Predictors of maternal and child double burden of malnutrition in rural Indonesia and Bangladesh
1–3Vanessa M Oddo, Jee H Rah, Richard D Semba, Kai Sun, Nasima Akhter, Mayang Sari, Saskia de Pee, Regina Moench-Pfanner, Martin Bloem, and Klaus Kraemer
ABSTRACT
Background:Many developing countries now face the double bur- den of malnutrition, defined as the coexistence of a stunted child and overweight mother within the same household.
Objective:This study sought to estimate the prevalence of the double burden of malnutrition and to identify associated maternal, child, and household characteristics in rural Indonesia and Bangladesh.
Design:A total of 247,126 rural households that participated in the Indonesia Nutrition Surveillance System (2000–2003) and 168,317 rural households in the Bangladesh Nutritional Surveillance Project (2003–2006) were included in the analysis. Maternal and child double burden (MCDB) and its association with individual and household characteristics were determined by using logistic regres- sion models.
Results:MCDB was observed in 11% and 4% of the households in rural Indonesia and Bangladesh, respectively. Maternal short stature [Indonesia (OR: 2.32; 95% CI: 2.25, 2.40); Bangladesh (OR: 2.11;
95% CI: 1.96, 2.26)], and older age were strong predictors of MCDB. Child characteristics such as older age and being female were associated with an increased odds of MCDB, whereas cur- rently being breastfed was protective against MCDB [Indonesia (OR: 0.84; 95% CI: 0.81, 0.84); Bangladesh (OR: 0.55; 95% CI:
0.52, 0.58)]. A large family size and higher weekly per capita household expenditure predicted MCDB [Indonesia (OR: 1.34;
95% CI: 1.28, 1.40); Bangladesh (OR: 1.94; 95% CI: 1.77, 2.12)].
Conclusions:Double burden is not exclusive to urban areas. Fu- ture policies and interventions should address under- and over- weight simultaneously in both rural and urban developing country settings. Am J Clin Nutr 2012;95:951–8.
INTRODUCTION
An estimated 1.5 billion adults are overweight worldwide (1), with recent trends showing a shift in prevalence from higher to middle- and low-income countries (2–4). Consequently, developing nations are now confronted with the double burden of malnutrition (5), characteristically defined by the coexistence of under- and overnutrition (6). Research has reported this phenomenon within the same country (7–9), within the same household (10–12), and among maternal-child pairs (5, 13–15).
Double burden stands to further threaten the health and eco- nomic stability of these already resource-strained countries.
MCDB4, commonly defined as an overweight mother paired with an undernourished child, is of concern because under- and overnutrition are thought to operate through independent causal
pathways and have typically been treated as distinct health problems (12–14). Child undernutrition is associated with an increased risk of childhood mortality and poor cognition (16), which may result in an increased risk of NCDs and obesity in adulthood (16–19). Several mechanisms have been proposed to explain the link between child undernutrition and obesity in adulthood. Undernutrition in childhood has been associated with an increased risk of high glucose concentrations, high blood pressure, higher susceptibility to gain central fat, and harmful lipid profiles—all of which are linked to NCDs in adulthood (19–21). NCDs cause approximately two-thirds of deaths globally and are thought to be a major barrier to achieving the Millennium Development Goals by 2015 without comprehensive actions by countries.
Although research is limited, an increased prevalence of MCDB has been observed in countries that are in the midst of their nutrition transition (2, 5, 22). Several studies have shown that MCDB is strongly tied to economic factors (11, 15). MCDB has also been associated with older maternal age (14), maternal short stature (15), and increased family size (12–14). However, most previous research on MCDB has focused on urban areas, and the association has not been adequately explored in rural areas of low income countries, where evidence suggests an in- creasing prevalence of overweight and a shift in dietary patterns (9, 19).
1From Mathematica Policy Research Inc, Cambridge, MA (VMO);
SIGHT AND LIFE, Basel, Switzerland (JHR and KK); the Department of Ophthalmology, Johns Hopkins University School of Medicine, Baltimore, MD (RDS and KS); Helen Keller International Asia Pacific, Dhaka, Bangla- desh (NA); Helen Keller International, New York, NY (MS); the Nutrition Service, Policy, Strategy and Programme Support Division, World Food Programme, Rome, Italy (SdP and MB); and Global Alliance for Improved Nutrition, Geneva, Switzerland (RM-P).
2This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
3Address correspondence and reprint requests to VM Oddo, Mathematica Policy Research Inc, 955 Massachusetts Avenue, Cambridge, MA 02139.
E-mail: [email protected].
4Abbreviations used: HKI, Helen Keller International; MCDB, maternal and child double burden; NCD, noncommunicable disease; NSP, Nutrition Surveillance Project; NSS, Nutrition Surveillance System; SES, socioeco- nomic status.
Received August 27, 2011. Accepted for publication January 13, 2012.
First published online February 22, 2012; doi: 10.3945/ajcn.111.026070.
