Modification of the CKD Epidemiology Collaboration (CKD-EPI) Equation for Japanese: Accuracy and Use for Population Estimates

Introduction We previously reported a modification to the Modification of Diet in Renal Disease (MDRD) Study equation for use in Japan. Recently, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) developed a new equation that is more accurate and yields a lower CKD prevalence estimate in the United States than the MDRD Study equation. We modified the CKD-EPI equation for use in Japan, compared its accuracy with the Japanese modification of the MDRD Study equation, and compared the prevalence of CKD in Japan using both equations. Design A diagnostic test study comparing the Japanese coefficient–modified CKD-EPI equation and Japanese coefficient–modified MDRD Study equation and a cross-sectional study comparing distribution of estimated glomerular filtration rate and prevalence of CKD in participants in a Japanese annual health check program. Setting & Participants 763 Japanese patients (413 for development and 350 for validation) were included. Prevalence estimates were based on 574,024 participants from the annual health check program. Index Test Japanese modification of the MDRD Study and CKD-EPI equations. Reference Test Inulin clearance. Results The Japanese coefficient of the modified CKD-EPI equation was 0.813 (95% CI, 0.794-0.833). In the validation data set, the modified CKD-EPI equation performed better than the modified MDRD Study equation. Bias (measured GFR [mGFR] − eGFR) was 0.4 ± 17.8 (SD) versus 1.3 ± 19.8 mL/min/1.73 m2 overall, respectively (P = 0.02); 7.3 ± 20.6 versus 7.8 ± 22.2 mL/min/1.73 m2 for participants with mGFR ≥60 mL/min/1.73 m2, respectively (P < 0.001); and −4.4 ± 13.8 versus −3.3 ± 15.6 mL/min/1.73 m2 for participants with mGFR <60 mL/min/1.73 m2, respectively (P = 0.5). The modified CKD-EPI equation yields a lower estimated prevalence of CKD than the modified MDRD Study equation (7.9% vs 10.0%), primarily because of a lower estimated prevalence of stage 3 (5.2% vs 7.5%). Limitation Most study participants had CKD. The study population contained a limited number of participants with mGFR ≥90 mL/min/1.73 m2. Conclusion The Japanese coefficient–modified CKD-EPI equation is more accurate than the Japanese coefficient–modified MDRD Study equation and leads to a lower estimated prevalence of CKD in Japan.