• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of wtpaEurope PMCEurope PMC Funders GroupSubmit a Manuscript
Scand J Public Health Suppl. Author manuscript; available in PMC Mar 1, 2010.
Published in final edited form as:
Scand J Public Health Suppl. Aug 2007; 69: 96–106.
doi:  10.1080/14034950701356435
PMCID: PMC2830109
EMSID: UKMS28816

What’s new? Investigating risk factors for severe childhood malnutrition in a high HIV prevalence South African setting1

Abstract

Aim

To identify risk factors for severe childhood malnutrition in a rural South African district with a high HIV/AIDS prevalence.

Design

Case-control study.

Setting

Bushbuckridge District, Limpopo Province, South Africa.

Participants

100 children with severe malnutrition (marasmus, kwashiorkor, and marasmic kwashiorkor) were compared with 200 better nourished (>−2 SD weight-for-age) controls, matched by age and village of residence. Bivariate and multivariate analyses were conducted on a variety of biological and social risk factors.

Results

HIV status was known only for a minority of cases (39%), of whom 87% were HIV positive, while 45% of controls were stunted. In multivariate analysis, risk factors for severe malnutrition included suspicion of HIV in the family (parents or children) (OR 217.7, 95% CI 22.7–2091.3), poor weaning practices (OR 3.0, 95% CI 2.0–4.6), parental death (OR 38.0, 95% CI 3.8–385.3), male sex (OR 2.7, 95% CI 1.2–6.0), and higher birth order (third child or higher) (OR 2.3, 95% CI 1.0–5.1). Protective factors included a diverse food intake (OR 0.53, 95% CI 0.41–0.67) and receipt of a state child support grant (OR 0.44, 95% CI 0.20–0.97). A borderline association existed for family wealth (OR 0.9 per unit, 95% CI 0.83–1.0), father smoking marijuana (OR 3.9, 95% CI 1.1–14.5), and history of a pulmonary tuberculosis contact (OR 3.2, 95% CI 0.9–11.0).

Conclusions

Despite the increasing contribution of HIV to the development of severe malnutrition, traditional risk factors such as poor nutrition, parental disadvantage and illness, poverty, and social inequity remain important contributors to the prevalence of severe malnutrition. Interventions aiming to prevent and reduce severe childhood malnutrition in high HIV prevalence settings need to encompass the various dimensions of the disease: nutritional, economic, and social, and address the prevention and treatment of HIV/AIDS.

An external file that holds a picture, illustration, etc.
Object name is ukmss-28816-f0001.jpg
Keywords: Case-control study, drug abuse, food intake, HIV/AIDS, kwashiorkor, malnutrition, marasmus, poverty, risk factors, South Africa

Background

Several studies have investigated the risk factors for malnutrition among young children in developing countries. A Pubmed search using the terms “child”, “malnutrition”, and “risk factors” identified 26 relevant papers on this topic since 1976, with studies conducted in 17 different countries [1-26]. Established risk factors vary widely in different settings with few consistent epidemiological patterns emerging globally. Leading risk factors include low family socioeconomic and education levels, difficult family dynamics (e.g. mother pregnant with another child, father unemployed, or separated), sub-optimal nutrition (e.g. lower breastfeeding rates or stopping breastfeeding at earlier age, lack of food) and poor environmental and hygiene conditions (e.g. inadequate housing, absence of water and sanitation). Few studies, however, have investigated risk factors for severe, rather than mild or moderate, malnutrition [6,7,14,15,17], and only one has described these in a high HIV prevalence setting [4]. Furthermore, most have used hospital rather than community controls, and therefore may have missed key risk factors.

While South Africa is categorized as a middle-income country, it also has one of the highest inequities in societal wealth distribution globally [27]. Consequently, malnutrition rates are still high among the poorest social strata, e.g. rural, black African children. At the study site in rural Limpopo Province, protein-energy malnutrition (mainly kwashiorkor) was the second leading cause of death in children under five years between 1992 and 1995, accounting for 15% of deaths in this age group [28]. Since then, the advent of the HIV/AIDS pandemic has escalated both the number of malnourished children in the area, and resultant mortality.

Debates continue regarding the most appropriate and cost-effective way to reduce childhood severe malnutrition. The identification of important causal factors in particular settings is one of the important steps in the development of effective intervention programmes [29]. Recognizing the high malnutrition rates (in an area with a relatively low infant and child mortality rate) [28], and the particularly high (>20% among pregnant women) prevalence of HIV in our study area [30], we felt an imperative to better understand the underlying causes and determinants of severe childhood malnutrition. This was necessary before we expended considerable effort on designing and implementing potentially cost-effective and sustainable interventions. We therefore conducted a case-control study comparing hospitalized severely malnourished children with better-nourished peers living in the same disadvantaged environment, to identify risk factors for severe malnutrition.

