Predictive factors of type 2 diabetes remission after bariatric surgery: A multivariate analysis

Jerry T Dang, BSc, MD1, Caroline Sheppard, BSc1, David H Kim, BSc2, Xinzhe Shi, MPH, MMSc, BSc3, Chris de Gara, MB, MS, FRCS, Ed, Eng, , C, FACS3, Shahzeer Karmali, MD, FRCSC, FACS3, Daniel W Birch, MSc, MD, FRCSC, FACS3. 1Department of Surgery, University of Alberta, Edmonton, Alberta, Canada, 2Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta, Canada, 3Centre for the Advancement of Minimally Invasive Surgery (CAMIS), Royal Alexandra Hospital, Edmonton, Alberta, Canada

INTRODUCTION: The aim of this study is to identify predictive factors for remission of type 2 diabetes (T2DM) after bariatric surgery. Bariatric surgery has been shown to induce diabetic remission in some severely obese patients. However, patients respond differently to surgery. Some patients achieve complete remission, while others remain diabetic. There is limited data on the factors that predict remission and previous research has identified preoperative factors such as time from diabetes diagnosis and Hba1c. As such, identifying predictors of diabetic remission after bariatric surgery is useful to prognosticate and decide which diabetic patients will benefit most from bariatric surgery.

METHODS AND PROCEDURES: A retrospective review of all T2DM patients undergoing laparoscopic adjustable gastric band (LAGB), laparoscopic Roux-en-Y gastric bypass (LRYGB) and laparoscopic sleeve gastrectomy (LSG), from January 2008 to July 2014 was performed. Patient preoperative and postoperative clinical and biochemical data were collected and analyzed using univariate and multivariate logistic regression analysis to identify preoperative predictive factors of diabetic remission. Diabetic remission was defined as: absence of hypoglycemic medications, fasting blood glucose < 7 mmol/L and HbA1c < 6.5%.

RESULTS: Two hundred and forty-one T2DM patients underwent bariatric surgery. Twelve were excluded because of missing clinical or biochemical data. Two hundred and twenty-nine were included in logistic regression analysis with 46.2% (106) of patients achieving diabetic remission at one year.

In univariate analysis, patients with diabetic remission were younger, using less preoperative insulin and oral hypoglycemics and were less frequently diagnosed with hypertension than those without remission. They also had lower HbA1c, fasting blood glucose, creatinine and LDL. Both LRYGB and LSG had higher remission rates than LAGB, with remission rates of 58.7%, 39.0% and 11.5% for LRYGB, LSG and LAGB respectively.

Multivariate analysis confirmed that LRYGB had the highest odds of remission (p <0.001), while LSG had second highest (p = 0.035). Additionally, shorter T2DM duration (p = 0.018), less number of preoperative oral hypoglycemics (p = 0.013) and absence of long-acting insulin (p = 0.018) predicted diabetic remission in multivariate analysis after controlling for confounders and interactions.

CONCLUSIONS: Diabetic remission occurred in 46.2% of diabetic patients by one year post-bariatric surgery. Type of bariatric procedure (LRYGB and LSG), more recent diagnosis of T2DM, less preoperative oral hypoglycemics and absence of long-acting insulin were independent predictors of remission. This potentially helps clinicians in predicting which patients will achieve diabetic remission following bariatric surgery.

« Return to SAGES 2016 abstract archive