Podoconiosis (endemic non-filarial elephantiasis) is a geochemical disease occurring in individuals exposed to red clay soil derived from alkalic volcanic rock. It is a chronic, debilitating disorder and a considerable public health problem in at least 10 countries in tropical Africa, Central America and northern India. Only a small proportion of individuals exposed to red clay develop disease and familial clustering of cases occurs, so we tested the hypothesis that disease occurs in genetically susceptible individuals on exposure to an environmental element in soil. Using multiple statistical genetic techniques we estimated sibling recurrence risk ratio (λs) and heritability for podoconiosis, and conducted segregation analysis on 59 multigenerational affected families from Wolaitta Zone, southern Ethiopia. We estimated the λs to be 5.07. The heritability of podoconiosis was estimated to be 0.629 (SE 0.069, P = 1 × 10−7). Segregation analysis showed that the most parsimonious model was that of an autosomal co-dominant major gene. Age and use of footwear were significant covariates in the final model. Host genetic factors are important determinants of susceptibility to podoconiosis. Identification of the gene(s) involved will lead to better understanding of the gene–environment interactions involved in the pathogenesis of podoconiosis and other complex multifactorial conditions.
Podoconiosis (endemic non-filarial elephantiasis) is a geochemical disease occurring in individuals exposed to red clay soil derived from alkalic volcanic rock (Fuller, 2005; Price, 1976a). It is a chronic and debilitating disorder that, although rarely recognized outside endemic regions, is a considerable public health problem in at least 10 countries in tropical Africa, Central America and northern India (Crivelli, 1986; Corachan et al., 1988; de Lalla et al., 1988; Onapa et al., 2001; Price, 1990; Ruiz et al., 1994). In Ethiopia alone, 11 million people (18% of the population) are at risk through exposure to irritant soil, and a recent estimate based on prevalence data from an endemic area in southern Ethiopia suggests that between 500 000 and 1 million people are affected nationwide (Destas et al., 2003). Podoconiosis is thus more common than HIV infection in these regions. Up to 64% of affected individuals are from the most economically active age groups (Destas et al., 2003), and a recent study estimated direct and productivity costs of podoconiosis in Wolaitta Zone alone at US$16 million per annum (Tekola et al., 2006). Social stigma towards people with podoconiosis is pronounced, leading to exclusion from school, religious and community gatherings, and a complete bar on marriage into non-affected families.
The epidemiological and clinical features are distinct and well described: the disorder occurs in barefooted farmers and other occupational groups in which exposure to volcanic soil is common (Price, 1973, 1974a, 1976a). Affected populations tend to reside at high altitudes, and there is a geographical correlation with alkalic-type volcanoes (Price, 1976a; Price and Bailey, 1984). Men and women are equally affected, and while the majority of cases develop signs and symptoms in the second and third decades, individuals as young as 4 years old may show early signs. Disease is bilateral, but asymmetrical, and almost always limited to below the knee. In terms of the pathogenesis, silica particles absorbed through the foot are thought to induce an inflammatory response in the lymphatic vessels, leading ultimately to fibrosis and obstruction of the vessel lumen (Fyfe and Price, 1985; Price, 1977; Price and Henderson, 1978). This leads initially to oedema of the foot and lower leg, which progresses to severe elephantiasis (Figure 1 ), resembling that observed in Wuchereria bancrofti infection (Price, 1983). However, podoconiosis occurs at altitudes above 1500 m, which exceeds that at which filarial transmission occurs and both midnight blood sampling and filarial antigen testing have consistently excluded filariasis as the cause of podoconiosis (Birrie et al., 1997; Cohen, 1960; Onapa et al., 2001; Price, 1974b). Podoconiosis is avoided by minimizing exposure to irritant soil through the use of appropriate footwear; the condition was present in Ireland, Scotland and France before footwear was widely adopted (Price, 1990). However, shoes are rarely used in Wolaitta Zone, because they are beyond the means of most of the population there and because simple preventive education messages are missing.
One of the striking features of podoconiosis is that only a small proportion of individuals who are exposed to red clay develop disease. Familial clustering of cases has been noted in affected communities in Ethiopia, Rwanda and Burundi, and by professionals working in clinics in Addis Ababa (Price, 1972, 1976b). Price studied a series of 90 families with more than one affected member who presented at the Princess Zenebeworq Memorial Hospital in Addis Ababa. His data supported an autosomal recessive mode of inheritance for the trait, with an estimated gene frequency of between 15 and 40% (Price, 1972). We aimed to elaborate the combined roles of genetics and environmental exposure in the development of podoconiosis using segregation analysis in family pedigrees gathered in an area of high prevalence in southern Ethiopia.
