Anastomotic leak after laparoscopic Roux-en-Y gastric bypass (LGB) is a major complication that must be recognized and treated early for best results. There is controversy in the literature regarding the reliability of upper GI series (UGI) in diagnosing leaks. LGB was performed in patients meeting NIH criteria for the surgical treatment of morbid obesity. All leaks identified at the time of surgery were repaired with suture and retested. Drains were placed at the surgeon's discretion. Postoperatively, UGI was performed by an experienced radiologist if there was a clinical suspicion of leak. From September 2001 until October 2004, a total of 553 patients (age 40.4 +/- 9.2 years, BMI 48.6 +/- 7.2) underwent LGB at UAB. Seventy-eight per cent (431 of 553) of patients had no clinical evidence suggesting anastomotic leak and were managed expectantly. Twenty-two per cent (122 of 553) of patients met at least one inclusion criteria for leak and underwent UGI. Four of 122 patients (3.2%) had a leak, two from anastomosis and two from the perforation of the stapled end of the Roux limb. No patient returned to the operating room without a positive UGI. High clinical suspicion and selectively performed UGI based on clinical evidence is reliable in detecting leaks.