A pilot exploratory study was undertaken to collect preliminary information relating to safety and overall outcome in using intravenous fluconazole (FLUC) for managing antibiotic resistant neutropenic fever (ARNF), with the objective of assessing feasibility of performing a larger prospective controlled study. Patients who were neutropenic from treatment for leukaemia or bone marrow transplantation, received either fluconazole (FLUC) or amphotericin B (AB). Eight of 16 patients (50%) on FLUC and 21 of 25 patients (84%) on AB defervesced; the mean time to defervescence was 11.0 +/- 10.0 days for FLUC compared to 7.7 +/- 6.3 days for AB, and a similar proportion in each treatment group defervesced within 5 days (50% vs. 52%), respectively. Six of 16 patients (37.5%) on FLUC and three of 25 patients (12%) on AB developed overt invasive fungal disease, including pulmonary aspergillosis (FLUC 4 cases, AB 2 cases) and invasive candidiasis (FLUC 2 cases, AB 0 cases). The mean time to these events was 19.5 +/- 13.4 (FLUC) and 9.0 +/- 3.6 (AB) days. The fungal related mortality rates were higher in the FLUC group: five of 16 patients (31%) vs. two of 25 patients (18%) died respectively; the time to fungal death was 43.2 +/- 18.2 (FLUC) and 25.0 +/- 18.4 (AB) days. This tendency towards a more favourable outcome in patients on AB may have been due to absence of prior fluconazole prophylaxis in patients subsequently receiving IV FLUC. Analysis of a small subgroup of patients who had all received prior prophylaxis with clotrimazole only, indicated that a greater number of patients subsequently receiving IV FLUC died from fungal disease (5/16 vs.0/6, P = 0.09).(ABSTRACT TRUNCATED AT 250 WORDS)