Phaeohyphomycoses in kidney transplant recipients: A 10-year retrospective study from the French West Indies

Abstract

Phaeohyphomycoses (PHM), infections caused by pigmented fungi, are increasingly reported in kidney transplant recipients (KTR), but data remains scarce. We retrospectively reviewed all episodes of PHM in KTR at the Guadeloupe University Hospital (French West Indies) from January 1st, 2012, to December 31st, 2022. Eighteen episodes of PHM occurred in 17 KTRs, corresponding to an incidence of 1.56 cases per 100 person-years. Median time from transplantation to PHM onset was 15 months (interquartile range: 9.0–23.5 months). Most episodes were localized cutaneous infections (n = 16/18). Twelve fungal species were identified; Paraconiothyrium cyclothyrioides was the most frequent (n = 6/18, 33%). This fungus it is not yet acknowledged as pathogenic in the current European guidelines. Surgery was performed in 41% of episodes with available data—some data are missing due to the loss of paper files. Eighty-two percent of patients received antifungal therapy (mainly azoles). Overall outcome was favorable in 82%. Three deaths occurred during follow-up; one was directly attributable to disseminated PHM. Diagnosing PHM is not that unusual in this caribbean transplant center, mainly presented as localized skin infections, with a distinctive species distribution characterized by predominance of P. cyclothyrioides and absence of Alternaria spp. With accurate diagnosis, combine surgical/medical therapy and control of transplant organ management, the vast majority of patients can be successfully managed.

Introduction

Dematiaceous fungal organisms, commonly referred to as black fungi, are characterized by their dark pigmentation, which contributes to their virulence and ability to evade host defenses.¹ These fungi are ubiquitous, frequently found in plants and soil, and are more prevalent in tropical regions.^2–5 Infections caused by these melanized, septate filamentous fungi are known as phaeohyphomycoses (PHM), with a complex taxonomy including over 100 species.¹

Although typically opportunistic, PHM can occur in immunocompetent individuals through direct inoculation following skin barrier injuries. Dematiaceous fungi can also colonize the airways of lung transplant recipients.⁴ Since October 2022 fungal diseases have been classified as a public health threat by the World Health Organization (WHO), as their incidence is increasing worldwide.^6–8 While PHM are usually localized, dissemination may occur, especially in immunocompromised patients,^5,9,10 such as kidney transplant recipients (KTR).^2–5 PHM gained attention during a 2012 meningitis outbreak in immunocompetent patients due to contaminated corticosteroid in the United States of America (USA),^11–13 and were linked to postoperative heart surgery infections.¹⁴

Data on PHM following kidney transplantation remain limited. Most knowledge about PHM in KTR comes from single-center studies and from literature reviews.^{2,7,10,15–18} According to the Mycoses Study Group registry, conducted across 18 hospitals in the USA, Peru, and Australia over seven years, only 18 cases were identified in KTR, highlighting the rarity of these infections in this population.¹⁹ Some cases of Exophiala spp. infections were described in French Guiana, ²⁰ while it is not infrequent in mainland France. Nevertheless, data on PHM in KTR from the Caribbean region remain scarce.²¹

To address this gap, we aimed to investigate the frequency, characteristics, and outcomes of PHM diagnosed in KTR at our Caribbean transplant center. As the only transplant center in the French West Indies, an insular setting, the Guadeloupe University Hospital (GUH) provides a unique setting to study these infections.

Materials and methods

Study design, setting, population, and ethical aspects

We conducted a single-center retrospective study at the GUH to describe the frequency, characteristics, management, and outcomes of PHM in KTRs from January 1st, 2012, to December 31st, 2022. The GUH is a 656-bed academic hospital located in the French West Indies. Patients were eligible for study inclusion if they met the following criteria: (1) age ≥ 18 years, (2) KTR, (3) presence of signs and/or symptoms consistent with dematiaceous fungal infection, (4) microbiological confirmation on skin biopsy performed on the study period at the GUH, and (5) follow-up at the GUH. Patients were excluded if they did not consent to the use of their data for research purposes, or if their medical records were unavailable. The study was approved by the GUH Ethics Committee (reference: A65_21_09_13_TREIN). The processing of personal data related to the protocol of this study complies with the requirements of the Commission Nationale Informatique et Libertés, under the European General Data Protection Regulation, as attested by the delegate for data protection (MR-004, 09/13/21).

