Mapping of Podoconiosis Cases and Risk Factors in Kenya: A Nationwide Cross-sectional Study

ABSTRACT.

Podoconiosis is a type of tropical lymphedema that is clinically distinguished from lymphatic filariasis (LF) because it is ascending and commonly bilateral but asymmetric. The disease is a result of a genetically determined inflammatory reaction to long-term exposure to mineral particles in irritant red clay soils derived mainly from volcanic soils. We conducted the first nationwide mapping of the prevalence and risk factors of podoconiosis in Kenya. We performed a population-based cross-sectional survey to determine the national prevalence of podoconiosis and included 6,228 individuals from 48 villages in 24 sub-counties across 15 counties. Participants answered a questionnaire about the history of symptoms compatible with podoconiosis, received a point-of-care antigen test, and underwent a physical examination if they had lymphedema. A confirmed case of podoconiosis was defined as a case in a resident of the study village who had lower limb bilateral and asymmetric lymphedema lasting more than 1 year, negative test results for Wuchereria bancrofti antigen, and other causes of lymphedema ruled out. Of all the individuals surveyed, 89 had lymphedema; of those, 16 of 6228 (0.3%; 95% confidence interval [CI], 0.1–0.5) were confirmed to have podoconiosis. A high prevalence of podoconiosis was found in western (Siaya, 3.1%; Busia, 0.9%) and central (Meru, 1.1%) regions, and a low prevalence was observed in northern (Marsabit, 0.2%), eastern (Makueni, 0.2%), and coastal (Tana River, 0.1%) regions. The identified risk factors were age 56 years or older (adjusted odds ratio [aOR], 5.66; 95% CI, 2.32–13.83; P < 0.001) and rarely wearing shoes (aOR, 18.92; 95% CI, 4.55–78.71; P < 0.001). These results indicated that the podoconiosis prevalence is low and localized in Kenya; therefore, elimination is achievable if appropriate disease prevention, management, and behavioral strategies are promoted.

INTRODUCTION

The WHO defines podoconiosis as a type of tropical lymphedema that is clinically distinguished from lymphatic filariasis (LF) because it is ascending and commonly bilateral but asymmetric.^1,2 Podoconiosis is thought to be the result of a genetically determined inflammatory reaction to long-term exposure to mineral particles in irritant red clay soils derived mainly from volcanic deposits.^3,4

During the past 9 years since the WHO included podoconiosis on its list of neglected tropical diseases (NTDs) in February 2011, much progress has been made globally in terms of disease advocacy, research, and control.⁵ Therefore, many studies have updated the global distribution of this ascending, geo-chemical lymphedema and made advances in disease assessment, treatment, and advocacy; furthermore, these studies have successfully propelled podoconiosis to the global limelight necessary for its targeted control and elimination.⁵ However, the WHO has not formally set any global targets for its elimination.^2,5

Podoconiosis was first identified by Ernest Price in the early 1970s through his extensive research in Ethiopia and the eastern Africa region of the etiology, natural, history, distribution, and management of nonfilarial lymphedema.⁶ Further work performed by Wanji et al.⁷ and Ruberanziza et al.⁸ showed lymphedema in the absence of filarial parasites during their mapping surveys in Cameroon and Rwanda, respectively. Since then, a new staging system for podoconiosis has been developed, validated, and used to assess the simple treatment regimen developed during a community-based control program in southern Ethiopia.^9,10 Ethical approaches to podoconiosis research, especially in the remote community areas where patients affected with podoconiosis live, have been developed.¹¹ The locus of a susceptibility gene and characterization of minerals that trigger podoconiosis have also been investigated.¹² Furthermore, other studies have investigated immunological changes associated with podoconiosis and documented its overlap with common NTDs.⁵

Podoconiosis is clinically diagnosed based on the individual’s location (altitude), history, clinical examination, and absence of microfilaria antigen based on the filariasis test strip (FTS) results, and by the exclusion of infectious and hereditary causes of lymphedema.¹³ Usually, this disease starts in the individual’s foot and progresses up the leg (usually bilaterally) to the knee, but it rarely involves the groin.¹⁴ The disease has been correlated with altitudes higher than 1000 m above sea level and with annual precipitation exceeding 1,000 mm.¹⁵ Conversely, it is distinguishable from LF, which is found mainly at lower altitudes and for which changes are often first noticed in the groin.¹⁶

