Nonulcer dyspepsia remains a difficult disorder to treat because it is a heterogeneous syndrome. Once patients with the irritable bowel syndrome, esophagitis, and other organic diseases are excluded, there remain patients with dyspepsia of unknown cause (termed "essential dyspepsia") and patients with dyspepsia plus symptoms of gastroesophageal reflux without esophagitis. The aim of this study was to determine whether cimetidine or pirenzepine is efficacious in relieving the symptoms of these latter subgroups. Sixty-two consecutive patients were studied who had chronic upper abdominal pain or nausea where endoscopy had shown no evidence of peptic ulceration, esophagitis, or malignancy; 47 had essential dyspepsia, and 15 had dyspepsia plus gastroesophageal reflux. They were initially randomized to either cimetidine or placebo, or pirenzepine or placebo. Patients continued each medication for 1 mo, and, after a washout period, crossed over when again symptomatic; 51 patients completed cimetidine and placebo, and 50 completed pirenzepine and placebo. The results showed that cimetidine was superior to placebo in decreasing the number of upper abdominal pain episodes weekly and the severity of pain, but the absolute improvement was small. Pirenzepine was not superior to placebo in decreasing symptoms.