Data collection and definitions

Clinical and demographic data were extracted from electronic medical records. Collected information included patient age, gender, underlying medical conditions, kidney transplant-related data (duration of pre-transplant dialysis, date of transplant, type of immunosuppressive therapy, preservation fluid culture results), time from transplant to PHM (with the date of diagnosis defined as the day of collection of the first positive clinical sample [skin biopsy]), immunosuppression status at time of PHM diagnosis, and PHM-related data (clinical presentation, diagnostic method, histopathological findings, and therapeutic management). PHM were classified as early onset if they developed in the first 90 days post-transplant; late onset if they occurred at least 90 days post-transplant. PHM were also classified as either single or multiple, localized or disseminated. Disseminated phaeohyphomycosis was defined as involvement of more than one noncontiguous site of infection, following the criteria used in the TRANSNET study (made by the Transplant Associated Infection Surveillance Network, to monitor data on phaeohyphomycoses in transplant recipient). .¹⁶

The immunosuppression status at time of diagnosis encompassed ongoing anti-rejection treatment, tacrolimus dosages (target concentration: between 5 and 12 ng/ml, subtherapeutic levels: strictly < 5 ng/ml, supratherapeutic levels strictly above 12 ng/ml), gammaglobulinemia levels (normal: 8–16 g/l) with hypogammaglobulinemia defined as gammaglobulinemia levels < 8 g/l or presence of a monoclonal-like spike ≥ 20 g/l. Diabetes status was determined from medical records, and glycemic control was assessed using glycated hemoglobin (HbA1c) levels. Cytomegalovirus (CMV) and BK virus (BKV) reactivations were diagnosed by polymerase-chain reaction in total blood and defined as viral loads > 3 log in previously negative patients. CMV and BKV reactivations were screened on the last three months before diagnosis. Recurrent graft pyelonephritis was defined as > 3 episodes in the year before PHM diagnosis .²²

Kidney function was assessed using serum creatinine levels collected at diagnosis, at the end of treatment (6 months after diagnosis), and at the end of study follow-up.

Drug monitoring was assessed when available. As defined in guidelines, the voriconazole recommended therapeutic range was 1–5.5 mg/l,²³ and the posaconazole recommended therapeutic range was > 1 mg/l.²⁴

Outcomes included clinical response, defined as the disappearance of the lesion, and microbiological cure, considered as absence of growth on samples. Relapse was defined as reinfection at the same organism while on antifungal therapy in the less or equal six first months; recurrence, defined as a subsequent episode with same fungus, occurring after completion of antifungal treatment for a minimum of 6 months regardless of occurrence delay, and new infection, caused by another organism. Impaired kidney graft function defined as serum creatinine levels greater than 125 µmol/l or dialysis, and survival status at time of last follow-up. An independent committee reviewed details related to death and assessed attributability to PHM.

Mycological investigations

Species identification was initially performed using standard mycological techniques, including cultures on Sabouraud chloramphenicol gentamicin agar, May-Grunwald-Giemsa and Ziehl-Neelsen staining, and matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF Vitek MS, Biomérieux). All strains were then sent to the French National Reference Center for Invasive Mycoses and Antifungals (NRCMA), Institut Pasteur, Paris for definitive identification, applying a polyphasic approach: first, phenotypical study, and second, sequencing of the internal transcribed spacer barcoding region and of the D1/D2 region of the 28S ribosomal RNA gene as previously described.^24,25 Because of limitation in routine identification method, particularly of MALDI-TOF MS databases on dematiaceous fungi, discrepancies occurred in 10 out of 18 identifications. The MALDI-TOF databases used at the time for identifying black fungi did not allow for precise discrimination. For many fungal cultures, mycologists classified as “Phaeohyphomycete” without identification and the strain was sent to the CNR. Identification came later with the return from the CNR. In vitro susceptibility testing was performed according to the European Committee on Antimicrobial Susceptibility Testing procedure.²⁶ A total of seven antifungal agents (ALSACHIM, Illkirch, France) were used to determine the minimal inhibitory concentrations (MICs) for itraconazole, voriconazole, posaconazole, isavuconazole, and amphotericin B or minimal effective concentrations (MECs) for caspofungin, and micafungin. Blood culture were grown in BacT/ALERT® FA PLUS, BacT/ALERT® FN PLUS et BacT/ALERT® FP PLUS (Biomérieux ®) bottles, aerobia, anaerobia, et pediatric respectively. Cultures were pursued for 14 days, incubated in the Virtuo BacT/ALERT (Biomérieux ®).

Statistical analysis

Categorical data were summarized as counts and percentages, and compared using χ² or Fisher’s exact test. Continuous variables were summarized as medians and interquartile ranges (IQRs) and compared using Wilcoxon-Mann-Whitney rank test. All statistical analyses were conducted using GraphPad Prism (v.9.3.1) and Rstudio (v.2024.04.1).

Results

Characteristics of study population

A total number of 18 PHM episodes occurred in 17 patients with KTR during the study period, of whom 11 were males (65%) (Table 1). Median age at time of infection was 62 years (IQR: 59–69 years). Individual characteristics of KTR are described in Table 2. Most patients had diabetes mellitus (59%, 10/17), with a median HbA1c of 6.4% (IQR: 6.2%–6.9%). The median duration of dialysis before transplant was 43.0 months (IQR: 32.5–119.3 months).

Table 1.

Characteristics of patients.

