Vitamin A Supplementation and Stunting Levels Among Two Year Olds in Kenya : Evidence from the 2008-09 Kenya Demographic and Health Survey

Background: High levels of undernutrition, particularly stunting, have persisted in Kenya, like in other developing countries. The relationship betweeen vitamin A supplementation and growth of children in Kenya has not been established, while there are context-specific variations on the relationship. This study explores this relationship in the Kenyan context. Methods: The study uses data from the 2008-09 Kenya Demographic and Health Survey, involving children aged 2435 months, a weighted sample of 1029 children. Descriptive and logistic regression analyses were conducted. The outcome variable of interest is stunting, while the exposure variable of interest is ever receiving a dose of vitamin A supplement. Secondary outcomes include underweight and wasting status. Results: The prevalence of stunting in the study group was 46%; underweight 20%; and wasting 6%. The prevalence of ever receiving vitamin A supplement was 78%. Receiving vitamin A supplement was significantly negatively associated with stunting and underweight status, adjusting for other co-risk factors. The odds of stunting were 50% higher (p=0.038), while for underweight were 75% higher (p=0.013) among children who did not receive Vitamin A supplement compared with those who did. Conclusion: This study demonstrates that receiving vitamin A supplement may be beneficial to growth of young children in Kenya. However, though freely offered through immunization services to children 6-59 months, some children do not receive it, particularly after completing the immunization schedule. There is need to establish innovative and effective ways of maximizing utilization of this intervention, particularly for children who have completed their immunization schedule.


INTRODUCTION
Childhood undernutrition remains a pervasive public health problem in the developing countries.It is associated with adverse short-term and long-term effects.It is a serious risk factor for ill health, associated with increased morbidity and mortality, and contributes substantially to the burden of disease in the developing countries [1][2][3].Undernutrition, particularly stunting is also associated with compromised cognitive development and future economic productivity; and higher risk of metabolic diseases later in the life course [4][5][6].The Millennium Development Goals (MDGs) prioritize reduction in undernutrition worldwide, however, young children are often vulnerable to undernutrition and its consequences, including *Address corresponding to this author at the African Population and Health Research Center, Nairobi, Kenya, P.O.Box 10787, 00100, Nairobi Kenya; Tel: +254-724322193; Fax: +254 (020) 4001101; E-mail: ekimani@aphrc.org,lizmurage@gmail.comstunting, underweight and wasting, which are prevalent in many low-income countries worldwide.In particular, estimates of stunting for children under the age of five are greatest in Africa, with East Africa carrying the highest burden, where close to 50% of young children in the region are afflicted [7].In Kenya, the prevalence of stunting for children aged less than five years is 35% [8].
Infant and young child nutrition, including micronutrient supplementation, has gained much attention over the previous decades, in part due to an increase in the understanding of the health benefits of nutrition in the first 1000 days of life.Breastfeeding, supplemental feeding and micronutrient fortification practices can benefit the health of young children through improved immune system response and adequate protection against infections [9,10].Micronutrient deficiency, specifically vitamin A deficiency, is estimated to cause more than a half million deaths among children under age five [1].It is clear that poor health conditions are the result of a multitude of factors including child nutrition and maternal and household characteristics [11].The continued high prevalence of stunting, underweight, and wasting across East Africa, including Kenya, emphasizes the importance of distinguishing the specific determinants of poor health status.By fully identifying and understanding the key risk factors that influence child health, it is possible to guide national policy and to better meet the needs of children.
The focus of this study was to determine the potential effects of nutritional supplementation on the growth of children age 24-35 months in Kenya.Primarily, the effects of receiving vitamin A supplements on stunting status was explored, and secondarily the effect on wasting and underweight status of young children, using data from the 2008-09 Demographic and Health Survey (DHS) in Kenya.Several studies have looked at micronutrient deficiencies and health status, showing the benefits of vitamin A, iron, zinc and calcium [9,10,12,13].However, other factors in different environmental contexts also influence growth of young children, including breastfeeding practices, parasitic infections and childhood diseases [14,15].This complexity has resulted in inconsistent results regarding the effects of vitamin A supplementation on child growth in developing countries [16][17][18] and indicates the need for context-specific studies.
Therefore, the study's goal was to better understand the relationship between vitamin A supplementation and growth of children in the Kenyan context.Children were evaluated for stunting as a key outcome of interest.The analyses had the overarching goal to address the burden of poor health and thus inform policy so that the negative cycle of poverty that poor health status perpetuates can be broken.The emphasis on alleviating this public health problem is critical in Kenya and throughout East Africa, and the long-term benefits of policy decisions and actions taken to improve the health of young children should be based on the key determinants of malnutrition.

