PEDIATRIC HIGHLIGHT Anthropometric relationships ... - Nature

International Journal of Obesity (2005) 29, 1222–1229 & 2005 Nature Publishing Group All rights reserved 0307-0565/05 $30.00 www.nature.com/ijo

PEDIATRIC HIGHLIGHT Anthropometric relationships between parents and children throughout childhood: the Fleurbaix–Laventie Ville Sante´ Study B Heude1*, A Kettaneh1, R Rakotovao1, JL Bresson2, JM Borys3, P Ducimetie`re1, MA Charles1 and the Fleurbaix–Laventie Ville Sante´ Group3 1

Institut National de la Sante´ Et de la Recherche Me´dicale (INSERM), Unite´ 258-IFR69, Faculte´ de Me´decine Paris Sud, Villejuif Cedex, France; 2INSERM CIC930 and Universite´ Paris V, Hoˆpital Necker – Enfants Malades, Paris; and 3Fleurbaix Laventie Ville Sante´ Association, Laventie, France BACKGROUND: The study of parent–child anthropometric relationships and their evolution over time may help to better understand familial risk factors for childhood obesity. METHODS: In a population-based cohort of 124 nuclear families (Fleurbaix–Laventie Ville Sante´ Study (FLVS) I and II), various anthropometric parameters were measured in both parents and their children, first when the children were prepubescent and again at the end of puberty. Troncular adiposity repartition was estimated by calculating troncular to peripheral skinfolds ratio and waist-to-hip circumferences ratio. Birth and infancy heights and weights were also obtained from the children’s health booklets. Parent–child correlations were estimated in infancy, before and at the end of the child’s puberty. A prospective analysis was performed to predict the changes in the children’s measurements over puberty according to their parents’ corresponding baseline values. RESULTS: BMI and weight correlations at birth were high (40.30) with the mother and low (o0.10) with the father, then they converged to an intermediate level at 2 y and remained between 0.2 and 0.3 thereafter. Correlations for waist circumference were already present at the prepubertal period and persisted on the same level at the postpubertal period, whereas correlations for subcutaneous adiposity – measured by four skinfolds – and for adiposity distribution – measured by ratios – were higher at the postpubertal period. Moreover, the prospective approach showed that mother’s BMI predicted the evolution of her children’s BMI over puberty, whereas this relationship was observed more specifically with the father concerning adiposity distribution parameters. CONCLUSION: Maternal adiposity may act early in life on the adiposity of the child. Maternal and paternal adiposities seem to have quite distinct effects at two key periods of the child’s adiposity development such as the prenatal and pubertal periods. International Journal of Obesity (2005) 29, 1222–1229. doi:10.1038/sj.ijo.0802920; published online 29 March 2005 Keywords: birth weight; body fat; children; familial correlation; growth; puberty

Introduction Obesity in children is becoming a major stake for public health, first because of immediate school and psychosocial repercussions,1 but above all because of its long-term consequences on adult morbidity.2 Parental obesity has been identified as a predominant risk factor for childhood over-

*Correspondence: Dr B Heude, Institut National de la Sante´ Et de la Recherche Me´dicale (INSERM), Unite´ 258-IFR69, Faculte´ de Me´decine Paris Sud, 16, Avenue Paul Vaillant-Couturier, 94807 Villejuif Cedex, France. E-mail: [email protected] Received 26 July 2004; revised 15 December 2004; accepted 27 December 2004; published online 29 March 2005

weight and obesity.3 Parents and children share both genes and familial environment. The heritability for BMI in children and adults has been established by a large number of studies to be around 40%.4 The obesity epidemic of the last decades suggests, however, that environmental factors are strongly implicated in this phenomenon. The study of parent–child adiposity relationships and their evolution over time in childhood may help to better understand familial risk factors for childhood obesity. As Stunkard et al5 previously underlined in a review of the literature, a relationship between mothers’ and offsprings’ weights is often found at birth and correlations vary within a narrow range (0.2–0.3). After birth, there are discordant conclusions whether a relationship exists before 3 y between

