Abstract
Background: Latin America has experienced increases in obesity. Little is known about the role of early life factors on body mass index (BMI) gain over the life course.
Objective: The objective of this research was to examine the role of early life factors [specifically, nutrition supplementation during the first 1000 d (from conception to 2 y of age) and childhood household socioeconomic status (SES)] on the pattern of BMI gain from birth or early childhood through midadulthood by using latent class growth analysis.
Methods: Study participants (711 women, 742 men) who were born in 4 villages in Guatemala (1962–1977) were followed prospectively since participating in a randomized nutrition supplementation trial as children. Sex-specific BMI latent class trajectories were derived from 22 possible measures of height and weight from 1969 to 2004. To characterize early life determinants of BMI latent class membership, we used logistic regression modeling and estimated the difference-in-difference (DD) effect of nutrition supplementation during the first 1000 d.
Results: We identified 2 BMI latent classes in women [low (57%) and high (43%)] and 3 classes in men [low (38%), medium (47%), and high (15%)]. Nutrition supplementation during the first 1000 d after conception was not associated with BMI latent class membership (DD test: P > 0.15 for men and women), whereas higher SES was associated with increased odds of high BMI latent class membership in both men (OR: 1.98; 95% CI: 1.09, 3.61) and women (OR: 1.62; 95% CI: 1.07, 2.45) for the highest relative to the lowest tertile.
Conclusions: In a cohort of Guatemalan men and women, nutrition supplementation provided during the first 1000 d was not significantly associated with higher BMI trajectory. Higher childhood household SES was associated with increased odds of high BMI latent class membership relative to the poorest households. The pathways through which this operates still need to be explored.
Keywords: child nutrition, developmental origins, Guatemala, latent class growth analysis, obesity, socioeconomic status, 1000 days
Introduction
In the past 2 decades, Latin America has experienced decreases in underweight and increases in overweight (1–3). The age-standardized prevalence of adult overweight (20–70 y) in Central America increased from 45% in 1994 to 59% in 2000 (4). In high-income countries, rapid weight gain during the first 2 y and high protein intake increase the risk of childhood overweight and adiposity (5–7). Conversely, the period of the first 1000 d after conception is a critical window to prevent stunting and to achieve optimal growth for health and cognitive development (8, 9). Despite this potential conflict, there is a paucity of data to assess the long-term effects of child nutrition interventions on BMI gain in low- and middle-income countries (LMICs)8.
Furthermore, the role of socioeconomic status (SES) in the development of obesity in developing countries is debated. Overweight in LMICs has traditionally been considered a disease of affluence; however, there is increasing evidence that the burden of obesity shifts to lower SES groups as countries undergo economic development (10, 11). Existing studies of SES and obesity in LMICs have focused on adults (12, 13); however, SES likely influences BMI gain early in the life course.
With the use of experimental data from a longitudinal cohort study in Guatemalan men and women, we estimated the association of an energy- and protein-containing nutrition supplement (atole) during the first 1000 d and of early-childhood household SES on the timing and rate of BMI gain from infancy through midadulthood by using latent class growth analysis (LCGA). Understanding heterogeneity in BMI gain over the life course and its early life predictors could help inform nutrition programming in countries facing the dual burden of stunting and increasing obesity.
Methods
Study population
Study participants were born in 4 villages in southeastern Guatemala from 1962 to 1977 and participated in the Institute of Nutrition of Central America and Panama (INCAP) Oriente Longitudinal Study (1969–1977) and its follow-up studies (1989–2004) (14). The original trial was designed to assess the influence of early life undernutrition on cognitive and physical development. The follow-up studies were implemented to assess longer-term effects of the original intervention. In 1969–1977, 2 sets of matched villages were randomly assigned to receive atole, a dietary supplement made from Incaparina (a vegetable protein mixture developed by INCAP), dry skim milk, and sugar (6.4 g protein, 0.4 g fat, and 90 kcal energy/100 mL), or fresco, a low-energy beverage with no protein or fat (33 kcal/100 mL; all calories from sugar). Both supplements were fortified with micronutrients in equal quantities by volume. The supplement was available in a central location in each village twice daily. Children could be exposed prenatally through maternal supplement intake and/or postnatally through breast milk or the child’s own consumption through age 7 y. Children were followed up through the study’s end or age 7, whichever came first. Three follow-up waves (1988–1989, 1998, and 2002–2004) provided anthropometric measures in adulthood.
Because a minimum of 3 BMI values are needed for model stability in LCGA (15), individuals were excluded if they had <2 BMI values in childhood (0–84 mo) and no nonpregnant adult BMI values (10–42 y). Of the 2392 individuals who participated in the original trial, 9% (205) were excluded for having no height and/or weight measurements, 20% (483) were excluded for having <2 childhood BMI values, and 11% (251) were excluded for having no nonpregnant adult BMI values (Figure 1). The final analyses included 1443 (60%) participants (742 men and 711 women).
