Background

Tungiasis is a highly neglected tropical skin disease (NTSD) caused by the female sand flea, Tunga penetrans, which penetrate the skin, usually of the feet, of their mammalian hosts and stay embedded for their remaining life [1]. The flea grows 2000-fold in size over seven days as a result of eggs develo** in the abdomen. A small opening is maintained in the skin through which the last segments of the abdomen stay in contact with the environment. Via this opening, the male copulates with the embedded female [2] and the female expels eggs, respires, and defecates. Eggs fall to the ground and, if conditions are favorable, the larvae hatch and grow over several larval stages, pupate and emerge as adults over the course of 3–4 weeks [3]. The embedded female dies after egg-laying and is removed by skin repair mechanisms if not extracted by the host.

Tunga penetrans is endemic in the tropics of the Americas and in sub-Saharan Africa with an estimated 668 million people considered to be at risk of infection in sub-Saharan Africa alone [4, 5]. In endemic areas, tungiasis is heterogeneously distributed with the poorest part of the population bearing the highest burden [6]. In Kenya, human tungiasis is considered a significant individual and public health threat, with an estimated two million people currently infected [7] although there is no systematically collected data from national level surveillance. In general, males, children, elderly people and people with disabilities carry the highest disease burden [6] with prevalence ranging between 7 and 60% in affected villages and schools [8,9,10]. Intensity of infection is also heterogeneously distributed among the infected population, with the majority of patients having only a few embedded fleas, while a few individuals have over 100 fleas [9].

Sand fleas cause severe morbidity in humans, companion animals and livestock [11,12,13]. Morbidity results from the intense inflammatory response around the rapidly growing female sand fleas firmly embedded in the epidermis [11]. The inflammation is further intensified by frequent bacterial superinfection of the lesions and bacterial superinfection may result in tetanus, gangrene or septicemia [14].

A study conducted in Brazil 20 years ago carefully documented the clinical features of tungiasis [15]. Acute symptoms included itching, pain, edema, erythema, warmness, desquamation, ulcers and fissures. Chronic symptoms, thought to be the result of repeated infection with large numbers of fleas, included hyperkeratosis, peri-ungual hypertrophy, deformation and loss of nails [15]. The authors developed systematic methods to quantify the severity of disease with severity scores for acute tungiasis (SSAT) and chronic tungiasis (SSCT). Modified versions of these scores have also been applied to estimate morbidity in pigs and dogs and to quantify the effects of successful treatment on disease severity [16, 17].

Overall, the SSAT and SSCT have only rarely been used since they were published. One of the reasons may be that edema, erythema and warmness-to-touch, the signs of inflammation, are difficult to assess, particularly on a dark skin and by non-clinical staff. The SSAT also included scores for the number of sites with flea clusters, and the patient’s experience of pain, itching and sleep disturbance. It was demonstrated that those individuals with the highest infection rate (newly embedded live fleas over time) had the highest SSAT [15]. A separate study showed that the number of live fleas and the SSAT scores were negatively correlated with quality of life as measured using the modified Dermatological Quality of Life Index (mDQLI) [18].

Now that tungiasis has been added to the World Health Organization (WHO) list of NTDs (under scabies and other ectoparasites), governments and organizations will start to plan surveillance and intervention programs, and these will need carefully defined indicators and targets. As has been implemented for other NTDs, initially it will be appropriate to target those individuals with the highest morbidity. For instance, the WHO NTD Roadmap 2021–2030 [19] recommends targets using the prevalence of medium and high infection intensity for schistosomiasis and soil-transmitted helminths. The thresholds to define low, medium, and high intensity infection, such as eggs per gram of stool, were set by the WHO Technical Working Groups and were based on the morbidity caused by differing levels of infection intensity [19,20,21]. Similar measures and thresholds will be needed for tungiasis.

To date, disease intensity for tungiasis has been classified in three-tiers, defined using embedded flea counts; mild cases being defined as having 1–5 fleas, moderate cases having 6–30 fleas and severe cases more than 30 fleas [9, 10, 22]. However, the first studies describing this classification did not explain the reasoning behind the thresholds and they have not been associated with levels of morbidity. As has happened for some other NTDs whose targets focus on medium and high intensity disease [23], i.e. reducing the classification from three groups to two, it may be more appropriate to define two disease groups for tungiasis rather than three.