Parent–offspring relationships of adiposity throughout childhood B Heude et al

1223 parents’ (fathers included) and offsprings’ BMI.5–10 In a recent study reporting serial parent–offspring BMI correlations from birth to 8 y, it was shown that correlations between maternal and paternal BMI and their offsprings first emerged at the age of 7 y and were around 0.25 kg/m2.10 BMI is the measurement generally used for familial correlations studies, whereas it is well established that, after puberty particularly, BMI measures relatively more lean mass than fat mass in boys compared with girls.11 To our knowledge, only one study examined serial correlations between parents and children for various fatness measurements, including skinfolds thicknesses, from infancy to late adolescence,12 but it concerned children who were born in the 1950s, before the current childhood obesity epidemic. It showed that mother– child correlations for fatness were generally higher than father–child correlations, and that mother–child correlations tended to increase with time up to mid-adolescence. Actually, little is known about the role of puberty in the evolution of the parent–child resemblance. In this context, our purpose was to investigate crosssectional and prospective relationships between children’s and parents’ various adiposity measurements at selected periods: in infancy, before puberty and at the end of puberty.

Material and methods Population Fleurbaix and Laventie were two neighbouring towns in northern France with 2488 and 4426 inhabitants, respectively, in 1992, when the Fleurbaix–Laventie Ville Sante´ Study (FLVS) took place. The first part of the study, FLVS I, was a 5-y follow-up of children involved in a nutritional education programme at school. In both towns, in 1992–1993, all children from the last section of nursery school (from 5 to 6 y of age) to the last section of primary school (10 to 11 y of age) and their families were asked to participate.13 From October 1992 to June 1993, 826 children aged 5–11 y from 579 families were examined in school by two school doctors. Their parents were examined once between 1992 and 1997 by their general practitioners. The second part, FLVS II, is an epidemiological study whose purpose was to explore the risk factors for weight gain and change in adiposity in children and adults. The FLVS II Study was proposed to every family who had participated in the FLVS I Study. Between April and September 1999, we recruited and examined 294 families (1113 subjects) among the 393 families still living in the two towns who could be contacted (ie 75%). All family members aged over 8 y were examined at home in 1999 and again 2 y later in 2001 (271 remaining families), by one of the seven study physicians. Written consent was obtained from parents in each study; the FLVS I Study was approved by the Ethical Committee of Lens (France) and the FLVS II Study was approved by the Ethical Committee of Lille (France). For this paper, the analysis was performed on children whose parents had both been examined during the FLVS I

Study and at the second FLVS II examination in 2001. Among them, children who were prepubescent (Tanner stage ¼ 1) in 1992 and were at the end of puberty (Tanner stage ¼ 4 or 5) in 2001 were selected. We ended up with 124 families, comprising 77 girls and 90 boys. In all, 31 families had two children, and six families three children.

Measurements Anthropometric data were collected by trained physicians (FLVS II). Weight was measured in light clothes to the nearest 0.1 kg, and height to the nearest cm. Waist and hip circumferences were measured to the nearest 0.5 cm. Waist circumference was defined as the smallest diameter between iliac crest and the last rib, and hip circumference as the largest diameter over the buttocks. Four skinfold thicknesses were measured to the nearest 0.1 mm, in triplicate in FLVS I and in duplicate in FLVS II, using Harpenden callipers, on the left side of the body: tricipital (posterior aspect of the arm, at midpoint between the acromion and the olecranon), bicipital (anterior aspect of the arm, at midpoint between the acromion and the olecranon), subscapular (1 cm below the inferior angle of the scapula), supra-iliac (1 cm over the iliac crest, at the midaxillary line). The average value of each measurement was used, and then the sum of skinfolds, the ratio of troncular (subscapular þ supra-iliac) to peripheral (triceps þ biceps) (TPR), the waist-to-hip (WHR) ratio and the body mass index (BMI ¼ weight in kg/height2 in m2) were calculated. Pubertal stage was determined according to the Tanner classification. The IOTF gender- and age-specific BMI cutoffs were used to define overweight and obesity in children.14 In parents, overweight and obesity were defined as BMI Z25 and 30 kg/m2, respectively. Moreover, the child’s health record booklet was reviewed and weight and height at birth, 9 months and 24 months were recorded for each child, when available (123 children, 65 boys and 58 girls).

