2705-LB: Insulin Dose and Long-Term Kidney Function in Type 1 Diabetes: A Causal Inference Analysis

Introduction and Objective: High exogenous insulin doses in Type 1 diabetes (T1D) may impact renal health through direct pathways or decreased insulin sensitivity. Assessing this is complicated by the feedback loop where insulin dose and HbA1c mutually influence each other over time. We used the parametric g-formula to estimate longitudinal causal effect of sustained insulin dosing on eGFR within the Pittsburgh Epidemiology of Diabetes Complications (EDC) study. Methods: We analyzed participants in the Pittsburgh EDC cohort, followed since 1986-88. We simulated three approximately 26-year sustained insulin dose strategies (units/kg/day): (1) <0.5, (2) 0.5-0.7, and (3) ≥0.7. The g-formula accounted for time-varying confounding by HbA1c and hypertension and adjusted for baseline BMI, age, and sex. Results: Mean eGFR for insulin dose <0.5 (units/kg/day) was 79.16 (95% CI: 72.24-95.61), for 0.5-0.7 was 77.94 (95% CI: 70.76-130.32), and for ≥0.7 was 76.78 (95% CI: 59.40-179.88). Lower sustained insulin requirements were associated with better long-term preservation of eGFR, independent of glycemic levels. Conclusion: Over 26 years, a lower weight-normalized insulin strategy was associated with better renal function. By resolving the HbA1c-insulin feedback loop in the EDC cohort, this study highlights the potential role of insulin dosing as a long-term determinant of renal health. Disclosure M. Mosslemi: None.