Study of the Risk Factors for the Development of Nutritional Oedema in North Kivu, DRC

By Mark Myatt and Frances Mason

Mark Myatt is a consultant epidemiologist and senior research fellow at the Division of Epidemiology, Institute of Opthalmology, University College London. His areas of expertise include infectious disease, nutrition, and survey design. He is currently working in Somalia.

Frances Mason is currently working part time as emergency nutrition advisor for Save the Children UK. Previously she spent three years as a consultant, following seven years with ACF - mainly as head of the nutrition and food security unit in Action Against Hunger UK.

This article is based on findings of a report written by Mark Myatt which, in turn, is based on the findings of research undertaken by Save the Children UK in DRC.

The Democratic Republic of Congo (DRC) in the Great Lakes region of Africa has suffered a civil war since 1998. Prior to this, in 1994, a large influx of Rwandans sought refuge in Eastern DRC following the genocide. Despite the potential wealth and fertility of the east of the country, rates of severe acute malnutrition are often high, with kwashiorkor being the predominant form¹, ².

In response to increasing concerns about this high prevalence of nutritional oedema, Save the Children UK undertook a case-control study of the risk factors for the development of nutritional oedema in North Kivu, in the north eastern part of DRC³. The study was carried out in the Masisi Territoire which comprises of displaced Congolese, returnees and Rwandan refugees (many of whom live in villages in eastern Masisi).

Method

Data were collected on diet, breastfeeding (for children aged 24 months or below only), and disease. The study dataset consisted of 243 cases of nutritional oedema paired with 243 matched (i.e. 1:1 matched) controls, aged 6-65 months. Asub-set of the data, children aged 24 months or below (118 cases of nutritional oedema paired with 118 matched (ie 1:1 matched) controls), included data on breastfeeding which was analysed separately. Data analysis using appropriate methods (i.e. paired t-tests, Mantel-Haenszel methods, and conditional logistic regression) was carried out, controlling for sex, age and socio-economic status.

Dietary data were collected on whether certain listed groups of food items were consumed within either the previous 24 hours or the previous seven days. Data were re-coded so that reported consumption in the previous 7 days (as recorded) included reported consumption in the previous 24 hours. Associations between the listed food items (and indices based upon the reported consumption of the listed food items) and nutritional oedema are presented as pairwise associations between the reported consumption of a given food item and case or control status. This does not control for potential confounding effects in the data. Hence multiple conditional logistic regression (using the cLOGISTIC programme, an EpiInfo add-in programme⁴) was subsequently used, presenting the results as adjusted odds rations, their 95% confidence intervals, and the p-value for the maximum likelihood ratio test statistic. Only those variables with significant pairwise associations (i.e. p < 0.05) were included in the conditional logistic regression model. Non-significant variables were removed from the model using backwards stepwise elimination. The remaining significant associations are termed independent associations.

Main findings

The case-control study found positive associations (i.e. increased risk) between nutritional oedema and the consumption of soya beans and sweetened or flavoured water, tea, infusion, or other liquids (including soups and broth). These latter fluids are often used for hunger abatement, providing children with a short-lived energy boost, replacing more nutritious meals and to 'supplement' breastmilk in non-weaned infants. Reported consumption of sweetened or flavoured water, tea or infusion, or other liquids (including soups and broth) in the previous seven days was independently associated with case or control status in children aged 6-24 months or below.

An increased risk also existed where the mother of the oedematous child was currently pregnant, hence cutting short the period of breastfeeding and potentially the time for childcare practices, which might otherwise enhance or protect the infant or young child's nutritional status. According to a survey Save the Children undertook in May 2005⁵, it was revealed that local beliefs are to stop breastfeeding when a mother becomes pregnant. However, this same survey showed:

• 0-5 months: Exclusive breastfeeding: 62.8% (CI 95%: 56.6-68.9%)
• 6-8 months: Introduction of complementary foods: 84.9% (CI 95%: 77.7-92.1%)
• 6-23 months: Continuation of breastfeeding: 78.1% (CI 95%: 72,7-83,4%)

Cases were significantly more likely to have been weaned or to have been weaned for a longer period than controls. The data is, therefore, consistent with earlier weaning in cases than in controls. Further risk factors included if the child had suffered a recent episode of disease (fever, intestinal parasites, respiratory problems and diarrhoea).

