点击显示 收起
1 From the Medical Research Council Epidemiology Unit, Cambridge, United Kingdom (UE, SB, and NJW); the Department of Paediatrics, University of Cambridge, Addenbrooke's Hospital, Cambridge, United Kingdom (KO and DBD); and the Obesity Unit, Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden (YL, MN, and SR)
2 Supported by the European Commission, Quality of Life and Management of Living Resources, Key action 1 Food, Nutrition and Health Programme as part of the project entitled "Dietary and genetic influences on susceptibility or resistance to weight gain on a high fat diet" (QLK1-2000-00515). MN was supported by Arbetsmarknadens Försäkrings och Aktiebolag. 3 Reprints not available. Address correspondence to U Ekelund, Medical Research Council Epidemiology Unit, Elsie Widdowson Laboratory, Fulbourn Road, CB1 9NL, Cambridge, United Kingdom. E-mail: ue202{at}medschl.cam.ac.uk.
ABSTRACT
Background: Rapid early postnatal weight gain predicts increased subsequent obesity and related disease risks. However, the exact timing of adverse rapid postnatal weight gain is unclear.
Objective: The objective was to examine the associations between rapid weight gain in infancy and in early childhood in relation to body composition at age 17 y.
Design: This prospective cohort study was conducted in 248 (103 males) singletons and their mothers. Height and weight were measured at birth, 6 mo, and 3 and 6 y. The rates of weight gain during infancy (0–6 mo) and early childhood (3–6 y) were calculated as changes in sex- and age-adjusted weight SD scores during these time periods. At 17 y, body composition was measured by air-displacement plethysmography.
Results: Increasing weight gain during infancy and early childhood were both independently associated with larger body mass index, fat mass, relative fat mass, fat-free mass, and waist circumference at 17 y (P < 0.005 for all; adjusted for sex, birth weight, gestational age, current height, maternal socioeconomic status, and maternal fat mass). Rapid weight gain in infancy, but not in early childhood, also predicted taller height at 17 y (P < 0.001).
Conclusions: Rapid weight gain in both infancy and early childhood is a risk factor for adult adiposity and obesity. Rapid weight gain in infancy also predicted taller adult height. We hypothesize that rapid weight gains in infancy and early childhood are different processes and may allow separate opportunities for early intervention against obesity risk later in life.
Key Words: Air-displacement plethysmography • catch-up growth • weight gain • obesity • Stockholm Weight Development Study
INTRODUCTION
Overweight and obesity in children and young people are rapidly increasing in prevalence worldwide and are related to a wide range of adverse outcomes (1). Perinatal life has been identified as a key period for the development of obesity (2). High birth weight is associated with an increased risk of later obesity (3). Conversely, lower birth weight and thinner size at birth have been linked to increased central fat, impaired glucose tolerance, and features of the metabolic syndrome later in life (4–7). Low-birth-weight intrauterine growth–restricted infants usually compensate by showing rapid catch-up growth during the first year of life. Postnatal rapid weight gain has been suggested to be a risk factor for later obesity (8–12), elevated blood pressure in adolescent males (13), impaired glucose tolerance in young adults (14), and increased mortality from coronary heart disease (15). However, the exact timing of rapid postnatal weight gain in relation to later obesity risk is unclear.
Rapid weight gain, or upward percentile crossing, during the first 2 y of life has been linked to general and central adiposity at age 5 y (9). More recently, rapid weight gain during the first week or months of life was shown to be associated with an increased risk of overweight and obesity, on the basis of body mass index (BMI; in kg/m2), in childhood (10) and in young adulthood (11, 12). However, it has been debated whether rapid weight gain during infancy or early childhood (the period of the "adiposity rebound" at 3–6 y) may contribute to obesity and the adult metabolic syndrome (16). Furthermore, it is unknown whether rapid postnatal weight gain is associated with a general increase in body size [ie, increase in fat mass (FM), fat-free mass (FFM), and height] or a selective increase in these components of body composition, because a detailed characterization of body composition was not available in previous studies (11, 12).
In a cohort of 17-y-old Swedish adolescents followed prospectively from birth, we therefore examined whether rapid weight gain during infancy or early childhood, or both, was associated with FM, percentage FM, FFM, BMI, and waist circumference later in life, after the adjustment for possible confounding factors.