Podoconiosis control strategies may be divided into three categories: primary prevention, secondary prevention, and tertiary prevention. Primary preventive measures focus on the prevention of contact between the feet and minerals in the irritant soil that trigger the inflammatory processes and consist of avoiding or minimizing exposure to irritant soils through frequent and consistent wearing of shoes or boots, regular foot hygiene, and covering of earthen house floor surfaces. Secondary and tertiary measures focus on the management of the lymphedema-related morbidity and involve simple lymphedema treatment regimens similar to those used during LF lymphedema management, such as daily foot hygiene (washing with water, soap, and antiseptic), elevation of the legs, controlled exercises, and wearing of shoes and socks.^1,2,17

Studies have shown that podoconiosis is endemic in 32 countries globally and distributed mainly in the highland areas of tropical Africa, Latin America, and Southeast Asia.^3,4,18,19 Although reliable and detailed data regarding the prevalence and distribution of podoconiosis globally are scarce, targeted population-based surveys suggest that the prevalence of podoconiosis is between 5% and 10% and mainly affects people who walk while barefoot and live on irritant soil.³ Recent estimates indicate that more than four million people live with podoconiosis globally.⁴ The majority of these people live in Africa, mainly Ethiopia, Cameroon, Uganda, Burundi, and Rwanda.^1,20,21 People living with podoconiosis are disproportionately severely stigmatized through exclusion from schools, local meetings, churches, and mosques. They are often denied marriage with uninfected individuals,^19,22–24 have a reduced quality of life,^25,26 lower productivity, and increased rates of depression.^27–29

Several studies have identified individual and environmental correlates thought to be associated with the development of podoconiosis.^30–32 However, few studies have clearly and simultaneously evaluated through a nationwide single study the individual, household, climatic, and environmental factors correlated with podoconiosis.¹³ Such a nationwide analysis of risk factors could be useful for comprehensively understanding their relative contributions to the underlying spatial heterogeneity of the disease.¹³

In Kenya, the current national burden and distribution of podoconiosis is unknown. We conducted the first nationwide, community-wide, population-based mapping of the prevalence of podoconiosis in the country. Additionally, we aimed to quantify the individual, behavioral, and household factors associated with the disease. The results of this mapping exercise provide essential input to the Kenyan and WHO action plans for the prevention, management, and elimination of podoconiosis.

MATERIALS AND METHODS

Study design and survey sites.

We conducted a nationwide population-based cross-sectional survey in 48 villages in 24 sub-counties across 15 counties covering the western, Nyanza, eastern, northeastern, Rift Valley, and coast regions of Kenya. The selected sub-counties had been suspected to have active podoconiosis and lymphedema cases. We selected two villages in each sub-county as per the previous approach for integrated mapping of podoconiosis in Ethiopia.³³ The selection of the villages was performed before the survey was initiated by the planning and coordination team using expert opinion and health information from each sub-county identified for mapping based on the previous approach in Ethiopia.³³ Furthermore, selection was performed in close collaboration with the county health management teams and other relevant organizations working in the area.

Study population and sampling.

The study population comprised individuals 5 years and older who were residents of the selected villages since birth. In each selected village, 50 households were selected using the systematic random sampling technique. Starting from a random point within the village, every k^th household was randomly selected (k denotes the sampling interval calculated by dividing the total number of households in the village by the total sampled households in the village). Within a household, all individuals 5 years or older who were residents in the village since birth (for those younger than 10 years) and for more than 10 years (for those older than 10 years) were targeted until a final sample size of 100 individuals per village was achieved. The individuals selected were enrolled to participate in the survey only after informed consent was obtained. When we failed to achieve the required sample size in the selected village, the sampling was extended to the neighboring village. Therefore, the final sample size of 2,400 households was calculated. We estimated an average of two individuals in each household to yield a final sample size of 4,800 study participants.

Before the household survey, communities were sensitized about the activities followed by random selection of households. In each of the selected village, a household survey was implemented. Sensitization meetings were held with relevant county-level health and administrative authorities, the chiefs and assistant chiefs of the study locations, and other key community leaders to inform them about the study.

