The Chronic Kidney Disease Epidemiology Collaboration equation outperforms the Modification of Diet in Renal Disease equation for estimating glomerular filtration rate in chronic systolic heart failure

Abstract

The Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula estimates glomerular filtration rate (GFR) better than the simplified Modification of Diet in Renal Disease (sMDRD) formula in numerous populations. It has not previously been validated in heart failure patients. The GFR was measured in 120 patients with chronic systolic heart failure (CHF) using [(125)I]iothalamate clearance (GFR(IOTH)) and estimated using the sMDRD and CKD-EPI equations. Accuracy, bias, and prognostic performance were compared. Cockcroft-Gault, CKD-EPI serum cystatin C, and CKD-EPI creatinine-serum cystatin C equations were compared in secondary analyses. Mean age was 59 ± 12 years, 80% were male. Mean LVEF was 29 ± 10%. Mean GFR(IOTH) was 74 ± 27 mL/min/1.73 m(2), and mean estimated GFR was 66 ± 23 mL/min/1.73 m(2) (CKD-EPI) and 63 ± 21 mL/min/1.73m(2) (sMDRD). CKD-EPI showed less bias than sMDRD (-8 ± 15 vs. -11 ± 16 mL/min/1.73 m(2), P < 0.001). Both equations underestimate at higher and overestimate at lower GFR(IOTH). Eleven patients (9%) were accurately reclassified into lower CKD classes with CKD-EPI. Cockcroft-Gault showed lower bias (-3 ± 16 mL/min/1.73 m(2)) but worse precision and accuracy. Cystatin C-based estimation showed the lowest bias (-3 ± 14 mL/min/1.73 m(2)) and the best precision and accuracy. Prognostic value did not differ between all GFR estimates The CKD-EPI equation more accurately estimates measured GFR than the sMDRD equation in CHF patients, with less bias and greater accuracy and precision. The prognostic power of all GFR assessments was equivalent. Based on better performance and equal risk prediction, we believe the CKI-EPI equation should be the preferred creatinine-based GFR estimation method in heart failure patients, particularly those with preserved or moderately impaired renal function.