Conceptual Framework
The conceptual framework (Figure 1), adapted from the UNICEF 1990 framework of causes of malnutrition [19] describes immediate and underlying causes of malnutrition.In this study, the focus was on the association of the immediate causes of malnutrition, specifically vitamin A supplementation, on nutritional status, particularly stunting, among children age two.The effects of other immediate factors including duration of breastfeeding, deworming and child morbidity, and other co-risk factors at the child, maternal and household levels were controlled for.by the Kenya National Bureau of Statistics in collaboration with government and non-government institutions.It covered 9,057 interviewed households, in which 8,444 women were interviewed.

The
The KDHS adopts a two-stage sampling design.The first stage involves selection of data collection clusters (enumeration areas).The second stage involves the systematic sampling of households from an updated list of households in each cluster.Regions with low population density are over-sampled, while those with high population density are under-sampled.Sampling weights are later applied to the data to obtain a nationally representative sample.All women age 15-49 who are either usual residents or visitors present in the households on the night before the survey are eligible to be interviewed in the survey.
This study analyzed data from the Kids Register (KR file).This file consists of records of interviewed de facto mothers (i.e.mothers who slept in the household the night before).Analysis involved children age 24-35 months whose anthropometric measurements (weight and height) were taken and were not considered erroneous (as detailed below).A sub-sample of 1,048 children age 24-35 months was filtered from the original sample of children aged below five years (unweighted N=6,079 children).To obtain nationally representative estimates, sampling weights were used.The final weighted sample included a total of 1,029 children age 24-35 months.

Anthropometric Measurements
All children below age five had their weight and length/height measured.These anthropometric measurements were taken by trained teams, consisting of two interviewers and one health worker.Weight was measured to the nearest 0.1kg using lightweight, bathroom-type scales with a digital screen designed and manufactured under the authority of UNICEF, and length/height was measured to the nearest 0.1 cm using a calibrated height board.

Outcome Variables
The outcome measures for this analysis include stunting, wasting and underweight, defined here as a height-for-age z-scores (HAZ), weight-for-height zscores (WHZ) and weight-for-age z-scores (WAZ) less than -2.The HAZ, WHZ and WAZ were were generated using the World Health Organization's (WHO) 2006 growth standards [20].Children with HAZ and WAZ from -5 and below and WHZ -6 and below were excluded from the analysis as they were considered to be extreme or incorrect.A total of 20 such cases were excluded from the analysis.The outcome variables were coded as "not stunted"/"stunted", "not wasted"/"wasted" and "not underweight"/ "underweight" respectively.

Explanatory Variables
The key explanatory variable of interest was ever receiving vitamin A supplement.The Kenyan government policy on vitamin A supplementation requires that the supplement be given to all children from the age of six months up to 59 months, at sixmonth intervals.The 2008-09 KDHS determined vitamin A supplementation status among these children through information from the caregivers as well as verification from documentation on the child health cards.For this variable, children who had ever received a dose of vitamin A supplement and those who have never received the supplement were coded as "Yes"/"No" respectively.
Other explanatory variables in the analysis include important risk factors for undernutrition classified as child, maternal and household factors, as described below.

