Predictors of children s body mass index: a longitudinal study of diet and growth in children aged 2 8 y

Transcription

1 (2004) 28, & 2004 Nature Publishing Group All rights reserved /04 $ PEDIATRIC HIGHLIGHT Predictors of children s body mass index: a longitudinal study of diet and growth in children aged 2 8 y JD Skinner 1 *, W Bounds 2, BR Carruth 1, M Morris 1 and P Ziegler 3 1 Nutrition Department, University of Tennessee, Knoxville, TN, USA; 2 Center for Nutrition and Food Systems, College of Health, University of Southern Mississippi, Hattiesburg, MS, USA; and 3 Gerber Products Co., Kimball Drive, Parsippany, NJ, USA OBJECTIVE: To identify longitudinal variables related to children s body mass index (BMI) (kg/m 2 ) at age 8 y. DESIGN: A longitudinal design, with nine interviews per child from ages 2 to 8 y. SUBJECTS: In all, 70 white children (37 males, 33 females) who were continuous participants since infancy in the longitudinal study. Families were primarily middle and upper socioeconomic status. MEASUREMENTS: At each interview, children s height and weight were measured, and mothers provided 3 days of the child s intake data (a 24-h recall and 2 days of food records). ANALYSES: Analyses used were means7s.d., correlations, repeated measures analysis of variance, and forward stepwise regression. BMI at each interview was calculated and age of adiposity rebound was determined. RESULTS: Children s BMI at 8 y was negatively predicted by age of adiposity rebound and positively predicted by their BMI at 2 y. Additionally, each model included one longitudinal dietary variable; mean protein and fat intakes recorded between 2 and 8 y were positive predictors of BMI at 8 y; mean carbohydrate intake over the same time period was negatively related to BMI at 8y.R 2 values indicated that these three-variable models predicted 41 43% of the variability in BMI among children. BMI of 23% of the children exceeded the 85th CDC percentile. CONCLUSIONS: The results of this study show that factors in early life are associated with children s BMI at age 8 y. (2004) 28, doi: /sj.ijo Published online 2 March 2004 Keywords: children s BMI; adiposity rebound; dietary intake Introduction Overweight during childhood is an increasing problem in the United States and many other countries because childhood overweight increases risk of adult obesity and its related health problems. 1 5 Currently, at least 22% of American youth are categorized as overweight or at risk of overweight, and the percentage has increased from 15% since the 1960s. 1 A variety of factors are related to the development of childhood overweight, including ethnicity, 1,2 family socioeconomic status, 5 dietary excesses (eg energy, fat, protein), 6,7 lack of physical activity, 8,9 time spent watching television/videos/computer games, 8 10 short sleep duration, 10 birth weight/infant fatness, 10,11 duration of *Correspondence: Dr JD Skinner, Nutrition Department, University of Tennessee, 229 JHB, 1215 Cumberland Av., Knoxville, TN , USA. skinner@utk.edu Received 9 October 2002; revised 1 May 2003; accepted 9 May 2003 breastfeeding, 12 previous overweight, 5,13,14 age of adiposity rebound (AR), parental obesity, 10,11 parental feeding styles, 18 and genetics. 19,20 It is obvious from these studies that the causes of childhood overweight are multifactorial, and many inter-relationships among factors are likely. Nonetheless, dietary energy intake and physiological energy output, including the 2% of energy needs associated with growth, 2 must be balanced throughout childhood for normal weight patterns to develop. 1,8,11 As overestimation of children s energy requirements may be a prescription for childhood obesity, recommendations for children s energy intake have been a topic of considerable debate in the United States and elsewhere Using doubly labeled water (DLW) methodology, which measures total energy expenditure, results from multiple studies in various countries indicate that energy recommendations developed in the 1980s and early 1990s for infants and children to age 3 y may overestimate requirements by 10 20%. 23,24 Similarly, young children s (ie ages 4 6 y) energy recommen-

