Systematic review and meta-analysis of the prevalence of prescription opioid dependence in patients with chronic non cancer pain (CNCP) treated with opioid painkillers

Background: Over the last two decades there has been a progressive rise in the use of opioid painkillers to treat chronic non cancer pain (CNCP) despite a lack of evidence for their effectiveness in the treatment of this type of pain. Long term opioid painkiller use is associated with substantial harms, including accidental overdose, dependence and abuse. We aimed to assess the prevalence of problematic opioid use including opioid painkiller dependence, misuse, addiction, abuse and aberrant behaviour in CNCP patients prescribed opioids.

Methods :We conducted systematic reviews and meta-analyses of prevalence for each of the previously described outcomes. Database searches (MEDLINE, Embase, PsycINFO, Ovid) were conducted from inception to January 2021 without language restrictions. We used random effects models for the meta-analyses and explored variation by age, sex, geographical location, study design, study setting and study quality.

Results: 5859 records were screened and data from 160 studies were included. Prevalence rates were as follows: 13%, 95% Confidence Interval (CI) 11%-17% for dependence, 32% (95% CI 21%-43%) for misuse, 15% (95% CI 7%-25%) for addiction, 6% (95% CI 5%-7%) for abuse and 32% (95% CI 23%-41%) for aberrant behaviour with high heterogeneity. Subgroup analyses showed differences in prevalence estimates by geography for all outcomes, by study design for abuse, dependence and misuse and by study setting for aberrant behaviour and dependence. Prevalence estimates did not differ by study quality, age or sex.

Conclusion: Problematic opioid use is common in CNCP patients treated with opioids. The variation in prevalence estimates by outcome may reflect the lack of clarity and consistency used when measuring and reporting problematic opioid use. Effective strategies are needed to prevent and treat problematic opioid use.