Overweight and Obesity, Cardiometabolic Health, and Body Composition: Findings From the Follow-Up Studies of the INCAP Longitudinal Study

Abstract

Background:

There has been increased interest in the hypothesis that undernutrition in early life predisposes to cardiometabolic disease risk in adulthood. The Institute of Nutrition of Central America and Panama Longitudinal Study is able to address one critical aspect of this field, specifically whether improvements in nutrition can prevent this increased risk.

Objective:

To describe the main findings on obesity and body composition across 5 waves of field work (1988–1989, 1991–1994, 1998–1999, 2002–2004, and 2015–2017) and on cardiometabolic health across 3 waves (1998–1999, 2002–2004, and 2015–2017).

Results:

Body weight and body fat increased considerably in adulthood, especially among women with sedentary occupations. Adiposity and weight in adulthood were strongly predicted by weight gain after the first 1000 days of life. On the other hand, exposure to improved nutrition in early life reduced diabetes risk by approximately 50% but increased the risk of overweight and obesity.

Conclusions:

Future research will aid in clarifying the underlying mechanisms that drive the opposite associations among diabetes and obesity with early-life nutrition.

Introduction

Guatemala is undergoing rapid nutrition and epidemiologic transitions. The study cohort was born into an impoverished setting. Over the decades, there have been major economic changes—rural electrification, improved road infrastructure, housing, water supply and sanitation, telecommunications, employment opportunities, as well as civil war, migration, and recently, gang-related violence. All of these affect people’s well-being. The cohort provides a window to explore how these factors affect physical well-being and health and the extent to which the initial intervention might modify the impact of these factors.

In this article, we reviewed all peer-reviewed published articles available from the Institute of Nutrition of Central America and Panama (INCAP) Longitudinal Study related to obesity, body composition, and cardiometabolic disease outcomes in adulthood, as well as their relation with early life improved nutrition. The objective of this article is to summarize the contributions of the INCAP nutrition supplementation trial on the aforementioned research areas.

Results

Antropometry

As described elsewhere in this volume, the initial premise of the study was that the intervention would contribute to improved cognition, with improvements in linear growth being an indicator of intervention efficacy, as it was assumed that improvements in nutrition would improve growth. Indeed, this was observed, with a net increment of 1.6 cm when lengths of children at age 3 year were compared between the atole and fresco villages.¹ Interestingly, the most recent exploration suggested that this difference did not persist into adulthood.²

Excess Weight

The cohort has become overweight, as they have moved into adulthood. The primary measure of overweight is the body mass index (BMI, body weight in kg divided by height squared in m²). In adults, a BMI of 25 kg/m² or above is considered overweight, and a BMI of 30 kg/m² or above is considered to be obese. In the 2015 to 2017 wave of follow-up, mean BMI was 28.7 kg/m² for women and 26.3 kg/m² for men, with 40% of women and 19% of men having a BMI of 30 kg/m² or more.²

Differences between people in the development of overweight and obesity were apparent already in childhood. Ford et al showed that BMI tracked over the life course.³ Using latent class growth analysis, it was observed that there were 3 classes in men and 2 classes in women. The classes differed in their BMI in early childhood, in the rate at which BMI increased over the life course, and in the median age at which individuals became overweight or obese. The median ages at onset of overweight for the higher and lower trajectory classes were 29 years and 23 years for the women and 34 years and 29 years for the men, respectively. Class membership could be predicted statistically by socioeconomic status of the household in childhood, with children born to households of higher socioeconomic status being more likely to be of higher BMI, and there was a suggestion, although it was not statistically significant, that exposure to atole in early life was associated with being in the class in which BMI increased the fastest.

The association of early-life exposure to atole with obesity became more pronounced using the 2015 to 2017 data wave. Specifically, using a formal difference-in-difference methodology, Ford et al showed that BMI was 1.29 kg/m² higher (95% confidence interval [95% CI] 0.08–2.50) and the odds of obesity were approximately doubled (odds ratio [OR] 1.94; 95% CI 1.11–3.40) among those who were exposed to atole from conception through their second birthday (a period known as the first thousand days) as compared to individuals exposed to atole at other periods or exposed to fresco.² That analysis did not explore the pathways through which improved nutrition acted. These are likely to reflect differences in patterns of dietary and physical activity behaviors that resulted from the improved life circumstances brought about by the early-life intervention and described elsewhere in the volume.

