Predictors of Type 2 Diabetes Remission and Relapse After Metabolic-Bariatric Surgery or Medical-Lifestyle Therapy in the ARMMS-T2D Study.
Abstract
Objective
Metabolic-bariatric surgery (MBS) is effective in promoting remission of type 2 diabetes, but this outcome is not universally achieved. We identified baseline patient characteristics that predict diabetes remission and subsequent relapse after various MBS procedures versus intensive medical-lifestyle intervention. RESEARCH
Design
AND
Methods
Alliance of Randomized Trials of Medicine vs. Metabolic Surgery in Type 2 Diabetes (ARMMS-T2D) is a 14-year, multicenter study of health outcomes in 262 patients with type 2 diabetes (68% women, age 49.9 ± 8.3 years, BMI 36.4 ± 3.5 kg/m2, HbA1c 8.5 ± 1.5%) previously randomized to MBS (n = 166) or medical-lifestyle management (n = 96). Predictors of remission (defined as HbA1c <6.5% on no glucose-lowering medications) and relapse were determined using univariate time-to-event analyses of prespecified candidate predictors plus agnostic, multivariate least absolute shrinkage and selection operator.
Results
Throughout a median of 11 years (maximum 14 years) of follow-up, 58% of participants randomized to MBS achieved diabetes remission versus 3% with medical-lifestyle management. Predictors of diabetes remission in univariate analyses included shorter diabetes duration, no insulin use, fewer glucose-lowering medication classes used, greater body weight, higher fasting C-peptide and insulin levels, and weight loss at 1 year (all P < 0.05). Diabetes remission predictors in multivariate analyses included baseline diabetes duration, insulin use, and weight change at 1 year (all P < 0.01). Diabetes relapse after initial remission in the MBS group was predicted by longer baseline diabetes duration and greater number of glucose-lowering medication classes used.
Conclusions
Baseline predictors of diabetes remission and subsequent relapse among participants randomized to one of three MBS procedures or a medical-lifestyle intervention should help inform personalized selection of diabetes treatment.