Risk factors and nomogram of overt hepatic encephalopathy and its impact on outcome in patients with acute-on-chronic liver failure

Overt hepatic encephalopathy (OHE) is a common and devastating complication of liver failure. This study aims to investigate the characteristics of OHE and its impact on outcome in patients with acute-on-chronic liver failure (ACLF), and to construct and validate a nomogram-based prediction model for OHE development. A total of 330 ACLF patients who met the World Gastroenterology Organization (WGO) criteria (ChiCTR1800017991; registered at The Third Central Hospital of Tianjin on August 25, 2018) were enrolled as a training cohort to identify the risk factors of OHE occurrence and to build an OHE model for predicting its development. Another 333 ACLF patients from a multicenter retrospective cohort in China (ChiCTR1900021539; registered at The Third Central Hospital of Tianjin on February 26, 2019) were enrolled as a validation cohort to externally validate the OHE model. Among the 330 patients in the training cohort, 158 (47.9%) suffered from OHE. The prevalence of grade 2, 3, and 4 OHE during hospitalization was 27.2%, 45.6%, and 27.2%, respectively. Among these OHE patients, 79.1%, 13.9%, and 7.0% were episodic, recurrent, and persistent OHE, respectively. Patients with OHE were older, had worse baseline hepatic and renal function, more frequent liver-related complications, and higher 28- and 90-day mortality than those without OHE. The multivariable logistic regression model (OHE model) to predict the occurrence of OHE included age, gastrointestinal (GI) bleeding, bacterial infection, ammonia, and INR. The AUROC of the OHE model for predicting the occurrence of OHE in ACLF patients was higher than those of the CTP, MELD, MELD-Na and CLIF-SOFA scores in both internal and external validations. The 90-day mortality was 59% in patients with high OHE model scores. OHE was associated with higher short-term mortality in ACLF patients. The predictive model of OHE occurrence, which includes age, GI bleeding, bacterial infection, ammonia and INR, may help clinicians identify patients at high risk of OHE and optimize management strategies.