FIGURE 1.
Participant retention across follow-up periods for the INCAP Nutrition Supplementation Trial longitudinal cohort. INCAP, Institute of Nutrition of Central America and Panama.
Data collection and variable specification
BMI.
From 1969 to 1977, child weight and length were measured by trained personnel with the use of standard procedures (16). For subsequent study waves, trained fieldworkers collected weight and height. Height was measured to the nearest 0.1 cm and weight to the nearest 10 g. All measurements were taken in duplicate; if the difference exceeded 500 g for weight or 1 cm for height, a third measurement was taken and the average of the closest measurements was used. BMI was calculated as weight divided by height squared (kg/m2).
Nutrition supplement and timing of exposure.
Treatment assignment (atole or fresco) was based on birth village. Age of exposure was determined on the basis of the child’s date of birth and the trial start (1 March 1969) and end (28 February 1977) dates. We assumed all participants were born at term. Participants born between 28 November 1969 and 11 March 1975 were considered exposed for the first 1000 d, whereas children conceived before 1 March 1969 would not have been fully exposed during gestation, and children conceived after 14 June 1974 would not have been exposed through age 24 mo (17).
Childhood household SES.
Data on socioeconomic factors were collected by interview. SES was a cumulative score developed from principal components analyses of household characteristics and consumer durable goods measured in a 1967 survey of participant households (18). We categorized household SES score into tertiles.
Covariates.
Maternal age in years and maternal years of schooling were ascertained by interview and specified as continuous variables in the models. All of the data collection followed protocols that were approved by the institutional review boards of Emory University (Atlanta, Georgia) and INCAP (Guatemala City, Guatemala). All of the participants or their parents, as appropriate, gave written informed consent.
Statistical methods
LCGA can help identify distinct growth patterns in cohort subgroups that are not readily identifiable by using other modeling techniques. In non–latent class type growth modeling, a single curve is used to describe average growth patterns—potentially obscuring heterogeneity in growth (19). With LCGA, similar individuals are grouped together on the basis of their growth characteristics, and each latent class has its own growth curve (20).
We derived BMI latent class trajectories from up to 22 possible measures of height and weight: 5 from 1 to 12 mo, 3 from 13 to 23 mo, 5 from 24 to 50 mo, 3 from 51 to 84 mo, 2 from 10 to 20 y, and 4 from 21 to 42 y (Supplemental Figure 1). Among included participants, 5% had 3 measurements, 39% had 4–9 measurements, 34% had 10–14 measurements, and 23% had ≥15 measurements. Due to potential sex differences in the physiologic responses to early life nutrition, we modeled sex-specific trajectories (21).
Variance and covariance estimates for growth factors were fixed to zero, so within a given latent class, participants had the same slope and intercept (15). We developed models with the use of all available data and robust maximum likelihood estimation. To minimize local solutions, we specified 200 random starting values. We assessed overall model fit with the use of the Bayesian information criterion, the Bootstrap Likelihood Ratio Test, and the Lo-Mendell-Rubin Likelihood Ratio Test, and took the interpretability of classes into account in determining the final model (15). To assess the quality of classification, we used entropy (a statistic ranging from 0 to 1 assessing the accuracy of classification, where a higher value indicates greater classification accuracy) and posterior probabilities (probability of assigning observations to groups given the data) (19, 22). Age at onset of overweight was visually assessed by graphing the mean BMI for each class at the available ages and assessing the age at which mean BMI (in kg/m2) exceeded 25.
To characterize early life determinants of BMI latent class membership, we used binomial and multinomial logistic regression for women and men, respectively. We estimated the difference-in-difference (DD) effect of atole relative to fresco during the first 1000 d. Model 1 included a variable for supplement (coded as 1 for atole and 0 for fresco), a variable for age of exposure (coded as 1 for exposure during the first 1000 d and 0 for exposure at other ages), and an interaction term representing the differential effect of exposure to atole compared with fresco during the first 1000 d, after subtracting the difference between participants exposed to atole or fresco at other ages (i.e., those coded as 0 for “age of exposure”) (17). Analyses focused on the estimate and significance of this interaction term. Model 2 included household SES tertile, maternal age, and maternal education. Model 3 included all variables from models 1 and 2.
In sensitivity analyses, we estimated the DD effect of atole during the first 1000 d on obesity (BMI ≥30) at the last age observed, controlling for birth year. We also derived latent class trajectories excluding BMI values taken during the first 1000 d and re-conducted all multivariate logistic analyses.
Analyses were conducted by using MPlus version 7.3 (Muthén & Muthén). Because 84% of participants had ≥1 sibling in the trial, SEs were adjusted by using the CLUSTER option to account for within-family correlation (23). Significance was set a priori at P < 0.05. All P values were 2-sided.
