Chronic kidney disease prevalence and secular trends in a UK population: the impact of MDRD and CKD-EPI formulae

Abstract

Introduction: Most UK laboratories use the MDRD4 formula to estimate glomerular filtration rate (eGFR), but this may exaggerate chronic kidney disease (CKD) prevalence. In a large adult population, we examined the impact of the more accurate CKD-EPI formulae on prevalence estimates, and on secular trends in prevalence.

Methods: We extracted all serum creatinine (SCr) results for adults, processed in our laboratory during two 1-year periods (2004, 2009-10). To minimize the effect of acute illness, a patient's lowest SCr was used for each period. eGFR (traceable to isotope dilution mass spectrometry value) was calculated using the MDRD4 and CKD-EPI formulae. Prevalence estimates were compared, with sub-group analysis by age and sex.

Results: In 2004, 102 322 patients had SCr tested (35.4% of the adult population), rising to 123 121 (42.3%) in 2009-10. The proportion tested rose with age to 86% of 85- to 89-year olds. The prevalence of CKD stages 3-5 was lower with the CKD-EPI formulae than the MDRD4 formula. The CKD-EPI formulae reclassified 17 014 patients (5.8%) to milder stages of CKD, most commonly from eGFR 60-89 ml/min/1.73m(2) and CKD stage 3A, in women, and in those <70 years old. 5172 patients (1.8%), mostly elderly women, were reclassified to more severe stages of CKD. Between the two time periods, the prevalence of CKD stages 3-5 rose from 5.44% to 5.63% of the population using MDRD4, but was static at 4.94% with CKD-EPI.

Conclusion: The CKD-EPI formulae, which are more accurate than the MDRD4 formula at higher GFR, reduced the estimated prevalence of CKD stages 3-5 by 0.5% in 2004 and 0.7% in 2009-10. The greatest reclassification was seen in CKD 3A, particularly amongst middle-aged females. The minor rise in CKD prevalence between 2004 and 2009-10 seen with the MDRD4 formula was not confirmed with the CKD-EPI formulae. The CKD-EPI formulae may reduce overdiagnosis of CKD, but further assessment in the elderly is required before widespread implementation.