Conceived and designed the experiments: SG MM LF. Performed the experiments: SG GK SL BKI LF. Analyzed the data: SG LF. Contributed reagents/materials/analysis tools: MD. Wrote the paper: SG LF. Contributed substantially in the interpretation of the data. Specifically involved in the study design, the methods used, and in the manuscript revision: MM. Contributed substantially in the interpretation of the data: DL ES. Critical revision of the paper: PJG.
The authors have declared that no competing interests exist.
Until the 1970s the prevalence of non-venereal trepanomatosis, including yaws, was greatly reduced after worldwide mass treatment. In 2005, cases were again reported in the Democratic Republic of the Congo. We carried out a survey to estimate the village-level prevalence of yaws in the region of Equator in the north of the country in order to define appropriate strategies to effectively treat the affected population.
We designed a community-based survey using the Lot Quality Assurance Sampling method to classify the prevalence of active yaws in 14 groups of villages (lots). The classification into high, moderate, or low yaws prevalence corresponded to World Health Organization prevalence thresholds for identifying appropriate operational treatment strategies. Active yaws cases were defined by suggestive clinical signs and positive rapid plasma reagin and Treponema pallidum hemagglutination serological tests. The overall prevalence in the study area was 4.7% (95% confidence interval: 3.4–6.0). Two of 14 lots had high prevalence (>10%), three moderate prevalence (5–10%) and nine low prevalence (<5%.).
Although yaws is no longer a World Health Organization priority disease, the presence of yaws in a region where it was supposed to be eradicated demonstrates the importance of continued surveillance and control efforts. Yaws should remain a public health priority in countries where previously it was known to be endemic. The integration of sensitive surveillance systems together with free access to effective treatment is recommended. As a consequence of our study results, more than 16,000 people received free treatment against yaws.
Yaws is a communicable, non-venereal treponematosis caused by the bacteria
The evolution of yaws is revealed in three stages of clinical manifestations
In the early 1950s an estimated 50 to 100 million people were infected in tropical areas particularly in Africa, Southeast Asia and South America
However, difficulties encountered by many countries in integrating continued control measures into local health services have led to a gradual build-up and extension of the treponemal reservoir. Thus, after enthusiastic mass treatment campaigns to eradicate yaws until mid-1960s, there has been a strong recrudescence of this disease in the last decades
In DRC yaws was thought to be eradicated in the early 1960s. However, in 1975, a study revealed yaws cases among the pygmies in the north of DRC, one of the socio-economically most disadvantaged populations, with up to 90% of the population showing serological evidence of infection at that time
In early 2005, an increased number of yaws cases confirmed through clinical screening were reported from the rural Wasolo health zone, a very remote and isolated region in the north of DRC Equator province (
MSF and Epicentre conducted a study to estimate the village-level prevalence in the population of the rural Wasolo health zone using Lot Quality Assurance Sampling (LQAS)
Between February 16 and March 1, 2005, we carried out a community-based survey in the rural Wasolo health zone, Equator province (
All persons older than 6 months living in the 35 villages on the first day of the survey were eligible for inclusion. Children under 6 months of age were excluded for ethical reasons (e.g. taking blood samples under basic field conditions). Written, informed consent was obtained from all participants or his/her legal guardian before inclusion.
We used the LQAS method to classify the prevalence of active yaws in groups of villages (lots). The LQAS method has been shown to be useful in identifying areas with high prevalence of disease
where
To ensure lots had similar population sizes, we grouped between one to five villages together according to their geographical proximity, natural boundaries such as rivers or fields and population size (
We developed a systematic sampling scheme to randomly select 84 households (defined as a group of people living together and eating from the same pot) from each lot. The size of the sample taken from each village was weighted by the village population so that larger villages contributed more to the lot sample than smaller villages. A uniform sampling fraction was used. The overall lot sample was, therefore, a proportionate, stratified sample
As yaws is known to cluster in households, we decided to select only one person per household for participation in the study. This simplified sampling because it allowed households, rather than individuals, to be sampled. This approach has been validated previously in LQAS surveys of trachoma prevalence
All of the villages in the survey area consisted of a single ribbon of houses situated along a road. This made it possible to use a straightforward systematic sampling method to select households. The number of households in the village was estimated by counting doors. A sampling interval was calculated as the ratio of the estimated number of households to the required sample size in the village.