Survey procedure and diagnosis of podoconiosis.

A confirmed case of podoconiosis was defined as a case in a resident of a study village since birth (for those younger than 10 years) or for at least 10 years (for those older than 10 years) involving lower limb bilateral and asymmetric lymphedema lasting more than 1 year, negative test results for Wuchereria bancrofti antigen (determined by the FTS results), and had other causes of lymphedema ruled out through physical and medical examinations. Therefore, for participants with lymphedema, additional questions on a checklist were administered to rule out other possible causes of lymphedema (leprosy, heart disease, hereditary lymphedema, and onchocerciasis) to confirm a diagnosis of podoconiosis. Additionally, a physical examination was conducted for all participants with lymphedema to verify clinical differential diagnoses (e.g., cardiovascular examination for suspected heart failure, loss of sensation in the toes for leprosy, scrotal involvement for LF, and visual loss for Onchocerca volvulus).³³ Medical examinations to determine causes of lymphedema were conducted by trained clinicians and nurses at local health facilities in the study area. In addition, a myriad of factors related to the causes of podoconiosis were collected to help determine risk factors related to podoconiosis in Kenya (Figure 1). Furthermore, a summary of individuals positive for LF according to the FTS results is shown in Supplemental Table 1. It is important to note that the FTS results showed that some people had positive results for LF and were confirmed to have negative results for microfilarial; therefore, they did not require treatment.

Figure 1.

Conceptual framework showing the potential factors related to podoconiosis in the 15 counties of Kenya. This figure appears in color at www.ajtmh.org.

Data collection and management.

The study was conducted in 15 counties between November 11, 2019 and November 30, 2019. Study tools including the questionnaire and screening tools were programmed into the Secure Data Kit (SDK) system and administered using smartphone; the system included predefined data quality checks to minimize data entry errors.³⁴ The study questionnaire was administered to each consenting participant and included information regarding socio-demographic, economic, and household characteristics, history of residence, deworming, and shoe-wearing behavior. Individual factors included age, sex, occupation, shoe-wearing behaviors (including the age when an individual first wore shoes and the frequency of wearing shoes), and feet hygiene. Household-level factors included the average household monthly income and household conditions. Data regarding blood collection and FTS results were first recorded on a paper questionnaire before being transferred to the SDK system. All study data were transmitted daily to a central server at the Kenya Medical Research Institute in Nairobi, and paper records were stored securely under lock and key in the NTD laboratory within the National Public Health laboratories. Participants were identified primarily by their study identification number. No individual identities were used in any reports or publications resulting from the study. The geographical positioning system (GPS) coordinates of all the study households and villages were collected during the survey.

Ethics approval and consent to participate.

The survey was part of the monitoring and evaluation activities of the NTD unit. Ethical approval for the survey was obtained from the Amref Health Africa Ethics and Scientific Review Committee (ESRC) (approval no. AMREF-ESRC P468/2018) and the Brighton and Sussex Medical School, Research Governance and Ethics Committee as part of the Global Atlas of Podoconiosis. Community informed consent was obtained after discussions with the respective county health management teams and community leaders. All communities were sensitized before the sample collection, and verbal consent was sought from all the individuals. Written informed consent was obtained from the household head on behalf of the selected household members who were also sensitized regarding the purpose of the activity. All individuals participated voluntarily, and the names were anonymized after the sample collection procedures.

Statistical analysis.

The podoconiosis prevalence was calculated and the 95% confidence intervals (CIs) were determined by using a binomial regression model and considering clustering by households. Overall, the factors associated with podoconiosis occurrence were analyzed using a univariable analysis and described as odds ratios (ORs) using a mixed effects logistic regression model at three levels: individuals nested within villages selected within sub-counties and counties. To select the minimum adequate variables for the multivariable analysis, an inclusion criterion of P < 0.1 was prespecified as a sequential (block-wise) variable selection method that selected covariates meeting the set criterion. Adjusted ORs (aORs) of the most parsimonious model were obtained by mutually adjusting all minimum generated variables by using a multivariable mixed effects logistic regression model at 95% CI and considering the hierarchical nature of the data. All statistical analyses were performed using STATA version 15.1 (STATA Corporation, College Station, TX). All graphs were developed using the ggplot package implemented in R. Village locations and the geographical distribution of the prevalence were mapped using ArcGIS Desktop version 10.2.2 software (Environmental Systems Research Institute Inc., Redlands, CA).

