2257. BDG-Guided Management of Empirical Antifungal Therapy: a Real-life Experience in a Hospital-Wide Context with High Incidence of Non-albicans Candida Infection

Abstract Background BDG-guided management of empirical antifungal therapy (AT) has been suggested as a tool to rule out invasive candidiasis (IC) and discontinue AT in critically ill patients. However, some authors reported lower BDG sensibility for non-albicans Candida (NAC) infection. Impact of BDG use in a hospital wide setting has yet to be determined. Methods We performed a retrospective observational study of consecutive patients admitted to a 1535-bed teaching hospital from November 2015 to August 2017. Adult patients starting empirical AT and performing at least one BDG test for a suspected fungal infection were included. According to first BDG result and AT management, patients were classified in 3 groups: (G1) negative index BDG and early AT withdrawal; (G2) negative index BDG and AT continued; (G3) positive index BDG and AT continued. IC was defined as monomicrobial Candida spp. isolation from blood cultures and/or surgical specimen. Comparison of the 3 groups was made using post-hoc Bonferroni correction. Univariate and multivariate analyses of risk factors for all-cause 30 days mortality were performed using binary logistic regression model. Results Study population consisted of 208 patients, of which 46 (22.1%) were included in G1, 79 (38.0%) in G2, and 83 (39.9%) in G3. NAC species were more commonly isolated from patients with IC and negative BDG (P = 0.005). IC was diagnosed in 2.2%, 13.9%, and 19.3% of G1, G2, and G3, respectively (P < 0.01 for G1 vs. G3). Median AT DDD were 8, 28, and 20 (P < 0.01 for G1 vs. G2 and G1 vs. G3) and 30-day mortality rate was 21.4%, 16.5%, and 30.1%, respectively. Factors associated with 30-day mortality were age, Charlson Comorbidity Index (CCI), ICU admission, SOFA score, septic shock, orotracheal intubation, CVVH and index BDG ≥ 80 pg/mL. At multivariate model, independent risk factors for 30-day mortality were CCI (OR 1.4, 95% CI 1.2–1.6, P < 0.001), SOFA score (OR 1.2, 95% CI 1.1–1.3, P < 0.001) and index BDG≥80 pg/mL (OR 2.4, 95% CI 1.2–4.8, p = 0.012). Model fit was P = 0.65 by Hosmer–Lemeshow test and accuracy according with ROC analysis was 0.82 (95% CI 0.76–0.88). Conclusion BDG positivity is a strong predictor of poor outcome, but its accuracy for NAC infection may be suboptimal. Caution may be necessary for AT discontinuation based on a negative BDG result in patients at high risk for NAC infection. Disclosures All authors: No reported disclosures.