Background: Chronic kidney disease is a widely recognized cardiovascular risk factor. Its detection within large populations depends upon the method used to estimate glomerular filtration. The Cockcroft and MDRD equations are widely used, although their accuracy is limited in certain cases.
Methods: The present study analyzes glomerular filtration values in 674 young, healthy subjects using five methods: Cockcroft, Cockcroft corrected for body surface, MDRD-4 Lund-Malmö and Sawyer formulas. Glomerular filtration values obtained with the first three methods were compared using ANOVA. The Spearman coefficient was calculated to estimate the correlation between MDRD-4 and Cockcroft values, and between Cockcroft values and body mass index.
Results: There was a slight glomerular filtration rate decrease (< 90 ml/min) seen in 394 subjects using the Cockcroft equation, and in 344 subjects using the MDRD-4 formula. The prevalence of chronic kidney disease (glomerular filtration < 60 ml/min) was seen in 3 subjects using the MDRD-4 equation and 161 subjects using the Cockcroft formula. There was significant discordance, by method, between values obtained, with 40% of the population being classified into different stages (> 90 or < 90 ml/min) depending on the formula used. In 8% of the population there was even greater discordance, because they had strictly normal renal function according to MDRD-4 (> 90 ml/min) but fell into chronic kidney disease Stage 3 (< 60 ml/min) according to the Cockcroft formula. There was poor correlation between glomerular filtration rates obtained using the Cockcroft and MDRD-4 equations, suggesting that the subjects with a glomerular filtration rate decrease detected by the two methods were not the same ones. There was correlation between body mass index and glomerular filtration rates obtained with Cockcroft, Cockcroft corrected for body surface and Sawyer formulas and not with MDRD and Lund-Malmö equations.
Conclusions: There are important discrepancies between the methods used to assess renal function in healthy populations. These limitations must be taken into account when deciding on strategies for diagnosis and control of occult chronic kidney disease in the general population.