End-stage renal disease (ESRD) is the most severe form of chronic kidney disease (CKD) and requires dialysis or renal transplant for survival. ESRD affects more than 600,000 individuals in the United States and over 40% of cases are diabetes-related. Results from recent clinical trials consistently showed that sodium-glucose cotransporter-2 (SGLT2) inhibitors improve renal outcomes among patients with type 2 diabetes mellitus (T2DM). Cost considerations, however, emphasize the need for designing cost-effective strategies to use the medicine in this class for ESRD prevention. Therefore, the goal of this project was to develop a risk prediction model that would help identify T2DM patients at high risk to develop ESRD. The study is based on a retrospective cohort of patients with T2DM who received health care services at a large, integrated healthcare system for at least two years between 2001 and 2016. Cox regression was used to identify the strongest predictors of ESRD among 93 candidate predictors (e.g., medical history, laboratory tests) obtained from the electronic medical records. The “high-risk” group included the highest risk 10% of patients according to the final model predictors. The study sample included 58,428 patients with a mean age of 61 years. Of these, 6350 (11%) had CKD at baseline. During a mean follow-up of 6.8 years, 2208 (3.8%) patients were diagnosed with ESRD. Estimated event rates were 0.4% at 1 year and 2.7% at 5 years. The corresponding rates for the high-risk group were 2.8% and 15.1%. The predictors of ESRD included in the final model were: blood urea nitrogen, CKD, albumin, pulse pressure, hemoglobin A1c, anemia, glomerular filtration rate, and preexisting diabetes (vs. newly diagnosed) (c-statistic: 0.706). The proposed risk prediction model for ESRD demonstrates strong predictive capabilities. Applied in clinical settings, the model could help clinicians identify T2DM patients most likely to benefit from SGLT2 inhibitors to prevent or delay progression to ESRD.
D. Geba: None. J.M. Cordova: Employee; Self; Boehringer Ingelheim Pharmaceuticals, Inc. S. Shetty: None. B. Williams: None.