Body Composition

Body composition refers to the in vivo quantification of the components in which the body can be divided and their distribution in it.⁴ The 2-compartment model, specifically fat mass (FM) and fat-free mass (FFM), is the most common. Fat-free mass can in turn be subdivided into multiple components (muscle, bone, etc). The proportion of total body fat and its distribution at the abdominal level are better indicators of cardiometabolic risk than is BMI alone, because the latter cannot identify which body component (fat vs muscle) is physiologically relevant.

Studies have shown that poor nutrition during pregnancy and the first years of life can cause excess body fat in adulthood.⁵ The intrauterine period is part of what is known as the “window of opportunity” or critical period, based on the concept that exposure to certain environments in this period will produce permanent changes in the organism.⁶ Birth weight is a widely used indicator of intrauterine or prenatal development.⁷ Studies of individuals exposed to famines in critical periods have shown a greater accumulation of abdominal fat in adult life in those who had low birth weight.^7,8 However, evidence to date on this relationship is mixed and inconclusive. Several longitudinal studies with large sample sizes have found no association between low birth weight and adiposity in adulthood.^9,10 Some degree of caution should be taken, since there is great variability between methods used to measure the distribution of body fat.^5,11,12

Methods to Determine Body Composition in the INCAP Longitudinal Study

To facilitate the estimation of body fat in the study cohort, it was necessary to develop simple and low-cost methods. Body fat prediction equations were generated based on anthropometric measures specific for sex and age. In the 1988 to 1989 and 1998 to 1999 follow-up waves, the equations developed by Conlisk et al, which were validated in a population of adolescents and young adults living in Guatemala City, were used.¹³ In the 2002 to 2004 follow-up, prediction equations developed by Ramirez-Zea et al, among a sample of rural and urban young adults, were used.¹⁴ Both studies used a 2-compartment model, determining body density by the underwater weight method, subsequently deriving FFM and FM using constant factors published in the literature.^15,16

Thanks to the recent availability at INCAP of isotope dilution methods for total body water measurement, which can be done in the field, it was possible to make direct measurement of body composition in the 2015 to 2017 follow-up.^17,18 In brief, after providing a saliva sample to establish a baseline level, the participant is asked to drink a known volume of water enriched with deuterium. Three hours later, a saliva sample is obtained, and body deuterium enrichment is computed from the relative enrichment. Using the hydration constant for adults (73.2%), FFM is derived, and by difference, FM.¹⁸ Table 1 describes methods and variables derived in each follow-up.

Table 1.

Methods to Determine Abdominal Obesity, Fat Free Mass, and Body Fat Across Waves in the INCAP Longitudinal Study.

Wage Year	Method	Variables
1988–1989	Sex-and-age-specific equations derived from Conlisk et al, 1992 (reference method: hydrostatic weighing)	% Body fat
		Waist-to-hip ratio
		Body mass index
1991–1994 (women)	Sex-and-age-specific equations derived from Conlisk et al, 1992 (reference method: hydrostatic weighing)	% Body fat
		Waist-to-hip ratioa
		Body mass index
1998–1999	Sex-and-age-specific equations derived from Conlisk et al 1992 (reference method: hydrostatic weighing)	% Body fat
		Fat free mass
		Waist-to-hip ratio
		Body mass index
2002–2004	Sex-and-age-specific equations derived from Ramirez-Zea et al, 2005 (reference method: hydrostatic weighing)	% Body fat
		Waist circumference
		Fat mass
		Fat free mass
2015–2017	Deuterium dilution technique	Fat free mass
		% Body fat
		Waist circumference
		Waist-to-height ratiob
		Body mass index

^aWaist-to-hip ratio = waist circumference/hip circumference.

^bWaist-to-height ratio = waist circumference/height.

Relationship of Growth in Early Childhood With Body Composition in Adulthood

The association between growth during pregnancy and the first years of life with body composition in adulthood was examined when cohort participants were young adults (1988–1994, 1998–1999, and 2000–2004) and middle-aged adults (2015–2017). Evidence in the scientific literature about poor prenatal and postnatal growth as risk factors for increased adiposity in adulthood is inconsistent, as are some of the results that have emerged from the INCAP Longitudinal Study.

