Invasive fungal sinusitis in patients with hematological malignancy: 15 years experience in a single university hospital in Taiwan

National Taiwan University Hospital is a 2200-bed teaching hospital that provides both primary and tertiary care in metropolitan Taipei. More than 600 admissions annually comprise adult patients with hematological malignancies for in-patient chemotherapy. Lymphoma, acute myeloid leukemia and myeloma were the three most common hematological diagnoses [13].

The search terms "sinusitis" and "sinonasal infection" were used to identify reports in the discharge summary database of all patients admitted to the Department of Internal Medicine between January 1995 and December 2009. Patients with hematopoietic stem cell transplants, hematological disease (including severe aplastic anemia, pure red cell aplasia, and hematological malignancy), and sinusitis diagnosed during their hospital stay were enrolled in this study. Patients with hematological disorders and chronic non-invasive sinusitis admitted in the same period of time cohort were enrolled as controls. Demographic features, hematological disease status, underlying medical diseases, chemotherapy regimen before diagnosis of fungal sinusitis, evidence (clinical, microbiological, pathological, and radiological) showing fungal sinusitis, antifungal therapy, and outcome were collected and analyzed retrospectively. This research conformed to the Helsinki Declaration and local legislation, and approved by the local ethics committee.

During the study period, patients with hematological malignancy developed sinusitis were routinely to take sinus X-ray and otolaryngist was consulted for focal evaluation and tissue culture. CT and MRI sinus study, surgical biopsy and debridement were performed according to clinicians' decision.

Physicians clinically diagnosed IFS in all these patients according to EORTC/MSG-NIAID consensus criteria in 2008 [5, 6]. The host factors included prolonged neutropenia (< 500 neutrophils/mm³ for > 10 days) temporally related to the onset of fungal disease, receipt of an allogeneic stem cell transplant, prolonged use of corticosteroids, immunosuppressive agents, or nucleoside analogues during the past 90 days. The clinical criteria included imaging showing sinusitis plus at least one of the following three signs: acute localized pain, nasal ulcer with black eschar, extension from the paranasal sinus across the bony barrier. The microbiological criteria included culture or isolation of fungus from surgical material or sinus aspirate samples, and detection of Aspergillus galactomannan antigens in serum. In our institution, an optical density (OD) index of 0.5 or higher is considered positive for aspergillus galactomannan test.

Proven, probable, and possible IFS were defined mainly according to the EORTC/MSG-NIAID criteria [5, 6]. Proven IFS was defined by the presence of fungi associated tissue damage on histopathologic examination of a biopsy specimen; or positive culture result, consistent with infection, from a sample obtained aseptically from a clinically or radiologically abnormal site. Probable IFS was defined by the presence of at least one host factor criterion, one microbiological criterion, and one clinical criterion. Possible IFS was defined by the presence of at least one host factor criterion and one clinical criterion.

The survival is defined from diagnosis of invasive fungal sinusitis to patients died. Attributable mortality to invasive fungal sinusitis is defined if patient died within 6 weeks, and there were no other attributable cause of death.

Comparisons were made using the chi-square test. Predictors of survival were identified using a Cox regression model. Factors which p level is below 0.2 in the univariate analysis were selected for multivariate Cox regression analysis. All statistical analyses were performed using SPSS 13.0 for Windows (SPSS, Chicago, IL, USA). A p value less than 0.05 was considered statistically significant.

The diagnosis of invasive fungal sinusitis was made in 46 patients with hematological disorders, and 64 patients with hematological disorders and chronic non-invasive sinusitis admitted in the same period of time cohort were enrolled as controls. There was no outbreak of invasive fungal sinusitis detected in this study. The overall incidence of IFS is 1.77 per 1000 patients. The incidence and trend of IFS in adult hematological patients was shown in Table 1. There were trends of decreasing incidence and mortality during the study period, although these did not attain statistically significance (p = 0.199 and 0.223).

