TABLE 3.

Analysis of factors related to mortality

VariableUnivariate analysisaMultivariate analysisa
OR95% CIP valueOR95% CIP value
Age of >63b4.41.7-11.70.0038.92.0-38.90.004
Male sex0.70.3-1.60.35
White race1.10.4-2.60.90
APACHE IIc1.21.08-1.30.00021.21.03-1.30.01
Malignancy0.30.09-1.20.10
Neutropeniad0.20.03-1.90.17
Intubated3.51.4-9.00.008
Candida colonization3.81.5-9.60.006
ICU locatione3.81.5-9.60.006
Time to fluconazolef1.20.8-1.80.31
Fluconazole resistant (MIC4850)g4.71.2-18.40.035.30.8-33.40.08
  • a For multivariate logistic regression analysis, variables with P of <0.20 on univariate analysis were included in a stepwise regression model in addition to variables for age, race, sex, and time to fluconazole administration. When a continuous variable was significantly related to mortality as such or as a categorical variable, the most significant of these variables was evaluated by logistic regression analysis. One MIC (MIC4850) was chosen for the final model based on strength of association with mortality. The best model fit was with the values of the MIC4850 or MIC4880, which showed identical results. Results are listed for significant variables (P < 0.05) and the MIC variable. P values obtained are two tailed.

  • b The value 63 years was based on breakpoint from CART analysis.

  • c At time of culture as acontinuous variable (1-point increments).

  • d Defined as absolute neutrophil count of <500 at time of culture.

  • e Location at time blood culture was drawn. ICU, intensive care unit.

  • f Defined as time from culture to start of fluconazole (days).

  • g Fluconazole resistance was defined as a MIC of ≥64 μg/ml.