Spontaneous loss of hepatitis C virus RNA from serum is associated with genotype 1 and younger age at exposure^†

INTRODUCTION

Spontaneous resolution of hepatitis C virus (HCV) infection, defined by normalization of liver function tests (LFTs) and loss of HCV-RNA from serum, occurs in 26–36% of HCV-infected individuals [Santantonio et al., 2003, 2006; Micallef et al., 2006]. This group is of particular interest since an understanding of the mechanisms associated with spontaneous loss of HCV-RNA from serum may aid vaccine development. A number of factors are linked to spontaneous loss of HCV-RNA from serum including clinical factors such as symptomatic infection [Gerlach et al., 2003], HBV co-infection [Piasecki et al., 2004], and host factors including female gender [Alric et al., 2000; Micallef et al., 2006] and white ethnicity [Thomas et al., 2000; Piasecki et al., 2004].

Innate and host immune responses to HCV are likely to play an important role in spontaneous loss of HCV-RNA from serum, which has been linked in some, but not all studies with: genetic polymorphisms that modify IL-10, TNF-α, or IFN-γ production [Lio et al., 2003; Mangia et al., 2004; Minton et al., 2005; Kusumoto et al., 2006]; polymorphisms in interferon-induced genes [Knapp et al., 2003] and HLA class II alleles [Minton et al., 1998]. Recent studies have demonstrated a significant association between a single polymorphism upstream of the IL28B gene (which encodes the type III interferon IFN-λ3) and both spontaneous [Thomas et al., 2009] and treatment-induced [Ge et al., 2009] HCV resolution. In a large cohort of women infected with HCV from a single source, HLA-B27 allele was present more frequently in non-viremic individuals compared to those with viremia [McKiernan et al., 2004].

Viral factors have also been reported to influence spontaneous loss of HCV-RNA from serum. A prospective study of 92 anti-HCV positive males with a history of injecting drug use demonstrated a greater prevalence of genotype 3 in individuals who cleared HCV-RNA from serum spontaneously (86% of type 3 cleared compared with 7% of type 1) [Lehmann et al., 2004]. Infection with genotype 1, specifically subtype 1b, has been linked to chronic infection [Amoroso et al., 1998; Hwang et al., 2001]. Conversely, in a large study, Harris et al. [2007] suggested that HCV genotype 1 infection was more likely to be associated with spontaneous loss of HCV-RNA from serum than non-genotype 1 infection. A recent meta-analysis of 31 longitudinal studies with 675 study subjects concluded that the only factors significantly associated with loss of HCV-RNA from serum were female gender and acute symptomatic HCV infection [Micallef et al., 2006].

The aim of this study was to describe the prevalent HCV genotypes in a cohort of patients presenting to the hepatology service at Addenbrooke's Hospital, Cambridge and to investigate whether spontaneous loss of HCV-RNA from serum was linked to infecting HCV genotype.

MATERIALS AND METHODS

RESULTS

Of the 321 specimens, 37 were untypable, 4 were mixed genotypes, 138 were type 1, 118 were type 3, and 24 were other types. For the primary analysis those that were untypable or of mixed type were not included, leaving 280 typed specimens of which 62 were from non-viremic patients (Table I). The univariable analyses demonstrated that spontaneous loss was more likely in those who were UK born, who were younger at acquisition and who were infected with type 1 virus (Table II). In the multivariable analysis the effect of type 1 remained similar with a RR compared to type 3 of 2.07 (95% CI: 1.25, 3.43). Genotypes 2, 4, 5, and 6 showed some evidence of having a higher spontaneous loss than type 3 (RR = 2.94), although this was based on small numbers (n = 24) and was not significant when using logistic regression (P = 0.10) (Table II). In the multivariable analysis, likelihood of spontaneous loss remained significantly associated with those born in the UK and those who were <20 compared to ≥20 years old at acquisition (RR = 0.42; 95% CI: 0.25, 0.72, for ≥20 vs. <20) (Table II).

