Validation of equations to estimate kidney function with and without adjustment by race/color in Brazilian adults (ELSA-Brazil)

Abstract

Objective: To evaluate accuracy and agreement between creatinine clearance (CrCl) measured in 12-h urine and glomerular filtration rate (GFR) calculated by the Modification of Diet in Renal Disease (MDRD-4) and the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formulas, with and without adjustment for race/color. Methods: Baseline data from the Longitudinal Study of Adult Health (ELSA-Brazil) in adults (35-74 years of age) of both genders were used. Serum creatinine was measured in fasting blood and urinary creatinine was measured in an overnight 12-h urine collect. The agreement between CrCl and the calculated GFR was analyzed by the Bland-Altman method. One-way analysis of variance (ANOVA) with race/color factor was used to verify differences between means of CrCl and GFR with and without correction for race/color. Statistical significance was accepted for p<0.05. Results: From 15,105 participants in the ELSA-Brazil, 12,813 had a validated urine collect. The Bland-Altman diagrams showed that formulas and CrCl agree with each other with a better accuracy for GFR <90 mL/.min x 1.73m2. The adjustment by race/color increased data dispersion. In this range, one-way ANOVA of CrCl with race/color factor showed similarity between groups (p=0.27). Conclusion: MDRD-4 and CKD-EPI are useful formulas for screening cases of chronic kidney disease, and correction by race/color, only in blacks or in black and brown subjects, proved to be unnecessary and reduced the reliability of the equations.