Results
The study sample was 51% male (Table 1). Twenty-five percent of females and males were exposed to atole during the first 1000 d.
TABLE 1.
Selected characteristics of the study population, by sex, in the INCAP Nutrition Supplementation Trial longitudinal cohort1
| Women (n = 711) |
Men (n = 742) |
|||||
| Atole during full first 1000 d after conception (n = 176) | Other2 (n = 535) | P3 | Atole during full first 1000 d after conception (n = 183) | Other2 (n = 559) | P3 | |
| Childhood household SES tertile, % | 0.1 | 0.09 | ||||
| Poorest | 28.8 | 37.1 | 29.5 | 37.9 | ||
| Middle | 36.4 | 30.4 | 35.5 | 29.3 | ||
| Wealthiest | 34.7 | 32.4 | 34.9 | 32.7 | ||
| Maternal age, y | 27.3 ± 7.3 | 27.1 ± 7.0 | 0.7 | 27.5 ± 7.2 | 27.5 ± 7.1 | 0.9 |
| Maternal schooling, y | 0.9 ± 1.4 | 1.0 ± 1.4 | 0.5 | 1.0 ± 1.3 | 1.1 ± 1.4 | 0.4 |
| BMI latent class, % | <0.0001 | <0.0001 | ||||
| Low | 33.5 | 65.1 | 28.4 | 42.6 | ||
| Medium | — | — | 48.6 | 46.4 | ||
| High | 66.4 | 34.8 | 22.9 | 10.9 | ||
Values are means ± SDs unless otherwise indicated. INCAP, Institute of Nutrition of Central America and Panama; SES, socioeconomic status.
“Other” includes children exposed to atole at ages other than the first 1000 d after conception and all children exposed to fresco.
For categorical variables, P values were calculated by chi-square tests for equality of proportions across levels of BMI latent class. All P values are 2-sided.
On the basis of model fit and quality of classification, we identified 2 BMI latent classes for women (Supplemental Table 1): low (n = 407; 57%) and high (n = 304; 43%) (Figure 2A). For men, we identified 3 classes: low (n = 290; 389%), medium (n = 348; 47%), and high (n = 104; 15%) (Figure 2B). The Lo-Mendell-Rubin Likelihood Ratio Test and the Bootstrap Likelihood Ratio Test suggested that the inclusion of an additional class did not improve model fit (Supplemental Table 2). The 2-class model for women and the 3-class model for men had the highest entropy (0.76 and 0.77, respectively) and the highest posterior probabilities (∼0.93 and ∼0.91, respectively) of the candidate models, suggesting high class separation.
FIGURE 2.
Mean BMI (in kg/m2) by latent class group in women (A) and men (B) in the INCAP Nutrition Supplementation Trial longitudinal cohort. Sex-specific BMI latent class trajectories were derived from 22 possible measures of height and weight from 1969 to 2004. INCAP, Institute of Nutrition of Central America and Panama.
The median age at onset of overweight among women was 29 y and 23 y for the low- and high-BMI latent classes, respectively. Among men, the median age at onset of overweight was 34 y for the low-BMI latent class and 29 y for the medium- and high-BMI classes. At 42 y, the mean difference in mean BMI between the low- and high-BMI latent classes was 5.2 for women and 6.0 for men.
The interaction term (estimating the DD effect of atole relative to fresco during the first 1000 d on BMI latent class membership) was not significant in women (high- compared with low-BMI latent classes, P = 0.16; Table 2) or in men (middle- and high-BMI latent classes compared with the low-BMI latent class: P = 0.29 and P = 0.43, respectively; Table 3). In sensitivity analyses, estimates were robust to restricting trajectories to BMI values after the first 1000 d (DD: P > 0.58 for men and women). Furthermore, atole during the first 1000 d was not significantly associated with obesity at the last age observed (DD: P > 0.30 for men and women) in multivariate analyses.
TABLE 2.
Multivariate binomial logistic regression predicting BMI latent class membership based on atole exposure during the first 1000 d after conception and childhood household SES in women in the INCAP Nutrition Supplementation Trial longitudinal cohort1
| Model 1 | Model 2 | Model 3 | |
| High- vs. low-BMI latent class | |||
| Supplement type (atole vs. fresco) | 1.51 (0.98, 2.32) | — | 1.40 (0.82, 2.29) |
| Exposure during full first 1000 d (vs. exposure at other ages) | 2.28 (1.43, 3.61) | — | 2.20 (1.38, 3.68) |
| Interaction (exposure to atole during full first 1000 d) | 1.582 (0.84, 2.96) | — | 1.632 (0.81, 3.26) |
| Childhood household SES tertile | |||
| Middle vs. poorest | — | 1.76 (1.17, 2.65) | 1.63 (1.06, 2.51) |
| Wealthiest vs. poorest | — | 2.19 (1.48, 3.29) | 1.97 (1.29, 3.00) |
Values are ORs (95% CIs), n = 711. Model 1 included a dummy variable for supplement assignment and birth village (atole vs. fresco), a dummy variable for age of exposure (conception to 2 y vs. other), and a multiplicative term for the interaction between supplement type and age of exposure. Model 2 included SES tertile, maternal age in years, and reported maternal years of schooling. Model 3 included all components of models 1 and 2. We controlled for clustering of participants within households. For all models, low-BMI trajectory class is the reference. INCAP, Institute of Nutrition of Central America and Panama; SES, socioeconomic status.