Material and methods

Study area

The Bushbuckridge District in Limpopo Province, South Africa, has a population of over half a million people. It is situated in the southeast corner of South Africa’s northernmost province adjacent to its border with Mozambique. The area is dry and overpopulated, with plots too small to support subsistence farming. Unemployment rates are high (>30%) and many who find employment work as migrant labourers elsewhere. The area’s health needs are served by three district hospitals, two health centres, and a network of some 45 clinics.

Study design

Risk factors were determined using a case-control study, a design most appropriate for events with relatively low prevalence, such as severe malnutrition. We enrolled 100 hospitalized, severely malnourished children and matched them with 200 controls of similar age and living in the same village as the case. Both cases and controls had to be less than five years old and live in the Bushbuckridge district. Fieldwork was conducted over a one-year period (April 2003 to March 2004).

Classification

The diagnosis of severe malnutrition was based on the Wellcome classification of malnutrition, i.e. weight-for-age less than 80% of median with oedema (kwashiorkor); or weight for age less than 60% of median, with (marasmic kwashiorkor) or without oedema (marasmus) [31]. A child was defined as wasted if his/her weight-for-height was below −2 standard deviation (SD) scores (z-scores) from the Centers for Disease Control (CDC) 2000 reference value [32]. Similarly, stunting and underweight-for-age was defined as children below −2 z-score for height-for-age and weight-for age, respectively. A child was classified as being HIV infected if the HIV Elisa test was positive (in a child older than 12 months) or if the child had clinical signs of HIV and his/her mother was HIV positive (in a child younger than 12 months).

Inclusion criteria were age under five years and residence in the Bushbuckridge district; cases were children severely malnourished at admission, and controls were neither underweight nor overweight at the time of examination.

Recruitment

Cases

Cases were identified at the three district referral hospitals (Mapulaneng, Matikwana and Tintswalo) by hospital staff. Primary caregivers were interviewed by trained study fieldworkers in the hospital ward. In situations where a child died before an interview was possible, the interview was performed at the child’s home.

Controls

Each case was matched with two controls of a similar age (but not sex) living in the same neighbourhood (village). The fieldworkers identified the controls by travelling to the village from which the case originated, identifying the home of the relevant case and asking passers-by if they knew of any child of that age group in the immediate area. Controls were matched to the same age ± one month if the case was less than one year old, ± three months if one to three years old, and ± six months if three to five years old. The process was repeated until two children who were within normal weight-for-age centiles (>−2SD and <+2SD) were identified. Identified children’s caregivers were then interviewed using the same questionnaire as that used for the cases.

Questionnaire

Four local fieldworkers were trained to conduct the interviews and to undertake the anthropometrical measurements. A 172-item, pre-coded questionnaire was designed to include information on anthropometrics, diet and weaning, health of child and mother, parental characteristics, household composition, socioeconomic status, and behaviour patterns. This questionnaire was translated and back-translated by research assistants with language expertise in the two local languages – XiTsonga and Sepedi – with the interview conducted in the caregiver’s preferred language. Children’s date of birth, birth weight, and immunization status were obtained from their Road to Health Card whenever possible (available in 75% of instances). Dietary diversity was calculated by summing the number of unique foods (grouped into eight categories) consumed by the child in the preceding seven-day period. The maximum score was 8, with one point given for each group consumed [33].

Weighing and measuring

Fieldworkers obtained the anthropometric measurements. The project supervisor (TdM) trained the fieldworkers and monitored subsequent performance. Electronic weighing scales (Hanson, UK) were used and calibrated before each session. The children were weighed with minimum clothing (diapers for babies and underclothes for older children) and the average of two readings was taken. Heights were measured using measuring tapes. Children younger than two years were measured supine. Weights were measured to the nearest 0.01 kg and height to the nearest 0.1 cm.

Ethical considerations

Ethical clearance was obtained from the University of the Witwatersrand’s Committee for Research on Human Subjects (Medical) and from the Limpopo Province Research Committee. Community consent and feedback of research findings receives continuous attention in all projects conducted at the study site.

Statistical analysis

The sample size was estimated to be sufficient to detect a minimum odds ratio of 2.0 for a prevalence of 2%, at the 95% confidence level and with a power of 80%. Data were entered into a Microsoft Access database. The univariate analysis was performed with STATA version 5 (Stata Corp., College Station, TX, USA) and the multivariate analysis using SPSS version 11.0 (SPSS Inc, 2001). EpiInfo 6.0 (Centers for Disease Control and Prevention, Atlanta, GA, USA) was used to convert anthropometrical measurements to percentage weight-forage and height-for-age, and to z-scores (using the CDC 2000 reference standard) [32].

