Gastric bypass better for type 2 diabetics

Roux-en-Y gastric bypass (RYGB) leads to greater weight loss and a greater rate of remission of diabetes than sleeve gastrectomy or medical therapy, according to a paper, ‘Estimated Cost-effectiveness of Medical Therapy, Sleeve Gastrectomy, and Gastric Bypass in Patients With Severe Obesity and Type 2 Diabetes’, published in JAMA Network Open. Interestingly, gastric bypass surgery also was projected to produce the best results regardless of diabetes severity.

Using data from previous studies and databases, researchers at Columbia University's Vagelos College of Physicians and Surgeons, led by Dr Chin Hur, created a model to investigate the effectiveness, quality of life gains, costs, and complications of gastric bypass, sleeve gastrectomy, and medical therapy among patients over a five-year period. The authors believe that this study is the first to consider diabetes severity in a comparison of gastric bypass and sleeve gastrectomy.

"Determining which groups may benefit from a specific strategy is an important step toward personalized medicine," explained Hur. "Our study suggests that in most cases, gastric bypass is the preferred strategy when looking at a five-year time frame, despite higher upfront surgical costs and complications, and becomes even more cost-effective when considered over 10 or 30 years."


For the study, the researchers sought to estimate the cost-effectiveness of medical therapy, sleeve gastrectomy and RYGB among patients with severe obesity and T2D, stratified by T2D severity. Their economic evaluation used a microsimulation model to project health and cost outcomes of medical therapy, sleeve and RYGB over five years. Time horizons varied between 10 and 30 years in sensitivity analyses.


The model inputs were derived from clinical trials, large cohort studies, national databases and published literature. Probabilistic sampling of model inputs accounted for parameter uncertainty. Estimates of US adults with severe obesity and T2D were derived from the National Health and Nutrition Examination Survey. Data analysis was performed from January 2020 to August 2021.

Quality-adjusted life-years (QALYs), costs (in 2020 US dollars) and incremental cost-effectiveness ratios (ICERs) were projected, with future cost and QALYs discounted 3.0% annually. It was determined as cost-effective if the ICER was less than US$100,000 per QALY. The preferred strategy resulted in the greatest number of QALYs gained while being cost-effective.


The model simulated 1,000 cohorts of 10,000 patients, of whom 16% had mild T2DM, 56% had moderate T2DM and 28% had severe T2DM at baseline. The mean age of simulated patients was 54.6 years (95% CI, 54.2-55.0 years), 61.6% (95% CI, 60.1%-63.4%) were female and 65.1% (95% CI, 63.6%-66.7%) were non-Hispanic White.


The outcomes revealed that compared with medical therapy over five years, RYGB was associated with the most QALYs gained in the overall population (mean, 0.44 QALY; 95% CI, 0.21-0.86 QALY) and when stratified by baseline T2DM severity: mild (mean, 0.59 QALY; 95% CI, 0.35-0.98 QALY), moderate (mean, 0.50 QALY; 95% CI, 0.25-0.88 QALY) and severe (mean, 0.30 QALY; 95% CI, 0.07-0.79 QALY). RYGB was the preferred strategy in the overall population (ICER, US$46 877 per QALY; 83.0% probability preferred) and when stratified by baseline T2D severity: mild (ICER, US$36 479 per QALY; 73.7% probability preferred), moderate (ICER, US$37 056 per QALY; 85.6% probability preferred), and severe (ICER, US$98 940 per QALY; 40.2% probability preferred).

Crucially, they noted that the cost-effectiveness of RYGB improved over a longer time horizon.

“In this study, over five years, RYGB was projected to result in greater weight loss and T2D remission rates than SG and medical therapy in US adults with severe obesity (BMI ≥40) and T2D, regardless of T2D severity at baseline,” the authors concluded. “Despite its higher upfront surgical costs, RYGB was estimated to be the most cost-effective treatment over 5 years and became even more cost-effective over longer time horizons (eg, 10 and 30 years).”

To access this paper, please click here