Tracking of body mass index during childhood: a 15-year prospective population-based family study in eastern Finland

Transcription

1 (2003) 27, & 2003 Nature Publishing Group All rights reserved /03 $ PAPER : a 15-year prospective population-based family study in eastern Finland RM Fuentes 1 *, I-L Notkola 2, S Shemeikka 1, J Tuomilehto 2,3 and A Nissinen 2 1 Department of Public Health and General Practice, Faculty of Medicine, University of Kuopio, Kuopio, Finland; 2 Department of Epidemiology and Health Promotion, National Public Health Institute, Helsinki, Finland; and 3 Department of Public Health, University of Helsinki, Helsinki, Finland OBJECTIVES: To investigate the tracking of body mass index (BMI) during childhood. The effect of birth weight and family history of obesity on BMI development during childhood was also evaluated. METHODS: All children born during in a rural community of eastern Finland were followed at ages 6 months, 7 and 15 y (-6 m, -7y, -15y). Out of 205 children, 138 completed the full follow-up period, of which 100 (45 girls) were included in the analysis with complete data. RESULTS: BMI-6 m was significantly associated with BMI-7y (r ¼ 0.320; P-value ¼ 0.001), but no longer with BMI-15y. BMI-7y was significantly associated with BMI-15y (r ¼ 0.686; P-value o0.001). Children in the highest tertile of BMI-6 m did not have a higher risk of being in the highest tertile of either BMI-7y or BMI-15y compared with children in other tertiles of BMI-6 m. Children in the highest tertile of BMI-7y had a significantly higher risk of being in the highest tertile of BMI-15y (relative risk ¼ 3.6 ( )) compared with children in other tertiles of BMI-7y. BMI-7y was predicted negatively by parents education and male gender and positively by BMI-6 m. BMI-15y was predicted positively by BMI-7y, the difference in BMI between ages 7 y and 6 months and the mean of BMI between ages 6 months and 7 y. Birth weight was not a good predictor of BMI during childhood. Children with at least one obese parent seemed to have higher BMI during childhood; however, this association did not reach a significant level. CONCLUSION: The study confirmed the tracking of BMI during childhood. Neither birth weight nor family history of obesity was found a good predictor of BMI during childhood. The risk of obesity in adolescence can be determined during middle childhood and obese children may be targeted in lifestyle advice to reverse this trend. Parental education may have a key role in the prevention of obesity during childhood. (2003) 27, doi: /sj.ijo Keywords: childhood; birth-cohort; tracking; familial aggregation; birth weight Introduction Adult obesity, measured by relative weight or body mass index (BMI), is a strong risk predictor of cardiovascular diseases (CVD), diabetes mellitus and mortality from all causes, CVD and some cancers. 1 3 Childhood obesity is associated with an increased morbidity because of chronic diseases in adulthood, and it is also associated with increased *Correspondence: Dr RM Fuentes, Department of Public Health and General Practice, University of Kuopio, POB 1627, FIN Kuopio, Finland. ricardo.fuentes@messi.uku.fi Received 17 October 2002; revised 17 December 2002; accepted 2 January 2003 premature mortality, mainly because of coronary heart disease and some cancers. 4 6 Recently, attention has been drawn to the increase of the incidence and prevalence of type II diabetes in children and young adults, and its frequent association with obesity as a predisposing factor. 7 The tracking, or the tendency for an individual to maintain his/her disease risk factor rank level relative to his/her peers through time, has important implications in terms of disease causality determination as well as for the prevention of future-related morbidity and mortality. The evidence of significant tracking of obesity from childhood to adulthood is numerous, but the comparison of studies is complicated by differences in study design, definition of obesity and analytical methods used. 8