Risk factors for severe malnutrition were investigated by comparing cases and controls, and using linear logistic regression (logit model). Factors contributing to severe malnutrition were grouped in large categories: socioeconomic risk factors, nutritional risk factors, factors linked with HIV/AIDS, and associated morbidity. A wealth index was built, by adding dummy variables associated with 29 features of modern housing and ownership of various household appliances and goods, with a scale from 1 to 24 (1 representing the lowest value and 24 the maximum found in the sample). We combined the average level of education of mother and father by combining the mean number of years schooling into a single variable labelled “Parents’ education”. Variables linked with breastfeeding were combined into a single variable labelled “Breastfeeding practices”, which included early (<12 months) and abrupt breastfeeding cessation, and weaned for abnormal causes. Variables linked with feeding practices were combined into a single variable labelled “Food diversity”, which included consumption of vegetables, green leaves, fruits, eggs, dairy products, fish or meat in the past week. Mothers’ and fathers’ deaths were combined into a “Parents’ death” and illnesses into a “Parents’ illness” category. Evidence of TB in either parents or children was combined into “TB in family” category. An “HIV in family” category was based on either the mother or child being HIV positive (obtained through hospital records). In addition, we recoded data to indicate any suspicion of HIV in either the mother or father, by studying responses to reasons for stopping breastfeeding, reported chronic illness(es) in the parents, and reasons for any hospitalization of parents.

Results

One hundred severely malnourished cases and 200 controls were enrolled. The biological mother was interviewed in 87% of instances. There were no significant differences between the two groups in age and sex composition. As expected, anthropometric measures were all significantly poorer among cases (Table I). High stunting rates (45%) were noted among controls.

Table I
Demographic and anthropometric characteristics of study children, Bushbuckridge district, South Africa, 2003–04.

Classification of cases

Of the 100 severely malnourished children, 40 (40%) had kwashiorkor, 38 (38%) were marasmic, and 22 (22%) presented with marasmic kwashiorkor. HIV testing was done at the discretion of the attending doctor, who usually performed the test if clinical signs indicative of HIV were recognized. HIV ELISA results were available for only 39 (39%) cases, of whom 34 (87%) tested positive. HIV prevalence varied somewhat among tested children according to pathology: marasmus (94%), marasmic kwashiorkor (91%), and kwashiorkor (73%). A quarter (25%) of the cases died during their hospital stay.

Birth characteristics

Cases were lighter at birth than controls (2.79 vs. 2.98 kg, p=0.01), and were more likely to be of higher birth order (3 or more). They were somewhat more likely to be born at home, but the difference was only borderline (14% vs. 6.5%, OR 2.3, 95% CI 0.9–5.7, p=0.05). More cases had a deceased sibling (Table II).

Table II
Bivariate analysis of risk factors for severe malnutrition, Bushbuckridge district, South Africa, 2003–04.

Medical history and associated morbidity

Cases being severely malnourished by definition, it was expected that they would have higher morbidity than controls. Cases were more likely to have been previously hospitalized, particularly for diarrhoea, and to have experienced an illness in the past month. While complete immunization rates were similar in the two groups (51% vs. 63%), cases were more likely to have consulted a traditional healer. Almost a quarter of cases had a TB contact (at home or through daily contact), four times higher than controls (see Table II).

On multivariate analysis, diarrhoea in the past 12 months (prior to hospitalization for cases, or prior to home visits for controls) was an obvious risk factor (OR=2.73) (Table III). Furthermore, the duration of the diarrhoea was significantly longer in cases. Exposure to a TB contact was important (OR=5.32) as was previous hospitalization (OR=6.78). All these risk factors remained at the same significance level after controlling for wealth.

Table III
Multivariate analysis of risk factors for severe malnutrition, Bushbuckridge district, South Africa, 2003–04.

Nutritional history

The questionnaire included a series of questions about breastfeeding, feeding practices, and foods consumed in the past week. Most mothers (>99%) in both groups breastfed, but a majority had introduced either fluids (51%) or solids (70%) within four months of birth. Fewer cases were still breastfeeding, and in children who were no longer breastfeeding, cases stopped breastfeeding about four months earlier on average (15 vs. 19 months, p<0.01). Caregivers of cases were more likely to admit that the child went hungry occasionally and these children were more likely to consume three or fewer meals per day. Cases had significantly lower diversity in their diets, consuming about 1.7 fewer food types per week (see Table II).

