To determine the relative efficacy and adverse effects of antidepressants and anticonvulsants in the treatment of diabetic neuroapathy and postherpetic neuralgia, published reports were identified from a variety of electronic databases, including Medline, EMBASE, the Cochrane Library and the Oxford Pain Relief Database, and from two previously published reviews. Additional studies were identified from the reference lists of retrieved reports. The relative benefit (RB) and number-needed-to-treat (NNT) for one patient to achieve at least 50 % pain relief was calculated from available dichotomous data, as was the relative risk (RR) and number-needed-to-harm (NH) for minor adverse effects and drug related study withdrawal. In diabetic neuropathy, 16 reports compared antidepressants with placebo (491 patient episodes) and three compared anticonvulsants with placebo (321). The NNT for at least 50 % pain relief with antidepressants was 3.4 (95 % confidence interval 2.6-4. 7) and with anticonvulsants 2. 7 (2. 2-3. 8). In postherpetic neuralgia, three reports compared antidepressants with placebo (145 patient episodes) and one compared anticonvulsants with placebo (225), giving an NNT with antidepressants of 2.1 (1. 7-3) and with anticonvulsants 3.2 (2.4-5). There was little difference in the incidence of minor adverse effects with either antidepressants or anticonvulsants compared with placebo, with 1VH (minor) values of about 3. For drug-related study withdrawal, antidepressants had an NNH (major) of 17 (11-43) compared with placebo, whereas with anticonvulsants there was no significant difference from placebo. Antidepressants and anticonvulsants had the same efficacy and incidence of minor adverse effects in these tzoo neuropathic pain conditions. There was no evidence that selective serotonin reuptake inhibitors (SSRIs) were better than older antidepressants, and no evidence that gabapentin was better than older anticonvulsants. In these trials patients were more likely to stop taking antidepressants than anticonvulsants because of adverse effects.