Fallacy of indexing renal and systemic hemodynamic measurements for body surface area

Renal plasma flow, glomerular filtration rate, and cardiac output are traditionally indexed for body surface area by expressing these traits as per-surface-area ratios. Indexing is intended to remove interindividual variation attributable to differences in body size. Regression is an alternative method commonly used to adjust other biological traits for the effects of a covariate, such as body surface area. The purpose of this study was to compare the indexing and regression methods of adjusting renal plasma flow, glomerular filtration rate, and cardiac output for interindividual differences in body surface area. We estimated renal plasma flow by the clearance of p-aminohippurate, glomerular filtration rate by clearance of inulin, and cardiac output by thoracic electrical impedance in a sample of 78 unrelated females and 78 unrelated males (ages 20-49.9 yr) from the general population of Rochester, MN. The indexing method created negative dependencies of renal plasma flow and cardiac output on body surface area and failed to eliminate the positive dependency of glomerular filtration rate on body surface area. Moreover, indexing obscured differences in mean renal plasma flow between females and males and created differences in mean cardiac output between the genders. In contrast, the regression method consistently eliminated dependencies of each trait on body surface area and did not lead to inappropriate inferences about mean differences in these traits between females and males. We conclude that the indexing method of adjusting renal plasma flow, glomerular filtration rate, and cardiac output for interindividual differences in body surface area should be abandoned and replaced by use of the regression method.