One end of the village was selected by tossing a coin. A random number between one and the sampling interval (inclusive) was read from a table of random numbers. Study teams walked towards the opposite end of the village, counting the houses that they passed, sampling the house indicated by the random number. Subsequent houses were sampled by repeated application of the sampling interval. Sampling stopped when the surveyors reached the calculated number of households for that village.
When a house was sampled, household members were listed and assigned a sequential number. One household member was then sampled at random using a random number table.
The day before study teams arrived in the villages, we visited the chief of the village to seek his permission to survey and to sensitise the population to be present on the day of the study. In the very few cases that households were empty and household members could not be located within a reasonable amount of time, the nearest household was sampled.
For the detection of yaws lesions, the study population was examined following a standardized checklist for clinical signs suggestive of yaws. Primary yaws chancre was defined as chronic non-genital, painless, non-tender papule (mother yaw), which bursts open and may present as a traumatic ulcer with raised margin located on the legs, feet or buttocks. Secondary lesions, which might be concomitant to the primary lesion, included painful raspberry-like cutaneous lesions (papilloma and papules) in various numbers eventually affecting palms or soles, and painful tender osteoperiostitis affecting the fingers (polydactilitis), nose (“goundou”), tibia or forearm. Tertiary late lesions were defined as mutilating facial ulcer around the nose (“gangosa”), skin gummata, hyperkeratosis of the palms or soles, juxta-articular nodules, or bone deformity including sabre shin tibia. Four teams of three persons each (medical officer, nurse and assistant) carried out the field investigation under the supervision of both an epidemiologist and a dermatologist. Each team took between three and four days to finish one lot. Prior to initiating the survey, the teams received two days of on-site training with special focus on the clinical diagnosis using illustrative pictures and including half a day to pilot the questionnaire.
In the case of clinical suspicion of yaws, individuals or their caretakers/parents if the individual was younger than 15 years were asked about the duration of the lesions and a blood sample (5 ml) was collected for rapid plasma reagin (RPR) and
Blood samples were centrifuged and stored at 2–8°C. Using a cold chain they were sent to the MSF laboratory in Botetenza (DRC) where an MSF laboratory technician performed all RPR (non-treponemal antigen, RPR-nosticon® II, Biomérieux, Holland) tests. For all RPR positive blood samples, a TPHA (treponemal antigen, Syphilis TPHA liquid®, Human, Germany) test was performed afterwards.
Participants were defined as having active yaws if they presented with cutaneous lesions clinically suggestive of yaws and scored positive on both the non-specific RPR test and the specific TPHA test.
Data were entered into the database using Epi-Data 3.0 software (The EpiData Association, Odense, Denmark). Double data entry and cleaning were performed to check for inconsistencies and errors were corrected. The analysis was done using Stata 8.0 (Stata Corporation, College Station; Texas, USA). In order to calculate the overall prevalence of yaws in the study population, the number of cases in the different lots was weighted according to each lot's population size and combined. Symmetric confidence intervals (95%CI) were calculated after weighting according to each lot's population size. All data presented take into account weighting if required.
