A simple covert hepatic encephalopathy screening model based on blood biochemical parameters in patients with cirrhosis.

Abstract

Covert hepatic encephalopathy (CHE) adversely affects clinical outcomes in patients with liver cirrhosis, although its diagnosis is difficult. This study aimed to establish a simple CHE screening model based on blood-related biochemical parameters. This retrospective study enrolled 439 patients who were assessed for CHE using a neuropsychiatric test between January 2011 and June 2019. A simple CHE (sCHE) score was calculated with hypoalbuminemia (≤ 3.5 g/dL) and hyperammonemia (≥ 80 μg/dL) as 1 point each. The association between sCHE score and CHE or overt hepatic encephalopathy (OHE) was assessed using logistic regression and Fine-Gray competing risk regression models. Of 381 eligible patients, 79 (21%) were diagnosed with CHE. The distribution of sCHE scores was 48% with 0 point, 33% with 1 point, and 19% with 2 points. Patients with sCHE score ≥ 1 point had a higher prevalence of CHE than those with sCHE score of 0 (27% vs. 14%, P = 0.002). A cut-off value of 1 point showed high discriminative ability for identifying CHE, with a sensitivity of 0.67, specificity of 0.56, positive predictive value of 0.27, and negative predictive value of 0.86. During the median follow-up period of 2.2 years, 58 (15%) patients developed OHE. Multivariate analysis showed that sCHE score ≥ 1 (sub-distribution hazard ratio [SHR], 2.69; 95% confidence interval [CI], 1.41-5.15) and CHE (SHR, 2.17; 95% CI, 1.26-3.73) independently predicted OHE. The sCHE score is a useful screening model for identifying patients with CHE and for predicting OHE occurrence.