An evaluation of the body surface area correction for 51Cr-EDTA measurements of glomerular filtration rate

Abstract

To interpret 51Cr-EDTA measurements of glomerular filtration rate (GFR) it is usual to correct results for a patient's body size by scaling values to a standard body surface area (BSA) of 1.73 m(2). To use 51Cr-EDTA data for a large group of healthy subjects to derive the optimum mathematical function for the body size correction that minimized the variance of the corrected GFR values. This function was then compared with the widely used Du Bois and Haycock BSA formulas to determine which of these two equations provided the better correction. GFR data for 428 healthy adults (218 female, 210 male) undergoing assessment as live kidney donors were evaluated. The body size correction was assumed to scale as the product of power laws of body mass index (BMI = weight/height(2)) and height. The corrected GFR figures were fitted to a model in which GFR was constant in subjects below 40 years of age and decreased linearly with age in older subjects. The scatter about the best fitting model was expressed as a percentage of the mean GFR in the younger group and the optimum power law indices derived from a least-squares fit were compared with the values for the Du Bois and Haycock formulas. The least-squares fit gave values of 0.444 (95% CI, 0.297-0.590) for the BMI power law index and 1.416 (95% CI, 1.027-1.802) for the height index with a 95% confidence error figure that included the points representing both the-Du Bois ([delta][chi]2 = 1.161, P = 0.560) and Haycock ([delta][chi]2 = 2.524, P = 0.283) formulas. For subjects of average height and a BMI in the middle of the normal range the Du Bois and Haycock equations agreed closely. Differences in BMI were found to explain 89% of the variance in the BSA estimates between the two formulas. Compared with the Du Bois formula the Haycock equation gave a 5% increase in corrected GFR in subjects with a BMI of 15 kg x m(-2) and a 10% decrease in subjects with a BMI of 40 kg x m(-2). Within the statistical errors both the Du Bois and Haycock BSA formulas were consistent with the optimum power law function that minimized the variance of the corrected GFR values for a group of 428 healthy adults.