	Overall number (n = 17)
Demographic characteristics
Age at diagnosis (years) median [IQR]	62.0 [59.0–69.0]
Male gender, n (%)	11 (65%)
Diabetes, n (%)	10 (59%)
HbA1c (%), median [IQR]	6.4% [6.2–6.9]
HIV infection, n (%)	0 (0%)
HTLV infection, n (%)	1 (6%)
Kidney transplant caracteristics
Place of graft, n (%)
Guadeloupe	16 (94%)
Mainland France	1 (6%)
Median time on dialysis, months, median [IQR]	43 (32.5–119.3)
Primary kidney disease diagnosis, n (%)	17 (100%)
Vascular	3 (18%)
Diabetic	2 (12%)
Interstitial	1 (6%)
Amyloid	1 (6%)
Unknown	10 (59%)
Immunosuppression medication, n (%)	17 (100%)
Tacrolimus	16 (94%)
Corticosteroids	14 (82%)
MMF	15 (88%)
Ciclosporin	2 (12%)
Biological characteristics at diagnosis, median [IQR]
Total lymphocyte count (G/l)	0.5 [0.3–0.8]
CD4 count (mm3)	176 [88–242]
CD4 to CD8 ratio	0.61 [0.49–0.91]
Gamma globulin levels (g/l)	11.8 [9.2–14.0]
Creatinine at diagnosis (µmol/l)	163.0 [135.5–207.5]
Creatinine at end of treatment, (µmol/l)	164.0 [134.0–204.0]
Creatinine at end of follow-up, (µmol/l)	238.0 [127.0–322.0]
Through tacrolimus level (ng/ml)	10.1 [7.5–15.1]
Post transplant infectious complications
Graft fluid infection, n (%)	6 (35%)
Concomitant CMV reactivation, n (%)	4 (24%)
Concomitant BKV reactivation, n (%)	0 (0%)
Recurrent graft pyelonephritis, n (%)	5 (29%)
Number of recurrent graft pyelonephritis, median [IQR]	4 [3-8]
Other fungal infection during management, n (%)	4 (24%)

BKV: BK Virus; CMV: cytomegalovirus; HbA1c: glycated hemoglobin; HIV: Human Immunodeficiency Virus; HTLV: Human T-lymphotropic Virus; IQR: Interquartile Range; MMF: Mycophenolate Mofetil.

Table 2.

Individual patient’s characteristics.

	Patient 1	Patient 2	Patient 3	Patient 4	Patient 5	Patient 6	Patient 7	Patient 8
Year of kidney transplantation	2011	2012	2014	2015	2015	2016	2016	2016
Year of diagnosis	2012	2014	2016	2016	2016	2016	2016	2017
Age at diagnosis, years	71	69	60	69	55	61	61	57
Delay after KTR, months	12	22	18	12	12	8	2	6
Number of locations at diagnosis	1	1	1	1	1	3	1	1
Number of locations during follow-up	4	1	1	1	1	3	4	1
Main location	Right third finger	Right hand	Right knee	Left ear	Right hand	Right hand and arm	Right leg	Third finger
Concomitant blood culture	Negative	Negative	Not done	Negative	Negative	Not done	Positive S. capitis (42 h)	Not done
Direct examination	Positive: Mycelial hyphae	Positive: Numerous mycelial hyphae	Positive: Few hyphae. Numerous conidias	Missing data	Missing data	Positive: Few hyphae. A few conidias.	Positive: Numerous mycelial hyphae.	Positive: Numerous hyphae
Fungal culture	Phaeoacremonium parasiticum, Paraconiothyrium cyclothyrioides	Phialemoniopsis ocularis	Paraconiothyrium cyclothyrioides	Curvularia hawaiiensis	Colletotrichum gigasporum	Exophiala oligosperma	Phaeoacremonium venezuelense Phaeoacremonium parasiticum	Trematosphaeria grisea (syn. Madurella grisea)
Localized/disseminated	Disseminated	Localized	Localized	Localized	Localized	Localized	Localized	Localized
Tacrolimus supratherapeutic level after treatment	Unknown. Stopped after diagnosis.	No	No	Missing data	No	Yes	No	No
Treatment (molecule + duration)	Voriconazole + AmphoB	Voriconazole, unknown duration	Posaconazole, 6 months	No antifungal treatment	No antifungal treatment	Voriconazole 6 months then Posaconazole, for 6 months	Voriconazole, 6 months	Voriconazole, 6 months
Surgery	Yes	Unknown	Yes	Yes	Yes	Yes	Unknown	Yes
Relapse (< 6 months)	Yes	No	No	No	No	Yes	Yes	Yes
Recurence	No	No	No	No	No	No	Yes (3)	No
Outcome (cure/death)	Death	Cure	Cure	Cure	Cure	Cure	Cure	Cure
	Patient 9	Patient 10	Patient 11	Patient 12	Patient 13	Patient 14	Patient 15	Patient 16	Patient 17
Year of kidney transplantation	2015	2017	2017	2013	2014	2017	2017	2019	2018
Year of diagnosis	2017	2018	2018	2018	2018	2019	2019	2020	2021
Age at diagnosis, years	74	46	62	59	57	68	63	78	71
Delay after KTR, months	17	13	8	60	55	21	24–30	7	29
Number of locations at diagnosis	2	1	1	1	2	1	2	1	1
Number of locations during follow-up	2	1	1	1	2	1	2	2	1
Main location	Right hand and right shoulder	Undetermined	Hand	Forearm	Left leg	Right leg	Left leg	Left Hallux	Ankle
Concomittant blood culture	Negative	Not done	Not done	Not done	Not done	Not done	Not done	Negative	Not done
Direct examination	Positive: Numerous hyphae. Numerous conidias.	Positive: rare yeasts	Positive: rare yeasts, numerous mycelial hyphae	Positive: some yeasts and some mycelial hyphae	Positive: Numerous mycelial hyphae	Positive: many encapsulated yeasts, numerous mycelial hyphae	Positive for the two locations: some yeast; numerous mycelial hyphae	Positive: many encapsulated yeasts	Missing data
Fungal culture	Paraconiothyrium cyclothyrioides	Exophiala spinifera	Exophiala spinifera	Phialophora chinensis	Diaporthe sp	Exophiala spinifera	Paraconiothyrium cyclothyrioides. Neoscytalidium dimidiatum	Paraconiothyrium cyclothyrioides	Paraconiothyrium cyclothyrioides
Localized/disseminated	Localized	Localized	Localized	Localized	Localized	Localized	Localized	Localized	Localized
Tacrolimus supratherapeutic level at diagnosis	Yes	No	Yes	No	Yes	No	No	Yes	Yes
Treatment (molecule + duration)	Voriconazole, unknown duration	Voriconazole, unknown duration	Voriconazole, unknown duration	Voriconazole, 6 months	Voriconazole, unknown duration	Voriconazole, unknown duration	Voriconazole, unknown duration	Voriconazole, then Posaconazole	Missing data
Surgery	Yes	Unknown	Unknown	Unknown	Unknown	Unknown	Unknown	Unknown	Unknown
Relapse (< 6 months)	No	No	No	No	No	No	Yes	No	No
Recurence	No	No	No	No	Yes (2)	No	No	Yes	No
Outcome (cure/death)	Death	Cure	Cure	Cure	Cure	Cure	Cure	Death	Cure

