Infection in Hospitalized Cirrhosis Patients: Changing Epidemiology and Clinical Features

Background: Patients with cirrhosis are uniquely predisposed to infections, which can lead to acute decompensation and an increase in mortality rates. We hypothesized that not only are cirrhotic patients more likely to develop certain infections, but that specific infections are associated with poorer outcomes. Therefore, we aimed to examine the epidemiology, bacteriology, and outcomes of infections in cirrhotic patients admitted to the hospital. Methods: In this single center observational retrospective cohort study, we identified admissions in which patients had an infection from a group of all admissions of cirrhotics from 2011-2016. Infections were categorized by the primary source of infection, and rigorous clinical and bacteriologic definitions were used. Results: We identified 1,208 admissions in 877 unique patients during the study period. The most common infections identified were as follows: urinary tract infections (33%), pneumonia (23%), spontaneous bacterial peritonitis (14%), and bacteremia (11%). Gram-positive organisms were most commonly isolated in patients with spontaneous bacterial peritonitis and bacteremia, whereas gram-negative bacteria were most prevalent in urinary tract infections and pneumonia. Candida infections were common and identified in the following proportions: spontaneous bacterial peritonitis (16%), pneumonia (14%), bacteremia (13%), and urinary tract infections (9%). Pneumonia, spontaneous bacterial peritonitis, and meningitis were associated with increased mortality rates (29%, 32%, and 67%, respectively), compared to the overall mortality rate of 20% (p-value < 0.05). Conclusions: In summary, infections were common in patients with cirrhosis and were associated with poor outcomes, particularly in the presence of evidence of sepsis. Spontaneous bacterial peritonitis and bacteremia are now most commonly due to gram-positive organisms and fungal infections appear to be rising in prevalence.