Statistical analysis Comparisons between the subjects selected or not selected for the analysis used t-tests for continuous traits, and w2 or Fisher’s exact tests for qualitative data. The significance of the changes in measurements between baseline and followup was tested by paired t-tests for continuous traits. Scheme for the analysis of familial associations. For the analysis, the participants were grouped as mothers, fathers and offsprings. First, correlations between spouses and between parents and offsprings were estimated cross-sectionally at the preand postpubertal periods of the children. Whenever available, the child’s height, weight and BMI at birth and in infancy were also correlated with the corresponding parental parameters measured at their first examination in the FLVS I Study. In a second analysis, correlations between changes over time, between spouses and between parents and International Journal of Obesity


1224 offsprings were computed. Changes were assessed for parents and children by annual variation rates, that is, (Parameter T2Parameter T1)/(Age T2Age T1). Finally, a prospective analysis aimed at predicting the changes in the children’s adiposity measurements over puberty, by their parents’ corresponding measurements at baseline. Model for familial correlations. For the computation of correlations between spouses and between parents and offspring, we used the method of Donner and Koval, which consists in a multivariate analysis of familial data.15 Adjustment for age was performed, by introducing regression terms in the model, specific for fathers, mothers, daughters and sons. Additional adjustment for height was performed for waist and hip circumferences. The hypothesis of equality of correlations was tested using the likelihood ratio test between nested models. The models testing different parent–offspring correlations for boys and girls were never significantly different from the models assuming the equality of correlations; therefore, the results are presented without making this distinction. Model for the prospective analysis. For the prediction of the changes in children’s adiposity measurements over puberty, a linear regression model was used to explain the child’s follow-up variable by the corresponding baseline parental variable adjusted for the baseline child’s value, as suggested by several authors.16,17 To control for effects of age (and height, for waist and hip circumferences), a linear regression on age (and height) was first performed for all variables, separately by gender; the residuals were transformed to

Table 1

standardized Z-scores for each group and were used as the variable of interest for the analysis. A random familial effect was introduced at the level of the intercept in the linear model, to take into account the remaining familial correlations between children. In FLVS II, all members of the same family were examined by the same investigator; thus, analyses of the parameters measured in 2001 were also adjusted for the investigator in order to avoid overestimation of familial correlations. For the familial correlation model, we used a programme developed in our laboratory, which used the GEMINI maximization procedure.18 We used SAS software, version 8.2 (SAS Institute Inc., Cary, NC, USA), for all other statistical analyses.

Results Descriptive characteristics Children selected for the present analysis had a lower mean BMI (P ¼ 0.06) with a lower prevalence of obesity or overweight (P ¼ 0.04) than those unselected. There was no significant difference between mean characteristics of parents selected or not for the analysis (results not shown). At baseline in 1992, selected children were aged between 5 and 13 y, mothers between 30 and 48 y and fathers between 32 and 51 y. In all, 6% of children were overweight at baseline and 9% over follow-up (Table 1). As expected, children presented a significant increase in all measurements, except for WHR in boys (P ¼ 0.76). In girls, the WHR decreased during this period, with hip circumference increasing relatively more than waist circumference (Po0.0001).

Age, Tanner stage and anthropometric characteristics of girls and boys at baseline (1992) and follow-up (2001) (the FLVS Study) Girls, N ¼ 77 1992

Age (y) Weight (kg) Height (cm) BMI (kg/m2) Waist circumference (cm) Hip circumference (cm) WHR Peripheral skinfolds (mm) Troncular skinfolds (mm) Sum of skinfolds (mm) TPR Weight status (N (%)) Overweight Obesity Tanner stage 1 4 5

7.8 25.3 126.8 15.5 54.9 66.5 0.83 17.5 13.8 31.4 0.79

(1.5) (6.1) (10.3) (1.8) (4.7) (7.0) (0.05) (6.2) (6.2) (11.8) (0.16)

Boys, N ¼ 90 2001 16.2 54.1 164.4 20.0 67.1 86.4 0.78 23.1 27.3 50.3 1.21

(1.5) (7.6) (5.7) (2.2) (5.3) (8.0) (0.05) (7.4)w (10.3) (15.7) (0.38)

1992 8.2 27.3 131.1 15.7 57.7 67.0 0.86 14.9 11.6 26.6 0.79

(1.6) (5.8) (9.5) (1.7) (4.1) (5.7) (0.04) (4.6) (3.7) (7.8) (0.15)

16.7 64.9 176.7 20.8 74.6 87.0 0.86 16.7 23.5 40.2 1.43

(1.6) (11.2) (7.2) (3.2) (7.4) (8.6) (0.05)* (10.4)w (16.3) (25.4) (0.42)

5 (6%) 1 (1%)

6 (8%)

4 (4%)

8 (9%) 2 (2%)

77 (100%) F F

F 27 (35%) 50 (65%)

90 (100%) F F

F 37 (41%) 53 (59%)

Values are mean (s.d.). *,w2001 values were all significantly different from 1992 values except for *P ¼ 0.76 and wP ¼ 0.06.