2 The Finnish population is a well-characterized genetic isolate 9,10 in which more than 30 recessive diseases, the Finnish disease heritage, have been found more prevalent than in other populations. 11,12 However, when compared to molecular epidemiological research, relatively few epidemiological family studies of adult noncommunicable chronic diseases, including CVD and related risk factors, have been performed in Finland. The evaluation of familial aggregation and tracking of BMI during childhood in a sample of families from a genetic isolate, using a prospective birth-cohort design in which children are followed periodically at the same age from infancy up to adolescence should contribute important information to support the notion of the primordial prevention of obesity. In this study, we report results on tracking of BMI during a 15-year follow-up period in a birth-cohort of 100 children living with both biological parents. The effect of birth weight and family history of obesity on BMI development during childhood was also evaluated. Methods A more detailed description of the study methods has been published previously. 13 The Project From childhood to teenage years is a follow-up study on children s development, health and life situation. It has consisted of periodical cross-sectional surveys of all children born during living in a rural community in eastern Finland. A total of eight cross-sectional surveys have been carried out (at ages of 6 months, 1, 2, 3, 5, 7, 12, and 15 y). For the first survey, the eligible population included 210 children born during living in the community at the age of 6 months. In all, 205 of these families agreed to participate. The last survey was carried out during when children were 15 y old. A total of 138 children completed the full 15-y follow-up period (67%). The parents of these children were evaluated at the survey carried out during , when children were 12 y old. Since BMI measurements taken closely apart are highly correlated, which could affect the tracking analysis using the regression model, only BMI measurements at ages 6 months, 7 and 15 y were considered in this study. For this study, the inclusion criteria were: (1) family informed consent (five families did not give consent during the first survey), (2) children with full data on BMI at ages 6 months, 7 and 15 y (four children had BMI missing at age 7 y, 42 children had BMI missing at age 15 y and 30 children had BMI missing at both follow-up ages), (3) children living with both biological parents (29 children were not living with both biological parents at the age of 12 y). Altogether 100 children fulfilled the inclusion criteria and were included in the tracking analysis of BMI. The study sample represented 48% (100/210) of the eligible population. Using the Pearson correlation as the estimate of tracking we calculated that a sample size of 84 subjects was needed to report significant correlations of r ¼ 0.30 with an a ¼ 0.05 and a power of 80%. The sample size of children in our study conforms to this calculation. We found no statistically significant difference in gender proportion and BMI between included and excluded children at baseline, when they were 6 months old. The participants answered self-administrated health and lifestyle questionnaires during (parents) and (children). Parents education was determined by asking parents to state the number of years of school or full-time life studies, with which the composed variable parents education was created (values from 0, both parents with r7 y of lifetime studies, up to 5, both parents with Z13 y of lifetime studies). The parents were also asked if his/ her mother or father has had acute myocardial infarction (AMI) or stroke before the age of 60 y, with which the composed variable family history of CVD was created (0, no grandparent with history of CVD; 1, at least one grandparent with history of CVD). Since there were no parents with history of premature AMI or stroke in this sample, this variable was not used in the analysis. Weight was measured using a digital scale with an accuracy of 0.1 kg. The child was naked and lying down during the measurement at age 6 months. The child was standing wearing light indoor clothes without shoes during the measurements at ages 7 and 15 y. Height was measured with an accuracy of 0.1 cm. Length was measured at age 6 months using a wooden scale with the child lying down. Height was measured using a stadiometer on bare-footed children at ages 7 and 15 y. Weight and height measurements on parents followed the same methodology described for children at ages 7 and 15 y. Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m). Obesity in parents was defined according to the NHANES II criteria (BMIZ27.8 kg/m 2 in fathers and Z27.3 kg/m 2 in mothers). A child was regarded as having family history of obesity when at least one parent had obesity according to the above criteria. Gestational age and birth weight were obtained from routine obstetrical records. Since all selected children were born full term (Z37 weeks) gestational age was not included in the data analysis. Tracking of BMI during childhood was evaluated by three statistical methods. First, Pearson product-moment correlations were calculated between the child s BMI measured at ages 6 months, 7 and 15 y without adjustment for confounders. Then partial correlations were calculated between the child s BMI measured at ages 6 months, 7 and 15 y controlling for the child s gender, birth weight, parents education, family obesity and family history of CVD. Second, the child s relative risk of being in the highest tertile of the gender-specific BMI distribution at ages 7 and 15 y was calculated according to the child s BMI category at ages 6 months and 7 y, respectively. In this analysis not being in the highest age gender-specific BMI tertile was regarded as the reference category. Third, the main predictors of BMI at ages 6 months, 7 and 15 y were evaluated using linear regression 717

3 718 models. For example, the regression model for the child s BMI at age 15 y included the child s gender, birth weight, parents education, family obesity, family history of CVD and the child s BMI at 6 months or BMI at 7 y or the difference of BMI between ages 7 y and 6 months or the mean of BMI between ages 6 months and 7 y. The independent variables were entered into the above analyses according to the enter method. Results were considered statistically significant with P-value o0.05. All the statistical analyses were carried out using the SPSS program version 10 (SPSS Inc., Chicago, IL, USA). The study was approved by the Ethic Committee of the University of Kuopio and the University Hospital of Kuopio. Results At age 6 months BMI was significantly higher in boys than in girls (Table 1). At ages 7 and 15 y BMI was higher in girls than in boys, but these differences were not significant. From age 6 months to 7 y BMI increased in girls and decreased in boys, and a significant age gender interaction was observed (repeated measures ANOVA, age gender P-value o0.001). Table 1 Body mass index and other variables in a birth-cohort of 100 children and their biological parents selected for tracking analysis Parents Mothers (n ¼ 100) Fathers (n ¼ 100) Variable Mean s.d. Mean s.d. P-value Age (y) o0.001 Education (y) BMI (kg/m 2 ) % % P-value Obesity FCVD Variable Children Girls (n ¼ 45) Boys (n ¼ 55) Mean s.d. Mean s.d. P-value BMI (kg/m 2 ) 6 months (y) (y) Birth weight (kg) % % P-value Family obesity FCVD BMI, body mass index; obesity, BMI Z27.3 kg/m 2 for mothers and BMI Z27.8 kg/m 2 for fathers; family obesity, at least one parent with obesity defined as above; FCVD, family history of CVD, at least one parent (grandparent of the index child) with diagnosis of acute myocardial infarction or stroke. Figure 1 Mean body mass index by age and gender with 95% confidence intervals. NS, not significant. From age 7 15 y BMI tended to increase more steeply in girls than in boys, but no significant age gender interaction was found. From age 6 months to 15 y BMI increased more steeply in girls than in boys with a significant age gender interaction (Figure 1). BMI at 6 months was significantly associated with BMI at 7 y, but no longer with BMI at 15 y (Table 2). BMI at 7 y was significantly associated with BMI at 15 y. These correlations did not change significantly after accounting for confounders. Children in the highest tertile of BMI at age 6 months did not have a higher risk of being in the highest tertile of BMI either at age 7 y or at age 15 y compared with children in other tertiles of BMI at age 6 months (Table 3). Children in the highest tertile of BMI at age 7 y had a significantly higher risk of being in the highest tertile of BMI at age 15 y compared with children in other tertiles of BMI at age 7 y. At age 6 months, male gender was associated positively with BMI (Table 4). At age 7 y, male gender and parents education were associated negatively with BMI and BMI at 6 months was associated positively. At age 7 y, the association between birth weight and BMI was negative, though statistically not significant. At age 15 y, BMI at 7 y, the difference in BMI between ages 7 y and 6 months and the mean of BMI between ages 6 months and 7 y were all associated positively with BMI. At age 15 y the association between birth weight and BMI was again positive, as during infancy, but it did not reach a statistically significant level. Children with a family history of obesity seemed to have higher BMI during childhood. However, this association reached a marginal significance only in the regression model for BMI at age 7 y where the child s gender, parents education and BMI at 6 months were the main predictors. Discussion Longitudinal data on BMI from as early as infancy up to adolescence or young adulthood had been reported

