Podoconiosis: Clinical spectrum and microscopic presentations

Background Podoconiosis is a skin Neglected Tropical Disease (skin NTD) that causes lymphoedema, and affects barefooted subsistence farmers in some tropical countries. The clinical presentation and histopathologic correlates of podoconiosis have been understudied. Here, we systematically document the clinical and histopathologic spectrum of podoconiosis. Methods This is a cross-sectional study in Durbete, Ethiopia from February 2018 to October 2019. Dermatologists performed a patient history, physical examination, filariasis test strip, and skin biopsy for histopathologic examination. The results were summarised and a descriptive statistical analysis and Wilcoxon rank sum test with continuity correction was done. Results We recruited 289 patients for the study, 178 (61.6%) had stage 1 or 2 podoconiosis, and 111(38.4%) stage 3 to 5 podoconiosis. 188 (64.1%) had a family history of podoconiosis. In 251 (86.9%) patients, both legs were affected by podoconiosis and in 38 (13.1%) only one leg was affected. 220 (77.5%) patients had warty lesions, 114 (39.4%) had nodules. The median number of episodes of Acute Dermato-Lymphangio-Adenitis (ADLA) reported by the patients in the last three months was 2 (interquartile range (IQR) 1–4). Increased episodes of ADLA were significantly associated with stage 3–5 podoconiosis (P = 0.002), while burning pain in the feet was more common in stage 1 or 2 podoconiosis. Stage 3–5 disease was histopathologically characterised by epidermal and dermal thickening, verrucous acanthosis, inflammatory cell infiltrates (predominantly lymphoplasmacytic), dilated and ectatic and a reduced number of lymphatic vessels, eccrine ductal hyperplasia, and sclerosis such as thickened collagen bundles. Conclusion We provide a detailed description of the different clinical patterns, associated clinical findings and the histopathologic spectrum of podoconiosis at different stages of the disease. Our observations should serve as a guide to classifying patients with podoconiosis for prognostic assessment and treatment decision.


Introduction
Podoconiosis is a neglected tropical disease (NTD) affecting barefooted subsistence farmers mainly in tropical countries of Africa, Southeast Asia, Central and South America. Globally the disease affects about 4 million peoples. Ethiopia is the most affected country with an estimated 1.5 million cases [1,2]. It causes chronic non-filarial lymphoedema that has a profound and negative impact on numerous health-related, psychosocial and economic aspects of the lives of affected patients and communities in endemic regions [3][4][5].
The aetiology of podoconiosis is not known, yet based on the geographic distribution, the affected communities, and its clinical presentation, podoconiosis is likely caused by long-term exposure to red clay soil derived from volcanic rock in rural highland areas [3].
Podoconiosis is a localised disease, primarily starting from the feet, progressively involving the legs, yet in most cases remaining below the knee. Clinically the disease severity is classified in 5 stages (S1 Table) [6]. In stages 1 and 2, simple lymphoedema management can completely reverse the clinical changes, but if the condition is not managed with appropriate foot care and consistent shoe wearing, it can lead to subsequent irreversible stages with progressive swelling and serious disfigurement [3]. Over half of the patients report stigma, impaired mobility and daily activity, resulting in considerable morbidity and limitation of economic performance [7].
Over the past half a century many aspects of the disease epidemiology, inheritance, psychosocial impact and therapeutic management have been studied [7][8][9][10][11][12][13]. Yet, clinical studies on aetiology, histopathology, and the systematic documentation of the course of the disease by trained physicians are sparse. The systematic description of clinical findings and the corresponding histopathological characteristics are essential to further clarify the aetiology and pathophysiology of podoconiosis. To the best of our knowledge, to date only three studies investigated the histopathology of podoconiosis, of which the first is limited to specific morphologic descriptions while the other two are brief case reports [13][14][15].
Here, we examined the clinical presentation and associated complications of patients with podoconiosis, and their histopathologic correlates in the North of Ethiopia. The aim of this study is a systematic documentation of the clinical spectrum and histopathology of podoconiosis to advance patient management, therapeutic follow-up, prioritisation of resources, epidemiological surveillance and to further future research to improve the quality of life of patients with podoconiosis.