Am J Clin Nutr 2012;95:951–8. Printed in USA.Ó 2012 American Society for Nutrition 951
by guest on August 23, 2016ajcn.nutrition.orgDownloaded from
In rural Bangladesh and Indonesia, a high prevalence of poverty and undernutrition is well documented (23). However, both countries also boast high mortality rates for cardiovascular disease and number of estimated diabetes cases (24). Increasing prevalence in overweight and NCDs translates into rising health care costs in Bangladesh and Indonesia, 2 countries with large rural area populations (23–25). The current analysis sought to determine the prevalence of MCDB in rural Bangladesh and Indonesia. Child, maternal, and household characteristics asso- ciated with MCDB were also explored using the nationally representative nutrition surveillance data from each country.
Understanding the prevalence and predictors of MCDB among these populations will allow for informed development of policy and intervention strategies.
SUBJECTS AND METHODS
Data
For nearly 2 decades, the NSS of Indonesia and the NSP of Bangladesh have been carried out jointly by the HKI and the governments of Indonesia and Bangladesh, respectively. Details of the NSS and NSP are described elsewhere (9, 26, 27). Briefly, both surveillance systems are based on UNICEF’s conceptual framework on the causes of malnutrition (28). Relevant infor- mation was collected from households with children aged,5 y in both urban and rural areas. A stratified multistage cluster sampling method was used to identify eligible households in both the NSS and NSP. The sampling schemes were designed to represent Indonesia and Bangladesh, both nationally and di- visionally. In Indonesia, rural data were collected every 3 mo from the provinces of Lampung, Banten, West Java, Central Java, East Java, Lombok, South Sulawesi, and West Sumatra. In Bangladesh, data were collected every 2 mo. Rural data were collected from 4 subdistricts in each of the 6 divisions, namely Barisal, Chittagong (including Chittagong Hill Tracts), Dhaka, Khulna, Rajshahi, and Sylhet. A household is defined as a group of individuals eating from the same kitchen, and new households were selected for every round of both surveys. This analysis includes a total of 415,443 rural households that participated in the NSS from 2000 to 2003 (n = 247,126) and in the NSP from 2003 to 2006 (n = 168,317).
Data collection
Data collection methods were similar in Indonesia and Ban- gladesh. Eligible respondents in each of the selected households were visited in their home by 2-person field teams, who con- ducted interviews and made anthropometric measurements.
Written consent was sought from each respondent. Parents or guardians provided consent for children under 5. Interview and assessments were conducted only after consent was obtained.
Using a structured questionnaire, the mother of the child or other caretaker was asked to provide information on the child’s age, sex, morbidity in the past week(s), vaccination, breast- feeding status, and dietary intake. Maternal data including age, education, morbidity, reproductive history, personal hygiene, smoking status, and alcohol consumption were collected. In- formation on household characteristics such as food composition, agricultural practice, expenditures, and other SES, environmental,
and health indicators were collected. Price variables were col- lected in Bangladesh taka or Indonesian rupiah.
Anthropometric measurements, including height/length and weight of the children and mothers were taken following standard procedures (29). Weight was measured by using a Precision Health Scale (TANITA Corporation) to the nearest 0.1 kg. Height/
length was assessed to the nearest 0.1 cm by using a height-and- length board. Age was verified by using immunizations cards or home records. If this information was unavailable, mother’s recall or the local events calendar was used to elicit the age of the child.
The field staff collecting the data was from local non- governmental organizations and were trained by HKI before each round of data collection. Monitoring teams from HKI supervised the field staff during data collection. To ensure data reliability, 5–10% of the interview and anthropometric data were recollected the following day. Nonresponse and refusal to participate were minimal (;3%) for both surveillance systems, which suggested that the samples are representative of Indonesia and Bangladesh.
These study protocols complied with the principles expressed in the Helsinki Declaration (30) and were approved by the ethical review committees of the Ministry of Health, Government of Indonesia and the Bangladesh Medical Research Council. Sec- ondary data analysis was approved by the institutional review board of the Johns Hopkins University School of Medicine.
Statistical analysis
Stunting was defined as a height-for-age z score ,22 ac- cording to the WHO growth standards (31) in AnthroPlus 2009 software. Maternal BMI was calculated as a ratio of weight (kg)/
height (m)2. Given our aim of informing policy and interven- tion strategies, “maternal overweight” was classified as a BMI (in kg/m2)23 to capture those at increased risk of NCDs (9, 32).
MCDB was defined as the coexistence of a stunted child and an overweight mother within the same household. This analysis in- cluded only the youngest child (6–59 mo of age) from each of the selected households. Children,6 mo of age and women with a BMI,12 or .50 were excluded from the analysis (9, 15). All analyses were weighted according to the population size and adjusted for the multistage cluster design of the NSS and NSP.
Weekly household per capita expenditure was used as the primary indicator of SES. The weekly expenditures on grain food items were added to nongrain food items. The monetary value of in-kind rice produced, received as in-kind payment for labor, or as a gift for the amount consumed by the household was added to the expenditure on grain foods. The sum of household expenditures on nonfood items, such as education and housing, were also calculated. Subsequently, the total weekly expenditures for all nonfood items, grain foods, and nongrain foods were summed and divided by household size to estimate a total household weekly per capita expenditure. Household size was measured as a simple count of household members. Quintiles of weekly per capita expenditure were created. Additional details on the cal- culation of household expenditure are described elsewhere (33).
