Aims To examine trends in prescription opioid (PO) injection and to assess its association with hepatitis C virus (HCV) seroconversion among injection drug users (IDUs).
Design Prospective cohort study.
Setting Montreal, Canada.
Participants HCV-negative IDUs at baseline, reporting injection in the past month.
Measurements Semi-annual visits included HCV antibody testing and an interview-administered questionnaire assessing risk behaviours. HCV incidence rate was calculated using the person–time method. Time-updated Cox regression models were conducted to examine predictors of HCV incidence.
Findings The proportion of IDUs reporting PO injection increased from 21% to 75% between 2004 and 2009 (P < 0.001). Of the 246 participants (81.6% male; mean age 34.5 years; mean follow-up time 23 months), 83 seroconverted to HCV [incidence rate: 17.9 per 100 person-years; 95% confidence interval (CI) 14.3, 22.1]. Compared to non-PO injectors, PO injectors were more likely to become infected [adjusted hazard ratio (AHR): 1.87; 95%CI:1.16, 3.03]. An effect modification was also found: PO injectors who did not inject heroin were more likely to become infected (AHR: 2.88; 95%CI: 1.52, 5.45) whereas no association was found for participants using both drugs (AHR: 1.19; 95% CI: 0.61, 2.30). Other independent predictors of HCV incidence were: cocaine injection, recent incarceration and >30 injections per month.
Conclusions Prescription opioid injectors who do not inject heroin are at greater risk for HCV seroconversion than are those injecting both heroin and prescription opioids. Important differences in age, behaviour and social context suggest a need for targeted outreach strategies to this population.