Never injected, but hepatitis C virus-infected: a study among self-declared never-injecting drug users from the Amsterdam Cohort Studies

Introduction

Acute hepatitis C virus (HCV) infection is usually asymptomatic, and leads to chronic infection in 50–80% of patients [1]. Decades of chronic HCV infection can lead to liver cirrhosis and, in 1–5% of these patients, eventually to hepatocellular carcinoma as well [2]. In recent years, treatment success rates have substantially improved [3]. The most important mode of HCV transmission is through exposure to infected blood [1,4], and although sexual and household transmission have been described, they appear to happen only occasionally [5–7]. While never-injecting drug users (DU) do not share needles and/or syringes, their HCV prevalence is still higher than in the general population. Some studies suggest that HCV infection in never-injecting DU is associated with the sharing of drug-use paraphernalia, especially utilities used for consumption of crack, but others could not confirm these findings (reviewed in Ref. [8]) Alternatively, never-injecting DU might become infected with HCV through needle-stick accidents, household transmission, or sexual exposure. Recent review of research describing HCV among noninjecting DU points to a substantial gap in our knowledge of HCV in never-injecting DU, as no uniform risk factors could be identified [8].

The Amsterdam Cohort Study (ACS) among DU comprises a large group of never-injecting DU. It was designed to evaluate the sexual and blood borne transmission of HIV, other blood borne pathogens, and sexually transmitted diseases, as well as the determinants of transition to injecting drug use. This design has the potential to determine prevalence, incidence and risk factors for HCV infection among never-injecting DU. Additionally, we used phylogenetic analysis to investigate whether HCV strains isolated from never-injecting DU were closely related to strains circulating among injecting DU, or whether separate introductions had occurred through unrelated modes of transmission [9].

Materials and methods

The ACS among DU is an open, prospective cohort study initiated in 1985 [9]. Participation in the ACS is voluntary, and informed consent is obtained for every participant at intake. Recruitment is ongoing and in recent years has been directed in particular towards young DU. Both injecting and noninjecting DU are included and visit the Amsterdam Health Service every 4–6 months. Each study visit standardized questionnaires on (injecting) drug use and sexual risk behaviour are administered by trained research nurses and blood is drawn for prospective HIV testing and storage of serum. To study HCV prevalence and incidence we retrospectively tested stored serum from all participants having at least two visits between December 1985 and November 2005 (n = 1276), using the first available sample in each case. Individuals who were HCV-negative at ACS entry were tested for HCV antibodies at their last ACS visit before November 2005. On finding HCV seroconversion (defined as the presence of HCV antibodies in a previously seronegative individual), we tested samples taken between these two visits to determine the moment of seroconversion (defined as the midpoint between the last HCV seronegative sample and the first seropositive visit) [10]. Third generation commercial microparticle EIA system tests were used to detect HCV antibodies (AxSym HCV version 3.0; Abbott, Wiesbaden, Germany). 28.9% of the seropositive participants were tested at two study visits or more, all with consistent positive HCV antibody test results. Presence of HCV antibodies in all never-injecting DU was confirmed with Western blot (Deciscan HCV Plus immunoblot; BioRad, Marnes la Coquette, France). All ACS samples were stored at −80 °C.

All ACS participants since 1985 (n = 1640) were tested for HIV antibodies by enzyme-linked immunosorbent assays (ELISA), since 2003 (AxSym HIV Ag/Ab Combo, Abbott, Chicago, IL, USA), at each study visit. Results were confirmed by Western blot, since 1986, by HIV Blot version 2.2 (Genelab Diagnostics, pte Ltd, Singapore Science Park, Singapore).

