Infection with Scedosporium apiospermum and S. prolificans, Australia

S. prolificans has become a major pathogen in immunocompromised patients.

In the fi nal months of 1999 during construction work, an increased number of Scedosporium spp. isolates and S. prolifi cans were noted at Alfred Hospital, a university hospital in Prahran, Victoria, Australia, that provides statewide trauma, burn, cystic fi brosis, heart and lung transplant, and HIV services. The records of all patients from whom Scedosprium spp. were isolated from June 30, 1997, through December 31, 2003, were reviewed to describe the epidemiology, clinical features, and outcomes of these infections.

Data Collection
Records of all patients with Scedosporium spp. cultured between June 30, 1997, andDecember 31, 2003, were reviewed retrospectively, and demographics, primary illness, antifungal therapy, and presence of other pathogens, and outcomes were recorded. Immunocompromised patients were defi ned as those with impairment of either or both natural and specifi c immunity to infection (22). Those with localized impaired host defenses caused by underlying diseases such as cystic fi brosis were classifi ed as immunocompetent. Invasive infection was defi ned as a tissue biopsy specimen with hyphae plus culture of the organism (23). Disseminated infection was defi ned as positive blood cultures or infection at >2 noncontiguous sites within 7 days. Neutropenia was defi ned as an absolute neutrophil count <0.5 cells/mm 3 . Descriptive statistics and odds ratios were calculated by using STATA release 8.0 (Stata Corporation, College Station, TX, USA.)

Laboratory Methods and Epidemiology
Scedosporium spp. were identifi ed after culture onto horse blood agar and Sabaroud dextrose agar (SDA) and incubation at 30°C and 37°C. Colonies were rapid-growing, gray-white, and downy, with a gray-black reverse side. Species were differentiated by microscopic morphology (24). Susceptibility testing was performed according to Clinical and Laboratory Standards Institute methods (25) and synergy studies (2-dimensional 2-agent microdilution checkerboard method) (26). Synergy was a summation of the fractional inhibitory concentration <0.500.
For comparison of temporal patterns of mold isolates, the number of persons with Scedosporium spp. during the 3 years subsequent to this review (2004)(2005)(2006) and Aspergillus spp. for 1999-2005 were extracted from laboratory databases. Rates of detection were expressed as per 1,000 separations and per 100,000 inpatient-patient days.

Environmental Sampling
From March 2001 through July 2002, environmental samples were collected from hospital public access areas within and adjacent to the construction site on 7 separate occasions. Additional sampling was conducted in ward corridors, nursing stations, patient rooms, and the patient car park. Air sampling was conducted with a portable highvolume air sampler (MAS 100; Merck, Darmstadt, Germany). The volume of air sampled at each site was 1,000 L/10 min (1 m 3 ). Settle plates were placed in ward corridors and patient rooms for 60 min of exposure. Dust was collected from horizontal surfaces with sterile swabs moistened with sterile saline. Dust and soil specimens were directly placed on standard SDA and selective SDA containing amphotericin B and incubated at 27°C.

Results
From June 1997 through December 2003, a total of 59 isolates of Scedosporium spp. were cultured from 56 patients. S. apiospermum was isolated from 31 patients, and S. prolifi cans was isolated from 28 patients. Both species were isolated from 3 patients at separate times (12 days, 13 months, and 27 months apart, respectively). Demographic information, coinfections, and outcomes of the patients are shown according to the species isolated (Table 1). During the period of this review, an average of 28 allogeneic stem cell, 35 lung, 27 heart, and 15 renal transplants were performed each year.

