Discussion
In this work, we showed that host factors, particularly neutropenia and hematologic malignancies, are determinant in the survival times of children affected with IFD. The accelerated failure time models estimate the effects of the variables on acceleration or deceleration on the survival times: so, we demonstrated that the children with leukemia, besides having multiplied the hazard for death due to IFD by 2.5, also could have their survival time shortened by a factor of 0.006. Among the several reasons why patients with leukemia are at risk of dying from IFD, one of the most important is neutropenia, which was prolonged in this cohort, and was a determinant factor for shortening survival in our analysis. Many patients with leukemia have severe, prolonged neutropenia caused by the disease itself or chemotherapy, and mucosal lesions that can predispose to infection.6 We analyzed few cases of lymphoma, but this diagnosis was also a factor to shorten survival in patients with IFD.
The overall mortality reported here, around 28%, is identical to that reported by Castagnola et al, in 2006.7 After 14 years, we are using the same therapeutic options, and little progress has been made in the development of new antifungal agents.8 Another point of similarity with the study of Castagnola et al. was the lower mortality in patients with fungemia, compared to other IFD.7 The high mortality observed in pulmonary infections can be explained by several factors: as noted, pulmonary involvement was correlated with the conditions that most favored mortality from IFD, such as the diagnosis of leukemia, cancer recurrences and neutropenia. The most life-threatening form of aspergillus infection, pulmonary angioinvasive aspergillosis, occurs in immunosuppressed or neutropenic patients: the lung shows a necrotizing pneumonia with areas of hemorrhage and acute and granulomatous inflammation.9 Over the last years, there has been a reduction of mortality due to invasive aspergillosis in those patients with myeloid leukemia, in whom remission of malignancy and effective antifungal treatments can be attained, but mortality can grew up to 54% in relapsed/refractory disease, reflecting the compromised immune status of these patients.10 Pulmonary zygomycoses andFusarium spp. pneumonia also have high fatality rates.11,12
The use of empirical antifungal drugs at the time of diagnosis of IFD affected the survival negatively, but the prophylaxis with fluconazole did not. This finding probably means that the children who received empirical treatment were at higher risk of severe infections, or presented a clinical deterioration, requiring the coverage with antifungals. The risk of death was also 3.8-fold higher in patients already receiving antifungals at the time of diagnosis of IFD in the study of Castagnola et al., but these authors analyzed prophylaxis and empiric use as a single variable.7
Candidemia is now a frequent cause of bloodstream infection, due to the widespread use of central venous catheters, parenteral nutrition, broad-spectrum antibiotics, and immunosuppressive therapy. Rare opportunistic yeasts are emerging as cause of IFD.13We observed 16 episodes of fungemia caused by these rare yeasts, but mortality was similar to observed in candidemia.
Allogeneic unrelated HSCT or mismatched donor HSCT put the patient in the highest risk group for IFD.14 Compared with allogeneic HSCT, autologous HSCT is associated with a lower risk of invasive aspergillosis because the period of pre-engraftment neutropenia is shorter and the possibility of graft-versus-host disease is absent.15 Although the mortality of our HSCT patients with IFD was still high (25.8%), and the fact that most were allogeneic, the transplantation was not a factor for shortening survival. One explanation is that HSCT is the best chance of cure for diseases like myeloid leukemia,16 and most of the patients will recover from neutropenia.
The most obvious weakness of this study, common to several others of its kind, is the absence of sensitivity tests to antifungals, not routinely performed in our institution. The results are also limited to patients treated in a single center, and may not reflect practices from other parts of our country or the world.