Descriptive statistics were used to examine the full distribution of variables. Using appropriate cutoffs, categorical variables were created for maternal education (no education, primary school, secondary school, or. secondary school); maternal height (,145, 145–149.9, or150 cm; maternal short stature, ,145 cm); birth order (1–3 or 4); family size (2–3, 4–6, or 7); place of
by guest on August 23, 2016ajcn.nutrition.orgDownloaded from
defecation (open/closed latrine or other); and source of tap water (tap, hand pump, well, or other). The analyses were performed separately for rural Indonesia and Bangladesh. To determine factors associated with MCDB, bivariate analyses were con- ducted for all the various risk factors by using a chi-square test.
Risk factors were initially selected on the basis of their known association with maternal and child nutrition, as reflected in UNICEF’s conceptual framework on the causes of malnutrition.
These causes include child characteristics such as demography, morbidity, birth weight, dietary intake, breastfeeding, and vacci- nation status; maternal characteristics such as age, education, morbidity, reproductive history, personal hygiene, childcare prac- tices, and tobacco use; household composition; expenditure; and environmental indicators, among others. Only child, maternal, and household characteristics that were significantly associated with
MCDB in the bivariate analyses (P, 0.05) were included in the multiple logistic regression models. Separate logistic regression models were developed for Indonesia and Bangladesh with MCDB as the dependent variable and potential predictors as the in- dependent variables. ORs and corresponding 95% CIs were cal- culated with statistical significance defined as P , 0.05. All analyses were performed by using Predictive Analytics SoftWare Statistics (PASW, version 18.0; SPSS Inc).
RESULTS Indonesia
The mean (6SE) age of the mother included in the analysis was 28.06 0.01 y (Table 1). Most of the mothers had no or only
TABLE 1
Characteristics of mothers and their children aged 6–59 mo included in the sample Indonesia1 (n = 247,126)
Bangladesh2 (n = 168,317) Maternal characteristics
Age (y) 28.06 0.013 27.06 0.02
Education (%)
No schooling 6 43
Primary school 58 27
Secondary school 20 23
. Secondary school 16 7
Short stature,,145 cm (%) 17 14
BMI (kg/m2) 22.06 0.01 19.86 0.01
Overweight BMI,23 kg/m2(%) 32 12
Child characteristics
Age (mo) 25.36 0.27 30.16 0.04
Male (%) 51 52
Birth order (%)
1–3 84 75
4 16 25
Stunted height-for-age z score,,22 (%)4 37 42
Birth weight (kg) 3.176 0.52 2.886 0.69
Had diarrhea at least once in the past week(s) (%) 7 18
Currently breastfed (%)5 87 96
Received high-dose vitamin A capsule in past 6 mo (%) 66 64
Household characteristics (%) Family size6
2–3 17 16
4–6 62 59
7 21 25
Place of defecation
Open/closed latrine 67 85
Other 33 15
Source of water
Tap 10 1
Hand pump 23 47
Well 47 49
Other 20 3
1Missing values existed in the Indonesian sample, including the following: BMI (n = 9), child diarrhea (n = 1149), receipt of vitamin A capsule (n = 1462), and birth weight (n = 55,563).
2Missing values existed in the Bangladeshi sample, including the following: maternal age (n = 57), child diarrhea (n = 8), breastfeeding status (n = 28), receipt of vitamin A capsule (n = 435), place of defecation (n = 7), and birth weight (n = 162,947).
3Mean6 SE (all such values).
4Estimated by using 2006 WHO growth reference.
5Percentage of children 6–23 mo of age who were breastfed at the time of interview.
6Number of people eating from the same cooking pot in the same household.
by guest on August 23, 2016ajcn.nutrition.orgDownloaded from
primary education (64%), and ;17% had short stature. The mean (6 SE) BMI was 22 6 0.01, and 32% of the mothers were overweight. The mean (SE) age of children was 25.3 (0.3) mo with both sexes equally represented. Approximately 37% of children were stunted, and a large majority were currently being breastfed. Two thirds (66%) of the children had received a high- dose vitamin A capsule in the previous 6 mo, and slightly.7%
had had diarrhea in the past week.
An overweight mother and stunted child pair was observed in 11% of the households in rural Indonesia (Table 2). More specifically, 5.3% of households had a stunted child paired with a mother whose BMI was 23–24.9. Nearly 6% of the population of Indonesia had a stunted child in the same household as a mother who had a BMI25. Overweight persisted among an additional 21% of the population.
Several maternal, child, and household characteristics were associated with MCDB (Table 3). In multivariate analyses, maternal short stature and older age strongly predicted MCDB, whereas higher levels of maternal education were protective against MCDB. Child characteristics, such as older age, higher birth order, and being female, were associated with an increased odds of MCDB, whereas being breastfed at the time of interview was associated with a lower odds of MCDB. A large family size and a higher weekly per capita household expenditure were positively associated with MCDB (Table 4).
Bangladesh
The mean (6SE) age of the mothers was 27.0 6 0.02 y (Table 1). Approximately 70% of the mothers had no or only primary education, and 14% had short stature. The mean (6SE) BMI was 19.86 0.01, and 12% were overweight. The mean (6SE) age of the children was 30.16 0.04 mo, and about one-half (52%) were male. Approximately 42% of the children were stunted, and 96%
were breastfed at the time of interview. Nearly two-thirds (64%) of children had received a high-dose vitamin A capsule in the previous 6 mo, and 18% had had diarrhea in the past 2 wk (Table 1).