Findings of the first follow-up study (1988–1994), when cohort participants were still young adults (17–28 years of age), were consistent with the hypothesis that early growth failure predispose to overweight. The strongest predictor of increased abdominal obesity (waist/hip ratio) in adulthood was low birth weight, indicating that poor prenatal growth was more important than postnatal growth.¹⁹ Migration to urban areas was associated with an increase in waist/hip ratio, more so in in women with severe short stature. The observed inverse association between postnatal growth (height for age) and abdominal adiposity disappeared when adjusted for birth weight, suggesting that it was mostly due to poor intrauterine growth. Other findings from that wave of field work were that men and women who were stunted at 3 years of age were shorter and thinner in adulthood. In addition, stunted men had lower body fat in adulthood, a relationship that was not found in women. Another interesting finding was that, compared to US reference values, stunted participants had a higher proportion of central fat, with the subscapular/tricipital skinfolds ratio between 50 and 90 percentiles of the reference medians from childhood to adolescence.²⁰

Results in later waves of follow-up did not confirm the findings of this first study. In the Cardiovascular Disease (CVD) follow-up (1998–1999), when participants were between 21 and 27 years, it was found that as birth weight increased, the waist/hip ratio also increased.²¹ In both sexes, birth weight (an indicator of prenatal growth) and height at 2 years of age (an indicator pre- and postnatal growth) were positively associated with height, weight, and FFM in adulthood. Both prenatal and postnatal growth were positively associated with FM and % body fat in women, whereas in men this association was evident only with postnatal growth. Men who were below the sample mean in their prenatal and postnatal growth were 9 cm shorter, 6.4 kg lighter, and had 6.5 kg less FFM than those with growth above the mean. The corresponding values for women were 5.0 cm for height, 11.8 kg for weight, and 4.9 kg for FFM. These differences may have important functional implications.

In the area of body composition, the main finding of the Human Capital Study follow-up (2002–2004), when participants were between 26 and 41 years of age, was that linear growth in the first 3 years of life was associated with higher FFM in adulthood and weakly associated with FM and abdominal circumference. On the other hand, an increase in BMI between 3 and 7 years of age was strongly associated with higher FM and abdominal fat.²² This result highlights the importance of monitoring changes in BMI in the preschool age to prevent excessive adiposity in adult life. The data from this follow-up were used in a joint analysis with 4 other low- to middle-income country cohorts (Brazil, India, Philippines, and South Africa) and consistently demonstrated that faster linear growth and relative weight gain after early childhood were associated with greater overweight in adulthood.²³

Body Composition, Physical Performance, and Residence

The first study on the association between body composition and physical performance was made during the 1988 and 1989 follow-up.²⁴ At the time of measurement in adolescence, participants who were exposed to atole in the first 3 years of life were taller and heavier than those exposed to fresco.²⁵ Men exposed to atole throughout pregnancy and the first 3 years of life had better physical performance, as measured by oxygen uptake (VO_2max), than those exposed partially.²⁴ These results showed that improving early nutrition has a positive effect on physical performance in adulthood. In women, no difference was found in physical performance between these 2 groups, most likely because physical activity level was generally very low.

In the 2002 to 2004 follow-up, it was observed that men still residing in the original study villages were lighter, had better physical fitness, and were more physically active compared to those living in Guatemala City.²⁶ In addition, men who lived in Guatemala City had higher % body fat compared to those who lived in the villages (23% vs 20%) and greater abdominal circumference (90.4 vs 85.6 cm). Interestingly, in women, no difference was found in all these variables. These results could be explained by changes in occupation, since the only group who maintained a better physical fitness and body composition profile were those who still engaged in physically demanding activities, specifically rural men engaged in agriculture.²⁷ Moreover, occupation appears more important than rurality in the increase of adiposity and associated risk factors.²⁸ These changes have occurred rapidly, with cohort participants experiencing a rapid increase in total body weight (more than 5 kg of weight gain) and %fat (4.2% in men and 3.2% in women) in the 5 years between the follow-ups of 1997 and 1998 and 2002 to 2004.²⁹

These results are most likely associated with significant lifestyle changes, such as consumption of unhealthy diets and sedentary occupations. Migration to the city results an improvement in income but also a change in dietary patterns, since migrants reported higher consumption of high sugar and fat foods.³⁰

Through the different follow-ups (Table 2), there has been a sustained increase in body weight and accumulation of total and abdominal fat in this population, especially among women. Unhealthy dietary patterns and low physical activity, in an obesogenic environment, might largely explain the accelerated increase in total and central adiposity.³¹

Table 2.