4 Table 2 Correlations of BMI measured longitudinally at ages 6 months, 7 and 15 y in a birth-cohort of 100 children 719 7y 15 y Crude r (P-value) Partial r (P-value) Crude r (P-value) Partial r (P-value) 6 months (0.001) (o0.001) (0.274) (0.278) 7 (y) (o0.001) (o0.001) Crude r, crude Pearson correlation; partial r, partial correlation, BMI controlled for the child s gender, birth weight, parents education, family obesity and family history of cardiovascular disease. Table 3 Relative risk (RR) of being in the high level of BMI during childhood in a birth-cohort of 100 children followed at ages 6 months, 7 and 15 y RR (95% CI) High level-7y High level-15y High level-6m 1.5 ( ) 0.7 ( ) High level-7y 3.6 ( ) (95% CI), 95% confidence interval; high level defined as being at the highest tertile of BMI according to the child s age and gender; not being at the highlevel category is the reference category for the relative risk calculation. Characteristically, BMI rises during infancy, peaking at about 9 months before falling to a nadir at about the age of 6 y, which is followed by a second rise lasting until adulthood (referred to as adiposity rebound). He et al 22 followed, retrospectively, BMI from birth to age 18 y in a populationbased sample of 3650 adolescents born in the early 1970s, in Göteborg, Sweden. In that sample mean values of BMI at ages 6 months, 7 and 15 y for girls were 16.77, and and for boys were 15.76, and 19.85, respectively. The mean values and the evolution of BMI in this sample were similar to those described in the above study. Inferences from the comparison of longitudinal data sets from different populations have to be drawn cautiously. Differences may be due, for example, to differences in sampling methods, statistical transformation of the data and population-specific secular trends in growth and development. 23 The correlation of BMI between ages 6 months and 7 y found in our study was positive, statistically significant and similar to that found by Gasser et al. 18 The lower correlation of BMI observed between ages 6 months and 15 y was expected since it covers a long follow-up period in which both infancy and adolescence were included. Similar low correlation levels of BMI between infancy and adolescence or young adulthood had been reported. 18,21,22 The correlation of BMI between ages 7 and 15 y found in our study was of the same order as those published previously. 18,20 22,24,25 The relative risk analysis demonstrated significant tracking of BMI during childhood only between ages 7 and 15 y, confirming the results of previous reports, 21,26,27 but disagreeing with other reports showing a significant tracking of BMI during childhood from as early as infancy. 14,16,28 The relative risk analysis has the additional advantage of overcoming the regression toward the mean phenomenon observed when working with repeated measurements of continuous variables in prospective studies. 29 The linear regression analysis confirmed the tracking of BMI during childhood and adolescence found by both the correlation and the relative risk analyses and gave important Table 4 Linear regression analysis of BMI measured longitudinally in a birth-cohort of 100 children Independent BMI-6m BMI-7y BMI-15y BMI-15y BMI-15y BMI-15y variables b (P-value) b (P-value) b (P-value) b (P-value) b (P-value) b (P-value) Constant (o0.001) 8.83 (0.004) (0.003) 2.67 (0.476) (o0.001) 2.25 (o0.635) Gender 0.87 (0.018) 1.29 (0.011) 1.38 (0.113) 0.36 (0.594)) 0.36 (0.629) 1.21 (0.103) Birth weight (kg) 0.82 (0.057) 0.11 (0.854) 0.68 (0.503) 0.53 (0.498) 1.31 (0.117) 0.23 (0.794) Parents education 0.12 (0.418) 0.43 (0.033) 0.44 (0.203) 0.02 (0.937) 0.13 (0.658) 0.13 (0.669) Family obesity 0.22 (0.578) 0.96 (0.067) 1.24 (0.171) 0.35 (0.627) 0.23 (0.769) 0.85 (0.287) FCVD 0.13 (0.719) 0.69 (0.163) 1.66 (0.057) 0.93 (0.182) 1.12 (0.128) 1.20 (0.120) BMI-6m (kg/m 2 ) 0.53 (0.001) 0.22 (0.399) BMI-7y (kg/m 2 ) 0.99 (o0.001) (BMI-7y) - (BMI-6m) (kg/m 2 ) 0.91 (o0.001) mean BMI-6m-7y (kg/m 2 ) 1.10 (o0.001) Adjusted R b, unstandardized linear regression beta; mean BMI-6m-7y, mean BMI between ages 6 months and 7 y; Gender (0, female; 1, male); Parents education (values from 0, both parents with o7 y of lifetime studies up to 5, both parents with >13 years of lifetime studies); family obesity (0, no obese parent ; 1, at least one obese parent); FCVD, family history of cardiovascular disease (0, no grandparent with history of CVD; 1, at least one grandparent with history of CVD).