Ethical considerations
The study was conducted according to the principles of the Declaration of Helsinki and approval was granted from The National Research Ethics Review Committee Study information sheets were read to the patients in the local language (Amharic). The patients were then asked to sign a consent form written in Amharic. Patient names and identifiable information were removed, and only aggregate non-identifiable data were used for the publication. The electronic databases with patient identifiable information were kept on the password protected computer of the principal investigator.

Study design
This is a cross-sectional study of the clinical features and histopathology of podoconiosis patients at the Durbete Podoconiosis Prevention and Treatment Centre in Durbete, Ethiopia from February 2018 to October 2019. The reporting is in line with the "Strengthening the Reporting of Observational Studies in Epidemiology" (STROBE) guidelines [16].
town. The centre manages 400 to 600 podoconiosis patients annually and supports the local podoconiosis patient association by awareness raising, shoe distribution and community education.

Study population and procedure
We invited all new patients who had not previously received lymphoedema morbidity management. We recorded relevant patient history, examined the patient (with a focus on the feet and lower legs), and staged the disease according to a validated podoconiosis staging system (S1 Table) [4]. To standardise the recording, we used written case definitions for type of lymphoedema, mossy foot, ADLA and signs of fungal infection (S1 Text  [23]. For economic reasons and considering the inconvenience of invasive procedures, we performed a Filariasis Strip Test (FST), and a 6 mm tissue sample from the dorsum of foot of every fourth patient who consented to these procedures (S2 Text). In two previous studies on surgical wide excision (nodulectomy) for podoconiosis, we documented that wound healing was satisfactory, also in patients with lymphoedema. We therefore considered that the risk of complications such as wound infection, ADLA, or hyperpigmentation following 6 mm skin punch biopsy would be low [22,24]. Specimens were stained with haematoxylin and eosin (HE) and Periodic Acid-Schiff (PAS) procedures. WE and a general pathologist (BA) performed the microscopic evaluation of each slide. Discrepancies were resolved by a dermatopathologist (ABA). Information on the histopathological presentation of the epidermis, dermis, lympho-vasculature, and inflammatory cell infiltration was documented, and the presence of double breaking foreign body material was ruled out using polarised light microscopy.

Data variables
Patient history and clinical findings (S2 Text) were recorded using a standardised questionnaire and pictures were taken. Two dermatologists independently reviewed the photographs from suspected lesions for most of the patients, and consensus was reached on the cutaneous conditions and type of lymphoedema.

Data collection and statistical analysis
Blood was collected and Filarial antigen tests (FTS) (BinaxNOW Filariasis card test, India) were performed by a trained laboratory technician. Tissue samples for biopsy were collected either by a dermatologist or a surgeon (S2 Text). Patient history and clinical data were entered into Microsoft Excel database by a dedicated person and cross-validated by the main investigator.
We first summarised sociodemographic characteristics, medical history, and clinical findings for the entire patient population using descriptive statistics. Next, we summarised histopathological findings for the subgroup of patients with available results. Finally, we assessed the association between clinical characteristics and podoconiosis-related complications, staging (stage 1&2 versus 3-5), and histopathologic features and expressed the strength of association with odds ratios and 95% confidence intervals, and using the Wilcoxon rank sum test with continuity correction. All statistical analyses were done with R version 3.5.0 (R Foundation for Statistical Computing, Vienna, Austria) [24].

Socio-demographic characteristics of podoconiosis patients
In total, 289 podoconiosis patients were involved in the study and 167 (57.8%) were male. The median age was 50 (IQR 35-65) years, and 85.2% of the participants had no formal education.

Diagnostic definitions
1. Definitive podoconiosis: 3 major; or 2 major plus 2 minor; or 1 major plus 5 minor criteria 2. Probable podoconiosis: 2 major; or 1 major plus 2 minor; or 5 minor criteria Only 40.1% of the patients had visited a health facility in the past for their leg swelling (without receiving treatment). For those who previously visited a health facility for their leg swelling, the median time between disease onset and care seeking was 15 (IQR 5-28) years. More than three fourths (75.4%) of the patients were farmers or shepherds (Table 1). All the 289 patients were barefoot or wore shoes inconsistently in the years before symptoms appeared. About two-thirds of patients (65.0%) had a family history of podoconiosis.

