Search tips
Search criteria 


Logo of wtpaEurope PMCEurope PMC Funders GroupSubmit a Manuscript
Trop Med Int Health. Author manuscript; available in PMC 2010 November 29.
Published in final edited form as:
PMCID: PMC2992944

Development and testing of a de novo clinical staging system for podoconiosis (endemic non-filarial elephantiasis)

Fasil Tekola, MPH, Research Fellow, Zewdu Ayele, MD, Assistant Professor, Dereje HaileMariam, MD, General Practitioner, Claire Fuller, FRCP, Consultant Dermatologist, and Gail Davey, MD, Associate Professor



Podoconiosis (endemic non-filarial elephantiasis) is a geochemical disease in individuals exposed to red-clay soil. Despite the prevalence and public health importance of podoconiosis, there is as yet no accepted clinical staging system.


We aimed to develop and test a robust clinical staging system for podoconiosis.


We adapted the Dreyer system for staging filarial lymphoedema and tested it in four re-iterative field tests conducted in an area of high podoconiosis prevalence in Southern Ethiopia. The system finally arrived at has five stages according to proximal spread of disease and presence of dermal nodules, ridges and bands. We measured the one-week repeatability and the inter-observer agreement of the final staging system.


We have developed a five-stage system that is readily understood by community workers with little health training. Kappa for one-week repeatability was 0.88 (95% CI 0.80 to 0.96), Kappa for agreement between health professionals was 0.71 (95% CI 0.60 to 0.82), while that between health professionals and community podoconiosis agents without formal health training averaged 0.64 (95% CI 0.52 to 0.78).


A simple staging system with good inter-observer agreement and repeatability has been developed to assist in the management and further study of podoconiosis.

Keywords: podoconiosis, non-filarial elephantiasis, agreement, repeatability, community worker, staging system


Podoconiosis (endemic non-filarial elephantiasis) is a geochemical disease occurring in individuals exposed to red clay soil derived from alkalic volcanic rock (Price 1976). It is a chronic and debilitating disorder of considerable public health importance in areas of high prevalence. These areas include parts of Uganda (Onapa et al. 2001), Tanzania (de Lalla et al. 1988), Kenya (Crivelli 1986), Rwanda, Burundi, Sudan and Ethiopia (Price & Bailey 1984), in the east of Africa and Equatorial Guinea (Corachan et al. 1988), Cameroon, and the islands of Bioko, Sao Tomé, Principe (Ruiz et al. 1994) and Cape Verde to the west. The condition has also been reported in the central American highlands from Mexico to Ecuador. Although filarial elephantiasis predominates in India, podoconiosis has been reported from north-west India, Sri Lanka and Indonesia. Podoconiosis was previously common in North Africa (Algeria, Tunisia, Morocco and the Canary Islands) and Europe (France, Ireland and Scotland), but is longer found in these areas (Price 1990).

In Ethiopia alone, 11 million people (18% of the population) are at risk through exposure to the 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 (Destas et al. 2003). Although rarely recognized outside endemic regions, podoconiosis is more common than HIV infection, TB or filarial elephantiasis in these regions. The majority of individuals with podoconiosis are from the economically active section of society. Recent research has demonstrated that people with podoconiosis lose 45% of total productive work days when compared to unaffected people, costing a single zone of 1.5 million inhabitants more than US$16m per year (Tekola et al. 2006). Podoconiosis thus contributes significantly to poverty in affected communities.

Despite the prevalence and public health importance of podoconiosis, there is as yet no accepted clinical staging system. A validated clinical staging system would enable clinicians to assess the results of medical and surgical treatment, and would enable investigators to document the effects of public health preventive interventions.

Globally, many different systems are used to stage lymphedema (Ryan, 2004), giving rise to difficulties comparing studies and trials. The Dreyer system for staging filarial lymphedema (Dreyer et al., 2002) has been used in a range of tropical countries to refine and improve management of lymphedema, and observer agreement has recently been assessed and found to be very good (McPherson et al., 2006). The system describes seven stages, which though not necessarily progressive, are related to the risk of acute attacks and to the complexity of lymphedema management. Since lymphedema is a major manifestation of podoconiosis, and the Dreyer system was developed as a practical tool for use in the community, we decided to adapt it for use in podoconiosis, and then to assess the inter-observer agreement and repeatability of the adapted staging system.