Descriptive Body Composition Measures^a Across Different Follow-Ups.^b

Follow-Up	1988–1989 1991–1994 (women) Schroeder et al, 199919		1998–1999 Li et al, 200321		2002–2004 Ramirez-Zea, 200614		2002–2004 Corvalan et al, 200730,c	2015–2017 Ford et al, 20182
Gender	Men	Women	Men	Women	Men	Women	Overall	Men	Women
Sample size, n	161	372	136	131	386	470	382	435	662
Age, y	18–24	17–28	24.2 ± 1.6	24.3 + 1.5			32.9 ± 4.2	45 ± 4.0	45 ± 3.0
Fat free mass—FFM, kg			51.0 ± 5.1	37.6 ± 4.1	51.3 ± 5.2	38.8 ± 3.5	44.9 ± 7.7
Fat mass—FM, kg			8.7 ± 4.7	17.1 + 6.8			18.4 ± 9.0
Body fat, %	12.8	26.2	14.0 ± 5.9	30.3 + 6.5	19.5 ± 6.4	34.6 ± 7.0	28.1	29.2 ± 4.1	42.4 ± 3.7
Waist circumference—WC, cm					85.4 ± 8.8	91.6 ± 11.7	89.7 ±11.1	93.5 ± 13.3	100.2 ± 15
Waist-to-hip ratio—WHR	0.89–0.90	0.89–0.91	0.90 ± 0.04	0.80 ± 0.1
Waist-to-height ratio—WHtR								0.57 ± 0.4	0.66 ± 0.6
Body mass index—BMI, kg/m2	20.3–21.3	21.8–22.2	22.2	23.8	24.2 ± 3.4	26.6 ± 4.6	25.8 ± 4.4	26.3 ± 2.7	28.7 ± 3.5

Abbreviation: SD, standard deviation.

^aDescriptive statistics are intervals or mean ± SD as reported in each manuscript.

^bCorvalan et al, 2007³⁰: data correspond to all participants for whom it is available cross-sectional data at 0, 1, 2, 5, 7 years and adulthood during the specific follow-up.

^cWaist-to-hip ratio = waist circumference/hip circumference; waist-to-height ratio = waist circumference /height.

Improved Nutrition in the First 1000 Days and Body Composition in Adulthood

Results from the 1988 and 1989 follow-up suggested that participants in the atole villages who were exposed to better nutrition in pregnancy and the first 3 years of life were taller, heavier, and had more FFM than those in the fresco villages when they were between 11 and 27 years of age.²⁵ In men, the largest FFM was fully explained by its larger size, but not in women, which involves a direct effect of early nutrition in FFM, independent of height.

In the 2015 to 2017 follow-up (known internally as project META), Ford and colleagues observed that individuals exposed to atole during the first 1000 days (from conception to 2 years) had 1.73% more body fat, an increase of 1.29 kg/m² in their BMI, were almost twice as likely to be obese, and had greater central adiposity (waist/height ratio) in adulthood, compared to those who received atole at other periods or received fresco.² Although there is no clear explanation for this phenomenon, the increased education and income associated with improving early nutrition may have facilitated the adoption of obesogenic lifestyles.

These results suggest a challenge for food and nutrition programs. Our underlying hypothesis is that the effects on overweight and obesity reflect lifestyle changes consequent to increased economic opportunities in an obesogenic environment. Interventions in early childhood that have a number of beneficial effects need to be supported by efforts to improve the nutrition and physical activity environment for older children and adults.³²

Cardiometabolic Health

Cardiometabolic diseases (cardiovascular disease, stroke, diabetes) are, collectively, the leading cause of death globally. Since the early work of Barker,³³ there has been interest in the role of early-life undernutrition in the programming of later disease. Much of this literature has focused on the circumstances of deprivation, whether measured as exposure to famine or other shocks, as stunting or as other measures of poor nutritional status. Few, if any, studies have effectively tested the potential for improvements in nutrition in early life to mitigate in adulthood the consequences of prenatal or early-childhood programming.