The study found a negative association (i.e. decreased risk of or protection against) of nutritional oedema and the consumption of a varied diet. This particularly included consuming maize, other cereals, bananas/plantain, avocados, sweet potatoes, Irish potatoes, oil/butter/ghee and sugar or honey. Good infant and child feeding practices were also found to be protective, particularly breastfeeding.

In younger children, cases were more likely to be reported as having had a common childhood disease (ie diarrhoea, fever, ARI) before the onset of oedema (or, for controls, one month prior to data being collected) than controls. In all children, cases were also more likely to have had intestinal parasites (a condition that may be more severe in older children) or to have suffered from other (unspecified) diseases before the onset of oedema (or, for controls, one month prior to data being collected) than controls.

Subsequent (ie conditional logistic regression) analysis of the data revealed that a lack of dietary diversity (the overall number and presence or absence of specific food items in the diet), possibly at particular times of the year, is associated with oedematous malnutrition.

Discussion

The data indicates that food insecurity, poor social care and limited access to public health are all potentially at play in determining the risk of a child having nutritional oedema. Agriculture and petty trade remain the principal economic activities in the region⁶. The main problems faced by the population include limited access to land (a fundamental cause of inter-ethnic conflict as, over the past few decades, poorer households have been forced off their land by the large landowners), crop disease and ash from the eruption of the volcano Nyamulagira (both of which have particularly affected the production of potatoes and sorghum). A lack of access to the prewar markets in Kinshasa and Western Congo has significantly reduced the prices of agricultural products since the supply is greater than demand. The area, therefore, is more cashpoor than food-poor. This particularly affects the poorest households who are reliant on their own harvest production and are unable to buy more diverse foods in the market. Alack of cash may also reduce access to health care - hence increasing the risk of disease - and may force the child's carer to find labour, thus further reducing time for child care and household food production. Time is already limited, particularly in the mountainous regions of North Kivu where most families travel long distances for water (taking up to 25% of their time in the day). Many of these households have no adequate storage which may further contribute to limited availability of clean water, resulting in cases of diarrhoea, worms and other intestinal diseases.

The positive association between illness and kwashiorkor suggests that kwashiorkor is part of a broader public health problem. Studies⁷,⁸ have shown that the poorest households are unable to purchase necessary items to ensure health care and a sanitary environment. In the latter of these two studies, 65-75% of income goes towards expenditure on basic food sources, reducing the available resources for healthcare even further than previous years.

For many years, kwashiorkor has been believed to be attributable, at least in part, to a deficiency in anti-oxidant nutrients⁹, resulting in high oxidative stress, primarily as a result of infection. The production of anti-oxidant enzymes that can protect the body from the harmful effects of this oxidative stress depends upon the presence of sufficient trace elements, such as selenium and zinc in the diet. Soya beans, which in the study are found to increase the risk of nutritional oedema, are high in phytic acid which, in turn, can block the uptake of essential minerals, including zinc. Only a long period of fermentation will significantly reduce the phytate content of soya beans. Soya is currently under promotion in Masisi and is used to make flour for porridge and soya milk used in tea, particularly amongst the poorer households who do not have access to cow's milk. Hence, the positive association between consumption of soya beans and nutritional oedema may be an indicator of household food insecurity, as well as the potential impact of anti-nutrient activity. Soya is also a component of the fortified flour (corn soya blend) given to malnourished children in feeding programmes. Anti-oxidant nutrients are strongly maintained in breast milk, which may also contribute to the protective nature of good infant feeding practices, demonstrated by the negative association with nutritional oedema. However, recent research has concluded that antioxidant supplementation of vitamin E, selenium, cysteine or riboflavin does not prevent the onset of kwashiorkor¹⁰. This study notes that previous research had entailed comparisons of biochemical variables in small groups of severely ill, malnourished children in hospital settings. This new research was a prospective trial investigating the role of antioxidants in the aetiology of kwashiorkor - hence indicating that antioxidant depletion may be a consequence rather than a cause of kwashiorkor.