Materials and methods

Study area

The study was conducted in Wolaita Zone, Southern Ethiopia where the prevalence of podoconiosis is more than 5% (Destas et al. 2003). In Wolaita zone, a local non-governmental organization, the Mossy Foot Treatment and Prevention Association (MFTPA) is active and currently works in 14 outreach sites. The MFTPA works with the community at grass-roots level and many patients in the zone have become beneficiaries of the services provided.

Study design

i) Development of clinical staging system

The clinical staging system was developed in a series of four re-iterative field tests between November 2006 and June 2007. The initial draft was based on the 7-stage Dreyer scale, namely: swelling reversible overnight; swelling not reversible overnight; shallow skin folds; knobs; deep skin folds; mossy lesions and social disability. At each field test, the medical epidemiologist (GD) was present to record the responses of the other experts (MFTPA field director (MA)); tropical dermatologist (ZA); or international dermatologist (CF) as subsequent versions of the staging system were field tested. Each field test was held at two or more field sites to ensure that at least 100 patients could be reviewed. Sites were selected according to the regular schedule of the MFTPA. New and returning adult patients were assessed, and where possible, digital photos were reviewed to follow disease course.

After each field test, health professionals debriefed and revised the staging system, taking into account -

  1. features associated with duration of disease
  2. features associated with severity of disease or complications
  3. features associated with regression of disease with treatment

After each revision, the revised staging system including photos was circulated for comment to two international experts (Dr Gerusa Dreyer and Prof Rod Hay). Their comments were incorporated into the version used for the next field test.

ii) Assessment of repeatability and reliability

Study setting and population

Once the final staging system had been decided, a cross-sectional study was performed to assess inter-rater reliability and a follow-up study to assess one-week repeatability. Four MFTPA sites were used: Boditi, Shanto, Gesuba and Dekaya, and new adult patients were examined at each site.

Sample size

The following formula was used to calculate sample size, based on the precision with which the standard error of Kappa can be measured13:


We assumed the proportion of pairs agreeing by chance (pe) on a 1-5 scale would be 0.2, thus given 50 patients (100 legs), we would be able to calculate a Kappa of 0.60 with a standard error of 0.06, and thus 95% confidence intervals of +/−0.12.

Data collection

A public health specialist (FT) and a medical doctor (DH) trained four Community Podoconiosis Agents (CPAs, treated patients with no formal health training who act as outreach clinic co-ordinators) in use of the staging system. Training involved a brief (two-hour) session in which each stage was defined, and practically demonstrated with patients attending the outreach clinic. The two health professionals and one CPA at each site then examined both legs of patients not included in the training exercise, and staged them blind to each other’s assessments. Demographic data, duration of signs and symptoms and history of shoe-wearing were also recorded for each patient.

The stage of the disease was graded again by one of the health professionals on separate forms one week after the initial scoring on 31 patients. The examiner was blind to the stage given previously.

Data analysis

SPSS V.11.05 was used to derive simple frequency distributions and cross tabulations. Since the marginal distributions of stages did not differ significantly between observers, Kappa analyses were carried out in SPSS version 11.05 to determine the degree of intra-observer and inter-observer agreement. Inter-observer agreement was assessed between the two health professionals, and between the CPAs and the health professionals. The international experts considered that weighting of stages was unnecessary, so weighted Kappa was not calculated.

The sensitivity and specificity of CPAs’ staging was assessed with reference to a ‘gold standard’ diagnosis, taken to be those instances on which both health professionals agreed a stage.


i) Development of staging system

Four field tests were performed, with between 119 and 146 patients examined during each field test. Between each field test, the staging system was reviewed by the two international experts. The staging system thus went through six revisions, the seventh draft not being substantially changed after the fourth field test.

Final version of staging system

The seventh and final version of the staging system is presented in Annex A. One version for CPAs and a more detailed version for health professionals are shown. The staging system has five mutually exclusive levels according to the proximal spread of swelling and dermal nodules, bands and ridges. Proximal spread is measured in relation to easily identifiable landmarks (the malleoli and the upper border of the patella). When recording the stage, the presence or absence of mossy changes and the greatest below-knee circumference are also noted.

ii) Assessment of repeatability and inter-observer reliability

Characteristics of the study subjects

Fifty-two new patients attending the four outreach sites were each examined by two health professionals and one CPA, giving a sample of 104 legs. Thirty-one of these patients returned the following week for re-assessment by the first health professional. Table 1 illustrates the characteristics of study participants. Most were women, most did not wear protective shoes, and most had had podoconiosis for less than 10 years.

