256 - Serratia Marcescens Bacteremia and Endocarditis: A Treatment Assessment from an Academic Medical Center

Background: Serratia Marcescens (SM) is often an opportunist that has been associated as a cause of healthcare-associated infection and in some people who inject drugs (PWID). Decisions about the treatment of SM infections are difficult given the small clinical studies available and concerns for multidrug resistance. SM has the ability to produce inducible AmpC b-lactamase and may acquire extended-spectrum b-lactamase (ESBL).  Evidence-based guidance is lacking in terms of identifying preferred antimicrobial therapy of SM bacteremia and endocarditis. Compared to other reports, our hospital has one of the largest SM data sets to compare.

Methods: This observation study included adult patients admitted to our hospital (2016-2019) with SM bloodstream infections, including endocarditis. Patients were excluded from the analysis, if they had a concomitant infection with another gram-negative organism. Our evaluation was designed to: compare outcomes associated with different antibiotic regimens, evaluate how care differed in PWID patients versus others, and identify factors associated with obtaining infectious diseases expert consultations (ID Consult).

Results: Forty-three patients met study inclusion/exclusion. Twenty-eight patients (65.1%) had ID Consults.  Twenty-four (55.8%) were PWID. Endocarditis was diagnosed in 30.2% of patients. The most common regimen was cefepime +/- aminoglycoside, followed by a carbapenem +/- aminoglycoside. Combination therapy was only recommended during ID Consult. Piperacillin-tazobactam was used in 11.6% of patients. No regimen displayed an efficacy or safety advantage over another. Most patients (90.7%) cleared their blood stream within 48 hours of antibiotic start. Phenotypic susceptibility testing did not identify either ESBL or AmpC production in any of the isolates, including recurrences. Multi-drug resistance was not appreciated. Significant factors associated with obtaining ID Consult were: PWID (p=0.004), endocarditis (p=0.0002), sepsis (p=0.022), surgical intervention (p=0.003).  

Conclusion: We could not identify an advantage with any particular antibiotic treatment regimen in this study. Induction of AmpC or selection of ESBL organisms was not displayed by any of the organisms.