To estimate the frequency of adverse effects associated with the use of the transdermal nicotine patch, we abstracted and analysed data from 47 reports of 35 clinical trials. The meta-analysis presented here represents a synthesis of data from 41 groups of nicotine patch recipients totalling 5501 patients, and 33 groups of placebo recipients totalling 3752 patients. Smoking abstinence was the primary outcome in 32 of the trials, and relief of colitis symptoms was the primary outcome in 2 of the trials; 1 study of contact sensitisation was included in the skin irritation analysis. The patch was clearly effective as an aid to smoking abstinence. Despite the large number of patients in the analysis, few adverse cardiovascular outcomes (myocardial infarction, stroke, tachycardia, arrhythmia, angina) were reported, and no excess of these outcomes was detected among patients assigned to nicotine-patch use. The incidences of several minor adverse effects were clearly elevated among the nicotine-patch groups, especially sleep disturbances, nausea or vomiting, localised skin irritation and respiratory symptoms, but the background rates and risk ratios varied considerably across studies. The incidence of nausea or vomiting appeared to be lowest when the patch dose was tapered. The results of this meta-analysis indicate that very large studies would be needed to assess the effect of the patch, if any, on serious, rare outcomes. These results also suggest that the rate of minor adverse effects might be lowered by modifying patch-use protocols.