Fig. 3

Residual eGFR variance after adjusting for age, sex, comorbidities and PGS in a stepwise fashion in elderly and general adults. We applied linear regression models on eGFRcrea and eGFRcys increasing the number of covariables in a stepwise fashion: adjusting for age, sex, BMI, DM, hypertension, CAD, high-ceiling diuretics intake, and PGS. We show the residual variance of eGFRcrea (A) or eGFRcys (B) in AugUR (n = 2,272) and KORA S4 (n = 2,900). The colored area is thus reflecting the proportion of the variance explained by the respective covariable adjusted for the covariables in the list above