Child Factors
Child's age was analyzed as a continuous variable; child's sex was coded as "Female" and "Male"; birth order was coded as 1 st , 2 nd , 3 rd , 4 th and above; child's birth size as perceived by the mother was coded as "Average", "Larger than average" and "Smaller than average"; deworming status in the six months preceding the survey was coded as "Yes"/"No"; duration of breastfeeding in months was used a continuous variable in the regression analysis but coded as "24+ months", "0-5 months", "6-11 months", "12-18 months" and "18-23 months" in the descriptive analysis; child morbidity as reported by the mother including acute diarrhoea, fever and cough within the previous 24 hours for fever or within the two weeks preceding the survey for diarrhoea and cough.These were coded as "No"/"Yes".

Maternal Factors
Maternal variables included the age at birth of the child, coded as "under 20", "20-24", "25-29", "30-34" and "35 and above"; marital status was coded as "In union" if married or cohabiting, and "Not in union" if never married, separated, divorced or widowed; maternal nutrition was measured by the body mass index -BMI, computed as weight in kilograms divided by the square of height in metres, and coded as "Normal BMI" (18.5-24.9),"Low BMI" (<18.5kgs/m 2 ) and "Overweight or obese" (BMI 25kg/m 2 ).Pregnant mothers and those two months postpartum were placed in a separate category.Maternal education was coded as "Secondary and above", "No formal education", and "Primary"; work status was coded as "Not working currently" and "Working currently".

Household Factors
Household residence was coded as either rural or urban, while region was coded as Nairobi, Central, Coast, Eastern, Nyanza, Rift Valley, Western and North Eastern provinces.Household wealth index was used as an indicator of socioeconomic status based on household assets [21].Each asset was assigned a factor score generated through principal components analysis.The resultant asset scores were standardized in relation to a normal distribution with a mean of zero and standard deviation of one.Each household was then assigned a score for each asset, and the scores were summed for each household.The wealth index was coded as "Lowest", "Lower", "Middle", "Higher" and "Highest".Other household factors included were: drinking water source, coded as "Improved" (including piped water, bottled water, tubewell/borehole and protected wells and springs) and "Unimproved" (including unprotected wells and springs, rain water and surface water sources); toilet facilities, coded as "Improved" (including water closets and ventilated improved pit latrines), "Unimproved" (including traditional pit latrines) and "No facilities" (including nontoilet faecal disposal methods mentioned in the survey such as bush and field); and number of children under age five living in the household, used as a continuous variable in the regression analysis but coded as "1", "2" and "3 and above" in descriptive analysis.De facto children (children visiting the household at the time of the survey but who were not usual residents in the household) (n=14) were assigned drinking water source and toilet facility of the host household.Additionally, they were assigned the number of children under age five that live with their mother.

Statistical Analysis
Data analysis was undertaken using Stata version 11.1.For bivariate analysis, Pearson's chi-square test was used to determine differences in proportions by background characteristics at the child, maternal and household levels.Multiple logistic regression was conducted with the outcome variables (stunting, underweight and wasting) and explanatory variables at the child, matenal and household levels as described above.A two-sided p-value of <0.05 was considered statistically significant.To account for the cluster sampling design, svyset commands in Stata were used, using pweights.

RESULTS
This study involves 1,029 children age 24-35 months, 488 girls (47%) and 541 boys (53%).The prevalence of stunting was the highest among the three nutritional indices, at 46%; prevalence of underweight was 20%; while prevalence of wasting was lowest, at 6%.The prevalence of ever receiving a dose of vitamin A supplement in the study group was 78%.
Table 1 shows the prevalence of stunting and other forms of malnutrition (underweight and wasting) by background characteristics for children age 24-35 months in Kenya.The results of the bivariate analysis indicate that ever receiving vitamin A supplement was significantly negatively associated with stunting (p=0.004).The prevalence of stunting was 56% among children who had not received vitamin A supplement compared with 43% among those who had received the supplement.Ever receiving vitamin A supplement was also significantly negatively associated with underweight (p=0.002),but not wasting.
Other factors with a significant association with stunting, underweight and wasting in the bivariate analysis level included perceived size at birth (for stunting and underweight), duration of breastfeeding (for wasting), morbidity from cough (marginal for underweight i.e. p=0.056),BMI status of the mother (marginal for stunting, i.e. p=0.051), maternal education (for all three indices), urban or rural residence (for stunting and underweight), region of residence (for wasting and underweight), wealth index (for all three indices), source of drinking water (for stunting and underweight), type of toilet facility (for all three indices), and number of children under age 5 living in the household (for wasting and underweight) (Table 1).

