The cirrhosis-hospital frailty risk score predicts mortality in patients with chronic liver disease: Analysis of a nationally representative database.

Abstract

The prevalence of cirrhosis is increasing in the United States, coinciding with a demographic shift towards older and more medically complex patients. The Hospital Frailty Risk Score (HFRS) is widely used to measure comorbidity burden in administrative data. However, it is not specific to cirrhosis patients. Our group developed the cirrhosis-HFRS (cHFRS), a simplified tool that predicts in-hospital mortality among cirrhosis patients undergoing abdominal surgery. This study aims to externally validate the cHFRS in a broader cohort of hospitalized patients with chronic liver disease. Adult patients with cirrhosis were identified from the National Inpatient Sample, 2016-2018. The predictive performances of cHFRS and HFRS were compared by discrimination and calibration. Adjusted associations between cHFRS and in-hospital mortality were computed using logistic regression. Stratified analyses and formal tests for interaction assessed whether age, sex, and liver disease severity modified the association between cHFRS and mortality. Among 407,739 cirrhosis inpatients, cHFRS demonstrated superior predictive accuracy for in-hospital mortality (AUC=0.86 vs. 0.75 for HFRS, p<0.001). In adjusted models, a high cHFRS score (≥89th percentile) was associated with an increased risk of in-hospital mortality (OR=30.24, p<0.001). Interaction analysis demonstrated an increase in the magnitude of the association between cHFRS and mortality among younger patients compared to older patients (OR=41.95 for <65 years vs. OR=19.58 for ≥65 years, p<0.001). Among patients with cirrhosis, the cHFRS demonstrated superior prediction of in-hospital mortality, compared to the original HFRS score. These findings support the use of cHFRS rather than HFRS as a comorbidity index in future studies of cirrhosis patients utilizing administrative data.