Regressing an optimal variable to scale glomerular filtration rate: more variations in glomerular filtration rate are explained.

Background: To compare glomerular filtration rate (GFR) among individuals, GFR is usually scaled to body surface area (BSA) based on the ratio method, which has been debated for its accuracy in recent years. Reference to the BSA as a normalization standard is the most common method currently in use but has limitations. This study was designed to a better variable to scale GFR.

Methods: We measured 99mTc- diethylene triamine pentaacetic acid plasma clearance (uncorrected GFR, uGFR) for 322 healthy adults who were enrolled according to the SENIEUR protocol. The individuals were randomly grouped into A and B for regressing and validating the optimal variable, respectively. Nonlinear regression was performed against uGFR, and the selected independent variables were body weight, height, age, and sex.

Results: Among several tested models, the regression coefficients of weight-age formula (W-A) were in narrower 95% confidence interval (CI). The coefficient of determination of the regression line between W-A and uGFR, as an indicator to explain the percentage of variations of GFR, was higher than that of other variables in both groups. The coefficient of determination of the regression line between W-A and uGFR was 0.571, which was higher than that of BSA (0.203) or TBW (0.241).

Conclusion: The index variable, based on both body weight and age, has a better statistical relationship to uGFR and is a better variable to scale GFR in adults.