Invasive fungal infections: epidemiology and analysis of antifungal prescriptions in onco-haematology

P>What is known and Objective: Invasive fungal infections (IFI) are associated with high rates of morbidity and mortality, particularly in onco-haematology patients. We aimed to study the epidemiology of IFI in neutropenic patients and estimate the economic impact of treatment of those infections. Methods: All patients hospitalized in onco-haematology, and treated with antifungal agents, in 2005 were investigated. Four features were studied: the diagnosis for each patient, the antifungal drugs used, the thoracic densitometry reports and the sero-mycological data. Infectious episodes were stratified according to the EORTC 2008 classification criteria (10). Results and Discussion: Of the 1130 patients surveyed, 192 patients received systemic antifungal agents. Of these 46% had acute leukaemia, 29% bone-marrow allografts, 7% lymphoma and 18% other malignant haemopathies. Using the EORTC 2008 criteria (10), there were 8 proved IFI (3 aspergillosis, 3 candidosis and 2 other IFI), 17 probable IFI (11 aspergillosis, 6 candidosis) and 16 possible aspergillosis. The incidence of IFI was 2 center dot 1%. Eighty patients (41 center dot 7%) had received prophylaxis: 56 with fluconazole and 24 with voriconazole. Treatment was most often empirical (n = 127, 66 center dot 1%). Combination of two antifungals was used in 17 cases. The mean duration of prophylactic, empirical, proved/probable aspergillosis-directed, candidaemia-directed and combination treatment was 19, 19, 46, 32 and 27 days, respectively. The cost of antifungal treatment in 2005 reached almost 2 000 000euro, including 427 000euro for documented infections (proved and probable), 1 246 000euro for empirical treatment and 58 300euro for prophylaxis. What is new and Conclusion: The incidence of IFI is low but the pharmacoeconomic impact is extremely high. Improved strategies are required to reduce the frequency and duration of empirical treatment without compromising beneficial outcome.