Results

HCV RNA and phylogenetic analysis

Of 22 HCV-antibody positive never-injecting DU at ACS entry, 15 (68.2%) had detectable HCV RNA. The most frequent HCV genotype found was 3a (53.3%), followed by genotype 1a (26.7%) (Table 1). HCV genotypes 1a and 3a are generally associated with injecting drug use, and in injecting DU in the ACS they account for 252/317 (79%) of HCV infections for which genotyping was performed. Hence, the proportion of injection-related HCV genotypes was comparable among injecting DU and never-injecting DU ([11,14], unpublished data). Figure 1 shows a phylogenetic tree of HCV genotype 3a, comprising the 8 NS5B sequences obtained from never-injecting DU together with all available genotype 3a NS5B sequences (n = 65) from injecting DU ([11,14], unpublished data). Comparable to a pedigree, a phylogenetic tree illustrates the evolutionary relationships between genes or organisms or, in our case, the relationship among aligned NS5B sequences of several HCV genotype 3a viral variants. The more related two sequences are, the smaller the horizontal distance between those sequences in the tree. Based on phylogenetic analysis, sequences from never-injecting DU could not be distinguished from those of injecting DU. Sequences derived from never-injecting DU were interspersed with those of injecting DU, and they were not distinct phylogenetic isolates, nor did they form separate never-injecting DU clusters. This was observed also in HCV genotype 1a sequences (data not shown). The 3 never-injecting DU not infected with HCV genotype 1a or 3a harboured distinct strains of genotype 1b and 2a, which in the Netherlands and Belgium are linked to blood transfusion and nosocomial transmission rather than injecting drug use [15,16]. The proportion of never-injecting DU infected with these types (20%) was somewhat larger than the proportion observed among injecting DU (9%) in the ACS, but the difference was not statistically significant (P = 0.26, Pearson chi-square).

Interestingly, only one never-injecting (male) DU seroconverted during follow-up despite denying injecting drug use. He has regularly reported a steady sexual relationship with an injecting (female) DU who also participates in the ACS. She is a known injecting DU and became chronically infected with HCV genotype 2b at least 2.7 years before her male never-injecting DU sexual partner seroconverted for HCV. When comparing their two HCV sequences, the sequences were 100% identical (data not shown), making accidental exposure during household contacts or sexual transmission the likely route of transmission in this couple.

Discussion

In this cohort of never-injecting DU, the HCV prevalence was 6.3% (95% CI 3.7–8.8%). Although much lower than the prevalence in injecting DU in the same cohort (83.5%, [10]), this is substantially higher than in the general population in The Netherlands (estimated to be 0.1–0.4%, [17]). In literature, the HCV prevalence in never-injecting DU ranges from 2.3 to 35.3% [8]. However many studies were not specifically designed to measure HCV prevalence in never-injecting DU and often did not include questions on noninjection drug use risk factors for HCV.

The observed HCV incidence was very low at 0.049/100 PY, sixteen-fold lower than the HCV incidence estimated from the prevalent cases at study entry (0.79/100 PY). This suggests underreporting of past injecting drug use, which may have led to misclassification of injecting DU as never-injecting DU. However, this estimated HCV incidence has limitations: it does not take into account losses to follow-up in the unknown cohort that the prevalence sample is supposed to represent. Nor does it take differential recruitment of rates of healthy and infected subjects into account. However, when interpreted with caution, it could support our hypothesis of underreporting of injection drug use. Especially when injecting was incidental or stopped before entry in the ACS, participants may deny past risk behaviour, as has been described for HCV-positive blood donors in the Netherlands [15].

Starting injection later during follow-up was independently associated with a higher prevalence of HCV antibodies at entry. Of 352 never-injecting DU, 47 switched to injecting drug use after a median of 56 months (IQR 20–58 months). Of the 47, seven were among the 22 found to be HCV seroprevalent at entry. Again, this finding could suggest that some injecting DU were misclassified as never-injecting DU. They might have given socially desirable answers and denied injecting, since it is perceived among DU as damaging to their appearance and as overstepping a limit in the drug-using scene in Amsterdam [18]. Alternatively, the DU who started injecting during follow-up were already actively participating in the scene of injecting DU and were therefore more likely to become exposed to HCV through routes other than injecting drug use, such as needle stick accidents. Since HIV and HCV share transmission routes, the finding that HIV-positive never-injecting DU had a higher HCV prevalence at entry compared to HIV-negative participants, could imply that HIV-positive status is an indicator of unreported injecting drug use. On the other hand, HIV is transmitted sexually much more efficiently than HCV, and HCV might be transmitted more easily to and/or from HIV-positive individuals, compared to HIV-negative individuals, since HIV co-infection is associated with higher HCV RNA viral load [19].

