Temporal trends in the nutritional status of women and children under five years of age in sub-Saharan African countries: ecological study

ABSTRACT BACKGROUND: While the global prevalence of obesity is rapidly increasing, this pandemic has received less attention in sub-Saharan Africa, particularly in the light of the persistent undernutrition that exists in the context of maternal and child health. We aimed to describe obesity trends among women of childbearing age over recent decades, along with trends in over and undernutrition among children under five years of age, in sub-Saharan African countries. DESIGN AND SETTING: Ecological study with temporal trend analysis in 13 sub-Saharan African countries. METHODS: This was a description of temporal trends in nutritional status: adult obesity, childhood overweight, low height-for-age (stunting), low weight-for-height (wasting), low weight-for-age (underweight) and low birth weight. Publicly available data from repeated cross-sectional national surveys (demographic and health surveys and multiple-indicator cluster surveys) were used. We chose 13 sub-Saharan African countries from which at least four surveys conducted since 1993 were available. We investigated women aged 15-49 years and children under five years of age. RESULTS: In multilevel linear models, the prevalence of obesity increased by an estimated 6 percentage points over 20 years among women of childbearing age, while the prevalence of overweight among children under 5 years old was stable. A major decrease in stunting and, to a lesser extent, wasting accompanied these findings. CONCLUSIONS: The upward trend in obesity among women of childbearing age in the context of highly prevalent childhood undernutrition suggests that the focus of maternal and child health in sub-Saharan Africa needs to be expanded to consider not only nutritional deficiencies but also nutritional excess.


INTRODUCTION
Obesity, a complex condition that affects all ages and socioeconomic groups, has become one of the world's most challenging public health problems. Because it has been raising the prevalence of noncommunicable diseases in low and middle-income countries that are still burdened with infections and nutritional deficiencies, it has helped to create a "double burden" of disease that threatens to overwhelm healthcare services. 1-3 Its prevalence is increasing rapidly worldwide, and this trend is believed to be related to dietary excess, physical inactivity 4 and increasing urbanization. 5 Childhood overweight has also been increasing at an alarming rate. This is defined in accordance with the World Health Organization child growth standards as a weight-for-height z-score ≥ 2 standard deviations above the median of the reference population. 6 Its worldwide prevalence increased from 4.2% (95% confidence interval, CI, 3.2%-5.2%) in 1990 to 7.8% (95% CI, 6.4%-9.1%) in 2015. This trend is expected to reach 9.1% (95% CI, 7.3%-10.9%) in 2020. 7 Currently, at least 41 million children under five years of age are obese or overweight, and the greatest rise is being seen in low and middle-income countries. 7 In Africa, the number of overweight children rose from 5 million in 1990 to 10 million in 2014. 8 While tackling obesity is a current public health priority in most of the world, in sub-Saharan Africa the issue has received less attention, especially within the context of maternal and child health. In this region, the main nutritional concern continues to be undernutrition, particularly among children.
Most data available on the nutritional situation in sub-Saharan African countries are derived from standardized national surveys. However, to our knowledge, no studies focusing on sub-Saharan Africa have investigated trends using nationally representative data from several of these countries to characterize trends in excess weight among women of childbearing age and children under five years of age.
The aim of the present study was thus to describe obesity trends among women of childbearing age and overweight trends among children under five, over the past two decades, within the context of the continuing general picture of undernutrition among young children living in sub-Saharan Africa.

METHODS
To determine temporal trends regarding the prevalence of over and undernutrition in sub-Saharan African countries, we analyzed secondary data from demographic and health surveys (DHS). 9 These are nationally representative surveys of sizes ranging from 5,000 to 30,000 households that were funded by the United States Agency for International Development (USAID).
They were usually conducted every five years to collect data on several topics from selected countries.
Standard data collection procedures and manuals had been used to guide the household survey process, and the data had been processed and presented in reports that described the situation in each country. We also used secondary data from multiple indicator cluster surveys (MICS), 10 which are nationally representative cross-sectional household surveys that were funded by the United Nations Children's Fund (UNICEF). These were conducted on an average of 11,000 households and provided information about maternal and child health.
For the present study, the sub-Saharan African countries from which at least four surveys providing information on the prevalence of obesity in women and/or anthropometric data on children under five years of age had been conducted since the early 1990s were eligible for inclusion. Thirteen countries were thus eligible, from which a total of 60 MICS or DHS conducted between 1993 and 2014 were available.
In the DHS, objective height and weight measurements had been made on women aged 15-49 years and on children under five years of age. Height had been measured using portable stadiometers. For children up to the age of two years, height had been measured in a lying position. Weight had been measured using portable digital scales. 11 In MICS, similar direct measurements had been made on children, but maternal height and weight had not been obtained. 10 Thus, the data on obesity among women of childbearing age for this study came only from the DHS.
Women were classified as obese if their body mass index was ≥ 30 kg/m 2 . In separate analyses, children under five years of age were classified with regard to wasting (low weight-for-height), stunting (low height-for-age), underweight (low weight-for-age) and overweight (high weight-for-height). These were based on z-scores for height and weight that were calculated in accordance with the World Health Organization child growth standards. 6 Wasting, stunting and underweight were defined as two or more standard deviations below the median and overweight as two or more standard deviations above the median of the reference population. 12 Birth weight had been obtained through interviews with parents/guardians for children under two years of age. Low birth weight was defined as < 2500 g, 13 which is a birth weight below the third percentile, according to the World Health Organization child growth standards.
Prevalence rates, with their respective confidence intervals, were estimated for each country at each point in time, using weighting to account for aspects of the survey design, including cluster effects, 14  Overall temporal trends regarding the prevalence of the different outcomes in the 13 countries were analyzed using a two-level hierarchical linear model. In this model, the country was the contextual variable, and both the intercept and the slope coefficient were treated as random variables. This analysis was also performed using Stata, version 13.0. Given the diversity of settings both among the countries included and among those not included in this study, we felt that treating these countries as random representatives of the sub-Saharan region was the best approach. Thus, this analysis was performed without additional weighting for country size.
All analyses were based on publicly available data from national surveys. Ethical clearance was the responsibility of the institutions that administered those surveys.