International Journal of Obesity

2001


1225 The number of overweight or obese parents increased between 1992–1997 and 2001, the mean increase in BMI was 1.1 kg/m2 in mothers (Po0.001) and 0.8 kg/m2 in fathers (Po0.001) (Table 2). Mothers had a significant increase in all anthropometric measurements. In fathers, the increase in adiposity was essentially troncular (on average, þ 2.5 cm for waist circumference, Po0.0001, and þ 4 cm for troncular skinfolds, Po0.001), and peripheral skinfold thicknesses did not change over 9 y. Therefore, the mean value of the TPR increased in fathers (Po0.02) and did not change in mothers (P ¼ 0.99).

Table 3. For BMI, waist and hip circumferences, correlations between parents and their children were positive and significantly different from zero (resp. 0.30, 0.28, 0.29, Po0.001). Peripheral skinfolds were correlated between

0.6

Height

0.5 0.4 0.3 0.2 0.1

** †‡ **

*** ***

*** ***

*** ***

*** ***

0.0 Birth

Intrafamily correlation study The evolution, from birth to the end of puberty, of the correlations between parental height, BMI and weight and their offspring’s corresponding measurements is presented in Figure 1. At each age, the resemblance in height between parents and children was the same for mothers or fathers. A strong correlation was established by the age of 2 y (resp. r ¼ 0.48, Po0.001; r ¼ 0.50, Po0.001) and remained stable thereafter. Correlation between mothers’ and children’s weight was high at birth (r ¼ 0.35, Po0.001), and decreased thereafter reaching r ¼ 0.19 (P ¼ 0.05) at 24 months. Correlation between fathers’ and children’s weight was low at birth (r ¼ 0.07, P ¼ 0.48), and increased thereafter. At birth, the difference between mother–child and father–child correlations for weight was significant (P ¼ 0.02). For BMI, correlations were stronger with mothers than with fathers at each age, but this difference was significant only at birth (P ¼ 0.02), when there was a strong correlation between mothers and children (r ¼ 0.34, Po0.001) and no correlation between fathers and children (r ¼ 0.04, P ¼ 0.65). The cross-sectional correlation analysis at the child’s prepubertal period for all measurements is presented in

Table 2

0.6

Weight

0.5

†

ns *** ‡

9 months

24 months

Pre-Pubertal

Post-Pubertal

*** ***

** **

ns *

* *

9 months

24 months

Pre-Pubertal

Post-Pubertal

ns **

*

* ***

* **

9 months

24 months

Pre-Pubertal

Post-Pubertal

0.4 0.3 @

0.2 0.1 0.0 Birth 0.6 0.5

ns † *** ‡

*

BMI

0.4 0.3 0.2

@

0.1 0.0 Birth

Figure 1 Evolution of correlations between fathers’ and offspring’s parameters (plain lines) and between mothers’ and offspring’s parameters (hatch lines) from birth to the end of puberty of the child (N ¼ 123 children). w,z P-value for significance of father–offspring and mother–offspring correlation as follows: nsNon significant, *Po0.05, **Po0.01, ***Po0.001. @The difference between mother or father correlation was significant, P ¼ 0.02.

Characteristics of mothers and fathers at baseline (1992) and follow-up (2001) (the FLVS Study) Mothers, N ¼ 124 1992–1997

Age (y) Weight (kg) Height (cm) BMI (kg/m2) Waist circumference (cm) Hip circumference (cm) WHR Peripheral skinfolds (mm) Troncular skinfolds (mm) Sum of skinfolds (mm) TPR Weight status (N (%)) Overweight Obesity

37.5 62.3 162.8 23.5 76.5 95.2 0.80 30.9 33.7 64.6 1.12

(4.1) (10.6) (6.1) (4.1) (10.9) (9.1) (0.07) (12.7) (16.8) (27.1) (0.43)

25 (20%) 8 (6%)