S. apiospermum
S. apiospermum was isolated from 32 specimens collected from 31 patients. Twenty-nine isolates were from the respiratory tract; sputum (n = 22), bronchoalveolar lavage (BAL) (n = 5), sinus (n = 1), and lung tissue (n = 1). The remaining 3 isolates were from brain tissue, a central ve-nous catheter tip, and an ear swab, respectively. No blood cultures were positive for S. apiospermum. S. apiospermum was isolated concurrently with Aspergillus spp. from the respiratory tracts of 9 (29%) of the 31 patients.
Three patients were lost to follow-up after isolation of S. apiospermum. The remaining 28 patients were followed up for a total of 981 months. The mean duration of followup was 35 months (median 16 months, range 1-96 months). Two (6%) of the 31 patients had invasive infections. The fi rst patient was a woman who received a lung transplant 5 years earlier and was previously colonized with Aspergillus spp. but not S. apiospermum. She had pulmonary and cerebral lesions and was treated with amphotericin B deoxycholate and itraconazole for presumed aspergillosis. S. apiospermum was isolated from postmortem lung and brain tissue. The second patient was a man who had advanced HIV infection with fungal sinusitis on biopsy. He was treated with voriconazole and surgery but died 3 months later from infection with cytomegalovirus. There were 6 deaths within 1 month of isolation of S. apiospermum. All deaths occurred in immunocompromised patients but only 1 was directly attributable to S. apiospermum. The other deaths resulted from the underlying condition.
Four patients with chronic lung disease were receiving itraconazole for treatment of Aspergillus spp. infection at the time S. apiospermum was isolated. Three of these patients died of respiratory failure 7, 9, and 16 months, respectively, after isolation of S. apiospermum, and 1 was alive when last seen 96 months after fungal isolation. Seven patients received antifungal therapy after isolation of S. apiospermum from respiratory tract samples (4 received voriconazole and 3 received itraconazole). These 7 patients had HIV infection and sinusitis (n = 2), had undergone lung transplantation (n = 3) or had bronchiectasis or cystic fi brosis (n = 2). Both patients with HIV infection had sinusitis and died within 7 months of complications of HIV. Of the other 5 patients, 4 who received azole therapy remained well without invasive infection at 32, 41, 48, and 88 months, respectively, of follow-up. The remaining treated patient died 15 months later; death was not attributed to fungal infection. The median duration of follow-up of those treated with azoles at the time of fungal isolation or subsequent to isolation was 16 months (range 3-96 months); S. apiospermum was not subsequently detected in these patients. The median duration of follow-up for patients receiving no treatment after fungal isolation was 19 months (range 1-84 months); S. apiospermum was isolated from 4 of these patients 1, 18, 30, and 36 months, respectively, after initial fungal isolation.

S. prolifi cans
S. prolifi cans was isolated from 46 specimens obtained from 28 patients. Fourteen (50%) of the 28 patients were immunocompetent. Most (12/14, [86%]) specimens from immunocompetent patients were from the respiratory tract; cystic fi brosis (n = 6), chronic airway disease (n = 3), nasal discharge and sinus aspirates with chronic sinusitis (n = 3). Patients with cystic fi brosis or airway disease were not considered to have invasive disease and none received antifungal therapy. All 3 patients with chronic sinusitis were treated with surgery, and 2 received itraconazole. One isolate was from knee cartilage of a patient with hemophilia, osteoarthritis, and a knee replacement; the patient underwent surgery and received itraconazole. S. prolifi cans was also isolated from a skin-biopsy specimen from a patient with multiple-trauma and cellulitis who was treated only with surgery. There were no cases of disseminated infection or deaths resulting from S. prolifi cans in these immunocompetent patients.
S. prolifi cans was isolated from 14 immunocompromised patients. Immunodefi ciency was caused by lung transplantation (n = 6), hematopoietic stem cell transplantation (HSCT) (n = 6), acute myeloid leukemia (AML), and myelodysplastic syndrome (MDS). S. prolifi cans was isolated from BAL of 6 lung transplant recipients. One patient was receiving itraconazole for Aspergillus spp. infection, 3 were receiving voriconazole, and 2 were receiving itraconazole and voriconazole with terbinafi ne. Invasive disease did not develop in any of the 4 lung transplant recipients who received antifungal treatment and none died. Two lung transplant recipients did not receive antifungal treatment. One of these patients was lost to follow-up and 1 survived >25 months without developing invasive infection, although S. prolifi cans was isolated again 1 year later. S. prolifi cans was isolated from August 2000 through September 2002 from 6 patients undergoing allogeneic HSCT and from 2 patients with hematologic malignancy. Four of the 6 HSCT recipients had positive blood cultures, and 4 recipients had skin lesions (multiple, nodular). Findings for 8 patients in whom invasive disease developed are summarized in Table 2.
Patients 1, 2, and 6 had HSCT complicated by chronic extensive refractory graft versus host disease (GVHD). All 3 died of invasive infection despite treatment with itraconazole or voriconazole and terbinafi ne; 2 had positive blood cultures. Infections developed in patients 3, 4, and 5 during neutropenia after HSCT. Two of them had positive blood cultures, and both died. Patient 4 received no antifungal treatment, and patient 5 died despite replacement of prophylactic itraconazole with empiric amphotericin B. Patient 3, described elsewhere (8), survived after treatment with voriconazole, terbinafi ne, surgery, and neutrophil recovery. Patient 7 had MDS and preceding idiopathic CD4-cell lymphocytopenia. S. prolifi cans was isolated from sputum, but not BAL, and this patient was not treated. One month later, S. prolifi cans sinusitis was diagnosed. Despite surgery and treatment with itraconazole, followed by voriconazole and terbinafi ne, the patient died. Patient 8, a woman, was neutropenic after remission-induction chemotherapy for AML. A swab from a Hickman catheter exit site was positive for S. prolifi cans. Subsequently, chest wall cellulitis, deep soft tissue infection, and multiple skin nodules developed. She was treated with surgery, voriconazole, and terbinafi ne. Her neutropenia resolved and she recovered.

