MDRD or CKD-EPI study equations for estimating prevalence of stage 3 CKD in epidemiological studies: which difference? Is this difference relevant?

Background: Prevalence of stage 3 chronic kidney disease (CKD) is increasing according to the NHANES study. Prevalence has been calculated using the MDRD study equation for estimating glomerular filtration rate (GFR). Recently, a new estimator based on creatinine, the CKD-EPI equation, has been proposed which is presumed to better perform in normal GFR ranges. The aim of the study was to measure the difference in prevalence of stage 3 CKD in a population using either the MDRD or the CKD-EPI study equations.
Methods: CKDscreening is organized in the Province of Liège, Belgium. On a voluntary basis, people aged between 45 and 75 years are invited to be screened. GFR is estimated by the MDRD study equation and by the "new" CKD-EPI equations.
Results: The population screened consisted in 1992 people (47% of men). Mean serum creatinine was 0.86 ± 0.20 mg/dl. The prevalence of stage 3 CKD in this population using the MDRD or the CKD-EPI equations was 11.04 and 7.98%, respectively. The prevalence of stage 3 CKD is significantly higher with the MDRD study equation (p <0,0012).
Conclusions: Prevalence of stage 3 CKDvaries strongly following the method used for estimating GFR, MDRD or CKDEPI study equations. Such discrepancies are of importance and must be confirmed and explained by additional studies using GFR measured with a reference method.