Table 1
Characteristics of the study subjects (n=52)

Inter-observer agreement

There was good agreement between the two health professionals (K=0.71, 95% CI 0.60 to 0.82), and lower (but still ‘good’ according to Landis & Koch, 1977) agreement between the health professionals and the CPAs (average K=0.64, 95% CI 0.52 to 0.78). Raw data are shown in Table 2. For all the raters, more than a quarter of the sampled legs were rated as either stage one or two. No leg examined was considered stage 4.

Table 2
Inter-observer agreements

One week re-test repeatability

Table 3 shows the initial and one week re-test ratings for one health professional (62 legs). There was a very high level of agreement in the two ratings (K=0.88, 95% CI 0.80 to 0.96).

Table 3
First and second ratings for health professional 1 (n=62)

Sensitivity and specificity

Eighty-two (80.4%) of 102 legs were given an identical stage by both health professionals, and were considered the ‘gold standard’ for testing the sensitivity and specificity of staging by the CPAs. The sensitivity and specificity of the CPAs’ staging were high (81.3% to 100%, and 85.7% to 100%, respectively). Close to or exactly 100% sensitivity and specificity was found for stages 0, 3 and 5 (Table 4).

Table 4
Sensitivity and specificity of ratings by community podoconiosis agents (n=82)

Discussion and conclusion

The purpose of this study was to develop a clinical staging system for podoconiosis, to test the ease of its use in the field, and to quantify the inter-observer reliability and one-week repeatability.

We have developed a straightforward staging system which reflects disease duration, severity and response to treatment and has content (face) validity as assessed by the lymphedema experts reviewing it. Demonstrating content validity is a necessary step, but is not itself sufficient (Flaherty et al, 1988). Construct validity would be difficult to measure, since no strong conceptual basis existed previously for podoconiosis staging. Measurement of concurrent validity would also be problematic given the lack of gold standard for podoconiosis staging. We therefore suggest that demonstration of predictive validity through follow-up studies will be important in the future.

The staging system was easy to convey to community podoconiosis agents familiar with the condition but without formal health qualifications: a two-hour training session with patients was sufficient for the four CPAs in the reliability study. We intend to develop a laminated A5-size card with photographs illustrating each stage for future training and easy reference in the field. The system was feasible to use under field circumstances, in which clinics of approximately of between eighty and one hundred and fifty patients are reviewed under trees or in very simple shelters. The trained CPAs spent two to three minutes evaluating each leg, which is acceptable if the scale is used for treatment evaluations in future.

Measurement of reliability used unweighted Kappa because the experts consulted did not consider that disagreements between different stages should be allocated different weights. Inter-observer agreement (between the health professionals and between the health professionals and the CPAs) was good (0.71 and 0.64, respectively), but slightly lower than the agreement measured between a physician and a clinic nurse using the Dreyer scale in filarial lymphedema (McPherson et al, 2006). The authors comment that agreement may have been increased by the use of a clinical questionnaire in addition to the Dreyer scale. Intra-observer repeatability was very good (0.88), and the diagnostic validity (sensitivity and specificity) of the community workers in assigning correct clinical stages to the legs of the patients was also high. More extended training and practice would be expected to improve the ability of the CPAs to use the staging system in a reliable and repeatable manner.

Application of this grading system will be vital to the Mossy Foot Treatment and Prevention Association or any other group developing or refining podoconiosis intervention programs. Since it reflects disease duration and severity, it may also act as a proxy marker for access to care: patients in communities with adequate access to podoconiosis care would be expected to have lower average scores than those without access to care. Although this staging system has been shown to be reliable and repeatable, feedback from clinicians and others working with podoconiosis patients will be warmly welcomed to further refine and develop it. We also expect that wider use of the staging system will lead to the development of clinical algorithms to guide management of patients at each stage of disease.


We are endebted to Dr Gerusa Dreyer and Professor Rod Hay for critically reviewing several drafts of the staging system and for ascribing weights to the final system.

Funding and independence.

This study was funded by the UK Association of Physicians Links with Developing Countries Scheme, 2006/2007, and the Wellcome Trust, UK, project grant 079791. Neither funding body had any role in the study design, the collection or analysis of data, or the writing or submission of this paper.



