Hepatic encephalopathy (HE) is a watershed moment in the natural history of cirrhosis portending decreased quality of life and worsening prognosis. Recent strides in the management of HE have been made to decrease symptom burden and readmissions. The impact of these interventions requires a contemporary re-examination of the natural history of HE and its clinical implications.
We examined data from a 20% random sample of US Medicare enrollees with cirrhosis and continuous Part D prescription coverage from 2008–2014. Those with a diagnosis of HE prior to or within 3 months after the diagnosis of cirrhosis was made were excluded. Incident HE was defined by ICD-9 code 572.2 and/or the initiation of a prescription for an HE-specific treatment (Neomycin, Lactulose, or Rifaximin). Outcomes included transplant-free survival and hospital-days or 30-day readmissions per person-year. Multivariate analysis was performed for survival (hazard ratios, HR, Cox regression) and hospital utilization (incidence rate ratios, IRR, negative binomial regression).
Among 186,160 Medicare-enrollees (median age 65 years) with cirrhosis, 49,164 experienced HE (26.4%). The median survival following cohort entry of those who did and did not develop HE was 5.78 and 3.4 years, respectively (P < 0.001). Multivariate analysis identified decreased survival with older age (HR 1.02, CI: 1.02–1.03), male sex (HR 1.21, CI: 1.19–1.24), ESRD (HR 1.08, CI: 1.01–1.14), and increasing Charlson Comorbidity Index (HR 1.2–1.42, CI: 1.17–1.48). Cirrhosis etiologies of HCV and alcohol were associated with improved survival (HR 0.87 CI: 0.85–0.90 and HR 0.82 CI: 0.79–0.85, respectively) while NAFLD was linked to increased mortality after HE (HR 1.07, CI: 1.02–1.12). Hospital-days per person year were 11.8 in patients with HE compared to 2.9 in those without (P < 0.001). Factors that were inversely associated with hospital utilization were Rifaximin use (HR 0.40, CI: 0.39–0.42) and gastroenterology consultation (HR 0.73, CI: 0.67–0.80). Rifaximin use was associated with decreased hospital-days (IRR 0.35, CI: 0.33–0.37) and fewer 30-day readmissions (HR 0.18, CI: 0.08–0.40), while gastroenterology consultation was associated only with a decreased risk of 30-day readmissions (HR 0.71: CI: 0.57–0.88) but not overall hospitalizations.
The outcomes after HE in this contemporary Medicare-insured population are poor. The development of HE increases hospitalization utilization and worsens survival with few, potentially modifiable, targets for prospective study intervention.