SUBJECTS AND METHODS
Subjects
SWEDES is a prospective study of weight development in the offspring of mothers participating in the Stockholm Weight and Pregnancy Development Study (SPAWN) (17, 18). Briefly, 2342 mothers were invited to participate in 1984–1985 and were followed during and after their pregnancies; 1423 of these mothers completed the study at the 1-y follow-up, and 481 mothers and their children participated in the follow-up study (SWEDES) after 17 y. Of these participants, complete data (including weight and height development during the first 6 y of life) were available for 248 (103 males) singletons and constitutes the sample for the present study.
A detailed dropout analysis between those mothers who were initially invited but not participating in the present study (n = 2094) and participants (n = 248) showed no significant differences in the number of previous pregnancies, age, weight, height, and BMI before pregnancy, total weight gain during pregnancy, and BMI at 6 and at 12 mo after pregnancy (P > 0.05 for all). Birth weight (3461 ± 494 compared with 3453 ± 563 g) and length of gestation (39.5 ± 1.6 compared with 39.4 ± 1.9 wk) did not differ significantly between the groups of children (P > 0.05).
Measurements in children
Birth weight and height were noted from hospital records. The ponderal index was calculated as birth weight/length3 (kg/m3). The heights and weights of the infants were measured by standard clinical procedures during routine visits at the child welfare center 4 times during the first year (birth and 6, 9, and 12 mo) of life and annually thereafter until age 6 y. Gestation was estimated from the date of the last menstrual period reported by the mothers; 242 infants were born full term (>36 wk), and 9 children were born preterm (33–36 wk). Analysis of the data excluding these 9 preterm individuals did not alter the results.
At follow-up, sexual maturity was assessed by using the 5-stage scale for breast development in females and for pubic hair in males, according to Tanner (19). A dichotomous variable, puberty passed (Tanner stage = 5) or not passed (Tanner stage < 5), was created.
Standing height was measured to the nearest 0.5 cm against a wall-mounted stadiometer, and body weight was measured to the nearest 0.1 kg with a BodPod scale (Life Measurement Instruments, Concord, CA). BMI was determined as weight/height2 (kg/m2), and adolescents were classified as normal weight, overweight, and obese according to the age-adjusted cutoffs described by Cole et al (20). Waist circumference was measured in duplicate, at the minimum circumference between the iliac crest and the rib cage, with subjects standing dressed in underwear and rounded to the nearest 0.5 cm. Body volume was measured by air-displacement plethysmography with the BodPod, after adjustments for predicted thoracic lung volume and estimated surface area artifact (21). FM, relative FM, and FFM were calculated according to the equation of Siri (22) by using the software provided by the manufacturer. Body volume was measured in duplicate or triplicate when the initial 2 measures differed by >150 mL. All subjects were measured while wearing tight-fitting underwear, or a swimsuit, and a swim cap.
Measurements in mothers
Maternal birth weight, smoking during pregnancy, and breastfeeding patterns were recorded on questionnaires during and after pregnancy. At follow-up, maternal education, occupation and monthly income were recorded on a questionnaire. Occupation was coded on a scale from 1 to 6 (according to Statistics Sweden) and used as an indicator of socioeconomic status (SES) (23). This variable is a commonly used indicator of SES in Sweden and was recently used in the Health-related Habits of Life report (23). Maternal body weight and weight development during and 1 y after pregnancy were measured by using standard clinical procedures (17). Standing height, body weight, waist circumference, and body composition were measured by using the same methods as described above. Mothers with a BMI > 25 were considered overweight, and mothers with a BMI > 30 were considered obese. The same procedures were adopted for the mothers and the study participants, and the mothers and their children were measured on the same day at follow-up. The local Ethical Committee of Huddinge University Hospital approved the study, and informed consent was obtained from each mother and each child.
Calculations and statistical analyses
Using all recorded body weight data, we generated internal SD scores, independent of sex and age, by subtracting the sample mean from the individual mean and then dividing by the sample SD for weight at birth, 6 mo, and 3 and 6 y. For an alternative validation, we also generated SD scores by comparing them against the Swedish growth reference (24). The rates of weight gain during infancy (0–6 mo) and early childhood (3–6 y) were calculated as changes in weight SD scores during these time periods. The analyses were also performed with the rate of weight gain stratified as "rapid" (gain in weight SD score >0.67), "slow" (decrease in weight SD score >0.67), and "no change" (SD change between –0.67 and 0.67). These groups are equivalent to "upward," "downward," and "average" in weight percentile crossing, because an SD of 0.67 represents the distance between each displayed percentile line on standard growth charts (ie, 2nd, 9th, 25th, 50th, 75th, 91st, and 98th percentile lines) (9).