Association of Vitamin A Supplementation with Stunting and other Forms of Malnutrition
As Table 2 shows, receiving vitamin A supplement remained significantly negatively associated with  Self-reported morbidity from diarrhoea, or fever or cough in the last two weeks before the survey. 3 Non-dejure defacto children in the household (n=14) were assigned the water source and toilet facility of the household they were living in. 4 Improved sources includes piped water sources, tubewell/borehole, protected wells and springs, and bottled water; Unimproved sources include all other sources such as surface water and rain water.Improved facility includes toilets with a flush to sewer, septic tank or to a pit latrine; no facility includes no facility/bush, flush to somewhere else; buckets , composting and hanging toilets.Non-dejure defacto children in the household (n=14) were assigned the number of children under five years that live with their mother.
stunting even after adjusting for other co-risk factors at the child, maternal and household levels.The odds of stunting were 50% higher among children who had not received vitamin A supplement compared with children who had received the supplement (p=0.038).Additionally, the odds of underweight status were 75% higher among children who had not received vitamin A supplement (p=0.013).

Co-Risk Factors of Malnutrition at the Child, Maternal and Household Levels
As shown on Table 2, at the child level, after adjusting for other factors, age of the child, perceived size at birth and reported morbidity in the last two weeks were significantly associated with either stunting, underweight or wasting (p<0.05).At the maternal level, only the education level of the mother was significantly associated with the nutritional status (stunting and underweight status, p<0.05).At the household level, region of residence, source of drinking water, type of toilet facility used by the household, socio-economic status and number of children under the age of five living in the household were all significantly associated with either stunting, underweight or wasting (p<0.05).However, these associations are beyond the focus of this study, and the inclusion of these co-risk factors in the models was mainly to control for their effects.

