A cross-sectoral approach to addressing Konzo in DRC

By Dr. Emery Kasongo and Muriel Calo

Emery Kasongo is the ACFUSA Project Manager who led the study described. Previously he has worked as national consultant on several food security studies in the Democratic Republic of the Congo (DRC) and was a lecturer at the University of Lubumbashi.

Muriel CALO is a Food Security & Livelihood Advisor for ACFUSA, supporting the DRC-Kinshasa mission.

Special thanks go to the whole of the ACF study team, namely Benjamin Kitiko, Guillain Yumba Kabenge, Guy Munsansa , Damien Naimana, Césarine Kuwa, Nono Bumba, Paul Bahati, Jean Ngoie Mandaku, Roger Mubake, Rodolphe Mwamukenza, Richard Muhongo, Jacques Mahunda, Axel Matondo, Timothée Mutalala and Sylvain Lumbala. Thanks also to Pierre Kadet, the PRONANUT (Programme National de Nutrition in DRC), and the communities who participated in the study. The funding support of the European Union (EU) Food Facility Programme is gratefully acknowledged.

In November 2009, Action Against Hunger (ACF-USA) launched a 22-month long intervention in the Bandundu province of the Democratic Republic of Congo (DRC) to address several factors underlying the Konzo epidemic affecting the population of Kwango district (see Figure 1). The 'Integrated Programme for the Eradication of Konzo in the Territory of Kwango in DRC' project is financed by the European Union (EU) Food Facility. It aims to eradicate the disease through a crosssectoral approach that focuses on nutrition education and training, dietary diversification, improved water access and agricultural processing. A total of 12,500 households are expected to benefit from these activities.

Prior to implementing these activities, ACFUSA conducted both a baseline epidemiological study and a project impact study to gather information that would inform the design of the intervention. The initial baseline study was completed in August 2010 and involved 113 villages surveyed across 11 health zones. A total of 2,388 suspected Konzo cases were screened. Forty-six focus groups and 35 semi-structured interviews were held with key informants. The findings of the study have so far been disseminated to key stakeholders in Bandundu and Kinshasa, and are shared in this article. A follow-up article will be published at the end of the project (October 2011) to share results on the impact of the intervention.

What is Konzo?

Symptoms

Konzo is a sudden epidemic spastic paraparesis (paralytic) disease which leads to a permanent, symmetrical (but non progressive) paralysis of the affected person's lower limbs. It is a neurological ailment triggered by sustained dietary exposure to the cyanide present in improperly processed cassava. Symptoms of mild cases include trembling in both legs and walking difficulties. Severely affected individuals are incapable of walking without support and suffer from speech impediments as well as visual impairments. Konzo itself is not fatal, but its debilitating effects heighten the risk of morbidity and mortality from other diseases. Furthermore its disabling effects result in practical, social and economic challenges for individuals, and families of individuals, living with the limited physical capacity induced by Konzo. The disease usually appears in clusters within households, as exposure comes from food consumed as a family meal.z

Aetiology of the disease

While the full aetiology of Konzo remains unclear, medical studies have attributed the appearance of the disease in DRC, Central African Republic, Tanzania, Mozambique and elsewhere to diets combining both high levels of cyanide and low quantities of sulphur proteins (pulse crops such as peas and beans)¹. As a natural defence mechanism, cassava tubers and leaves produce two types of cyanogenic glycosides: linamarin and lotaustralin. These glycosides are decomposed by the linamarase enzyme, naturally present in cassava, to liberate hydrocyanic acid (HCN). Amino acids, which are the building blocks of protein, contribute to the elimination of certain dietary toxins. Diets deficient in protein can therefore lead to an accumulation of cyanate in the blood, as the body does not have sufficient sulphur-based amino acids to break down and eliminate the cyanide.

A Konzo case in Kwango District, Bandundu Province, DRC

While cyanide is naturally present in all kinds of cassava, bitter varieties contain much higher levels than sweet varieties (1g cyanide/kg fresh tuber vs. 20mg/kg fresh tuber, respectively). Additionally, cassavas grown during drought are known to contain particularly high levels of cyanide. As a result of this natural toxicity, cassavas must be properly processed before they can be safely consumed. This process requires that they be peeled, grated and soaked in warm water for several days. This allows the linamarase enzyme to convert the linamarin into sugar and cyanide gas (which usually disperses harmlessly). The cassava should then be completely dried before being consumed or turned into flour.