The values are reported are means ± SDs unless otherwise stated. Differences between the sexes were tested by analysis of variance. Relations between variables were assessed by correlation and partial correlation coefficients. The independent associations between weight gain during infancy and early childhood and body-composition variables at age 17 y were tested by general linear modeling (analysis of covariance) and adjusted for sex, birth weight, gestational age, height, and maternal SES and FM. When BMI was modeled as the outcome variable, height was excluded as an exposure variable because height is one of the components of BMI (outcome variable). Similarly, the independent associations between weight gain and FFM or height were tested by GLM. When FFM was modeled as the outcome variable, adjustments were made for sex, birth weight, gestational age, and maternal SES and FFM. When height was modeled as the outcome variable, adjustments were made for sex, birth length, gestational age, and maternal SES and height. Model building was performed by first introducing infancy weight gain and then early childhood weight gain and finally by testing the interaction between weight gain during infancy and early childhood. We also analyzed our data by comparing groups of children who showed clinically significant rapid weight gain (ie, >0.67 SD) and average (no change) and decelerated weight gain (ie, <0.67 SD) during the 2 periods, with adjustment for the same cofactors and with the use of the same model building as described above. The statistical analysis was performed with the use of SPSS for WINDOWS (version 11.0; SPSS Inc, Chicago, IL), and a P value < 0.05 denoted statistical significance.
RESULTS
The physical characteristics of the subjects are provided in Table 1. Birth weight, BMI at birth, and BMI in the girls at age 17 y did not differ significantly from Swedish reference data (24, 25), whereas the BMI in the boys at age 17 y was slightly lower than the Swedish reference data (20.5 ± 2.5 compared with 21.3 ± 2.6; P = 0.002).
View this table:
Table 1. Physical characteristics of the subjects at follow-up and size at birth1
Weight gain in infancy (0–6 mo)
The gain in weight SD score in infancy (between 0–6 mo) was inversely related to birth weight (partial r = –0.48, P < 0.001) and ponderal index at birth (partial r = –0.50; P < 0.001; adjusted for gestational age and maternal weight gain during pregnancy). Overall, between 0 and 6 mo, 25.4% (n = 63) of all children showed rapid weight gain in infancy (>0.67 SD scores), whereas 24.2% (n = 60) showed relatively slow weight gain (> –0.67 SD scores).
Weight gain in early childhood (3–6 y)
The gain in weight SD score between 3 and 6 y was not significantly associated with size at birth (not shown) but was inversely related to body weight and BMI at age 3 y (r = –0.28 and r = –0.35, P < 0.001). Only 8.8% (n = 22) of all subjects showed rapid weight gain during early childhood (>0.67 SD scores), and 8.5% (n = 21) showed relatively slow weight gain during this period.
Weight gain in infancy was inversely related to rapid weight gain in early childhood (partial r = –0.25, P < 0.001), which indicated that weight gain between 3 and 6 y tended to slow down in those who showed rapid weight gain between 0 and 6 mo. Only 2 children had a rapid weight gain during both periods; one child had a slow weight gain during both periods. The exclusion of these subjects from the analyses did not change the results.
Body composition at age 17 y
Increasing rate of weight gain during infancy and early childhood (as continuous variables) were both significantly and independently (of each other) associated with larger FM, relative FM, FFM, waist circumference, and BMI at age 17 y (P < 0.005 for all; adjusted for sex, birth weight, gestational age, current height, and maternal FM and SES; Table 2). Weight gain during infancy (P < 0.0001), but not during early childhood (P = 0.21), predicted taller stature at age 17 y; however, the effects of weight gain in infancy on later body composition were independent of height.
View this table:
Table 2. Independent determinants of obesity indicators, fat-free mass (FFM), and height at age 17 y1
Weight gains (as FM, waist circumference, BMI, FFM, and height) in infancy and early childhood, stratified as clinically significant "rapid," "no change" (or "average"), and "slow," are shown in Figure 1. Rapid weight gain in infancy was associated with larger FM, waist circumference, BMI, FFM, and height at age 17 y (P < 0.001 for all). Rapid weight gain in early childhood was associated with larger FM, waist circumference, BMI, and FFM (P < 0.001 for all) but not with height at age 17 y.
