Validation of Ad-DiaRem and ABCD Diabetes Remission Prediction Scores at 1-Year After Roux-en-Y Gastric Bypass and Sleeve Gastrectomy in the Randomized Controlled Oseberg Trial.

Abstract

Prediction of type 2 diabetes (T2DM) remission is an important part of risk-benefit assessment before bariatric surgery. Advanced-DiaRem (Ad-DiaRem) and ABCD diabetes remission scores for sleeve gastrectomy (SG) and Roux-en-Y gastric bypass (RYGB) were calculated using baseline data. Differences in model discrimination using area under the curve of receiver operating curve (AUC-ROC) and model calibration were tested for complete remission (HbA1c ≤ 6.0% without antidiabetic medications) in the two groups. Optimal cutoff scores were calculated using the Youden index. We randomized 109 patients to either SG or RYGB. With one patient lost to follow-up in each group, the scores were calculated for 54 patients in the SG group and 53 patients in the RYGB group. Both models showed moderate predictive power without any significant difference between the groups: AUC-ROCs (95% CI) for the Ad-DiaRem score (SG versus RYGB) were 0.872 (0.780-0.964) versus 0.843 (0.733-0.954), p = 0.69, and for the ABCD score 0.849 (0.752-0.946) versus 0.750 (0.580-0.920), p = 0.32, respectively. Using optimal cutoff points derived from the whole study population, the actual proportion of diabetes remission was significantly higher than predicted for both the Ad-DiaRem and ABCD scores in the RYGB group. Diabetes duration and glycated haemoglobin predicted diabetes remission in the entire Oseberg population. Both the Ad-DiaRem and ABCD scores showed moderate ability to discriminate between those who achieved remission of T2DM and those who did not after SG and RYGB. Larger studies are needed for the identification of procedure-specific optimal cutoffs. Trial Registration ClinicalTrials.gov Identifier: NCT01778738.