Maternal nutrition and appropriate gestational weight gain

Few studies have explored gestational weight gain knowledge and beliefs in LMIC. One study, conducted in Nigeria, found that 40·1 and 35·4 % of mothers were classified as overweight and obese, respectively, according to pre-pregnancy BMI⁽ Reference Olagbuji, Olofinbiyi and Akintayo ^{13 )}. However, none of the obese mothers perceived themselves as obese⁽ Reference Olagbuji, Olofinbiyi and Akintayo ^{13 )}. In contrast, 32·0 % of the overweight mothers perceived themselves as overweight, and those who perceived themselves as overweight were more likely to gain the recommended amount of weight during pregnancy⁽ Reference Olagbuji, Olofinbiyi and Akintayo ^{13 )}. A qualitative study in Egypt found that women were not aware of and had no concept of appropriate gestational weight gain due to a lack of counselling by health-care providers⁽ Reference Kavle, Mehanna and Khan ^{14 )}. Given that maternal obesity and excess gestational weight gain have been associated with increased adiposity in offspring, largely in studies conducted in developed countries⁽ Reference Poston ^{15 )}, this is an important research gap in LMIC. In addition, insufficient maternal intakes of energy and micronutrients have been associated with increased adiposity in offspring, particularly female offspring – as reviewed by Yang and Huffman⁽ Reference Yang and Huffman ^{16 )} and Christian and Stewart⁽ Reference Christian and Stewart ^{17 )}. For example, maternal energy intake of less than 3766 kJ/d (900 kcal/d) during the Dutch famine of 1944–1945 was associated with greater fat deposition in female offspring (average age at follow-up approximately 60 years), but not males⁽ Reference Stein, Kahn and Rundle ^{18 )}. Two secondary analyses of trials, one in Nepal⁽ Reference Stewart, Christian and LeClerq ^{19 )} and one in Peru⁽ Reference Iannotti, Zavaleta and Leon ^{20 )}, found increased linear growth (Nepal) and lean body mass (Peru), and, in the case of Nepal, decreased adiposity, among offspring of women who received Zn+Fe+folic acid during pregnancy v. offspring of women who received Fe+folic acid.

Infant and young child nutrition and healthy growth trajectories

Fetal growth and growth in the first 2 years of life are important predictors of adult weight status⁽ Reference Darnton-Hill, Nishida and James ^{21 )}. A recent analysis of the Consortium on Health Oriented Research in Transitional Societies (COHORTS) study, which includes five LMIC (Brazil, Guatemala, India, the Philippines and South Africa), found that birth weight was more strongly associated with adult lean body mass than with adult fat mass⁽ Reference Kuzawa, Hallal and Adair ^{22 )}. Similarly, a study in Pune, India, found that the small birth size of babies was largely due to reduced skeletal muscle rather than reduced body fat⁽ Reference Yajnik, Lubree and Rege ^{23 )}, and a study in New Delhi, India, found that birth weight was positively associated with adult lean body mass but not adult central adiposity⁽ Reference Sachdev, Fall and Osmond ^{24 )}. In contrast, with respect to postnatal growth, a meta-analysis of fifteen studies reported an 84 % increase in risk of obesity at 10 years of age per 0·67 sd gain of weight in infancy⁽ Reference Ong and Loos ^{25 )}. Data from the COHORTS study suggest that both faster relative weight gain and faster linear growth among children were associated with increased risk of adult overweight⁽ Reference Adair, Fall and Osmond ^{26 )}. However, it is important to note that the magnitude of these associations in the COHORTS study was small: a 1 sd change in weight gain up to 12 or 24 months predicted less than a 1% change in adult body composition⁽ Reference Kuzawa, Hallal and Adair ^{22 )}. Together, these results suggest two key points:

1. 1. improved birth weight, as an indicator of fetal growth, may be protective against overweight via building adult lean body mass; and

2. 2. the known benefits of growth in the first 2 years of life on health and function⁽ Reference Victora, Adair and Fall ^{27 )} and the known adverse consequences of growth faltering in early life on adult stature⁽ Reference Li, Stein and Barnhart ^{28 )} are likely to outweigh any negative effects on adult obesity resulting from rapid ‘catch-up’ growth during this period⁽ Reference Adair, Fall and Osmond ^{26 )}.

