Estimation of glomerular filtration rate in patients with cirrhosis: evaluation of equations currently used in clinical practice and validation of Royal Free Hospital cirrhosis glomerular filtration rate

• Published in: European journal of gastroenterology & hepatology Vol. 34; no. 1; p. 84
• Main Authors: Protopapas, Adonis A, Papagiouvanni, Ioanna, Fragkou, Nikolaos, Alevroudis, Emmanouil, Sinakos, Emmanouil, Goulis, Ioannis
• Format: Journal Article
• Language: English
• Published: England 01.01.2022
• Subjects: Creatinine; Fibrosis; Glomerular Filtration Rate; Hospitals; Humans; Liver Cirrhosis - complications; Liver Cirrhosis - diagnosis; Renal Insufficiency, Chronic - diagnosis; Renal Insufficiency, Chronic - epidemiology; Retrospective Studies
• ISSN: 1473-5687
• DOI: 10.1097/MEG.0000000000001935

Conventional creatinine-based glomerular filtration rate (GFR) equations have been reported to overestimate renal function in patients with cirrhosis. The Royal Free Hospital (RFH) cirrhosis GFR equation was developed to accurately estimate GFR in this population. The aim of this study was to evaluate the ability of widely available equations [Chronic Kidney Disease Epidemiology Collaboration equation (CKD-EPI), Modification of Diet in Renal Disease equations (MDRD-4, MDRD-6)] and the RFH equation to correctly estimate the GFR of patients with cirrhosis. We retrospectively analyzed data from patients with cirrhosis who underwent measurement of GFR with the use of 51Cr-EDTA (GFR-M). The CKD-EPI, MDRD-4, MDRD-6 and RFH equations were calculated, while bias, precision and accuracy were estimated for each one of them and then compared with paired t-tests. Bias was defined as the mean difference between the GFR-M and the result of each equation; precision was defined as the SD of the differences and accuracy was defined as the square root of the mean squared error (mean of the squared differences). Higher values are associated with worse bias and better precision/accuracy. One-hundred and thirty-four cirrhotic patients were included. Bias was estimated for CKD-EPI, MDRD-4, MDRD-6 and RFH at -5.91, -3.13, 0.92 and 18.24, respectively. Significant differences were observed between all equations (P < 0.001). Regarding precision, only the comparison between MDRD-4 (20.81) and RFH (16.6) yielded a statistically significant result (P = 0.037). Finally, CKD-EPI (19.32) and MDRD-6 (18.81) exhibited better accuracy than GFR-RFH (24.61) (P = 0.006 and 0.001). RFH demonstrates inferior accuracy in predicting renal function in patients with cirrhosis, in comparison to conventional equations.