Seven factors remained significant in the multivariate analysis (see Table III). A first group of risk factors was constituted by breastfeeding variables. Breastfeeding being stopped before 12 months or abruptly, and for causes other than what are regarded as the norm were all strong risk factors for severe malnutrition (OR=2.57, 2.26, and 2.40 respectively). “Abnormal” causes included breast feeding being stopped because of child illness; mothers becoming pregnant, ill, leaving or dying; and the fear of HIV transmission (only three cases). The second group was constituted by feeding variables. In particular, a balanced diet in the past week produced significantly lower risks of developing severe malnutrition: vegetables (OR=0.35), green leaves (OR=0.54), fruits (OR=0.23). In addition, “having to cut the size of children’s meals or skipping meals because there was not enough money to buy food”, an indicator of severe poverty, was also significant (OR=2.09). Another variable had the same effect, “child said he/she was hungry because not enough food”, but the effect disappeared in the multivariate analysis. All effects remained similar after adding an indicator of poverty (income class or wealth index). Eating wild foods (vegetables, insects, fruits, herbs) was not significant after controlling for other factors, and the sign of each variable was positive, suggesting that eating wild food is above all an indicator of severe poverty.

Family health and dynamics

Mothers of severely malnourished children were more likely to be unmarried, not to have completed secondary education (matriculated), and to have been admitted to hospital in the preceding year (see Table II). There was no difference in age, employment status, or parity between the two groups of mothers. Fathers of cases were more likely to be deceased or to be living at a different residence from that of the child. They were also less likely to be the head of the household and more likely to smoke marijuana (see Table II). There was no difference in age, employment status, or education levels between the two groups of fathers.

Many families in the study were headed by a single parent, but cases were more likely to belong to a single-parent household. The father was not the head of the household in more than two-thirds of homes, with this being the situation for more than three-quarters of cases (see Table II). There was no significant difference in residence status, abuse of alcohol, spousal or child abuse, or encounters with the law between the two groups.

Multivariate analysis revealed five significant risk factors associated with parental health (see Table III). Two factors were associated with the mother: mother had an illness that forced the child to stop breastfeeding (OR=4.80), and mother had experienced the loss of a child other than the index child (OR=2.47). Three factors were associated with the father: father had died (OR=33.0), father was hospitalized in the past 12 months (OR=3.84), and father smoked marijuana (OR=4.83).

Suspicion of parental HIV infection also appeared as a strong risk factor (OR=9.10). These risk factors stayed robust after controlling for wealth, and the two borderline risk factors now became significant at the p=0.05 threshold. Other factors were not significant after controlling for the five factors described above. In particular pulmonary tuberculosis was insignificant after controlling for HIV, as was the case for alcohol abuse (mother or father) and domestic violence (against women or children).

Household and economic factors

Most (63%) households earned less than R1,000 (South African rand: ≈US$150) per month, well below the poverty line (defined in 2001 as an income of R1,541 per month for a household of five) [34]. Cases generally lived in poorer households with fewer household assets and lower incomes. Only a third of cases were recipients of government-provided child social support grants, compared with more than one-half of controls. Unemployment rates were high in both groups, but cases were less likely to reside in households where at least one member was employed (see Table II).

Six socioeconomic factors for severe malnutrition were identified as noteworthy on multivariate analysis (see Table III). Three demographic factors were important: multiple births had a higher risk than single births (OR=4.35), as did births of order 3 and more (OR=2.66), and married mothers had a lower risk than unmarried mothers (OR=0.30). Father’s education was not significant, although at least 10 years of paternal education reduced the risk of severe malnutrition by about half (OR=0.52). Mother’s education had a similar effect originally, which disappeared once the other factors were taken into account.

Wealth was an obvious and strong risk factor: a wealth index of 10 or more (less poor households) reduced the risk of severe malnutrition by two-thirds (OR=0.35). Interestingly, receiving a child support grant still had a positive effect after controlling for wealth (OR=0.47), suggesting that offering this grant to poor families strongly reduces the risk of severe malnutrition. Income classes had the same effect as the wealth index, and both could be used interchangeably in the regression analyses, although the wealth index had a higher variance and offered results that were more robust. The effect of income class was no longer significant when wealth was controlled for.

Synthesis

A final synthesis of the leading risk factors of severe malnutrition was undertaken where a few variables were grouped together to simplify the analysis (as outlined in the methods section). Factors associated with the medical history were excluded because of reverse causality. This analysis resulted in little change to the overall picture (Table IV).

Table IV
Synthesis of risk factors for severe malnutrition, Bushbuckridge district, South Africa, 2003–04.

