Comparative effectiveness of extended-release naltrexone versus buprenorphine-naloxone on treatment interruption: Comparing findings from a reanalysis of the X:BOT RCT and harmonized target trial emulation using population-based observational data

Background and aims: It is unclear if findings from randomized controlled trials (RCT) of medications for opioid use disorder apply to real-world settings. We estimated the effectiveness of buprenorphine-naloxone (BUP-NX) versus extended-release naltrexone (XR-NTX) on treatment interruption in a RCT and an observational study based on real-world data. Design: Target trial emulation to harmonize the protocol and statistical analyses of X:BOT (target trial) and the observational study (observational emulation). Baseline was randomization in the target trial and medically managed opioid withdrawal (MMOW) discharge in the observational emulation. Settings: X:BOT trial and Massachusetts Public Health Data Warehouse observational data (United States). Participants: The target trial included all X:BOT participants. The observational emulation trial included MMOW discharges from January 2014 to May 2016. Measurements: Treatment strategies were BUP-NX versus XR-NTX initiation within 28 days of baseline. The outcome was treatment interruption (earliest of treatment discontinuation, incarceration, MMOW readmission, death). We estimated the 24-week risk and risk difference. Findings: In the target trial, 94% (269/287) and 66% (187/283) of participants randomized to BUP-NX or XR-NTX initiated their assigned treatment within 28 days, respectively. In the observational emulation, BUP-NX and XR-NTX were initiated within 28 days in 9% (5209/59 076) and 3% (1813/59 076) of MMOW discharges, respectively. The adjusted 24-week treatment interruption risks (95% confidence interval) for BUP-NX and XR-NTX were 68% (60%,77%) and 72% (60%,83%) in the target trial [risk difference, −4 percentage points (pp; −17 pp,11 pp)] and 82% (81%,83%) and 93% (92%,95%) in the observational emulation [risk difference,-11 pp (−13 pp,-10 pp)]. Conclusions: Buprenorphine-naloxone might be superior to extended-release naltrexone in real-world settings where the majority of people struggle to remain on medications for opioid use disorder. Buprenorphine-naloxone initiators had a lower risk of treatment interruption than extended-release naltrexone initiators in an observational emulation, but similar risks in a randomized controlled trial, although confidence intervals were wide. Trial participation, study size and residual confounding may explain these differences.