Performance of MDRD study and CKD-EPI equations for long-term follow-up of nondiabetic patients with chronic kidney disease

Chronic kidney disease (CKD) typically extends over decades. Longitudinal monitoring of kidney function in CKD is thus of great importance. Here, we retrospectively evaluate use of the Modification of Diet in Renal Disease (MDRD) study and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations to monitor long-term course of kidney function and to identify individuals with progressive kidney function loss. Patients were selected from our outpatient clinic for having four glomerular filtration rate measurements (mGFR, (125)I-iothalamate) and at least ≥ 4 years of follow-up. Renal function slopes were obtained by within-individual linear regression. Sixty-five nondiabetic CKD patients (40 male, mean baseline age 44 ± 12 years) with a median (range) of 9 (4-16) mGFR measurements and a median follow-up of 11 (4-33) years were included. Both equations significantly underestimated mGFR/(BSA) at baseline and at the end of follow-up. mGFR slope was significantly underestimated by the MDRD study but not by CKD-EPI equation (slopes -1.41 ± 2.06, -1.07 ± 1.72 and -1.39 ± 1.77 mL/min/1.73 m(2)/year, respectively). Sensitivity and specificity to identify progressive kidney function loss (mGFR/(BSA) slope > 1.5 mL/min/1.73 m(2)/ year, n = 23) were 78 and 88% for the MDRD study and 91 and 80% for CKD-EPI equation. In the subgroup of progressors, both MDRD study and CKD-EPI equation underestimated the rate of mGFR loss (P < 0.05) Long-term course of mGFR is reasonably well estimated by CKD-EPI and slightly underestimated by MDRD study equation. Patients with progressive kidney function loss may, however, not be reliably identified, so caution is warranted when using these equations in clinical practice.