A first group of variables played a large role in childhood malnutrition owing to the high risk or prevalence of the factor involved: evidence of HIV in the family (parents or children) (OR=217.7), weaning practices (OR=3.0), and food diversity (OR=0.53). A second group appeared to have a moderate role: sex (males were now at significantly higher risk than females) (OR=2.7), higher birth order (3 or more) (OR=2.3), and the absence of a child support grant (OR=0.44). A final group of three variables played a more modest role: wealth of the family (OR=0.91), father smoking marijuana (OR=3.9), and evidence of pulmonary tuberculosis in the family (OR=3.2).

Discussion

The study has identified a number of alternative ways that severe malnutrition can be tackled in this predominantly poor, rural community. Diverse risk factors such as poor household food security, unhealthy feeding practices, sub-optimal access to quality health services, disruptions of family structure (influenced by the effects of HIV), and inadequate access to child support grants all contributed to the problem and lend themselves to potential interventions. None is mutually exclusive, and the need to focus on multiple interventions, including economic, educational, health, and social welfare services, in the prevention and management of severe malnutrition is made overt.

Two independent determinants, nevertheless, predominated. First, the role of sub-optimal feeding (particularly breastfeeding and weaning) practices and poor nutrient intake (low food diversity or lack of food associated with poverty), even after controlling for many other potentially confounding factors, including income, was paramount. Second, and of equal importance, was the role of HIV, either as a direct (infected child) or indirect (parents infected, sick, or dead) influence.

The dominance of kwashiorkor as a cause of severe malnutrition and of under-five mortality is noteworthy considering the extremely high percentage of HIV positivity in the children who were tested for this disease. Marasmic malnutrition is often more common in HIV-infected children [35-37], with kwashiorkor comprising about 22–33% of hospital admissions for severe malnutrition in other African countries. While the association between kwashiorkor and HIV is well recognized, this finding reaffirms the view that primary nutritional deficiencies still contribute substantially to severe malnutrition in this high HIV prevalence area. The high (25%) mortality from severe malnutrition, despite reasonable standards of hospital care, can also be attributed to the effects of HIV. Our findings support the contention that attaining <5% case-fatality rates in children with severe malnutrition, as proposed by the World Health Organization, is hard to achieve in high HIV settings [38,39].

The 45% stunting rate found in controls is a manifestation of the high levels of chronic under-nutrition in this region. The rate is almost double that found in local and national surveys conducted in other parts of South Africa [40]. The finding has important implications for both the interpretation of the study findings and the design of malnutrition interventions in the region. There is increasing acceptance that the aetiology of stunting may be different from that of acute wasting [41]. Some countries have witnessed a dramatic fall in the prevalence of wasting owing to improved health service delivery and identification and treatment of acutely malnourished children, despite stunting rates remaining high [42]. Thus, while severe malnutrition (and wasting) in the region may be amenable to short- and medium-term health system interventions, preventing mild and moderate childhood malnutrition may require greater attention to many of the “distal” determinants (political, economic, cultural, and social factors) that rarely cause ill health by themselves, and are highlighted in this paper. Further, the disadvantages (risks) identified for severely malnourished children in this setting are accentuated if one considers that the controls they were being compared with were far from “normal” by developed world standards and bore a high burden of stunting.

The strong association between severe malnutrition and poor food security, limited food diversity, and sub-optimal breastfeeding and weaning practices was expected and has been shown by many researchers previously. What is striking, in addition, is the poor exclusive breastfeeding rates and early introduction of supplementary feeds in both groups. Dietary diversity is an indicator of household food security and a dietary diversity score of less than six has been proposed as the cut-off to identify children with an inadequate diet [33]. Our results suggest that this cut-off is also helpful in predicting the development of severe malnutrition among children who are of normal weight and possibly stunted.

Another disturbing finding is the preponderance of single-parent households in this area, with half of controls also having this status. This phenomenon is common in South Africa [43], owing to late marriage and low contraceptive use among adolescents, and is, to a certain extent, a residual legacy of apartheid. In the past, legislation forced working-age men, many of them fathers, to seek employment in the cities and towns and prevented them from marrying early or, when married, from taking their spouses with them. Many cohabited with their spouses but never married because of their inability to pay a “lobola” or dowry to the women’s family. In this study, the absence of a father, whether through death, urban migration, or non-marriage, contributed significantly to the risk of malnutrition.

A strongly consistent finding in studies done in different settings and continents is the benefits of women’s education in preventing child malnutrition [44]. The small or negligible effect of mother’s and father’s education in our study is worth noting. Obviously, parental education is important, but its effects seem to be mediated through a variety of other determinants, such as income and food practices. Furthermore, the average level of education in this area is rather high by African standards, and the mean level of education differed only slightly between cases and controls (8.0 versus 9.5 years of schooling), which is likely to dampen the effect of education per se.

