Sporotrichosis in the Highlands of Madagascar, 2013–2017

Sporotrichosis is a saprozoonotic fungal infection found mostly in tropical and subtropical areas. Few case reports in Madagascar have been published. To document sporotrichosis epidemiology in Madagascar, we conducted a cross-sectional study. During March 2013–June 2017, we recruited from select hospitals in Madagascar patients with chronic cutaneous lesions suggestive of dermatomycosis. Sporotrichosis was diagnosed for 63 (42.5%) of 148 patients. All but 1 patient came from the central highlands, where the prevalence was 0.21 cases/100,000 inhabitants. Frequency was high (64.7%) among patients <18 years of age. Sporotrichosis was diagnosed for 73.8% of patients with arm lesions, 32.3% with leg lesions, and 15.4% with lesions at other sites. Molecular identification identified 53 Sporothrix schenckii isolates. Among the 32 patients who were followed up, response to itraconazole was complete or major for 15 and minor for 17. Overall, endemicity of sporotrichosis in Madagascar was high, concentrated in the highlands.

S porotrichosis is a chronic fungal infection of humans and animals, found mostly in tropical and subtropical regions. The causal fungi develop in the soil or on plants and infect mammals through wounds, either directly (wounds from spiky plants or thorns) or through contact with contaminated soil or infected animals. Thus, sporotrichosis is a so-called implantation mycosis, affecting principally rural populations, particularly those who work with bare hands and feet (1)(2)(3). The disease is caused by a dimorphic fungus of the genus Sporothrix. These fungi display a high degree of genomic diversity, leading to the description of at least 6 cryptic species: S. schenckii, S. brasiliensis, S. globosa, S. luriei, S. mexicana and S. albicans (formerly S. pallida). S. mexicana, and S. albicans are mostly environmental (saprophytic and nonpathogenic) (4)(5)(6)(7). The infection generally occurs as a lymphocutaneous form with an ulcerated subcutaneous nodule at the inoculation site and similar secondary lesions arising along the lymphatic route (1,5). Mucosal or primary pulmonary forms are less common (8). Some cases occur as disseminated forms with multiorgan involvement, most notably in HIV-infected persons (9,10).
Sporotrichosis is widespread throughout the world; several areas of known hyperendemicity are Brazil, Mexico, Peru, and China. Outbreaks from various environmental sources, involving thousands of persons, have been reported (1)(2)(3)11,12). In Brazil, a large zoonotic outbreak associated with cats is ongoing; it has been suggested that a strain of S. brasiliensis with enhanced virulence is involved (13)(14)(15).
In Madagascar, no epidemiologic data are available for evaluation of the sporotrichosis burden. Dermatologists and infectious disease specialists have reported encountering a large number of suspected cases during their routine medical consultations; however, the cases have not been biologically confirmed. In 2007, the dermatology department of Antananarivo University Hospital in the capital of Antananarivo confirmed a series of cases and reported 1 case (16). Since 2013, we conducted a cross-sectional study to document the current epidemiology of this fungal infection in Madagascar. To diagnose, confirm, and identify the fungal species responsible, we used conventional mycology and molecular biology methods, including matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry. We describe the average annual prevalence and clinical presentation of sporotrichosis in Madagascar, along with patient outcomes. We also report the species-level identification, genetic relatedness, and antifungal susceptibility of the clinical isolates.

Study Design and Patient Recruitment
We performed a cross-sectional study of patients with clinically suspected sporotrichosis or another chronic dermatomycosis (i.e., unique or multiple nodular, budding, wart-like, or plaque-like skin lesions following a lymphatic vessel, with or without ulceration, enduring for >1 month). Patients were recruited during March 2013-June 2017 (4 years and 3 months, hereafter referred to as a 4-year period) from the Dermatology Department of the Joseph Raseta Befelatanana University Hospital (CHUJRB) in Antananarivo or during advanced consultation campaigns in regional hospitals ( Figure 1, panel A). These campaigns were preceded by announcements made via radio, social media, and posters that invited patients with cutaneous or subcutaneous lesions to come to these consultations. Healthcare providers completed clinical and demographic information forms that asked about the patient's age, sex, and occupation; the probable area of contamination; anatomic location and appearance of the lesions; and treatments received. We excluded from the study 2 patients for whom this information could not be obtained. The study was approved by the Ethics Committee for Biomedical Research of the Ministry of Public Health of Madagascar (authorization no. 66-MSANP/CE). After sampling,

Statistical Methods
We compared sporotrichosis cases and other nonsporotrichosis cases by using χ 2 or Fisher exact tests for qualitative variables and Student t tests for quantitative variables and a χ 2 test for trend to compare annual frequencies. We estimated the average annual prevalence as the mean number of cases per year (considering the period as a 4-year period) over the mean number of inhabitants. The mean number of inhabitants was calculated from the available figures in 2013 (obtained from the National Institute of Statistics of Madagascar, https://www.instat.mg) and adjusted according to the estimate of the World Bank for demographic growth of 2.7% per year (https://donnees.banquemondiale. org/indicateur/SP.POP.GROW). In our evaluation of the prevalence in the highlands of Madagascar, we excluded the region of Haute Matsiatra because no patients were recruited from there. We analyzed data and generated maps by using EpiInfo version 7.2.2.1 (https://www.cdc.gov/ epiinfo/index.html) and RStudio version 1.0.153 (https:// www.r-project.org).

