Results
We analyzed 152 IFD (133 proven and 19 probable). All probable IFD were
pulmonary aspergillosis, with high-resolution computerized tomography
scans presenting dense, well-circumscribed lesions with a halo sign, air
crescent sign, cavities or segmental or lobar consolidation, with
galactomannan antigen detected in plasma. Table 1 depicts the
demographics and other characteristics of patients. There were 133
fungal isolates in 133 patients (table 2).
[Table 1]
[Table 2]
There were 43 deaths attributed to IFD (28.3%), and survival
probabilities were lower for pulmonary IFD (46.9%, log rank test p =
0.0017), for the diagnosis of leukemia (62.5%, p = 0.004), and
neutropenia <500 cells/mm3 (55.4%, p < 0.0001). For
Candida fungemia, survival probabilities were higher (76.6%, p =
0.043). Figure 1 depicts the Kaplan Meier survival curves for these
conditions. Table 3 shows the Weibull hazard ratios and accelerated
failure time models for the event “death caused by IFD”.
[Figure 1]
[Table 3]
Most of the cases of leukemia (51/56, 91%) were in children older than
4 years; about 72% of the patients were older than 4 years, but out of
the 43 deaths, 38 (88%) were in this age group. Non-Candida and
non-Cryptococcus yeasts caused fungemia in 16 patients, with
three deaths (18.7%). In four patients, there were pulmonary lesions
accompanying Candida fungemia, and Cryptococcus was
isolated in lung biopsy in one case; all other cases of fungal pneumonia
were associated with filamentous fungi. Pulmonary involvement could be
correlated in chi-square tests with the diagnosis of leukemia (18/56
cases, 32%, p = 0.01), relapse of disease (17/41 cases, 41%, p =
0.0002), and neutropenia (23/74 cases, 31%, p = 0.0002).
In the HSCT patients, there were eight deaths (25.8%). Fungemia was
diagnosed in 16 patients (51.6%) and pulmonary fungal infections in 12
(38.7%).