Should CKD EPI equation be used for estimation of glomerular filtration rate in obese subjects?

Introduction: Pandemic of obesity is strongly related to increase of chronic kidney disease (CKD) prevalence. Currently recommended CKD EPI equation has several serious limitations particularly in obese subjects who have high body surface area (BSA). The aim of our study was to analyze differences in the prevalence of CKD between CKD EPI and de-indexed equations where individual BSA was used.Methods: In total of 2058 subjects (random sample from general rural population, 29.65% obese) BSA was estimated using Dubois&Dubois and Moesteller equations and included into the de-indexed equations (CKD Di, CKD Mi). CKD was classified according to the KDIGO guidelines and glomerular hyperfiltration (GHF) was defined as 95th percentile according the gender and age decade.Results: In obese subjects, prevalence of CKD was significantly higher with CKD EPI than with CKD Di and CKD Mi equations (9.5%, 6.1%, 5.3%, respectively; p<0.001), while prevalence of GHF was significantly lower (3.8%, 12.3%, 12.8%, respectively; p<0.001). Opposite results were observed in subjects with BMI<25 kg/m2 for CKD (5%, 7.1%, 7.2%; p=0.07) and GHF prevalence (6.1%, 1%, 0.6%; p<0.001). Discussion/Conclusions: The prevalence of CKD is overestimated, and the prevalence of GHF is under-estimated in obese subjects using CKD EPI equation, i.e. CKD EPI equation is unreliable in one third of population. De-indexed equations should be recommended instead of CKD EPI equation in epidemiological studies until directly measurement of GFR become more available.