None of the selected households refused to participate in the study. Demographic characteristics of the study population (n = 1176) are shown in
Study population | Persons with cutaneous lesions clinically suggestive of yaws | Persons with active yaws confirmed by RPR |
|
N = 1176 | N = 383 | N = 55 |
|
Median age in years (IQR |
23 (9–45) | 22 (8–45) | 21 (10–55) |
Range (Minimum, Maximum) | 1, 80 | 7, 72 | 2, 72 |
<5 (%) | 146 (12.4) | 50 (13.1) | 3 (5.6) |
5–14 (%) | 249 (21.2) | 94 (24.5) | 19 (34.5) |
15–29 (%) | 287 (24.4) | 86 (22.5) | 7 (12.7) |
30–45 (%) | 186 (15.8) | 54 (14.1) | 7 (12.7) |
≥45 (%) | 292 (24.8) | 99 (25.8) | 19 (34.5) |
Unknown (%) | 16 (1.4) | – | |
Sex ratio (male/female) | 1.2 (643/526) |
1.6 (236/146) |
1.8 (35/20) |
Median household size [IQR |
7 |
7 |
7 |
Main household activity | |||
Farming (%) | 852 (72.4) | 281 (73.4) | 42 (76.4) |
Rapid Plasma Reagin test.
1 missing data for all listed variables.
Inter-quartile range.
7 missing data for sex.
1 missing data for sex.
Two of 14 lots presented more than seven confirmed active yaws cases, indicating a high prevalence of over 10%. In these lots, total mass treatment was the chosen treatment strategy. For three lots, the number of confirmed active yaws cases was between four and seven (inclusive), indicating a moderate prevalence of 5–10%. Here, juvenile mass treatment was chosen as the treatment strategy. For the remaining nine lots, the number of confirmed active yaws cases was three or lower, indicating a low prevalence of below 5%. Selective mass treatment was chosen as the treatment strategy (
Study lot Village name number | Number of active yaws cases | Prevalence | Treatment strategy used |
|||
suspected | confirmed | |||||
clinically |
+ RPR |
+ RPR + TPHA |
||||
1 | Vadale, Lima, Vato | 15 | 7 | 3 | low | SMT |
2 | Toli, Lunza, Vangbandi, Tongunde | 18 | 7 | 1 | low | SMT |
3 | Liboko, PEFA I+II, Bige, Ndesogo | 22 | 13 | 2 | low | SMT |
4 | Ngaba, Soli | 15 | 10 | 0 | low | SMT |
5 | Tiri, Tari Bondi, Ndanu | 39 | 21 | 3 | low | SMT |
6 | Torunga, Konzi, Ngbongbo Kondia | 33 | 25 | 4 | medium | JMT |
7 | Ngbongbo Vatiri, Ngbongbo Sida, Ngbongbo Tongonze | 36 | 18 | 11 | high | TMT |
8 | Ngbongbo Tokonzi, Mando | 36 | 26 | 15 | high | TMT |
9 | Modale I | 24 | 11 | 4 | medium | JMT |
10 | Modale II, Nzinga | 20 | 6 | 1 | low | SMT |
11 | Bige, Lupu, Ngwado | 42 | 28 | 6 | medium | JMT |
12 | Ndangba, Kuzangu | 14 | 9 | 3 | low | SMT |
13 | Vamburu - Beta/Koto | 33 | 25 | 1 | low | SMT |
14 | Wasolo mission | 36 | 17 | 2 | low | SMT |
TOTAL | 383 | 223 | 56 |
Diagnosed clinically.
Rapid Plasma Reagin test.
TMT (total mass treatment); JMT (juvenile mass treatment, treatment of active cases, their direct contacts and all children under 15 years); SMT (selective mass treatment, treatment of both active cases and direct contacts).
Among 1176 persons examined, 383 (32.6%) presented with cutaneous lesions clinically suggestive of yaws and were tested by RPR. Of these, 223 (58.2%) were RPR positive. Out of the 223 RPR positive individuals, 56 (25.1%) were TPHA positive and defined as having truly active yaws. The overall prevalence of active yaws in the rural Wasolo health zone was 4.7% (95%CI: 3.4–6.0, 56/1176). When separated in two age groups, the overall prevalence of active yaws was 6.1% (95%CI: 3.4–8.9, 22/395) for persons less than 15 years of age and 4.0% (95%CI: 2.6–5.5, 33/765) for those being 15 years and older (age data missing for 16 study participants in the denominator and one case in the numerator).