Consequently, in the current study we set out to (i) describe the prevalence and intensity of infection in two regions of Kenya with suspected high disease burden; (ii) simplify the clinical scoring method; (iii) test the use of thermography to assess inflammation; (iv) evaluate the past disease severity classification with respect to symptoms in these populations in Kenya and develop a new classification with only two disease severity groups. The data reported here is part of a larger project aimed at characterizing the disease ecology of tungiasis in East Africa, including a better understanding of the parasite, risk factors for severe tungiasis and its impact on child development and well-being.

Methods

Study design

A cross-sectional observation study of children in primary schools, their families, and homesteads in two regions of Kenya was carried out between February 2020 and April 2021. Primary school children between the age of 8 and 14 years were selected since this is the age group most affected by tungiasis in Kenya [22].

Study area and population

Surveys, following the same procedures, were conducted in the sub-counties of Matuga and Msambweni of Kwale county on the south east coast of Kenya, and in Ugenya sub-county of Siaya county in western Kenya near the border with Uganda (Fig. 1). These counties were chosen since they lie in similar ecological zones but are inhabited by people with different ethnicity, cultures and livestock-kee** habits. Both counties are major areas of sugar cane-production and thus a proportion of the population is engaged in this agro-industry. Both areas were among the counties with the highest tungiasis prevalence rate according to the Kenyan Ministry of Health in 2014 [7].

Fig. 1
figure 1

Map of the study sites to show the location in Kenya, schools and catchment areas by survey round and the prevalence of tungiasis in the schools during round 1 and 3 (squares and triangles, respectively). Prevalence is not indicated for round 2 catchment areas surveyed during COVID-19 restrictions (blue circles)

Full size image

Average rainfall amount (mm) for each month of the study from October 2019 to April 2021 was retrieved from World Weather Online [24] for the nearest possible sites with available data; Siaya town near Ugenya sub-county and Matuga town in Kwale. For data analyses the average rainfall for the two months prior to the survey month was calculated, for example the average rainfall in mm for April and May for all households visited in June. Both measures, average rainfall in the month of the visit and average rainfall in the two months before the visit, were converted from mm to cm (by dividing by 10) of rainfall for ease of interpretation of the outcomes from regression analyses.

Sampling procedure

Within Kwale and Siaya counties, sub-counties were selected that were known by the county Department of Health to have a high tungiasis burden. Lists of all existing public primary schools in the sub-counties were provided by the county education departments and 35 schools randomly selected using a paper lottery approach. In each school, 51 boys and 51 girls between the age of 8 and 14 years (the age with the highest prevalence and intensity of tungiasis [22]) were quasi-randomly selected by lining the pupils up into three age groups (8 and 9 years; 10 and 11 years; 12–14 years) and by sex within each age group. Every nth (depending on the total number in the group) pupil was then selected in each sex and age group until 17 from each was reached, to a total of 102. All selected pupils were examined for tungiasis, and a small questionnaire was administered asking about the floor of the house they sleep in and whether other people in the family are infected. The examinations and interview were conducted by field enumerators trained and supervised by the authors. Out of all tungiasis infected pupils identified in a school, a maximum of ten who reported living in a home with an unsealed soil or sand floor were randomly (paper lottery method) selected for household surveys. If only ten or less than ten were identified in a school, all were selected. We specifically targeted households with unsealed floors since this risk factor has been well-established for tungiasis [8, 25, 26] and for our risk factor study to be reported elsewhere we wanted to explore other determinants of disease. This strategy was used from February to mid-March 2020.