2 dations may be 25% higher than energy requirements. 21,25 Dietary energy recommendations are particularly relevant because child nutrition programs, such as the National School Breakfast and Lunch Programs and many preschool programs, serve amounts of food based on the latest energy (and nutrient) recommendations. However, underestimation of children s energy requirements also would have negative consequences because child nutrition programs could serve amounts of food that did not provide adequate energy for children s normal growth and development. The study was planned to describe longitudinal growth and energy intakes from ages 2 to 8 y in 70 children and to identify factors related to the children s body mass index (BMI), which is an accepted and frequently used measure of adiposity. 26,27 Additional growth data and food intake variables from the longitudinal data set (eg birth to age 8 y) were included in initial analyses. Methods Participants This study included 70 white children, 37 males and 33 females, who were continuous participants since infancy in a longitudinal study of children s food patterns and related factors;62ofthechildrenbeganthestudyat2or3monthsof age, two joined the longitudinal study prior to age 1 y, and six children participated in a similar infant study from this laboratory and joined this longitudinal study at age 2 y. 28 All children were born in 1992, with the majority of births in June, July, and August. As described elsewhere, 29 the children lived in a medium-sized southern city in the United States, most were from middle or upper socioeconomic status (SES) families, all were without birth anomalies, and none had continuing health problems that required medication. Most parents were college educated, and many held graduate degrees. Families with these characteristics were selected purposively to include parents with sufficient resources to provide children adequate diets and health care for normal growth and development. A single racial group was selected to avoid differences in growth patterns and body composition related to race. The university s institutional review board approved all study protocol. Mothers signed consent forms for all interviews and children signed assent forms for the later interviews. Interviews All interviews were conducted in the child s home by one of two interviewers, both of whom were registered dietitians (RD) with Master of Science degrees and clinical experience in a medical setting. Additionally, interviewers were trained in study protocol prior to each set of interviews. Each interviewer continued throughout the first 5 y with the same set of families. The other interviewer remained with the study throughout the 8 y period, and did all the interviews when children were ages 6, 7, and 8 y. One interviewer participated for the first 5 y of the study. A randomized incomplete block design was used to schedule early interviews (2 36 months) because interviews were more frequent during this period. The randomized data points were months 2, 3, 4, 6, 8, 10, 12, 16, 20, 24, 28, 32, and 36. However, all children were interviewed at 24 months either as part of the 12 to 24 month sequence or the month sequence. Each child had four to five interviews during the first year of life, two to three during the second year, and two during the third year. Beginning at age 3.5 y, each child was interviewed at each of the seven data points (ages 3.5, 4.0, 4.5, 5.0, 6.0, 7.0, and 8.0 y). Thus, each child had sets of growth and dietary data. Additional information, such as breastfeeding duration; picky eating practices; average h per day spent watching television, videos, and computer games; and children s food preferences, were collected at some but not all interviews. At each interview, the RD weighed and measured the child, using established protocol. 30,31 Initially, the infant was weighed to the nearest 0.25 ounce, using a Detecto Doctor s Infant Scale (Detecto Scale Company, Webb City, MO, USA), and recumbent length was measured to the nearest 0.1 cm with a board designed like the one used in health clinics in this area. Later, the child was weighed to the nearest 0.1 pound using a standard scale (model 707 SECA, Columbia, MD, USA); both scales were periodically checked with standards of known weight. The child s standing height was measured with a steel tape to the nearest 0.1 cm using a wall or doorway in the child s home and a square (ie two boards attached at a right angle). All measures were taken in triplicate and averaged. Parental BMIs were measured for mothers and reported (by mothers) for fathers in the interviews when children were 8 y old. The child s food and beverage intake was provided by the mother at each interview. During the more frequent early interviews (2 24 months), a single 24-h recall was collected at each interview. Beginning with the third year interviews (the month sequence), mothers provided the child s dietary data for three nonconsecutive days (two food records and a 24-h recall). Data collection days were assigned by the research team to include one weekend day with each set of 3-day dietary data. Mothers were taught during the initial interviews about the necessity of detailed and complete descriptions of the child s food and beverage intake and the importance of accurate estimates of portion sizes. Additionally, detailed instructions were provided with each set of food record forms. During each interview, the RD reviewed the food records with the mother for ambiguities, possible omissions, and potential errors in estimating amounts of food or beverages. Analyses SAS System for UNIX (SAS version 8.1, Cary, NC, USA) was used for computations and analyses. A significance level of Pp0.05 was established for all analyses. BMI at each interview was calculated for each individual, beginning at age 2 y, and group means7s.d. were calculated by gender. Children s weight, height, and BMI percentiles were calculated with an SAS program 32 based on the newest growth charts from The Center for Disease Control (CDC)