In rural Bangladesh, MCDB was observed in ;4% of the households (Table 2). Approximately 2% of the population had a stunted child paired with a mother whose BMI was 23–24.9. In total,;12% of the mothers surveyed in the sample were over- weight.
In multivariate analyses, maternal short stature and older age were associated with an increased odds of MCDB (Table 4).
Higher maternal education also predicted MCDB. Child char- acteristics, such as older age and being female, predicted MCDB, whereas being breastfed at the time of interview was protective against MCDB. A larger family size and a higher weekly per capita household expenditure were associated with an increased odds of MCDB (Table 4).
DISCUSSION
In the current analysis, we examined the prevalence and pre- dictors of MCDB in rural Indonesia and Bangladesh. To our knowledge, this article is the first to present trends in the co- existence of under- and overnutrition exclusively in a rural, de- veloping country setting using nationally representative data sets.
Our results suggest that the prevalence of MCDB is higher than previous estimates, particularly in Indonesia. In Indonesia and
Bangladesh, previous literature estimated the prevalence of double burden households to be 11% (11) and 2% (14), re- spectively. However, these studies used smaller samples from urban areas. Specifically in rural Asian countries, the prevalence of MCDB was estimated to be;1–6% of the population (5).
Overall, our findings confirm previous results that overweight among adult women (4, 9, 34, 35) and child stunting (5, 36, 37) persist in rural settings. Overweight among adult women in Asia is well documented in several developing countries (24). Nota- bly, this trend is also documented in several Asian rural pop- ulations, including Kazakhstan (36%), Uzbekistan (26%), China (15%), and India (6%) (4). The increasing prevalence of over- weight may be a result of shifts in eating and lifestyle with in- creased access to energy-dense foods. Moreover, mechanization in rural areas (ie, use of motorized vehicles to do agricultural work) has contributed to physical inactivity in developing countries (38). Similarly, child stunting is well documented in the developing world (14, 39, 40), and evidence suggests that it can be higher in the rural setting (5, 36). In Asia specifically, 31% of children younger than 5 y are stunted, and an estimated 16 countries have reported a child stunting prevalence of40%
(41). The health consequences of maternal overweight and child stunting are considerable (16, 20, 42, 43).
In both Indonesia and Bangladesh, maternal characteristics such as age, stature, and education level were strongly associated with the risk of MCDB. Most notably, maternal short stature increased the risk of MCDB in both countries. This result is consistent with that of Sichieri et al (44), who reported that BMI gain was significantly higher among short-statured women.
Relevant research shows that maternal short stature, reflecting malnutrition in early life, is associated with an increased risk of having cephalopelvic disproportion and stunted child (44, 45).
Child stunting, in turn, is also associated with an increased risk of obesity and chronic diseases in adulthood (45). In particular, Victora et al (16) suggest that rapid postnatal weight gain is linked to high glucose concentrations, blood pressure, and harmful lipid profiles. This intergenerational cycle of maternal short stature and child stunting helps to explain the association between maternal short stature and MCBD.
In the Indonesian sample, higher levels of maternal education were protective against MCDB. This association corroborates previous findings (14) indicating that discordant mother-child
TABLE 2
Mother-child pair with double burden of malnutrition in rural Indonesia and Bangladesh by maternal BMI1
Maternal BMI (kg/m2)
23–24.9 25
Child height-for-age z score [n (%)]2 Indonesia3
22 23,738 (9.6) 28,835 (11.7)
,22 13,072 (5.3) 14,271 (5.8)
Bangladesh4
22 7886 (4.7) 6459 (3.8)
,22 3737 (2.2) 2412 (1.4)
1Calculated by using cross tabulation.
2Estimated by using the 2006 WHO growth reference.
3Total sample size for Indonesia was 247,126.
4Total sample size for Bangladesh was 168,317.
by guest on August 23, 2016ajcn.nutrition.orgDownloaded from
pairs were significantly less likely to occur in households in which the mother had some type of formal education relative to those in which the mother has no formal education. On the other hand, in rural Bangladesh, higher levels of maternal education were associated with an increased risk of MCDB. Evidence in- dicates that the association between education and obesity in a developing country setting is complex and may even vary within the same country (46). Discordant results for education related to obesity in women suggest that traditional primary and secondary education may not be sufficient to reduce the risk of MCDB, but rather that specific education on nutrition and child feeding practices may be a more effective intervention strategy to reduce MCDB in rural populations.
Several important child characteristics were associated with an increased likelihood of MCDB, including older age, later birth, and
being female. A greater prevalence of stunting among females has been reported elsewhere (15, 36). Lee et al (15) reported that in households with a stunted child and an overweight mother, a majority (55.3%) had a female child. In addition to nutritional status, poverty, inequity, and social status are among the underlying social factors that may contribute to stunting. An analysis of nu- trition indicators globally indicates that there might be differences in the feeding and care of girls compared with boys, which leads to an increased prevalence of stunting among girls (36). This may stem from power relations and social norms that perpetuate discrimi- natory attitudes and practices toward females. The increased prevalence of stunting among females may account for the asso- ciation of being female with a higher odds of MCDB in this sample.
