Early adequate empiric antifungal therapy and 28-day mortality in ICU patients with candidemia: a pooled patient-level analysis of EUROBACT-1 and EUROBACT-2 cohorts

Background The optimal timing of empiric antifungal therapy in patients with suspected candidemia remains uncertain. In practice, many antifungal treatments initiated empirically are eventually found to be unnecessary, exposing patients to potential toxicity, selection of non- albicans Candida species, emergence of resistance, and additional costs. The objective of the study was to evaluate whether adequate empiric antifungal therapy initiated within 2 days after blood culture draw was associated with 28-day mortality. Methods We performed a pooled secondary analysis of two prospective worldwide observational cohorts (EUROBACT-1 and EUROBACT-2) including adult patients with Candida bloodstream infection. The exposure was defined as receipt of adequate antifungal therapy within 2 days after blood culture draw. Propensity scores for receiving early adequate antifungal therapy were estimated using prespecified covariates, and inverse probability of treatment weights were applied in a weighted Cox proportional hazards model to assess the association with 28-day mortality. Prespecified subgroup analyses were conducted according to study cohort, Candida species, and presence of hypotension at the time of bloodstream infection. As a sensitivity analysis, a frailty Cox proportional hazards model including a random center effect was performed. Results The final study population included 272 patients with candidemia from 158 centers, including 81 from EUROBACT-1 and 191 from EUROBACT-2. One hundred and ten patients received early adequate antifungal therapy and 162 did not. Overall, 38.6% of patients (105/272) died within 28 days, including 40.9% (45/110) in the early adequate therapy group and 37.0% (60/162) in the non-early adequate therapy group. In the weighted Cox model, early adequate antifungal therapy was not significantly associated with 28-day mortality (HR 1.23, 95% CI 0.82–1.83, p = 0.314). Results were consistent across subgroup analyses according to study cohort, Candida species, the presence of hypotension at the time of bloodstream infection, as well as in the frailty Cox sensitivity analysis. Conclusions Early adequate antifungal therapy was not independently associated with improved 28-day survival in ICU patients with candidemia. These findings support a more targeted approach to early empirical antifungal treatment and underscore the need to identify patient subgroups most likely to benefit from immediate therapy.