RESULTS

Demographic information and shoe-wearing behavior.

In each village, an average of 43 households (range, 12–120 households; standard deviation [SD], 57 households) and 129 participants (range, 44–311; SD, 77 participants) were surveyed. Overall, data were collected from 2,024 households, 6,228 participants between 5 and 105 years of age (median, 30 years; interquartile range, 30 years). Of all the participants surveyed, 3,381 (54.3%) were females and 2,847 (45.7%) were males. The demographic information as well as the number of participants and villages (and households) surveyed per county and sub-county are shown in Table 1. Furthermore, the number of individuals examined and the number of those with podoconiosis in each village are provided in Supplemental Table 2.

Table 1.

Demographic characteristics, numbers of villages (households) and participants surveyed, and prevalence of podoconiosis cases diagnosed in 24 sub-counties (15 counties) in Kenya

Demographics	Villages, no. (no. of households*)	Participants, no. (%)	Podoconiosis cases (N = 16)
			No. with podoconiosis	Prevalence % (95% CI)
Overall	48 (2,024)	6,228 (100%)	16	0.3 (0.1–0.5)
County (sub-county)
Baringo		401 (6.4%)	0	0
Baringo South	2 (72)	201	0	0
Tiaty	2 (55)	200	0	0
Busia		221 (3.6%)	2	0.9 (0.2–4.2)
Teso North	2 (105)	221	2	0.9 (0.2–4.2)
Homabay		199 (3.2%)	0	0
7Homabay Town	2 (195)	199	0	0
Makueni		411 (6.6%)	1	0.2 (0-1.6)
Kibwezi East	2 (35)	208	1	0.5 (0.1–2.9)
Kibwezi West	2 (39)	203	0	0
Marsabit		547 (8.8%)	1	0.2 (0–1.2)
Moyale	2 (24)	147	0	0
North Horr	2 (27)	199	0	0
Saku	2 (45)	201	1	0.5 (0.1–3.4)
Meru		362 (5.8%)	4	1.1 (0.5–2.6)
Igembe Central	2 (22)	205	3	1.5 (0.8–2.7)
South Imenti	2 (29)	157	1	0.6 (0–9.2)
Nakuru		404 (6.5%)	0	0
Naivasha	2 (107)	203	0	0
Subukia	2 (81)	201	0	0
Narok		141 (2.3%)	0	0
Narok East	2 (141)	141	0	0
Siaya		194 (3.1%)	6	3.1 (2.8–3.4)
Alego Usonga	2 (129)	194	6	3.1 (2.8–3.4)
Trans Nzoia		205 (3.3%)	0	0
Endebes	2 (36)	205	0	0
Turkana		367 (5.9%)	0	0
Turkana North	2 (39)	160	0	0
Turkana West	2 (52)	207	0	0
Wajir		206 (3.3%)	0	0
Tarbaj	2 (175)	206	0	0
West Pokot		207 (3.3%)	0	0
Kacheliba	2 (29)	207	0	0
Taita Taveta		605 (9.7%)	0	0
Taveta	2 (133)	605	0	0
Tana River		1,758 (28.2%)	2	0.1 (0–0.4)
Bura	2 (97)	610	0	0
Galole	2 (16)	529	1	0.2 (0–1.1)
Garsen	2 (341)	619	1	0.2 (0–1.2)
Sex
Male	48 (922)	2,847 (45.7%)	4	0.1 (0.1–0.4)
Female	48 (1,102)	3,381 (54.3%)	12	0.4 (0.2–0.7)
Age group (in years)
< 15	13 (212)	779 (12.5%)	2	0.3 (0.1–0.9)
15–35	48 (1,262)	2,798 (44.9%)	3	0.1 (0–0.3)
36–56	48 (926)	1,561 (25.1%)	2	0.1 (0–0.5)
> 56	48 (678)	1,090 (17.9%)	9	0.8 (0.4–1.9)
Shoe-wearing behavior
Ever worn shoes
Yes	46 (1,988)	6,097 (97.9%)	16	0.3 (0.1–0.5)
No	2 (36)	131 (2.1%)	0	0
Age when first wore shoes
< 18 years	44 (1,855)	5,663 (92.9%)	13	0.2 (0.1–0.5)
≥ 18 years	2 (133)	434 (7.1%)	3	0.7 (0.2–1.9)

CI = confidence interval.

