Prediction of Long-Term Diabetes Remission After RYGB, Sleeve Gastrectomy, and Adjustable Gastric Banding Using DiaRem and Advanced-DiaRem Scores

Purpose: DiaRem is a clinical scoring system designed to predict diabetes remission (DR) 1-year post-Roux-en-Y gastric bypass (RYGB). We examined long-term (2- and 5-year) postoperative DR prediction by DiaRem and an advanced-DiaRem (Ad-DiaRem) score following RYGB, sleeve gastrectomy (SG), and gastric banding (GB). Methods: We accessed data from a computerized database of persons with type 2 diabetes and BMI ≥ 30 kg/m ² who underwent RYGB, SG, or GB, and determined DR status 2- and 5-year postoperative according to preoperative DiaRem and the Ad-DiaRem calculated scores. Results: Among 1459 patients with 5-year postoperative diabetes status data, 53.6% exhibited DR. For RYGB, Ad-DiaRem trended to exhibit mildly improved predictive capacity 5-year postoperatively compared to DiaRem: Areas under receiver operating characteristic [AUROC] curves were 0.85 (0.76–0.93) and 0.78 (0.69–0.88), respectively. The positive predictive values (PPVs) detecting > 80% of those achieving DR (i.e., sensitivity ≥ 0.8) were 78.2% and 73.2%, respectively, and higher Ad-DiaRem scores more consistently associated with decreased DR rates. Following SG, both scores had an AUROC of 0.82, but Ad-DiaRem still had a higher PPV for predicting > 80% of those with 5-year postoperative DR (76.2% and 71.0%). Predictive capacity parameters were comparatively lower, for both scores, when considering DR 5-year post-GB (AUROC: 0.73 for both scores, PPV: 66.3% and 64.3%, respectively). Conclusions: Ad-DiaRem provides modest improvement compared to DiaRem in predicting long-term DR 5-years post-RYGB. Both scores similarly provide fair predictive capacity for 5-year postoperative DR after SG.