Long-Term Follow-Up Of Invasive Aspergillosis In Allogeneic Stem Cell Transplantation Recipients and Leukemia Patients: Differences In Risk Factors and Outcoumes

Invasive aspergillosis (IA) remains a major concern among patients with hematological malignancies who receive cytotoxic chemotherapy or allogeneic stem cell transplantation (allo-SCT). With a view toward better understanding the differences in terms of clinical significance and outcomes of IA between allo-SCT recipients and patients treated for leukemia, we report a single-center study of 739 unselected consecutive patients treated in our unit between August 2000 and February 2004. IA episodes were classified according to the 2008 EORTC/MSG criteria.

Patients were hospitalized either for allo-SCT (n=135), or for the treatment of acute leukemia (n=378), chronic lymphocytic leukemia (n=116), or myelodysplasic syndrome (n=105). Probable or proven IA were observed in 29 patients, among whom 7 were undergoing allo-SCT (5.2%), 20 were treated for acute leukemia (5.3%), 1 for chronic lymphocytic leukemia (0.8%), and 1 for myelodysplastic syndrome (0.95%). Most of the clinical signs were mild and nonspecific, especially in allo-SCT recipients. Chronologically, thoracic computerized tomography (CT) showed the first signs of IA in 38% of cases, and was positive the same day as at least one biological test (bronchoalveolar lavage [BAL] and serum galactomannan antigen) in 41% of cases. BAL was contributive to the diagnosis in 66% of the 29 cases.

Comparison of patients undergoing allo-SCT to the others revealed significant differences. In allo-SCT recipients, the onset of IA occurred later than in the other patients, after the neutropenic period. The median time between the last hematologic treatment and the diagnosis of IA were 231 days (range, 68-341) in allo-SCT recipients and 17 days (6-57) in the other patients (p<0.001). At the time of AI onset, the median absolute neutrophil count were 1.8 x 10⁹ cells/L (0.7-5) and 0 x 10⁹cells/L (0-2.1) in allo-SCT patients and the others, respectively (p<0.0001). In addition, the median C-reactive protein level was lower in allo-SCT recipients (53 mg/L, 13-146) compared to the other patients (143.5 mg/L, 22.5-285), p=0.023. It is noteworthy that the 7 cases of IA in allo-SCT recipients occurred only in patients treated with corticosteroids for graft-versus-host disease (GVHD) (n=42).

First line therapy was liposomal amphotericin B (n=22), caspofungin (n=8), voriconazole (n=6), or itraconazole (n=1), either alone (n=21) or in combination (n=8). Complete remission of IA on CT was observed in 12 cases, stable lesions in 8 cases and progression in 9 cases. Mortality directly related to IA was 24% in our study and occurred early after the onset of infection. The 100-day, 1-year, 2-year and 10-year overall survival (OS) were 42.9%, 28.6%, 0%, 0% for allo-SCT recipient, respectively, compared to 68.1%, 40.9%, 18.2%, 13.6% for the other patients, respectively (P-values ≥ 0.05). These poor outcomes were mainly attributable to the non-relapse mortality (NRM), especially in allo-SCT recipients, who had a 2-year NRM of 100% compared to 54.6% in the other patients, p=0.034 (2-year cumulative incidence of relapse were 0% and 70.2%, respectively).

In conclusion, our data seem to distinguish 2 types of IA. One has an early onset in neutropenic leukemia patients. The second, observed in allo-SCT recipients, occurs later, after GVHD development and is associated with a very high NRM incidence, suggesting that those patients remain fragile. Thus, systematic broad-spectrum anti-fungal prophylaxis is highly recommended in patients undergoing allo-SCT who develop GVHD.