Case Definitions
We used a list of clinical, mycologic, histologic, and prognostic criteria to classify cases in this study (Table 1). Cases were identified after a monthly consultation among clinicians of the Department of Dermatology of the CHU-JRB and teams of mycologists from the Charles Mérieux Infectiology Center of Antananarivo, Madagascar, and Université Grenoble Alpes, Grenoble, France.

Clinical Samples
After obtaining patient consent, we collected specimens consisting of biopsy material, pus, or flakes of skin from all patients. All samples were sent to the laboratory of the Charles Mérieux Infectiology Center of Antananarivo, Budding yeast cells with the characteristic cigar-shaped buds observed on direct microscopic examination or histologic analysis. Direct examination of pus and/or histologic analysis showing asteroid bodies (Splendore-Hoeppli reaction). Positive culture of a fungus morphologically suggestive of Sporothrix spp. from a clinical sample without molecular or MALDI-TOF mass spectrometry confirmation. Classification Confirmed >1 of the major clinical criteria and >1 of the major mycologic criteria or 1 minor clinical criterion and >1 of the major mycologic criteria. Probable >1 of the major clinical criteria and 1 minor mycologic or histologic criterion and a complete or partial response to antifungal therapy. Possible >1 of the major clinical criteria without any (major or minor) mycologic or histologic criteria or >1 of the minor clinical criteria without any (major or minor) mycologic or histologic criteria and a complete or partial response to antifungal therapy. Clinical response to antifungal therapy Cure Complete resolution of all lesions. Major response Substantial improvement of most lesions with a substantial decrease in subcutaneous nodules.

Minor response
Mild improvement of most lesions with a smaller decrease in subcutaneous nodules than for a major response. Failure Stabilization of the lesions after >3 months of antifungal therapy or worsening of the lesions after >3 months of antifungal therapy. *ITS, internal transcribed spacer; MALDI-TOF, matrix-assisted laser desorption/ionization time-of-flight.
where they were processed either immediately or after 24 to 48 hours of storage at 2°-8°C.

Mycologic Analyses
We directly examined the clinical specimens under a microscope (5,11). The samples were used to inoculate Sabouraud medium supplemented with chloramphenicol and incubated at 30°C for 2-3 weeks. For positive cultures, we identified the fungal isolates morphologically, extracted DNA, and froze the culture at -80°C.

Molecular Analyses
We used the QIAamp DNA Blood Mini Kit (https://www. qiagen.com) according to the manufacturer's instructions for DNA purification. Colonies and biopsies were crushed before processing. PCR amplification was performed in 2 steps. The first step comprised 2 panfungal PCRs targeting internal transcribed spacer (ITS) regions with the primers ITS1/ITS4 and D1D2 with the primers NL-1/NL-4 and NL-3/NL-4 (17)(18)(19). The second step was a specific S. schenckii PCR targeting the topoisomerase II gene with SSHF31/SSHR97 primers (20). We sequenced panfungal PCR products by LGC Genomics GmbH (https://www.biosearchtech.com) by using the same primers and aligned the sequences obtained with the reference sequences in the International Society of Human and Animal Mycology (ISHAM) Barcoding Database (http://its.mycologylab.org) (17) for the ITS region and the National Center for Biotechnology Information database for the D1D2 and ITS regions. We constructed the phylogenetic tree by using MEGA7 software (https://www.megasoftware.net).

MALDI-TOF Mass Spectrometry Analysis
We created a main spectrum profile (MSP) in-house  (22).
Patients with sporotrichosis were younger than other patients (38 vs. 43 years of age), although this difference was not statistically significant (p = 0.10). Analysis by age group showed that this trend was linked to a high frequency of sporotrichosis in persons <18 years of age (64.7%) compared with the rest of the population (39.7%; p = 0.08).
The location of lesions differed significantly between sporotrichosis patients and other patients ( Table 2). The principal site affected was the arms (49.2%) for patients with sporotrichosis. Sporotrichosis was diagnosed more frequently for patients with lesions on the arm (73.8%) than for patients with lesions on the leg (32.3%) or other body sites (15.4%) (p<0.0001; Table 2). The sporotrichosis lesions had been present for <1 year for 71% of patients, 1-2 years for 14.5% of patients, and >2 years for 14.5% of patients. The most frequent type of lesion for sporotrichosis patients was lymphocutaneous (69.3%). The other forms were characterized by vegetative or verrucous lesions, infiltrated plaques, or tuberous lesions (Figure 2). We observed no fixed cutaneous forms or extracutaneous forms.
Sporotrichosis patients were predominantly farmers (52.4%), but a large number were craftsmen and tradesmen ( Table 2). Sporotrichosis was diagnosed for 71.4% of the craftsmen and tradesmen and 39% of patients in other professions grouped together (p = 0.04).