Fathers, N ¼ 124 2001

44.1 65.3 163.0 24.6 79.2 96.9 0.82 35.7 39.9 75.6 1.11

(4.0) (11.7)*** (6.2) (4.5)*** (11.4)*** (10.8)** (0.05)* (14.7)*** (22.0)*** (33.9)*** (0.41)

33 (27%) 14 (11%)

1992–1997 39.5 77.9 176.1 25.1 89.0 95.9 0.93 19.2 38.1 57.3 2.06

(4.7) (12.1) (6.4) (3.4) (10.4) (7.0) (0.07) (9.3) (16.0) (23.1) (0.61)

54 (44%) 10 (8%)

2001 46.0 80.4 176.0 25.9 91.7 96.6 0.95 19.1 41.8 60.7 2.24

(4.5) (12.2)*** (6.3) (3.5)*** (10.2)*** (7.5) (0.06)*** (8.1) (20.3)* (25.8)* (0.76)*

53 (43%) 21 (17%)

Values are mean (s.d.). *Po0.05, **Po0.01, ***Po0.001.



1226 Table 3 Correlation coefficients between spouses and between parents and offspring for anthropometric measurements, adjusted for age, at the child’s prepubescent period (167 children from 124 nuclear families)

Table 5 Correlation coefficients between spouses and between parents and offspring for the changes in anthropometric measurements between baseline and follow-up (167 children from 124 nuclear families)

Parameter

Between spouses

Parent– child

Father– child

Mother– child

Parameter

BMI (kg/m2) Waist circumference (cm)a Hip circumference (cm)a Peripheral skinfolds (mm) Troncular skinfolds (mm) TPR WHR

0.14 0.18* 0.18* 0.07 0.04 0.08 0.24**

0.30*** 0.28*** 0.29*** 0.21*** 0.14** 0.12* 0.10

0.27** 0.28*** 0.24** 0.18* 0.14 0.11 0.13

0.33*** 0.28*** 0.33*** 0.23** 0.15 0.13 0.06

BMI (kg/m2) Waist circumference (cm)a Hip circumference (cm)a Peripheral skinfolds (mm) Troncular skinfolds (mm) TPR WHR

Between spouses

Parent– child

Father– child

Mother– child

0.21* 0.20* 0.17 0.06 0.29** 0.26** 0.20

0.14* 0.07 0.07 0.01 0.11 0.02 0.10

0.19* 0.05 0.09 0.12 0.07 0.04 0.10

0.10 0.08 0.05 0.10 0.14 0.07 0.00

*Po0.05, **Po0.01, ***Po0.001. aAdditionally adjusted for height.

*Po0.05, **Po0.01, ***Po0.001. aAdditionally adjusted for height.

Table 4 Correlation coefficients between spouses and between parents and offspring for anthropometric measurements, adjusted for age and physician, at the end of the child’s puberty (167 children from 124 nuclear families)

Table 6 Prospective effect of parental parameters on the changes of the corresponding parameter of their offspring from pre- to postpubescent periods (linear regression model of mother’s or father’s parameter on the child’s parameter at follow-up, adjusted for the child’s parameter at baseline, and on the physician at follow-up)

Parameter BMI (kg/m2) Waist circumference (cm)a Hip circumference (cm)a Peripheral skinfolds (mm) Troncular skinfolds (mm) TPR WHRb

Between spouses

Parent– child

Father– child

Mother– child

0.11 0.16 0.07 0.07 0.12 0.17 0.16

0.32*** 0.29*** 0.21** 0.28*** 0.26*** 0.15* 0.17**

0.30*** 0.31*** 0.17* 0.18* 0.17* 0.20* 0.28**

0.34*** 0.27** 0.25** 0.36*** 0.34*** 0.10 0.07

Mother–offspring

*Po0.05, **Po0.01, ***Po0.001. aAdditionally adjusted for height. bDifference between mother–child and father–child correlations was at the limit of significance (P ¼ 0.05).

a

b (s.d.) Height (cm) BMI (kg/m2) Waist circumference (cm)b Hip circumference (cm)b Peripheral skinfolds (mm) Troncular skinfolds (mm) TPR WHR

0.15 0.18 0.09 0.09 0.26 0.25 0.04 0.04

(0.04) (0.06) (0.06) (0.06) (0.07) (0.07) (0.07) (0.06)