Overall, vulnerability to Konzo is heightened by the combination of low protein intake (associated with low diet diversity), poor soil conditions (which favour the cultivation and consumption of bitter cassava varieties), and a lack of sufficient water resources for thorough processing.

Characteristics of Bandundu province

Kwango district is situated in the south of the Bandundu province and comprises 14 health zones, themselves sub-divided into health areas. Kwango is characterised by very poor soils, which explains the prevalence of bitter cassava varieties in the district. Bitter cassava has greater pest and disease resistance and yields higher outputs even in poor soils.

A July 2007 nutrition survey conducted by MEMISA (an international non-governmental organisation (INGO) active in the area) determined that global acute malnutrition (GAM) rates were 21.1% (NCHS, 95% CI: 12.9%-29.4%) in Kahemba Health Zone, Kwango district. Following these findings, ACF implemented a nutritional programme in Kahemba from September 2007 to July 2009, as well as two food security and livelihood programmes between 2008 and 2009. While these interventions sought to improve the nutritional situation of the population, they did not exclusively target Konzo affected individuals and households. An ACF led nutrition survey conducted in December 2008 determined that the nutritional situation had improved in Kahemba (GAM 10.2% [NCHS, 95% CI: 7.0%-13.3%]), but that Konzo still represented a major challenge in the area.

Konzo accelerates and worsens malnutrition, and can be fatal if victims remain untreated. In 2010, reports from Kahemba continued to signal the appearance of new Konzo cases. From January to June 2010 alone about sixty new cases were reported.

The baseline study

Objectives of the study

The objective of the study carried out at the outset of the programme was to provide key baseline information to inform the design of programme activities. The overall programme would integrate three of ACF's technical areas of expertise: Food Security and Livelihoods (FSL), Nutrition/Health, and Water, Sanitation and Hygiene (WASH). Prevention is a central activity in the management of the spread of the disease. In particular, the study gathered data on:

• The incidence and epidemiology of Konzo disease in the district of Kwango
• Level of cyanide consumed by the population
• Nutritional status of Konzo patients
• Local beliefs surrounding the origin of Konzo disease
• Dietary diversity, beliefs and practices
• Agricultural crops and processing techniques
• Food access and availability
• Water sources and access

The study sought to establish correlations between these factors to understand better the underlying causality of the disease.

Methodology

The study was conducted in 11 of the 14 health zones in the Kwango district: Boko, Popokabaka, Kasongo Lunda, Wamba Luadi, Kitenda, Kenge, Feshi, Panzi, Kajiji, Kisanji and Kahemba (total population: 1,670,029) in the months of March through May 2010. See Figure 2 for study sites.

Data were collected using the following methods:

a) Control group studies This involved comparing households with Konzo cases, to two control groups:

• Control households without Konzo cases in affected villages
• Control households in non-affected villages

Villages with high incidence levels of Konzo were identified using a purposive sampling method, with local leaders assisting ACF staff in listing any person with walking difficulties. A total of 2,218 Konzo cases were confirmed in 113 studied villages and in-depth physical examinations were carried out on about one-third of these cases. Urine samples and cassava flour samples were collected on 15% of surveyed households (Konzo-affected and control groups). A questionnaire was administered to about one-third of households (Konzoaffected and control groups) to better understand causal elements and other socio-economic factors underpinning the appearance of the disease (namely food consumption, cassava processing, water access and overall food security).

b) Focus group discussions Two focus groups (one male and one female) were organised per village (with or without Konzo cases). A total of 46 focus group discussions were carried out across Kwango to gather first hand information on local perceptions of the disease, local agricultural and nutritional practices, as well as water sources and access.

c) Semi-structured interviews Key informants (territory administrators, city and village chiefs, priests/pastors, teachers and agricultural experts) were interviewed better to understand the aetiology of the disease and obtain local advice about feasible interventions. A total of 35 semi-structured interviews were conducted across 11 health zones.

Nutritional data collection and analysis

Data collection was undertaken by three teams (four people each), responsible for one specific area of technical expertise over a period of 75 days. Local investigators were hired and supervised by experts from Kinshasa.

Nutritional data was analysed using Excel and EPI INFO software. Weight/height ratio was calculated for a total of 374 Konzo affected individuals under 18 years of age and Body Mass Index (BMI) for a total of 412 affected individuals over 18 years of age. As a precautionary measure it must be noted that the muscular atrophy linked to Konzo can result in biased anthropometric results and lead to an overestimate in the prevalence of malnutrition.