Regarding immunosuppressive therapy at the time of diagnosis, all patients received calcineurin inhibitors, mostly tacrolimus (n = 16, 94%). The median tacrolimus concentration after antifungal treatment introduction was 10.1 ng/ml (IQR: 7.5–15.1 ng/ml), with 6 patients being overexposed (35%), highlighting the known interaction between antifungals and immunosuppressants. Other commonly used immunosuppressive agents were mycophenolate mofetil (n = 15, 88%) and corticosteroids (n = 14, 82%). Doses of steroids at diagnosis were available for 2/14 patients, and were between 10 mg per day and 12.5 mg per day. Hypogammaglobulinemia was detected in two patients (12%). The median total lymphocyte count was 0.52 G/l (IQR: 0.34–0.79 G/l), and the median CD4 lymphocytes count was 176/mm³ (IQR: 88–242/mm³).

Concerning associated infections at diagnosis, 24% (n = 4/17) of patients experienced CMV reactivation, while none showed BKV replication in blood. Additionally, 29% (n = 5/17) had a history of recurrent graft pyelonephritis, with a median of 4 episodes in the previous year (IQR: 3–8).

Characteristics of dematiaceous fungal infections and microbiological data

Among these 18 cases of PHM, seven of them experienced multiple infections, with multiple locations on the body: 4 with Paraconiothyrium cyclothyrioides and 2 with Phaeoacremonium spp. (Table 2). With a total of 335 kidney transplants followed up during the study period, the overall cumulative episode-based incidence of PHM among KTR was 5.4%, and the incidence in KTR was 1.56 cases per 100 person-years (Supplementary Table 1). The median time from transplant to PHM was 15.0 months (IQR: 9.0–23.5 months). One infection occurred early (in the first 90 days post-transplant); others were late-onset episodes. Most patients (n = 16/17, 94%) presented with localized infections, with acral predominance. Upper limb involvement was observed in 47% (n = 8/17) of patient, including 41% (n = 7/17) affecting the hands, while lower limb involvement was noted in 47% (n = 8/17) of patients. One infection involved the ear (6%), and one undertermined. One patient (6%) had disseminated infection, initially complicating an injury to the right hand, with pulmonary and cerebral involvement. Disseminated infection was related to a dual infection caused by P. cyclothyrioides and Phaeoacremonium parasiticum. The characteristics of PHM, including laboratory test results at the time of diagnostic are detailed in Supplementary Table 2.