Clinical characteristics of podoconiosis
At the time of examination, the median time patients had leg swelling was 20 (IQR 10-32) years. There was no significant difference in disease duration between patients diagnosed with mild (stage 1 & 2) versus advanced (stages 3-5) podoconiosis (p-value = 0.1) and similarly, there was no association between disease duration and the type of lymphoedema (pvalue = 0.2) (Fig 1).178 (61.6%) patients had mild stage podoconiosis while the rest had advanced stage.251 (86.9%) patients had both legs affected, while 38 (13.1%) had podoconiosis only in one leg. 104 (36%) patients had asymmetrical stages of the two legs and 51 (18.5%) also had asymmetry in the type of lymphoedema (one leg waterbag and the other fibrotic/sclerotic). A small number (1.4%) of patients had leg swelling extending above the knee. 220 (77.5%) patients had hyperkeratotic (mossy) lesions on one or both feet. 114 (39.4%) patients had nodules (Figs 1 and 2). The reported median number of episodes of ADLA in the last three months was 2 (IQR 1-4). About 100 (34.7%) patients had oozing foot lesions and 130 (45.1%) had foul-smelling feet. 248 (85.8%) patients had associated signs of fungal infection on their feet at the time of examination. Those with fungal infections had an increase incidence (OR = 7.3) of ADLA than those without (P<0.05). Increased episodes of ADLA were significantly associated with stage 3-5 disease (P< 0.05), while burning pain on the feet was much more common in stages 1 & 2 (Tables 2 and 3).

