BackgroundAllogeneic hematopoietic stem cell transplant (allo-HSCT) recipients are exposed to an increased risk of invasive fungal infections (IFIs) due to neutropenia, immunosuppressive treatments, graft-versus-host disease (GvHD) and incomplete immune reconstitution. Although clinical benefit from antifungal prophylaxis has been demonstrated, IFIs remain a leading cause of morbidity and mortality in these patients. In the last decades, attention has also been focused on potential risk factors for IFI to tailor an antifungal prevention strategy based on risk stratification.Aim of the StudyThis retrospective single-center study aimed to assess the epidemiology and the prognostic factors of IFI in a large cohort of allo-HSCT patients.MethodsBetween January 2004 and December 2020, 563 patients with hematological malignancies received an allo-HSCT at the Stem Cell Transplant Unit in Turin: 191 patients (34%) received grafts from a matched sibling donor, 284 (50.5%) from a matched unrelated donor, and 87 (15.5%) from an haploidentical family member. The graft source was peripheral blood in 81.5% of the patients. Our policy for antifungal prophylaxis included fluconazole in matched related and unrelated donors, while micafungin was administered in patients receiving haploidentical transplant. According to this practice, fluconazole was administered in 441 patients (79.6%) and micafungin in 62 (11.2%), while only 9 patients received mold-active prophylaxis. Galactomannan testing was routinely performed twice a week; patients with persisting fever unresponsive to broad spectrum antibiotics were evaluated with lung high-resolution computed tomography (HRCT) scan. In case of imaging suggestive of IFI, bronchoalveolar lavage (BAL) was performed whenever feasible.Statistical AnalysisOnly probable/proven IFI (PP-IFI) occurring during the first 12 months after transplant have been evaluated. IFIs were classified as probable or proven according to the new revised European Organization for Research and Treatment of Cancer (EORTC)/Mycoses Study Group (MSG) consensus criteria. Multivariate competing risk regression, binary logistic, and proportional hazard models were performed to identify risk factors for PP-IFI.ResultsA total of 58 PP-IFIs (n = 47 probable; n = 11 proven) occurred in our patients resulting in a cumulative incidence of 4.1%, 8.1%, and 9.6% at 30, 180, and 365 days, respectively. Molds were the predominant agents (n = 50 Aspergillus; n = 1 Mucor), followed by invasive candidemia (n = 5 non-albicans Candida; n = 1 Candida albicans; n = 1 Trichosporon). Lung was the most frequent site involved in patients with mold infections (47/51, 92.2%). Median time from HSCT to IFI was 98.44 days (0–365 days). Only 34.5% of patients with IFI were neutropenic at the time of infection. The presence of IFI had a significant impact on overall survival at 1 year (IFI, 32.8% vs. non-IFI, 54.6%; p < 0.001). IFI-related mortality rate was 20.7% in the overall population, 17% in patients with probable IFI, and 36% in patients with proven IFI. Multivariate competing risk regression revealed that donor type was the factor significantly associated to the risk of IFI [subdistribution hazard ratio (SDHR), 1.91, IC 1.13–3.20; p = 0.015]. BAL was informative in a consistent number of cases (36/57, 63.2%) leading to the identification of fungal (21), bacterial (4), viral (3), and polymicrobial (8) infections. Overall, 79 patients (14%) received a diagnostic-driven treatment, and 63 patients (11.2%) received a fever-driven treatment. Liposomal amphoteric B was the drug used in the majority of patients receiving diagnostic-driven therapy (30/79, 38%), while caspofungin was administered more frequently in patients who received a fever-driven strategy (27/63, 42.9%).ConclusionAccording to our experience, a non-mold active prophylaxis in patients undergoing allo-HSCT is feasible when combined with an intensive diagnostic work-up including CT scan and BAL. BAL performed at the onset of the disease may provide informative results in most patients. A diagnostic-driven treatment strategy may contribute to limit the use of costly antifungal therapies.