Creatinine fluctuation has a greater effect than the formula to estimate glomerular filtration rate on the prevalence of chronic kidney disease

Abstract

Background/aims: Cases of chronic kidney disease (CKD) are defined by the estimated glomerular filtration rate (eGFR), calculated using the Modified Diet in Renal Disease (MDRD) or, more recently, the CKD Epidemiology Collaboration (CKD-EPI) formula. This study set out to promote a systematic approach to reporting CKD prevalence.

Design, setting, participants and measurements: The study explores the impact of the way in which eGFR is calculated on the prevalence of CKD. We took into account whether including (1) ethnicity, (2) using a single eGFR, (3) using more than 1 eGFR value or (4) using the CKD-EPI formula affected the estimates of prevalence.

Sample: Of 930,997 registered patients, 36% (332,891) have their eGFR defined (63% of those aged 50-74 years, 81% >75 years).

Results: The prevalence of stage 3-5 CKD is 5.41% (n = 50,331). (1) Not including ethnicity data the prevalence would be 5.49%, (2) just using the latest eGFR 6.4%, (3) excluding intermediary values 5.55% and (4) using the CKD-EPI equation 4.8%. All changes in eGFR (t test) and the proportion with CKD (χ(2) test) were significant (p < 0.001). Using serum-creatinine-calculated eGFR instead of laboratory data reduced the prevalence of stage 3-5 CKD by around 0.01%. Sixty-six percent of people with stage 3-5 disease have cardiovascular disease and 4.0% significant proteinuria using the MDRD formula; the corresponding figures using CKD-EPI are 74 and 4.6%.

Conclusions: A standardised approach to reporting case finding would allow a better comparison of prevalence estimates. Using a single eGFR tends to inflate the reported prevalence of CKD by ignoring creatinine fluctuation; this effect is greater than the difference between MDRD and CKD-EPI.