FIGURE 1.. Adjusted mean fat mass (FM), waist circumference, BMI, fat-free mass (FFM), and height at age 17 y in the subjects, stratified according to rapid (increase in SD score >0.67), no change (change in SD score between –0.67 and 0.67), and slow (decrease in SD score >0.67) weight gain during infancy (0-6 mo) and early childhood (3-6 y). Error bars indicate 95% CIs. These stratifications are equivalent to upward, average, and downward weight percentile crossings, respectively, on standard growth charts. Rapid weight gain in infancy and in early childhood was associated with greater FM, FFM, waist circumference, and BMI (P for trend < 0.001 for all). Rapid weight gain in infancy was associated with height (P for trend < 0.005). Bars with different lowercase letters are significantly different, P < 0.05 (analysis of covariance with Bonferroni adjustment for multiple comparisons). Data were adjusted for sex, birth weight (or length), gestational age, maternal fat mass (or BMI or height), and maternal socioeconomic status. Data for waist circumference and FM were additionally adjusted for height.
No significant interactions were observed between birth weight and rapid weight gain in infancy and early childhood for any of the models. The results were essentially unchanged after further adjustment for sexual maturity (222 participants were postpubertal and 26 were classified as Tanner stage 4). We thereafter reanalyzed all our data by using externally calculated SD scores (24), and these results were also unchanged.
On the basis of recent international BMI standards (20), 10.5% (26/248) of all subjects were overweight at age 17 y (of whom 3 were obese). Fifty percent of these (n = 13) subjects had clinically significant rapid weight gain in infancy (n = 10) or in early childhood (n = 5); 2 subjects had rapid weight gain during both periods). Thus, 15.9% (10/63) of the subjects with rapid weight gain in infancy became overweight or obese at age 17 y (unadjusted relative risk: 1.8; 95% CI: 0.9, 3.8), and 22.7% (5/22) of the subjects with rapid weight gain in early childhood became overweight or obese at age 17 y (unadjusted relative risk: 2.5; 95% CI: 1.0, 5.8).
DISCUSSION
In a healthy contemporary birth cohort, rapid weight gain during both infancy (0–6 mo) and early childhood (3–6 y) were related to long-term changes in body composition. A gain in weight of 1 SD score in infancy and in early childhood was associated with an increase in FM of 1.8 and 3.4 kg, respectively, and with an increase in FFM of 1.0 and 1.4 kg, respectively, at age 17 y. Upward weight percentile crossing at these early ages also predicted larger BMIs and waist circumferences in young adulthood, whereas rapid weight gain in infancy also predicted taller adult height.
Although the effect of weight gain in early childhood on obesity risk appeared to be greater than that of weight gain in infancy, rapid weight gain occurred more commonly during the shorter infancy period (25.4%) than during early childhood (8.8%) in keeping with other studies (9). Thus, the population attributable risk (an estimate of the proportion of all overweight subjects that could be prevented if each risk factor were removed) was slightly higher for rapid weight gain in infancy (15.7%) than for rapid weight gain in early childhood (11.7%).
A growing number of contemporary studies have linked rapid weight gain in infancy to later obesity risk (8–12, 26, 27). The major advantages of our study were the long-term follow-up and the validated measure of body composition (ie, air-displacement plethysmography), which allowed important discrimination between the effects of rapid weight gain on FM and FFM. However, we cannot exclude the possibility that the entire childhood period, from birth to 6 y of age, is a critical period for the development of later obesity risk, as recently suggested (28).
Regardless, our data suggest that rapid weight gain in infancy and early childhood are at least partly mediated by different mechanisms, as indicated by their inverse interrelation and by their differing effects on childhood height. Rapid weight gain in infancy, or "catch-up," is suggested to be a compensatory mechanism following intrauterine growth restraint (9, 29). Accordingly, children who showed rapid weight gains in infancy were shorter, lighter, and thinner at birth than were other children. In contrast, children who had rapid weight gains in early childhood did not differ in size at birth from other children. A previous study suggested that rapid weight gain in infancy is mediated by decreased satiety (29), and another recent study reported that energy intake and sucking behavior, but not total and nonsleeping energy expenditure, predict infancy weight gain (30). It has been suggested that maternal prepregnant BMI and an interaction between the duration of breastfeeding and the introduction of complementary foods are associated with weight gain between birth and 1 y (31). In our study, maternal weight gain during pregnancy, but not maternal BMIs before pregnancy, was associated with weight gain in infancy; however, this association disappeared after adjustment for gestational age and birth weight (data not shown). Lower weight gain in infancy may partly underlie the protective effects of breastfeeding and of lower-nutrient milk formulas on later obesity and cardiovascular disease risk (32, 33). We did observe a negative association between the degree of breastfeeding at 2.5 mo and infancy weight gain during the first 6 mo. However, the inclusion of breastfeeding to the models did not substantially alter any of the other findings.