Discussion
We describe here the different clinical patterns, associated clinical presentations and histopathologic presentations of podoconiosis. Based on his outpatient clinic experience, EW Price in the 1980s described some of the clinical patterns, but this knowledge needed a timely update and validation by community based studies [1]. All the other recent studies on clinical presentations are based on case reports, lack photography or use inadequate terms or morphologic descriptions [2,[9][10][11][23][24][25][26]. Given the large number (based on an initial power calculation) and representative distribution of participants, we feel confident that the clinical and histopathological findings presented here truly reflect the podoconiosis diseases spectrum. splaying of the right forefoot, accentuation of the skin markings on the metatarsophalangeal joint. C) Stage 2: bilateral lymphoedema below the knee. Area of scarring is seen on the lower leg and foot (from traditional bloodletting). Hyperpigmented, warty hyperkeratotic papules covering the dorsum of the feet extending to the ankle. Nodules on the right forefoot. D) Stage 2: Pitting lymphoedema below the knee, with dry skin. E) Stage 2(right foot) and stage 3 (left foot). Right foot: pitting lymphoedema on the shin, few papules and nodules on the dorsum anterior one third of the foot and toes. Left foot: smooth surface with multiple nodules and tumorous masses covering the whole foot and flexural ankle with extension to the anterior distal one third of the lower leg. The foot is totally deformed and the toes are no longer visible. Based on the extension of the nodules the right foot is stage 2 (nodules below the ankle only) and the left foot stage 3 (nodules extending above the ankle). F) Stage 3: Bilateral lymphoedema. On the right leg the swelling is pitting on the proximal half of the lower leg and non-pitting fibrotic edema on the distal one third of the lower leg and foot. Scattered, infiltrative papules on the dorsal distal one third of the foot and toes. Hyperkeratotic and Nodules can occur from stage 2 onwards and 39.4% of our patients had nodules on the feet or lower legs. The programmatic implication of nodules on the foot is that most of those patients require surgical intervention [13,27,28]. In previous publications, oedema in podoconiosis was considered not to extend above the knee [12,28,29]. However, in our study 4 (1.4%) patients had swelling above the knee (stage 4 disease). Documentation from the 1990s reported that 2.4% of podoconiosis patients had above-knee lymphoedema and other recent studies, reported 5.9%-18% of patients with stage 4 podoconiosis [9,10,23].
In most studies, the standard description of podoconiosis is that it results in "progressive bilateral swelling of the lower legs" and most studies that have documented stages of podoconiosis have failed to explore on the evidence of the variation/symmetry of involvement [1,2,8,10]. This study has showed that asymmetry in leg involvement is common as far as the type of lymphoedema and hyperkeratosis between both legs involved. These findings suggest that even though podoconiosis is linked to genetic factors, the type of lymphoedema manifested is unlikely to be genetically linked. The finding also poses a question of why such significant asymmetric manifestations should have an environmental cause, if both feet are exposed equally to the same environment.
Price stated that most patients presented late [1]. The late presentation in our study also suggests that the structural and individual barriers preventing patients seeking care need to be identified and addressed. Stigma could be an important factor in late presentations. Like many other studies [30][31][32], we found a large proportion of patients to have a family history of podoconiosis. Family history of similar illness is an important factor in differentiating podoconiosis from lymphatic filariasis.
In our study, 77.5% of patients had hyperkeratotic papules in a 'slipper' distribution on the feet. Other studies reported a wide variation (38.5 to 97.95) in the clinical presentation [9,10,33,34]. From our clinical observation, hyperkeratotic papules are common in podoconiosis with no previous treatment history and among those who do not wear protective shoes consistently. The wide variation in prevalence seen across the different studies may be explained by the absence of a standard clinical definition and/or the limited clinical experience of data collectors.
Distinguishing the characteristics by disease stage can help the clinical management of podoconiosis, such as the choice of treatment option, and follow-up after treatment. Moreover, clear disease staging criteria are critical to further our understanding of podoconiosis pathophysiology, compile evidence-based prevention measures and as endpoints in the evaluation of innovative therapies. The existing podoconiosis staging criteria has been used since 2008 [6]. Considering the findings in this article, we recommend revision of the clinical staging. In the new staging criteria in addition to clarification/rephrasing the existing clinical stages we highlighted the necessary criteria for each stage, removed the redundancy and included additional clinical findings (S1 Table).
As to the prevalence of ulcers, similar inconsistencies with earlier literature were found. In our study 9.7% of the patients had an associated leg ulcer. One study in West Ethiopia documented 37.7% of patients examined to have associated ulcers [10], but another study in the same population reported 7.2% of patients to have leg ulcers [32]. An epidemiological study in Northern Ethiopia reported that 53% of patients had an ulcer [9]. depigmented rough papules and nodules covering the distal one third of the left lower leg, and dorsum of the foot. G) Stage 4 (right) and stage 2 (left). Right foot: non pitting oedematous swelling of the leg extending above the knee with an area of skin depigmentation around the ankle. Left foot: non pitting swelling of the leg below the knee. H) Stage 5: fibrotic globular swelling of the left foot and ankle with ankylosis of the ankle joint, multiple areas of scarring (from traditional bloodletting). I) Stage 5: right leg with nodules and rubbery to woody hard tumors on the dorsum of the foot with band like redundant skin on flexural ankle (pillowy oedema) and ankle fixation.
