Background: Although bariatric surgery is known to be effective in the short term, the durability of that effect has not been convincingly demonstrated over the medium term (>3 years) and the long term (>10 years). The authors studied the durability of weight loss after bariatric surgery based on a systematic review of the published literature. Methods: All reports published up to September, 2005 were included if they were full papers in refereed journals published in English, of outcomes after Roux-en-Y gastric bypass (RYGBP), and its hybrid procedures of banded bypass (Banded RYGBP) and longlimb bypass (LL-RYGBP), biliopancreatic diversion with or without duodenal switch (BPD±DS) or laparoscopic adjustable gastric banding (LAGB). All reports that had at least 100 patients at commencement, and provided ≥3 years of follow-up data were included. Results: From a total of 1,703 reports extracted, 43 reports fulfilled the entry criteria (18 RYGBP; 18 LAGB; 7 BPD). Pooled data from all the bariatric operations showed effective and durable weight loss to 10 years. Mean %EWL for standard RYGBP was higher than for LAGB at years 1 and 2 (67 vs 42; 67 vs 53) but not different at 3, 4, 5, 6 or 7 years (62 vs 55; 58 vs 55; 58 vs 55; 53 vs 50; and 55 vs 51). There was 59 %EWL for LAGB at 8 years, and 52 %EWL for RYGBP at 10 years. Both the BPD±DS and the Banded RYGBP appeared to show better weight loss than standard RYGBP and LAGB, but with statistically significant differences present at year 5 alone. The LL-RYGBP was not associated with improved %EWL. Important limitations include lack of data on loss to follow-up, failure to identify numbers of patients measured at each data point and lack of data beyond 10 years. Conclusions: All current bariatric operations lead to major weight loss in the medium term. BPD and Banded RYGBP appear to be more effective than both RYGBP and LAGB which are equal in the medium term.