Households where a child was being breastfed at the time of interview were less likely to have a discordant mother-child pair,
TABLE 3
Predictors of mother-child pairs with a double burden of malnutrition in rural Indonesia1
n
Crude OR
(95% CI) P
Adjusted OR
(95% CI)2 P
Maternal characteristics Age
,20 y 9765 1.0 (Reference) — 1.0 (Reference) —
20–24 y 68,216 1.65 (1.50, 1.82) ,0.0001 1.40 (1.27, 1.54) ,0.0001
25–29 y 71,381 2.27 (2.07, 2.50) ,0.0001 1.79 (1.63, 1.97) ,0.0001
30 y 97,763 3.11 (2.83, 3.41) ,0.0001 2.22 (2.02, 2.44) ,0.0001
Education
No schooling 13,554 1.0 (Reference) — 1.0 (Reference) —
Primary school 144,404 0.87 (0.83, 0.92) ,0.0001 1.07 (1.02, 1.13) 0.011
Secondary school 49,233 0.71 (0.67, 0.75) ,0.0001 0.97 (0.91, 1.03) 0.322
. Secondary school 39,935 0.63 (0.59, 0.67) ,0.0001 0.78 (0.73, 0.84) ,0.0001
Height
,145 cm 40,710 2.28 (2.20, 2.35) ,0.0001 2.32 (2.25, 2.40) ,0.0001
145–149.9 cm 86,872 1.61 (1.57, 1.66) ,0.0001 1.63 (1.59, 1.68) ,0.0001
150 cm 119,544 1.0 (Reference) — 1.0 (Reference) —
Child characteristics Sex
Male 126,005 1.0 (Reference) — 1.0 (Reference) —
Female 121,121 1.06 (1.03, 1.08) ,0.0001 1.05 (1.02, 1.07) ,0.0001
Age
6–11 mo 45,329 1.0 (Reference) — 1.0 (Reference) —
12–23 mo 79,764 2.79 (2.65, 2.95) ,0.0001 2.70 (2.56, 2.84) ,0.0001
24 mo 122,033 4.12 (3.92, 4.33) ,0.0001 3.45 (3.26, 3.65) ,0.0001
Birth order
1–3 208,327 1.0 (Reference) — 1.0 (Reference) —
4 38,799 1.56 (1.51, 1.61) ,0.0001 1.20 (1.15, 1.24) ,0.0001
Breastfeeding status
Not breastfed 110,788 1.0 (Reference) — 1.0 (Reference) —
Currently breastfed 136,337 0.58 (0.57, 0.60) ,0.0001 0.84 (0.81, 0.87) ,0.0001
Household characteristics Family size
2–3 42,378 1.0 (Reference) — 1.0 (Reference) —
4–6 153,405 1.30 (1.25, 1.35) ,0.0001 1.18 (1.13, 1.23) ,0.0001
7 51,343 1.29 (1.24, 1.35) ,0.0001 1.25 (1.19, 1.31) ,0.0001
Household weekly per capita expenditure3
Quintile 1 48,774 1.0 (Reference) — 1.0 (Reference) —
Quintile 2 49,284 1.09 (1.04, 1.13) ,0.0001 1.11 (1.07, 1.16) ,0.0001
Quintile 3 49,558 1.12 (1.08, 1.17) ,0.0001 1.17 (1.12, 1.22) ,0.0001
Quintile 4 49,663 1.13 (1.08, 1.17) ,0.0001 1.21 (1.16, 1.26) ,0.0001
Quintile 5 49,847 1.20 (1.15, 1.25) ,0.0001 1.34 (1.28, 1.40) ,0.0001
1Total sample size for Indonesia was 247,126.
2Derived by using multiple logistic regression.
3Quintile 1 = poorest and quintile 5 = wealthiest.
by guest on August 23, 2016ajcn.nutrition.orgDownloaded from
which reinforces the notion that breastfeeding affects the nu- tritional status of both mother and child (47). Breastfeeding results in a greater likelihood of the child meeting their daily nutrient requirements; thus, they have a decreased prevalence of stunting. Maternal health outcomes have also been linked to breastfeeding, with the risk of maternal overweight being inversely associated with breastfeeding. Several possible ex- planations have been proposed, including the notion that breastfeeding increases the calorie expenditure of the mother. We recognize that this finding is primarily applicable to those children younger than 24 mo of age. In Indonesia and Bangla- desh, 87% and 96% of children aged 6–23 mo, respectively, were breastfed at the time of interview. Although not presented, multivariate logistic regression analyses were explored with a
stratified sample (6–23 mo compared with 24–59 mo of age), and the findings were similar to those presented in this article.