COVID-19 restrictions forced the closure of schools from March to December 2020. Field research activities were however free to recommence from July 2020. Consequently, the survey strategy was adapted to recruit cases through a household-based survey only. In the catchment areas of 11 schools already selected randomly before COVID-19 closures, community health volunteers (CHVs) were asked to invite any children aged 8 to 14 years they considered to be infected, to a location where they could be screened by the study team. Out of all infected children seen through this approach, a maximum of ten were randomly selected for household surveys. This survey strategy was used between August and October 2020 (survey round 2) returning to school-based surveys in January to May 2021 (survey round 3). No school or household was visited more than once. Figure 1 shows the location of the schools screened in round 1 and 3 and the catchment areas in round 2, demonstrating their equal distribution across the study area.

Clinical assessment procedures

The clinical assessment procedures were the same for the pupils in schools and for the household members of the selected pupils.

The feet of the 102 children in each school were washed and dried and systematically examined for the presence of tungiasis by observing nine zones on each foot in order from the largest toe to the smallest toe, the medial side, lateral side, the sole and the heel. Those pupils found to have sand fleas embedded in the feet, were assessed for intensity of infection by counting the number of fleas that were: alive (round white lesion with dark spot at center); dead (black, irregular-shaped lesion); manipulated (lesion from where a flea had clearly been removed) or large clusters of embedded fleas in which individual fleas could not be counted as they were too close together.

Infected individuals were also assessed for acute and chronic symptoms by modifying the technique described by Kehr et al. [15]. Both feet were examined systematically by nine zones each (five toes, medial side, lateral side, sole and heel). For each of the zones, the presence or absence of each symptom was recorded. These included desquamation, fissures, ulcers and abscess for acute symptoms and hyperkeratosis, peri-ungual hypertrophy, deformed nails and lost nails for chronic symptoms. In comparison to the original description of the score by Kehr et al. [15], the signs of inflammation; edema, erythema and warmness, were not recorded due to the difficulty of assessing these for non-clinicians.

Infected individuals were also asked to report the amount of pain and, separately, the amount of itching they felt in their feet associated with the embedded fleas using the options: “none at all”, “a little”, “some” or “a lot”.

Infra-red thermography

A low-resolution (220 × 160) infra-red camera detecting wavelengths of 8–14 µm (Hti Thermal Imaging Camera, HT-A1, Dongguan **/living in their caregiver’s house and have to care for themselves [36]. It is also possible that in spending time together in close proximity in a crowded classroom or elsewhere, these children could infect each other since free-living, host-seeking female fleas have been shown to move between hosts [31]. The higher prevalence among the youngest and oldest members of the community may also reflect their inability to remove penetrating fleas with a sharp instrument [22]. The higher odds of infection among pupils with unsealed sand or mud floors was also as expected from past studies[22, 25, 26, 29] and likely reflects the developmental needs of the off-host stages and high odds of the emerging adults finding a suitable host.

As might be expected, the intensity of infection in school pupils was positively correlated to the prevalence, albeit with some outlier schools. The more children are infected in a community, the more contamination of the environment with off-host stages and the higher the exposure of others to infection. Even if most transmission is happening within homes [8], children of this age group often spend time visiting and even slee** in the homes of friends and family, particularly during school holidays/closures (personal observations), and thus might be exposed to infection from other households. Further enquiries of the research team regarding the outlier schools where there was a prevalence of less than 7% but a high median intensity, found that in three of the six schools, two of the cases were siblings and some from families who were members of religious sects which do not accept any modern health care. In addition, these schools had received tungiasis interventions in the recent past and it is not inconceivable that these children/families refused treatment at that time, as they did when our study teams visited. Explaining these anomalies will probably require in-depth anthropological studies.

While some past studies have described tungiasis as a highly aggregated disease where the majority of cases had an intensity of infection of 1–5 fleas and only a few cases had a high intensity of infection with more than 30 fleas [9, 10, 22, 35], the current study found the majority of cases (47%) had 6–30 fleas, and 18% had more than 30 fleas. Our median intensity of infection of 13 among all infected individuals was also considerably higher than reported for other studies, which varied from a median of 2.5 in Cameroon [35] to 6 in Nigeria [33]. This suggests an overall higher burden of disease in the current study, and yet there was a lower prevalence of disease compared to past studies. This difference may have been caused by our intensity of infection measure incorporating a count of five for every flea cluster, and not a straight flea count. Other studies do not mention how flea clusters were incorporated in the flea counts, but possibly the enumerators attempted to count the closely packed fleas, which we felt was unlikely to be accurate in our circumstances.

