Assessment of glomerular filtration rate in diabetic nephropathy using the plasma clearance of 51Cr-EDTA.

Plasma clearance of 51Cr-EDTA is widely used to assess the glomerular filtration rate (GFR) in diabetic nephropathy. Originally, the ratio between the intravenously injected amount of tracer and the total area under the plasma concentration curve was used for the calculation of total 51Cr-EDTA plasma clearance (C(T)). Simplified methods, using the final mono-exponential part of the plasma curve, have been suggested, e.g. four samples, taken 180 to 240 min after injection (C(IV)), or using one sample taken at 240 min (C(I)). Our aim was to evaluate the agreement between measurements of GFR and rate of decline in GFR based upon these three methods. Bland & Altman plots were used to illustrate the range of agreement. We investigated 76 insulin-dependent diabetic (IDDM) patients with microalbuminuria or diabetic nephropathy. GFR was measured after a single intravenous injection of 3.7 MBq 51Cr-EDTA by determining the radioactivity in venous blood samples taken 5, 7, 10, 15, 30, 45, 60, 90, 120, 150, 180, 200, 220, and 240 min after the injection. Rate of decline in GFR was assessed using 12 (6-17) determinations of GFR over a period of time of 8 (4-10) years. Mean (SD) GFRT was 123 (21) ml x min(-1) compared to GFR(IV) 123 (21) ml x min(-1) (NS) and GFR(I) 115 (17) ml x min(-1) (p < 0.00001). The mean difference (95% limits of agreement) between GFR(T) and GFR(IV) was +0.6 (-16.6 to +17.7) ml x min(-1), and between GFR(T) and GFR(I) +8.0 (-6.0 to +22.2) ml x min(-1). The difference between GFR(T) and GFR(I) was significantly correlated with their mean value (r = 0.56, p < 0.00001), indicating increasing underestimation by GFR(I) with increasing GFR levels. The mean (SD) rate of decline in GFR(T) was 2.3 (3.9) ml x min(-1) x year(-1), compared to a mean rate of decline in GFR(IV) of 2.4 (3.6) ml x min(-1) x year(-1) (NS), and a mean rate of decline in GFR(I) of 2.2 (3.5) ml x min(-1) x year(-1) (NS). The mean difference (95% limits of agreement) between rate of decline in GFR(T) and rate of decline in GFR(IV) was +0.16 (-1.59 to +1.91) ml x min(-1) x year(-1), and between rate of decline in GFR(T) and rate of decline in GFR(I) -0.01 (-1.64 to +1.61) ml x min(-1) x year(-1), respectively. In conclusion, our cross-sectional study revealed a close agreement between GFR(T) and GFR(IV) with acceptable limits of agreement (precision), while GFR(I) lacked accuracy. However, a close agreement between rate of decline in GFR(T) and rate of decline in GFR(IV), and between rate of decline in GFR(T) and rate of decline in GFR(I), with acceptable limits of agreement (precision), suggests that both simplified methods are applicable for long-term follow-up.