DESCRIPTION (provided by applicant): In hemodialysis (HD) patients with diabetes mellitus (DM), the risk of developing diabetes- related complications and mortality is high. However, data on the associations among diabetes control, treatment, and diabetes-related complications in HD patients with DM are extremely limited. In addition, racial and ethnic differences in these relationships have not been documented. Objectives: The objective of this observational study is to assess the relationship among glycemic control, oral antidiabetic drug (OAD) therapy, and both cardiovascular and non-cardiovascular outcomes in HD patients with DM. We will also explore racial and ethnic differences in these associations. Methods: We will use data from the U.S. Renal Data System and electronic medical records of DaVita, Inc., to conduct a retrospective cohort study of patients on HD with diabetes mellitus at a large national provider of dialysis services to examine the associations among HbA1c and various clinical outcomes (Specific Aims 1 and 2). A subset of these patients will be used for a comparative effectiveness study of patients receiving TZD or sulfonylurea monotherapy (Specific Aim 3). For the first two aims, we will use Cox proportional hazards models to assess the associations between HbA1c and time to primary outcomes of interest, and test for effect modification by race and ethnicity by including interaction terms and performing a Wald test. We will further explore contrasts between different racial and ethnic groups upon finding significant interaction terms. For Specific Aim 3, we will compare prevalent and incident users of TZDs to users of sulfonylureas by limiting all analyses to patients receiving monotherapy with one of these two drug classes. We will examine the association between OAD monotherapy (TZD vs. sulfonylurea) and HbA1c by performing a repeated-measures analysis using linear regression models with a robust sandwich estimator. We will also analyze the association between OAD monotherapy (TZD vs. sulfonylurea) and time to cardiovascular events and death using Cox proportional hazards models. We will further conduct as-treated and intention-to-treat analyses in addition to other sensitivity analyses, and employ advanced statistical methods including propensity score techniques. Racial and ethnic differences in these associations will be investigated using the method described above for the first two aims. Significance: This research study will take advantage of an unusually large and detailed dataset of patients initiating HD. Due to the limited data on the associations among diabetes management and treatment, including glycemic control and OAD therapy, and diabetes complications in HD patients with DM, additional studies are warranted in this high-risk population. Moreover, our study will have the ability to investigat racial and ethnic differences in these associations as we capitalize on the enormous size of the database that will be available for analysis.

PUBLIC HEALTH RELEVANCE: Diabetes mellitus and end stage renal disease synergistically contribute to the incidence of cardiovascular and non-cardiovascular outcomes and mortality. In addition, diabetic hemodialysis patients of particular race or ethnicity, especially blacks and Hispanics, may carry a heavier disease burden than patients of other racial/ethnic groups. Results of this proposed research may hold important implications for race/ethnicity- specific diabetes control and treatment practices that have the potential to reduce diabetes-related morbidity and mortality in hemodialysis patients with diabetes.