While several systematic reviews and meta-analyses, including a WHO-led review⁽ Reference Horta ^{29 )}, have found protective effects of breast-feeding on obesity in developed countries, including the USA, Canada, UK, Germany, Australia, New Zealand and former Czechoslovakia, there is a need for additional studies to provide stronger evidence because these studies did not account for publication bias and residual confounding by maternal socio-economic status and lifestyle habits⁽ Reference Casazza, Fernandez and Allison ^{30 )}. One randomized controlled trial on promotion of exclusive breast-feeding in Belarus did not find any effect on childhood obesity⁽ Reference Kramer, Matush and Vanilovich ^{31 )}. A significant evidence gap remains on whether any relationship between breast-feeding and overweight exists in LMIC. A study in Brazil did not find a significant association between breast-feeding duration and BMI⁽ Reference Brion, Lawlor and Matijasevich ^{32 )}, while a study in India found only a weak association between breast-feeding duration and BMI and no association with skinfold thickness⁽ Reference Caleyachetty, Krishnaveni and Veena ^{33 )}. Regardless of whether breast-feeding is or is not protective against overweight, the strong evidence supporting other benefits of breast-feeding^{( 34 )} warrants continued support of breast-feeding recommendations^{( 35 )}.

With regard to complementary feeding, the limited data available from LMIC suggest that suboptimal complementary feeding practices are common and include energy-dense, nutrient-poor foods. For example, a survey of 700 mothers of 6- to 18-month-old infants in Ibadan, Nigeria, found that, on a daily basis, 65·0 % regularly gave biscuits to their infants, 16·1 % gave soft drinks and 9·6 % gave commercial fruit juice⁽ Reference Bankole, Aderinokun and Odenloye ^{36 )}. In addition, 57·0 % of mothers sweetened their infants’ pap (maize-based porridge) with sugar⁽ Reference Bankole, Aderinokun and Odenloye ^{36 )}. In a study of 4299 children living in two slums outside Nairobi, Kenya, 41 % had received sweetened/flavoured water in the first 6 months of life⁽ Reference Kimani-Murage, Madise and Fotso ^{37 )}. With regard to breast-milk substitutes, high-protein infant formulas in the first 2 years of life have been associated with increased adiposity in childhood relative to control formulas in several randomized controlled trials conducted in Europe⁽ Reference Koletzko, von Kries and Closa ^{38 ,} Reference Singhal, Kennedy and Lanigan ^{39 )}, but this relationship has not been explored in LMIC.

Value and social standing

Few studies have explored drivers of food choice in LMIC. One qualitative study, conducted in rural Kerala, India, found that the two most important drivers of food choice at the household level were affordability and taste preferences of the children and husband⁽ Reference Daivadanam, Wahlström and Thankappan ^{40 )}. For example, one woman said, ‘… we use lot of oil … difficult to reduce, he [husband] likes lot of fried things’⁽ Reference Daivadanam, Wahlström and Thankappan ^{40 )}. Processed foods were perceived as ‘higher value’ based on the media’s portrayal of these foods as necessary for healthy growth and development of children⁽ Reference Daivadanam, Wahlström and Thankappan ^{40 )}. This can often lead to higher-than-affordable expenditures on these foods: for example, one woman said, ‘… spend an average four-thousand rupees on food per month … have to spend an extra one to two thousand on biscuits and powder items [health drinks] for small children in the house’⁽ Reference Daivadanam, Wahlström and Thankappan ^{40 )}. Similarly, expensive foods such as restaurant foods were considered ‘better’ than foods that were cheaper and could be prepared at home⁽ Reference Daivadanam, Wahlström and Thankappan ^{40 )}. A study in Nigeria also reported that consumption of fast foods was associated with higher social standing⁽ Reference Akpan and Ekpenyong ^{41 )}. In contrast, in Egypt, processed junk foods for children were not perceived as expensive and were within economic reach of most families in both rural and urban areas⁽ Reference Kavle, Mehanna and Saleh ^{42 )}. Also in Egypt, junk food consumption was not just an issue among children; mothers and other family members also consumed these foods⁽ Reference Kavle, Mehanna and Saleh ^{42 )}.