Blood cultures were performed in 41% (n = 7/17) of patients at diagnosis, with no case of fungemia detected. Antifungal marker assays were not routinely performed: serum galactomannan antigen was tested in three patients and detected in none of them; and serum 1,3 β-D-glucan (BDG) was tested in only one patient and was positive (level > 500 pg/ml). The only patient tested for BDG had disseminated PHM infection due to both P. cyclothyrioides and P. parasiticum, with an initial lesion on the right hand that developed into extensive cellulitis, but also pulmonary and cerebral lesions. Microbiological fungal cultures of bone marrow, blood and sputum were negative after 3 weeks incubation, and the patient presented as acute kidney failure requiring hemodialysis. Direct examination of skin samples was positive in 78% (n = 14/18) of cases. The most common finding on direct examination was the presence of numerous hyphae in 44% (n = 8/18) of cases. In 11% (2/18), only a few yeast cells were observed. Histopathological findings were suggestive of PHM or fungal infection in 82% of tested cases (n = 9/11) but suggested an alternative diagnosis in two other cases (18%): one capillary hemangioma and one invasive squamous cell carcinoma. Twelve different fungal species were identified from skin lesions (Table 2), among which the more frequent were P. cyclothyrioides (n = 6/18, 33%, Figs 1 and  2), Exophiala spp. (n = 4/18, 22%) - including Exophiala oligosperma (n = 1, Fig. 3) and E. spinifera (n = 3), Phaeoacremonium spp. (n = 3/18, 17%) - including P. venezuelense (n = 1) and P. parasiticum (n = 2). Other identified species (n = 1) included: Colletotrichum gigasporum, Curvularia hawaiiensis, Diaporthe spp., Neoscytalidium dimidiatum, Phialemoniopsis ocularis, Phialophora chinensis, and Trematosphaeria grisea (syn. Madurella grisea). No case of Alternaria spp. was detected.

Figure 1.

Right knee skin lesion related to Paraconiothyrium cyclothyrioides. (A) Lesion before treatment. (B) Lesion after surgery and treatment with posaconazole. (C) and (D) May-Grunwald-Giemsa staining, examined at x10 magnification, revealing P. cyclothyrioides.

Figure 2.

Phaeohyphomycosis caused by Paraconiothyrium cyclothyrioides. (A) Clinical lesion. (B) Direct examination after culture with lactophenol blue x50. (C) May-Grunwald-Giemsa staining, examined at x10 magnification, revealing Paraconiothyrium cyclothyrioides.

Figure 3.

Phaeohyphomycosis caused by Exophiala oligosperma. (A) Skin lesion due to E. oligosperma. (B) May-Grunwald-Giemsa staining, examined at x10 magnification, revealing E. oligosperma. (C) Macroscopic culture of E. oligosperma. (D) Direct examination at x50 magnification with lactophenol blue after culture.

Antifungal susceptibility testing results are described in Supplementary Table 3 and testing could be performed in 72% (n = 13/18) of strains and was missing in the absence of sporulation.

Treatment and outcomes

Treatment details are summarized in Table 2 and Supplementary Table 2. Information regarding surgical status was collected for 7 patients, and all of them underwent surgical operation. Surgery alone was performed for 12% (n = 2/17) of patients, for infection to C. gigasporum and C. hawaiiensis. 29% (n = (5/17) received combined surgical and antifungal therapy. The surgery consisted of excising the skin lesions caused by opportunistic fungal infections, where technically possible, with healthy tissue margins as for basal cell carcinoma (3 mm), with the aim of reducing the duration of antifungal drug treatment. Most patients (n = 14/17) were treated by antifungal therapy, primarily with voriconazole (n = 13/17, 76%) or posaconazole (n = 3/17, 18%). Of the 13 subjects who received voriconazole, 1 received a dual therapy of voriconazole and liposomal amphotericin B (L-AmB) for a total of 20 days: the patient with a disseminated coinfection due to P. cyclothyioides and P. parasiticum. 2 of these patients who had received voriconazole were subsequently given posaconazole, in response to a case of relapse at the second month due to Exophiala oligosperma, and a case of recurrence at the tenth month due to P. cyclothyrioides. Drug monitoring was available for most of voriconazole patients (n = 11/13, 85%), and for one third of patient treated by posaconazole (n = 1/3, 33%). Median voriconazole levels were in the therapeutic range (1.8 mg/l, IQR: 0.9–4.7 mg/l) even if 2 patients were overdosed, and 3 underdosed. About posaconazole, the only dosage we have was under the therapeutic recommended level (0.88 mg/l). All treated patients completed at least 6 months of antifungal therapy. The median duration of antifungal treatment was 6 months (IQR: 6.0 –6.0). The median serum creatinine level was steady, from 163.0 μmol/l (IQR: 135.5–207.5 μmol/l) at diagnosis to 164.0 μmol/l (IQR: 134.0–204.0 μmol/l) at the end of antifungal therapy. At the end of follow-up of 47 months in median (IQR: 35.5–73.3), the median creatinine level was 238.0 μmol/l (IQR: 127.0–322.0). Renal function worsened by a median of 45.5 μmol/l (IQR: 0.0–112.5), corresponding to a mean increase of 27.9% between diagnosis and end of the study. We found no statistical difference between the renal function at the beginning and at the end of the antifungal therapy (P = .670), and between the beginning of the treatment and the end of follow-up (P = .098). Among all patients, six (35%) had a supratherapeutic level of tacrolimus at the beginning of antifungal therapy. Among those with impaired renal function (n = 14/17, 82%), four patients had tacrolimus supratherapeutic levels. Details are shown in Table 1. Fisher test was performed to detect any association between supratherapeutic level of Tacrolimus and renal impairment, and was not significative (P = .52 for renal impairment at the end of treatment, and P = .60 at the end of follow-up). Similarly, we found no association between the use of voriconazole and worsening of renal function. (P = .99 for renal impairment at the end of treatment, and P = .51 at the end of follow-up).

