Utility of the hepatic encephalopathy scoring algorithm (HESA) for diagnosing hepatic encephalopathy in a randomized, controlled trial of rifaximin vs. Placebo

Abstract

Background: The Conn score (CS) is the most widely used criteria in assessing the severity of hepatic encephalopathy (HE) in clinical practice and clinical trials. Although clinically simple, it is subjective, lacks specific definitions, and can be inaccurate in differentiating grades of HE. The Hepatic Encephalopathy Scoring Algorithm (HESA) was developed to capitalize on the strengths of the CS and improve it by incorporating brief, objective, and validated indicators of cognitive impairment that are independent of age and education. HESA relies primarily on clinical examination in the more severe grades, and more heavily on objective cognitive testing in the lower grades in which dysfunction may be subtle and less obvious in clinical examination. The HESA was recently incorporated into a large, multicenter, randomized controlled trial of rifaximin (RFX) vs. placebo. The objective of this analysis was to assess the utility of the HESA as a tool for enhancing HE grading in randomized, controlled trials. Methods: The efficacy of RFX for maintenance of remission from overt HE was examined in a multicenter, multinational, placebo-controlled clinical trial in cirrhotic patients (n=299) with recurrent HE and in remission at enrollment (CS of 0 or 1). Breakthrough overt HE was defined as an increase in CS to ≥2, or an increase in both CS and asterixis score of 1 grade each for patients entering with a CS of 0. The HESA was used to derive the CS score (see Figure). The ability of HESA parameters to differentiate HE grades at baseline and post-baseline was examined. Results: 299 participants (mean age: 56.2 years, 60.9% male) were randomized in 70 study centers in the US, Canada, and Russia. The mean (SD) duration of advanced liver disease was 56.2 (58.2) months, and mean (SD) time since last HE episode was 71.1 (49.6) days. 200 patients (66.9%) had CS score of 0 at baseline and 99 patients (33.1%) had a score of 1. Significant differences were observed at baseline between patients classified as 0 or 1 on the CS for all clinical (P <0.0001) and neuropsychological (P<0.05) HESA indicators (Table). Post-baseline measurements indicated that 18 patients had progressed to a CS of 2, indicative of breakthrough HE. Significant differences were observed at post-baseline between CS of 0 vs. CS of 1 and CS of 0 vs. CS of 2 for most HESA clinical and neuropsychological indicators. Fewer differences were observed between CS of 1 vs. CS of 2 for HESA indicators. There was no significant variability within or among study sites, suggesting good inter-rater reliability of the HESA. Conclusions: The HESA, an easily administered, standardized HE grading tool, was found to have good precision in differentiating CS grades 0 and 1 and correlated well with the Conn score. HESA may have utility for enhancing the objectivity of HE grading in large, multicenter trials. (Table Presented).