Time and effort for food preparation

As women enter the workforce, their time to prepare food becomes limited. For example, in the aforementioned qualitative study conducted in rural Kerala, India, one mother stated: ‘I come back from work and I just try to make something fast for the children as they are hungry by then’⁽ Reference Daivadanam, Wahlström and Thankappan ^{40 )}. In contrast, in lower-income households in these communities in rural Kerala, the limitation was affordability, not time: ‘Most people complain that it takes a lot of time and effort to cut vegetables. That is only for those who don’t have time. Here, we have time, but no vegetables and no money to buy it’⁽ Reference Daivadanam, Wahlström and Thankappan ^{40 )}. In Egypt, reducing consumption of junk food is challenging because mothers often give these foods out of convenience, stating that, ‘I have no time to cook for my children’, ‘I have no free time’ and ‘I felt lazy’⁽ Reference Kavle, Mehanna and Saleh ^{42 )}. Moreover, data from Egypt suggest that older siblings care for younger children and play an important role in feeding young children when mothers are away from home⁽ Reference Kavle, Mehanna and Saleh ^{42 )}. A recent analysis of data from Cambodia found that women increasingly have sales and service jobs, especially in urban settings, and that this is linked to overweight and obesity, potentially through time pressures in addition to increased sedentary time⁽ Reference Espeut, Coffey and Nyaku ^{43 )}. Further research is needed that focuses on the impact of women entering the workforce on weight status in LMIC and on potential mechanistic pathways other than sedentary time, such as stress, time pressures and economic independence.

Cultural beliefs about body size

Studies in some LMIC indicate that cultural beliefs about body size may be an important barrier to programmes aimed at preventing overweight and obesity. For example, in urban areas of Pakistan, where 64 % of adult study participants were overweight, most did not perceive themselves as overweight⁽ Reference Bhanji, Khuwaja and Siddiqui ^{44 ,} Reference Saleem, Ahmed and Mulla ^{45 )}. Two cross-sectional studies of overweight female university students residing in urban Pakistan reported that up to 18 % believed they were of normal weight⁽ Reference Saleem, Ahmed and Mulla ^{45 ,} Reference Sirang, Bashir and Jalil ^{46 )}. Findings also reveal that a Pakistani adult who is happy or does not think about his/her weight is six times more likely to misperceive him/herself as not being overweight⁽ Reference Bhanji, Khuwaja and Siddiqui ^{44 )}. In some sub-Saharan African countries, such as Nigeria, obesity is perceived as a sign of ‘power, respect, and an evidence of good living’⁽ Reference Akpan and Ekpenyong ^{41 )}. Further research is needed to understand beliefs relating to overweight and obesity, particularly in areas with a high prevalence of HIV/AIDS, as there may be a stigma attached to thinness in these contexts⁽ Reference Kruger, Puoane and Senekal ^{47 )}.

Integration of overweight and obesity into national nutrition plans

Strong nutrition governance, including setting SMART (specific, measureable, achievable, relevant and time-bound) targets, has been linked to achieving undernutrition goals such as stunting^{( 4 ,} Reference Sunguya, Ong and Dhakal ^{51 )}. Few LMIC address overweight and obesity in their national nutrition plans. Kenya is an example of one such country. The Kenyan National Nutrition Action Plan (2012–2017) outlines specific activities to address the increase in overweight and obesity in Kenya, including the following: review, develop and disseminate a comprehensive strategy and guidelines for preventing, managing and controlling nutrition-related non-communicable diseases; train service providers and create public awareness on the importance of preventing, managing and controlling nutrition-related non-communicable diseases; scale up community screening of BMI and waist circumference; and improve nutrition in schools. Kenya’s 2013 National Maternal, Infant, and Young Child Nutrition Policy Guidelines also state that childhood obesity is an emerging public health problem.

Food systems

The food system is a critical underlying determinant of dietary intake, and includes food production, distribution, processing, packaging, marketing and retail. Two aspects of food systems that are particularly important for addressing overweight and obesity in LMIC include: (i) implementing the WHO recommendations on marketing of food and beverages to children; and (ii) policies that promote healthy diets and discourage unhealthy diets.

According to the WHO, introducing policies that reduce the marketing of foods high in saturated fat, trans-fat, free sugars or salt to children is likely to be one of the most cost-effective interventions available to governments to address obesity^{( 49 , 52 )}. Three upper-middle-income countries (Brazil, Thailand and South Africa) have drafted resolutions that address food marketing to children⁽ Reference Hawkes and Lobstein ^{53 )}, but these policies are currently suspended in both Brazil⁽ Reference Silva, Bortolini and Jaime ^{54 )} and South Africa⁽ Reference Igumbor, Sanders and Puoane ^{55 )}. Many LMIC have already adopted policies that regulate food industry marketing of breast-milk substitutes^{( 56 )} and lessons learned from that process may prove to be informative for adopting resolutions to regulate marketing of unhealthy foods to children⁽ Reference Lutter ^{57 )}. Lessons learned from high-income countries that have already adopted marketing restrictions could also be useful to policy makers in LMIC. For example, in the UK, where marketing restrictions have been in place since 2007, an analysis of advertising exposure to high-fat, -salt or -sugar products among children under 16 years of age found almost universal adherence to the restrictions, but that exposure did not change before and after the adoption of the restrictions⁽ Reference Adams, Tyrrell and Adamson ^{58 )}, suggesting that stronger restrictions targeting a wider range of advertisements is needed.