Discussion
This study describes the epidemiology of Scedosporium infection in a cohort of contemporaneous patients at a university hospital. The single-center approach, with cases identifi ed by laboratory detection, allows capture of the full spectrum of infection from colonization to invasive infection. Scedosporium spp. were detected in a broad range of patients and clinical settings. However, there were distinct differences in epidemiology, clinical manifestations, antifungal susceptibility patterns, and outcomes between S. prolifi cans and S. apiospermum.
Previous reviews have focused on invasive cases reported (10,27,28), but this approach is limited by both selection and publication bias (29) and does not describe the natural history of infection or the prevalence of these infections. Our series enables the annual incidence of cases, ratio of colonized to infected patients, and natural history of colonization to be determined for each species. We showed that invasive infections accounted for only 6% of S. apiospermum isolates and for 46% of S. prolifi cans isolates, and that isolation of Aspergillus spp. was 20-30 times more frequent than that of Scedosporium spp.
S. apiospermum and S. prolifi cans are colonizers of abnormal airways caused by bronchiectasis, cystic fi brosis, chronic obstructive pulmonary disease, or lung transplantation (1,2,6,18,30,31). S. apiospermum was an airway colonizer in <10% of patients with cystic fi brosis in France and Australia (1,2), and similar proportions of lung transplant recipients were colonized with 1 or both species at 1 center in Australia (31). However, there are few reports that S. prolifi cans colonized patients with lung disease (6,32). In the present study, S. prolifi cans was identifi ed as an airway colonizer only after December 1999. The onset of invasive infections in HSCT and neutropenic patients occurred in August 2000. S. prolifi cans was fi rst detected at the M.D. Anderson Cancer Center in Houston, Texas, after 2001 (19). Emergence of S. prolifi cans may be related to an environmental source that was not identifi ed or selection pressure caused by changes in antifungal prophylaxis practices (33). This fi nding did not appear to be the explanation in our patients because fl uconazole remains standard prophylaxis for neutropenia, and itraconazole was used in only 1 of 8 patients. Other possible explanations include use of more aggressive chemotherapy regimens in patients with acute leukemia and the advent of nonmyeloablative allografts, which change characteristics of patients selected for transplants. The reason for persistence of this organism over several years after its initial appearance is also unclear but has also been observed by others (19). In patients with respiratory tract disease, both Scedosporium spp. were isolated in comparable numbers. However, the only invasive infection in this diverse group was disseminated S. apiospermum 5 years after lung transplantion. The outcome of 19 patients who had undergone lung or heart lung transplantation and were colonized with this fungus was comparable to the 17 patients in the immunocompetent group with airways disease caused by cystic fi brosis, bronchiectasis, or chronic obstructive pulmonary disease. As in other reports, other opportunistic pathogens, especially Aspergillus spp. (present in one third), were commonly cultured simultaneously (1,18,30,31). In lung transplant recipients, Scedosporium spp. in BAL was associated with advanced bronchiolitis obliterans and airway stenosis (31), which emphasizes the diffi culty of interpreting the role of Scedosporium spp. in this group. Whether colonization follows airway damage, immunosuppression, or antifungal therapy to treat Aspergillus spp. colonization or is itself an ongoing cause of airway damage such as bronchiolitis obliterans is unknown. In our study group, most patients with respiratory tract colonization had not received antifungal therapy. Although antifungal treatment was used in ≈50% of patients with S. apiospermum airway colonization, there appeared to be no survival advantage in patients treated; our study was not a randomized comparison. With S. prolifi cans respiratory tract colonization, too few patients were treated with antifungal drugs for valid conclusions to be made. There was no reduction in survival in 5 untreated patients with isolates of either S. apiospermum or S. prolifi cans. The role of Scedosporium spp. respiratory tract colonization requires further evaluation.
Allogeneic HSCT and AML/MDS were the only settings in which S. prolifi cans was more common than S. apiospermum. There were no S. apiospermum infections in these patients, although in at 1 US cancer center, S. apiospermum infections were more common (19). This may refl ect different geographic distributions of Scedosporium spp. In our study, S. prolifi cans caused illness with high mortality rates in HSCT and leukemia patients. In HSCT recipients, S. prolifi cans infections occurred equally in those with neutropenia (early after transplant) and GVHD (later after transplant). Thus, host factors unique to AML and HSCT, such as neutropenia, GVHD, associated macrophage dysfunction, and environmental exposure, may play a role in the propensity for invasive infection with S. prolifi cans in this group.