3 478 The first step in estimating each child s age at adiposity rebound (AR), that is, the age at which BMI increased following the lowest BMI, was to graph each child s BMI from birth to age 8 y. Four researchers then individually identified the lowest BMI for each child; there was 100% agreement among researchers for 42 of the 70 children and 75% agreement for 17. The remaining 11 graphs were reviewed again until consensus was reached among the researchers and the nadir verified with calculated BMI. Visual determination of the age of the BMI nadir as a proxy for the AR was also used by Rolland-Cachera et al 15 and Dorosty et al. 34 Energy and macronutrient intakes were determined for each day for each individual using a nutrient analysis program (Nutritionist IV, version 3.5, N-Squared Computing, and version 4.1, First DataBank, San Bruno, CA, USA). Each individual s 3-day dietary intakes were averaged to provide a representative day for each of the nine interview times. The nine representative days were totaled for each individual for further statistical analyses. This total of nine representative days were termed longitudinal intake for this study. Thus, longitudinal intakes were based on 27 days of dietary data per child, collected in nine sets of 3-days data from age 2 to 8 y. Group means 7 s.d. by gender at each age were calculated and compared with the age appropriate energy recommendations, both the 2002 recommendations 35 and the 1989 recommendations, 36 which were in effect at the time these data were collected. Group changes in energy intake over time and gender differences in energy intake were tested with general linear model (GLM) repeated measures analysis of variance. Predictive models Based on the literature and selected data from the longitudinal study, the following variables were considered for inclusion in statistical models to predict children s BMI at age 8 y: gender, birth weight, breastfeeding duration, age that cereal was introduced, BMI at age 2 y, estimated AR, longitudinal dietary energy intake, longitudinal intakes of each macronutrient, longitudinal percentages of energy from each macronutrient, mother s perception of child as a picky eater at age 6 y, 37 the number of foods liked at age 8 y, as assessed on the Food Preference Questionnaire, 38 average daily time spent at ages 6 and 7 y watching television, videos, or computer games (as estimated by mothers), average dietary variety score at ages y, 39 and BMI of both parents. Potential relationships among these variables and between these variables and children s BMI at 8 y were investigated with correlation analyses and the RSQUARE procedure in SAS. The RSQUARE procedure considers all variables simultaneously to maximize R 2, while identifying potentially significant variables in sequential models with two, three, four, and five, etc variables. These procedures were followed by forward selection stepwise regression to test the significance of all variables in the selected models. Only those models in which all variables were significant at Pp0.05 will be reported. Results Growth indices Group means7s.d. by gender for children s weight, height, BMI, and percentiles of each from ages 2 8 y are presented in Table 1. Yearly differences in growth indices between time periods were significant. Between 2 and 8 y, children gained about 2 4 kg per year in weight and added about 6 8 cm in Table 1 Children s growth indices, ages 2 8 y. Values are means7s.d. and mean percentiles from CDC growth charts Age (y) N a Weight (kg) Height (cm) BMI (kg/m 2 ) (Wt percentile) (Ht percentile) (BMI percentile) Males (51.8) (65.5) (49.1) (56.5) (60.4) (51.1) (57.3) (64.1) (45.8) (56.5) (61.0) (48.1) (59.8) (63.1) (47.6) (60.6) (63.5) (51.3) (58.2) (65.7) (48.5) (59.6) (66.2) (52.3) (60.4) (69.4) (51.4) (63.5) (67.8) (56.3) (68.8) (69.0) (61.1) Females (48.8) (60.0) (39.2) (48.0) (51.2) (45.6) (48.0) (42.3) (50.0) (52.4) (62.3) (42.0) (50.7) (57.0) (53.0) (58.4) (49.5) (53.2) (56.5) (54.8) (55.4) (58.2) (56.0) (58.0) (58.0) (55.6) (60.8) (54.0) (60.0) (62.6) (54.7) (61.2) a From age 2 to 36 months, a randomized incomplete block design was used, so the total n is o70 except for 24 months when all children participated. 2.3 y ¼ average of first toddler interview at age 24, 28, or 32 months, 2.8 y ¼ average of second toddler interview at age 28, 32 or 36 months.

