In our analytic sample comprising the most recent year of surveys across South Asia, approximately 2.6% of children aged 0–59 months in the sample were overweight, ranging from 1.3% in Nepal to 5.8% in the Maldives. While this is slightly lower than the estimated overall South Asia average for 2016 of 4.3% (2.4% to 7.5%), the rate of growth is certainly worrisome. The prevalence of overweight among children has increased considerably in South Asia since 1990, and it is likely to continue. In highlighting the two- to five-fold rise in overweight and obesity in Afghanistan, Bangladesh and India since the early 2000s, Aguayo and Paintal (2017) pointed out that the diet of adolescent girls and younger children remains poor, and that nutrition programmes do not focus on managing the seriousness of the double burden of malnutrition across South Asia [
39].
4.1. Pre-School Aged Children
The literature presents some inconsistencies regarding the relationship between family history of obesity and child obesity in South Asia. For example, two independent studies among school-aged children in urban Bangladesh and affluent India each concluded that a family history of obesity increased children’s likelihood of being obese [
40,
41]. Yet, another study among affluent school-aged children in India found no significant association between child and parental obesity [
42]. The current analysis provides clear and consistent results across countries of a significant association between maternal overweight and child WHZ as well as child overweight in South Asia. Furthermore, in a similar pooled analysis examining child wasting, maternal underweight was associated with child wasting, further supporting this relationship [
27]. Prior research supports that both environmental and genetic factors may in part explain the association between maternal overweight and child WHZ and overweight. Maternal overweight and obesity puts women at a greater risk of pre-term delivery (<32 weeks gestation) [
43,
44,
45]. Premature birth in turn increases children’s risk of morbidity, mortality, and cognitive deficits [
12,
46,
47,
48]. Furthermore, the metabolic consequences that result from premature birth can be exacerbated in growing obesogenic environments, for instance regions undergoing rapid urbanization and nutrition transition [
12,
49]. Thus, the mechanisms through which maternal overweight and obesity impact child WHZ and overweight, including epigenetics and fetal programming, need to be further explored.
MDD and MAD, two indicators of appropriate young child feeding practices, were positively associated with increased mean WHZ in children and with a higher prevalence of overweight; at the same time, MMF was only associated with increased mean WHZ and not with overweight. In a region where the mean WHZ in under-fives is significantly below the mean WHZ of a well-nourished reference population, the significant association of complementary feeding indicators with increased mean WHZ is encouraging. However, the association of MDD and MAD with overweight needs further exploration; it suggests that while achieving diet diversity as a policy goal typically aimed at tackling undernutrition, how it is achieved may matter in the context of risk factors for overweight and obesity. As the prevalence of childhood overweight and obesity rise, monitoring infant and young child feeding practices among different populations in relation to these outcomes will be important. For instance, breastfeeding is known to have protective effects against overweight and obesity and positive survival, growth and development outcomes universally [
50,
51]. In contrast, some parental feeding styles (i.e., indulgent or restrictive) are associated with child weight, although the specific parental feeding styles that are associated with child weight vary across contexts [
52,
53,
54].
4.2. Adolescent Girls
Mean BMI among adolescent girls from the most recent year of surveys in this study ranged from 19.4 (±3.0) kg/m2 in India (2006) to 21.9 (±4.6) kg/m2 in the Maldives, and the prevalence of overweight ranged between 4.4% in Nepal and 24.5% in the Maldives. In our sample, the prevalence of overweight in adolescent girls was 4.6% using BMI ≥25.0 kg/m2 and 11.6% using BMI ≥23.0 kg/m2.
