Prevalence and Impact of Ceftriaxone-Resistant Pathogens on Mortality in Spontaneous Bacterial Peritonitis

Background: Ceftriaxone (CRO) is often used empirically for spontaneous bacterial peritonitis (SBP). Recent studies outside of the United States (US) have shown increased rates of SBP due to CRO-resistant pathogens (16-67%). Little is known about the epidemiology of SBP in the US. This study aims to assess the prevalence and impact of CRO-resistant pathogens on all-cause 30-day mortality in patients with SBP at a large tertiary center in the US.

Methods: This retrospective cohort study included adult patients with liver cirrhosis and culture-positive SBP (polymorphonuclear cells ≥250/mm3 and ≥1 organism in ascitic fluid culture) from 2011-2015. Patients with skin contaminants on ascitic fluid cultures or secondary peritonitis were excluded. CRO-resistant pathogen was defined as a pathogen intrinsically resistant to CRO or classified as intermediate or resistant to CRO per the Clinical and Laboratory Standards Institute breakpoints. Baseline characteristics, variables associated with mortality based on prior literature, and treatment variables were recorded.

Results: A total of 44 isolates from 42 unique patients were included. The prevalence of CRO-resistant pathogens was 43% (19/44): extended-spectrum beta-lactamase-producing Enterobacteriaceae (n=8), Pseudomonas aeruginosa (n=4), methicillin-resistant Staphylococcus aureus (n=1), Enterococcus species (n=4), and Candida species (n=2). The median age was 58 years (range, 27-76), and the median Model for End-Stage Liver Disease score was 25 (range, 9-46). All-cause 30-day mortality was 21% (9/42), which was higher in patients with CRO-resistant pathogens (42% [8/19]) than in those with CRO-susceptible pathogens (4% [1/23]), p=0.006. More patients with CRO-resistant pathogens received inappropriate empiric therapy (58% [11/19]) than those with CRO-susceptible pathogens (17% [4/23]), p=0.01. Multivariate regression showed CRO-resistant pathogens are independently associated with mortality (adjusted odds ratio, 10.73; 95% confidence interval, 1.06-108.63; p=0.044) after adjusting for confounders.

Conclusion: CRO-resistant pathogens were common causative agents in SBP and independently associated with mortality in cirrhotic patients at a tertiary center in the US.