HCV seroconversion among never-injecting heroin users at baseline: No predictors identified other than starting injection

BACKGROUND: Heroin users who do not inject constitute a large pool of drug users with a potentially important impact on public health. We aimed to estimate the incidence of hepatitis C virus (HCV) among heroin users who had never injected (NIDUS) at baseline, and the effect of starting injecting during follow-up, other percutaneous exposures, sharing snorting paraphernalia, cocaine/crack use, and risky sexual behaviour on HCV-seroconversion. METHODS: Prospective cohort of 305 HCV-negative NIDUs at baseline, aged 18-30 and street-recruited in three Spanish cities in 2001-2003. Computer-assisted personal interviews were conducted and dried blood-spot samples were collected. Bivariate and multivariable Poisson models were used. RESULTS: Among the 305 never-injectors who were HCV-negative at baseline, 197 (64.6%) were followed-up and 21 seroconverted [HCV-incidence rate=5.8/100 person-years at risk (pyar) (95% CI: 3.6-8.9)]. HCV incidence in new-injectors was 28.4/100pyar [(95% CI, 14.7-49.7) vs. 2.8/100pyar (95% CI, 1.3-5.4)] among NIDUs. Of the risk exposures considered, starting injecting was the only predictor of HCV-seroconversion [adjusted relative risk=10.1, 95% CI: 3.8-26.7]. CONCLUSION: The HCV-seroconversion rate was 10 times higher among new-injectors than never-injectors. No predictors other than starting injecting were found for HCV-seroconversion. Harm reduction interventions to prevent HCV infection should include prevention of drug injection.