Overall, adolescent girls were more likely to be overweight if living in urban households, even controlling for wealth. This finding is consistent with the work of Jaaks et al. (2015) which considered the double burden through 53 Demographic and Health surveys globally; they found that 38% of urban areas had both an underweight and overweight prevalence exceeding 10% [
55]. That said, there is ample evidence that overweight and obesity are increasing in rural as well as poorer households across Asia [
1,
39]. The rising prevalence of overweight in our Bangladesh sample is of particular concern since the prevalence among adolescents increased from <1% in 1996 to >8% in 2014; that is a rate of change of 0.16 percentage points per year in the first 10 years, followed by 0.14 percentage points per year in the last 7 years (2007–2014). The rise in obesity was steeper still, at 0.44 percentage points in the years up to 2014.
Our analysis, agreeing with that of Leroy et al. (2018) on Bangladesh, suggests that wealth and education are not consistently significant moderators of the rising burden of overweight among girls or adolescents [
56]. In other words, while rapid poverty reduction across Bangladesh has driven a successful lowering of child stunting in recent decades, rising incomes and access to education have not prevented the concomitant acceleration in obesity outcomes.
That said, across our sample as a whole we found that formal education was not significantly associated with (i.e., not promoting) overweight among adolescent girls once stratified by country, despite the fact that higher education tends to be associated with more overweight and obesity elsewhere [
14]. Ideally, formal education systems should be negatively correlated with both forms of malnutrition, but directly via improved knowledge of optimal nutrition and health practices, and indirectly via income-earning potential. South Asia’s education systems are in effect passive bystanders to the unfolding of a hugely significant nutrition and health crisis; they must do more to inform, educate and promote improved dietary choices, physical activity and healthy behaviours [
39].
Finally, it has been highlighted that overweight and obesity among adolescent girls is inconsistently measured or reported, and that when reported reports often include different age groups and definitions for overweight and obesity [
57]. This was evident across the surveys included in our sample, with some surveys reporting on all adolescent girls and, in most, the category of adolescent girls was extrapolated from the sample of women of reproductive age (15- to 49-year-olds), generating an important data gap among younger adolescent girls. Thus, South Asian countries need to collect nationally representative data on the nutrition of adolescent girls and boys aged 10–19 disaggregated by sex, age group, geographic region, and socioeconomic status. Data need to show the extent and severity of the triple burden of poor nutrition in adolescents: growth failure (stunting and wasting), micronutrient deficiencies and anaemia, and overweight and obesity. There must also be measures of the adequacy of adolescents’ diets, and the socioeconomic determinants affecting this. The effect of national policies and programmes for adolescent nutrition must be measured, including their coverage, effectiveness, and equity. This evidence should be used to develop national policies and to scale up cost-effective programmes [
39].
4.3. Adult Women
Adult women in this sample had a mean BMI of 22.3 (±4.3) kg/m
2 but 22% of the women were overweight with a BMI ≥25.0 kg/m
2. India had the lowest prevalence of overweight among women in the region (24%), while the Maldives (46%) and Pakistan (41%) had levels that are significantly above the global estimate of overweight among women (38%) [
1]. Women with some formal education, from wealthier homes and living in urban areas were more likely to be overweight. Similar findings have been reported for Sri Lanka among women aged 35 to 64 years [
9]. Furthermore, the overall association between wealth and overweight in low and middle-income countries has been reported in the literature [
58,
59].
In our analysis, women’s mean BMI increased with survey year in Bangladesh, India and Nepal at rates between <0.001 and 0.011 kg/m
2 per year; this is lower than the recent global estimate presented by Popkin et al. of 0.4 to 0.5 kg/m
2 per year increase over past 3 decades [
12]. The rate of change in overweight and obesity among Bangladeshi and Nepalese women in this study between the mid−2000 s and the 2010 s was approximately 0.1 percentage point and 0.14–0.18 percentage point per year, respectively, suggesting that obesity may be increasing at a faster rate than overweight. The rate of change in mean BMI, overweight and obesity appears to be lower in India compared with Bangladesh and Nepal, yet the most recent prevalence of both overweight and obesity are similar across these three countries.
