OBJECTIVE To assess whether adding codeine to paracetamol has an additive analgesic effect; to assess the safety of paracetamol-codeine combinations versus paracetamol alone. METHODS Systematic literature review with meta-analysis, methodological quality of published trials being scored by means of 13 predefined criteria. METHODS 24 of 29 trials that met the inclusion criteria. Models studied in the trials were postsurgical pain (21), postpartum pain (one), osteoarthritic pain (one), and experimentally induced pain (one). METHODS Dosages ranged from 400 to 1000 mg paracetamol and 10 to 60 mg codeine. METHODS The sum pain intensity difference (efficacy analysis) and the proportion of patients reporting a side effect (safety analysis). RESULTS Most trials were considered of good to very good quality. Only the single dose studies could be combined for analysis of analgesic efficacy. Pooled efficacy results indicated that codeine added to paracetamol provided a 5% increase in analgesia on the sum pain intensity difference. This effect was comparable to the difference in analgesic effect between codeine and placebo. The cumulative incidence of side effects with each treatment was comparable in the single dose trials. In the multidose studies a significantly higher proportion of side effects occurred with paracetamol-codeine preparations. CONCLUSIONS The difference is analgesic effect between paracetamol-codeine combinations and paracetamol alone was small but statistically significant. In the multidose studies the proportion of patients reporting a side effect was significantly higher with paracetamol-codeine combinations. For occasional pain relief a paracetamol-codeine combination might be appropriate but repeated use increases the occurrence of side effects.