Abstract: Background and Objective: A higher risk of overweight has been widely described in other populations but very limited information was available on Indonesian school-aged children (ISC). The aim of this study was to analyze the association between stunting and overweight ISC. Materials and Methods: This study used electronic data of Basic Health Research (BHR) 2010 which was designed as a cross sectional survey. A total of 8,599 ISC aged 6-12 years from eight provinces in Indonesia were selected for the analysis. Overweight and obese were defined as body mass index (BMI) for age Z-score (BAZ) of ≥+1 and ≥+2 standard deviation (SD), while stunting was defined as a height-for-age Z-score (HAZ) of <-2 SD based on WHO cut-off point. Results: The prevalence of overweight, obesity and stunting in ISC were 19.5, 7.9 and 28.0%, respectively. There were 7.5% of ISC categorized as having concurrent stunting and overweight. Most of them were boys and at younger age. They also had low economic status, low parents education and lived in rural areas (p<0.05). In multivariate logistic regression analyses adjusted for all factors, stunting was associated with overweight (OR = 2.33, 95% CI: 2.06-2.61). Conclusion: It is apparent that there is a significant association between stunting and overweight in ISC. This association implies that the double-burden of malnutrition (DBM) does exist in Indonesia and comprehensive planning is needed to overcome the problem.
INTRODUCTION
Indonesia is currently facing the double burden of malnutrition, in which undernutrition and overnutrition occur at the same time1. It is related to nutrition transition, the changes in nutritional status profile among population determined by economic, demographic, environmental and cultural changes2. During the period of 1993-2007, the prevalence of obesity in Indonesia increased significantly in all population groups, including rural population, low-income groups and school-aged children3. Some nutrition disorders at school age are undernutrition (i.e. stunting and wasting) and overnutrition (i.e. overweight and obesity).
Based on the results of Basic Health Research (BHR)4, 12.2% of children aged 6-12 years were wasted and 9.2% of them were overweight. The most prominent nutrition disorder was stunting because 35.6% or more than one third of children aged 6-12 years were stunted, in which 15.1% were categorized as severe stunted and 20.5% as moderate stunted4. The latest results of BHR5 showed that 30.7 and 18.8% of children aged 5-12 years were stunted and overweight, respectively.
There are only limited researches in Indonesia that showed a significant association between undernutrition and overnutrition, particularly stunting and overweight. Therefore, this study aimed to identify risk factors for overweight and analyze the association between stunting and overweight of Indonesian school-aged children (ISC) in eight provinces of Indonesia. Specific objectives of this study were to analyze characteristics of the children and their families, analyze nutritional status of the children, analyze the relation of characteristics of the children and their families with their nutritional status; and analyze risk factors for overweight among children.
MATERIALS AND METHODS
Data, design, time and location: This was a descriptive national-scale cross-sectional study. Electronic files (secondary data) was used from BHR4 conducted in 2010 by National Institute of Health Research and Development (NIHRD), Ministry of Health. The research sites consisted of eight provinces based on the category of stunting prevalence6, i.e. very high (≥40%) in East Nusa Tenggara and North Sumatra; high (30-39%) in West Nusa Tenggara and West Java; moderate (20-29%) in Bangka Belitung, Special Capital Region of Jakarta and Special Region of Yogyakarta and low (<20%) in Bali. Data collection in several regions conducted by BHR’s data collection team was started from May to August 2010. Data was processed, analyzed and interpreted in November 2013 in Bogor, West Java.
Total sample and sampling method: Population, according to Basic Health Research4 by Ministry of Health, were all households representing all provinces in Indonesia. The household sample was selected based on 2010 Citizen Census listing. Household selection was conducted by Central Bureau of Statistics with two-stage sampling.
The children in this study were boys and girls aged 6-12 years from eight provinces found in 2010 BHR’s electronic data4. There were 11,335 children contained in the electronic files. The total number of children that was processed and analyzed became 8,599.
Children’ inclusion criteria in this analysis were as follows: aged 6-12 years with a body mass index (BMI) for age Z-score (BAZ) of -5 SD ≤Z-score ≤+5 SD, energy adequacy level of 40-300% and had complete data7. The exclusion criteria were the ones with BAZ of < -5 SD and > +5 SD; energy adequacy level of <40 and >300%, were having a diet, fasting or celebration day; and had incomplete data.
Data types and collection method: All data in this study were secondary data taken from BHR4 2010, obtained in the form of electronic files. The data we used included children characteristics (age and sex), family characteristics (parents age, parents education level, parents BMI, family size and household expenditure), children’ food consumption (type of food and amount of food consumed in grams), children’ body weights and heights.
The data were collected by BHR team from NIHRD, Ministry of Health which was conducted from May to August 2010. Socioeconomic and demographic data were obtained by interviewing the children directly using structured questionnaires equipped with manual instructions on filling out the questionnaire. Food consumption data were collected by a 24 h recall method. Anthropometric data were collected by direct measurement.