Prevalence and Geographic Distribution of Sporotrichosis
We determined the geographic origin of patients with sporotrichosis, corresponding to the presumed origin of contamination ( Figure 1, panel C). The concentration of sporotrichosis cases in the highlands was very high; almost all patients with sporotrichosis (n = 62, 98.4%) originated from these areas. The frequency of sporotrichosis cases was much higher in the highlands (62/90, 68.9%) than on the rest of the island (1/58, 1.7%; p<0.0001). In all highland areas, the frequency of sporotrichosis was similar and very high (66.6%-100%; p = 1.00) ( Table 2). The average annual prevalence of sporotrichosis on the high plateaus of the Analamanga, Amoron'i Mania, Bongolava, Itasy, and Vakinankaratra regions was evaluated at 0.21 cases/100,000 inhabitants. Prevalence was highest in the Analamanga (0.27/100,000 inhabitants) and Amoron'i Mania (0.25/100,000 inhabitants) regions (Table 3; Figure  1, panel C).

Culture Results
We obtained 172 cultures, including 72 established with the samples of the 63 sporotrichosis patients. Overall, macroscopic and microscopic examination exhibited morphologic features consistent with Sporothrix spp. for 90.2% (65/72) of the cultures.

Molecular Results
Sensitivities for the 2 panfungal PCRs, for D1D2 and ITS, were lower for clinical specimens than for cultures, and the ITS PCR was less sensitive than the D1D2 PCR for clinical specimens and cultures (Appendix 2 Figure 1, https:// wwwnc.cdc.gov/EID/article/25/10/19-0700-App2.pdf). The specific S. schenckii topoisomerase II PCR was unable to confirm identification for any of the clinical specimens, whereas its sensitivity for cultures was 89.2% (58/65), with features suggestive of Sporothrix spp.
The alignment of the ITS sequences from cultures identified 13 isolates as S. schenckii, according to data from the ISHAM database (Appendix 1 Table 1) (17). The phylogenetic analysis showed that the isolates from patients with sporotrichosis in our study were grouped in the S. schenckii clade, together with clinical strains from other regions of the world (Figure 3; Appendix 2 Figure  2). Our results confirm that ITS sequencing is suitable for separating the cryptic species of clinical importance from the strictly environmental ones (25).
The MSP in-house Sporothrix library performed well for S. schenckii identification; the score for 87.9% (29/33) of the strains was >2 and for 4 strains was >1.8. Unfortunately, 2 strains were contaminated by Candida spp. and could not be identified. For sporotrichosis identification, in-house MSPs were systematically the first choice in the list of MSPs, confirming their superiority for identification at the species level over the 2 MSPs in the Bruker database and the 1 MSP in the NIH database. Comparison of the spectra obtained by using the MSP in-house Sporothrix library with those obtained from identification to the external MSI platform indicated that the most likely identification was S. schenckii (Appendix 1 Table 2). Percentages of similarities were consistent with accurate identification to the species level (>20%) for 94.1% (48/51) (Appendix 1 Table  1); only 3 strains were not formally identified.
Taking together all results of the molecular analyses, we identified 53 S. schenckii strains: 51 by MALDI-TOF mass spectrometry, of which 46 were identified also by the specific S. schenckii topoisomerase II PCR and 13 were   identified also by ITS sequencing. The 2 contaminated isolates not identified by MALDI-TOF mass spectrometry were confirmed by the specific S. schenckii topoisomerase II PCR.

Susceptibility of Sporothrix schenckii to Antifungal Drugs and Patient Outcomes
A total of 46 S. schenckii isolates were culturable after thawing for MIC determination. We determined MICs and their geometric means for the 5 antifungal drugs tested ( All 63 patients received treatment, but only 32 were monitored for >2 months after treatment began (23 with lymphocutaneous disease and 9 with minor forms of disease). Among those patients, the response was complete or major for 15 (47%) after 4-7 months of treatment and minor for 17 (53%) after prolonged (9 months) treatment. The remaining 31 patients received treatment for 2 months and then did not return for follow-up visits.