Podoconiosis Staging Sheet (Podoconiosis-Endemic Areas) Staging For Field Workers


The field worker is expected to look at and examine the right and left leg of each patient in turn, and give a score to each leg separately.

  • ‘Swelling’ here means a general increase in size of part of the foot or leg.
  • ‘Reversible swelling’ here means a swelling that is not present when the patient first gets up in the morning and becomes more marked as the day advances.
  • ‘Persistent swelling’ here means a swelling that is present all the time.
  • ‘Knob or bump’ here means a discrete, hard lump that can be seen or felt to protrude from the rest of the foot or leg.
  • ‘Ankle’ here means the level of the two ankle bones when the patient is standing.
  • ‘Knee’ here means the level of the top of the knee cap when the patient is standing.

In addition to the numerical stage, the field worker should measure the greatest below-knee circumference and record the presence (M+) or absence (M−) of mossy changes. For example, if a patient’s right leg has irreversible below-knee swelling, nodules below the ankle, mossy changes around the heel and a circumference of 48cm, the staging should be recorded as Stage 2, M+, 48.

Stage 1. Swelling reversible overnight

The swelling is not present when the patient first gets up in the morning.

Stage 2. Below-knee swelling that is not completely reversible overnight; if present, knobs/bumps are below the ankle ONLY

Persistent swelling that does not reach above the knee. If knobs or bumps are seen or felt, they are only present below the ankle, NOT above the ankle.

Stage 3. Below-knee swelling that is not completely reversible overnight; knobs/bumps present above the ankle

Persistent swelling that does not reach above the knee. Knobs or bumps can be seen or felt above the ankle as well as below.

Stage 4. Above-knee swelling that is not completely reversible overnight; knobs/bumps present at any location

Persistent swelling that is present above the knee. Knobs or bumps can be seen or felt at any place on the foot or leg.

Stage 5. Joint fixation; swelling at any place in the foot or leg

The ankle or toe joints becomes fixed and difficult to flex or dorsiflex. This may be accompanied by apparent shortening of the toes.

Podoconiosis Staging Sheet (Podoconiosis-Endemic Areas) Description For Health Professionals

A more detailed and technical description of the changes that may be present at each stage is given, though the definitions for each stage remain the same. The stages represent severity of disease, and do not necessarily represent the disease process: it is possible, for example, for an individual to have stage 5 disease but never to have had above-knee swelling. The following terms are used in the descriptions:

  • Dermal nodules: elevated, non-translucent lesions >0.5cm diameter, with width approximately equal to length.
  • Dermal ridges: elevated lesions >0.5cm width, with length greater than width
  • Dermal bands: palpable, but non-elevated ridges
  • Mossy changes: round or fusiform, either fluid filled (and hence translucent) lesions, or papillomatous hyperkeratotic horny lesions giving the skin surface a rough velvet-like appearance.

Stage 1. Swelling reversible overnight

The swelling is not present when the patient first gets up in the morning. Changes such as hyperpigmentation and nail dystrophy are unusual, but may be seen. The swelling is usually confined beneath the ankle.

Stage 2. Below-knee swelling that is not completely reversible overnight; if present, knobs/bumps are below the ankle ONLY

Persistent swelling that does not reach above the knee. If present, knobs or bumps do not extend beyond the ankle.

The ‘knobs or bumps’ may take the form of dermal nodules, ridges or bands. Tourniquet-like effects may be observed at this stage or any subsequent stage, depending on the position of dermal ridges and nodules in relation to joints.

Mossy changes may be apparent, but their presence depends on a range of factors including the use of plastic footwear. Interdigital maceration and hyperpigmentation are often present at this stage, and nail dystrophy almost always present.

Stage 3. Below-knee swelling that is not completely reversible overnight; knobs/bumps present above the ankle

Persistent swelling that does not reach above the knee. Dermal nodules, ridges or bands are seen or felt above the ankle. Tourniquet-like effects are frequently observed at this stage

Any of the other changes mentioned for Stage 2 may also be present.

Stage 4. Above-knee swelling that is not completely reversible overnight; knobs/bumps present at any location

Persistent swelling that is present above the knee.

Any of the other changes mentioned for Stage 2 may also be present. In addition, signs of lymphectasia may be apparent, particularly on the thigh.

