The estimation of glomerular filtration rate in an Australian and New Zealand cohort

Abstract

Accurate estimation of glomerular filtration rate (GFR) allows early detection of renal disease and maximizes opportunity for intervention. To assess the accuracy of estimated GFR (eGFR) in an Australian and New Zealand cohort with chronic kidney disease using the 4-variable Modification of Diet in Renal Disease equation (MDRD(4V)), the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations, and the Cockcroft and Gault equation with actual and ideal body weight. Retrospective review of patients who had measured GFR (mGFR) by 51Cr-EDTA clearance and simultaneous measurements of serum biochemistry and anthropometrics. eGFR was compared with mGFR using the concordance correlation coefficient (CCC) and Bland-Altman measures of agreement. 178 patients had 441 radioisotope measurements of GFR. Mean mGFR of was 22.6 mL/min per 1.73 m(2) . The MDRD(4V) equation using the 'black' correction factor was most accurate with a mean eGFR of 19.74 (CCC 0.733, bias -2.86). The CKD-EPI equations also using the 'black' correction factors were almost as good at 19.11 (CCC 0.719, bias -3.49). The Cockcroft-Gault creatinine clearance values had the poorest agreement with mGFR. In the 18 nonwhite non-Asian patients, the MDRD(4V) and CKD-EPI equations were generally less accurate although the use of the 'black' correction factor resulted in greater accuracy for both equations. The MDRD(4V) equation was the most accurate. However, its accuracy might be less for nonwhite non-Asian patients if the 'black' correction factor is omitted. Further study of the estimation of GFR in Australian and New Zealand ethnic subgroups would be helpful.