A key factor affecting all underlying determinants of malnutrition is poverty. The effects of poverty on child malnutrition are pervasive. Poor households and individuals are unable to achieve food security, have inadequate resources for care, and are less able to avail themselves of modern health services (sometimes using traditional healers instead, as in this study). South Africa is one of the few countries in sub-Saharan Africa that offers state-funded child social support grants (R190 [US$28] per month) to children younger than 14 years of age in needy families. Most caregivers claim to use the money to feed children. Notwithstanding the relatively small amount of money provided, our data confirm the beneficial effect of being a grant recipient in protecting against severe malnutrition, reducing the risk by more than half, everything else being equal. However, the low grant uptake in an area where more than 90% of households would be expected to qualify for grants offers another potential intervention target, i.e. expanding efforts to increase grant access.

A particular strength of this study was the effort made to match cases with controls living in the same setting (village and location). This reduces confounders and allows stronger inferences to be made regarding the potential role of identified risks. Despite the care taken in the sampling scheme, selection bias might have occurred in the selection of controls as we depended on neighbours to direct us to appropriate households rather than pre-selecting controls. Furthermore, the high stunting rates in controls might be attributed to a measurement bias (inadequate measurement technique in the field). However, field staff were carefully trained and their practice regularly monitored; further, a province-wide stunting rate of 36% has previously been described [40]. We were unable to obtain HIV results in more than half the cases and depended solely on an HIV ELISA test (performed mainly in cases older than 15 months of age). None of the control’s HIV status was known. This obviously biases any comparisons. Finally, we recognize that we may have omitted to measure determinants that are difficult to assess and are typically unobserved, but are important to nutritional outcome, such as cultural factors influencing caring behaviours.

The study has identified a number of new risk factors that are worthy of further investigation. Negative outcomes associated with higher birth order (3 or more) require further investigation to determine whether this arises from child competition, or is rather the consequence of benign neglect from caretakers. Similarly, fathers smoking marijuana remained a consistent risk factor in all multivariate models. This could be related to a number of negative attitudes or behaviours, including insufficient money and food being available for children owing to the diversion of resources to enable continuation of the father’s habit, lack of attention to childcare, and so forth. The precise role of HIV and TB in the aetiology of severe childhood malnutrition remains to be better delineated. However, even an incomplete initial attempt (by this study) to define the link suggests that they are likely to be important, explaining the increasingly severe malnutrition-related mortality noted over the past five years in the study area.

How can the health service best respond to the study findings? Potential areas for intervention could include more intense attention, through activities such as regular growth monitoring and culturally appropriate nutrition counselling, to underweight children or those who are failing to thrive [45]. The latter could arguably be best achieved by supporting the wider and more thorough implementation of Integrated Management of Childhood Illness (IMCI) services at all primary healthcare centres in the region [46]. Better monitoring of all hospitalized children following discharge (probably best done through home visits by health promoters) [47], may be a worthwhile investment. Community-wide exclusive breastfeeding promotion – at least for the first four months of life – as well as weaning education is warranted, recognizing that the task is complicated by the high HIV prevalence in mothers in the study area. Attention to removing barriers to child support grants could also stem severe malnutrition in the region. Finally, the primacy of prevention of mother-to-child HIV transmission strategies, whether to prevent severe malnutrition or other diseases contributing to childhood mortality and morbidity, as well as the treatment of HIV infected children and parents, cannot be overemphasized.

Acknowledgements

The authors would like to thank the Stella and Paul Loewenstein Educational Trust and the South African Medical Research Council for funding the study, and Professors Steve Tollman and John Pettifor for assisting in the conceptualizing of the study and in reviewing the final draft of the paper.

Footnotes

1This paper has been independently peer-reviewed according to the usual Scand J Public Health practice and accepted as an original article.

The authors have no conflict of interest to declare.