5 720 additional results. First, gender is a good predictor of BMI at ages 6 months and 7 y. This may reflect the association between gender and adiposity during infancy and adiposity rebound during middle childhood, with girls having a lower BMI during infancy and boys having a later adiposity rebound during childhood. The lack of association between gender and BMI at age 15 y in our study most probably is due to the similar BMI level at this age between girls and boys in the sample. Second, birth weight was not a good predictor of BMI during childhood. There are inconsistent reports on the association between birth weight and BMI during childhood, adolescence and young adulthood, with some studies showing no significant association 19,30 and others showing a significant positive association Third, subject s characteristics like gender and previous BMI seem to be relatively more important in predicting BMI during childhood than family history of obesity. In this study, the association between family history of obesity and childhood BMI level has been evaluated at different ages using a multivariate analysis in which the child s gender, birth weight and previous BMI readings have been accounted for in the same model, thus taking full advantage of its prospective design. Differences in sampling and analysis methodologies may explain partially the disagreement with previous crosssectional and longitudinal studies showing a significant positive association between family history of obesity and BMI level in children and adult offspring. 30,34 39 Fourth, family history of CVD was not a good predictor of BMI during childhood. The disagreement with previously published reports from studies performed both during childhood and adulthood most probably is due to methodological differences There is no other study in which such multivariate analyses had been performed. Fifth, parental education had a significant inverse association with BMI at age 7 y, the time of adiposity rebound. Thus, parents education, as a modifiable environmental factor, may have a great value for the prevention of obesity during this critical period of human growth. The use of BMI as an indicator of fatness in childhood has been questioned because of its dependency on age, since the body composition and body frame change rapidly during growth and puberty. However, recently the International Obesity Task Force concluded that BMI is a reasonable measure with which to assess fatness in children and adolescents. 44 The dependency of BMI on age is likely not a source of bias in this study either in the familial aggregation analysis or in the tracking analysis, because a cohort of children was examined at the same age at baseline and at each follow-up. In conclusion, we found significant tracking of BMI during the first 15 y of life, which was more consistently shown from age 7 y onward. Neither birth weight nor family history of obesity was found to be a good predictor of BMI during childhood. The identification of children at risk of developing high BMI during adolescence may consider the child s gender and previous BMI values during childhood. The risk of obesity in adolescence can be determined during middle childhood and obese children may be targeted in lifestyle advice to reverse this trend. Parental education may have a key role in the prevention of obesity during childhood. Acknowledgements This study was funded by the Juho Vainio Foundation, the Finnish Cultural Foundation of Northern Savo, the Finnish Foundation