LAUSR

OBJECTIVE To estimate all-cause and overdose crude mortality rates and standardized mortality ratios among people prescribed opioids for chronic noncancer pain and risk of overdose death in this population relative to people with similar clinical profiles but not prescribed opioids. DESIGN Systematic review and meta-analysis. METHODS Medline, Embase, and PsycINFO were searched in February 2018 and October 2019 for articles published beginning 2009. Due to limitations in published studies, we revised our inclusion criteria to include cohort studies of people prescribed opioids, excluding those studies where people were explicitly prescribed opioids for the treatment of opioid use disorder or acute cancer or palliative pain. We estimated pooled all-cause and overdose crude mortality rates using random effects meta-analysis models. No studies reported standardized mortality ratios or relative risks. RESULTS We included 13 cohorts with 6,029,810 participants. The pooled all-cause crude mortality rate, based on 10 cohorts, was 28.8 per 1000 person-years (95% CI = 17.9-46.4), with substantial heterogeneity (I2 = 99.9%). The pooled overdose crude mortality rate, based on six cohorts, was 1.1 per 1000 person-years (95% CI = 0.4-3.4), with substantial heterogeneity (I2 = 99.5%), but indications for opioid prescribing and opioid exposure were poorly ascertained. We were unable to estimate mortality in this population relative to clinically similar populations not prescribed opioids. CONCLUSIONS Methodological limitations in the identified literature complicate efforts to determine the overdose mortality risk of people prescribed opioids. There is a need for large-scale clinical trials to assess adverse outcomes in opioid prescribing, especially for chronic noncancer pain.