As in previous reports (10,32), neutropenic patients had sepsis, positive blood cultures, and overwhelming infection (often associated with disseminated rash). In nonneutropenic patients, infection was more likely to be localized initially in lungs, joints, or sinuses. The ability of S. prolifi cans to grow in blood cultures in AML and HSCT patients is well recognized (10,34), and this diagnosis, or infection with Fusarium spp., should be considered when a mold is cultured from blood. In our study, S. prolifi cans was the only species to grow in blood culture, although others have reported S. apiospermum, albeit, less frequently (10,18,19). Although colonization with S. prolifi cans without progression to disease was observed in patients with respiratory disease and after lung transplantation, this was not observed in 3 patients with AML/MDS or HSCT. A nonsterile site swab or sputum culture yielding S. prolificans was soon followed by a diagnosis of invasive infection, which indicated that in these patients a culture result, even from a nonsterile site, should be viewed seriously.
The clinical fi ndings of S. prolifi cans in our patients suggest that this fungus is more invasive than S. apiospermum, as has been found in mice (35). In addition to disseminated infection in AML patients and HSCT recipients, S. prolifi cans also caused locally invasive disease in patients who were not immunocompromised, such as those with posttrauma cellulitis, septic arthritis in a damaged joint, and sinus disease. One proposed mechanism for virulence of S. prolifi cans is melanin in the cell wall (27).
Antifungal therapy is problematic with S. prolifi cans with high MICs for amphotericin B, echinocandins, and azoles. This fi nding has stimulated interest in combinations of voriconazole and terbinafi ne (26), voriconazole and echinocandin (36), or polyene and echinocandin (36). However, these infections are rare and experience is limited to case reports (8,9,36). In vitro treatment with interferon-γ and granulocyte-macrophage-colony-stimulating factor has improved neutrophil function against S. prolifi cans hyphae (37). Surgery appears to be an important factor in survival in our cases, as well as other series (10). Although surgery and antifungal therapy were successful in 2 neutropenic patients with invasive S. prolifi cans infections, these patients also had concomitant neutrophil recovery that could explain their survival.
Much needs to be learned about the epidemiology, transmission, pathogenesis, and environmental niche of Scedosporium spp., especially S. prolifi cans. A source for emergence of S. prolifi cans at our hospital was investigated, but none was identifi ed. Our inability to detect the source of emergent S. prolifi cans was puzzling, although exposure and colonization may have occurred by the time the fi rst case was seen. Cultures may have been obtained too late to identify environmental contamination, although changes in immunosuppression in the population at risk may also have contributed. S. prolifi cans appeared fi rst as a colonizer of abnormal airways 8 months before the fi rst invasive infection in an HSCT recipient. Shifts in the epidemiology of colonizing organisms in this patient group during hospital construction may be worthy of closer attention and surveillance. Molecular typing methods such as PCR (38) and inter-simple sequence repeat PCR fi ngerprinting (39) have distinguished genotypes from outbreaks and different geographic regions. Recently, S. apiospermum complex has been shown to include several individual species indistinguishable morphologically (40). These methods will be used for further investigation of our isolates as part of a larger Australian surveillance study.
In conclusion, S. prolifi cans and S. apiospermum are pathogenic fungi that demonstrate distinct clinical features dependent on the immune function of the host and the type of species isolated. They are usually found in patients with underlying disease, although occasionally after trauma. S. prolifi cans emerged as a major pathogen in allogeneic HSCT recipients and as a colonizer of patients with underlying lung disease. Madura foot is now rare and was not observed in this series. The high attributable mortality rate of invasive infection with both Scedosporium spp., limitations of antifungal therapy, necessity for aggressive and deforming surgery to treat infections with S. prolifi cans, and uncertainty over the role of airway colonization emphasize the need to better understand the epidemiology and pathogenesis of this infection.