Further analyses on all individuals initially included in the study (n = 321), which took into account the 41 that were non-typable or of mixed type, were undertaken (Table II). Age at exposure and country of birth remained significantly associated with spontaneous loss of HCV-RNA from serum with broadly similar RR estimates. Additionally, excess alcohol was associated with a reduced likelihood of spontaneous loss of HCV RNA in this group (RR = 0.61; 95% CI: 0.43, 0.85). Genotype could not be assessed in this analysis since those that were non-typable (with one exception) were non-viremic.

DISCUSSION

Early treatment of HCV can prevent chronic infection. However, antiviral therapy for HCV infection has substantial side effects and is costly. Spontaneous loss of HCV-RNA from serum occurs in around 20–30% of people with acute HCV infection [Santantonio et al., 2003; Micallef et al., 2006]. It would be advantageous to identify these people prior to therapy. Spontaneous loss of HCV-RNA from serum occurs within 3 months of infection [Gerlach et al., 2003; Santantonio et al., 2005], so delaying treatment of individuals to target only those who remain HCV-RNA positive is feasible [Santantonio et al., 2006]. For the strategy of delayed treatment to be possible, it would be helpful to identify factors, host or viral, that predict spontaneous loss of HCV-RNA from serum.

In this study, those without viremia were defined as negative for HCV-RNA on a number of occasions. It is presently unclear whether spontaneous loss of HCV-RNA from serum is associated with complete eradication of the virus including the liver, or whether low-level viral replication persists but is controlled by cellular and humoral immune responses. Several small studies suggest that HCV-RNA can be consistently undetectable in serum or plasma by laboratory assays, but be identifiable in serum and peripheral blood mononuclear cells (PBMCs) by molecular tests with enhanced sensitivity, in patients thought to have loss of HCV-RNA from serum [Pham et al., 2004; Radkowski et al., 2005]. Increasing the sensitivity of HCV-RNA detection is possible by ultracentrifugation of the serum prior to nucleic acid extraction [McHutchison et al., 1999; Forman and Valsamakis, 2004; Bartolome et al., 2007]. The HCV quantitation assay utilized in this study, demonstrated good sensitivity (25 IU/ml), comparable to commercial assays. HCV-RNA was not detected in any of those HCV-RNA negative in serum despite ultracentrifugation prior to extraction.

Thirty-seven isolates (36.3%) were negative by HCV serotyping, which was probably due to degradation of HCV-specific antibodies during storage. Serotyping failure could also be a result of loss of antibody response with time in resolved infection, especially in individuals with a long time period between exposure and subsequent testing. The assay should not exhibit different genotype sensitivity which might have resulted in bias. The possibility of bias was investigated by examining the optical density (OD) readings of all serotyped isolates. No significant difference in OD was observed between genotype 1 and type non-1 isolates. However, the possibility of over-representation of a particular serotype amongst those which failed cannot be completely ruled out.

Despite the steps taken to ensure that samples were HCV-RNA negative, the possibility of occult HCV infection in individuals presumed to have cleared the virus cannot be ruled out. Occult HCV infection has been recently reported and is defined as the presence of HCV-RNA in the liver without serum HCV-RNA [Castillo et al., 2004]. It has been reported in two clinical situations: in anti-HCV negative, serum HCV-RNA negative patients with abnormal LFTs and in anti-HCV positive subjects with normal LFTs without serum HCV-RNA [Carreño et al., 2006]. In a small study on 12 patients who were anti-HCV positive, serum HCV-RNA negative with normal LFTs, 83% of these individuals had both genomic and anti-genomic HCV-RNA strands detected in the liver, demonstrating HCV replication in the majority of these patients [Carreño et al., 2006]. Interestingly, all the infections were found to be with subtype 1b. Recently a study examining liver biopsies from HCV non-viremic individuals found evidence of an ongoing immune response in the liver, which also often demonstrated fibrosis, supporting the view that HCV persists in the liver of most serum HCV-RNA negative individuals [Hoare et al., 2008]. These findings suggest that the term “spontaneous clearance” must be used with care, as these individuals may still have genomic HCV-RNA in liver tissue.