P-value 0.0017 0.0065 0.16 0.16 0.0008c 0.0007 0.62 0.52

Father–Offspring b (s.d.) 0.14 0.09 0.08 0.04 0.17 0.08 0.26 0.14

(0.05) (0.06) (0.07) (0.06) (0.07) (0.07) (0.07) (0.06)

P-value 0.0049 0.13 0.20 0.51 0.0262 0.23d 0.0009 0.0387

a

parents and children (0.21, Po0.001). Among the anthropometric measurements, the child’s troncular skinfolds and TPR showed weak correlations with those of their parents (resp. 0.14, P ¼ 0.009; 0.12, P ¼ 0.04). There was no significant correlation between parents’ and offspring’s WHRs (0.10, P ¼ 0.15). Correlations with mothers were not statistically different from those with fathers. Cross-sectional correlations at the end of puberty are presented in Table 4. The patterns of correlations for BMI and waist circumference were similar to those at the prepubertal period. On the other hand, peripheral and troncular skinfolds measurements were more strongly correlated with those of their parents than at the prepubertal period (resp. 0.28, Po0.001; 0.26, Po0.001). For these parameters, correlations with mothers were higher than those with fathers, but the comparison did not reach statistical significance (P ¼ 0.13 for peripheral skinfolds and P ¼ 0.10 for troncular skinfolds). Correlations between fathers’ and offspring’s TPRs and WHRs were notably higher than at the prepubertal period (resp. 0.20, P ¼ 0.013; 0.28, P ¼ 0.001). The difference between mother– child and father–child correlations was at the limit of significance for WHR (P ¼ 0.05). There was no significant correlation between the annual rates of change in the parents’ and their offspring’s measurements, except for BMI (r ¼ 0.14, P ¼ 0.027) (Table 5). International Journal of Obesity

Coefficient age adjusted and standardized by using Z-scores of the residuals of the regression on age as dependent variable. bAdditionally adjusted for height. c,dDifference between boys and girls was significant or at the limit of significance 1. P ¼ 0.056 and 2. P ¼ 0.035, in the direction of a stronger association with girls.

Prospective study Mothers’ and fathers’ heights were positively associated with their children’s height increase over puberty (resp. P ¼ 0.001, P ¼ 0.002) (Table 6). Mothers’ BMI, peripheral and troncular skinfolds at baseline were positively and significantly associated with the change in the children’s corresponding measurements over puberty (resp. P ¼ 0.006, P ¼ 0.0008, P ¼ 0.007). The relationship for peripheral skinfolds was stronger in girls than in boys (P-value for interaction ¼ 0.056). Fathers’ subcutaneous peripheral adiposity at baseline was also positively associated with the change in their children’s corresponding parameters, and again, the relationship was stronger for girls but not significantly (P-value for interaction ¼ 0.30). Fathers’ subcutaneous troncular adiposity was not significantly associated with the change in their children’s corresponding measurements. There was a significant interaction with gender (P ¼ 0.035), showing an association with girls (P ¼ 0.062) but not with boys (P ¼ 0.53). Notably, fathers’, but not mothers’, TPR was significantly associated to the change in TPR of their


1227 children during puberty (resp. b ¼ 0.26, P ¼ 0.0009; b ¼ 0.04, P ¼ 0.62). The same pattern was observed for WHR (resp. b ¼ 0.14, P ¼ 0.039; b ¼ 0.04, P ¼ 0.52).

Between-spouse correlations At the first exam performed in parents, between 1992 and 1997, there were significant correlations between spouses for height (r ¼ 0.40, Po0.001), waist and hip circumferences (resp. 0.28, Po0.05; 0.18, Po0.05) and WHR (r ¼ 0.24, Po0.01) (Table 3). At the second exam, performed in 2001, correlations between spouses were similar to those of the first exam for height, BMI and waist circumference. Between-spouse correlations for troncular skinfolds and TPR were higher than at the first exam, but still not significant (Table 4). Significant between-spouse correlations of the rate of changes in the different variables between 1992–1997 and 2001 were found for BMI, waist circumference, troncular skinfold and TPR but not for peripheral skinfold variation (Table 5).

