Study Validates 2 Formulas Estimating Kidney Function for Adults in Brazil

Published on: 

These data helped to verify the MDRD-4 and CKD-EPI equations’ accuracy in the estimation of GFR among the Brazilian population.

The formulas known as the Modification of Diet in Renal Disease (MDRD-4) and the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) are accurate in the screening of chronic kidney disease (CKD), according to new findings, and adjusting for race/color may diminish the reliability of the equations.1

These findings were drawn from a new study designed to look into the accuracy and agreement between creatinine clearance (CrCl) that is measured in 12-h urine and glomerular filtration rate (GFR) which is calculated through the MDRD-4 and CKD-EPI formulas. The investigators looked at their accuracy with and then without an adjustment for race/color of the subjects.

Background and Findings

While the investigators acknowledged the widespread implementation of MDRD-4 and CKD-EPI in the population, they noted that there were no rigorous studies with a large enough sample.2 The team also added that the equations had used a race/color correction multiplier that they pointed out had been criticized.

To address this, the research was led by Wagner Luis da Cruz Almeida, from the Universidade Federal da Bahia’s School of Pharmacy in Brazil.

“The objective of this work was to verify the accuracy of MDRD-4 and CKD-EPI equations in estimating GFR in a robust sample of the Brazilian population through agreement with the CrCl calculated from 12-hour urine,” Almeida and colleagues wrote. “Furthermore, the adequacy of the adjustment by race/color in these equations was verified, to assess their applicability in the Brazilian population.”

The investigators utilized baseline data drawn from the Brazilian Longitudinal Study of Adult Health (ELSA-Brazil). ELSA-Brazil was a multicenter cohort that had looked at 15,105 public servants in the age range of 35 - 74.

They looked at anthropometric, sociodemographic, clinical, and laboratory data which had been gathered at 5 universities and a research institution. The team looked to have demographic diversity in the research, with 76% voluntary participation.

At the point of ELSA baseline, the subjects went through a 12-hour overnight period of urinary collection. The research team’s analysis excluded those that were shown to have invalid or untimely collections.
Sociodemographic data, some of which included lifestyle factors, self-declared race/color, and noncommunicable history of diseases, were identified through trained interviewers. Weight, height, and blood pressure were also looked at by the investigators.

The research team gathered venous blood following a period of fasting, and these samples were assessed at the ELSA Central Laboratory. Validation criteria for the urinary samples included diary reports, volume, collection time, and adjusted creatinine excretion.

The team measured various elements in the collected urine, including sodium, creatinine, potassium, and albumin. Subjects’ CrCl was found through a calculation of their urinary flow, concentrations of creatinine, and body surface adjustment.

The investigators focused on assessing the accuracy and the agreement between CrCl that was measured in 12-hour urine and GFR, which was calculated through the use of MDRD-4 and CKD-EPI formulas. They did this with and without race/color adjustment, involving data from 12,813 subjects that had validated urine collections.

Overall, the research team’s results showed that MDRD-4 and CKD-EPI may be considered reliable for the screening of CKD. They also reported that the adjustment made for race/color was shown to be unnecessary, and this was particularly the case for GFR <90 mL/min x 1.73m².

Among those in the group of 15,105 individuals featured in ELSA-Brazil, the investigators noted that 12,813 experienced a validated urine collection. The team also showed that Bland-Altman diagrams indicated better agreement between formulas and creatinine clearance (CrCl) for GFR <90 mL/min x 1.73m².

They added that adjusting for race/color had increased the dispersion of data dispersion, adding that 1-way ANOVA showed similarity between the different CrCl groups with race/color in this range (P=0.27).

“The ELSA-Brazil sample population is quite diverse, including both healthy adults and those with morbidities frequently found in the general population, such as: obesity, diabetes, hypertension, dyslipidemia, and CKD, among others, ensuring a sample with good similarity with the Brazilian population of the same age group,” they wrote. “Above all, the high methodological rigor aimed at achieving maximum quality of the data obtained stands out.”


  1. Almeida WLDC, Barreto SM, Vidigal PG, Mill JG. Validation of equations to estimate kidney function with and without adjustment by race/color in Brazilian adults (ELSA-Brazil). Rev Bras Epidemiol. 2023;26:e230057. Published 2023 Dec 11. doi:10.1590/1980-549720230057.
  2. Bastos MG, Kirsztajn GM. Doença renal crônica: importância do diagnóstico precoce, encaminhamento imediato e abordagem interdisciplinar estruturada para melhora do desfecho em pacientes ainda não submetidos à diálise. Braz J Nephrol 2011; 33(1): 93-108.