4 height. For males, BMI percentiles were near the CDC 50th percentile (Table 1) until age 8 y when group means increased to the 61st CDC percentile. For females, BMI percentile means increased to about the 55th CDC percentile at age y and the 60th CDC percentile at ages 7 and 8 y. Of the 70 children, 16 (23%), eight males and eight females, exceeded the 85th CDC BMI percentile at age 8 y, and of these 16, six (9%) children exceeded the 95th CDC BMI percentile. Children between the 85th and 95th CDC BMI percentiles are considered at risk of overweight, and children who exceed the 95th CDC BMI percentile are considered overweight. 33 Longitudinal energy intakes Children s mean energy and macronutrient intakes from 2 to 8 y of age are shown in Table 2. Although males energy intakes were generally ( kj) higher than females, differences were not statistically significant. Energy intakes increased gradually with children s age; differences between time periods were significant only between 54 and 60 months. Comparisons of children s energy intakes with the new estimated energy requirements (EERs) 35 and the 1989 recommendations for energy 36 are shown longitudinally in Figures 1 and 2. Average percentages of energy from fat, protein, and carbohydrate were 32, 14, and 56%, respectively. Predictors of BMI at age 8 y In addition to growth indices and energy and macronutrient intakes, several variables were initially included in analyses. Means7s.d. for these variables are presented in Table 3; mean values did not differ by gender. Table 4 includes four similar models that predicted children s BMI at age 8 y. All models included the AR (negatively related to BMI), BMI at age 2 y (positively related), and one dietary variable. Dietary fat (longitudinal intake in grams and longitudinal percent energy from fat) and dietary protein (longitudinal intake in grams) variables were positively related to BMI, whereas the percent energy from carbohydrate was negatively related. These models had R 2 values , with age of adiposity rebound contributing 19% of the variability, BMI at age 2 y 16%, and the dietary variables contributed 5 8%. Two variables did not contribute significantly to the predictive models, but they were significantly correlated with the child s BMI at age 8 y. mothers BMI, r ¼ 0.26, P ¼ 0.03; and TV/video/computer time, r ¼ 0.29, P ¼ Children s TV/video/computer time was also positively related to the child s BMI at age 2 y, (r ¼ 0.33, P ¼ 0.006), and both parents BMI, (mothers BMI r ¼ 0.24, P ¼ 0.04; fathers BMI r ¼ 0.32, P ¼ 0.006). Discussion The major findings of this study emphasize the longitudinal character of children s BMI at age 8 y, as predicted by the age of AR, BMI at age 2 y, and longitudinal (2 8 y of age) dietary intakes of macronutrients. It appears that children s early patterns may program their BMI later in childhood. The average ages of the BMI nadir (AR 4.7 y for males, 4.5 y for females) in this study are slightly lower than the 6 y reported by Rolland-Cachera et al, 15 the 5.1 y for males and 5.3 y for females reported by Siervogel et al, 14 the 6.0 y for 479 Table 2 Children s mean energy and macronutrient intakes, ages 2 8 y. Values are means7s.d. from 3 days of dietary data at each interview Age (y) Gender (37 M, 33 F) Energy (kj) (kcal) Fat (g) Protein (g) Carbohydrate (g) 2.3 a M ( d ) F ( ) a M ( ) F ( ) M ( ) F ( ) M ( ) F ( ) M ( ) F ( ) M ( ) F ( ) M ( ) F ( ) M ( ) F ( ) M ( ) F ( ) a 2.3 y ¼ average of first toddler interview at age 24, 28, or 32 months, 2.8 y ¼ average of second toddler interview at age 28, 32, or 36 months.

