Jun 4, 2013
A Systematic Review
Published In: JAMA, The Journal of the American Medical Association, v. 309, no. 21, June 5, 2013, p. 2250-2261
Posted on RAND.org on June 01, 2013
IMPORTANCE: Bariatric surgery is beneficial in persons with a body mass index (BMI) of 35 or greater with obesity-related comorbidities. There is interest in using these procedures in persons with lower BMI and diabetes. OBJECTIVE: To assess the association between bariatric surgery vs nonsurgical treatments and weight loss and glycemic control among patients with diabetes or impaired glucose tolerance and BMI of 30 to 35. EVIDENCE REVIEW: PubMed, EMBASE, and Cochrane Library databases were searched from January 1985 through September 2012. Of 1291 screened articles, we included 32 surgical studies, 11 systematic reviews on nonsurgical treatments, and 11 large nonsurgical studies published after those reviews. Weight loss, metabolic outcomes, and adverse events were abstracted by 2 independent reviewers. FINDINGS: Three randomized clinical trials (RCTs) (N = 290; including 1 trial of 150 patients with type 2 diabetes and mean BMI of 37, 1 trial of 80 patients without diabetes [38% with metabolic syndrome] and BMI of 30 to 35, and 1 trial of 60 patients with diabetes and BMI of 30 to 40 [13 patients with BMI <35]) found that surgery was associated with greater weight loss (range, 14.4-24 kg) and glycemic control (range, 0.9-1.43 point improvements in hemoglobin A1c levels) during 1 to 2 years of follow-up than nonsurgical treatment. Indirect comparisons of evidence from observational studies of bariatric procedures (n ≈ 600 patients) and meta-analyses of nonsurgical therapies (containing more than 300 RCTs) support this finding at 1 or 2 years of follow-up. However, there are no robust surgical data beyond 5 years of follow-up on outcomes of diabetes, glucose control, or macrovascular and microvascular outcomes. In contrast, some RCT data of nonsurgical therapies show benefits at 10 years of follow-up or more. Surgeon-reported adverse events were low (eg, hospital deaths of 0.3%-1.0%), but data were from select centers and surgeons. Long-term adverse events are unknown. CONCLUSIONS AND RELEVANCE: Current evidence suggests that, when compared with nonsurgical treatments, bariatric surgical procedures in patients with a BMI of 30 to 35 and diabetes are associated with greater short-term weight loss and better intermediate glucose outcomes. Evidence is insufficient to reach conclusions about the appropriate use of bariatric surgery in this population until more data are available about long-term outcomes and complications of surgery.