P = 0.1.
TABLE 3.
Multivariate multinomial regression predicting BMI latent class membership based on atole exposure during the first 1000 d after conception and childhood household SES in men in the INCAP Nutrition Supplementation Trial longitudinal cohort1
| Model 1 | Model 2 | Model 3 | |
| Medium- vs. low-BMI latent class | |||
| Supplement type (atole vs. fresco) | 1.63 (1.07, 2.48) | — | 1.91 (1.21, 3.03) |
| Exposure during full first 1000 d (vs. exposure at other ages) | 1.48 (0.95, 2.31) | — | 1.48 (0.92, 2.39) |
| Interaction (exposure to atole during full first 1000 d) | 0.862 (0.45, 1.62) | — | 0.693 (0.35, 1.37) |
| Childhood household SES tertile | |||
| Middle vs. poorest | — | 1.43 (0.96, 2.13) | 1.30 (0.86, 1.95) |
| Wealthiest vs. poorest | — | 1.69 (1.13, 2.55) | 1.62 (1.07, 2.45) |
| High- vs. low-BMI latent class | |||
| Supplement type (atole vs. fresco) | 1.65 (0.84, 3.23) | — | 1.87 (0.87, 4.00) |
| Exposure during full first 1000 d (vs. exposure at other ages) | 1.42 (0.69, 2.92) | — | 1.46 (0.66, 3.21) |
| Interaction (exposure to atole during full first 1000 d) | 1.764 (0.68, 4.54) | — | 1.515 (0.53, 4.25) |
| Childhood household SES tertile | |||
| Middle vs. poorest | — | 1.27 (0.68, 2.39) | 1.07 (0.56, 2.03) |
| Wealthiest vs. poorest | — | 2.18 (1.21, 3.92) | 1.98 (1.09, 3.61) |
Values are ORs (95% CIs), n = 742. Model 1 included a dummy variable for supplement assignment and birth village (atole vs. fresco), a dummy variable for age of exposure (conception to 2 y vs. other), and a multiplicative term for the interaction between supplement type and age of exposure. Model 2 included SES tertile, maternal age in years, and reported maternal years of schooling. Model 3 included all components of models 1 and 2. We controlled for clustering of participants within households. For all models, low-BMI trajectory class is the reference. INCAP, Institute of Nutrition of Central America and Panama; SES, socioeconomic status.
P = 0.7.
P = 0.2.
P = 0.6.
P = 0.4.
High-SES tertile was associated with increased odds of high-BMI latent class membership in both women (OR: 1.97; 95% CI: 1.29, 3.00) and men (OR: 1.98; 95% CI: 1.09, 3.61) relative to the lowest tertile. Among women, middle-SES tertile was also associated with increased odds of high-BMI latent class membership (OR: 1.63; 95% CI: 1.06, 2.51) relative to the lowest tertile.
Discussion
With experimental data from a longitudinal cohort of Guatemalan men and women, we examined the role of early childhood factors on the timing and rate of BMI gain with the use of LCGA. We identified 2 BMI latent classes in women and 3 in men. Early separation of the latent classes and lack of overlapping trajectories over the life course suggest that preconceptional or early life factors are important in establishing BMI trajectories or reflect persistent circumstances and/or behaviors. Exposure to atole during the first 1000 d was not associated with BMI latent class membership, whereas higher childhood household SES tertile was associated with increased odds of high-BMI latent class membership in both women and men.
Atole, a food-based protein-energy supplement, was associated with improved child growth among those exposed before age 3 y (24). Nevertheless, we did not find evidence to suggest that exposure to atole during the first 1000 d was associated with BMI latent class membership. Studies in high-income countries have shown an association between rapid growth before age 2 y and obesity risk (25–27); however, others found no increased risk when weight gain was appropriate for linear growth (28, 29). A study of infant BMI trajectories in the Philippines found that accelerated infant BMI did not have long-term consequences on young-adult BMI (30). In the original trial, atole exposure before age 3 y was positively associated with linear growth and weight, negatively with skinfolds, and had no association with weight-for-height (24, 31). Other studies in these villages also suggest that, to date, the increases in length-for-age have not been accompanied by increases in overweight (32, 33).
