Monitoring kidney function in diabetic nephropathy

Abstract

Progression in diabetic nephropathy is usually determined by repeated measurements of glomerular filtration rate and expressed as rate of decline in glomerular filtration rate. Our aim was to evaluate the agreement between rate of decline in glomerular filtration rate estimated from the Cockroft-Gault formula: (140-age)*K*body weight* (1/S-creatinine) and measured by the plasma clearance of 51CrEDTA. All insulin-dependent diabetic patients with diabetic nephropathy followed-up for at least 5 years with at least 5 simultaneous measurements of glomerular filtration rate, s-creatinine, and weight were included in the study. Forty-three patients (32 male/11 female), age 31 (18-61) years were enrolled. Observation period: 6.6 (5.1-9.9) years and number of investigations per patient 6 (5-16) (median(range)). Baseline glomerular filtration rate (ml/min) was 97 (30) measured and 107 (37) estimated (mean(SD))(p<0.001) and the 95% limits of agreement were -42.0 to 20.8 ml/min. Measured and estimated glomerular filtration rate correlated significantly (r = 0.91, p<0.00001). Rate of decline in kidney function ml · min-1 · year-1 was 4.7 (3.3) measured and 4.8 (3.5) estimated (mean(SD)) (NS), but the 95% limits of agreement showed a wide range -3.9 to 3.5 ml · min-1 · year-1. A significant correlation between rate of decline in measured and estimated glomerular filtration rate was present (r = 0.84, p<0.00001). In conclusion, glomerular filtration rate is overestimated by the Cockroft-Gault formula. The mean rates of decline in glomerular filtration rate are comparable, but the limits of agreement are wide, which make the Cockroft-Gault method unacceptable for clinical purposes, i.e. monitoring progression in kidney function in the individual patient. However, the estimated glomerular filtration rate may be used for comparison of groups in observational studies and in clinical trials with a long observation period. © 1994 Springer-Verlag.