5 480 Energy Intake (kcal) Males Energy Intake Over Time Children's Age in Months Males RDA EER Figure 1 Mean energy (kcal/day) intakes of 37 males from ages 2 to 8 y, compared to 1989 age- appropriate RDAs and 2002 EERs. Energy intake data were derived from 3- day food records/recalls obtained from mothers at each age. Data at 27 and 34 months are averages from randomly assigned interviews at 24, 28, or 32 months and 28, 32, or 36 months, respectively. Energy Intake (kcal) Females Energy Intake Over Time Children's Age in Months Females RDA EER Figure 2 Mean energy (kcal/day) intakes of 33 females from ages 2 to 8 y, compared to 1989 age- appropriate RDAs and 2002 EERs. Energy intake data were derived from 3- day food records/recalls obtained from mothers at each age. Data at 27 and 34 months are averages from randomly assigned interviews at 24, 28, or 32 months and 28, 32, or 36 months, respectively. males and the 5.6 y for females reported by Williams et al 17 or the 5.5 y reported by Whitaker. 4 Since the earlier reports 4,14,15,17 documenting the age of AR were from studies initiated several decades or more prior to the current study, it is possible that the earlier AR observed in this study represents societal changes. When individuals in the earlier studies were children, some of today s environmental influences known to be related to childhood obesity (eg television, videos, computer games) 8 10 were either not available or were available only to a limited extent. Additionally, it has been shown that lessactive children exhibit an earlier AR. 8 Thus, the increase in overweight in children and the decrease in AR age and physical activity may be inter-related and reflect temporal trends. Table 3 Additional independent variables included in predictive regression models. The dependent variable was children s BMI (kg/m 2 )atage8y Males Females Variable n ¼ 37 n ¼ 33 Birthweight (kg) Cereal introduction age (months) Breastfeeding duration (months) Dietary variety score months (%) a Adiposity rebound (age in y) Picky eater, 6 y (% of sample), mothers designation TV, video, computer time, ages 6 and 7 y, (h/day) Foods liked, 8 y (% of 196 foods) Mothers BMI (kg/m 2 ) Fathers BMI (kg/m 2 ) a Based on 100%, which indicates that child met minimum daily number of servings recommended in each food group (eg two fruit and three vegetable servings) according to the Food Guide Pyramid. Table 4 Regression models predicting children s a BMI at age 8 y. N ¼ 70 children Variable Parameter estimate7s.e. F Pr>F Model I (R 2 ¼ 0.43 F ¼ 16.6 Po0.0001) Adiposity rebound age o BMI at 2 y o Longitudinal dietary fat intake (g) a Model II (R 2 ¼ 0.42 F ¼ 15.6 Po0.0001) Adiposity rebound age o BMI at 2 y o Longitudinal dietary fat (% energy) a Model III (R 2 ¼ 0.41 F ¼ 15.6 Po0.0001) Adiposity rebound age o BMI at 2 y o Longitudinal dietary carbohydrate (% energy) a Model IV (R 2 ¼ 0.41 F ¼ Po0.0001) Adiposity rebound age o BMI at 2 y Longitudinal dietary protein (g) a a Calculated from nine sets of 3-day dietary data (a set ¼ a 24-h recall and two food records) collected from ages 2 to 8 y.