Excessive weight gain relative to linear growth could be the result of inappropriate or unbalanced nutrition. Atole had relatively higher protein and fat than fresco and higher protein but lower fat than lipid-based nutrient supplements (LiNSs) used to prevent undernutrition in some LMICs. Among children 15–36 mo of age, the mean energy contribution of atole to the total diet was 150 kcal, 10 g protein, and 0.6 g fat/d compared with 25 kcal/d and 0 g protein or fat for fresco, with minor substitution by the nutrition supplements to the home diet (31). The recommended dose of small-quantity LiNSs in infants and young children is 110–120 kcal/d, which provides 2.6 g protein and 9.6 g fat (34). To date, trials have not found conclusive evidence to support a positive effect of small-quantity LiNSs on linear growth (35–37). Our findings suggest that nutrition supplementation with sufficient protein provided during the first 1000 d can improve linear growth but is not significantly associated with higher BMI trajectory in these analyses.
The lack of association between nutrition supplementation and higher BMI trajectory in this study could be due to the substantial burden of undernutrition during the original trial. At age 2 y, 86% of the study population was stunted (height-for-age z scores <−2 SDs), suggesting a strong need for nutritional support (24). Studies from Chile found that a complementary feeding program was associated with increases in child weight-for-length without increases in length-for-age in a population with a mean height-for-age z score close to zero (1, 38); however, this national program targeted all children <6 y, including those not at high risk of nutrition deficiencies. This is typical in Latin America, where only 25% of food programs use anthropometric measurements to target beneficiaries (39). Garmendia et al. (40) found that only 5–6% of supplementary feeding program beneficiaries in the region were truly underweight. Ensuring that programs target high-risk populations, address underlying nutrition deficiencies, and focus on healthy growth (appropriate weight gain and linear growth) could help prevent inadvertent increases in child obesity in LMICs.
In both men and women, high childhood household SES was associated with increased odds of high-BMI latent class membership relative to the poorest SES tertile. Little is known about how childhood SES influences adult obesity risk in LMICs. For childhood overweight (BMI-for-age z score >2 SDs), prevalence does not differ substantially by household SES in most LMICs. Among 78 LMICs with data, the prevalence of child overweight was, on average, 1.31 times higher (95% CI: 0.55, 3.60) in the richest quintile relative to the poorest (41). In the INCAP cohort, <2% of participants were overweight in childhood (33). For adult overweight, high SES is positively associated with obesity in low-income countries (10). However, the burden of obesity often shifts from the wealthy to the poor as countries move to middle income—an important distinction for countries with developing economies (42). To inform obesity prevention, more research is needed to understand the mechanisms through which SES increases risk. Childhood household SES likely functions as a proxy for a number of important factors that influence the early life environment, such as access to health care (43).
Since the trial ended in 1977, the cohort has been exposed to the cumulative effects of a changing food environment, urbanization, and other factors that could influence BMI (44, 45). Better early growth and development lead to increased educational attainment and income (17), and these might be associated with lifestyle risk factors for weight gain, such as sedentarism or ability to acquire energy-dense diets. The early separation of the BMI latent classes suggests that later life influences, such as parity in women (46), might contribute to adult BMI gain but do not contribute to the establishment of BMI classes.
There was a high burden of overweight in this cohort, which increases the risk of chronic disease. For all latent classes, mean BMI exceeded 25 by age 34 y, and the mean BMI in the high-BMI latent classes exceeded 30 by 42 y. The Global Burden of Disease Study found that a BMI >23 increased the risk of cardiovascular disease and diabetes, among other diseases (47).There is a strong need for chronic disease risk factor prevention and management in this population.
This study used experimental data from a randomized trial of a nutrition supplement and its follow-up studies. The cohort has >40 y of follow-up with serial measures of clinically measured anthropometry. The cohort has experienced relatively low attrition, and previous studies have indicated that attrition has not biased estimates of early life exposures and adult outcomes (48). To our knowledge, this is the first study to explore BMI trajectories from early childhood through midadulthood in an LMIC population with the use of LCGA. Most studies of BMI from birth to adulthood have focused on high-income country populations; however, growth patterns in LMICs are likely distinct due to high rates of childhood undernutrition (49, 50). Studies from other LMIC cohorts have been limited by the relatively young age of the study participants (51, 52). In a region in which compensatory growth is promoted and an estimated 20% of the population receives food assistance from nutrition programs, understanding how nutrition supplementation is associated with BMI gain is important in Latin America’s landscape of stunting and emergent obesity (1). Finally, in sensitivity analyses, atole during the first 1000 d after conception was not associated with obesity at the oldest age observed, and all of the findings were robust to restricting trajectories to BMI values after the first 1000 d.
