye Version 3.1 chemistry and an Applied Biosystems 3730 automated capillary sequencer. Sequence analysis and multisequence alignment was carried out using Megalign software (DNASTAR, Madison, WI). Viral relatedness was assessed by constructing a phylogenetic tree of study sequences together with 74 sequences from the epidemic described by Danta et al2 and 96 unrelated sequences obtained from GenBank. The phylogenetic tree was estimated using maximum likelihood with the Genetic Algorithm for Rapid Likelihood Inference program.8 The substitution model used was General Time Reversible plus gamma rate heterogeneity with a proportion of invariable sites. Of all HIV/acute HCV coinfection clinic attendees, 22 individuals were identified as having a second episode of re-emergent viremia. All episodes were in men who have sex with men, none of whom reported a history of injecting drug use. The mean age at first HCV infection was 40 years (range, 26-51 years). The mean duration of HIV infection was 87 months (range, 8-183 months) with a mean CD4 count at first episode of viremia of 463 cells/mm3 (range, 195-834 cells/mm3). All but two of the 22 individuals were treated for the first episode of HCV infection. The two untreated individuals underwent spontaneous virologic clearance of more than 6 months. Of those treated, the majority received 24 weeks of pIFN/RBV with all but four meeting the criteria for an virologic response. Of the 22 affected, 15 were prescribed highly active antiretroviral therapy at the first episode of viremia. All individuals were asymptomatic at the second episode of viremia. Re-emergent viremia was detected after an increase in ALT or screening of HCV PCR. The median peak ALT at initial identification of re-emergent viremia was 12 times (range, 1-74 times) the upper limit of normal. The mean CD4 count at the second episode of viremia was 512 cells/mm3 (range, 224-748 cells/mm3). The mean duration of HCV RNA negativity after an ETR and/or apparent clearance before re-emergent viremia was 22 months (range, 3-54 months). STIs diagnosed during the period of the ETR and/or apparent clearance, reported by subjects or documented in medical notes, were recorded. Our data indicate that in the period after the ETR and/or apparent clearance and before re-emergent viremia, 27 STIs were recorded in 18 individuals. Infectious syphilis and gonorrhea were the STIs most frequently diagnosed. Of 22 individuals with re-emergent viremia, nine had paired amplifiable sequences available for phylogenetic analysis. One (Patient 15) was initially infected with HCV genotype 4 and had an virologic response to treatment lasting 19 months after which there was re-emergent viremia with HCV genotype 1a, suggesting reinfection rather than a very late relapse. The other eight patients had both initial and re-emergent infection with HCV genotype 1a. These, together with three sequences from patients on whom paired sequences were not acquired, 74 sequences from the epidemic of acute HCV described by Danta et al,2 and 96 unrelated sequences, were subjected to phylogenetic analysis and a tree constructed (Fig. 1). Six of the eight had genetically divergent samples with each of the paired samples falling into different monophyletic clades evolutionarily more distant than unrelated isolates, providing powerful evidence that these individuals were re-infected with a different strain of HCV.FIGURE 1: Genotype 1a Maximum Likelihood Phylogenetic Tree generated with GARLI using GTR model plus gamma rate heterogeneity with a proportion of invariable sites. Study subjects are in blue, related epidemic sequences are in red (74) and unrelated sequences in blue (96). The scale is in units of substitutions per nucleotide.In two individuals (Patients 2 and 10), re-emergent viremia was seen at 3 months and in a further two individuals (Patients 4 and 20) at 4 months after treatment cessation and an observed ETR and/or apparent clearance. However, in all four of these cases, re-emergent viremia was not observed until a mean of 7 months (range, 6-9 months) after the first episode of negative HCV RNA with each individual having at least two negative PCR results separated by more than 2 months. In two of these cases (Patients 2 and 4), phylogenetic analysis of paired samples indicates that HCV reinfection with a new strain occurred. A further three individuals (Patients 8, 19, and 22) without paired samples available for direct sequencing had a switch in genotype (4d to 1a, 1a to 1b, and 4d to 1) between the two episodes of viremia. In two individuals (Patients 6 and 7), the paired sequences were closely related, falling into the same monophyletic clade. However, each sequence in the pair is more closely related to reference strains than to each other, suggesting that reinfection is the most likely scenario rather than late relapse (although this cannot be ruled out entirely). In these two cases, the ALT increased to 8 and 18 × upper limit of normal at the time of the second viremia with the second episode occurring more than 4 years after the initial infection. Both received treatment with 24 weeks of pIFN/RBV for the initial episode. We report increasing numbers of HIV-positive men who have sex with men infected with HCV through sexual routes, experiencing second episodes of hepatitis C viremia after clearance of the initial infection. Viremia was associated with biochemical evidence of hepatic inflammation with genotype and phylogenetic data supporting the possibility of reinfection rather than late recurrence in the majority There is also evidence of high-risk sexual behavior as evidenced by concurrent STI diagnoses. Immune responses to HCV do not provide absolute protection against subsequent reinfection. Chimpanzees exposed to viral challenge with ensuing clearance are highly susceptible to a new challenge.9 Although reinfection with HCV has been reported in other clinical settings, this is the first report of a cohort reinfected after sexual exposure. In general, the incidence of reinfection reported varies widely with rates of one to eight cases per 100 person-years10 to 31 cases per 100 person-years11 having been published. Reinfection remains strongly related to levels of ongoing exposure risk. In two cases, we demonstrated that reinfection was the likely cause of viremia 3 to 4 months after an ETR. Previously this would have been considered likely to have been a relapse. This, therefore, has important implications. Although individuals with relapse have a low probability of responding to further courses of pIFN/RBV, this may not be the case for reinfection. It is also possible that reinfection, rather than relapse, is responsible for reappearance of viremia during interferon treatment as illustrated by a recent case report.12 In view of the important distinction between reinfection and relapse during antiviral treatment, we would consider it justified to assess individuals with re-emergent viremia for the possibility that viremia represents reinfection rather than treatment failure. Such evaluation should include inquiry of recent sexual activity and drug use, virus genotyping, and consideration of viral sequencing with phylogenetic analysis. The study may be limited by the use of direct sequencing of the amplified PCR products. This method has a tendency to introduce bias in favor of the detection of the most dominant sequence. It is possible that the second episode of viremia may represent re-emergence of minor sequences; however, the extensive time periods between first and second episodes, elevation of ALT, and the absence of an identified HCV sanctuary site are more suggestive of reinfection than late relapse. Without intervention, cohorts are at continuing high risk of contracting further HCV infection. Not only is there evidence of ongoing risk behavior, but it is occurring in networks with a high prevalence of HCV. As a result of poor cellular immune responses associated with HIV infection, these individuals may be more susceptible to reinfection. This convergence of risk is supported by the increasing number of cases of reinfection that we have recorded over the last 4 years. Most importantly, effective education and behavioral intervention are imperative in this high-risk group. Rachael Jones, MRCP* Mark Nelson, MD, FRCP* Emma Low, MRCP* Mark Atkins, MRCP* David Asboe, MRCP* David Brown, MPhil† Sanjay Bhagani, FRCP† Geoff Dusheiko, FRCP† Mark Danta, MD, FRACP‡ Oliver Pybus, PhD§ *Chelsea and Westminster NHS Foundation Trust, London, UK †Royal Free and University College Medical School, London, UK ‡University of New South Wales, Sydney, Australia §Oxford University, Oxford, UK