BACKGROUND: The importance of creating a small 10-20 ml upper pouch when performing a vertical banded gastroplasty (VBG) is often stated in the literature. In order to test the hypothesis that weight loss is superior in patients with a small upper pouch, we examined the weight loss curves for three different pouch sizes in our patients operated with VBG. METHOD: Retrospective analysis of patients operated with VBG at our institution between November 1986 and April 1994 was done. A modified Mason VBG was performed with intraoperative balloon measurements of the size of the upper pouch. Three groups were identified according to different pouch volumes: 20 ml (n = 65), 30 ml (n = 46), and >/= 40 ml (n = 47). RESULTS: Of the 198 patients operated with VBG, pouch volume measurement was successful in 158 patients. Mean pouch volume was 32 ml at 50.5 cm of water. Loss of body mass index at 6, 12, 24, 48, 60 months did not significantly differ in the three groups. The rate of late reoperative procedures was also similar in the three groups. The incidence of staple-line breakdown (SLB) and endoscopically verified esophagitis was higher in the >/= 40 ml group. CONCLUSION: We have been unable to demonstrate a difference in weight loss after VBG for differing pouch volumes. There is an increased rate of SLB and esophagitis in the group with largest pouch volume; however, length of follow-up was longest for this group.