Association of fluconazole pharmacodynamics with mortality in patients with candidemia

Abstract

Recent studies of nonneutropenic patients with candidemia or candidiasis suggest that fluconazole pharmacodynamic parameters correlate with clinical outcomes; however, additional data of correlation to mortality in patients with candidemia would be valuable. We assessed the impact of MICs for Candida, fluconazole pharmacodynamics, and patient characteristics on all-cause mortality with use of a prospective cohort of 96 hospitalized patients with candidemia. Among 84 patients for whom Candida isolates were available for testing, the most frequent Candida species isolated were Candida albicans (44%), followed by Candida parapsilosis (20.2%), and Candida glabrata (20.2%). Fluconazole resistance (MIC of >or=64 microg/ml) was present in 7 (8.3%) to 10 (11.9%) of 84 isolates, depending on the MIC endpoint determination method (50% or 80% inhibition read at 24 or 48 h). Overall mortality occurred in 27 (28.1%) of 96 patients, and nonsurvivors were more likely to have fluconazole-resistant isolates (25% versus 6.7%; P = 0.02). Multivariable analysis demonstrated an association between fluconazole resistance and mortality, but it did not reach statistical significance (odds ratio, 5.3; 95% confidence interval, 0.8 to 33.4; P = 0.08). By pharmacodynamic analysis, a fluconazole area under the concentration-time curve/MIC of <11.5 or MIC of >or=64 was associated with increased patient mortality (P <or= 0.09). These data support previous findings of an antifungal exposure-response relationship to mortality in patients with candidemia. In addition, similar MICs were obtained using a 24- or 48-h MIC endpoint determination, thus providing the opportunity to assess earlier the impact of isolate susceptibility on therapy.

FIG. 1.

Relationship between the fluconazole 24-h AUC/MIC and survival in patients with candidemia (n = 84).

FIG. 2.

Results of CART analysis for AUC/MIC and MIC values associated with survival based on MICs assessed at 24 (A) and 48 (B) h. Within each box, the proportion of patients surviving and the sample size (N) are shown.