Detection of Scedosporium spp.: Colonizer or pathogen? A retrospective analysis of clinical significance and management in a large tertiary center

Abstract Infections with Scedosporium spp. are emerging in the past two decades and are associated with a high mortality rate. Microbiological detection can be associated with either colonization or infection. Evolution from colonization into infection is difficult to predict and clinical management upon microbiological detection is complex. Microbiological samples from 2015 to 2021 were retrospectively analyzed in a single tertiary care center. Classification into colonization or infection was performed upon first microbiological detection. Clinical evolution was observed until July 2023. Further diagnostic procedures after initial detection were analyzed. Among 38 patients with microbiological detection of Scedosporium spp., 10 were diagnosed with an infection at the initial detection and two progressed from colonization to infection during the observation time. The main sites of infection were lung (5/12; 41.6%) followed by ocular sites (4/12; 33.3%). Imaging, bronchoscopy or biopsies upon detection were performed in a minority of patients. Overall mortality rate was similar in both groups initially classified as colonization or infection [30.7% and 33.3%, respectively (P = 1.0)]. In all patients where surgical debridement of site of infection was performed (5/12; 42%); no death was observed. Although death occurred more often in the group without eradication (3/4; 75%) compared with the group with successful eradication (1/8; 12.5%), statistical significance could not be reached (P = 0.053). As therapeutic management directly impacts patients’ outcome, a multidisciplinary approach upon microbiological detection of Scedosporium spp. should be encouraged. Data from larger cohorts are warranted in order to analyze contributing factors favoring the evolution from colonization into infection.


Introduction
Fungal infections caused by dematiaceous moulds of the genus Scedosporium are caused by different species belong-ing to the Scedosporium apiospermum complex (including amongst others Scedosporium apiospermum and Scedosporium boydii) and Scedosporium aurantiacum , which can be found in soil and polluted water. 1 The separation from the genus Lomentospora (formerly included to the genus of Scedosporium ) is clinically highly relevant due to different drug resistance patterns. 2 , 3Clinical classification ranges from colonization to localized infections to invasive disease.Invasive disease with dissemination typically occurs in immunocompromised patients. 4Localized infections are associated with inoculation trauma in immunocompetent patients and are often described in eye trauma. 5 , 6Scedosporium spp.interpreted as colonization is mostly linked to respiratory specimens of patients with cystic fibrosis (as the second most frequent fungus after Aspergillus spp.) or other chronic respiratory disease with perturbation of airway architecture and mucosal surface, including the ear, nose, and throat (ENT) area. 1 , 7-10 Due to the increasing numbers of immunocompromised patients, infections with Sce-dosporium spp.have been emerging over the last two decades. 7 , 11 , 12he progression from colonization to infection in immunocompetent and immunocompromised hosts is not fully predictable.Typical risk factors for invasion are hematological stem cell transplantation (HSCT), solid organ transplantation (SO T), and diabetes mellitus. 1 , 13 , 14Furthermore, the evolution of colonization is only partially predicted by antifungal prophylaxis: some lung transplant recipients colonized with Scedosporium spp.6][17] Data on the evolution from colonization to infection under immunosuppressive treatment are scarce as most patients are under universal antifungal prophylaxis.
The diagnostic approach depends on microbiological culture and, if possible, on histopathology to further distinguish infection from colonization. 17][20] Treatment options are limited, as Scedosporium spp.may exhibit resistance to multiple treatment options, especially to Amphotericin B if used as single therapy. 7 , 16Recommended first-line treatment is voriconazole in conjunction with surgical excision in localized infections.Additionally, reduction of immunosuppressive therapy is advised. 4 , 16 , 17n this retrospective observational study, we aimed to analyze the microbiological detection of Scedosporium spp. in a large tertiary center over 6 years, including the clinical evolution from colonization to infection and related outcome.Diagnostic and therapeutic management, as well as the factors influencing the outcome of patients with infection were collected from the medical chart.