IFS was proven in 25 patients (54%), probable in 5 patients (11%), and possible in 16 patients (35%). Their clinical characteristics are shown in Table 2. Compared with chronic non-specific sinusitis, IFS was more common in patients with disease subtype of acute myeloid leukemia (AML) than in patients with other types of hematological malignancy (p < 0.001) and more common in patients with absolute neutrophil count less than 500/mm³ for ≥10 days. The occurrence of IFS was not affected by underlying diabetes mellitus, hematological disease status, chemotherapy intensity, and allogeneic stem cell transplantation.

Among the 25 patients with proven invasive fungal sinusitis, fifteen patients had positive biopsy showed hyphae with Giemsa stain or PAS stain, and 14 patients showed positive fungal cultures. The fungal isolates from 25 patients with proven invasive fungal sinusitis included: Aspergillus flavus (12), Aspergillus fumigatus (2), Aspergillus vesicolor (1), Aspergillus sybowii (1), unidentified Aspergillus species (1), Mucor (4), Phaeohyphomyces (1), Fusarium (1), Penicillium (1), and unidentified fungal species (1).

Sixteen patients during 2005 to 2009 had serial follow-up of Aspergillus galactomannan antigen. The mean value (+/- SD) of Aspergillus galactomannan antigen test is 2.32 (+/- 1.86). The mean value (standard deviation) of Aspergillus galactomannan antigen in the seven patients with positive results was 2.32 (1.86), compared with those with negative results 0.13 (0.07). The p value (0.048) is below 0.05. Aspergillus galactomannan antigen was not elevated in three patients with mucormycosis sinusitis. Aspergillus galactomannan antigen was elevated in seven of eleven patients with aspergillus sinusitis. One patient with phaeohyphomycosis-related sinusitis and one patient with sinusitis caused by Fusarium and Penicillium marneffei also showed elevated aspergillus galactomannan antigen. The sensitivity of serum galactomannan antigen testing was 64% and the specificity was 60% for the detection of invasive aspergillus sinusitis. (Table 3)

Sinus imaging by computed tomography (CT) or magnetic resonance (MRI) were performed in 44 (96%) of patients with IFS and 37 (59%) of 63 patients with chronic non-specific sinusitis. Evidence of extra-sinus tissue involvement and/or bony destruction were detected in 15 (33%) of patients with IFS. In addition, brain abscess complicating IFS was found in two patients and cavernous sinus thrombosis in one patient.

Pre-emptive antifungal treatment was given in 43 (94%) of 46 patients with IFS. Fourteen of sixty-four patients with non-IFS received pre-emptive antifungal therapy initially, and these agents were discontinued after excluding the diagnosis of IFS. The overall mortality rate of patients with IFS was 41% (Table 2). The attributable mortality of patients with AML was also much higher than non-AML patients. All the patients died rapidly due to IFS (median survival, 15 days; 25^th and 75^th percentile, 10 and 34 days, respectively). Six-week survival was not significantly different between patients who had received antifungal combination therapy (amphotericin B plus capsofungin and amphotericin B plus voriconazole vs amphotericin B) compared to those who had received monotherapy (p = 0.74). All four patients with mucormycoses had received high dose antifungal therapy and aggressive surgical debridement, but three died within 6 weeks. Twelve patients with proven invasive fungal sinusitis and two patients with possible invasive fungal sinusitis underwent functional endoscopic sinus surgery for debridement. The median interval between diagnosis and surgical debridement is 11.5 days (range 1-38 days). Nineteen of 46 patients (41%) with IFS and 8 of 64 patients (13%) with non-invasive sinusitis died within 6 weeks after diagnosis of sinusitis. The 6-week mortality rate was significantly higher in patients with IFS than those with non-invasive sinusitis (p = 0.001). Univariate analysis of 6-week outcomes in patients with hematological disorders and invasive fungal sinusitis is summarized in Table 4. The acute myeloid leukemia disease subtype (p = 0.044) and refractory disease status were associated with worse outcomes (p = 0.05). The intensity of chemotherapy and allogeneic stem cell transplantation were not independent prognostic factors of invasive fungal infection in this study. The median interval between neutropenia and diagnosis of IFS was not significantly different between non-survivors and survivors (median 19 days vs 21 days, p = 0.677). Multivariate Cox regression analysis revealed functional endoscopic sinus surgery for debridement as the only independent prognostic factor for survival (p = 0.047).