Invasive Aspergillosis (IA) in Acute Myeloid Leukemia (AML): When Is Prophylaxis Appropriate?

Introduction: management strategies for IA in patients receiving intensive chemotherapy (CT) for AML remain a matter of debate. Although most centers adopted primary prophylaxis recommendations based on few trials with less-than-ideal comparators, it appears that the option between a preemptive/early diagnosis-driven strategy and a prophylactic one should take in account the local incidence of IA, as the number needed to treat differs considerably between institutions. Until 2012 our center followed a preemptive strategy, but the finding of an alarming increase in IA cases led us to adopt primary prophylaxis. Results were prospectively evaluated in the following years and are reported here.

Methods: from January/06 to March/12 308 patients received 629 first-line or subsequent CT courses (377 induction, 252 consolidation) in which only secondary antifungal prophylaxis was allowed. From April/12 to May/16 197 patients received 378 courses (230 induction, 148 consolidation) under prophylactic voriconazole. In both periods patients were nursed in rooms with positive pressure and HEPA filters, and CT regimens were unchanged. IA cases were classified according to EORTC/MSG criteria. Incidence of IA, overall mortality and attributable (specific) mortality at 12 weeks were compared between the 2 periods.

Results: in the first period 82 cases of IA (79 lung, 3 sinus) were diagnosed, 6 of which were proven, 43 probable and 33 possible. Half occured in induction, 33% in consolidation, and the remaining in second-line CT courses. Most cases were treated with voriconazole (only 5 received other antifungals, in a clinical trial setting) with a global success rate of 82%. Two of the 11 failures occured in complete remission (consolidation courses). Overall mortality at 12 weeks was 23% and attributable mortality was 12%. In this period 2 cases of rhinocerebral Mucor were diagnosed. In the second period 13 cases of IA (12 lung, 1 gut) occured, of which 1 was proven, 7 probable e 5 possible. Most (77%) arose during induction. Serious (grade 3/4) adverse effects of voriconazole needing interruption were infrequent and consisted mainly of hepatic enzymes increases in 5% of the patients. Success rate (diverse antifungals) was 85%. Overall mortality at 12 weeks was 23% with no attributable mortality. Incidence of IA (27% vs 6.6%) and attributable mortality were significantly different between the 2 periods (both p < 0.001).

Conclusion: preemptive/early diagnosis-driven strategy is atractive, as it spares most patients the side effects of useless antifungals. However, in centers with a high incidence of IA like ours, prophylaxis becomes highly rewarding, leading to a very significant reduction in these life-threatening fungal infections.