Examination of this potential in the INCAP study started with the CVD follow-up, with field work conducted in 1997. In this wave, cohort members who were born in the period 1970 to 1976, and hence were all exposed to the intervention from conception through at least their first birthday, were included. A basic clinical history was obtained, as were anthropometry, a fasting capillary blood sample, and blood pressure. Analyses were limited by the lack of within-village controls to address the challenge that there were only 2 pairs of villages in the original trial and the young age at study.³⁴ Nevertheless, the data were useful to show that migrants to Guatemala City had lower levels of physical activity than rural cohort members and were more likely to be overweight or obese.³⁵

The second wave of follow-up which was used to explore cardiometabolic risk was the Human Capital Study wave of 2002 to 2004. In this wave, which included all available original cohort members regardless of birth year, a physical examination was done, including measurement of blood pressure and anthropometry. A fasting capillary blood sample was obtained and was used to measure lipids and glucose. As described elsewhere in this volume, this was the first study wave to explicitly include a double-difference design in the analytic plan. Although still limited by the young age at examination and by the lack of a complete blood profile, the data showed that those exposed to atole had a reduced glucose level (7.0 mg/dL, 95% CI: 0.5–13.5) for exposure at ages 36 to 72 months; lower systolic blood pressure (3.0 mm Hg, 95% CI: 0.4–5.6) for exposure at ages 24 to 60 months; and a lower triglyceride level (sex-adjusted; 22.2 mg/dL, 95% CI: 0.4–44.1) and higher high-density lipoprotein cholesterol level (males only; 4.7 mg/dL, 95% CI: 1.5–7.9) for exposure prior to age 36 months; compared with those not exposed in early life.³⁶

The META follow-up, 2015 to 2017, was funded to test explicitly the hypothesis that improved nutrition in early life would lower the risk of cardiometabolic disease. Field work included a fasting venous blood draw, a meal challenge using a standardized drink made from Incaparina, sugar, skim milk, and safflower oil and followed by a repeat venous blood draw and a full clinical examination and medical history. The blood samples were tested in Guatemala, with aliquots frozen and shipped to Emory for further analysis and long-term storage. The primary result from this project was the finding that the prevalence at examination of diabetes was approximately 50% lower (OR 0.45, 95% CI 0.21–0.95) among individuals exposed to atole in their “first thousand days” as compared to those exposed to atole at other ages or to fresco.² This association was robust to adjustment for covariates, including perhaps most interestingly, concurrent BMI, and the waist to height ratio, suggesting that the mechanism through which early life exposure to atole acted on diabetes was not through obesity. Ongoing work is further exploring these pathways. There was no significant association of exposure to atole in the first thousand days with prevalent hypertension or the metabolic syndrome.

Conclusion

The INCAP nutrition supplementation trial and its longitudinal follow-up waves have played a key role in understanding the relationship between birth weight, weight gain, body composition, and cardiometabolic health. Physical development and optimal nutrition during pregnancy and the first years of life have an important effect on the body composition of adults, especially in the development of FFM. Postnatal weight gain, but not birth weight, was associated with overweight and obesity in adulthood and adiposity increased more rapidly in women due to their sedentary behaviors.

All of these relationships are occurring in the context of a transitioning society in which food availability and physical activity patterns are rapidly changing. This is especially evident in the high prevalence of overweight and obesity among the cohort.

Results of the study have important implications for programs and Guatemala and other low- or middle-income countries. As detailed elsewhere in this volume, there have been multiple beneficial effects of the intervention, with respect to schooling in women, cognition, and income. We have described the beneficial effect on diabetes risk. The impacts on obesity are a concern, but probably reflect the consequences of choices that the cohort participants make as a result of their increased economic opportunities. Thus, our data suggest a need for combining early-life interventions with effective strategies to improve nutrition and maintain levels of physical activity to promote cardiometabolic health throughout the life course.

The INCAP study cohort has provided, and will provide in the future, unique data on the role of early-life nutrition on body composition and cardiometabolic health. The cohort is now reaching the age at which the stressors of poverty and physical labor interact with obesogenic environments of a society in economic and epidemiologic transition. The divergent results for diabetes on the one hand and overweight/obesity on the other open up interesting avenues for further scientific exploration. What exactly is the mechanism through which improved nutrition reduces the risk of diabetes? Why is it specific to diabetes? Is the increased prevalence of obesity a function of the improved economic prospects described elsewhere? Over the coming years, we will be pursuing these and other questions.