Presented by Devon Nerstad
Masters Candidate in Biostatistics
Plan B Adviser: Dr. Erika Helgeson

An understanding of long-term cardiovascular disease (CVD) risk for living kidney donors (LKDs) is crucial for informing pre-donation screening. This study evaluated prognostic value of Framingham’s 10-year (FHS-10) and 30-year (FHS-30) risk models, the Pooled Cohort Equations (PCE), and the PREVENT 10-year (PREVENT-10) and 30-year (PREVENT-30) models in living kidney donors using risk factors measured pre-donation. Models were evaluated on discrimination, calibration, and average prediction error. 10-year models demonstrated good discriminative ability and yielded accurate prediction of observed CVD rates, with PCE and PREVENT-10 performing slightly better. 30-year models demonstrated similar discrimination, though each systematically underpredicted observed CVD risk. PCE and PREVENT models performed better in female than male LKDs. Recalibrating the FHS-30 significantly improved model calibration. Hazard ratios for CVD risk factors reported by Framingham did not significantly differ from those estimated using LKD data. Our findings suggest that PCE and PREVENT-10 models may be useful tools for assessing 10-year CVD risk of LKD candidates, and recalibrating FHS-30 predictions could provide prognostic value for assessing longer-term risk.