1110-P: Effect of Raised Alanine Transaminase (ALT) Levels on HbA1c in the Association of British Clinical Diabetologists (ABCD) Nationwide Audits of SGLT2 Inhibitors (SGLT2i)

Introduction: Alanine transaminase (ALT) is an important marker for Nonalcoholic Fatty Liver Disease (NAFLD), and its levels are reduced following treatment with SGLT2i. However, how raised ALT levels and hence NAFLD affects HbA1c response to SGLT2i is not known. Methods: Data for this study was obtained from the ABCD nationwide audits of patients with type 2 diabetes started on SGLT2i. We used a gradient boosting machine learning algorithm (GBM) to identify if ALT is an important predictor of glycaemic response to SGLT2i. The results of the GBM model were confirmed using linear regression analysis where the absolute drop in HbA1c was modelled as a dependent variable with baseline ALT as independent variable adjusted for relevant covariates. Results: We studied 9,609 patients initiated on Empagliflozin (n=5061), Dapagliflozin (n=3711) or Canagliflozin (n=837). The mean age of the study population was 59(±10.3), 62% male with a mean weight of 96.6(±18.8), baseline eGFR 81(±11) and baseline HbA1c 9.06% (±1.62). At a median 5.8 months follow-up, the mean HbA1c drop was 0.81% and was similar in all three types of SGLT2i. The drop in HbA1c was 0.62%, 0.78% and 1.01% in 1st 2nd and 3rd quartiles of baseline ALT, respectively (P-Anova <0.0001). In the GBM analysis the important predictors of glycaemic response were baseline weight (relative influence (RI)=71%), and baseline HbA1c (RI= 7.2%), baseline ALT (RI= 5.6%) baseline creatinine (RI= 4.9%). The model accuracy was 0.73 (0.71-0.75) and area under the curve was 0.83. In the linear regression model, higher baseline ALT was associated with a larger drop in HbA1c (beta=0.007, P<0.0001) after adjustments for baseline weight, weight loss, baseline HbA1c, and duration of diabetes. Conclusion: Higher baseline ALT levels are associated with a more significant SGLT2 induced HbA1c drop and SGLT2 inhibitors and are likely to be more effective in those with NAFLD. Disclosure H. Deshmukh: None. A. Bickerton: None. S.M. Phillips: None. A. Evans: Other Relationship; Self; AstraZeneca. R.P. Raghavan: Research Support; Self; Allergan plc. Speaker’s Bureau; Self; Abbott, AstraZeneca, Boehringer Ingelheim Pharmaceuticals, Inc., Lilly Diabetes, Napp Pharmaceuticals, Takeda UK. Other Relationship; Self; Napp Pharmaceuticals. D.K. Sennik: Speaker’s Bureau; Self; Boehringer Ingelheim Pharmaceuticals, Inc., Novo Nordisk Inc., Sanofi. A. Rohilla: None. K. Adamson: None. M. Yadagiri: None. I.W. Gallen: None. R.E. Ryder: Consultant; Self; GI Dynamics Inc. Other Relationship; Self; Novo Nordisk A/S. T. Sathyapalan: Research Support; Self; Amgen, Boehringer Ingelheim Pharmaceuticals, Inc., Eli Lilly and Company, Novo Nordisk A/S. Other Relationship; Self; Ipsen Biopharmaceuticals.