Background and Aims : Whether bariatric surgery should be encouraged for severely obese people with new or existing diabetes, specifically to control their diabetes, is a pressing question. Our goal was to utilize existing data from HealthPartners, Kaiser Permanente Northern California, and Kaiser Permanente Southern California to examine both the short and long-term outcomes for individuals with type 2 diabetes who have undergone bariatric surgery vs. individuals who have been treated non-surgically.
Methods : We are conducting a retrospective cohort study of adults with diabetes mellitus and a recorded BMI >35 who underwent bariatric surgery between 2002 to 2008 at one of three health systems and comparable diabetic patients who did not have bariatric surgery. Severe obesity was defined using height and weight data from the electronic medical record, and diabetes was defined using seven ICD-9, pharmacy, and laboratory criteria. Ongoing analyses will assess the association between bariatric surgery versus usual care on the following outcomes: resolution of diabetes, control of diabetes, diabetes medication discontinuation, change in fasting glucose and HbA1c, and change in weight (kg) and BMI. We will also examine the above outcomes in relation to common bariatric procedure type: open and lap gastric bypass vs. gastric banding vs. sleeve gastrectomy.
Results : We identified 4,693 adults with diabetes who underwent bariatric surgery between 2002 to 2008 at one of three health systems and 170,984 adults with diabetes and a BMI >35 who did not undergo surgery. Among the surgical cases, 193 (4.1%)subjects underwent lap gastric banding, 903 (19.2%) open gastric bypass, 3,460 (73.9%) lap gastric bypass, 123 (2.6%) gastric sleeve, and 14 (0.2%) biliopancreatic diversion. Only 2,643 bariatric cases had a BMI >35 available in the medical record prior to surgery for use in propensity modeling and multivariable analyses; these subjects will make up our final comparative cohort.
Conclusions : The HMORN provides a rich opportunity for bariatric surgery effectiveness research among patients with type 2 diabetes. Ongoing analyses (that will be completed in January 2009 and can be presented in March) will assess the rate of diabetes resolution, diabetes control, and adverse events across different bariatric procedures vs. usual care.