During the past two years, Masisi has witnessed an upsurge of humanitarian interventions. Programmes are principally related to the restoration process and in particular, infrastructure rehabilitation (i.e. building roads, health centres, schools, etc.) and the provision of basic materials to some of these structures (essential drugs, school materials, etc.). Several agencies have recently initiated livestock activities. Some agencies have supported households to grow soya beans. While this remains a good source of nutrients (particularly protein), it is essential that this be cooked well to reduce the phytate content.

Future interventions should take into consideration the need of the poorer households for cash in order to provide greater dietary diversity, access to health care and support to carers/ mothers to ensure that they can afford to take care of their young children. Other recommendations include:

• Better immunisation, vaccination and pro motion of seeking early treatment.
• Diarrhoea control through hygiene promotion - protecting water sources and providing effective treatment with oral rehydration solution (ORS), and implementing measures to improve water availability, accesssibility and utilisation. If water is more accessible to the house-hold, it is likely that it may be used for sanitary purposes as well as the minimum for drinking and cooking.
• Adequate facilities for dispoal of faeces.
• Mass de-worming campaign and helminth control through hygiene promotion.
• Promotion of best weaning practices, probably within a family-planning programme.
• Promotion of best possible infant and child feeding practices and an emphasis on the inclusion in the diet of foods shown to be protective in this study. The utility of a 'kitchen garden' initiative could be investigated.
• In the longer-term, access to land and affordable health care-particularly the issue of user fees for the poorest households -must be addressed.

For further information, contact Frances Mason, Nutrition Advisor, Save the Children UK, email:F.Mason@savethechildren.org.uk

¹In May 2005, SCUK undertook a nutrition assessment in which 2.8% severe acute malnutrition was found (95% CI 0.8%-4.6%) of which 1.9% was kwashiorkor and 0.1% was marasmic-kwashiorkor.

²UNSCN (2001). Report on the Nutrition Situation of Refugees and Displaced Populations. RNIS 34.

³Myatt M. (2004). Analysis of data from a matched casecontrol study of risk factors for the development of nutritional oedema in children in North Kivu, Democratic Republic of Congo. Save the Children UK.

⁴Dallal GE (1989). "cLOGISTIC: A Conditional Logistic Regression Program for the IBM-PC," The American Statistician, 43, 125.

⁵Save the Children UK (May 2005). Enquête nutritionnelle : Anthropométrie, mortalité et analyse causale de la malnutrition; Zone de Santé de Masisi (Province du Nord Kivu) ; République Démocratique du Congo; En collaboration avec le PRONANUT

⁶King A, Adams L (2000). Household Food Economy Assessment: Eastern Democratic Republic of Congo. Plaine de la Ruzizi Moyens Plateaux Savane, Foret Food Economy Zones of South Kivu Province and Zone Volcanique (Rutshuru) and Zone des Plateaux (Masisi) Food Economy Zones of North Kivu Province. Save the Children UK.

⁷SC UK (2000). Household Food Economy Assessment: Eastern Democratic Republic of Congo, Food Economy Zones of South and North Kivu Provinces. Save the Children UK.

⁸SC UK. DRC Programme (2003). Update of the Household Economy Analysis of the Rural Population of the Plateaux Zone, Masisi, North Kivu, Democratic Republic of Congo. Save the Children (UK).

⁹Golden MHN (1998). Oedematous malnutrition. British Medical Bulletin 1998:54 (No. 2): 433-444.

¹⁰Ciliberto H, et al (2005). Antioxidant supplementation for the prevention of kwashiorkor in Malawian children: randomised, double blind, placebo controlled trial. BMJ, doi:10.1136/bmj. 38427.404259.8F (published 25 April 2005). See summary in this issue of Field Exchange.