Nutritional status: Anthropometric data that were analyzed included height-for-age index (HAZ) and BAZ, using WHO AnthroPlus v1.0.47. Based on HAZ, the children were classified into two groups, i.e. normal nutritional status group and stunting (<-2SD) group. Overweight and obese were defined as BAZ of ≥+1 and ≥+2 SD8.
Data processing and analysis: The data were processed and analyzed by using Statistical Package for the Social Science (SPSS 17.0) for Windows (IBM Corporation, www.ibm.com/us-en/marketplace/spss-statistics). The processing stage included selecting variables that would be analyzed, cleaning and recoding them into categorical data. Data analyses performed in our study were univariate, bivariate and multivariate analyses. Comparative test was also conducted to observe the differences in variable values based on normal and overweight categories. Bivariate analysis was used to determine the association between two variables, i.e. dependent and one of independent variables by using chi-square test (χ2) for categorical data. Multivariate analysis was performed to determine the values of risk factors or odds ratio (OR) by using multiple logistic regression with Backward Wald method, results were considered statistically significant at p<0.05 and at 95% confidence interval (CI).
RESULTS
Characteristics of children and their families: Results showed that there were 19.5% overweight children and 7.9% obese children. On the other hand, the analysis also showed that 28% of the children were stunted. The analysis was also performed to determine the children with concurrent stunting and overweight and the results showed that there were 7.5% of children with concurrent stunting and overweight. Children’ mean age was 8.9 years and percentage proportion of boys and girls was almost the same, i.e. 49.2% (girls) and 50.2% (boys) (Table 1 and 2).
Most of the children were from lower middle class in which their household expenditures were in 3rd-5th quintiles (53.9%) and living in urban areas (55.4%). Most of them (54.3%) had large family (≥8 people). Mean mother age and father age were 36 and 41 years, respectively. Most of the parents had low education level. Only less than 10% of children’ parents were highly educated. Mean BMI of parents nutritional status was relatively normal, i.e. 22.5 kg/m2 for the fathers and 23.9 kg/m2 for the mothers. However, most of the children’ parents were underweight or BMI <18.5 kg/m2 (Table 1 and 2).
Children’ food consumption was also analyzed. The results indicated that mean energy and protein adequacy levels were still categorized as adequate. The values of all food consumption variables were significantly higher in overweight children than children with normal nutritional status (p<0.05). However, Healthy Eating Index (HEI) score were not significantly different in both groups (p>0.05). Mean HEI score showed that most of the children had poor diet quality (Table 1 and 2).
Nutritional status: Nutritional status of the children was determined based on indicator of BAZ and HAZ. The BAZ z-score of the stunted children (-0.01±1.70 SD) were significantly higher (p = 0.000) than the normal children (-0.36±1.49 SD), although their mean Z-scores were considered normal. HAZ z-score of the overweight children (-1.39±2.11 SD) were significantly lower (p = 0.000) than the normal children (-1.13±1.43 SD) (Table 1). These results were also supported by mean BMI. Mean BMI of the stunted children was 16.98±3.63 kg/m2 which were significantly higher (p = 0.000) than children with normal nutritional status (16.08±2.86 kg/m2).
Relationship between children’ characteristics, family characteristics and nutritional status: The relationship between children’ characteristics, family characteristics and BAZ was analyzed.
| Table 1: | Characteristics of children and their families based on nutritional status (normal and overweight) |
| *t-test, significant at p<0.05 | |
| Table 2: | Participant distribution by nutritional status (normal and overweight) |
| *chi-square test; significant at p<0.05 | |
Children’ characteristics (i.e. age and sex) had no significant relationship with overweight (p>0.05). The number of overweight children in the stunted group was also significantly higher than in the normal group. Most of the overweight children were from higher-class family (household expenditure was in 3rd-5th quintiles), living in urban areas and having medium-sized family (5-7 people) (Table 2).
Most of the overweight children had highly educated and overweight (BMI ≥25 kg/m2) parents. However, parents age was not significantly associated with overweight among the children (Table 2).
Results also showed that food consumption had a significant association with BAZ. Energy and protein adequacy levels, as well as percentage of energy intake from carbohydrates, fat and protein were significantly associated with overweight in the children (p<0.05). However, HEI score had no significant relationship with overweight. There were more overweight children who had good HEI scores (Table 2).
Table 3 showed the association between social-economic characteristics and the concurrent stunting and overweight among children. The analysis showed that concurrent stunting and overweight cases were more likely found among boys and in younger age (6-9 years). Most of the stunted-with-overweight children were from lower class family (household expenditure was in 1st-2nd quintiles) and living in rural areas. However, family size had no significant association with concurrent stunting and overweight.
Among all parents characteristics, only father education level and nutritional status that had significant association with the nutritional status of their children. Most of the children with concurrent stunting and overweight significantly had low-educated and underweight (<18.5 kg/m2) fathers. Mother’s education level and nutritional status were not significantly related to the concurrent stunting and overweight in the children. Parents age also had no significant association with concurrent stunting and overweight found among the children (Table 3).