RESULTS
The trends regarding the prevalence of obesity among women aged 15-49 years are shown in Table 1 and Figure 1

, Panel A.
A strong upward trend in obesity can be seen among women of childbearing age in most countries, especially in Cameroon, Ghana and Kenya, with increases of 0.55, 0.48 and 0.35 percentage points per year (pp/yr), respectively (P < 0.001). In recent  surveys, the prevalence of obesity among these women ranged from 3% in Burkina Faso (in 2010) to 15% in Ghana (in 2014).
Trends regarding the prevalence of overweight among children under five years of age are also shown in Table 1

DISCUSSION
In the present study, we describe a significant increase in the  17 The increase in prevalence reported for Africa was close to that found in the present study (5.6%), in which we considered both rural and urban areas together.
Although Jones-Smith et al. 18 observed that the prevalence of being overweight had a positive relationship with wealth and education, they noted that the rate of weight gain over time was frequently greater in groups of lower socioeconomic status. Razak et al. 19 reported that, while populations as a whole are gaining weight, the pattern of these gains is not uniform across the range of body mass indexes. The main increases were found to be concentrated in the overweight and obese parts of the spectrum, while there was frequently little or no change in the underweight and normal-weight parts of the nutritional status distribution.
Focusing on children and adolescents (aged < 20) in developing countries, Ng et al. 20  Moreover, the data suggest that high rates of overweight and obesity among sub-Saharan African adolescents and adults within a few decades are quite likely, even among those whose childhood was marked by wasting and stunting. Additionally, these high rates will most likely soon be accompanied by high rates of obesity-related complications, especially diabetes.
Additionally, studies have suggested that gestational diabetes, which frequently accompanies obesity during pregnancy, also increases these risks. 24 Pre-pregnancy obesity and gestational diabetes are major risk factors for large-for-gestational-age births, 25,26 and the risk of adult obesity is uniformly greater among those born with excess weight. 27 This is notably so in the context of a rapid nutritional transition. 28 Within the conceptual framework of the developmental origins of health and disease, several noncommunicable diseases originate in the fetal period and during early life. 27,29,30 These facts highlight how important it is for sub-Saharan   women to maintain a healthy weight before and during pregnancy.
They also highlight the importance of avoidance of later nutritional excess among sub-Saharans who previously suffered from undernutrition during fetal life and early infancy.
In sum, these findings suggest that the time has come for public health officials in sub-Saharan African countries to implement policies for controlling obesity that include attention to women of childbearing age and young children. In global terms, overweight and obesity have already replaced undernutrition and infectious diseases as the major cause of health-related problems in the 21 st century. 31,32 There is little reason to believe that sub-Saharan a key necessity. 30,33 Policies to combat economic inequality and poverty, which are striking features of present-day Africa; to combat obesogenic social changes, such as increasing advertising and availability of unhealthy foods and snacks to children; to encourage physical activity; to promote family farming; to support gender equality; and to improve access to health care and education 34 will also be beneficial in this effort. The complex challenge for many transitioning African countries will be to simultaneously address childhood undernutrition, on the one hand, and excess weight, on the other. 34 Some limitations of our study merit comment. The group of 13 countries analyzed is not fully representative of sub-Saharan Africa, since the countries were selected based on survey availability. Therefore, caution is needed when generalizing the results.
However, it is noteworthy that many of the sub-Saharan African countries with the highest obesity rates (those located in the south) were not included in our sample because they lacked a minimum number of surveys. Another limitation was our use of the body mass index to estimate cutoffs for overweight, given the widespread presence of stunting, because the progression to future overweight and obesity among children currently presenting stunting (a situation involving a large proportion of the children under five years of age in these countries) is less well understood. Nevertheless, we can highlight that one strength of our study was our use of available, high-quality data to summarize the current picture, thereby providing useful information for the public healthcare services of the countries involved.