Comparison of fungal and nonfungal infections in a broad-based surgical patient population

Background: Our aim was to compare fungal and nonfungal infections among a diverse surgical patient population. Methods: Data on all hospital-acquired infectious episodes among surgical intensive care unit and ward patients were collected prospectively over six years at a single university hospital. The relationships between fungal and nonfungal infection and over 100 variables were examined using univariate and multiple logistic regression analysis. Results: During the study period, 3,980 infectious episodes were identified; 554 were associated with fungal infection. Multiple logistic regression analysis demonstrated that markers of severity of acute illness (higher APACHE II scores and white blood cell counts, greater transfusion of cellular blood products, mechanical ventilator dependency, and prior infection) predicted fungal infection, whereas markers of chronic illness (comorbidities) did not independently predict either fungal or nonfungal infection. Patients with fungal infection were treated with more antibiotics for longer periods of time, had prolonged lengths of stay, and more often died compared with nonfungal infection patients. A separate multiple logistic regression analysis demonstrated that both fungal infection and the number of fungal sites of infection independently predicted mortality. Of all fungal isolates, only Candida albicans and Aspergillus spp. independently predicted mortality. Conclusions: Fungal infections differ significantly in character and outcomes from nonfungal infections among surgical patients.