^*In some areas, household visits were not possible; therefore, participants were surveyed at a central and accessible place in the village, such as the chief’s camp, church, school, or community field.

Of the 6,228 respondents, 131 (2.1%; 95% CI, 1.2–3.6) reported that they have never worn shoes, and 434 (7.1%; 95% CI, 5.1–9.9) reported first wearing shoes when they were adults (age 18 years or older). Accordingly, the majority of participants who had never worn shoes were from Turkana (16.1%) and Meru (2.8%) counties. Additionally, many of the participants who had first worn shoes when they were adults (18 years or older) were from Meru (29.8%) and Makueni (17.5%) counties (Table 2).

Table 2.

Shoe-wearing behavior in 15 counties in Kenya

County	Proportion of participants who have never worn shoes, % (95% CI)	Proportion of participants who first wore shoes at 18 years or older, % (95% CI)
Overall (all counties)	2.1 (1.2–3.6)	7.1 (5.1–9.9)
Baringo	2.5 (1.5–4.2)	7.9 (5.3–11.8)
Busia	0.9 (0.6–1.4)	9.6 (2.7–33.8)
Homabay	0	9.0 (9.0–9.1)
Makueni	0	17.5 (13.4–23.0)
Marsabit	0	2.7 (1.5–5.0)
Meru	2.8 (0.4–17.8)	29.8 (22.2–40.0)
Nakuru	0.2 (0–1.8)	14.9 (11.4–19.4)
Narok	0.7 (0–0.5)	2.1 (0.6–7.1)
Siaya	0	9.8 (5.4–17.7)
Taita Taveta	0.2 (0–1.2)	5.6 (3.5–9.0)
Tana River	2.3 (1.0–5.1)	1.7 (0.8–3.7)
Trans Nzoia	2.0 (0.7–5.1)	4.0 (2.5–6.4)
Turkana	16.1 (10.1–25.6)	1.3 (0.3–5.5)
Wajir	0	0.5 (0.1–3.6)
West Pokot	1.4 (0.7–2.8)	6.9 (3.8–12.2)

CI = confidence interval.

Prevalence of podoconiosis cases.

The diagnosis of podoconiosis was conducted as shown in Figure 2. Overall, 16 of 6,228 (0.3%; 95% CI, 0.1–0.5) of the participants were diagnosed with podoconiosis. Twelve of the 16 cases were observed among female participants: 12 of 3,381 (0.4%; 95% CI, 0.2–0.7) females and 4 of 2847 males (0.1%; 95% CI, 0.1–0.4). After categorizing the prevalence according to participants younger than 15 years and those 15 years or older, it was observed that the prevalence among the two age groups was nearly similar: 2 of 779 (0.3%; 95% CI, 0.1–0.9) and 14 of 5,449 (0.3%; 95% CI, 0.1–0.6), respectively. Furthermore, those 15 years and older were categorized as 15 to 35 years, 36 to 56 years, and older than 56 years and showed an increasing prevalence of 3 of 2,798 (0.1%; 95% CI, 0–0.3), 2 of 1,561 (0.1%; 95% CI, 0–0.5), and 9 of 1,090 (0.8%; 95% CI, 0.4–1.9), respectively (Table 1).

Figure 2.

Study profile showing the podoconiosis diagnosis based on history, physical examination, and disease-specific test results in the 15 counties of Kenya. This figure appears in color at www.ajtmh.org.