https://doi.org/10.1371/journal.pntd.0010057.g001 The most disabling problem associated with podoconiosis is ADLA, which needs emergency medical intervention and prevention strategies. About 86% of the cases in our study had at least one episode of ADLA during the three-month period preceding the date of examination. ADLA frequency was significantly higher in those with stage 3-5 podoconiosis and those with associated interdigital fungal infection. The frequency of ADLA varies in studies from 5 to 23.3 episodes per year [9][10][11] and interdigital lesions were found to be the main risk factor [34]. One factor explaining this wide variation in reported episodes may be recall bias. We tried to minimise this by asking the history over the last 3 months instead of one year. ADLA frequency data would be much more accurate if the study was prospective. More importantly, in almost all previous studies which have documented ADLA, an inadequate case definition (a reddish, hot, swollen leg with a painful groin) has been used, or the case definition is not clear [9][10][11]. In our study comprising patients with skin of colour, exfoliative dermatitis was the dominant clinical finding. Whereas most studies use erythema/red skin as hallmark sign for ADLA, this is not helpful in the dermatologic evaluations of individuals with a skin of colour [35]. The reported frequency of ADLA episodes in lymphoedema patients with podoconiosis is much more frequent than that of lymphoedema secondary to lymphatic filariasis [36] suggesting the need for public health programs to improve access to prevention and case management. Histopathologic presentations of podoconiosis have a lot of similarity to those seen in longstanding filarial lymphoedema in regard to epidermal and dermal thickening, epidermal verrucous acanthosis and sclerosis [15,[37][38]. However, in podoconiosis, infiltrates consisted predominantly of mast cells, plasma cells and lymphoplasmacytic cells, while lymphocytes predominate the infiltrates in filariasis and mast cells have not been emphasised in this condition [37]. Our previous study on the histology of nodules from podoconiosis patients demonstrated similar findings [13]. Mast cells are commonly seen in fibrotic skin disorders, and it has been hypothesised that they may be involved in the pathogenesis of fibrosis [39]. Podoconiosis also differs from filariasis in terms of vascular abnormalities. While filariasis shows dilated and tortuous lymphatics, in stages 3-5podoconiosis, the lymphatics are reduced and blood vessels are dilated, increased and often show a sclerotic wall [38][39][40][41][42]. In stage 1-2 podoconiosis, the cellular infiltrates are much more intense, probably showing a more active inflammatory process. Compared to the advanced stages of podoconiosis, in stages 1-2, the lymphatic vessels were more numerous and dilated, with indication of collateral formation. The reduction of lymphatic vessels, sclerotic blood and lymphatic vessel walls, eccrine ductal hyperplasia, thickening of collagen and sclerotic connective tissue are signs of a fully developed stage of the disease.
The strengths of this study are that each patient was evaluated by an expert dermatologist, and that the description of the clinical pattern was made based on the morphology. The clinical findings of a large number of mostly untreated patients with different stages of the disease were systematically documented and revision of the staging criteria was recommended. This is also the first study into the histopathologic substrate of the various podoconiosis stages, allowing a better understanding of the spectrum of clinical presentations and to correlate it with associated histopathology, morbidities and severity.
One limitation of this study is that patients were recruited from the treatment centre and hence only give a picture of podoconiosis patients seeking health care, and may not represent those who do not have access to or knowledge about the existence of such service. As information on the clinical history was collected retrospectively by patient interview, there may be some recall bias especially on the duration of the disease or the date of the first symptoms. The newly suggested clinical staging also needs to be field tested.
The findings of this study emphasise the need for standardisation of clinical presentations of podoconiosis, both in clinical and research settings. Second, the detailed morphologic description of the disease with histopathological correlates is expected to serve as a reference for clinicians for diagnosis and management. This information will also be important for public health in budgeting, resource allocation, and monitoring, but also in the design of awareness and prevention campaigns.

Conclusion
The clinical features of podoconiosis include bilateral but mostly asymmetrical leg lymphoedema. Clinically, symptoms of stages 1 & 2 podoconiosis are a burning sensation or/and itching of the forefoot, splaying of the forefoot, plantar oedema, hyperkeratosis and increased skin markings. Later, debilitating bilateral lymphoedema of the lower leg occurs with or without skin changes such as hyperkeratosis, nodules and 'mossy' papillomatosis. Accompanying oedema may be either soft or hard and fibrotic. Two out of five patients have multiple smooth surface or hyperkeratotic nodules or tumours dominantly on the dorsum of the foot extending to the ankle area. Other clinical variants are unilateral involvement, toe fusion, toe resorption, maceration, and ulcerative podoconiosis. The histopathologic changes demonstrate broad similarities with the lymphoedematous stage of filariasis but also have some differences in the vascular abnormalities and the type of infiltrates.
We suggest that further research includes the establishment of prospective cohorts, population based epidemiological studies, field testing of the revised staging and more in-depth histological studies to further the understanding of the pathogenesis of podoconiosis, including potential infectious associations.