There are several limitations to this study. The nutrition supplement was randomly assigned by village, so village-level effects on BMI are not adequately addressed by randomization but are captured within the DD design. We did not analyze the quantity of the supplement actually consumed. If the effect of atole during the first 1000 d on BMI latent class was small, we might have had limited power to detect an association. Finally, although LCGA helps identify heterogeneity in BMI gain, classes are not “real” but instead are heuristic, reflecting a continuum of growth in the population (53). Care should be taken to avoid oversimplification of the groups; instead, they should be considered a tool to help visualize variability within a global distribution.
In conclusion, early separation of the BMI latent classes suggests that early life factors are important in establishing BMI trajectory over the life course or reflect persistent circumstances and/or behaviors; however, exposure to atole during the first 1000 d was not associated with higher BMI latent class. This finding suggests that supplemental feeding programs during this critical developmental window are unlikely to adversely affect BMI trajectories. Higher childhood household SES was associated with increased odds of high-BMI latent class membership relative to the poorest households—the pathways through which this operates still need to be explored.
Acknowledgments
We thank Paúl Melgar at INCAP. We also thank Cria G Perrine for her feedback. NDF and ADS conceived of the original study idea, formulated the research question, and designed the study; RM led data collection activities during the 1988–1989 and 2002–2004 waves; NDF conducted all analyses, wrote the initial manuscript draft, and had primary responsibility for final content; and NDF, RM, NKM, MR-Z, and ADS interpreted the findings, contributed to the intellectual content of the work, and edited subsequent drafts. All authors read and approved the final manuscript.
Footnotes
Abbreviations used: DD, difference-in-difference; INCAP, Institute of Nutrition of Central America and Panama; LCGA, latent class growth analysis; LiNS, lipid-based nutrient supplement; LMIC, low- and middle-income country; SES, socioeconomic status.
References
- 1.Uauy R, Albala C, Kain J. Obesity trends in Latin America: transiting from under- to overweight. J Nutr 2001;131(Suppl):893S–9S. [DOI] [PubMed] [Google Scholar]
- 2.Jaacks LM, Slining MM, Popkin BM. Recent underweight and overweight trends by rural–urban residence among women in low- and middle-income countries. J Nutr 2015;145:352–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Popkin BM, Slining MM. New dynamics in global obesity facing low- and middle-income countries. Obes Rev 2013;14:11–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.PanAmerican Health Organization. Regional Health Observatory subregional health indicators [Internet]. Health Indicators Database. 2011 [cited 2015 Dec 29]. Available from: http://ais.paho.org/phip/viz/subregionalhealthindicators.asp.
- 5.Howe LD, Tilling K, Benfield L, Logue J, Sattar N, Ness AR, Smith GD, Lawlor DA. Changes in ponderal index and body mass index across childhood and their associations with fat mass and cardiovascular risk factors at age 15. PLoS One 2010;5:e15186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Belfort MB, Gillman MW, Buka SL, Casey PH, McCormick MC. Preterm infant linear growth and adiposity gain: trade-offs for later weight status and intelligence quotient. J Pediatr 2013;163:1564–9, e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Patro-Gołąb B, Zalewski BM, Kouwenhoven SM, Karaś J, Koletzko B, van Goudoever JB, Szajewska H. Protein concentration in milk formula, growth, and later risk of obesity: a systematic review. J Nutr 2016;146:551–64. [DOI] [PubMed] [Google Scholar]
- 8.Victora CG, de Onis M, Hallal PC, Blössner M, Shrimpton R. Worldwide timing of growth faltering: revisiting implications for interventions. Pediatrics 2010;125:e473–80. [DOI] [PubMed] [Google Scholar]
- 9.Martorell R. Results and implications of the INCAP follow-up study. J Nutr 1995;125(4 Suppl):1127S–38S. [DOI] [PubMed] [Google Scholar]
- 10.Sobal J, Stunkard AJ. Socioeconomic status and obesity: a review of the literature. Psychol Bull 1989;105:260–75. [DOI] [PubMed] [Google Scholar]
- 11.Monteiro CA, Moura EC, Conde WL, Popkin BM. Socioeconomic status and obesity in adult populations of developing countries: a review. Bull World Health Organ 2004;82:940–6. [PMC free article] [PubMed] [Google Scholar]