Analyses by county indicated that podoconiosis cases were prevalent in six counties: Siaya, 6 of 194 (3.1%; 95% CI, 2.8–3.4); Meru, 4 of 362 (1.1%; 95% CI, 0.5–2.6); Busia, 2 of 221 (0.9%; 95% CI, 0.2–4.2); Makueni, 1 of 411 (0.2%; 95% CI, 0–1.6); Marsabit, 1 of 547 (0.2%; 95% CI, 0–1.2); and Tana River 2 of 1,758 (0.1%; 95% CI, 0–0.4) (Table 1). Accordingly, analyses of the prevalence by sub-county revealed that podoconiosis cases were prevalent in eight sub-counties: Alego Usonga, 6 of 194 (3.1%; 95% CI, 2.8–3.4); Igembe Central, 3 of 205 (1.5%; 95% CI, 0.8–2.7); Teso North, 2 of 221 (0.9%; 95% CI, 0.2–4.2); South Imenti, 1 of 157 (0.6%; 95% CI, 0–9.2); Saku, 1 of 201 (0.5%; 95% CI, 0.1–3.4); Kibwezi East, 1 of 208 (0.5%; 95% CI, 0.1–2.9); Garsen 1 of 619 (0.2%; 95% CI, 0–1.2); and Galole, 1 of 529 (0.2%; 95% CI, 0–1.1) (Table 1). Additionally, the geographical distribution of the prevalence of podoconiosis cases as observed in all the 24 sub-counties is presented in Figure 3, which shows that podoconiosis cases are prevalent in pockets of villages in the western and central parts of Kenya. Supplemental Table 2 provides the village-level prevalence of podoconiosis cases.

Figure 3.

A map showing the geographical location of the sampled villages and distribution of the prevalence of podoconiosis cases in the 24 sub-counties of 15 counties in Kenya. This figure appears in color at www.ajtmh.org.

Risk factors associated with the prevalence of podoconiosis cases in Kenya.

In addition to the demographic factors, more predisposing factors were added to this analysis, including occupation, shoe-wearing frequency, foot-washing frequency, household factors such as house floor and wall type, and information regarding the previous residence. We noted that the majority of participants were unemployed or students 3,284 (52.7%); only 2,944 (47.3%) participated in an income-generating activity. The majority (4849; 79.5%) reported wearing shoes every day; however, 100 (1.6%) rarely wore shoes. Half (3,134; 50.3%) and approximately three-quarters (4,498; 72.2%) of the houses had walls and floors made of mud/earth or wood, respectively (Table 3). Small proportions of the participants were infected with filariasis (89; 1.4%) or hydrocele (17; 0.6%). Only 1,681 (27.0%) of the participants reportedly received treatment for worms during the past 6 months.

Table 3.

Univariable associations between predisposing factors and prevalence of podoconiosis cases among surveyed participants in 15 counties in Kenya

Factors	No. (%)	Univariable logistic regression
		OR (95% CI)	P value
Individual and behavioral factors
Age group, years
< 15	779 (12.5%)	2.40 (0.44–13.02)	0.311
15–35	2,798 (44.9%)	Reference
36–56	1,561 (25.1%)	1.19 (0.19–7.45)	0.849
> 56	1,090 (17.9%)	7.76 (3.30–18.22)	< 0.001*
Sex
Female	2,847 (45.7%)	2.53 (1.05–6.14)	0.040*
Male	3,381 (54.3%)	Reference
Occupation
Retired/business/farmer	2,944 (47.3%)	1.86 (0.62–5.61)	0.268
Unemployed/student	3,284 (52.7%)	Reference
Age when shoes were first worn
< 18 years	5,663 (92.9%)	Reference
≥ 18 years	434 (7.1%)	3.02 (0.85–10.71)	0.087
How frequently do you wear shoes?
Every day	4,849 (79.5%)	Reference
Often but not every day	1,148 (18.8%)	0.94 (0.20–4.38)	0.934
Rarely	100 (1.6%)	28.27 (6.61–81.88)	< 0.001*
How often do you wash your feet?
Once/more than once per day	5,670 (91.0%)	Reference
Once/more than once per week	558 (9.0%)	0.68 (0.08–5.39)	0.712
Household factors
Floor type
Cement/tile	1,730 (27.8%)	Reference
Mud/earth/wood	4,498 (72.2%)	1.16 (0.27–4.87)	0.844
Wall type
Cement/stone	1,470 (23.6%)	Reference
Iron sheet	451 (7.2%)	1.09 (0.09–13.11)	0.948
Mud/earth	3,134 (50.3%)	1.73 (0.33–9.02)	0.518
Wood	1,173 (18.8%)	0.42 (0.03–5.02)	0.491
History of residence
Respondent has lived in the current sub-county for the past 10 years or longer
Yes	5,821 (93.5%)	1.05 (0.13–8.66)	0.966
No	407 (6.5%)	Reference

CI = confidence interval; OR = odds ratio.