Discussion
This study provides recent epidemiologic data for sporotrichosis in Madagascar. We detected numerous cases and substantial endemicity despite previous reports of only sporadic cases or small series (16,26). We estimated an average annual prevalence in the highlands of 0.21 cases/100,000 inhabitants; 98% of the cases were concentrated in that area. Among sporotrichosis patients in Madagascar, we highlight the high infection risk for young persons (<18 years of age) and the particularly high frequency of lesions on the arms. On the basis of this study, we were able to develop and routinely implement molecular analyses in Madagascar, enabling positive identification of S. schenckii for all confirmed cases.
The high estimated prevalence (0.27 and 0.25/100,000 inhabitants) in 2 highland regions (Analamanga and Amoron'i Mania) reveals the high transmission rates in this part of the island. To date, the only published series of sporotrichosis cases in Madagascar have described disease-endemic foci in the highlands, particularly in the Analamanga region (16), but the concentration of cases in the highlands that we observed was unexpected and striking. This almost exclusive distribution may be the result of climate conditions in this region, which differ from those on the rest of the island. This region has a tropical climate, with a mean temperature of 19.5°C and substantial rainfall, which probably favors development of fungi on plants and in the soil. A phylogeographic study focusing strictly on S. schenckii showed that this species was present in temperate (United States), hot and humid (South Africa, Australia, Colombia, and Venezuela), hot and dry (Australia and Uruguay), cool and dry (Peru and South Africa), and cool and humid (Uruguay) zones (27). Findings of that study therefore seem to go against the notion of a single climatic factor. The frequency of sporotrichosis in some areas of the island to the west and southwest are unknown because these areas were not investigated; thus, sporotrichosis may not be totally distributed in the highlands. In addition, some cases could have been missed because of our mode of study recruitment, the low incomes of people living in remote areas, and the limited development of the healthcare system. The concentration of sporotrichosis in the highlands is probably not the result of better access to healthcare facilities because patients with other diagnoses most frequently do not live in the highlands (p<0.0001) ( Table 2). However, better access to medical care in the highlands does partly explain the sporotrichosis diagnoses made relatively early in the course of disease (71% in the first year after onset).
Patients found it difficult to remember when their lesions had appeared and to associate them with a particular injury or activity; however, our survey revealed considerable involvement in rural activities: farming (rice, cassava, corn), logging, trade, and craftsmanship. Not only artisans are exposed to contamination through manual work; tradespeople are also exposed because they practice activities other than selling for living. Contamination by activities associated with the manual production of charcoal and cutting wood for cooking and heating seems likely on the basis of the predominance of arm lesions, the concentration of the disease in the coldest region of the island, and the high frequency of infections among children (who practice these activities). In northeastern China, a risk associated with the use of  wood or other fuel has been proposed as an explanation for transmission patterns (28); the authors of that study thought that contamination occurred via maize stalks (where S. globosa has been found) used for heating and cooking. They hypothesized that the fermentation of the plants promotes the development of yeast forms of the fungus, increasing the risk for contamination during transport and storage at home. Other possible sources of contamination, such as soil or decaying plant material, are also possible in Madagascar. Neither we nor J.F. Carod et al. (16) observed any cases of zoonotic transmission; the identification of S. schenckii alone confirms the hypothesis of contamination by soil and plants.
The molecular methods that we developed in this study, including MALDI-TOF mass spectrometry, made it possible to confirm cases and to identify the species responsible (21,29). We found that it was easier to amplify the D1D2 domain (LSU) and that the amplicons obtained were easier to analyze by sequencing than were those of the ITS domain. However, the availability of a database with many verified ITS sequences and the more polymorphic and discriminant nature of these sequences makes them more suitable for cryptic species identification and phylogenetic analysis (17,25).
We added a rapid and inexpensive mass spectrometry identification approach to the molecular tools for identifying S. schenckii to the cryptic species level. Our results obtained by using a Bruker instrument confirm previous analyses performed with a Shimadzu instrument (29). The MSP in-house Sporothrix library yielded better S. schenckii identification scores than did the 3 preexisting MSPs. The excellent identification scores and the external validation with another mass spectrometry platform showed that S. schenckii identification at the species level with MALDI-TOF mass spectrometry is accurate and adapted for routine diagnoses in clinical laboratories.
In conclusion, our study reveals substantial endemicity of sporotrichosis in Madagascar. Sporotrichosis was particularly concentrated in the highlands, which have climate, vegetation, and lifestyle conditions that favor the development and transmission of the fungus. Using molecular methods and MALDI-TOF mass spectrometry, we were able to identify S. schenckii as the species responsible for sporotrichosis in Madagascar. Despite its high frequency, sporotrichosis remains neglected in Madagascar.

About the Author
Dr. Rasamoelina is a scientist in the Centre d'Infectiologie Charles Mérieux in Antananarivo, Madagascar. His research interest is the laboratory diagnosis of tropical mycoses.