Intragastric balloons are new but commonly used devices for the treatment of obesity; however, their safety and efficacy have not been established. We report our results of a small, double-blind, randomized trial in which the effectiveness of intragastric balloons was compared with that of conventional medical therapy for obesity. Twenty-two patients, who were 21 to 77% over ideal body weight, were studied. Eleven underwent insertion of an intragastric balloon, and 11 underwent sham procedures. One patient with a gastric balloon withdrew from the study after 3 days. Weight loss at 2 to 3 months in the conventional therapy group averaged 2.8 kg; in the balloon-treated group, the mean weight loss was 5.8 kg (P greater than 0.15). Of the 10 balloons, 8 spontaneously deflated, and 1 was passed in the stools. We noted gastric erosions in five patients and multiple gastric ulcers in one. We conclude that the intragastric balloon was not clearly effective in inducing weight loss, had a high rate of spontaneous deflation, and was damaging to the gastric mucosa. Controlled trials should be done before similar weight-reduction devices are used in routine clinical practice.