^*Indicates a statistically significant factor.

Univariable analyses of individual and household predisposing factors and the residence history revealed mixed associations between the prevalence of podoconiosis cases and many of the variables of interest that were considered (Table 3). Significant associations observed were limited because of the already very low prevalence of podoconiosis cases among the sampled participants. Despite this limitation, the results indicated that older (older than 56 years) participants (OR, 7.76; 95% CI, 3.30–18.22; P < 0.001) and female participants (OR, 2.53; 95% CI, 1.05–6.14; P = 0.040) were at significantly higher risk for podoconiosis. Additionally, participants who rarely wore shoes (OR, 28.27; 95% CI, 6.61–81.88), P < 0.001) were at increased risk for podoconiosis. However, first wearing shoes during adulthood (age 18 years or older) was mildly associated with an increased risk of podoconiosis (OR, 3.02; 95% CI, 0.85–10.71; P = 0.087).

Additionally, the multivariable analysis (Table 4) indicated that the only factors significantly associated with an increased risk of podoconiosis cases were age older than 56 years (aOR, 5.66; 95% CI, 2.32–13.83; P < 0.001) and rarely wearing shoes (aOR, 18.92; 95% CI, 4.55–78.71; P < 0.001). However, female sex was mildly associated with an increased risk of podoconiosis (aOR, 2.27; 95% CI, 0.85–5.81; P = 0.087).

Table 4.

Multivariable associations between predisposing factors and prevalence of podoconiosis cases among surveyed participants in 15 counties in Kenya

Factors	Multivariable logistic regression
	aOR (95% CI)	P value
Individual factors
Age group, years
< 15	1.81 (0.27–12.08)	0.538
15–35	Reference
36–56	1.03 (0.15–7.18)	0.980
> 56	5.66 (2.32–13.83)	< 0.001*
Sex
Female	2.27 (0.89–5.81)	0.087
Male	Reference
Age when shoes were first worn
< 18 years	Reference
≥ 18 years	1.18 (0.30–4.63)	0.814
How frequently do you wear shoes?
Every day	Reference
Often but not every day	0.81 (0.18–3.72)	0.760
Rarely	18.92 (4.55–78.71)	< 0.001*

aOR = adjusted odds ratio; CI = confidence interval.

^*Indicates a statistically significant factor.

DISCUSSION

We present the results of nationwide mapping of podoconiosis in Kenya. To our knowledge, this is the first national population-based survey of podoconiosis in Kenya. The results indicated that podoconiosis is not widespread in Kenya, and that it is mainly localized to the highlands of the western and central regions of Kenya. The national prevalence is low (0.3%), indicating that the disease could be easily eliminated in the country if appropriate prevention and management measures are applied.

Our mapping results indicated that the highest prevalence rates of podoconiosis were found mainly in the counties of Siaya and Busia in the western highlands and Meru in the central highlands. Low prevalence rates were identified in Makueni, Marsabit, and Tana River counties, which are in the lowland areas. The high prevalence rates recorded for Siaya, Busia, and Meru counties were not surprising because these regions are located near mountainous areas of Mount Elgon in the western part of Kenya and Mount Kenya in central Kenya; all have an altitude higher than 1,200 m and relatively high rainfall amounts.^35,36 The soil types in these areas consist of volcanic ash that may contain irritant minerals thought to trigger inflammatory processes.^37–39 However, more studies are necessary to confirm the geochemical composition of the soils in these regions. It is not clear what causes podoconiosis in the low-prevalence counties of Makueni, Marsabit, and Tana River. We can only speculate that the observed cases could have resulted from soil from high-altitude areas being washed down into lowland areas and increased human activities, such as mining exploration and sand harvesting, which are often associated with working barefoot.^40–43 Disease management efforts need to be initiated and upscaled in these regions.