Over a median follow-up of 47 months, 18% (n = 3/17) of patients died. Details are shown in Supplementary Table 4. Overall, three patients died during follow-up; one death was directly attributable to disseminated phaeohyphomycosis, while other deaths occurred in a context of severe comorbidity/immunosuppression and were considered non-attributable.

P. cyclothyrioides infections

Characteristics of patients with P. cyclothyroides infection are detailed in Supplementary Table 5. They were significatively older, with a median age of 71 years (IQR: 65.0–73.3, P = .018), and were mostly males (n = 5/6, 83%). PHMs appeared within a similar temporal window, in median 17.5 months (IQR 13.3–22.5). Multiple lesions were observed in 67% (4/6) of the cases, with 1 (16.7%) disseminated form. Levels of tacrolimus seemed to be higher in this group (median: 13.3 ng/ml). The treatment was surgery for 50% (n = 3/6), and 89% antifungal drugs (n = 5/6) including one treated by dual therapy by L-AmB and voriconazole, four by voriconazole, and two by posaconazole—including one who was switched from voriconazole to posaconazole due to a recurrence at month 10. Of note, among the three antifungal susceptibility tests available for P. cyclothyrioides, two reported MICs of 4 and 1 mg/l for L-AmB, respectively. Both had MECs > 4 mg/l for caspofungin while conserving a MEC of 0.5 mg/l for micafungin. The third isolate had MIC of 0.25 mg/l for L-AmB, and MECs of 0.5 mg/l for caspofungin and of 0.25 mg/l for micafungin. MIC ranges for azoles were low (< 0.016–0.03 mg/l for itraconazole and posaconazole, 0.06–0.25 mg/l for voriconazole and isavuconazole) for all three isolates.

Patients were followed up for a median of 38 months (IQR: 36.0 –46.0) months in median, as compared with 47 months of follow-up of the whole cohort. Outcomes were less favorable, with 17% (n = 1/6) of patients experiencing relapse, 17% (n = 1/6) recurrence, and this species was significantly associated with death (n = 3/6, 50%, P = .025).

Discussion

In this 10-year, single-center cohort from the only kidney transplant center in the French West Indies, we found that PHM were not exceptional in KTR, with an incidence increasing globally^1,15—Supplementary Table 6. It is, to our knowledge, the largest study in the French West Indies. PHM, often under-recognized by transplant teams, are among the most frequently encountered fungal infections in KTR.² Clinicians should consider the possibility of invasive mycoses in KTR presenting with chronic cutaneous or subcutaneous nodules.

The cumulated episode-based incidence of 5.4% for PHM in KTR in our cohort is clearly higher than the typically reported incidence of around 0.7% in other studies.^4,15 Schieffelin et al. found an incidence of 3.6% among kidney/pancreas transplant patients.¹⁵ This might be explained by the relatively high incidence of melanized fungi in tropical regions such as the Caribbean area.^27,28 Although the median time to onset of PHM in our study was 15 months post-transplantation, aligning with other studies,^2,16,29 patients were relatively old, with a median age of 62 years compared to 52 years in other cohorts. .^9,16 This trend of older age is commonly observed among KTR, as compared to recipients of other solid organ transplant (SOT) recipients in the literature; and may also be attributed to regional variations in demographics, healthcare access and availability.

In terms of diagnostic methods, direct examination of purulent fluid was positive in 76% of our cases, slightly less than the study by He et al., where the sensitivity reached 89%.⁷ This underscores the reliability of direct examination as a primary diagnostic tool for PHM: the high sensitivity of direct examinations emphasizes the importance of this simple diagnostic method. Disseminated forms were relatively uncommon in our study, occurring in only 6% of cases (1/17), compared to 55% in the TRANSNET study, which included both SOT recipients and hematology patients.¹⁶ This difference may be attributed to the specific immunosuppressive regimens used in KTR, compared to neutropenic patients. In the study by Schieffelin et al., including only SOT recipients, dissemination was reported in only 1/27 patients (3.7%). Neurological involvement was not systematically assessed by cerebral imaging in our patients, potentially leading to underestimation of disseminated disease prevalence; but PHM typically present in KTR as localized skin infections.³⁰ The risk of disseminated infection is higher in neutropenic hematology patients, highlighting the influence of different immunosuppressive regimens on patient outcomes.^31–33 The overall mortality rate in our study was 18% which is similar to the 18.5% reported by Schieffelin et al. in SOT recipients, underscoring the need for prompt diagnosis and aggressive management of PHM.