Household characteristics, including family size and SES, were also important factors explored in this analysis. Notably, higher SES as measured by weekly household expenditure was significantly associated with MCDB. Our findings were similar to those (11, 14) who reported that double-burden households were more likely to be among the wealthier quintiles. Expenditures on total food, nongrain food, and grain food were also explored (data not shown). Overweight among mothers was most prevalent among households in the highest quartiles of expenditure for total food, nongrain food, and grain food. Stunting and expenditures had an inverse relation. These results are consistent with the notion that there is a strong positive association between SES and
TABLE 4
Predictors of mother-child pairs with a double burden of malnutrition in rural Bangladesh1
n
Crude OR
(95% CI) P
Adjusted OR
(95% CI)2 P
Maternal characteristics Age
,20 y 11,333 1.0 (Reference) — 1.0 (Reference) —
20–24 y 54,786 2.17 (1.83, 2.57) ,0.0001 1.82 (1.53, 2.16) ,0.0001
25–29 y 50,294 3.23 (2.72, 3.82) ,0.0001 2.71 (2.27, 3.22) ,0.0001
30 y 51,903 3.85 (3.26, 4.56) ,0.0001 3.24 (2.70, 3.87) ,0.0001
Education
No schooling 73,454 1.0 (Reference) — 1.0 (Reference) —
Primary school 44,560 1.15 (1.07, 1.22) ,0.0001 1.21 (1.13, 1.29) ,0.0001
Secondary school 38,083 1.32 (1.23, 1.40) ,0.0001 1.49 (1.39, 1.60) ,0.0001
. Secondary school 12,220 1.63 (1.50, 1.79) ,0.0001 1.57 (1.43, 1.73) ,0.0001
Height
,145 cm 23,504 1.85 (1,72, 1.98) ,0.0001 2.11 (1.96, 2.26) ,0.0001
145–149.9 cm 52,851 1.49 (1.41, 1.58) ,0.0001 1.58 (1.50, 1.68) ,0.0001
150 cm 91,962 1.0 (Reference) — 1.0 (Reference) —
Child characteristics Sex
Male 88,271 1.0 (Reference) — 1.0 (Reference) —
Female 80,045 1.08 (1.02, 1.13) 0.004 1.08 (1.02, 1,14) 0.004
Age
6–11 mo 20,779 1.0 (Reference) — 1.0 (Reference) —
12–23 mo 43,717 1.92 (1.70, 2.18) ,0.0001 1.84 (1.63, 2.09) ,0.0001
24 mo 103,820 2.96 (2.64, 3.32) ,0.0001 1.85 (1.63, 2.09) ,0.0001
Birth order
1–3 126,392 1.0 (Reference) — 1.0 (Reference) —
4 41,925 1.28 (1.21, 1.35) ,0.0001 1.00 (0.93, 1.07) 0.946
Breastfeeding status
Not breastfed 58,319 1.0 (Reference) — 1.0 (Reference) —
Currently breastfed 109,970 0.45 (0.42, 0.47) ,0.0001 0.55 (0.52, 0.58) ,0.0001
Household characteristics Family size
2–3 26,702 1.0 (Reference) — 1.0 (Reference) —
4–6 99,341 1.28 (1.18, 1.39) ,0.0001 1.07 (0.98, 1.17) 0.118
7 42,274 1.61 (1.48, 1.76) ,0.0001 1.34 (1.22, 1.47) ,0.0001
Household weekly per capita expenditure3
Quintile 1 33,663 1.0 (Reference) — 1.0 (Reference) —
Quintile 2 33,663 1.15 (1.04, 1.26) 0.005 1.16 (1.05, 1.27) 0.002
Quintile 3 33,663 1.42 (1.30, 1.55) ,0.0001 1.44 (1.31, 1.58) ,0.0001
Quintile 4 33,663 1.78 (1.63, 1.94) ,0.0001 1.74 (1.59, 1.90) ,0.0001
Quintile 5 33,662 2.10 (1.93, 2.28) ,0.0001 1.94 (1.77, 2.12) ,0.0001
1Total sample size for Bangladesh was 168,317.
2Derived by using multiple logistic regression.
3Quintile 1 = poorest and quintile 5 = wealthiest.
by guest on August 23, 2016ajcn.nutrition.orgDownloaded from
overweight in less-developed countries (4), because those in wealthier quintiles may have increased calorie intakes and sedentary behavior. This supports the idea that Indonesia and Bangladesh are in the earlier phases of the economic and nutrition transition.
Some limitations to this study should be considered. First, this was a cross-sectional analysis that explored the predictors of MCDB, so causation cannot be established. Second, although low food variety characterized by micronutrient malnutrition may increase the risk of MCDB (48), the dietary intakes of women and children were not considered. Predictors of maternal obe- sity, such as history of physical activity, prepregnancy weight, and weight gain during pregnancy, were not available. An in- sufficient number of observations prohibited us from considering parity, frequency of breastfeeding, or age at which supplementary feeding commenced in the multivariate analysis. However, the literature (49, 50) suggests that the effects of postpartum weight gain may be minimal in our samples. Although an important variable to consider, the birth weight of children was not included in the multivariate analysis because it was not statistically sig- nificant when explored in bivariate analysis. Finally, operational definitions of double burden vary. MCDB, as defined in the current analysis, may also lack comparability with definitions in other countries and contexts.
Despite these limitations, examining the key risk factors of stunted children and overweight mothers is a critical step in developing nutrition intervention programs that target both extremes of the malnutrition paradox. The current analysis draws attention to the need to consider a broad definition of malnu- trition. These findings suggest that MCDB is not exclusive to urban settings and that maternal short stature may be a strong risk factor, whereas the role of maternal education needs further exploration. These findings reinforce evidence of an association between economic factors and malnutrition, yet suggest that future policies and interventions need to address malnutrition in both rural and urban areas.