Energy adequacy level had a significant relationship with concurrent stunting and overweight in the children. Table 3 showed that most of the stunted-with-overweight children were categorized as having energy intake <110%. HEI score also had no significant association with the concurrent stunting and overweight cases. The majority of stunted-with-overweight children had good HEI scores.
Risk factors for overweight: Based on the bivariate analysis which indicated a significant association between stunting and overweight, we assumed that stunting had an effect on overweight. Therefore, it was used as one the risk factors in the analysis of children’ overweight-related risk factors. The results were presented in Table 4.
The results of multiple logistic regression indicated that stunting was a risk factor for overweight with the highest OR value (2.33) among other variables analyzed. It meant that stunted children were 2.33 times at higher risk of being overweight than the ones with normal height.
DISCUSSION
Results of the present study showed that there were 19.5% overweight children and 7.9% obese children. These results were slightly higher than the latest BHR’s data in 20135; i.e. 18.8% of children aged 5-12 years were classified as overweight. The results of multiple logistic regression indicated that stunting was a risk factor for overweight with the highest OR value (2.33) among other variables analyzed.
A previous study showed that there was a significant association between stunting and overweight status (weight-for-height index) in children aged 3-9 years in four countries (Russia, Brazil, South Africa and China)9. Stunted children had 1.7-7.8 times higher risk to be overweight than the normal ones. Similar result was also found in another study which indicated that stunted babies had nearly three times higher risk of being overweight (OR = 2.7, 95% CI: 1.8-4.1)10.
| Table 3: | Participant distribution based on stunted-with-overweight and non-stunted-with-overweight nutritional statu |
| *This category included normal, overweight only and stunted only children, **Chi-square test, significant at p<0.0 | |
| Table 4: | Significant risk factors for overweight based on logistic regression analysis |
| 1logistic regression; significant at p<0.05 | |
Previous researchers also showed that there were 19% of 120 under-three children with concurrent stunting and overweight in South Africa2. However, result from this analysis was different from a study conducted in South Africa by different researchers who observed children aged 8-11 years and found there was no significant association between stunting and overweight (p>0.05)11.
The relationship between stunting and overweight/obesity can be explained by mechanism of growth retardation and changes in hormonal response combined with unhealthy food consumption. Stunted children have less lean body mass, resulting in decreased basal metabolic rate and physical activity. If energy intake is adequate, there is a difference between linear growth potential and adipose tissue deposition. It may happen for several reasons; i.e. the food consumed does not contain sufficient essential nutrients for linear growth but had adequate nutrients to increase adipose tissue deposition. Moreover, early nutrition programming is likely to produce some hormonal effects on limited linear growth but the potential for weight gain is not deprived of hormonal effects9. Stunting leads to a series of important changes such as lower energy expenditure, more vulnerable to the effects of high fat intake, lower fat oxidation, disruption of food consumption adjustment and impaired fat metabolism12-13. A 36-month longitudinal study on 30 girls aged 7-11 years in Brazil indicated that stunted group had lower resting metabolic rate during follow-up period with significant differences between 24-month and 36-month periods. It was related to the increased weight gain and decreased lean body mass if compared to the normal group. These conditions indicated risk of obesity in stunted group14.
Concurrent stunting and overweight in school-aged children can describe nutrition transition aspect11. In the past, stunting and limited access to food were closely related. However, the association is not as clear as in the present, especially in countries that undergo nutrition transition such as Indonesia. The transitions are, for instance, the changes in dietary pattern from traditional diet into Western-type diet (energy-dense, high-fat and low-fiber) or high-carbohydrate but low-protein diets; thus, the children who were originally stunted and underweight will become overweight but still stunted2.
CONCLUSION AND RECOMMENDATION
There were 7.5% of school-aged children with concurrent stunting and overweight in eight provinces that we observed. Bivariate and multivariate analyses also indicated the effect of stunting on overweight. The stunted children were 2.33 times at higher risk of being overweight than the normal children. Thus, it was clear that there was a significant association between stunting and overweight in school-aged children. This finding proves that nutrition transition and double burden of malnutrition occur in Indonesia. It implies the need for comprehensive planning to overcome the problems.
Comprehensive planning should be immediately designed and its implementation should be followed up. Nutrition programs -especially on adolescent girls, women of childbearing age, pregnant women and nursing mothers-need to be a top priority. The review is needed to determine the most appropriate window of opportunity among all life cycle aspects to break the chain of nutrition problems.
SIGNIFICANCE STATEMENT
This study discovered that there were 7.5% of school-aged children with concurrent stunting and overweight. Most of them were boys and at younger age. They also had low economic status, low father education and lived in rural areas. Stunted children were 2.33 times at higher risk of being overweight than the normal ones. This finding proves that nutrition transition and double burden of malnutrition occur in Indonesia. This study will help the researcher to uncover the critical areas of relationship of stunting and overweight that many researchers in Indonesia have not able to explore. Thus a new theory on relationship of stunting and overweight may be arrived at.
ACKNOWLEDGMENTS
The authors thanks to National Institute of Health Research and Development, Ministry of Health, Republic of Indonesia, who had given permission to use data from Basic Health Research 2010 for this research.