The pathogenesis of PHM in the absence of traumatic inoculation, involve fungal dormancy in subcutaneous tissues with opportunistic multiplication when host immunity is compromised by immunosuppression or diabetes.^32,33 The relatively high incidence of PHM observed in our study may be related to the large proportion of diabetic patients (59%), compared to 24% in Revankar’s study.¹⁹ Diabetes mellitus is a recognized risk factor for severe infections.¹⁹ The interplay between diabetes and immunosuppression likely facilitates fungal colonization and infection.

A major originality of our cohort was the species distribution. Twelve different species were identified, with Paraconiothyrium cyclothyrioides being the most frequently isolated organism (one-third of episodes) ³⁴ It underscores the importance of region-specific studies and guide appropriate treatment strategies. Conversely, Alternaria spp., often described in European series, was not identified.^35–39 In 2003, Desbois-Nogard et al found minimal Alternaria spores but detected some fungi such as Cladosporium spp., Curvularia spp., and Fusarium spp. in Martinique, < 200 km south of Guadeloupe, highlighting regional differences in airborne fungal spores.⁴⁰ Advances in diagnostic and molecular identification techniques may also account for discrepancies in species identification. The genus Paraconiothyrium was proposed in 2004 to accommodate fungal species that, because of their unique morphological and genetic traits, could not be reliably assigned to any of the existing genera, such as Coniothyrium, Microsphaeropsis, Cyclothyrium, or Cytoplea.^41,42 Several new species were proposed, including P. cyclothyrioides. This species, known for causing mainly cutaneous infections, is identified in tropical regions.^1,43 Previous reports have described P. cyclothyrioides primarily in geographic areas extending from Texas to the Caribbean, with cases also reported in La Réunion.^44–50 A study from the NRCMA recorded three cases in Martinique, Guadeloupe, and La Réunion.⁵¹ Lo Porto et al.² reported two cases of infection caused by P. cyclothyrioides, including the case described by Colombier et al.⁹ Nevertheless, data on this pathogen are scarce, and it may explain why it was not describe in the European recommendations. To date, P. cyclothyrioides yields only nine results on PubMed, our series is the most important one concerning this pathogen. In a study on Coelomycetes in clinical practice, published by the NRCMA in 2019, P. cyclothyrioides belonged to the four species most frequently identified in France and Spain. ⁵² Finally, some evidence suggest that the clinical susceptibility to voriconazole cannot be assumed to be universal.⁴⁶

Our findings highlight the variability in antifungal susceptibilities across species and strains, underscoring the necessity of performing antifungal susceptibility testing. The minimum inhibitory concentrations of these filamentous fungi vary significantly.⁹ The European Society of Clinical Microbiology and Infectious Diseases guidelines recommend itraconazole or voriconazole as first-line treatments due to the CNS penetration capabilities.⁵³ Data are based on few clinical observations, and biological experiments, due to rarity of these infections, and P. cyclothyrioides, the main fungi found in this study, is not yet described in these current guidelines. Finally interpreting antifungal susceptibility testing and MICs remains challenging due to the limited data available for many of these rare fungal species. Azoles were predominantly used in the TRANSNET study as well.¹⁶ However, in vitro sensitivity does not always translate into clinical efficacy, necessitating further research into new antifungal agents and combination therapies. In our study, only one severe case received dual therapy.⁵⁴ The limited use of dual therapy may reflect challenges in managing drug interactions and toxicity, particularly in transplant recipients.

The absence of established clinical breakpoints for dematiaceous fungi is a well-known limitation in clinical mycology. Nevertheless, the results of the sensitivity of P. cyclothyrioides to azoles in this article are identical to those found in the literature, ⁵² or even increased in vitro sensitivity to azoles .⁴⁴ A 2007 american article describes Exophiala species as all having approximately the same in vitro sensitivity ⁵⁵ except for E. oligosperma which has a higher MIC to voriconazole, and amphotericin B resistance in three isolates of E. attenuata. Overall, the MICs of Exophiala to azoles are comparable to those found in the literature,.⁵⁶ Numerous studies have been carried out in an attempt to explore post-antifungal effects ⁵⁷ or synergies ,⁵⁸ with no impact in clinical practice. In the case of N. dimidiatum, although these are individual results, there was increased resistance to AmphoB and posaconazole .⁴⁴ Overall, we observed no clear relationship between in vitro susceptibility (MIC/MEC) and clinical outcome: some isolates with low MICs were associated with unfavorable outcomes, whereas other isolates with higher MICs had favorable outcomes.

