Adverse events of opioids may restrict their use in non-cancer pain. Analysis of the incidence of common adverse events in trials conducted in non-cancer pain has usually been limited to opioids used to treat severe pain according to the WHO three-step ladder. To examine the incidence of common adverse events of opioids in non-cancer pain, a systematic review and meta-analysis of information from randomised trials of all opioids in non-cancer pain was undertaken. Studies used were published randomised trials of oral opioid in non-cancer pain, with placebo or active comparator. Thirty-four trials with 5,546 patients were included with 4,212 patients contributing some information on opioid adverse events. Most opioids used (accounting for 90% of patients) were for treating moderate rather than severe pain. Including trials without a placebo increased the amount of information available by 1.4 times. Because of clinical heterogeneity in condition, opioid, opioid dose, duration, and use of titration, only broad results could be calculated. Use of any oral opioid produced higher rates of adverse events than did placebo. Dry mouth (affecting 25% of patients), nausea (21%), and constipation (15%) were the most common adverse events. A substantial proportion of patients on opioids (22%) withdrew because of adverse events. Because most trials were short, less than four weeks, and because few titrated the dose, these results have limited applicability to longer-term use of opioids in clinical practice. Suggestions for improved studies are made.