Sixty patients with nonulcer dyspepsia took cimetidine 1 g daily or a placebo for 5 weeks in a double-blind trial. Thirty-four men and 26 women with normal double-contrast barium meals were included. Five percent had the endoscopic appearances of duodenitis. Symptoms improved during treatment in 62% of those taking cimetidine and in 54% of the placebo group (NS, p = 0.5). We suggest that significant duodenitis is uncommon in the absence of active or quiescent duodenal ulcer disease. From this study there is no evidence that cimetidine benefits those patients with normal double-contrast barium meals and no duodenitis endoscopically who nonetheless have symptoms similar to those of peptic ulcer disease.