Antifungal susceptibility of invasive yeast isolates in Italy: the GISIA3 study in critically ill patients

BACKGROUND Yeasts are a common cause of invasive fungal infections in critically ill patients. Antifungal susceptibility testing results of clinically significant fungal strains are of interest to physicians, enabling them to adopt appropriate strategies for empiric and prophylactic therapies. We investigated the antifungal susceptibility of yeasts isolated over a 2-year period from hospitalised patients with invasive yeast infections. METHODS 638 yeasts were isolated from the blood, central venous catheters and sterile fluids of 578 patients on general and surgical intensive care units and surgical wards. Etest strips and Sensititre panels were used to test the susceptibility of the isolates to amphotericin B, anidulafungin, caspofungin, fluconazole, itraconazole, posaconazole and voriconazole in 13 laboratories centres (LC) and two co-ordinating centres (CC). The Clinical and Laboratory Standards Institute (CLSI) reference broth microdilution method was used at the CCs for comparison. RESULTS Etest and Sensititre (LC/CC) MIC90 values were, respectively: amphotericin B 0.5/0.38, 1/1 mg/L; anidulafungin 2/1.5 and 1/1 mg/L; caspofungin 1/0.75 and 0.5/0.5 mg/L; fluconazole 12/8 and 16/16 mg/L; itraconazole 1/1.5, 0.5/0.5 mg/L; posaconazole 0.5 mg/L and voriconazole 0.25 mg/L for all. The overall MIC90 values were influenced by the reduced susceptibility of Candida parapsilosis isolates to echinocandins and a reduced or lack of susceptibility of Candida glabrata and Candida krusei to azoles, in particular fluconazole and itraconazole. Comparison of the LC and CC results showed good Essential Agreement (90.3% for Etest and 92.9% for Sensititre), and even higher Categorical Agreement (93.9% for Etest and 96% for Sensititre); differences were observed according to the species, method, and antifungal drug. No cross-resistance between echinocandins and triazoles was detected. CONCLUSIONS Our data confirm the different antifungal susceptibility patterns among species, and highlight the need to perform antifungal susceptibility testing of clinically relevant yeasts. With the exception of a few species (e.g. C. glabrata for azoles and C. parapsilosis for echinocandins), the findings of our study suggest that two of the most widely used commercial methods (Etest and Sensititre) provide valid and reproducible results.