Spectrophotometric detection of azole-resistant Aspergillus fumigatus with the EUCAST broth microdilution method: is it time for automated MIC reading of EUCAST antifungal susceptibility testing of Aspergillus species?

OBJECTIVES: Current reference susceptibility testing methods of Aspergillus require visual reading, which is subjective and necessitates experienced staff. We compared spectrophotometric and visual MIC reading of EUCAST E.Def 9.3.2 susceptibility testing of Aspergillus fumigatus for a large collection of isolates with different azole resistance mechanisms. METHODS: A. fumigatus (n = 200) were examined, including 62 WT and 138 non-WT with the following alterations: TR34/L98H (n = 57), TR46/Y121F/T289A (n = 54) or single point mutations (n = 27). EUCAST E.Def 9.3.2 susceptibility testing was performed for amphotericin B, itraconazole, voriconazole, posaconazole and isavuconazole. MICs were determined after 48 h of incubation visually and spectrophotometrically, as the lowest concentration corresponding to a 1%, 3%, 5%, 10% or 15% OD increase above the background OD. The best spectrophotometric endpoint (SPE) was identified based on the highest essential agreement (EA; ±1 two-fold dilution) and categorical agreement (CA) and fewer very major errors (VMEs) and major errors (MEs). RESULTS: Τhe best SPEs were 5% and 10% for all drugs. The best agreement between visual and spectrophotometric MICs was found with the 10% growth endpoint, which resulted in identical median MICs with 90% of differences being ≤1 two-fold and higher EA (91%-100%) and CA (100%) and no VMEs and MEs compared with the 5% endpoint (77%-100%, 96%-98%, 0% and 0%-4%, respectively). CONCLUSIONS: Spectrophotometric MIC reading can be used for A. fumigatus susceptibility testing and for detecting azole resistance. A visual inspection of the plate should be performed to confirm equal inoculation, absence of well contamination and proper growth, and to identify potential uncommon phenotypes or subpopulations.