Predictors of mortality among patients with compensated and decompensated liver cirrhosis: the role of bacterial infections and infection-related acute-on-chronic liver failure.

OBJECTIVE: Population-based data on the impact of bacterial infections on the course of compensated and decompensated cirrhosis as well as the occurrence, predictors of infection-related acute-on-chronic liver failure (ACLF) and its fatal outcome are limited.

MATERIAL AND METHODS: All patients with incident cirrhosis in the period 2001-2010, residing in an area of 600,000 inhabitants, were retrospectively identified. All serious bacterial infections (resulting in or occurring during an inpatient hospital episode) during this period were analyzed. Infection site and acquisition type, comorbid illness (Charlson comorbidity index) and infection severity features were analyzed. Patients were followed up until death, transplant, or the end of 2011.

RESULTS: Overall, 398 serious bacterial infections occurred in 241/633 (38%) patients (106/332 diagnosed with compensated and 135/301 with decompensated disease; follow-up time was 2276 patient-years). ACLF occurred in 95/398 (24%) serious infections with an in-hospital mortality of 50%. In logistic regression analysis, the model for end-stage liver disease score, active alcohol misuse and healthcare-associated infections were predictors of infection-related ACLF (p < 0.05 for all). In-hospital mortality in infections with ACLF was related to albumin levels, Charlson comorbidity index >1 and occurrence of one or more organ failures (p > 0.05 for all). In Cox regression analysis, infection-related ACLF was an independent negative predictor of transplant-free survival in decompensated patients (p = 0.049).

CONCLUSIONS: In a population-based cirrhotic cohort, infection-related ACLF was a negative predictor of survival in decompensated disease. Infection-related ACLF was frequent and related to cirrhosis severity and infection acquisition type, as well as to high inpatient mortality, in particular in patients with significant comorbidity.