199. Infections in VADers: A True Villain of the Force

Background Ventricular assist devices (VADs) are increasingly used for the management of end-stage heart failure, but infection is a complication that has not been thoroughly studied. The purpose of our study was to compare patients who had surgical debridement vs. medical therapy alone for VAD-related/specific infections. Methods We performed a retrospective chart review on patients at Duke University Hospital (DUH) from 2015 to 2017. Patients with VAD-related/specific infections were included, per 2011 ISHLT definitions. We reviewed electronic medical records for demographics, VAD implantation data, infectious episodes, surgical debridements and mortality. Descriptive statistics compared patients with and without debridement and compared with and without relapse. Results We found 94 infections in 72 patients. Descriptive statistics of the cohort and comparisons with and without debridement can be seen in Table 1. Sixty-one cases (65%) included debridement and 5 (5%) required pump exchange. Notably, patients with fever or bacteremia were more likely to undergo debridement. Of the patients that had a preoperative CT, sensitivity for deep infection (pump, pocket, or deep to the muscle) was 38%, yet specificity was 95%. For superficial infections (involving the driveline or superficial to the muscle), preoperative CT sensitivity was 95%; specificity 65%. Table 2 shows intraoperative culture data. When the preoperative driveline culture grew Staphylococcus species or Pseudomonas aeruginosa there was strong correlation with intraoperative organism (matched in >75% of cases). Table 3 compares treatments among patients with and without infective relapse. Relapse rate appeared the same if patients received 2, 4, or ≥6 weeks of intravenous antibiotics. Conclusion We present a large single-center cohort [DCWM1] examining VAD-related/-specific infections. While patients chosen for debridement may be sicker, these patients had a longer hospital stay and relapsed more often. Preoperative CT should be used with caution as it underestimates the extent of disease. However, preoperative driveline cultures correlated strongly with intraoperative cultures for most common pathogens. There was no association between initial intravenous therapy duration and infection relapse. Table 1. demographic characteristics of total cohort and comparisons among patients who underwent debridement for treatment of infection and patients who did not undergo debridement for treatment of infection Table 2. Organisms found intraoperatively and the number of organisms found preoperatively that are the same as the intraoperative cultures Table 3. Comparing treatment and cultures in patients who suffered an infection relapse Disclosures All authors: No reported disclosures.