The addition of an antifungal agent with activity against molds is considered standard of care for neutropenic patients with persistent fever treated with broad spectrum antibacterial antibiotics. The currently available intravenous agents suffer from high cost or high rates of toxicity. Voriconazole, a triazole with excellent activity against yeast and molds, has an oral bioavailability of 96% and comparable AUCs following IV and PO dosing. In 2003, we changed our standard policy to allow use of oral voriconazole in whole or in part for empiric antifungal coverage for neutropenic patients with uncomplicated persistent fever. Voriconazole was to be loaded at 6 mg/kg IV or PO q 12 hrs for 2 doses and followed by 200 mg PO BID. Over a 32-month period, 27 patients were treated for 31 episodes of persistent neutropenic fever. The cohort included 14 males and 13 females of median age 55 yrs (range, 19–78 yrs). Median weight at the time of first treatment was 84 kg (range, 62–122 kg). Nineteen patients had leukemia, 5 had lymphoma, and 3 had myeloma. Of the 31 episodes, 24 were for induction chemotherapy, 6 for autologous transplantation, and 1 for allogeneic transplantation. Standard prophylactic antibiotics included levofloxacin, fluconazole, valacyclovir and weekly nebulized amphotericin. The patients had been treated with vancomycin and cefepime, or a similar broad spectrum regimen, at the time of fever onset. The median time from fever to start of voriconazole was 4 days. All patients had an ANC<1500. The median WBC prior to voriconazole was 200 (range, 100–1700), and the median duration of ANC<500 prior to voriconazole was 9 days (range, 0–35 days). The median total duration of ANC<500 was 19 days (range, 5–86 days). The chest X-ray was abnormal in 8/30 episodes, CT scan was abnormal in 13/20 episodes, and the galactomannan assay was positive in 1/22 episodes. In 8 (26%) episodes, the patients fulfilled the criteria for possible invasive fungal infection, but none had a probable or definite invasive fungal infection at the start of voriconazole. The loading dose was given PO in 18 episodes and IV in 13 episodes. Median duration of therapy was 11 days (range, 2–54 days). Visual disturbances were reported in 0 (0%) episodes, rash in 2 (6%), creatinine ≥2.0 in 3 (10%), bilirubin ≥2.0 in 5 (16%), AST ≥100 in 2 (6%), and ALT ≥100 in 1 (3%). No patient discontinued drug because of toxicity. Success was described as resolution of fever without the need for a change in antibiotics. The success rate was 55% (95% CI, 36–73%). Fever resolved at a median of 4 days whether or not antibiotics were changed. Changes in the antibiotic regimen after voriconazole varied with the clinical circumstances; these included new antifungal agents in 7/14 and new antibacterial agents in 7/14. Gram positive cocci was the most common organism in this group of patients (10/14 (71%)), and a fungal infection (fusarium) was documented in only one patient. Day-90 survival was 81% (95% CI, 63–93%) for all patients. Neither treatment success nor Day-90 survival differed when compared by whether loading was IV vs PO, or whether the patient had a possible invasive fungal infection at the start of voriconazole. We conclude that oral voriconazole is safe for neutropenic patients with uncomplicated persistent fever, and its efficacy should be evaluated in a randomized trial.

Disclosures: We have assessed use of oral voriconazole for empiric treatment of fungal infections, an area for which the drug does not have FDA approval.

Author notes


Corresponding author