A prolonged randomized, prospective, double-blind, crossover study, including a sham-sham-treated group, was undertaken to evaluate the efficacy and safety of a 500-mL gastric bubble (Ballobes; DOT ApS, Rödovre, Denmark) as an adjunct to diet, physical training, and behavioral modification. Only supermorbidly obese patients who fulfilled the usual criteria for surgery were admitted. A weight loss of 38 kg in the first 17 weeks and another 12 kg in the second 18 weeks could be achieved. The body mass index, the percentage of overweight and the loss in percentage of initial weight, paralleled this impressive weight loss. In the second period, a plateau effect occurred after the massive changes in the first period, and only one third of the changes in all parameters was seen. Stratification into a sham-sham, sham-balloon, balloon-sham, and balloon-balloon group did not show any statistical difference for all parameters between the four groups. The double-blind nature of the study was affirmed by the patient's correct judgment of the presence or absence of a balloon in only 21% of the balloon and 44% of the sham procedures. Gastrointestinal complications were infrequent and consisted of erosions (three patients), asymptomatic reflux oesophagitis (one patient), and asymptomatic gastric ulcer (one patient). Only the latter patient had elevated gastrin levels. One patient could not tolerate the balloon. All balloons remained airtight during both parts of the study for a mean of 123 days. This study confirmed the safety of the balloon, but no additional benefit could be ascribed to the balloon compared with a very low-calorie diet and medical and dietary support.