We have conducted a meta-analysis of the clinical effects of morphine and hydromorphone to compare their benefit in analgesia. Embase and Medline were searched with an end-date of June 2009 for randomized, controlled trials or observational studies that addressed comparative analgesic and side-effects or particular side-effects. Two researchers independently identified included studies and extracted the data. Estimates of opioid effects were combined by using a random-effects model. Meta-analysis of eight studies suggested that hydromorphone (494 patients) provides slightly better (P=0.012) clinical analgesia than morphine (510 patients). The effect-size was small (Cohen's d=0.266) and disappeared when one study was removed, although the advantage of hydromorphone was more evident in studies of better quality (Jadad's rating). Side-effects were similar, for example, nausea (P=0.383, nine studies, 456 patients receiving hydromorphone and 460 morphine); vomiting (P=0.306, six studies, 246 patients receiving hydromorphone and 239 morphine); or itching (P=0.249, eight studies, 405 patients receiving hydromorphone, 410 morphine). This suggests some advantage of hydromorphone over morphine for analgesia. Additional potential clinical pharmacological advantages with regard to side-effects, such as safety in renal failure or during acute analgesia titration, are based on limited evidence and require substantiation by further studies.