References

[1] Kikafunda JK, Walker AF, Collett D, Tumwine JK. Risk factors for early childhood malnutrition in Uganda. Pediatrics. 1998;102:E45. [PubMed]
[2] Jeyaseelan L, Lakshman M. Risk factors for malnutrition in south Indian children. J Biosoc Sci. 1997;29:93–100. [PubMed]
[3] Ferrari AA, Solymos GM, Castillo RM, Sigulem DM. Risk factors for protein-energy malnutrition in pre-school shanty-town children in Sao Paulo, Brazil. Rev Paul Med. 1998;116:1654–60. [PubMed]
[4] Ticklay IM, Nathoo KJ, Siziya S, Brady JP. HIV infection in malnourished children in Harare, Zimbabwe. East Afr Med J. 1997;74:217–20. [PubMed]
[5] Rikimaru T, Yartey JE, Taniguchi K, Kennedy DO, Nkrumah FK. Risk factors for the prevalence of malnutrition among urban children in Ghana. J Nutr Sci Vitaminol (Tokyo) 1998;44:391–407. [PubMed]
[6] Sive AA, Subotzky EF, Malan H. The social, family and medical backgrounds of children with kwashiorkor presenting at a teaching hospital. S Afr Med J. 1993;83:180–3. [PubMed]
[7] Tolboom JJ, Ralitapole-Maruping AP, Kabir H, Molatseli P, Anderson J. Severe protein energy malnutrition in Lesotho, conditioning factors and death and survival in hospital. Trop Geogr Med. 1989;41:1–8. [PubMed]
[8] Anandaraja S, Manoharan U, Amaleswari A, Rotti SB, Soudarssanane MB, Felix AJ. Risk factors in protein-energy malnutrition among infants and toddlers [letter] J Trop Pediatr. 1994;40:313–14. [PubMed]
[9] Henry FJ, Briend A, Fauveau V, Huttly SR, Yunus M, Chakraborty J. Risk factors for clinical marasmus: A case-control study of Bangladeshi children. Int J Epidemiol. 1993;22:278–83. [PubMed]
[10] Ighogboja SI. Some factors contributing to protein–energy malnutrition in the middle belt of Nigeria. East Afr Med J. 1992;69:566–71. [PubMed]
[11] Jongpiputvanich S, Poomsuwan P, Phittayanon P. Prevalence and risk factors of protein energy malnutrition (PEM) in preschool children of Klong-Toey Slum, Bangkok, Thailand. J Med Assoc Thai. 1992;75:39–45. [PubMed]
[12] Reed BA, Habicht JP, Niameogo C. The effects of maternal education on child nutritional status depend on socio-environmental conditions. Int J Epidemiol. 1996;25:585–92. [PubMed]
[13] Chopra M. Risk factors for undernutrition of young children in a rural area of South Africa. Public Health Nutr. 2003;6:645–52. [PubMed]
[14] Ayaya SO, Esamai FO, Rotich J, Olwambula AR. Socio-economic factors predisposing under five-year-old children to severe protein energy malnutrition at the Moi Teaching and Referral Hospital, Eldoret, Kenya. East Afr Med J. 2004;81:415–21. [PubMed]
[15] Owor M, Tumwine JK, Kikafunda JK. Socio-economic risk factors for severe protein energy malnutrition among children in Mulago Hospital, Kampala. East Afr Med J. 2000;77:471–5. [PubMed]
[16] Getaneh T, Assefa A, Tadesse Z. Protein-energy malnutrition in urban children: prevalence and determinants. Ethiop Med J. 1998;36:153–66. [PubMed]
[17] Islam MA, Rahman MM, Mahalanabis D. Maternal and socioeconomic factors and the risk of severe malnutrition in a child: A case-control study. Eur J Clin Nutr. 1994;48:416–24. [PubMed]
[18] Bloss E, Wainaina F, Bailey RC. Prevalence and predictors of underweight, stunting, and wasting among children aged 5 and under in western Kenya. J Trop Pediatr. 2004;50:260–70. [PubMed]
[19] Sanghvi U, Thankappan KR, Sarma PS, Sali N. Assessing potential risk factors for child malnutrition in rural Kerala, India. J Trop Pediatr. 2001;47:350–5. [PubMed]
[20] Gallo PR, Amigo H, Claudio L. Risk factors for growth retardation in children of low economic and social level in Sao Paulo, Brazil. Arch Latinoam Nutr. 2000;50:121–5. [PubMed]
[21] Radebe BZ, Brady P, Siziya S, Todd H. Maternal risk factors for childhood malnutrition in the Mazowe District of Zimbabwe. Cent Afr J Med. 1996;42:240–4. [PubMed]
[22] Setswe G. Prevalence and risk factors for malnutrition among children aged 5 years and less in the Lefaragatlha village of Bophuthatswana. Curationis. 1994;17:33–5. [PubMed]
[23] Coulter JB, Omer MI, Suliman GI, Moody JB, Macfarlane SB, Hendrickse RG. Protein-energy malnutrition in northern Sudan: Prevalence, socio-economic factors and family background. Ann Trop Paediatr. 1988;8:96–102. [PubMed]
