Correction of the slope-intercept method for the measurement of glomerular filtration rate

Glomerular filtration rate (GFR) is frequently assessed using the slope-intercept method by fitting a single exponential to plasma samples obtained 2-5 h after injection. The body surface area (BSA)-corrected one-pool clearance (CO,BSA) overestimates true GFR (CT,BSA) because it fails to sample the full plasma curve, and values of CT,BSA are usually estimated from CO,BSA using the Brøchner-Mortensen (BM) equation. An improved equation, CT,BSA=CO,BSA/(1+fBSA×CO,BSA), with fBSA a fixed constant, was proposed by Fleming, but subsequently Jødal and Brøchner-Mortensen (JBM) reported that fBSA varies with BSA. We report data for a large group of individuals who underwent GFR investigations with sampling of the full plasma curve. The aims were to validate the JBM equation with independent data and assess whether replacing the BM equation with a BSA-dependent correction based on Fleming's equation can increase the accuracy of the slope-intercept method.

Plasma data were analysed for 142 children and adults aged 0.6-56 years who underwent technetium-99m-diethylenetriaminepentaacetic acid GFR investigations with blood samples taken between 5 min and 8 h after injection. Values of CO,BSA were calculated using the 2, 3 and 4 h data. Values of CT,BSA were calculated by integrating the plasma curve between 5 min and 4 h and extrapolating the terminal exponential. Individual values of fBSA were calculated using the relationship fBSA=1/CT,BSA-1/CO,BSA. Nonlinear regression was used to fit the function fBSA=f1×BSA and find the best-fit values for f1 and n. Scatter and Bland-Altman plots were drawn comparing the various formulae for correcting slope-intercept GFR.

The trend for fBSA to decrease with increasing BSA was highly significant (Spearman's test: RS=-0.31; P=0.0002). When the data were fitted by nonlinear regression, the best-fit values (95% confidence interval) of the model parameters were n=-0.13 (from -0.21 to -0.04) and f1=0.00191 (from 0.00183 to 0.00200).

The results confirm that fBSA varies with BSA and provide independent values of the parameters f1 and n. Differences from GFRs calculated using the original JBM equation were small and not clinically significant. The BM equation also performed well for CT,BSA less than 125 ml/min/1.73 m. However, there was a small number of children with CT,BSA greater than 150 ml/min/1.73 m for whom the JBM formula provided more accurate estimates of true GFR than did the BM equation.