LAUSR

OBJECTIVE Mycotic aortic aneurysms and aortic graft infections (aortic infections [AIs]) are rare but highly morbid conditions. Open surgical repair is the "gold standard" treatment, but endovascular repair (EVR) is increasingly being used in the management of AI because of the lower operative morbidity. Multiple organisms are associated with AI, and bacteriology may be an important indication of mortality. We describe the bacteriology and associated outcomes of a group of patients treated with an EVR-first approach for AI. METHODS All patients who underwent EVR for native aortic or aortic graft infections between 2005 and 2016 were retrospectively reviewed. Primary end points were 30-day mortality and overall mortality. The primary exposure variable was bacteria species. Logistic regression analysis was used to determine association with mortality. Kaplan-Meier survival analysis was used to estimate survival. RESULTS A total of 2038 EVRs were performed in 1989 unique and consecutive patients. Of those, 27 patients had undergone EVR for AI. Thirteen presented ruptured (48%). Eighteen (67%) were hemodynamically unstable. Ten had a gastrointestinal bleed (37%), whereas others presented with abdominal pain (33%), fever (22%), chest or back pain (18.5%), and hemothorax (3.7%). Twenty patients had a positive blood culture (74%), with the most common organism being methicillin-resistant Staphylococcus aureus (MRSA) isolated in 37% (10). Other organisms were Escherichia coli (3), Staphylococcus epidermidis (2), Streptococcus (2), Enterococcus faecalis (1), vancomycin-resistant Enterococcus (1), and Klebsiella (1). Thirteen patients had 4 to 6 weeks of postoperative antibiotic therapy, six of whom died after therapy. Fourteen were prescribed lifelong therapy; 10 died while receiving antibiotics. On univariate analysis for mortality, smoking history (P = .061) and aerodigestive bleeding on presentation (P = .109) approached significance, whereas MRSA infection (P = .001) was strongly associated with increased mortality. On multivariate analysis, MRSA remained a strong, independent predictor of mortality (adjusted odds ratio, 93.2; 95% confidence interval, 1.9-4643; P = .023). Overall 30-day mortality was 11%, all MRSA positive. At mean follow-up of 17.4 ± 28 months, overall mortality was 59%. Overall survival at 1 year, 3 years, and 5 years was 49%, 31%, and 23%. Kaplan-Meier survival analysis demonstrated that MRSA-positive patients had a significantly lower survival compared with other pathogens (1-year, 20% vs 71%; 5-year, 0% vs 44%; P = .0009). CONCLUSIONS In our series of AI, the most commonly isolated organism was MRSA. MRSA is highly virulent and is associated with increased mortality compared with all other organisms, regardless of treatment. Given our results, EVR for MRSA-positive AI was not a durable treatment option.