Phylogenetic analysis revealed that the HCV sequences of never-injecting DU did not cluster together, suggesting that they were not a uniform group that became infected through sharing of noninjection drug use paraphernalia. In contrast, the noninjecting DU clustered together with the sequences found in injecting DU in the ACS (Fig. 1), indicating that they have close links with injecting DU and possibly underreport injection drug use. So although these DU did not report injecting drug use, they were infected from the pool of injecting DU (Fig. 1). Although self-reported data on methadone prescription in this cohort have been investigated and shown to be consistent with data from the Dutch Central Methadone Registration, self-reported data on sexually transmitted diseases (STD) were shown to be less consistent with diagnosis of such diseases [20,21]. In this study, based on the findings from logistic regression and phylogenetic analysis, some misclassification of ever-injecting DU seems likely in this never-injecting DU population.

Female sex was also associated with a higher HCV prevalence at entry, possibly indicating that women having sex with an HCV-positive partner are at higher risk for sexual transmission than men, as has been shown for HIV [22,23]. However, this gender difference has not yet been described for HCV [24]. We did not find an association between the presence of HCV antibodies and sexual behaviour. Furthermore, we observed only one HCV seroconversion during >2000 PY of follow-up, indicating that the risk of sexual transmission – and also household transmission-- is very small as has been demonstrated in partner studies among discordant heterosexual couples [25,26]. Unfortunately we were not able to perform risk factor analysis based on just one HCV seroconversion, but such analysis of incident cases in a longitudinal study would be more robust than a cross-sectional analysis of prevalent cases.

HCV has been detected on drug-use paraphernalia, and it has been hypothesized that HCV can be transmitted via these utilities (e.g. straws used for cocaine snorting) [27]. In line with our phylogenetic finding of nonclustering of never-injecting DU, we did not find statistically significant associations between cocaine use and the presence of HCV antibodies. However, questions on snorting paraphernalia were not included in the ACS questionnaires used in our study period. Some questions (e.g. having a tattoo, having a piercing) were added to the questionnaires in 2001 and thus yield data for only a portion of participants included in this study. A similar limitation holds true for the data on having received a blood transfusion, a question not asked after 1989, shortly before HCV screening of donor blood was introduced in developed countries. Moreover, never-injecting DU might potentially have received a blood transfusion when travelling to countries where transfusion is not yet safe. Although the direction of the effect of having received a blood transfusion was as expected (i.e. higher risk for those who have received a blood transfusion compared to those who did not), the main HCV genotype related to transmission by blood transfusion is genotype 1b, whereas the main genotypes circulating among never-injecting and injecting DU are 1a and 3a. Remarkably, in The Netherlands between 1997 and 2002, genotypes 1a and 3a, were found in 9/18 (50%) of HCV RNA-positive new donor candidates who most likely acquired HCV through a contaminated blood transfusion in the past [15].

In conclusion, although the incidence of HCV was very low in this study among never-injecting DU, the prevalence was much higher than in the general population. In the methadone outposts of the Amsterdam Health Service, HCV screening is offered every year irrespective of recent injecting drug use. Although, we could not distinguish whether the increased risk of HCV infection in never-injecting DU was related to underreporting of injection or to household or sexual transmission, HCV strains of never-injecting DU cluster with those found among injecting DU. HCV treatment has improved substantially since 2000 and is effective in up to 80–90% of patients [3]. Therefore, whatever the route of transmission, it is clear that routine HCV testing and treatment should be extended to both never-injecting and injecting DU.

Financial support

The Amsterdam Cohort Studies on HIV infection and AIDS, a collaboration between the Health Service of Amsterdam, the Academic Medical Centre of the University of Amsterdam, Sanquin Blood Supply Foundation and the University Medical Centre Utrecht, are part of the Netherlands HIV Monitoring Foundation and financially supported by the Netherlands National Institute for Public Health and the Environment. Website: http://www.amsterdamcohortstudies.org/. FZ, AK and TL are funded by the Netherlands organization for health research and development (ZonMW, grant no. 61200016, 71150001 and 91203005).