Incidence of Invasive Fungal Infections in Acute Myeloid Leukemia Without Antifungal Prophylaxis

Antifungal prophylaxis during induction for acute myeloid leukemia (AML) varies according to local rates of invasive fungal infections (IFIs). We evaluated fluconazole prophylaxis and no antifungal prophylaxis, as a natural interrupted time-series study to assess survival and infection complications. We identified patients with AML ≥ 18 years old undergoing induction chemotherapy during 2 time periods: period 1, fluconazole prophylaxis from August 1, 2013 to September 30, 2015, and period 2, no prophylaxis from October 1, 2015 to December 31, 2017. The primary outcome was incidence of proven or probable IFI. Secondary outcomes included types of IFIs and 60-day overall survival (OS). IFI was defined by the 2002 European Organization for Research and Treatment of Cancer/Mycoses Study Group Consensus criteria. One hundred forty-four patients received induction chemotherapy over the 2 time periods. In the prophylaxis versus no-prophylaxis groups, the rate of proven or probable IFIs was 4 (5%) of 87 versus 12 (21%) of 57 (P = .01). The total number of proven IFIs was 3 (3%) of 87 versus 4 (7%) of 57 (P = .44), whereas probable IFIs were 1 (1%) of 87 versus 8 (14%) of 57 (P < .01). No difference was observed in fungemia. Incidence of IFIs was too low to detect resistance patterns. OS at 60 days was improved in with fluconazole prophylaxis compared with no prophylaxis (hazard ratio, 0.329; 95% confidence interval, 0.12-0.89; P = .028). Observed rates of proven or probable IFI were lower in the fluconazole prophylaxis group versus the no-prophylaxis group. Sixty-day OS was higher with fluconazole prophylaxis. Further study is required to evaluate how fluconazole may impart the differences in survival seen in this analysis. The role of antifungal prophylaxis in acute myeloid leukemia continues to be controversial, with national guidelines recommending prophylaxis based on incidence of invasive fungal infections (IFIs) to each region. We performed a retrospective, interrupted time-series study based on periods where fluconazole was used or not used as prophylaxis during induction for acute myeloid leukemia. There was a decrease in IFIs with fluconazole; however, this was largely driven by probable Aspergillus infection seen in our non-prophylaxis cohort. Our results suggest further investigation of azoles is warranted with consideration of including risk adapted strategies according to the rate of IFIs.