Methods
From 2015 to 2021, all microbiological samples positive for Scedosporium spp .at the Institute of Medical Microbiology in Zurich, Switzerland with a signed general consent form from the University Hospital Zurich were eligible for further evaluation.The study was approved by the local ethics commission of Zurich (BASEC Nr. 2022-01090).
Patient's data were analyzed in the electronic hospital health record system.At initial microbiological detection, a classification of Scedosporium spp.colonization or infection was performed.Colonization was defined as the presence of Scedosporium spp.without any clinical evidence of infection.Infection was defined according to the revised EORTC criteria of 2019 for invasive fungal infections, including histopathological evidence of hyphae in otherwise non-infected tissue or as per physician's assessment according to clinical presentation and/or imaging modalities, as the EORTC mentioned host factors for mould infections are not applicable owing to the fact that infection can also affect healthy hosts. 21Infection was further subdivided into 'localized infection', 'invasive infection' according to an extension to surrounding tissue or 'disseminated infection' in case of positive blood cultures or infection at ≥ 2 non-contiguous sites. 10 , 22dditional initial diagnostic investigations, such as biopsy or imaging modalities aiming to differentiate infection from colonization at the time of microbiological detection were reassessed.
Period of observation was until July 2023 or until date of death, respectively.Clinical evolution over time was classified into 'no evolution' or 'from colonization to infection'.In patients with recurrent microbiological detection, days of recurrence were calculated at least 14 days following initial detection.Detection after less than 14 days was not considered as recurrent.Repetition of sampling were performed individually by the treating physicians and did not follow a standardized protocol.
Treatment modalities were subdivided into antifungal therapy (combined or single), concomitant reduction of immunosuppressive treatment and/or surgical debridement. 18utcome was investigated for findings on imaging, occurrence of death, or eradication.Findings on imaging were classified into progression of infection, stable response, and regressive findings at six and 12 weeks after initial diagnosis, respectively. 23Days between first microbiological detection and death were calculated and cause of death was analyzed whether Scedosporium spp.infection was directly attributable or not.Eradication was assessed by negativity of repeated microbiological sampling and regression of clinical and imaging findings.
For the identification of filamentous fungi, clinical samples were analyzed by microscopy and cultured on different agar types as described previously. 24Briefly, samples were cultured on Sabouraud agar with gentamicin and chloramphenicol (Becton Dickinson AG, Switzerland) at 25 • C and 35 • C, on Mycosel agar (Becton Dickinson AG, Switzerland), and potato carrot agar (PCA), both at 25 • C. Phenotypic identification up to genus level was based on macro-and micro-morphological criteria, according to de Hoog G.S., 2009. 25olecular identification was done if it was deemed clinically relevant, if macro-and micro-morphological identification was not possible or when a respiratory specimen originated from a cystic fibrosis (CF) patient.Chromosomal DNA were extracted from isolates using the InstaGene extraction matrix (Biorad) according to the manufacturer's instructions.Identification was performed by amplification and sequencing of the inter spacer region (ITS) region as described previously. 26Sequencing data were analyzed using the ITS Smart-Gene custom database, applying the identification guidelines by Ciardo et al. 27 Antifungal susceptibility testing (AFST) was performed on request or if deemed clinically relevant by the treating physician.AFST was done by gradient diffusion strips using Etest (bioMérieux, France; Liofilchem S.r.I., Italy).Conidia from a well sporulated culture on PCA were resuspended in 0.85% NaCl with 0.01% Tween80.After 3-5 min of sedimentation, the supernatant, without the mycelium below, was transferred to a fresh tube and vortexed vigorously.A fungal suspension of McFarland 0.5-1 was prepared in fresh 0.85% NaCl and spread onto Roswell Park Memorial Institute (RPMI) agar plates (bioMérieux, France) with a cotton swab.Antifungal gradient strips were applied, and the plates were incubated at 35 • C. Minimal inhibitory concentrations (MIC) were read after 24-48 h of incubation.MICs for amphotericin B were read at complete growth inhibition, MICs for azoles (voriconazole, posaconazole, isavuconazole, and itraconazole) and echinocandins (anidulafungin and caspofungin) were read at 80% of growth inhibition (minimal effective concentration).Micafungin AFST was performed by broth mi-crodilution using the Sensititre YeastOne panel (TREK Diagnostic Systems, UK) according to the manufacturer's manual for Aspergillus spp .Etest MIC values were rounded up and adjusted to a binary log scale.
If several isolates per patient were tested, only the first isolate per patient per year and material were included in this study.
Statistical data were analyzed using SPSS version 29.0 (IBM Corp., USA) and R Version 2023.06.1 + 524.Baseline characteristics were calculated by frequency, average, or mean.Comparison analysis was performed by Fisher's exact test.A Pvalue of < 0.05 was considered statistically significant.Death, successful eradication, and strength of association with surgical debridement and antifungal therapy were analyzed by calculation of odds ratio, with determination of confidence interval.