Our study reported an overall lower prevalence of podoconiosis compared to that of other endemic countries that are found mainly in Africa; the estimated prevalence in Ethiopia was 2.7% to 7.5%,^15,44 that in Cameroon was 0.5% to 8.1%,^20,44 and that in Uganda was 0.1% to 4.5%.⁴⁴ However, Rwanda recorded an overall lower prevalence of 68.5 per 100,000 people (i.e., 0.000685%, but with widespread cases throughout the country).¹⁹ The high prevalence in these countries could be the result of suitable environmental conditions, mainly in the highland rural areas, high precipitation, and soil rich in silt and clay particles,¹³ along with low levels of footwear use and inadequate access to water for foot hygiene.⁴

We found that the prevalence was high among participants who first wore shoes at an older age (18 years or older). Age at the time when shoes were first worn is one of the factors that has been reported to be associated with podoconiosis in previous studies.^45,46 The delay in shoe-wearing could be seen as a limiting socio-economic factor because those who wore shoes at an older age might be poorer and have less access to facilities necessary for foot hygiene. The cumulative effect of the delay in shoe-wearing and the contact of uncovered feet with red soil increase the risk of podoconiosis.⁴⁷

During this study, we found that the risk of podoconiosis was significantly higher among female participants and individuals who rarely wore shoes. These results are not surprising because previous studies have shown that podoconiosis is more prevalent among women,⁴⁸ probably because men participate in fewer domestic activities than women, whose their daily activities include collecting water from the river/lake, collecting firewood, farming, and other predisposing household activities.³⁰ Our results indicate as much as an 18-times higher risk of developing podoconiosis for participants who rarely wear shoes. This finding is supported by previous small-scale studies performed in Ethiopia and Kenya that showed that frequently not wearing shoes increased the risk of podoconiosis by up to nine-times compared with that of individuals who wear shoes daily.^41,46 This indicates that shoe-wearing is an important risk factor that control programs can target to slow the prevalence of podoconiosis. We observed that the prevalence of podoconiosis cases increased steadily with advanced age. A significantly higher risk (up to five-fold) of podoconiosis was found among individuals 56 years and older. This finding corroborates those of previous studies performed in Ethiopia,^13,31 Uganda,³⁶ and Cameroon,⁷ which all documented the increased risk of podoconiosis with increased age.

A strength of this study was that we used a combination of historical, physical, clinical, and molecular examinations and diagnosis techniques to exclude other potential causes of lymphedema. These procedures followed a standardized clinical algorithm for diagnosing podoconiosis.⁴⁴ Additionally, this study provided statistically powered and spatially representative sampling, which enabled us to capture localized occurrences within the country. The results of this study add to the evidence that podoconiosis is uncommon among younger individuals.¹³

Additionally, this study had some limitations. In some instances, the study team members were not able to access some of the participants’ houses because of poor road networks, poor weather, and the remoteness of the areas. Therefore, in some villages, participants were surveyed at central locations within the village, such as the chief’s camp, churches, or schools. However, this posed a limitation to potential participants who were not able to travel to the study location because of mobility-related constraints and the associated stigma. This may have resulted in the under-identification of people with podoconiosis who were more likely to experience mobility constraints.

In conclusion, this country-wide mapping revealed a very low prevalence of podoconiosis cases in the country, with only 8 of the 24 sub-counties surveyed reporting cases. A village-level prevalence of more than 1% was mainly observed in the western and central parts of Kenya. This mapping exercise is important to the country’s NTD control program because it will help Kenya target resources, monitor the control progress, and advocate for investments in podoconiosis prevention, control, and elimination. Furthermore, these results indicate that podoconiosis is highly localized and not widespread in Kenya; therefore, elimination is achievable if appropriate disease prevention, management, and behavioral strategies, which comprise the NTD 2021–2030 goals, are promoted. In the sub-counties where the prevalence was found to be more than 1%, it is important to conduct micro-level mapping to estimate the actual burden of cases and identify risk factors with higher spatial resolution.

Note: Supplemental tables appear at www.ajtmh.org.