6 Another predictor of BMI at age 8 y was the child s BMI at age 2 y, which is consistent with the literature. Siervogel et al 14 reported significant correlations between BMI at 2 y and BMI at age 18 y, r ¼ 0.31 in males and 0.22 in females. Scaglioni et al 6 reported a positive association between BMIs at ages 1 and 5 y. Wang et al 5 tracked 6 13 y old Chinese children for 6 y; the results indicated that the initial BMI was significantly correlated with BMI at follow-up, r ¼ 0.42 for males, 0.36 for females. Guo et al 3 tracked 347 individuals from childhood to age 35 y; probabilities of BMIX25 Kg/m 2 (ie defined by the authors as overweight) at age 35 y were for those with BMIX85th percentile at age 3 y. Deheeger et al 13 found that 47% of children were in the same BMI percentage tertile at 10 months as at 8 y of age. These findings are also consistent with a previous report on the children in this longitudinal study that showed that infants weight prior to 12 months was the strongest predictor of their weight gain, months. 40 Only a few studies have simultaneously examined children s growth, a measure of adiposity, and longitudinal energy/macronutrient intakes. 6,7,11,13 A higher protein intake at ages 1 2 y predicted overweight at age 5 y 6 and an earlier AR, 7 which in turn was related to higher BMI at age 8 y. Early energy intake was not related to BMI at age 5 y, 6 although carbohydrate was negatively related. 6 Energy intake increased more between ages 4 and 6 y in overweight children than in lean and medium weight children; 13 interestingly, this time period coincides with the usual period of AR. These findings 6,7,11,13 are consistent with the current study, in which longitudinal dietary intakes of protein and fat were positively related to BMI, and carbohydrate was negatively related. Although statistically significant, longitudinal dietary factors predicted only a minor amount of the variability in children s BMI at age 8 y, and longitudinal energy intake did not appear as a significant variable in any of the four models. Children s macronutrient intakes were within the currently recommended ranges. 35 Group mean intakes were adequate for protein and carbohydrate; total dietary fat recommendations have not been defined. 35 Although children s mean energy intakes generally were less than 1989 recommendations, 36 their intakes over time were generally within about 100 kcal (420 kj) of the 2002 estimated energy requirements (EERs) for physically active children. 35 However, mean energy intakes of this group of children exceeded recommendations for sedentary and low activity children. The EER has four levels of energy intake recommendations based on four levels of energy expenditure; EERs for children over age 4 y generally are lower than the 1989 recommended daily allowance (RDA). Although the accuracy of food records/recalls is sometimes questioned, several aspects of this study suggest enhanced reliability of the data. Mothers in the study were highly educated, a characteristic associated with accurate record keeping. They were experienced in providing food intake data because they had participated in the longitudinal study since the child was 2 or 3 months old and had been interviewed six or seven times prior to the 24-month interview. 29 The two interviewers were highly qualified and experienced in the collection of food intake data, and they had established rapport with the mothers and children in the early months of the longitudinal study. Using face-to-face interviews allowed for probing and checking of data. The longitudinal data presented in this paper represent 27 days of food intake data collected over a 6 y period for each child; this procedure diminishes the effect of a specific day, or even a specific 3 day period, that was not representative of the child s usual intake. A weakness of this study is that there were no observations to verify food intake nor was total energy expenditure measured by DLW methods. However, several DLW studies have shown that parental reports of children s energy intakes, using records/recalls, are reasonably accurate, unlike self-reports in adolescents and adults. 