Stage 5. Joint fixation; swelling at any place in the foot or leg

The ankle or interphalangeal joints becomes fixed and difficult to flex or dorsiflex. This may be accompanied by adhesion and fusion of the toe web spaces, making the toes appear short or indistinct. Sensation is preserved. X-rays show tuft resorption and loss of bone density.


Conflicts of Interest

No author declares any conflict of interest associated with the publication of this manuscript.

Ethical approval.

Approval was granted by the ethical committee of the Faculty of Medicine, Addis Ababa University.

Contributor Information

Fasil Tekola, Brighton and Sussex Medical School, University of Sussex, Tel: +00251-911-12 60 43 Fax: +00251-16-627876 ; moc.oohay@aloketlisaf.

Zewdu Ayele, Department of Dermatovenereology, ALERT & Tikur Anbessa Hospitals, Addis Ababa, Tel: +00251-917-800610 ; ku.oc.oohay@eleyaudwez.

Dereje HaileMariam, Otona Hospital, Wolaita Sodo, Tel: +00251-911-054486, moc.oohay@311ejered.

Claire Fuller, East Kent Hospitals NHS Trust, Tel: +0044-7973-187025, ten.shn@relluf.erialc.

Gail Davey, Department of Community Health, Addis Ababa University, Addis Ababa, Ethiopia, Tel: +00251-911-388835 Fax: +00251-16-627876, te.tenoihte@rakruren.


  • Bland JM. An introduction to medical statistics. 3rd ed. Oxford University Press; 2000.
  • Corachan M, Tura JM, Campo E, Soley M, Traveria A. Podoconiosis in Equatorial Guinea. Report of two cases from different geological environments. Tropical and Geographical Medicine. 1988;40:359–364. [PubMed]
  • Crivelli P. Filarial elephantiasis in Nyambene range: a geochemical disease. East African Medical Journal. 1986;63:191–194. [PubMed]
  • de Lalla F, Zanoni P, Lunetta Q, Moltrasio G. Endemic non-filarial elephantiasis in Iringa District, Tanzania: a study of 30 patients. Transactions of the Royal Society of Tropical Medicine and Hygiene. 1988;82:895–897. [PubMed]
  • Destas K, Ashine M, Davey G. Prevalence of podoconiosis (endemic non-filarial elephantiasis) in Wolaitta, Southern Ethiopia. Tropical Doctor. 2003;32:217–220. [PubMed]
  • Dreyer G, Addiss D, Dreyer P, Noroes J. Basic Lymphedema Management. Hollis Publishing Co.; New Haven: 2002.
  • Flaherty JA, Gaviria FM, Pathak D, Mitchell T, Wintrob R, Richman JA, Birz S. Developing Instruments for Cross-Cultural Psychiatric Research. J Nervous & Mental Disease. 1988;176:257–263. [PubMed]
  • Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159–74. [PubMed]
  • McPherson T, Fay MP, Singh S, Penzer R, Hay R. Health workers’ agreement in clinical description of filarial lymphedema. American Journal of Tropical Medicine and Hygiene. 2006;74:500–504. [PubMed]
  • Onapa AW, Simonsen PE, Pedersen EM. Non-filarial elephantiasis in the Mt Elgon area (Kapchorwa District) of Uganda. Acta Tropica. 2001;78:171–176. [PubMed]
  • Price EW. The association of endemic elephantiasis of the lower legs in East Africa with soil derived from volcanic rocks. Transactions of the Royal Society of Tropical Medicine and Hygiene. 1976;70:288–95. [PubMed]
  • Price EW, Bailey D. Environmental factors in the etiology of endemic elephantiasis of the lower legs in tropical Africa. Tropical and Geographical Medicine. 1984;36:1–5. [PubMed]
  • Price EW. Podoconiosis: Non-filarial Elephantiasis. Oxford University Press; Oxford: 1990. pp. 1–35.
  • Ruiz L, Campo E, Corachan M. Elephantiasis in Sao Tome and Principe. Acta Tropica. 1994;57:29–34. [PubMed]
  • Ryan TJ. A search for consensus on the staging of lymphedema. Lymphology. 2004;37:180–181. [PubMed]
  • Tekola F, HaileMariam D, Davey G. Economic costs of endemic non-filarial elephantiasis in Wolaita Zone, Ethiopia. Tropical Medicine and International Health. 2006;11:1136–1144. [PubMed]