The authors’ responsibilities were as follows—VMO, JHR, and KK: con- ceptualized and designed the research; VMO and JR: analyzed the data, wrote the manuscript, and had major responsibility for the final content; and RDS and KS: provided statistical expertise. All authors contributed to the interpre- tation of the data, read the manuscript, made a substantial contribution to the revision, and approved the final manuscript. None of the authors declared a conflict of interest.
REFERENCES
1. World Health Organization. Overweight and obesity. 2011. Available from: http://www.who.int/mediacentre/factsheets/fs311/en/index.html (cited 8 April 2011).
2. Popkin BM. The nutrition transition and obesity in the developing world. J Nutr 2001;131:871S–3S.
3. Popkin BM, Gordon-Larsen P. The nutrition transition: worldwide obesity dynamics and their determinants. Int J Obes Relat Metab Disord 2004;28:S2–9.
4. Mendez MA, Monteiro CA, Popkin BM. Overweight exceeds un- derweight among women in most developing countries. Am J Clin Nutr 2005;81:714–21.
5. Garrett JL, Ruel MT. Stunted child-overweight mother pairs: preva- lence and association with economic development and urbanization.
Food Nutr Bull 2005;26:209–21.
6. Kennedy G, Nantel G, Shetty P. Assessment of the double burden of malnutrition in six case study countries. In: The double burden of malnutrition. Food and Nutrition Paper 84. Rome, Italy: Food and Agriculture Organization of the United Nations, 2006:1–18.
7. Shukla HC, Gupta PC, Mehta HC, Hebert JR. Descriptive epidemiol- ogy of body mass index of an urban adult population in western India.
J Epidemiol Community Health 2002;56:876–80.
8. Monteiro CA, Conde WL, Popkin BM. The burden of disease from undernutrition and overnutrition in countries undergoing rapid nutrition transition: a view from Brazil. Am J Public Health 2004;94:433–4.
9. Shafique S, Akhter N, Stallkamp G, de Pee S, Panagides D, Bloem MW. Trends of under- and overweight among rural and urban poor women indicate the double burden of malnutrition in Bangladesh. Int J Epidemiol 2007;36:449–57.
10. Floreˆncio TM, Ferreira HS, de Franca AP, Cavalcante JC, Sawaya AL.
Obesity and undernutrition in a very-low-income population in the city of Maceio, Northeastern Brazil. Br J Nutr 2001;86:277–84.
11. Doak CM, Adair LS, Bentley M, Monteiro C, Popkin BM. The dual burden household and the nutrition transition paradox. Int J Obes (Lond) 2005;29:129–36.
12. Saibul N, Shariff ZM, Lin KG, Kandiah M, Ghani NA, Rahman HA.
Food variety score is associated with dual burden of malnutrition in Orang Asli (Malaysian indigenous peoples) households: implications for health promotion. Asia Pac J Clin Nutr 2009;18:412–22.
13. Fernald LC, Neufeld L. Overweight with concurrent stunting in very young children from rural Mexico: prevalence and associated factors.
Eur J Clin Nutr 2007;61:623–32.
14. Jehn M, Brewis A. Paradoxical malnutrition in mother-child pairs:
untangling the phenomenon of over- and under-nutrition in under- developed economies. Econ Hum Biol 2009;7:28–35.
15. Lee J, Houser RF, Must A, de Fulladolsa PP, Bermudez OI. Disen- tangling nutritional factors and household characteristics related to child stunting and maternal overweight in Guatemala. Econ Hum Biol 2010;8:188–96.
16. Victora CG, Adair L, Fall C, Hallal PC, Martorell R, Richter L, Sachdey HS, Maternal and Child Undernutrition Study Group. Maternal and child undernutrition: consequences for adult health and human capital.
Lancet 2008;371:340–57.
17. Sawaya AL, Martins PA, Grillo LP, Florencio TT. Long-term effects of early malnutrition on body weight regulation. Nutr Rev 2004;62:
S127–33.
18. Sawaya AL, Sesso R, Florencio TM, Fernandes MT, Martins PA. As- sociation between chronic undernutrition and hypertension. Matern Child Nutr 2005;1:155–63.
19. Sawaya AL, Martins PA, Baccin Martins VJ, Floreˆncio TT, Hoffman D, do Carmo P, Franco M, das Neves J. Malnutrition, long-term health and the effect of nutritional recovery. Nestle Nutr Workshop Ser Pe- diatr Program 2009;63:95–105.
20. Prentice AM. The emerging epidemic of obesity in developing coun- tries. Int J Epidemiol 2006;35:93–9.
21. Prentice AM. Obesity in emerging nations: evolutionary origins and the impact of a rapid nutrition transition. Nestle Nutr Workshop Ser Pediatr Program 2009;63:47–54.
22. Monteiro CA. Trends in the burden of disease due to under- and over-nutrition in the developing world. Forum Nutr 2003;56:266–7.
23. World Bank. Countries and economies. 2011. Available from: http://
www.worldbank.org (cited 8 April 2011).
24. Khor GL. Food-based approaches to combat the double burden among the poor: challenges in the Asian context. Asia Pac J Clin Nutr 2008;
17:111–5.
25. Asian Development Bank. Key indicators 2006: measuring policy ef- fectiveness in health and education. 2006. Available from: http://www.
adb.org/ (cited 12 June 2011).