Among the 383 persons with cutaneous lesions clinically suggestive of yaws, 37.6% (144/383) were less than 15 years of age (
Fifty-six persons had concomitant RPR and TPHA positive tests confirming active yaws. Detailed information was obtained for 55 of them (
Knowledge of the prevalence of yaws is critical to plan appropriate treatment strategies according to current WHO recommendations. As it is well-known that the disease is clustered and that its prevalence can vary greatly between villages, we used the LQAS method as a rapid way of assessing the prevalence in groups of villages instead of using classical sampling in the entire population. The finding that some lots had a high prevalence of over 10% while the overall prevalence was 4.7% underlines the benefit of this strategy. If the treatment decision had been based solely on overall prevalence (defined as ‘moderate’ in our study), total mass treatment would not have been carried out in any of the villages, even though some displayed high prevalence of the disease. In addition, when considering only the overall prevalence estimation, it would have been difficult to follow treatment recommendations because the 95% confidence interval of 3.4 to 6.0 encompassed the critical 5% threshold level recommended by WHO to define appropriate distinct treatment strategies.
In April 2005 following this study, MSF launched a campaign offering free treatment to more than 16,000 people in the villages of the rural Wasolo health zone based on the prevalence estimates in the different groups of villages. Our study nicely illustrates the benefit of using LQAS as an operational research method to rapidly determine the prevalence of yaws in distinct areas in order to define appropriate treatment strategies.
Some data and methodological limitations of our study should be mentioned. The different study teams had varying levels of clinical skills in diagnosing the disease. In the context of a known yaws outbreak, clinical diagnosis may lead to an overestimation of potential active yaws cases. Systematic serological tests of suspected cases were performed to confirm the diagnosis and illustrated this overestimation problem. A relatively low percentage of suspected cases were confirmed serologically. For instance, lesions of the soles, which are very frequent in rural populations not wearing shoes, where most commonly clinically overdiagnosed as yaws but not confirmed by RPR and TPHA tests. However, these tests have themselves well-known limitations. Indeed,
Finally, we observed that only about one quarter of the RPR positive cases were confirmed by TPHA. It is known that the RPR test is not specific for antibodies to
Between December 2004 and February 2005 more than 2500 cases with cutaneous lesions clinically suggestive of yaws were reported by health personnel working in different health centres in the rural Wasolo health zone, an isolated region in the North of DRC (personal communication, main health office, rural health zone Wasolo, 2005). These findings were confirmed by our survey revealing an overall prevalence of active yaws of 4.7% (95%CI: 3.4–6.0). This prevalence estimate was similar to that reported in the few yaws outbreaks documented in other African countries, such as in Ghana in 1981 (4.0%)
The presence of yaws in the rural Wasolo health zone is very likely to be the consequence of a progressive increase of yaws cases in an area where the disease was supposed to be eradicated but where new cases had been reported in the past
Among measures for potential eradication of the disease, the implementation of a long term yaws control system with free access to treatment is urgently needed to avoid the further spread of this debilitating disease in the rural Wasolo health zone but also in neighbouring health zones and districts. Horizontal programmes in which community-based health workers are involved in all aspects of disease control have yielded very positive results in Ecuador
In conclusion, yaws is a highly curable neglected tropical disease. Our study showed the advantage of LQAS as a rather uncommon sampling method to rapidly estimate the prevalence of yaws in order to treat affected populations according to WHO recommendations. The control of yaws should remain a public health priority. The disability and disfigurement resulting from this disease cause individual suffering and will also have indirect and largely unrecorded effects on the morbidity and mortality of affected populations.
We would like to express our sincerest thanks to both national and local Ministries of Health for providing permission to carry out this study. We are very grateful to all the field teams of Médecins Sans Frontières who contributed to this effort, and to our very dedicated team of Congolese interviewers and translators. Very special thanks are addressed to the population in the rural Wasolo health zone who participated in the study.