for Cardiovascular Research, the Ministry of Education of Finland and the Academy of Finland. References 1 Pi-Sunyer FX. Health implications of obesity. Am J Clin Nutr 1991; 53: 1595S 1603S. 2 Kushner RF. Body weight and mortality. Nutr Rev 1993; 51: Solomon CG, Manson JE. Obesity and mortality: a review of the epidemiologic data. Am J Clin Nutr 1997; 66: 1044S 10450S. 4 Mossberg H. 40 year follow-up of overweight children. Lancet 1989; 2(8661): Must A, Jacques PF, Dallal GE, Bajema CJ, Dietz WH. Long-term morbidity and mortality of overweight adolescents. A follow-up of the Harvard Growth Study of 1922 to N Engl J Med 1992; 327: Gunnell DJ, Frankel SJ, Nanchahal K, Peters TJ, Smith GD. Childhood obesity and adult cardiovascular mortality: a 57-y follow-up study based on the Boyd Orr cohort. Am J Clin Nutr 1998; 67: American Diabetes Association. Type 2 diabetes in children and adolescents. Pediatrics 2000; 105: Serdula MK, Ivery D, Coates RJ, Freedman DS, Williamson DF, Byers T. Do obese children become obese adults? A review of the literature. Prev Med 1993; 22: Workman PL, Mielke JH, Nevanlinna HR. The genetic structure of Finland. Am J Phys Anthropol 1976; 44: Sajantila A, Salem AH, Savolainen P, Bauer K, Gierig C, Paabo S. Paternal and maternal DNA lineages reveal a bottleneck in the founding of the Finnish population. Proc Natl Acad Sci USA 1996; 93: de la Chapelle A, Wright FA. Linkage disequilibrium mapping in isolated populations: the example of Finland revisited. Proc Natl Acad Sci USA 1998; 95: Peltonen L, Jalanko A, Varilo T. Molecular genetics of the Finnish disease heritage. Hum Mol Genet 1999; 8: Fuentes R, Notkola I-L, Shemeikka S, Tuomilehto J, Nissinen A. Tracking of systolic blood pressure during childhood: a 15-year follow-up population-based family study in eastern Finland. J Hypertens 2002; 20: Rolland-Cachera MF, Deheeger M, Guilloud-Bataille M, Avons P, Patois E, Sempe M. Tracking the development of adiposity from one month of age to adulthood. Ann Hum Biol 1987; 14: Rolland-Cachera MF, Bellisle F, Sempe M. The prediction in boys and girls of the weight/height index and various skinfold measurements in adults: a two-decade follow-up study. Int J Obes Relat Metab Disord 1989; 13: Muramatsu S, Sato Y, Miyao M, Muramatsu T, Ito A. A longitudinal study of obesity in Japan: relationship of body habitus between at birth and at age 17. Int J Obes Relat Metab Disord 1990; 14: Prokopec M, Bellisle F. Adiposity in Czech children followed from 1 month of age to adulthood: analysis of individual BMI patterns. Ann Hum Biol 1993; 20:

6 18 Gasser T, Ziegler P, Seifert B, Molinari L, Largo RH, Prader A. Prediction of adult skinfolds and body mass from infancy through adolescence. Ann Hum Biol 1995; 22: Hulman S, Kushner H, Katz S, Falkner B. Can cardiovascular risk be predicted by newborn, childhood, and adolescent body size? An examination of longitudinal data in urban African Americans. J Pediatr 1998; 132: 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: He Q, Karlberg J. Prediction of adult overweight during the pediatric years. Pediatr Res 1999; 46: He Q, Albertsson-Wikland K, Karlberg J. Population-based body mass index reference values from Goteborg, Sweden: birth to 18 years of age. Acta Paediatr 2000; 89: Livingstone MB. Childhood obesity in Europe: a growing concern. Public Health Nutr 2001; 4: Kelly JL, Stanton WR, McGee R, Silva PA. Tracking relative weight in subjects studied longitudinally from ages 3 to 13 years. J Paediatr Child Health 1992; 