- 12.Jones-Smith JC, Gordon-Larsen P, Siddiqi A, Popkin BM. Is the burden of overweight shifting to the poor across the globe? Time trends among women in 39 low- and middle-income countries (1991–2008). Int J Obes (Lond) 2012;36(8):1114–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Subramanian SV, Perkins JM, Özaltin E, Davey Smith G. Weight of nations: a socioeconomic analysis of women in low- to middle-income countries. Am J Clin Nutr 2011;93:413–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Stein AD, Melgar P, Hoddinott J, Martorell R. Cohort profile: the Institute of Nutrition of Central America and Panama (INCAP) Nutrition Trial Cohort Study. Int J Epidemiol 2008;37:716–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Jung T, Wickrama KA. An introduction to latent class growth analysis and growth mixture modeling. Soc Personal Psychol Compass 2008;2:302–17. [Google Scholar]
- 16.Martorell R, Habicht JP, Rivera JA. History and design of the INCAP longitudinal study (1969–77) and its follow-up (1988–89). J Nutr 1995;125(4 Suppl):1027S–41S. [DOI] [PubMed] [Google Scholar]
- 17.Martorell R, Melgar P, Maluccio JA, Stein AD, Rivera JA. The nutrition intervention improved adult human capital and economic productivity. J Nutr 2010;140:411–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Maluccio JA, Murphy A, Yount KM. Research note: a socioeconomic index for the INCAP longitudinal study 1969–77. Food Nutr Bull 2005;26(2 Suppl 1):S120–4. [DOI] [PubMed] [Google Scholar]
- 19.Nylund KL. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Modeling 2007;14:535–69. [Google Scholar]
- 20.Hoekstra T, Barbosa-Leiker C, Koppes LL, Twisk JW. Developmental trajectories of body mass index throughout the life course: an application of latent class growth (mixture) modelling. Longit Life Course Stud 2011;2:319–30. [Google Scholar]
- 21.Case A, Menendez A. Sex differences in obesity rates in poor countries: evidence from South Africa. Econ Hum Biol 2009;7:271–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Berlin KS, Williams NA, Parra GR. An introduction to latent variable mixture modeling (part 1): overview and cross-sectional latent class and latent profile analyses. J Pediatr Psychol 2014;39:174–87. [DOI] [PubMed] [Google Scholar]
- 23.Liang K-L, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986;73:13–22. [Google Scholar]
- 24.Hoddinott J, Behrman JR, Maluccio JA, Melgar P, Quisumbing AR, Ramirez-Zea M, Stein AN, Yount KM, Martorell R. Adult consequences of growth failure in early childhood. Am J Clin Nutr 2013;98:1170–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Karaolis-Danckert N, Buyken AE, Bolzenius K, Perim de Faria C, Lentze MJ, Kroke A. Rapid growth among term children whose birth weight was appropriate for gestational age has a longer lasting effect on body fat percentage than on body mass index. Am J Clin Nutr 2006;84:1449–55. [DOI] [PubMed] [Google Scholar]
- 26.Angbratt M, Ekberg J, Walter L, Timpka T. Prediction of obesity from infancy to adolescence. Acta Paediatr 2011;100:1249–52. [DOI] [PubMed] [Google Scholar]
- 27.Malhotra R, Ostbye T, Riley CM, Finkelstein EA. Young adult weight trajectories through midlife by body mass category. Obesity (Silver Spring) 2013;21:1923–34. [DOI] [PubMed] [Google Scholar]
- 28.Adair LS, Fall CHD, Osmond C, Stein AD, Martorell R, Ramirez-Zea M, Sachdev HS, Dahly DL, Bas I, Norris SA, et al. Associations of linear growth and relative weight gain during early life with adult health and human capital in countries of low and middle income: findings from five birth cohort studies. Lancet 2013;382:525–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Taveras EM, Rifas-Shiman SL, Belfort MB, Kleinman KP, Oken E, Gillman MW. Weight status in the first 6 months of life and obesity at 3 years of age. Pediatrics 2009;123:1177–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Slining MM, Herring AH, Popkin BM, Mayer-Davis EJ, Adair LS. Infant BMI trajectories are associated with young adult body composition. J Dev Orig Health Dis 2013;4:56–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Martorell R, Habicht J-P, Klein RE. Anthropometric indicators of change in nutritional status in malnourished populations. In: Underwood BA, editor. Proceedings. Methodologies for human population studies in nutrition related to health. Bethesda (MD): US Department of Health and Human Services; 1982. p. 96–110. [Google Scholar]
- 32.Stein AD, Wang M, DiGirolamo A, Hoddinott J, Martorell R, Ramirez-Zea M, Yount KM. Height for age increased while body mass index for age remained stable between 1968 and 2007 among Guatemalan Children. J Nutr 2009;139:365–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Corvalán C, Gregory CO, Ramirez-Zea M, Martorell R, Stein AD. Size at birth, infant, early and later childhood growth and adult body composition: a prospective study in a stunted population. Int J Epidemiol 2007;36:550–7. [DOI] [PubMed] [Google Scholar]
