Introduction and Aims. Recent data suggest increasing prescription opioid and decreasing heroin use among street drug users, yet little is known on possible differential use characteristics and outcomes associated with these drugs. [While we recognise that, correctly, these populations would need to be labelled as opioid 'abusers' or 'non-medical users', we rely on the simpler terms 'use' and 'users' for the population under study within the wider context of them being engaged overall in illicit opioid use activities.] This study compared drug use, health, and socio-economic characteristics between heroin (H)-only, prescription opioid (PO)-only and mixed heroin and prescription (PO&H) users in a Canadian multi-site cohort of illicit opioid and other drug users (OPICAN). Design and Methods. Data from the most recent (2005) multi-component assessment of the H-only (n = 94), PO-only (n = 304) and PO&H (n = 86) cohort sub-samples were analysed. Based on bivariate analyses of variables of interest, a multinomial logistic regression analysis (MLRA) model was computed, comparing PO-only and PO&H groups to the H-only reference group, respectively. Results. H-only users were found in two of the seven study sites. Based on the MLRA, PO-only and PO&H users, compared to H-only users, were more likely to: be older, use benzodiazepines and cocaine, use drop-in shelters and less likely to use walk-in clinics. PO-only users were also more likely to: be white; receive legal income; use drugs by non-injection; have physical health problems; and use private physician services. Discussion and Conclusions. Our study underscores the increasing prevalence of PO compared to heroin use in the study population. Differences between PO-only and H-only users were more pronounced than differences between PO-only and PO&H users. PO-only use may be associated with lowered health risks and social burdens, yet concerns regarding polysubstance use and drug sourcing arise. Challenges for targeted interventions are discussed.