Five-year opioid treatment study found persistent heroin use linked to crack but not cannabis escalation

Continued high-level heroin use during OST predicted continued high-level crack cocaine use (RRR 58.7), continued high-level alcohol use (RRR 1.2), and increasing unspecified drug use (RRR 1.7), but was associated with less high-and-increasing cannabis use (RRR 0.5).

Prospective observational cohort of 7,717 adults continuously enrolled in opioid substitution treatment with methadone or buprenorphine from 2008/09 to 2013/14 in England. Multi-level latent class growth analysis identified substance use trajectories across 11 six-monthly clinical reviews.

Understanding how different substance use patterns cluster during long-term opioid treatment helps clinicians identify which patients need additional support. The finding that cannabis use did not escalate alongside heroin non-response adds nuance to polysubstance use assumptions.

Crack cocaine, not cannabis, was the substance most tightly linked to persistent heroin use and treatment failure. This suggests that crack cocaine co-use should be a primary clinical concern in long-term opioid substitution programs.

Observational design cannot determine causation. Cannabis and other drug use were clinician-assessed, not biochemically verified at every timepoint. English treatment population may not generalize to other healthcare systems.