Analysis of the relationship between fluconazole consumption and non-C. albicans Candida infections

The effect of fluconazole consumption on the incidence of nosocomial non-C. albicans Candida infections remains unclear. In this study we investigated such a relationship in an intensive care unit (Poland) over an 11-year period (2002-2012). Statistics relating to the number of candidiasis cases and the number of defined daily doses of fluconazole showed that only a very weak and not statistically significant linear correlation existed between these two variables (r(2) = 0.36, P = 0.052). However, the assumption of a 1-year delay in the infection response to changes in fluconazole concentrations resulted in a strong and statistically significant linear correlation (r(2) = 0.64, P = 0.0053). To more accurately determine the nature of this relationship, a simple epidemiological model was proposed that provided a better than linear correlation (r(2) = 0.78, P = 0.00077). We successfully used this approach to analyze results from the literature that were interpreted as evidence that fluconazole use is not a risk factor for development of non-C. albicans Candida infections. If a time delay in the infection response was assumed, a strong and statistically significant correlation was obtained. These findings suggest the need for a closer look at fluconazole therapy as a possible risk factor for development of non-C. albicans Candida infections.