423. Ten-Year Experience of Burkholderia pseudomallei Infections in a Singapore Tertiary Hospital

BACKGROUND: Burkholderia pseudomallei is endemic in the tropics and associated with high mortality. We performed a retrospective study analyzing the clinical and microbiologic features of melioidosis, and predictors of mortality. METHODS: Patients with culture-positive melioidosis from 2006 to 2016 were identified from microbiologic records. Clinical data including demographics, treatment, and outcomes were extracted from medical records. Categorical variables were compared using χ(2) test or Fisher exact test while continuous variables were compared using Student’s t-test or Mann–Whitney U test. RESULTS: Forty-three cases of melioidosis were identified. Presentations included fever (41.9%), respiratory symptoms (20.9%), and joint swelling (9.3%). 76.7% were bacteremic and 69.7% were culture-positive from a nonblood source. Mean time from presentation to positive microbiological data was 5.1 ± 6.4 days. Infection sites included pulmonary (62.8%), spleen (27.9%), skin/soft tissue (25.6%), and bone/joint (25.3%). Antibiotic susceptibility were as follows: ceftazidime (97.5%), imipenem (100.0%), trimethoprim–sulfamethoxazole (92.1%), amoxicillin–clavulanate (94.7%), and doxycycline (94.7%). Mean time from presentation to melioid-active coverage was 6.8 ± 9.1 days. Thirty-day all-cause mortality occurred in nine patients (from first positive culture); one patient died within 5 months. Univariable analysis associations with 30-day all-cause mortality were: intensive care unit (ICU) admission (OR 26.3, 95% CI 4.0–173.1, P < 0.01), mechanical ventilation (OR 15.0, 95% CI 2.6–85.0, P < 0.01), higher median Pitt Bacteremia Score (PBS) (4.0 vs. 2.0; P < 0.01), receipt of ceftazidime (vs. a carbapenem) as primary induction antibiotic therapy (OR 0.2, 95% CI 0.03–0.91, P = 0.047) and not receiving melioidosis-active induction intravenous antibiotics (P = 0.04). Multivariable analysis found mechanical ventilation to be an independent predictor for 30-day mortality (P = 0.003, OR 18.8, 95% CI 2.7–130.9). CONCLUSION: ICU admission, a high PBS, and in particular, receipt of mechanical ventilation may help identify patients with high mortality risk. Delays in melioid-active therapy were not uncommon. Prompt recognition of melioidosis and early institution of active therapy, especially in the critically ill, may reduce mortality. DISCLOSURES: All authors: No reported disclosures.