28: Power C, Lake JK, Cole TJ. Body mass index and height from childhood to adulthood in the 1958 British birth cohort. Am J Clin Nutr 1997; 66: Stark O, Atkins E, Wolff OH, Douglas JW. Longitudinal study of obesity in the National Survey of Health and Development. Br Med J (Clin Res Ed) 1981; 283: Mamalakis G, Kafatos A, Manios Y, Anagnostopoulou T, Apostolaki I. Obesity indices in a cohort of primary school children in Crete: a six year prospective study. Int J Obes Relat Metab Disord 2000; 24: Fisch RO, Bilek MK, Ulstrom R. Obesity and leanness at birth and their relationship to body habitus in later childhood. Pediatrics 1975; 56: Chuang-Stein C, Tong DM. The impact and implication of regression to the mean on the design and analysis of medical investigations. Stat Methods Med Res 1997; 6: Frisancho AR. Prenatal compared with parental origins of adolescent fatness. Am J Clin Nutr 2000; 72: Seidman DS, Laor A, Gale R, Stevenson DK, Danon YL. A longitudinal study of birth weight and being overweight in late adolescence. Am J Dis Child 1991; 145: Sorensen HT, Sabroe S, Rothman KJ, Gillman M, Fischer P, Sorensen TI. Relation between weight and length at birth and body mass index in young adulthood: cohort study. BMJ 1997; 315: Rasmussen F, Johansson M. The relation of weight, length and ponderal index at birth to body mass index and overweight among 18-year-old males in Sweden. Eur J Epidemiol 1998; 14: Fuentes R, Notkola I-L, Shemeikka S, Tuomilehto J, Nissinen A. Familial aggregation of body mass index: a population-based family study in eastern Finland. Horm Metab Res 2002; 7: Greenlund KJ, Liu K, Dyer AR, Kiefe CI, Burke GL, Yunis C. Body mass index in young adults: associations with parental body size and education in the CARDIA Study. Am J Public Health 1996; 86: Lake JK, Power C, Cole TJ. Child to adult body mass index in the 1958 British birth cohort: associations with parental obesity. Arch Dis Child 1997; 77: Whitaker RC, Wright JA, Pepe MS, Seidel KD, Dietz WH. Predicting obesity in young adulthood from childhood and parental obesity. N Engl J Med 1997; 337: Whitaker RC, Pepe MS, Wright JA, Seidel KD, Dietz WH. Early adiposity rebound and the risk of adult obesity. Pediatrics 1998; 101: E5. 39 Stettler N, Tershakovec AM, Zemel BS, Leonard MB, Boston RC, Katz SH, Stallings VA. Early risk factors for increased adiposity: a cohort study of African American subjects followed from birth to young adulthood. Am J Clin Nutr 2000; 72: Bao W, Srinivasan SR, Wattigney WA, Berenson GS. The relation of parental cardiovascular disease to risk factors in children and young adults. The Bogalusa Heart Study. Circulation 1995; 91: Bao W, Srinivasan SR, Valdez R, Greenlund KJ, Wattigney WA, Berenson GS. Longitudinal changes in cardiovascular risk from childhood to young adulthood in offspring of parents with coronary artery disease: the Bogalusa Heart Study. JAMA 1997; 278: Burke GL, Savage PJ, Sprafka JM, Selby JV, Jacobs Jr DR, Perkins LL, Roseman JM, Hughes GH, Fabsitz RR. Relation of risk factor levels in young adulthood to parental history of disease. The CARDIA study. Circulation 1991; 84: Pereira MA, Schreiner PJ, Pankow JS, Williams RR, Higgins M, Province MA, Rao DC. The Family Risk Score for coronary heart disease: associations with lipids, lipoproteins, and body habitus in a middle-aged bi-racial cohort: The ARIC study. Ann Epidemiol 2000; 10: Dietz WH, Bellizzi MC. Introduction: the use of body mass index to assess obesity in children. Am J Clin Nutr 1999; 70(Part 2): S123 S