Background: Infections are common in patients with liver cirrhosis and increase mortality. We explored the relationship between infection and liver dysfunction in their effects on mortality.
Methods: Single-center data on decompensated liver cirrhosis patients hospitalized between March 2014 and December 2017 (index period) were reviewed until death, liver transplantation or 31 December 2018. Infections were classified as community-acquired infection (CAi) or hospital/healthcare associated infection (HCAi). Child-Pugh, model for the end-stage liver disease (MELD) and chronic liver failure-organ failure (CLiF-OF) scores indicated liver (dys)function.
Results: We enrolled 155 patients (85% alcoholic liver disease), 65 without infection at first hospitalization, 48 with CAi and 42 with HCAi. Multidrug resistant agents were confirmed in 2/48 (4.2%) CAi and 10/42 (23.8%) HCAi patients. At first hospitalization, infection was independently associated with worse liver dysfunction and vice versa, and with higher 30-day mortality [odds ratio (OR) = 2.73, 95% confidence interval (CI) 1.07-6.94]. The association was reduced with adjustment for MELD/CLiF-OF scores, but mediation analysis detected an indirect (via liver dysfunction) association. Twenty-eight patients were repeatedly hospitalized, 11 with new HCAi. HCAi was independently associated with twice higher risk of medium-term mortality and added an additional risk to any level of liver dysfunction, considering all or patients who survived the first 30 days. In those repeatedly hospitalized, HCAi appeared independently associated with a higher probability of infection and higher MELD scores at subsequent hospitalizations.
Conclusion: Infection (particularly HCAi) adds mortality risk to any level of liver dysfunction in decompensated liver cirrhosis patients. Mechanisms of long(er)-term effects (in acute episode survivors) seemingly include enhanced deterioration of liver function.