[24] Reyes H, Perez-Cuevas R, Sandoval A, Castillo R, Santos JI, Doubova SV, et al. The family as a determinant of stunting in children living in conditions of extreme poverty: A case-control study. BMC Public Health. 2004;4:57. [PMC free article] [PubMed]
[25] Khin MU, Myo K, Nyunt NW, Nyi WH, Thein TM, Butler TC. Risk factors for persistent diarrhoea and malnutrition in Myanmar children, I: Socio-anthropological risk factors. J Trop Pediatr. 1994;40:41–3. [PubMed]
[26] Fletcher PD, Simeon DT, Grantham-McGregor SM. Risk indicators of childhood undernutrition in Kingston, Jamaica. Trans R Soc Trop Med Hyg. 1992;86:566–9. [PubMed]
[27] United Nations Development Programme of South Africa . South Africa: Health Development Record 2003. Oxford University Press; Cape Town: [cited 14 March 2007]. 2003. Available at: http://www.undp.org.za/NHDR2003/NHDRSumFull.pdf.
[28] Kahn K, Tollman SM, Garenne M, Gear JSS. Who dies from what? Determining cause of death in South Africa’s rural north-east. Trop Med Int Health. 1999;4:433–41. [PubMed]
[29] Sanders D. Success factors in community-based nutrition programmes. Food Nutr Bull. 1999;20:307–14.
[30] Department of Health . National HIV and Syphilis antenatal sero-prevalence survey in South Africa, 2005. Department of Health; Pretoria: [cited 14 March 2007]. 2006. Available at: http://www.doh.gov.za/docs/reports/2005/hiv.pdf.
[31] Classification of infantile malnutrition. Lancet. 1970;10:302–3. Lancet.
[32] Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, et al. CDC growth charts: United States. National Center for Health Statistics; Hyattsville, MD: 2000. (Advance data from vital and health statistics; Report No. 314). [PubMed]
[33] Hatlùy A, Torheim LA, Oshaug A. Food variety – a good indicator of nutritional adequacy of the diet? A case study from an urban area in Mali, West Africa. Eur J Clin Nutr. 1998;52:891–8. [PubMed]
[34] Human Sciences Research Council . Fact sheet: Poverty in South Africa. Human Sciences Research Council; Pretoria: 2004.
[35] Cartmell E, Natalal H, Francois I, Ferreira MH, Grahnquist L. Nutritional and clinical status of children admitted to the malnutrition ward, Maputo central hospital: A comparison of data from 2001 and 1983. J Trop Pediatr. 2005;51:102–5. [PubMed]
[36] Kessler L, Daley H, Malenga G, Graham S. The impact of the human immunodeficiency virus type 1 on the management of severe malnutrition in Malawi. Ann Trop Paediatr. 2000;20:50–6. [PubMed]
[37] Yeung S, Wilkinson D, Escott S, Gilks CF. Paediatric HIV infection in a rural South African district hospital. J Trop Pediatr. 2000;46:107–10. [PubMed]
[38] Ashworth A, Chopra M, McCoy D, Sanders D, Jackson D, Karaolis N, et al. WHO guidelines for management of severe malnutrition in rural South African hospitals: Effect on case fatality and the influence of operational factors. Lancet. 2004;363:1110–15. [PubMed]
[39] Deen JL, Funk M, Guevara VC, Saloojee H, Doe JY, Palmer A, et al. Implementation of WHO guidelines on management of severe malnutrition in hospitals in Africa. Bull World Health Organ. 2003;81:237–43. [PMC free article] [PubMed]
[40] Labadarios D. The National Food Consumption Survey (NFCS): Children aged 1–9 years, South Africa, 1999. Department of Health; Pretoria: 2000.
[41] Waterlow JS. Reflections on stunting. In: Gopalan C, editor. Recent trends in nutrition. Oxford University Press; Delhi: 1993.
[42] Keller W. The epidemiology of stunting. In: Waterlow JC, editor. Linear growth retardation in less developed countries. Nestle Nutrition/Raven Press; Vevey/New York: 1988.
[43] Ziehl SC. Families in South Africa. In: Adam BN, Tsai WY, editors. Handbook of world families. Sage Publications; Thousand Oaks, CA: 2005.
[44] Smith LC, Haddad LJ. Explaining child malnutrition in developing countries: A cross-country analysis. International Food Policy Research Institute; Washington, DC: 2000.
[45] Guldan GS, Fan HC, Ma X, Ni ZZ, Xiang X, Tang MZ. Culturally appropriate nutrition education improves infant feeding and growth in rural Sichuan, China. J Nutr. 2000;130:1204–11. [PubMed]
[46] Santos I, Victora CG, Martines J, Goncalves H, Gigante DP, Valle NJ, et al. Nutrition counseling increases weight gain among Brazilian children. J Nutr. 2001;131:2866–73. [PubMed]
[47] Devadas RP. Activating the community for nutritional improvement. Food Nutr Bull. 2002;23:119–32. [PubMed]
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...