Why do echinocandins fail? Identifying key predictors to improve clinical outcomes of candida bloodstream infections: a retrospective multicenter cohort study

ackground: Echinocandins represent first-line therapy for Candida Bloodstream Infections (C-BSIs). In- cidence of treatment failure (TF) remains high with unclear risk factors. Aim: to evaluate predictors of echinocandin TF for C-BSIs. Methods: Retrospective observational multicenter study, enrolling all patients with C-BSI treated with echinocandin from 01/06/2020 to 30/06/2023 in four Italian Hospitals. Primary outcome: to evaluate predictors of TF defined as a composite of: i)transfer to ICU or any worsen- ing in organ dysfunction at day 5 of therapy; ii)Persistent C-BSI; iii)Echinocandin discontinuation for any reason; iv)Onset of a new infection site by Candida spp. during treatment. Secondary outcome: 90-day all-cause mortality. Cox regression and treatment-effect were used, along with inverse-probability of treatment-weighting (IPTW) to adjust cohort treatment-assignment bias. Results: Overall, 218 patients were enrolled. Median (q1-q3) age was 72 (56-78), 55% male. In 33% and 63% of cases, septic shock at presentation and C-BSIs by non-albicans strains were reported. Importantly, 68 (31%) patients received high dosage echinocandin (“HDE”: increase of 30-50% of standard dosage), ac- cording to clinical judgement. Eighty-two (38%) experienced TF; 90-day all-cause mortality was 30%, sig- nificantly higher in TF-group (P < 0.001). At multivariable Cox-regression analysis, obesity, septic shock, and increased MIC to echinocandins were predictors of TF; presence of removable intravascular devices and HDE resulted protective. After adjustment by inverse-probability of treatment-weighting, HDE still reduced TF risk in patients admitted to the ICU, with SOFA score<6, BMI>30, or with serum albumin concentration ≤2,5gr/dL.