Current in vitro and in vivo data indicate that invasive aspergillosis due to Aspergillus terreus is resistant to treatment with amphotericin B. Because little clinical data are available to guide therapy, we performed a retrospective cohort study of cases of invasive A. terreus infections from 1997-2002 to determine whether the use of voriconazole, compared with use of other antifungal therapies, led to an improved patient outcome. We analyzed a total of 83 cases of proven or probable invasive A. terreus infection (47% and 53%, respectively). A total of 66.3% of patients (55 of 83) died during management of IA, with 55.8% mortality (19 of 34 patients) in the voriconazole group and 73.4% mortality (36 of 49) in the group that received therapy with other antifungals. By use of Cox proportional hazards modeling, decreased mortality at 12 weeks was observed in those patients who received voriconazole (hazard ratio, 0.29; 95% CI, 0.15-0.56). Voriconazole is likely to be a better treatment choice for A. terreus infection than is a polyene.