Regarding fungal markers, the BDG test was highly positive (> 500 pg/ml) in the only tested patient in our study. However, there is limited data in the literature regarding its utility in diagnosing PHM: Sakata’s study reported a negative BDG at diagnosis, with levels rising post-treatment.⁵⁴ Similarly, the galactomannan antigen is not considered a reliable marker for PHM and is seldom used.^9,59 Revankar’s 2017 prospective study highlighted the limited value of these markers in clinical practice for PHM, ¹⁹ while recent data shows that BDG levels can be elevated up to 100% in certain cases of PHM.^18,60 Serum fungal markers (e.g., β-D-glucan or galactomannan) may be supportive in selected cases but are neither specific nor consistently sensitive for PHM; therefore, diagnosis should primarily rely on direct mycological evidence from tissue or fluid samples and species-level identification when possible. In practice, these markers may be more helpful for differential diagnosis or for detecting co-infections than for confirming PHM.

Renal function, although not frequently discussed in PHM literature, is a critical marker of survival in renal transplant patients.⁶¹ Co-administration of voriconazole or posaconazole can increase tacrolimus levels, leading to nephrotoxicity.³¹ In our study, a median 27.9% increase in creatinine levels was observed at the end of follow-up; with tacrolimus supratherapeutic level in 35% of cases during PHM treatments. It is important to carefully monitoring of plasma tacrolimus concentrations, and adjusting dosages. Exploring therapeutic alternatives, such as topical treatments, could be beneficial.⁶² Surgical intervention also plays a crucial role in managing PHM, as demonstrated by its success in curing immunocompromised patients in our cohort.^63,64 This worsening of renal function could also be linked to the graft’s natural history: the mean decline is approximately -1.1 ml/min/year in renal transplant recipients. ⁶⁵

This study has several limitations, inherent to its retrospective, single-center design and the small number of episodes. Case ascertainment may have been incomplete because diagnostic practices evolved over time and some follow-up occurred outside our center; however, the insular setting and centralized transplant care likely improved case capture. Finally, diagnostic criteria and therapeutic strategies were not standardized. Multicenter studies are needed to refine incidence estimates and to better define optimal management, including the role of advanced molecular diagnostics and combination therapy

Conclusion

In this single-center Caribbean cohort, we found a significantly higher cumulative episode-based incidence rate of 5.4% of PHM in this KTR population, compared to the global average of 0.7%. They were predominantly localized cutaneous/subcutaneous infections. Most patients had favorable outcomes with surgical source control when feasible and azole therapy, provided close management of drug–drug interactions with calcineurin inhibitors. A distinct regional pattern of fungal species was observed, with a predominance of Paraconiothyrium cyclothyrioides, supporting the need for locally adapted diagnostic and treatment strategies.

Author contributions

Caroline Thomas (Conceptualization, Data curation, Formal analysis, Methodology, Software, Writing—original draft, Writing—review & editing), Charlotte Boullé (Conceptualization, Methodology, Supervision, Validation, Writing—review & editing), Pierre Abel Claude (Conceptualization, Data curation, Investigation, Methodology), Joelle Claudeon (Investigation, Methodology, Project administration), Yann Cossard (Investigation, Project administration, Supervision, Validation), Sandra Devatine (Investigation), Julie Antonowicz (Investigation), Muriel Nicolas (Data curation, Formal analysis, Investigation, Methodology, Supervision, Validation, Visualization), Rishma Amarsy (Formal analysis, Methodology) Samuel Markowicz (Validation, Writing—review & editing), Julien Coussement (Validation, Writing—review & editing), Kinda Schepers (Writing—review & editing), Marwan Haboub (Investigation, Methodology, Validation), Pascal Musson (Investigation, Methodology, Validation), Pascale Piednoir (Investigation, Methodology, Validation), Fanny Lanternier (Investigation, Methodology, Supervision, Validation), Loïc Epelboin (Investigation, Methodology, Project administration, Supervision, Validation, Writing—review & editing), Dea Garcia-Hermoso (Data curation, Investigation, Methodology, Project administration, Validation), and Laurene Tardieu (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Supervision, Validation, Writing—review & editing).

Funding

This research received no external funding. Identification and susceptibility testing were funded by the CNRMA (Centre National de Référence des Mycoses invasives et Antifongiques).

Declaration of interest

The authors have no relevant financial or non-financial interests to disclose.

Institutional review board

The study protocol was approved by the Guadeloupe University Hospital Ethics Committee (reference: A65_21_09_13_TREIN) and complies with the requirements of the Commission Nationale Informatique et Libertés (CNIL), under the European General Data Protection Regulation, as attested by the delegate for data protection (MR-004, 09/13/21).

Informed consent

Informed consent was waived due to the retrospective nature of the study and the use of anonymized patient data, in accordance with ethical guidelines and regulations.