41 Another weakness of the study is that no validated estimate of children s overall physical activity was obtained and, thus, it is difficult to determine which level of physical activity should be used to compare energy intake with the newest recommendations. 35 The EERs used were those for active children, which actually may exceed the energy needs of children in this study, who averaged almost 3 h per day of sedentary activity including watching television, videos, or computer games. Other researchers have related children s TV time with measures of adiposity, 8,9 and the Dietary Reference Intake Committee strongly advocates increased physical activity along with the lowered energy recommendations for children. 35 Implications of study The results of this study identify several indicators of developing overweight in young children. These indicators were child s BMI as early as 2 y of age and an early AR. Acknowledgements This study was funded by Gerber Products Company and the Tennessee Agricultural Experiment Station. References 1 Troiano RP, Flegal KM, Kuczmarski RJ, Campbell SM, Johnson CL. Overweight prevalence and trends for children and adolescents. Arch Pediatr Adolesc Med 1995; 149: Ogden CL, Troiano RP, Briefel RR, Kuczmarski RJ, Flegal KM, Johnson CL. Prevalence of overweight among preschool children in the United States, 1971 through Pediatrics 1997; 99: e1 3 Guo SS, Wu W, Chumlea WC, Roche AF. Predicting overweight and obesity in adulthood from body mass index values in childhood and adolescence. Am J Clin Nutr 2002; 76: Whitaker RC, Pepe MS, Wright JA, Seidel KD, Dietz WH. Early adiposity rebound and risk of adult obesity. Pediatrics 1998; 101:e5: 481

7 482 5 Wang Y, Ge K, Popkin BM. Tracking of body mass index from childhood to adolescence: a 6-year follow-up study in China. Am J Clin Nutr 2000; 72: Scaglioni S, Agostoni C, De Notaris R, Radaelli G, Radice N, Valenti M, Giovannini M, Riva E. Early macronutrient intake and overweight at five years of age. Int J Obes Relat Metab Disord 2000; 24: Rolland-Cachera MF, Deheeger M, Akrout M, Bellisle F. Influence of macronutrients on adiposity development: a follow up study of nutrition and growth from 10 months to 8 years of age. Int J Obes Relat Metab Disord 1995; 19: Deheeger M, Rolland-Cachera MF, Fontvielle AM. Physical activity and body composition in 10 year old French children: linkages with nutritional intake? Int J Obes Relat Metab Disord 1997; 21: Andersen RE, Crespo CJ, Bartlett SJ, Cheskin LJ, Pratt M. Relationship of physical activity and television watching with body weight and level of fatness among children. JAMA 1998; 279: Locard E, Mamelle N, Billette A, Miginiac M, Munoz F, Rey S. Risk factors of obesity in a five year old population. Parental versus environmental factors. Int J Obes Relat Metab Disord 1992; 16: Parsons TJ, Power C, Logan S, Summerbell CD. Childhood predictors of adult obesity: a systematic review. Int J Obes Relat Metab Disord 1999; 23: S1 S Tulldahl J, Pettersson K, Andersson SW, HulthWn L. Mode of infant feeding and achieved growth in adolescence: early feeding patterns in relation to growth and body composition in adolescence. Obes Res 1999; 7: Deheeger M, Akrout M, Bellisle F, Rossignol C, Rolland-Cachera MF. Individual patterns of food intake development in children: a 10 months to 8 years of age follow-up study of nutrition and growth. Physiol Behav 1996; 59: Siervogel RM, Roche AF, Guo S, Mukherjee D, Chumlea WC. Patterns of change in weight/stature from 2 to 18 years: findings from long-term serial data for children in the Fels longitudinal growth study. Int J Obes Relat Metab Disord 1991; 15: Rolland-Cachera MF, Deheeger M, Bellisle F, SempW M, Guilloud- Bataille M, Patois E. Adiposity rebound in children: a simple indicator for predicting obesity. Am J Clin Nutr 1984; 39: Dietz WH. Critical periods in childhood for the development of obesity. Am J Clin