Glomerular filtration rate estimation and mortality in an elderly population.

BACKGROUND Few studies have addressed the link between minor renal dysfunction and mortality in the elderly. AIM To compare three equations for estimated GFR (eGFR) in assessing renal dysfunction and predicting mortality in an elderly population. DESIGN Longitudinal observational study. METHODS We studied 441 people from the Jerusalem Seventy Year Olds Longitudinal Study who had measurements of serum creatinine, all of whom were aged 70 years at study initiation and were living in the community. GFR was estimated based on serum creatinine and using the Cockcroft-Gault (CG), the abbreviated Modification of Diet in Renal Disease (MDRD) and the Mayo Clinic equations. Twelve-year mortality was the main outcome measure. RESULTS The prevalence of reduced eGFR was 51% using the CG, 34% using MDRD and 16% using the Mayo Clinic equation. eGFR dichotomized by the definition of CKD significantly predicted mortality only with the Mayo Clinic equation (hazard ratio 1.56, 95%CI 1.01-2.39). When eGFR was divided into quartiles and the lowest compared to the highest, all equations predicted mortality. Hazard ratios (95%CI) were 5.48 (1.27-23.65), 7.47 (2.74-20.3), and 7.375 (3.13-17.36), for CG, MDRD, and Mayo Clinic, respectively. DISCUSSION Reduced eGFR was prevalent in this study group, and associated with mortality. This association was strongest using the Mayo Clinic equation.