Patient characteristics
Patient's baseline characteristics are presented in Table 1 .Out of 40 patients with positive microbiological samples for Scedosporium spp ., 38 were included in the analysis.In the patients classified as having infections ( n = 10), nine presented as localized infection and one as invasive infection (eye trauma with presence of hyphae in the surrounding periorbital tissue).After a polytrauma related to a traffic accident the infection was localized in the calf muscles.The majority of patients ( n = 25; 65.7%) with a positive sample had underlying lung disease.Out of these 25 patients, 22 (88%) were classified as colonization and three (12%) as localized infection at initial diagnosis.

Diagnostic steps initiated after the initial detection
Upon evidence of a positive culture histological additional diagnostic investigations were performed in 10 out of 38 patients (26.3%), bronchoscopy in five out of 25 patients with pulmonary pathologies (20%), and imaging (e.g., CT scan of lungs or brain) in 12 patients (31.6%) (cf.Table 2 ).In the vast majority of patients ( n = 28; 73.7%)only a single culture was positive, whereas in the remaining patients ( n = 10; 26.3%) cultures were repetitively positive beyond 14 days of initial detection, ranging from five to 10 times over a time period from 24 to 138 days (mean 67 days).

Microbiology
For the final identification and to exclude Lomentospora prolificans , 27 (71.1%)isolates were analyzed by sequencing of the ITS region.A total of 23 isolates belonged to the S. apiospermum complex (including S. boydii ), one isolate was identified as S. aurantiacum , and three isolates could only be identified to genus level, because sequencing of the ITS region cannot distinguish Scedosporium dehoogii from the S. apiospermum complex. 1 The remaining 11 isolates were only identified by microscopy and phenotypic features.
Antifungal resistance patterns of 21 Scedosporium species isolated from 17 patients are shown in Table 3 and detailed in Appendix 1 .Choice of antifungals tested varied, depending on consultation with clinicians.Furthermore, Scedosporium isolates from 10 patients were tested retrospectively on all shown antifungals, including micafungin, which was only available at our lab as part of the Sensititre YeastOne   was applied when isolates were only identified by microscopy and phenotypic features or when sequencing did not allow unequivocal identification based on the used database.TBS: tracheobronchial secretion; BAL: bronchoalveolar lavage.microdilution panel.All isolates were tested for voriconazole, which had the lowest MIC 90 (inhibition of 90% of the isolates) of 0.125 μg/ml and an MIC range of 0.008-0.5 μg/ml, followed by posaconazole ( n = 19) and isavuconazole ( n = 14) both with an MIC 90 of 1 μg/ml and MICs ranging from 0.016 to 2 μg/ml and 0.008 to 1 μg/ml, respectively.In vitro activity of amphotericin B ( n = 18) against Scedosporium sp. was low and MICs ranged from 0.5 to 32 μg/ml with a median MIC (MIC 50 ) of 8 μg/ml and a MIC 90 of 32 μg/ml.Total numbers of Scedosporium sp.tested for echinocandins were low; caspofungin ( n = 13), anidulafungin ( n = 11), and micafungin ( n = 10), and the echinocandins showed very low in vitro activity, with caspofungin being the most effective with MICs ranging from 0.064 to 32 μg/ml, an MIC 50 of 1 μg/ml, and an MIC 90 of 32 μg/ml.Micafungin showed no in vitro activity against Scedosporium sp.

Clinical evolution and diagnostic evaluation after initial detection
Evolution from colonization into infection was observed in a total of two patients (2/28; 7.1%).One developed an invasive infection from pulmonary colonization after lung transplantation with a septic knee arthritis 88 days after first microbiological detection (while no antifungal therapy was initiated at diagnosis of colonization).The other patient with CF developed an allergic bronchopulmonal mycosis (ABPM) 501 days after first detection.Both infections were caused by the S. apiospermum complex.
Follow-up imaging at six and 12 weeks was performed in 10 patients with initial colonization ( n = 5) or infection ( n = 5).At week six, progression of infection was demonstrated in two patients (40%), while two had stable (40%) and one regressive findings (20%).At week 12, no patient had progression, four had stable (66.7%), and two had regressive findings (33.3%).Of note, some patients only underwent follow-up imaging at weeks 6 or 12.