Nutr 1994; 59: Williams S, Davie G, Lam F. Predicting BMI in young adults from childhood data using two approaches to modelling adiposity rebound. Int J Obes Relat Metab Disord 1999; 23: Johnson SL, Birch LL. Parents and children s adiposity and eating style. Pediatrics 1994; 94: Safer DL, Agras WS, Bryson S, Hammer LD. Early body mass index and other anthropometric relationships between parents and children. Int J Obes Relat Metab Disord 2001; 25: Faith MS, Rha SS, Neale MC, Allison DB. Evidence for genetic influences on human energy intake: results from a twin study using measured observations. Behavior Genet 1999; 29: Davies PSW, Gregory J, White A. Energy expenditure in children aged 1.5 to 4.5 years: a comparison with current recommendations for energy intake. Eur J Clin Nutr 1995; 49: Cryan J, Johnson RK. Should the current recommendations for energy intake in infants and young children be lowered? Nutr Today 1997; 32: Alexy U, Kersting M, Sichert-Hellert W, Manz F, Schch G. Energy intake and growth of 3- to 36- month- old German infants and children. Ann Nutr Metab 1998; 42: Davies PSW. Energy requirements for growth and development in infancy. Am J Clin Nutr 1998; 68: 939S 943S. 25 Goran MI, Carpenter WH, Poehlman ET. Total energy expenditure in 4- to 6-yr-old children. Am J Physiol (Endocrinol Metab 27) 1993; 264: E706 E Pietrobelli A, Faith MS, Allison DB, Gallagher D, Chiumello G, Heymsfield SB. Body mass index as a measure of adiposity among children and adolescents: a validation study. J Pediatr 1998; 132: Dietz WH, Robinson TN. Use of the body mass index (BMI) as a measure of overweight in children and adolescent. J Pediatr 1998; 132: Carruth BR, Nevling W, Skinner JD. Developmental and food profiles of infants born to adolescent and adult mothers. J Adol Health 1997; 20: Skinner JD, Carruth BR, Houck K, Coletta F, Cotter R, Ott D, McLeod M. Longitudinal study of nutrient and food intakes of infants aged 2 to 24 months. J Am Diet Assoc 1997; 97: United States Department of Health Education and Welfare Public Health Service, Center for Disease Control. Evaluation of body size and physical growth 1 4. United States Government Printing Office: Washington, DC; Department of Health and Human Services, United States Department of Agriculture. WIC manual B22-B31. United States Government Printing Office: Washington, DC; CDC. A SAS program for the CDC growth charts. US Dept of Health and Human Services, Centers for Disease Control and Prevention, Division of Nutrition and Physical Activity: Atlanta, GA. Website sas.html Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, Flegal KM, Guo SS, Wei R, Mei Z, Curtin LR, Roche AF, Johnson CL. CDC growth charts: United States. In: Advance data from vital and health statistics, no National Center for Health Statistics: Hyattsville, MD; Dorosty AR, Emmett PM, Cowin IS, Reilly JJ. ALSPAC Study Team. Factors associated with early adiposity rebound. Pediatrics 2000; 105: National Academy of Sciences. Dietary reference intakes for energy, carbohydrates, fiber, fat, protein and amino acids (macronutrients). Institute of Medicine of the National Academies. The National Academies Press, Washington, DC; 2002, 36 Food and Nutrition Board. Recommended dietary allowances, 10th edn. National Academy Press: Washington, DC; Carruth BR, Skinner JD. Revisiting the picky eater phenomenon: neophobic behaviors of young children. J Am Coll Nutr 2000; 19: Skinner JD, Carruth BR, Bounds W, Ziegler PJ. Children s food preferences F a longitudinal analysis. J Am Diet Assoc 2002; 102: Skinner JD, Carruth BR, Houck K, Bounds W, Morris M, Cox DR, Moran III J, Coletta F. Longitudinal study of nutrient and food intakes of white preschool children aged 24 to 60 months. JAm Diet Assoc 1999; 99: Carruth BR, Skinner JD, Houck KS, Moran III JD. Addition of supplementary foods and infant growth (2 to 24 months). JAm Coll Nutr 2000; 19: Hill RJ, Davies PSW. The validity of self-reported energy intake as determined using the doubly labeled water technique. Br J Nutr 2001; 85: