Development and validation of prediction models of adverse kidney outcomes in the population with and without diabetes mellitus

Objective: To predict adverse kidney outcomes for use in optimizing medical management and clinical trial design.

Research Design and Methods: In this individual participant data meta-analysis, 43 cohorts (N=1,621,817) from research studies, electronic medical records, and clinical trials with global representation were separated into development and validation cohorts. Models were developed and validated within strata of diabetes mellitus (presence or absence) and eGFR (≥60 or <60 ml/min/1.73 m2) to predict a composite of ≥40% decline in eGFR or kidney failure (receipt of kidney replacement therapy) over 2-3 years.

Results: There were 17,399 and 24,591 events in development and validation cohorts, respectively. Models predicting ≥40% eGFR decline or kidney failure incorporated age, sex, eGFR, albuminuria, systolic blood pressure, anti-hypertensive medication use, history of heart failure, coronary heart disease, atrial fibrillation, smoking status, and body-mass index (and hemoglobin A1c, insulin use, and oral diabetes medication use in those with diabetes). The median C-statistic was 0.774 (interquartile range [IQR]: 0.753, 0.782) in the diabetes/higher eGFR validation cohorts, 0.769 (IQR: 0.758, 0.808) in the diabetes/lower eGFR validation cohorts, 0.740 (interquartile range [IQR]: 0.717, 0.763) in the no diabetes/higher eGFR validation cohorts, and 0.750 (IQR: 0.731, 0.785) in the no diabetes/lower eGFR validation cohorts. Incorporating previous 2-year eGFR slope minimally improved model performance, and only in the higher eGFR cohorts. 

Conclusions: Novel prediction equations for an eGFR decline of ≥40% eGFR can be applied successfully for use in the general population in persons with and without diabetes with higher or lower eGFR.