Faculty of Public Health - Andalas University - OCS, 13th IEA SEA Meeting and ICPH - SDev

Font Size: 
FACTORS ASSOCIATED WITH THE SURVIVAL OF ACUTE LIVER FAILURE PATIENTS USING REGRESSION MODELING
Anil Chankaramangalam Mathew, Leelakrishnan Venkatakrishnan

Last modified: 2018-08-27

Abstract


Background: The aim of the current study is to identify factors significantly associated with the survival of acute liver failure (ALF) patients using parametric regression modeling and to estimate  an optimum cutoff value for significant parameters.

Materials and Methods:. The factors studied were age, sex, total serum bilirubin, serum creatinine, serum albumin, urea, aspartate aminotransferase (AST), alanine aminotransferase (ALT), prothrombin time, grade of hepatic encephalopathy and recent hepatitis E virus infection. The hazard functions for these factors were estimated using  Cox proportional hazards model, Weibull regression model under Proportional Hazards (PH) assumption as well as Accelerated Failure Time (AFT) assumption.  

Results: The median survival in days was 42.It was found that Weibull model fits well for the survivorship of ALF patients. Also observed that patients with blood urea level > 50 mg/dL and ALT >470 IU/L had worse prognosis. In Receiver Operator Characteristic (ROC) Curve analysis, the optimum cutoff value of urea was found to be 42 mg/dL and ALT was found to be 400IU/L.

Conclusions: Prognostic markers in ALF such as blood urea and ALT may be used along with the existing criteria to rapidly assess the patient in need of early liver transplantation.

Acknowledgements: The authors are thankful to Mr. Alan Jose and Mr.Noel George, Trainee Biostatisticians, St. Thomas College, Pala, Kerala and Dr.Siddharth Dharamsi ,Post Graduate Student of the Department of Gastroenterology,PSG Hospitals for their support in this study.

Keywords: Survivorship, Acute Liver Failure patients,  Parametric Regression modeling