26. Bloem MW, Moench-Pfanner R, Panagides D. Health and nutritional surveillance for development. Singapore: Helen Keller Worldwide, 2001.
27. De Pee S, Bloem MW, Sari M, Kiess K, Yip R, Kosen S. The high prevalence of low hemoglobin concentration among indonesian infants aged 3-5 months is related to maternal anemia. J Nutr 2002;132:
2215–21.
28. de Pee S, Bloem MW. Assessing and communicating the impact of nutrition and health programs. In: Semba RD, Bloem MW, eds. Nu- trition and health in developing countries. Totowa, NJ: Humana Press, 2001:483–506.
29. Gibson RS. Principals of nutritional assessment. New York, NY:
Oxford University Press, 1990.
30. World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. Bull World Health Organ 2001;79:373–4.
by guest on August 23, 2016ajcn.nutrition.orgDownloaded from
31. World Health Organization. WHO child growth standards: methods and development. Geneva, Switzerland: WHO, 2006.
32. World Health Organization Expert Consultation. Appropriate body- mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004;363:157–63.
33. Thorne-Lyman AL, Valpiani N, Sun K, Semba RD, Klotz CL, Kraemer K, Akhter N, de Pee S, Moench-Pfanner R, Sari M, et al. Household dietary diversity and food expenditures are closely linked in rural Bangladesh, increasing the risk of malnutrition due to the financial crisis. J Nutr 2010;140:182S–8S.
34. Lim TO, Ding LM, Zaki M, Suleiman AB, Fatimah S, Siti S, Tahir A, Maimunah AH. Distribution of body weight, height and body mass index in a national sample of Malaysian adults. Med J Malaysia 2000;55:108–28.
35. Usfar AA, Lebenthal E, Atmarita E, Achadi E, Soekirman H. Obesity as a poverty-related emerging nutrition problems: the case of Indonesia.
Obes Rev 2010;11:924–8.
36. United Nations Children’s Fund. Tracking progress on child and ma- ternal nutrition. New York, NY: UNICEF, 2009.
37. Rah JH, Akhter N, Semba RD, de Pee S, Bloem MW, Campbell AA, Moench-Pfanner R, Sun K, Badham J, Kraemer K. Low dietary di- versity is a predictor of child stunting in rural Bangladesh. Eur J Clin Nutr 2010;64:1393–8.
38. Misra A, Khurana L. Obesity and the metabolic syndrome in de- veloping countries. J Clin Endocrinol Metab 2008;93:S9–30.
39. United Nations Children’s Fund. The state of the world’s children 2009. New York, NY: UNICEF, 2008.
40. Petrou S, Kupek E. Poverty and childhood undernutrition in developing countries: a multi-national cohort study. Soc Sci Med 2010;71:1366–73.
41. Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M, Mathers C, Rivera J, Maternal and Child Undernutrition Study Group.
Maternal and child undernutrition: global and regional exposures and health consequences. Lancet 2008;371:243–60.
42. World Health Organization. Diet, nutrition and the prevention of chronic diseases: report of a Joint WHO/FAO Expert Consultation.
Geneva, Switzerland: WHO, 2003.
43. Martins VJB, Toledo Floreˆncio TM, Grillo LP, Grillo LP, Franco Mdo C, Martins PA, Clemente AP, Santos CD, Vieira Mde F, Sawaya AL.
Long-lasting effects of undernutrition. Int J Environ Res Public Health 2011;8:1817–46.
44. Sichieri R, Silva CV, Moura AS. Combined effect of short stature and socioeconomic status on body mass index and weight gain during re- productive age in Brazilian women. Braz J Med Biol Res 2003;36:
1319–25.
45. Ferreira HS, Moura FA, Cabral CR Jr, Florencio TM, Vieira RC, de Assuncao ML. Short stature of mothers from an area endemic for undernutrition is associated with obesity, hypertension and stunted children: a population-based study in the semi-arid region of Alagoas, Northeast Brazil. Br J Nutr 2009;101:1239–45.
46. Monteiro CA, Conde WL, Popkin BM. Independent effects of income and education on the risk of obesity in the Brazilian adult population.
J Nutr 2001;131:881S–6S.
47. Agostoni C, Braegger C, Decsi T, Kolacek S, Koletzko B, Michaelsen KF, Mihatsch W, Moreno LA, Puntis J, Shamir R, et al. Breast-feeding:
a commentary by the ESPGHAN Committee on Nutrition. J Pediatr Gastroenterol Nutr 2009;49:112–25.
48. Hatløy A, Hallund J, Diarra MM, Oshaung A. Food variety, socio- economic status and nutritional status in urban and rural areas in Koutiala (Mali). Public Health Nutr 2000;3:57–65.
49. Winkvist A, Stenlund H, Hakimi M, Nurdiati D, Dibley M. Weight- gain patterns from prepregnancy until delivery among women in Central Java, Indonesia. Am J Clin Nutr 2002;75:1072–7.
50. Coitinho DC, Sichieri R, D’Aquino Benı´cio MH. Obesity and weight change related to parity and breast-feeding among parous women in Brazil. Public Health Nutr 2001;4:865–70.
by guest on August 23, 2016ajcn.nutrition.orgDownloaded from
