Delhi Model: A New Tool for Predicting Response to Terlipressin in Hepatorenal Syndrome—Acute Kidney Injury

• Published in: Journal of clinical and experimental hepatology 2024-01, Vol.14 (1), p.101266, Article 101266
• Main Authors: Agrawal, Jatin, Kumar, Ashish, Sharma, Praveen, Bansal, Naresh, Singla, Vikas, Sharma, Samarth, Arora, Anil
• Format: Article
• Language: English
• Subjects: acute kidney injury; acute on chronic liver failure; cirrhosis; hepatorenal syndrome; Original; terlipressin
• ISSN: 0973-6883; 2213-3453
• DOI: 10.1016/j.jceh.2023.08.002

International club of ascites (ICA) has introduced revised criteria for hepatorenal syndrome—acute kidney injury (HRS-AKI) with an aim to improve the response rate to treatment. We lack prospective trials to assess its positive impact on the response rate and factors influencing response rate. Thus, we conducted this study with the primary aim of identifying independent factors that predict treatment response to terlipressin. We prospectively included patients of HRS-AKI as per the revised ICA criteria. All were treated with terlipressin and albumin according to the defined protocol and were followed for 90 days, death or liver transplantation. Baseline parameters, as well as delta serum creatinine (sCr) at day 4 (DCD4), were investigated as predictive factors influencing response to terlipressin (primary endpoint). Secondary endpoints were the overall response rate to terlipressin, response in various subgroups of acute-on-chronic liver failure (ACLF) patients, need for readmission, and 90 days survival. The study included 114 patients with a median age of 52 years (83% males). 70 (61%) patients responded to terlipressin. Response rate among ACLF1, ACLF2, and ACLF3 were 62%, 48%, and 35%, respectively. On multivariate analysis, baseline creatinine (odds ratio [OR] 7.889, 95% confidence interval [CI] 3.335, 18.664), Child Turcotte Pugh (CTP) score (OR 1.470, 95% CI 1.026, 2.106), and the DCD4 (OR 0.048, 95% CI 0.015, 0.158) were independently predicting response. We also created a Delhi Model (DM) with an excellent predictive ability for response prediction at day 4 with an AUROC of 0.940 (95% CI 0.897, 0.982). Among responder group, 50% of patients required readmission within three months. The 90-days survival among responder and non-responder groups were 68.5% and 9% (P value