25. Royal free hospital cirrhosis GFR is not better than creatinine-based equations in predicting mortality and need for renal replacement therapy

Abstract

Background and Aims: Acute kidney injury (AKI) among cirrhotic patients carries poor prognosis. Various eGFR equations tend to over-estimate GFR in cirrhotics. Royal Free Hospital has developed new model for GFR assessment, which is supposedly close to measured GFR. We aim to evaluate the performance of various GFR equations, to predict the in-hospital mortality and requirement of Renal replacement therapy (RRT). Methods: From April 2017 to March 2018, consecutive cirrhotic patients with AKI admitted in ICU were included. Patients with CKD and ACLF were excluded. e- GFR was calculated using CKD-EPI, MDRD-4, MDRD-6 equations and Royal Free Hospital cirrhosis GFR (RFHC GFR) equation. In-hospital mortality and requirement of RRT was evaluated using ROC curve. Results: 120 patients [88.2% male, mean age 48.5 years ± 10.7 (SD)] with mean CTP score 10.84 ± 1.84 and mean MELD 27.8 ± 7.4 were included. There were 80.3%, 17.1% and 2.6% of patients in AKIN Stage I, Stage II and Stage III respectively. In-hospital mortality was 36.6%. 56.8% among non-survival group and 1.3% among survival group underwent RRT. Mean eGFR by RFHC GFR (18.36 ± 7.18 ml/min/1.73m2) was significantly lower than CKD-EPI (42.68 ± 18.88), MDRD4 (39.97 ± 16 .17) and MDRD6 (31.36 ± 13.64). AUROC failed to show significant differences among various equation in predicting the in-hospital mortality (CKD-EPI: 60.0%, MDRD4: 60.4%, MDRD6: 61.3%, RFHC GFR: 59.6%; p > 0.05) and requirement of RRT (CKD-EPI: 66.3%, MDRD4: 67.2%, MDRD6: 67.8%, RFHC GFR: 65.5%). Conclusions: Though the Royal free hospital cirrhosis GFR equation might be lowest and closest to the actual GFR, none of the equations failed to show superiority over other in predicting the in-hospital mortality and requirement of RRT among critically ill cirrhotic patients with AKI. The authors have none to declare.