Background: One important public health issue associated with opioid use today is the risk of hepatitis C (HCV) infection. Although methadone maintenance may help to decrease HCV-related risk practices, HCV risk behaviors persist and are strongly associated with specific substance use patterns, mental status and social context. The ANRS-Methaville study gave us the opportunity to better disentangle the different relationships between these various factors and HCV risk practices. Methods: The ANRS-Methaville multisite randomized trial was designed to assess the feasibility of initiating methadone in primary care by comparing it with methadone initiation in specialized centers. This study recruited 195 participants initiating methadone maintenance and followed up for 12 months. Longitudinal data from this trial was used to acquire a greater understanding of HCV risk practices and their pattern of correlates in this population. We selected 176 patients who had data on HCV risk practices at M0 and M12, accounting for 312 visits. HCV risk practices were defined as follows: sharing needles or syringes, sharing drug paraphernalia, getting a tattoo or having a piercing in a non-professional context, sharing toiletry items. To identify factors associated with HCV risk practices, we performed a mixed logistic regression analysis. Results: HCV risk practices were reported by 19% and 15% of participants at baseline and M12, respectively. After adjustment for age, cocaine use and alcohol dependence as well as suicidal risk, living in a couple with a non-drug user and in a couple with a drug user were both independent predictors of HCV risk practices (OR[CI95%] = 4.16 [1.42-12.12]; OR[CI95%] = 9.85 [3.13-31.06], respectively).