The NOURISHING framework developed by World Cancer Research Fund International is a tool that policy makers in LMIC can use for guidance on developing a comprehensive policy package to promote healthy diets and discourage unhealthy diets⁽ Reference Hawkes, Jewell and Allen ^{59 )}. Few LMIC have adopted policies consistent with the NOURISHING framework. This was made evident at the 2013 Bellagio Conference on Program and Policy Options for Preventing Obesity in Low- and Middle-Income Countries⁽ Reference Popkin, Monteiro and Swinburn ^{9 )}: of the thirteen countries represented at the Conference, only two were lower-middle-income countries (India and Bangladesh) and none were low-income countries. Nevertheless, some LMIC policies addressing the food system within the NOURISHING framework are worth mentioning here.

Several LMIC have policies that promote fruit and vegetable production; many were originally adopted in order to address micronutrient deficiencies, but now play the additional role of overweight and obesity prevention⁽ Reference Lachat, Otchere and Roberfroid ^{60 )}. Several LMIC have also adopted food-based dietary guidelines. For example, the Dietary Guidelines for the Brazilian Population, which, in conjunction with the National Food and Nutrition Security Policy and National Health Promotion Policy, aim to promote healthy eating and the prevention of malnutrition including micronutrient deficiencies and overweight and obesity^{( 61 )}. These Guidelines could be adapted to other countries as part of a comprehensive policy package to address malnutrition, including overweight and obesity.

In 2005, the Ghana Ministry of Health adopted the Regenerative Health and Nutrition Program with the primary objective of promoting healthy lifestyles, including diet and daily physical activity. Since 2006, over 50 000 community members have been trained, who in turn educate other community members. However, analysis of data from before and after implementing the programme shows an overall decline in unhealthy behaviours only among highly educated individuals^{( 62 ,} Reference Tagoe and Dake ^{63 )}. Moreover, a panel in the 2016 Global Nutrition Report suggests that the 2012 National Policy for the Prevention and Control of Chronic Non-Communicable Diseases in Ghana has not been ‘operationalized in any way’^{( 4 )}.

Finally, nutrition labelling has been recommended as a policy tool for addressing unhealthy food environments^{( 49 )}. However, the impact of interpretive nutrition labelling policies on food purchasing patterns has been evaluated only in the UK, where there was ‘no discernible effect on the relative healthiness of consumer purchases’⁽ Reference Sacks, Rayner and Swinburn ^{64 )}.

Education systems

A systematic review including twenty-two studies from LMIC found that school-based interventions have the potential to improve diet and physical activity behaviours and decrease BMI in these contexts⁽ Reference Verstraeten, Roberfroid and Lachat ^{65 )}. Schools offer a unique opportunity to change norms around nutritional practices and healthy body weights. The review also found that involving multiple stakeholders and integrating educational activities into existing school curricula worked best⁽ Reference Verstraeten, Roberfroid and Lachat ^{65 )}.

Two cities in Africa – Ouagadougou, Burkina Faso and Cotonou, Benin – have piloted the WHO Nutrition-Friendly School Initiative, which is a framework for the prevention of the double burden of undernutrition and overweight among schoolchildren. There are five conditions that must be met for a school to be considered nutrition-friendly, beginning with the formation of a School Nutrition Committee that involves key stakeholders. The School Nutrition Committee is responsible for meeting these five conditions and for monitoring and evaluation of the programme: (i) a written school policy on nutrition; (ii) building awareness and capacity of the surrounding community as they relate to nutrition; (iii) school curriculum adaptation to include nutrition; (iv) school environment that supports optimal nutrition and health; and (v) school nutrition and health services^{( 66 )}. In Ouagadougou, sensitizing communities to nutrition was identified as a priority for half of the schools⁽ Reference Delisle, Receveur and Agueh ^{67 )}, and many involved in the initiative felt that these policies should be implemented at the national level rather than the school level⁽ Reference Delisle, Receveur and Agueh ^{67 )}. Piloting the Nutrition-Friendly School Initiative in 2007 in New Delhi, India, an initial assessment found no written nutrition policies in the four participating schools and no integration of nutrition and health into the school curriculum⁽ Reference Parmar ^{68 )}.