Food Security and WASH data collection and analysis

The food security part of study was carried out on a sample of 722 Konzo affected households and 112 control group households. Data relating to both Food Security and WASH were analysed using Excel.

Baseline study findings

The study determined that women were slightly more affected by Konzo than men (see Table 1). Out of the 2,218 confirmed Konzo cases encountered, 53.3% were female and 46.7% were male. Children under the age of 15 years represented the overwhelming majority of cases (73.9%) while individuals over the age of 15 years only represented 26.1% of cases. This suggested a high mortality rate among Konzo cases, given the non-reversible nature of the disease. The average Konzo prevalence rate across the 51 health areas visited was found to be 1.07%. Strong disparities were observed between health zones as prevalence ranged from 2.47% in Kajiji to 0.10% in Kisanji. The vast majority of cases (83%) were located in savannas, while 1% lived in hillside areas and 4% in valleys. See Figure 3 for Konzo prevalence rate by health zone and Figure 4 for number of cases by health zone.

Nutritional status

More than one quarter (25.8%) of Konzo cases under 18 years were diagnosed with GAM - out of which 12.6% suffered from severe acute malnutrition. Nutritional oedema was observed in 8.7% of those affected. Malnutrition rates did not significantly vary across gender, but younger children (ages 5 to 11 years) were slightly more affected than older ones (ages 12 to 17 years). BMI measures indicated that over half of Konzo cases (56%) over 18 years suffered from severe acute malnutrition, while 13.3% of Konzo cases were diagnosed with moderate acute malnutrition. Of the 199 female Konzo cases over 18 years, 31.2% were severely malnourished (MUAC<21 cm).

Seasonality and Konzo occurrences

Our study determined that although new Konzo cases appear all year round, much higher numbers of new occurrences (incidence) were observed during the dry season (May to September) (see Figure 5). While the average over one year is 49 new cases per month, 51 were reported in May, 116 in June, 181 in July and 69 in August. Water access is most limited between the months of May and September (dry season) and food availability is very poor between June and September (lean season). Women are thus more inclined to use less water to process or ret2 their cassavas, and use the same water over and over again. Additionally, retting time is often cut down during the lean season as households have very little to eat and end up consuming their cassavas too quickly, before it has been properly detoxified. The appearance of new Konzo cases is clearly higher during this critical time of the year, and can most likely be linked to these seasonal hardships.

Local customs and Konzo incidence

A Konzo case in Kwango District, Bandundu Province, DRC

Local beliefs and traditional customs were also identified as factors surrounding the appearance of Konzo. Our team learned that the disease is in fact considered to be 'a bad luck spell'. This makes it difficult to sensitise the population to prevention techniques, as locals see the condition as 'beyond their control'. In addition, certain local eating customs that favour the male head of household may contribute to making women and children much more vulnerable to Konzo. The male head of household is always served the largest portion, regardless of the overall quantity of food available to the household. These eating practices make women and children much more vulnerable to malnutrition as they are forced to only eat whatever is left over. Their diets therefore lack diversity and are insufficient. This trend is further enhanced by a number of taboos prohibiting women - particularly during pregnancy - from consuming certain protein-rich food items (turtle, monkey, eggs, avocados and nuts). This finding suggests a positive correlation between local dietary practices and taboos, and the higher Konzo incidence rates faced by women and children.

Agricultural practices and Konzo incidence

Most households were found to rely primarily on agriculture for food and quasi-uniformly cultivate cassava as their principal crop (90% of surveyed households cultivate at least one cassava crop per year). Maize, groundnuts and beans were secondary crops in most cases, except in the Kisanji zone where millet was produced and consumed relatively heavily. Despite these rather uniform crop cultures, diverse environmental factors (such as soil fertility or soil water retention) do affect the quantity and quality of harvests across areas, and could ultimately have an impact on the nutrition of the populations in question.

In fact, Kisanji, which had the lowest Konzo incidence rate, is the only health zone to produce and consume millet (see Figure 6). This raises a hypothesis that should be further explored, whether the presence of millet in diets can explain lower Konzo rates. It is difficult to establish a clear link between the consumption of millet and Konzo incidence rates as the study only collected snapshot data. It would therefore be necessary to conduct a study over a longer period of time to further explore this relationship to ascertain whether the promotion of millet as a complementary or alternative staple food crop would be an effective strategy in the prevention of new Konzo cases in affected areas.

Water access

Water access is a critical factor in preventing further Konzo occurrences in the Kwango district. Accessibility is limited by two factors: distance to springs (reliable springs are often far away) and seasonality of springs (poor yield in dry season).

