Candida species have emerged as frequent causes of nosocomial bloodstream infection (BSI) in association with well-defined risk factors, including prolonged hospitalization, abdominal surgery, antibiotic treatment, neutropenia and central venous catheterization (14). Candidemia is associated with high rates of attributable mortality, prolongation of hospital stay, and excessive costs (28). In recent years, there has been a shift in the distribution of Candida species causing invasive infection, with non-albicans species now surpassing Candida albicans in many institutions (14,25). Of particular concern is the rising incidence of the azolenonsusceptible species C. glabrata and the inherently fluconazoleresistant species C. krusei (11,25,27).Fluconazole is often used as empirical treatment of candidemia. However, given the correlation between the survival rate and the timely initiation of appropriate treatment for candidemia (8), accurate assessment of the risk of fluconazole-resistant Candida (FRC) BSI is of prime importance. Patients who were recently treated with an azole drug are at increased risk of infection with FRC (9) and should be treated initially with an echinocandin agent according to current guidelines (18). However, experimental and clinical data support the notion that nonantifungal antimicrobial agents also affect the risk of colonization and infection with FRC (15,17,22). Since exposure to antibacterial drugs among at-risk patients far exceeds exposure to antifungal agents, even modest effects of individual antibacterials could translate into significant overall changes in the susceptibility patterns of Candida spp. Nevertheless, the collateral effects of antibacterial drugs on Candida spp. are poorly understood. To address this question, we analyzed prospectively collected data from a nationwide study of candidemia in Israel and examined the association between exposure to antifungal and antibacterial agents and the risk of infection with FRC.
MATERIALS AND METHODS
Study design.We performed a prospective nationwide study of candidemia in Israel from November 2005 through June 2007. Eighteen medical centers, which together account for 75% of the hospital beds in Israel, were included. All candidemia episodes that occurred in the participating centers during the study period were eligible for inclusion in this study. Clinical data were prospectively entered into standardized data forms by on-site investigators at each of the centers. The Candida sp. clinical isolates underwent preliminary identification and susceptibility testing in each center according to local practices. Subsequently, the isolates were transferred together with the corresponding data forms to the central study site, where species identification and susceptibility testing were performed as detailed below. The data forms were collected by the study coordinator, reviewed by the principal investigator, and entered into a computerized database. The study was approved by the ethics committee of each of the participating centers.Data...