Given the increased choice of therapeutic agents and the rising incidence of serious invasive disease, it is important that reliable in-vitro methods for detecting antifungal drug resistance in Aspergillus spp. are developed. Six clinical isolates of Aspergillus fumigatus, obtained from patients in whom the clinical outcome was known, were selected for study. Each was used to examine a range of parameters affecting agar dilution and broth microdilution susceptibility test results. The in-vitro results were compared with outcome in a neutropenic mouse model of invasive aspergillosis. Groups of animals were treated with itraconazole at 25 mg/kg and 75 mg/kg and survival rates and organ burdens were determined. Itraconazole was efficacious against four isolates (susceptible) but failed for two (resistant) in the animal model of infection. Both the resistant isolates had been obtained from patients receiving itraconazole treatment with good serum concentrations of the drug. Conditions for the agar dilution test which produced results that correlated best with our in-vivo observations included the use of RPMI agar with L-glutamine buffered to pH 7 with MOPS, inoculated with 10(6)-10(7) conidia/mL and incubated for 48-72 h at 28 or 35 degrees C with a no-growth endpoint. Optimal conditions for the broth microdilution method included the use of RPMI medium with L-glutamine and 2% glucose buffered to pH 7 with MOPS, an inoculum of 2 x 10(5) conidia in 200 microL incubated for 48 h at 35 degrees C with a growth (or trace) endpoint. The MICs for the susceptible isolates were 0.12-1.0 mg/L and > or = 16 mg/L for the resistant isolates. With careful selection and standardization of test conditions it is possible to generate reproducible in-vitro susceptibility data for Aspergillus spp. that will predict clinical outcome.