Households in Kwango have access to less than the minimum standard of 10L/person/ day, and women generally use large containers (basins between 5 and 25 litres) to collect and transport water. An average of 50 to 60 minutes is devoted for each water collection roundtrip, twice a day (morning and evening). Water coverage levels are very low, with 5% coverage in the Feshi health zone and 4.3% In Kajiji (based on standard norms of 350 people per source). In Kajiji, 16 out of 67 existing sources had been protected, but the majority of these water points remain extremely difficult to access as they are situated on slopes and ramps. In Feshi, only seven out 64 existing sources were protected and similar challenges applied in terms of accessibility.

Because of these particular challenges, women in rural areas most often prefer to directly soak the cassava on river banks, in ponds or in swampy areas in order to avoid carrying heavy quantities of water back to their homes. In urban areas, however, women prefer to ret the cassava in their homes (in buckets or barrels) due to the high likelihood of theft if the cassava is left overnight in a public area. This can be very challenging as the quantity of water available in urban areas is often insufficient, and women often re-use the same water three or four times to ret their cassavas. This increases the likelihood of cyanide intoxication. An alternative technique also used to reduce the risk of theft involves digging a hole in swamp or riverside sites near cassava growing locations, filling it with water to permit the retting process, and then covering it with earth to conceal it. This practice is also hazardous because the cyanide that should be harmlessly released in the retting process is prevented from escaping by the earth layer.

Recommendations

Based on the findings, ACF-USA issued the following key recommendations.

• Training of key health zone staff, certified nurses, local leaders and community mobilisers to raise awareness of factors that increase vulnerability to Konzo and of processing techniques that decrease the cassava toxicity.
• Water access to be improved in the urban areas of Kahemba, Kajiji and Feshi where water supply is extremely limited during the dry season.
• Retting techniques to be enhanced through awareness building and training activities in rural areas.
• Dietary diversity to be increased by facilitating access to leguminous seeds (beans and black-eyes peas) and sweet cassava cuttings.
• Cereal and cassava processing to be facilitated through the installation of villagebased mills to improve dietary diversity and increase value of household produce.
• Income generating activities to be developed and diversified across the entire district.
• Farm to market roads to be rehabilitated to support increased agricultural trade at prices beneficial to the producer.

ACF plans to carry out project activities in the highly affected areas of Kahemba, Kajiji, Feshi and Panzi and to a lesser extent in Kenge, Boko, Popokabaka, Kasongo Lunda, Wamba Luadi and Kitenda. Due to limited resources, the intervention will not focus on all Konzo locations, but rather on the most affected villages.

As a result of this limited coverage, ACFUSA encourages its partners to consider the following interventions:

• Conduct a baseline study on Konzo in the health zones of Tembo, Mwela Lemba and Kimbao which were left out of the ACF study due to time constraints and accessibility issues.
• Further investigate the epidemiological links between millet consumption and Konzo incidence.
• Address local water and sanitation needs other than those related to cassava retting. A list of all local water points, schools and health centres needing to be rehabilitated and sanitised is available upon request.
• Facilitate access to leguminous seeds and healthy cassava cuttings, along with other food security activities, in areas outside of ACF's coverage in the Kwango district.
• Integrate a Konzo sensitisation component in all ongoing and future nutrition interventions in the Bandundu district and other Konzo-affected provinces.

For more information, contact: Muriel Calo, email: mcalo@actionagainsthunger.org

Note: Field Exchange 17 previously featured a summary of a published paper on a Konzo outbreak in 1996 in Bandundu province. See http://fex.ennonline.net/17/analysis.aspx

¹Rosling H, Mlingi N, Tylleskär T, Banea M. 1993. Causal mechanisms behind human diseases induced by cyanide exposure from cassava. In: RocaWM, ThroAM, editors. Proceedings of the first international scientific meeting of Cassava Biotechnology Network, 25-28 August 1992. Cali, Colombia : Centro Internacional de Agricultura Tropical. p 366-75 Tilleskär T, Cooke RD, Banea M, Poulter NH, Bikangi N, Rosling H. 1992. Cassava cyanogens and konzo, an upper motoneuron disease found in Africa. Lancet 339:208-11. Tylleskär T., Banea M., Bikangi M., Fresco L., Persson L.A., and Rosling H. 1991. Epidemiological evidence from Zaire for a dietary etiology of konzo, an upper motor neuron disease. Bull World Health Organ. 69(5): 581-589.