OBJECTIVE The relationship between patient characteristics and posaconazole exposures was evaluated in a population pharmacokinetic (PK) model using trial data from neutropenic patients administered oral posaconazole suspension as antifungal prophylaxis. METHODS Data were analyzed using nonlinear mixed-effects modeling. Covariates were tested using the forward addition, Objective Function (OF) cut-off of 3.84, followed by the backward elimination (OF cut-off 10.88) steps in NONMEM. These covariates included demographics, mucositis, neutropenia, vomiting, diarrhea, proton pump inhibitor (PPI) or H(2)-receptor antagonist usage and baseline bilirubin or baseline gamma-glutamyl transferase (GGT) levels > or =2 x upper limit of normal (ULN). A correlation between posaconazole PK and the occurrence of invasive fungal infection (IFI) was also examined. RESULTS Statistically significant associations were demonstrated between posaconazole PK and diarrhea, PPI intake, race, and baseline GGT and bilirubin levels. These covariates did not predominate in patients who developed IFI. CONCLUSIONS This analysis provides information regarding the correlation of patient covariates with posaconazole exposures estimated in a clinical setting. The results of this analysis agree with previously reported analyses. However, because of the successful prophylaxis and the low number of posaconazole-treated patients with IFI proven or probable (IFIPP), the absence of a statistically significant relationship between IFIPP and exposure may not mean this relationship does not exist. A meta-analysis of several efficacy trials or exploring alternate composite endpoints for efficacy may be needed to answer this question.