The relationship between the key exposure variables evaluated in this study (formal education, household wealth, and urban residence) and the prevalence of overweight among women attenuated over time. On a global level, the gap in the prevalence of overweight between urban and rural women is narrowing, with the prevalence increasing at a quicker rate in rural areas compared with urban settings [
12]. In a longitudinal analysis of cross-sectional data from China, Jones-Smith et al. reported similar findings to those of the current study; over time, the prevalence of overweight increased at a greater rate among women with lower education than among women with higher education [
14]. A study assessing trends across 39 countries found that high socioeconomic status was associated with greater overweight prevalence among women, but that higher gross national product (GNP) was associated with a relatively greater overweight prevalence among women of lower socioeconomic status [
10].
It is possible that the trends observed in Bangladesh, India, and Nepal are also related to their status as economies emerging from a history of pervasive poverty and limited economic growth. While the prevalence of overweight and obesity remain higher among more educated, wealthier urban women in South Asia, these findings suggest that attention should also be paid to overweight and obesity among those with less education, from poorer households and in rural areas, as the prevalence of overweight seems to be increasing more rapidly in these subgroups that are also more vulnerable to stunting, wasting, micronutrient deficiencies, and diet-related non-communicable diseases (NCDs).
There are some important risk factors associated with overweight that could not be evaluated in this study. A balance of energy intake and expenditure translate into weight loss, maintenance or gain, and several individual studies have documented the impact of dietary choices and physical activity on overweight and obesity [
7,
60,
61]. As populations transition from labour-intensive jobs and high carbohydrate diets to more sedentary jobs and diets high in edible oils, sweeteners and animal source foods, diet-related non-communicable diseases tend to increase [
12]. On a global level, Green et al. found a significant association between energy available from meat, dairy products and vegetable oils with higher ischemic health disease [
62]. To address the growing concerns of overweight and obesity in South Asia, a critical examination of the food environment, access to good quality diets, and the effectiveness of food-based programming in relation to all forms of malnutrition is urgently needed.
Overweight has substantial health and economic costs to individuals and countries. Maternal overweight is associated with poor health outcomes and huge medical costs associated with gestational diabetes and pre-eclampsia, preterm birth and maternal and infant mortality [
63]. Economically, costs can be broken down to direct medical and non-medical costs and indirect costs of morbidity and mortality [
64]. One systematic review of economic costs has estimated that obese individuals had 30% greater medical costs compared to individuals with normal weight [
65].
South Asia has the largest number of undernourished children, anemic women and small for gestational age births globally [
63], all of which continue to have high social and economic implications on the region. The added burden of overweight and the associated implications are on the rise and require rapid policy and programme attention.
4.4. Limitations and Strengths
This study has several limitations and strengths. The data were aggregated from cross-sectional surveys and the survey weights from each dataset were not used in this pooled analysis, and thus should be interpreted as such. While we attempted to access data from all South Asian countries, similar data were not accessible from Bhutan and Sri Lanka at the time of the analysis. Furthermore, data on adolescent girls are lacking throughout the region. Standard DHS modules do not include an evaluation of the nutritional status of adolescents specifically, thus our sample of adolescent girls was limited to the 15- to 19-year-old group, which are included in the DHS standard survey module for “women of reproductive age”. For consistency, we used this age range from the Afghanistan survey as well.
This study highlights the importance of overweight as a growing public health issue in South Asia among pre-school age children, adolescent girls, and adult women. By pooling compatible datasets from national surveys across South Asia over time, we were able to examine which factors are significantly associated with overweight using a large sample.
The World Health Assembly has adopted six global nutrition targets for the year 2025, including the goal of “no increase in childhood overweight” [
66]. This target was established because of the serious health risks associated with child overweight and obesity, which in turn often leads to adult overweight and obesity. Most countries in South Asia are currently off-target. In addition, there is a voluntary global target in the WHO Global Monitoring Framework for Non-Communicable disease to halt the rise in obesity by 2025 and a call to monitor non-communicable disease [
67]. Policymakers in South Asia will need to urgently prioritize investments to reverse recent trends in overweight in children, adolescents and women and address a situation that is becoming of critical public health concern.