Reporting of adverse effects in clinical trials should be improved: lessons from acute postoperative pain.

We assessed the quality of assessment and reporting of adverse effects in randomized, double-blind clinical trials of single-dose acetaminophen or ibuprofen compared with placebo in moderate to severe postoperative pain. Reports were identified by systematic searching of a number of bibliographic databases (e.g., MEDLINE). Information on adverse effect assessment, severity and reporting, patient withdrawals, and anesthetic used was extracted. Compliance with former guidelines for adverse effect reporting was noted. Fifty-two studies were included; two made no mention of adverse effects. No method of assessment was given in 19 studies. Twenty trials failed to report the type of anesthetic used, eight made no mention of patient withdrawals, and nine did not state the severity of reported adverse effects. Only two studies described the method of assessment of adverse effect severity. When all adverse effect data were pooled, significantly more adverse effects were reported with active treatment than with placebo. For individual adverse effects, there was no difference between active (acetaminophen 1000 mg or ibuprofen 400 mg) and placebo; the exception was significantly more somnolence/drowsiness with ibuprofen 400 mg. Ninety percent of trials reporting somnolence/drowsiness with ibuprofen 400 mg were in dental pain. All studies published after 1994 complied with former guidelines for adverse effect reporting. Different methods of assessing adverse effects produce different reported incidence: patient diaries yielded significantly more adverse effects than other forms of assessment. We recommend guidelines for reporting adverse effect information in clinical trials.