Our simplified symptoms score is appropriate since both the acute and chronic symptoms scores correlated significantly with the infection intensity, and with each other. The most commonly reported symptom in previous studies was deformed nails, being as high as 98% of cases in a study focusing only on severe cases [15] and was seen on 73% of cases in the present study and correlated with infection intensity. Toenail loss and deformity are likely caused by fleas embedded in the nail bed and hyponychium causing direct physical damage as well as damage through inflammation. Some nail deformity may be the result of lost nails regrowing from damaged nailbeds. This suggests many of the cases in this study have been heavily infected for some time since deformed nails are the result of chronic and severe infection. This is striking since toenail loss and deformity are permanent if the nailbed is damaged and may remain as a mark of past disease and a source of stigma, shame, and discrimination for life. In fact, as part of a previous shoe donation program one of the investigators (LE) has had teenage girls explicitly say they are happy to receive shoes as they enable them to hide their toenails deformed by past infections.

To further simplify the assessment of inflammation we adapted the methodology of Schuster et al. [27] who demonstrated high resolution thermography can identify areas of inflammation by taking measurements of the temperature of the skin around embedded fleas and comparing that to other areas of the foot. Simple handheld infra-red cameras are affordable (costing as little as USD 170 for smart phone attachments or USD 385 for the model used in the current study) are readily available and can be used to locate areas of the skin that are hotter than others through a color transformation of the raw data. Instead of measuring temperatures of the skin we trained observers to record presence or absence of hot spots in the 18 zones of the feet used to record symptoms. The facts that the infra-red images revealed inflammation where no edema or erythema was visible as in Fig. 2 and that the number of sites with hot spots correlated significantly with the intensity of infection and the acute symptoms scores, suggest it is a good proxy measure for inflammation. Simple thermography such as this will likely be very useful in clinical trials to monitor the impact of treatment without the need of expensive equipment.

As governments and organizations begin to map tungiasis in their countries and to implement intervention programs, it will be important to have clear definitions for identifying target populations and intervention goals. Previously the three-tier classification of disease severity has been used, based only on flea counts with no description of how this correlated with symptoms. Since WHO guidelines for the control of other NTDs use targets based on two-tier levels of morbidity caused by different infection intensities, we propose a two-tier classification for tungiasis based on the Clinical score of 10. As with other NTDs, assessment of symptoms is time consuming, and parasite counts quicker and simpler to conduct for large scale public health projects, so we recommend a total flea count of 10 is an appropriate threshold for severe disease.

This study had three main limitations. While the pandemic accidentally afforded the opportunity to observe the impact of school closures, it also meant we could not continue the randomized sampling of pupils in all communities that was needed to obtain prevalence estimates across all sampling sites. Secondly, school-based sampling means the study would have missed children from the poorest families who cannot afford to send their children to school and given the association of tungiasis with poverty [6], possibly the most affected children. In addition, the survey would have missed those children who may have been unable to walk to school on account of having severe tungiasis. However, this was likely to have a minimal impact on the study outcomes since some children with high intensity infections were identified in the schools. Lastly, the study did not assess other measures of inflammation such as oedema, erythema and warmness of skin to compare with the thermographic measure.

Conclusions

Tungiasis is a highly heterogeneous disease with the prevalence in schools varying considerably. Prevalence of tungiasis was positively correlated with infection intensity and with morbidity. Simplified thermography is a valuable addition for assessing morbidity associated with tungiasis and will be useful to assess the efficacy of treatment in future clinical trials. Along with other pathologies, thermography helped to classify mild and severe disease which will be used in our future studies on the impact of tungiasis. Fortuitously, the survey spanned the COVID-19 school closures and demonstrated that when children spent an extended period out of school, the prevalence, intensity and morbidity of tungiasis increased significantly indicating prevention measures and education should target household level infrastructure and behavior.