Materials and methods
The study was conducted in Wolaitta Zone, Southern Ethiopia. The zonal capital, Sodo, is located 380 km southwest of the capital city, Addis Ababa, and lies at an altitude between 2000 and 2100 m, beneath the extinct volcano, Mount Damota (2908 m). The zonal population is about 1.5 million, of which 92.2% live in rural areas. The average family size is nine people, and Wolaitta Zone is one of the most densely populated areas in Ethiopia, with an average of 290 people/km2. The majority of the population earn their livelihood from crop production (61.3%), followed by livestock rearing (22.3%). The main ethnic group is Wolaita, whose mother tongue is Wolaytigna (CSA, 1998). Podoconiosis prevalence was measured at 5.46% of the general population in 2001 (Destas et al., 2003).
Study design and ethical approval
We planned to collect data from multigenerational affected families to determine the risk of disease among siblings of affected individuals compared to that in the general population, to estimate the relative contribution of genetic factors to disease aetiology and to do segregation analysis to determine the most likely mode of inheritance. Approval was obtained from ethical committees of the Faculty of Medicine, Addis Ababa University and the National Ethical Review Committee of the Ethiopian Science and Technology Commission.
A local non-government organization, the Mossy Foot Prevention and Treatment Association, has been working with podoconiosis patients and their families for more than 15 years. The Mossy Foot Prevention and Treatment Association is based in Sodo town and runs outreach clinics in 11 sites in the surrounding administrative areas, at distances of 9 to 45 km from Sodo town. We approached health workers at seven of these outreach sites and asked them to take the research team to the homes of potential probands. A proband was defined as a person with podoconiosis who had one or more affected first degree family members and was present for examination at the time of the visit of the research team. Family members living on the same compound were examined, but we relied on verbal reports of podoconiosis status for family members living at a distance. Given the stigma associated with disease, we consider that under-ascertainment of affected people was more likely than over-ascertainment. Distinction between podoconiosis and lymphatic filariasis relied on clinical examination (Price, 1990), the altitude of the communities visited (all >1600 m), and exclusion of migrants from lowland areas. Information on age, sex, relationship to proband, vital status, presence of podoconiosis, age of onset and use of footwear was gathered for living and dead family members over three or four generations and was entered into Microsoft Office Excel 2003 (Microsoft Corp., Seattle, WA, USA).
Pedigree statistics were generated using the Pedstats package (Wigginton and Abecasis, 2005). Estimation of heritability, defined as the proportion of the total phenotypic variation due to genetic factors, was done using the SOLAR software package (Almasy and Blangero, 1998). Affectation with podoconiosis was analysed under a polygenic model using the discrete trait algorithms implemented in SOLAR version 2.1.2. Age of onset was analysed under a polygenic model using a maximum-likelihood variance components approach, as implemented in SOLAR version 2.1.2. All heritability analyses were done with proband ascertainment correction (i.e., adjusting for the manner in which index cases were ascertained) and the inclusion of age, sex and use of footwear as covariates in the models. We also estimated the sibling recurrence risk (the risk of disease among siblings of probands) and sibling recurrence risk ratio (the recurrence risk of disease among siblings of an affected individual compared with the risk of disease in the general population).
We conducted segregation analysis using a variety of genetic and non-genetic models to determine which inheritance model was most consistent with the observed data. Segregation analysis was done using the Genetic Analysis Package (Epicenter Software, Pasadena, CA, USA). Several models were fitted to the data, including: single major gene Mendelian models (dominant, recessive, additive, and co-dominant), a polygenic model, an environmental model, a founder-non founder model and a general transmission model. Mixed models (single major gene with a polygenic component) were also tested. Adjustment for proband ascertainment was done in all the segregation analyses. Likelihood ratio tests (LRTs) were used to compare the general model to the Mendelian models. We also computed Akaike's Information Criteria (AIC) for each model as −2 (log-likelihood at the maximum likelihood estimator) +2 (number of model parameters estimated). A lower AIC indicates a more parsimonious model.
Data were collected on 59 pedigrees with 1400 family members. Characteristics are summarized in Table 1 . Nineteen percent of individuals whose disease status was recorded had podoconiosis (19.4% of males and 18.5% of females, P > 0.05). Mean (SD) family size was 23.7 (6.8) individuals, with a range from 14 to 48 family members. Twenty-five families had data collected over three generations, and 34 over four generations. Two hundred and thirty-five sib ships were recorded, with a mean (SD) size of 4.7 (2.6), a range from 1 to 16 siblings, and a total of 175 affected sib-pairs. Other types of relative pairs in the dataset included: 131 parent–offspring pairs, 12 half-sibling pairs, 142 avuncular pairs and 42 grandparent–grandchild pairs. A typical pedigree is illustrated in Figure 2 .