The findings of the present study support those of Harris et al. [2007]. In their study, the infecting type was ascertained in a large cohort of individuals who were viremic (n = 508) or had spontaneous loss of HCV-RNA from serum (n = 96). That study showed that the prevalence of genotype 1 in those who were HCV-RNA negative in serum was 69% and in those that had remained viremic was 51%, suggesting that individuals with HCV genotype 1 infection were more likely to become HCV-RNA negative (OR = 0.47; 95% CI: 0.29–0.78; P = 0.003). The study cohort consisted of patients enrolled in the UK HCV National Register and most of these patients (90%) had been traced during the National HCV Lookback Programme and had acquired their infection on a known date via transfusion of infected blood (prior to routine HCV screening) [Harris et al., 2000]. The population group in the present study is very different. Unlike the aforementioned study, many of the patients in this study would have been identified by virtue of their clinical symptoms and referred to the Hepatology clinic for further investigation post-testing. As such, cohorts of patients from tertiary referral centers are more likely to be biased towards symptomatic disease [Sweeting et al., 2006]. Also, in contrast to the study by Harris and colleagues, “date of infection” was often a crude estimate, as high-risk activities, like intravenous drug use, gave an uncertain date for HCV acquisition.

The results from the present study strengthen the evidence that patients with genotype 1 infections are more likely to become spontaneously HCV-RNA negative in serum. Infecting genotype was significantly associated with non-viremia in both univariable and multivariable statistical analyses. Individuals who had type 1 infections were more than twice (2.07×) as likely to become HCV-RNA negative. Over 91% of HCV infections in this study were genotype 1 or 3, which reflects UK epidemiology [Health Protection Agency, 2009] This study also demonstrated that genotypes 2, 4, 5, and 6 were associated with non-viremia (RR = 2.94) although due to the low numbers of these genotypes, the finding was not significant when using logistic regression (P = 0.10). Additionally those who acquired their infections at an older age were less likely to become HCV-RNA negative, with a RR of 0.42 for those infected at ≥20 years of age compared to <20 years of age.

The suggestion that genotype 1 infections are more likely to spontaneously lose HCV-RNA from serum than type 3 infections is contrary to certain other studies [Amoroso et al., 1998; Hwang et al., 2001; Lehmann et al., 2004]. Although spontaneous loss of HCV-RNA from serum may be more likely in genotype 1 infection, it is thought that, in those individuals with a chronic genotype 1 infection, disease may be more severe [Pozzato et al., 1994; Nousbaum et al., 1995; Lopez-Labrador et al., 1997; Tanaka et al., 1998; Harris et al., 2007]. It is plausible that genotype 1 infections might be more aggressive and therefore more likely to provoke an immune response leading to resolution, but if the immune response fails then the disease may be more severe. The association between younger age at exposure and spontaneous loss of HCV-RNA from serum demonstrated in this study has been reported in hemophiliacs [Messick et al., 2001; Zhang et al., 2006] and in childhood HCV infection [Yeung et al., 2007]. Older age at exposure is thought to be linked to increased risk of progression to cirrhosis [Minola et al., 2002]. It is pertinent to mention that in the study described, the age of exposure was, for the majority, a crude estimate which relied on each individual's recollection of risk incident. For those, such as intravenous drug users, who are involved continually in risk behavior, the date of infection may be inaccurate. The association between younger age at exposure and increased likelihood of spontaneous non-viremia is not surprising as the immune system is more responsive and able to fight infection at a young age and degenerates during the ageing process.

This study has demonstrated that age at exposure and HCV genotype are both associated with loss of HCV-RNA from serum. Thus, younger patients with acute genotype 1 HCV infection may be less likely to need antiviral treatment.