Rapid weight gain in infancy, but not in early childhood, predicted an increase in height at age 17 y. Rapid weight gain in infancy was shown to be associated with subsequent higher insulin-like growth factor I concentrations (34), and this major childhood growth factor could mediate the subsequent gains in lean body mass and height. Considering its high prevalence, rapid weight gain in infancy could have heterogeneous effects on subsequent statural growth and body composition. We suggest that interindividual differences in insulin-like growth factor I secretion or activity could lead to variations in height and lean mass gains and in disease risks in response to rapid weight gain in infancy.
Rapid weight gain in early childhood was also consistently and independently associated with an overall increase in FM and FFM but not in height. The determinants of rapid weight gain during early childhood are largely unproven. Compared with the infancy period, different environmental factors are likely to be important, including childhood food consumption patterns (35) and a sedentary lifestyle (36). However, the causal associations remain to be demonstrated.
In a recent longitudinal study, physical activity–related energy expenditure, assessed by the doubly labeled water method, did not predict future weight and FM gain in children (37). Future studies using other robust methods of assessing physical activity during daily living may help to clarify this issue. Food intake, especially of energy-dense foods, may also contribute. For example, the consumption of sugar-sweetened soft drinks among children has more than doubled during recent decades (38). Excessive intake of sweetened drinks is associated with greater childhood weight gain (39), and the odds of becoming obese may increase by as much as 60% with each additional daily serving of sugar-sweetened drinks (40). Similar findings were recently reported in adult women (41). Prospective studies incorporating careful characterization of childhood growth, body composition, and endocrine-, metabolic-, social-, nutritional-, and energy expenditure–related factors would advance our understanding of these complex associations. Identification of the mechanisms that underlie rapid weight gain in early life is likely to be important to understanding how to prevent obesity throughout life.
As with any observational study, caution should be exercised when inferring causality based on the findings. However, our observations of independent associations between rapid weight gain in infancy and childhood and body composition and height at age 17 y were consistent in both continuous and stratified analytic approaches. Our findings were consistent before and after adjustment for potential confounding by most known variables that are plausibly associated with infancy and childhood growth. Furthermore, all data were collected prospectively as part of a study aimed at addressing weight development in offspring, which decreased the risk of recall bias or unreliable measurements. Birth weight and body composition at age 17 y did not differ significantly between our study sample and the larger cohort on the basis of the availability of early growth measurements. Furthermore, birth weight and BMI at birth and at age 17 y in females were not significantly different from Swedish reference data (24, 25). Our relatively small sample size and the low prevalence of obesity reduced the power to robustly predict overweight and obesity risk related to rapid weight gain. However, the unadjusted relative risk of rapid weight gain in early childhood was significantly related to the risk in young adulthood, and the relative risks were similar in size to those reported in earlier larger studies (10, 27).
In conclusion, our data suggest that rapid weight gain, or upward percentile crossing, in infancy and in early childhood both predict increased adiposity and obesity risk in young adults. We hypothesize that rapid weight gain in infancy and childhood may be mediated by different factors and may provide separate opportunities for targeted interventions to prevent the development of obesity.
ACKNOWLEDGMENTS
We are grateful to Prevnut, CNT, NOVUM, Stockholm, Sweden for support with the BodPod equipment.
UE conceived the hypothesis for this article and conducted the data analyses. UE and KO were responsible for drafting the manuscript. YL and SR were primarily responsible for developing the study design for SWEDES, supervised the data collection, and provided critical input. MN, SB, NJW, and DBD provided critical input on the data analyses and on all versions of the manuscript. SR was the principal investigator, conceived the idea of the Stockholm Pregnancy and Weight Development Study and of the SWEDES, assisted with the study design, and provided critical input. All authors participated in the interpretation of the results and approved the final version of the manuscript. The sponsors had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. None of the authors had any conflicts of interest.
REFERENCES