- 34.Arimond M, Zeilani M, Jungjohann S, Brown KH, Ashorn P, Allen LH, Dewey KG. Considerations in developing lipid-based nutrient supplements for prevention of undernutrition: experience from the International Lipid-Based Nutrient Supplements (iLiNS) Project. Matern Child Nutr 2015;11:31–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Sayyad-Neerkorn J, Langendorf C, Roederer T, Doyon S, Mamaty A-A, Woi-Messe L, Manzo ML, Harouna S, de Pee S, Grais RF. Preventive effects of long-term supplementation with 2 nutritious food supplements in young children in Niger. J Nutr 2015;145:2596–603. [DOI] [PubMed] [Google Scholar]
- 36.Ashorn P, Alho L, Ashorn U, Cheung YB, Dewey KG, Gondwe A, Harjunmaa U, Lartey A, Phiri N, Phiri TE, et al. Supplementation of maternal diets during pregnancy and for 6 months postpartum and infant diets thereafter with small-quantity lipid-based nutrient supplements does not promote child growth by 18 months of age in rural Malawi: a randomized controlled trial. J Nutr 2015;145:1345–53. [DOI] [PubMed] [Google Scholar]
- 37.Mangani C, Maleta K, Phuka J, Cheung YB, Thakwalakwa C, Dewey K, Manary M, Puumalainen T, Ashorn P. Effect of complementary feeding with lipid-based nutrient supplements and corn–soy blend on the incidence of stunting and linear growth among 6- to 18-month-old infants and children in rural Malawi. Matern Child Nutr 2015;11:132–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Kain J, Pizarro F. Effect of an enhanced supplementary feeding program on infant’s length. Arch Latinoam Nutr 1997;47:101–4. [PubMed] [Google Scholar]
- 39.Uauy R, Rojas J, Corvalan C, Lera L, Kain J. Prevention and control of obesity in preschool children: importance of normative standards. J Pediatr Gastroenterol Nutr 2006;43:S26–37. [DOI] [PubMed] [Google Scholar]
- 40.Garmendia ML, Corvalan C, Uauy R. Addressing malnutrition while avoiding obesity: minding the balance. Eur J Clin Nutr 2013;67:513–7. [DOI] [PubMed] [Google Scholar]
- 41.Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, de Onis M, Ezzati M, Grantham-McGregor S, Katz J, Martorell R, et al. Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet 2013;382:427–51. [DOI] [PubMed] [Google Scholar]
- 42.Monteiro CA, Conde WL, Lu B, Popkin BM. Obesity and inequities in health in the developing world. Int J Obes Relat Metab Disord 2004;28:1181–6. [DOI] [PubMed] [Google Scholar]
- 43.Power C, Parsons T. Nutritional and other influences in childhood as predictors of adult obesity. Proc Nutr Soc 2000;59:267–72. [DOI] [PubMed] [Google Scholar]
- 44.Rivera JA, Barquera S, Gonzalez-Cossio T, Olaiz G, Sepulveda J. Nutrition transition in Mexico and in other Latin American countries. Nutr Rev 2004;62:S149–57. [DOI] [PubMed] [Google Scholar]
- 45.Swinburn BA, Sacks G, Hall KD, McPherson K, Finegood DT, Moodie ML, Gortmaker SL. The global obesity pandemic: shaped by global drivers and local environments. Lancet 2011;378:804–14. [DOI] [PubMed] [Google Scholar]
- 46.Kim SA, Stein AD, Martorell R. Country development and the association between parity and overweight. Int J Obes (Lond) 2007;31:805–12. [DOI] [PubMed] [Google Scholar]
- 47.Forouzanfar MH, Alexander L, Anderson HR, Bachman VF, Biryukov S, Brauer M, Burnett R, Case D, Coates MM, Cohen A, et al. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. The Lancet [Internet]. 2015 [cited 2015 Dec 7]. Available from: http://www.sciencedirect.com/science/article/pii/S0140673615001282. [DOI] [PMC free article] [PubMed]
- 48.Grajeda R, Behrman JR, Flores R, Maluccio JA, Martorell R, Stein AD. The Human Capital Study 2002–04: tracking, data collection, coverage, and attrition. Food Nutr Bull 2005;26(2 Suppl 1):S15–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Barker DJP, Osmond C, Kajantie E, Eriksson JG. Growth and chronic disease: findings in the Helsinki Birth Cohort. Ann Hum Biol 2009;36:445–58. [DOI] [PubMed] [Google Scholar]
- 50.Eriksson JG, Osmond C, Kajantie E, Forsén TJ, Barker DJP. Patterns of growth among children who later develop type 2 diabetes or its risk factors. Diabetologia 2006;49:2853–8. [DOI] [PubMed] [Google Scholar]
- 51.Stein AD, Thompson AM, Waters A. Childhood growth and chronic disease: evidence from countries undergoing the nutrition transition. Matern Child Nutr 2005;1:177–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Richter LM, Victora CG, Hallal PC, Adair LS, Bhargava SK, Fall CH, Lee N, Martorell R, Norris SA, Sachdev P, et al. Cohort profile: the consortium of health-orientated research in transitioning societies. Int J Epidemiol 2012;41:621–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Bauer DJ. Observations on the use of growth mixture models in psychological research. Multivariate Behav Res 2007;42:757–86. [Google Scholar]