The sibling recurrence risk (the risk of disease in a sibling of the proband) in this dataset (Ks) was estimated to be 0.277; that is, just over one-quarter of siblings of probands were themselves affected. The sibling recurrence risk ratio corrects the sibling recurrence risk for the prevalence of disease in the general population, and is defined as sibling recurrence risk/population prevalence. Given the population prevalence of 5.46% (Destas et al., 2003), we estimated the sibling recurrence risk ratio (λs) to be 5.07; thus the risk that a sibling is affected is five times that if he or she were a member of the general population.
Heritability of podoconiosis (with correction for proband ascertainment) was estimated to be 0.629 (SE 0.069, P = 1 × 10−7); that is, about 63% of phenotypic variance is accountable for by genetic factors. The heritability of the age of onset (as a continuous phenotype) was 0.621 (SE 0.036, P = 4.9 × 10−56), indicating that genetic factors influence when, as well as whether, the phenotype develops.
Segregation analysis showed that the most parsimonious model was that of an autosomal co-dominant major gene, with age as a significant covariate and footwear as a borderline significant covariate (Table 2 ). This model fitted the data better than a general, purely environmental, founder, and single major gene models and had the smallest log likelihood among the mixed (single major gene plus polygenic component) models.
The heritability and sibling recurrence risk we calculated provide striking evidence for a genetic basis to podoconiosis, and suggest that further study using linkage analysis may be productive in mapping genes predisposing to podoconiosis.
Even though reported age shows heaping towards 5- or 10-year peaks, as is common when eliciting age in a population without formal birth registration, the relationships we found between podoconiosis and both age and age-squared are consistent with the clinical picture. We do not believe that age reporting would be biased in respect to podoconiosis status. If non-differential misclassification of age has occurred, this would, if anything, dilute the effect seen, suggesting a greater association between podoconiosis and age in reality. Although the early stages of podoconiosis may be found in under 10-year-olds, mature disease is usually found after several years of exposure to irritant soil, when patients are in their twenties or thirties. The burden of disease has been shown to fall in the most economically active sector of the population (Destas et al., 2003), having severe impact on productivity (Tekola et al., 2006).
Segregation analysis also indicated the borderline significance of lack of footwear (a proxy for exposure to irritant soil) as a covariate in the final model. The measure of footwear use was relatively crude: participants were asked if they ever wore shoes or boots, and the response was given a binary code; thus, individuals who had worn footwear even once were considered as footwear users, which is likely to dilute any true protective effect of footwear, if it is assumed that misclassification is non-differential. However, it is possible that a form of social desirability bias operated in relation to reported use of footwear: affected individuals may have been more likely to know that use of footwear is a preventive strategy than unaffected people, and thus have been more likely to report footwear use. This bias would diminish any true protective effect.
The key intervention to prevent podoconiosis is already known and appears simple: consistent use of appropriate footwear while engaged in activities on irritant soil (Price, 1975). However, even were this simple measure widely known to rural Ethiopians, it remains beyond their means. The rationale for further studies of gene–environment interactions in predisposing to podoconiosis is twofold: susceptible families may be targeted for provision of free footwear to prevent disease, and through insight into mechanisms of disease, strategies to reverse early stages may be developed.
Relevance to other fibrotic diseases
Not every individual exposed to irritant soil develops podoconiosis. Investigations have demonstrated silicon and aluminium particles within lymph node macrophages of normal individuals living on irritant soil as well as those of people with podoconiosis (Price and Henderson, 1978). This suggests that podoconiosis represents an abnormal reaction to mineral particles absorbed into the lymph system, which in turn implies that identification of the genes predisposing to podoconiosis might prove valuable to other conditions, such as pneumoconiosis, in which abnormal reactions to minerals provoke fibrosis.
Relevance to other diseases exhibiting gene–environment interactions
Numerous important causes of morbidity and mortality result from the interaction between genes and environmental factors. However, many conditions exhibiting gene–environment interaction are polygenic, so complex study designs involving numerous subjects are required to identify all the factors involved. Podoconiosis could provide a model for gene–environment interactions, given the apparently simpler nature of interaction between a single environmental factor and few major genes.
Host genetic factors are important determinants of susceptibility to podoconiosis. Identification of the gene(s) involved will lead to better understanding of the gene–environment interactions involved in the pathogenesis of podoconiosis and other complex multifactorial conditions.
Conflicts of interest statement
The authors have no conflicts of interest concerning the work reported in this paper.
We gratefully acknowledge the financial support of the Association of Physicians of Great Britain and Northern Ireland under their Links with Developing Countries Scheme 2003/2004. We thank Ato Meskele Ashine of the Mossy Foot Prevention and Treatment Association for advice and logistic support.