Treatment
At initial detection, antifungal treatment was initiated in eight out of 10 patients categorized as infection (80%) (Fig. 1 ).Two patients, both with CF, had a treatment delay of 15 and 52 days, respectively.In one patient, it was due to unawareness of microbiological results, while in the second patient, no treatment indication was made as inflammation parameters were absent.Two patients (dacryocystitis and chronic rhinosinusitis) had no antifungal treatment at all and only surgical debridement was performed.
Voriconazole was the main initial antifungal treatment (9/10; 90%).In one patient, a combination of voriconazole and terbinafine was initiated (based on physician's assessment upon the finding of a disseminated infection with septic knee arthritis).Three patients (30%) developed a voriconazole induced drug toxicity requiring a switch to posaconazole ( n = 2) and isavuconazole ( n = 1) as second-line treatment.Antifungal treatment could be stopped in five patients (50%) due to regression of initial findings.If treatment could be stopped, it lasted for a median time of 98.5 days [SD 66 days; eye infections ( n = 3), ABPM ( n = 1); polytrauma ( n = 1)].In the remaining five patients (50%) in whom treatment could not be stopped, treatment duration lasted until day of death  (11-1168 days, median: 573 days) or is still ongoing at last follow-up.Surgical debridement was performed in five patients, including patients with eye infections ( n = 2), septic knee arthritis ( n = 1), polytrauma ( n = 1), and chronic rhinosinusitis ( n = 1).In the remaining five patients, only antifungal treatment was administered [cystic fibrosis ( n = 3), contact lensassociated and postoperative keratitis ( n = 2)].Concomitant reduction of immunosuppressive was not conducted in any of the two patients with infection.
Median time to death was 540 days after first microbiological detection in infection (12-1201, SD 489 days) and 691 days in colonization (62-1440; SD 466 days).Eradication of initial infection could be achieved in eight out of 12 patients (66.6%) (Fig. 1 ).In the group with successful eradication, death occurred in one out of eight patients (12.5%) compared to three out of four patients (75%), where no eradication was achieved (OR 21.0; CI 0.96-458.86;P = 0.053).Only in one patient (lung transplantation due to CF with disseminated S. apiospermum infection) death was related to the mould infection, while in the other three patients causes of death were unrelated.
Surgical debridement was performed in four out of eight (50%) patients with successful eradication (two patients were concomitantly treated with antifungals) and in one out of four (25%) patients without eradication (OR 3.0; CI 0.21-42.62;p = 0.58).No death occurred in patients undergoing surgical debridement.
Of the 10 patients receiving antifungal treatment, four patients died (40%), including one patient receiving a combination of voriconazole and terbinafine.Eradication was achieved in six patients (60%) after antifungal treatment.No eradication could be achieved in the two patients with CF where treatment initiation was delayed.

