Systemic antifungal therapy in critically ill patients without invasive fungal infection*

To determine the number of adult or pediatric intensive care unit patients without documented invasive fungal infection who receive systemic antifungal therapy.

A 1-day cross-sectional cohort study.

One hundred sixty-nine intensive care units in France and Belgium.

All patients staying in the participating intensive care units.

None.

A hierarchical mixed model was used to identify center-based and patient-based determinants of systemic antifungal therapy use. Day 28 mortality was compared in patients with and without systemic antifungal therapy. Two thousand forty-seven patients were recruited. Systemic antifungal therapy was used in 154 (7.5%) patients, including 100 without and 54 with a proven invasive fungal infection. Overall, systemic antifungal therapy consisted of monotherapy of fluconazole (60%), caspofungin (24%), voriconazole (8%), or liposomal amphotericin B (5%). Independent predictors of systemic antifungal therapy included patient-related factors (severity, emergency surgery, malignancy, Candida colonization, and severe sepsis) and center-related factors (hospital with <800 beds, solid organ transplantation activity, higher annual incidence of candidemia, uncontrolled use of fluoroquinolones, and routine systemic antifungal therapy in patients with unresolved documented or undocumented sepsis). The group given systemic antifungal therapy had greater disease severity and higher rates of sepsis and organ failures. Nevertheless, crude 28-day mortality in the systemic antifungal therapy group was not significantly higher than in the group not given systemic antifungal therapy (20% vs. 19.2%; hazard ratio, 0.97 [0.61-1.52]; p = .88).

Systemic antifungal therapy was used in 7.5% of intensive care unit patients. Two-thirds of patients given systemic antifungal therapy had no documented invasive fungal infection. Our results warrant a trial of systemic antifungal therapy in severely ill intensive care unit septic patients without documented invasive fungal infection based on their severity of illness and the presence of Candida colonization.