Abstract
Background Mortality among patients with liver cirrhosis has recently increased in Indonesia. However, predicting the prognosis of patients hospitalized with liver cirrhosis remains a clinical challenge due to its variability and dependence on multiple factors. A simple and accurate method is required to identify high-risk patients. This study aims to build a predictive scoring system of in-hospital mortality in patients hospitalized with liver cirrhosis for clinical application. Methods A retrospective cohort study was done to collect data on patients with liver cirrhosis from November 2021 to January 2022. The study involved 110 patients hospitalized with liver cirrhosis. A multivariate regression analysis was performed to identify factors predicting in-hospital mortality. Each variable's score was determined by applying the (B/SE)/lowest B/SE formula. The overall probability was calculated using the equation 1/1+exp(-y). Analysis of area under the curve (AUC) was conducted to evaluate the sensitivity and specificity of the scoring system. Results A total of 52 patients (47.3%) died during hospitalization. The median age of the patient was 54.5 (30-82). A final model involving the presence of hepatic encephalopathy (HE) (p = 0.001), ascites (p = 0.025), diabetes mellitus type II (p = 0.003), acute kidney injury (p = 0.017), alanine transaminase (ALT) ≥ 68 (p = 0.001), creatinine level ≥ 1.25 (p = 0.011), and abnormal international normalized ratio (INR) (p = 0.047). The "ADRECIA" score was developed, consisting of ascites, type II diabetes mellitus, renal injury, hepatic encephalopathy, creatinine serum, INR, and ALT. The presence of HE and ALT ≥68 was scored as two, and the rest variables were scored as one. The best-discriminating value was at a cut-off point ≥ 2.5, with a sensitivity of 90.4%, and a specificity of 74.1%, and an AUC of 0.913 (95%CI: 0.862-0.964). The scoring system was categorized as low risk (score of zero to three) with a 1.4-43.6% probability of death and high risk (score of 4-9) with 74.7-99.9% probability of death. Conclusion This scoring system provides good accuracy in predicting in-hospital mortality in patients with liver cirrhosis. Therefore, treatment can be modified according to the score to reduce mortality rates.