Discussion
Lung as the main site of infection is in line with reports from earlier cohort studies. 13 , 28-30The second most frequent sites were ocular infections, also described in previous studies. 220][31] Only one out of 12 infections (8.3%) progressed to disseminated disease.This is less than described in the literature, where rates of dissemination ranged from 11.5% up to 30%.The low prevalence of severely immunocompromised patients (specifically, no neutropenic patient nor after HSCT) may be one of the main factors.Other explanations for this low rate of disseminated disease might be the intensified prevention strategies for mould infections in the last two decades, including mould-specific antifungal prophylaxis after solid organ transplantation with activity against Scedosporium spp.(but limited activity against Lomentospora prolificans ).
Upon microbiological detection, further diagnostic steps were limited to a minority of patients and not following recommendations from current guidelines with the goal to differentiate between colonization and infection, or to assess the extent of infection. 18 , 21Only one-third of all patients underwent imaging modalities to assess for signs of invasive mould infections, even though imaging is strongly encouraged by different recommendations. 4 , 18Histologic examinations yielded an invasive infection in one out of 10 biopsies, (endophthalmitis with infiltration of periorbital tissue).Despite this low yield, exclusion of an invasive disease can guide further management.One possible explanation for this very low adherence to guidelines is that the limited awareness about the potential danger does not trigger further examinations.
Surgical debridement represents one of the cornerstones in the initial management in localized lesions. 7In our population, it was performed in five out of 12 patients with infection (41.6%).In patients experiencing surgical debridement, no death was observed during the observation period.Small numbers preclude any strong recommendations.While certain case series did not demonstrate an improvement in overall survival following surgery, others showed an association with decreased mortality rates, thus limiting general conclusions regarding the impact of surgical debridement on mortality. 22 , 31eduction of immunosuppression, a further cornerstone in treatment of mould infections, was not performed in the two patients with infection under immunosuppressive treatment.Awareness of this management strategy should be reinforced using an interdisciplinary approach whenever this reduction can be taken into consideration.In two cases, of which one with immunosuppression, treatment was delayed despite diagnosis of infection, which represents an independent predictor for mortality at 3 months. 30This further supports the importance of an infectious disease consultation upon microbiological detection.
As expected, the antifungal susceptibility testing results were concordant to published results with lowest MIC for voriconazole followed by posaconazole. 24 , 32Despite the weak recommendation supporting its usage in the treatment of scedosporiosis, isavuconazole was administered in one patient as a second-line treatment due to drug toxicity with voriconazole. 18In general, AFST should not primarily guide the choice of antifungal treatment and is only recommended with moderate evidence as clinical breakpoints are missing. 9However, it can help choosing a second-line antifungal treatment as in this case.Combination therapy with terbinafine and azoles, as applied as first-line treatment in one patient, exhibits an in vitro synergistic activity. 34 , 35Due to poor tissue penetration of terbinafine and scarce clinical data on outcome, the clinical significance of this combination remains unclear and is moderately recommended as first-line alternative over voriconazole monotherapy. 18 , 24t remains unclear why certain patients progress from colonization to infection, as this evolution can be observed in immunocompetent and immunocompromised patients.Furthermore, infections can still be observed despite antifungal prophylaxis, especially in Scedosporium spp.colonization after lung transplantation, under life-long antifungal prophylaxis. 15A very recent study included colonization in their cohort, but no information was provided on the clinical evolution. 29Other cohorts actively excluded colonization or were focused on patients with colonization in CF or lung transplant recipients only. 22 , 30-34he overall mortality rate was similar in the colonization and infection groups (30.7% and 33.3%, respectively) and higher in the subgroups receiving immunosuppressive therapy (50% in colonization and infection) compared to immunocompetent patients (30% in infection, 18.8% in colonization).In a very recent study, airway colonization with Scedosporium spp. in CF patients has been associated with higher numbers of hospitalizations and bacterial infections as compared to CF patients without colonization. 35Infection with S. apiospermum complex was directly attributed to scedosporiosis in only one patient of our cohort.Published series described infection attributed mortality of up to 76%. 22 , 36Overall mortality from these large cohorts ranged from 43.6% to 70% in immunocompromised and 30.7% in immunocompetent patients. 22 , 30 , 37Here again, our population is biased by the fact, that only few patients were immunocompromised, had disseminated disease or central nervous system involvement, which are all representing predictors for mortality. 22 , 31ssessment of eradication of infection is representing a new aspect for evaluating Scedosporium spp.infections outside the field of CF or lung transplantation, where especially eradication of colonization by antifungal treatment has been studied, but mostly with poor results. 38 , 39In this cohort, even though eradication and surgical debridement were associated with a lower number of deaths, a major bias is represented by the relatively high numbers of ocular infections, corresponding to a less comorbid patient population than patients with pulmonary pathologies.
One of the study's main limitations, next to the already above-mentioned biases, is the small and very heterogeneous cohort size.This restricts the interpretation for significant results especially regarding outcome data or the performance of a multivariate analysis.Thus, no general conclusions can be drawn and studies in larger multicentric cohorts are warranted, especially to explore risk factors for clinical evolution from colonization to infection.
Scedosporium spp.colonization and infections are representing a major marker for morbidity and mortality.As the management upon microbiological detection is complex and can directly influence patient's outcome, a multidisciplinary approach should be encouraged in particular to define further diagnostics and treatment options.

Figure 1 .
Figure 1.Outcome and interventions of initially or during observation time diagnosed infections with Scedosporium spp.Antifungal treatment (AF) was initiated in ten out of twelve patients, while in the remaining two patients (with chronic sinusitis resp.dacryocystitis) surgical debridement was considered sufficient for eradication.

Table 1 .
Baseline characteristics of 38 patients with Scedosporium spp.detection in microbiological samples.

Table 2 .
Microbiological diagnostics and results, further diagnostics at initial microbiological detection.

Table 3 . Antibiotic resistance patterns of 21 Scedosporium spp. isolated from 17 patients.
In three patients with recurring or disseminated Scedosporium infections one AFST per year or specimen was included in the analyses.