DISCUSSION
This study set out to determine if there is an association between vitamin A supplementation and stunting among young children age two -the age group with the highest level of stunting in Kenya, according to the 2008-09 KDHS [8].The study shows that receiving vitamin A supplement is significantly associated with children's growth.This suggests that vitamin A supplementation may be protective against stunting, or may reverse stunting [16,22].This is important given the negative consequences of stunting at an early age and the fact that vitamin A supplementation is a public health intervention in Kenya that is freely provided to children between the ages of six months and five years but that some children do not receive it.
The association between vitamin A and child growth has been documented in observational studies [16,22].This association is linked to several mechanisms including the inverse relationship between vitamin A intake and incidence and severity of childhood infections, which may lead to growth faltering due to reduced nutritional intake, malabsorption of nutrients and increased nutrient expenditure during illness [23], and also the role of vitamin A in cell differentiation and function, therefore showing biological plausibility [24].Despite this established relationship between vitamin A and child growth, studies regarding vitamin A supplementation and growth in developing countries have found inconsistent results [16][17][18].For example, in a randomized controlled study among pre-school Indonesian children age 6-48 months, Hadi et al. (2000) found that vitamin A supplementation improved linear growth.This effect was age specific, higher among children age 24 months or older compared with children under age 24 months.Additionally, this effect was higher among children who were not breastfed and those who were deficient of vitamin A [16].Another study, however, in a randomized controlled trial in Northern Ghana found no significant association between vitamin A supplementation and linear growth of children age three or older, despite substantial reduction in severe morbidity and mortality in the same group [17].Some studies have suggested that the inconsistencies are due to the effect of other factors in the environment, for example parasitic and respiratory infections, which may result in a lower effect on growth of infected children, and dietary vitamin A intake, also with a negative effect on the relationship between vitamin A supplementation and growth [14,15].Such variations in study findings suggest the importance of environmental context and hence the need for contextspecific studies on the association between vitamin A supplementation and child growth.
The positive association between receiving vitamin A supplements and stunting observed in this study indicates that over and above any interactions between vitamin A supplementation and other factors affecting child morbidity and mortality, there may still be a net beneficial effect of providing vitatmin A supplementation.Continued government support of this important primary health intervention is essential, not only in addressing morbidity and mortality, but also in addressing malnutrition.This is particularly important with regards to stunting, given its short-term and longterm adverse effects, and which has persisted as a major public health concern in Kenya for many years [8,25].It is unfortunate though that the coverage of vitamin A supplementation for the targeted age group (6-59 months) is very low at 30%, with coverage being particularly low after the second year of life, and decreases with age [8].The low coverage is despite the fact that vitamin A is readily available in MCH facilities, i.e. 94% of facilities offering vaccinations also have vitamin A [26].
In this study our main focus was on the association between receiving vitamin A supplement and stunting.However, we also established an even stronger negative relationship between receiving vitamin A supplement and underweight status.This relationship adds to the potential net benefit of vitamin A supplementation on growth among children in Kenya that should inform existing and future interventions on vitamin A supplementation.Additionally, we were able to establish other potential co-risk factors of undernutrition at the child, maternal and household levels.However, this was beyond the focus of this study, and these factors have been explained in other studies in Kenya [27,28], and elsewhere [29][30][31][32].
Despite the importance of this study in determining the relationship between receiving vitamin A supplements and undernutrition, we acknowledge possible endogeineity of receiving the supplement in that it may be influenced by other explanatory variables in the model including socio-economic status and mother's education, hence potential simultaneity bias.More complex analytic models beyond the scope of this study may be worthwhile in further explorations of this relationship.Additionally we are not able to establish a causal relationship, given the cross-sectional nature of the data.What we do know from the data, however, is that vitamin A supplements were given prior to the date when the outcome variables for stunting, wasting and underweight status were recorded.Further longitudinal studies to determine causal relationship are recommended.The Kenya demographic and health survey did not obtain data for all vitamin A supplementation doses.We therefore could not establish the dose-response relationship.
In summary, this study has demonstrated that vitamin A supplementation may be beneficial to the growth of young children, potentially either through protective mechanisms or through reversal of malnutrition [16,22].This is an important finding that should guide policy and programs regarding vitamin A supplementation among children under age five through the primary health care system in Kenya and elsewhere in sub-Saharan Africa.However, further studies, as mentioned earlier are needed to better understand this relationship for more effective and targeted policies and programs.

Policy Implications
The prevalence of stunting for children aged less than five years, particularly those aged two years in Kenya is worrying.The low coverage of vitamin A supplementation, only at 30% among children 6-59 months in the six months preceding the survey [8] is also of great concern, given that it is freely available for all children aged 6-59 months.The coverage of vitamin A supplementation among children aged three to five years is currently (since 2011) being maximized through provision at early child development (ECD) Centres.However, children aged two years are left out.This is because they usually have not joined the ECD Centres, and because they have generally completed the immunization schedule, they are hardly in contact with the primary health care facilities where they would be offered the supplements.There is need to find innovative and effective ways of reaching this group, and also those aged one year, who are also left out as most have also completed their immunization schedule by the age of one year.We recommend task sharing in vitamin A supplementation, allowing vitamin A supplementation to be given through the community strategy by community health workers who are in regular contact with children in the community.
International and the National Council for Population and Development (NCPD), Kenya.We particularly acknowledge technical support from Sarah Staveteig and Sarah Bradley of ICF International.The initial manuscript was reviewed by Bryant Robey of ICF International and Geroge Kichamu of NCPD.We also acknowledge contribution of Margaret Kungu of NCPD in recoding of variables.Elizabeth Kimani-Murage is a Wellcome Trust Fellow.
study uses secondary data from the 2008-09 Kenya Demographic and Health Survey (KDHS).The Demographic and Health Surveys are nationally representative household surveys conducted in approximately 75 countries that provide data on a wide range of indicators including family planning, health and child nutrition.The 2008-09 KDHS was conducted

Figure 1 :
Figure 1: Immediate and underlying causes of malnutrition among children in Kenya.(Adapted from UNICEF's 1990 Framework of Causes of Malnutrition [19]).