A 6-month nutrition education programme focused on nutrition, physical activity, non-communicable diseases and healthy cooking practices was implemented in both private and public schools of three northern Indian cities (New Delhi, Agra and Jaipur)⁽ Reference Shah, Misra and Gupta ^{69 )}. At baseline, only 25–55 % of students considered deep-fried Indian snack foods to be junk foods and 25–35 % of students thought that consuming butter improves bone strength and health⁽ Reference Shah, Misra and Gupta ^{69 )}. Results of the intervention were promising: a pre/post statistical comparison of responses of a random subset of students to a questionnaire on health and nutrition-related knowledge found significant improvements in both private and public schools, particularly on questions relating to physical activity and healthy cooking practices⁽ Reference Shah, Misra and Gupta ^{69 )}.

Government policies to support the provision of healthy foods in schools could build off the Purchase from Africans for Africa programme (Ethiopia, Malawi, Mozambique, Niger and Senegal), which encourages the consumption of non-processed, nutrient-dense foods (cereals, pulses and legumes, fruits, vegetables and animal products)^{( 70 )}. For example, as part of the programme, Malawi smallholder farmers now provide staple grains, groundnuts, bananas and fish to ten primary schools where enrolment has increased by 15 %^{( 71 )}. Prevention of overweight and obesity could be an important co-benefit of such programmes.

Health systems

Currently, programming during routine contact points with women and children at the community and facility levels in LMIC focuses on reducing undernutrition without consideration of overweight and obesity. However, these are unique opportunities to address maternal diet and weight gain during pregnancy, as well as infant and young child feeding practices that are important risk factors for overweight and obesity.

Few studies in LMIC have evaluated the impact of antenatal counselling on appropriate gestational weight gain. In Nigeria, receiving information on gestational weight gain from multiple health-care providers was significantly associated with increased levels of maternal knowledge on risks of inappropriate weight gain⁽ Reference Olagbuji, Olofinbiyi and Akintayo ^{13 )}. However, although 57 % of Nigerian mothers were counselled on the risks associated with excess weight gain, many mothers were not aware of the recommended amount of weight gain during pregnancy⁽ Reference Olagbuji, Olofinbiyi and Akintayo ^{13 )}. In Egypt, women were also not counselled on weight gain during pregnancy and had little understanding of the appropriate amount of weight gain⁽ Reference Kavle, Mehanna and Khan ^{14 )}. A small study of women attending a clinic in Accra, Ghana, found that they were willing to reduce their body size in order to reduce the risk of obesity-related illnesses such as hypertension, diabetes, CVD and myocardial infarction⁽ Reference Duda, Jumah and Hill ^{72 )}. Education, income, marital status and number of children did not influence their response, although women who were younger than 50 years of age tended to be more likely to report being willing to reduce their body size to improve health⁽ Reference Duda, Jumah and Hill ^{72 )}. Thus, providing counselling on maternal diet and weight gain during pregnancy during antenatal care can potentially address overweight and obesity in LMIC.

Another important point of entry is existing health promotion-based interventions such as complementary feeding interventions for infants and young children. Health-care providers and families need to understand healthy v. unhealthy weight gain and can learn to monitor excessive and/or rapid weight gain within the context of undernutrition programming. In addition, families should be advised that junk foods are detrimental to the growth of children and the entire family’s health, and providers should be trained to provide nutrition counselling for families during routine growth monitoring visits^{( 73 )}. A recent study in Egypt used Trials for Improved Practices (TIPs) methodology to address poor infant and young child feeding practices over the course of three home visits (Fig. 2)⁽ Reference Kavle, Mehanna and Saleh ^{42 )}. During the first visit, researchers discussed behaviours, practices and motivations for infant and young child feeding with mothers, focusing on junk foods. During the second visit, mothers jointly decided with the research team what specific feeding practices they would be willing to try over a one-week period. During the third and final visit, mothers discussed their experiences with TIPs: which recommended practices worked, which did not work and any modifications that were made. Results supported the importance of reducing junk food consumption keeping cultural influences and beliefs in mind when designing interventions.

Fig. 2 Example of how Trials for Improved Practices (TIPs) addressed junk food as an infant feeding problem in Egypt