Discussion The main results of this study were that parent–offspring correlations for global adiposity measurements were already present before the puberty period, whereas correlations for subcutaneous adiposity and fat distribution were higher at the end of puberty. Moreover, mothers’ BMI and weight were specially associated with those of their children at birth. Children selected for the analysis were less overweight and had a lower BMI than the others. Hence, our results are not explained by severe child obesity and apply to adiposity values commonly found in a general population. One limitation of this study is the relatively small number of subjects that may have limited our ability to detect differences between genders in children and between fathers’ and mothers’ correlations. Parent–child correlations for height were similar between fathers and mothers at any time during childhood and adolescence. They were low at birth, increased and reached their maximum during the second year and remained stable thereafter. Several former studies had explored parent–child correlations for height at birth or in early infancy.6,19–21 Hewitt20 found that parent–child correlations for height were, like in our study, around 0.40 at 2 y and remained stable until 5 y, whereas Ounsted et al21 found that there was a gradual rise in correlations for height for both mothers and fathers between birth and 4 y. It is thought that parent–child resemblance in height is principally due to genetic factors. Even if results have to be interpreted with caution because of high measurement error for height at birth and early ages, our pattern of correlations during infancy suggests that the genetic effect is not fully expressed until the second year of life.

Several other studies have examined parent–offspring correlations in BMI at different periods of childhood, crosssectionally or longitudinally.5,6,10,22 Safer et al10 examined the correlation between BMI in parents and their biological offspring from birth to age 8 and found the emergence of statistically significant correlations at the age of 7 y – after the adiposity rebound – even if correlations with mothers’ BMI seemed to appear sooner. Lake et al22 assessed relationships between the adiposity of children and their parents from childhood to adulthood. They showed that mothers– offspring correlations in BMI reached their maximum when children were 11 y (rE0.23) and father–offspring correlations when children were 16 y (rE0.20). In our study, BMI correlations were high at birth with the mother, decreased in the first years, probably until school age and did not change much between before and after puberty. Correlations with the father, quite low at birth, reached substantial levels at school age and remained similar during puberty. In the same way, in their study, Stunkard et al5 found that the relationship between the body weights of mothers and their offspring may be weaker during the first 2 y of life than it is at birth. We can hypothesize that the first years of life may then represent a transition period characterized by the decrease in the prenatal environmental influences and the emergence of the genetic and shared environmental parental influences. Adiposity measurements like BMI, height-adjusted waist and hip circumferences are complex phenotypes representing the amount of fat mass, lean mass and body build. Parent–child correlations for these parameters exist at the prepubertal period and do not change much with puberty. They may be due to a familial resemblance in lean mass as well as in fat mass. Indeed, results from Ferrari et al’s23 study suggested that familial resemblance for most bone traits is already present between daughters and their mothers before puberty. Moreover, Treuth et al24 found that familial resemblance of parental obesity and leanness was clearly evident in daughters at prepubertal ages. Both regional fat tissue and fat-free tissue in the offspring at prepubertal ages were influenced by the body composition of the parents. Our results on the pre- and postpubertal periods suggest that familial factors affecting the development of subcutaneous adiposity and troncular adiposity strengthen during puberty, unlike those for the overall corpulence parameters. A literature review already underlined that there was a familial influence on the pattern of subcutaneous fat distribution in adulthood.4 Our results suggest that these familial resemblances would be established only after puberty. Another longitudinal study examined parental influences on children’s overall fatness and distribution of subcutaneous fat and showed that parent–child correlations for fat patterning components varied more with time than those for overall fatness.12 Several phenomena could account for this pattern of correlations. First, a cumulative effect of a shared familial environment could strengthen the parent–child correlation International Journal of Obesity


1228 over time. Second, with puberty, additional genetic factors may be involved in hormone secretions related to subcutaneous adiposity development and then participate in these stronger parent–child correlations in post- than prepubertal period. In our study, only change in children’s BMI was correlated with the corresponding changes of their parents. Obviously, this is a very complex association because of the intrinsic growth of the child during puberty, which may be due to genetic or physiological factors; their effects may predominate over those of environmental factors shared in the family. However, this association with BMI can be underlined as well with the strong significant correlations between almost all spouses’ adiposity parameters, and even between their respective evolutions over 9 y. Cross-sectional resemblances between spouses could be due to assortative mating phenomena, which involve genes and environmental factors indistinguishably.25 However, longitudinal results suggest an important contribution of a common environment to the familial aggregation of fatness. Hunt et al26 also showed that changes in BMI and adiposity among Canadian families significantly aggregated within families. Important factors that parents and children share are food intake and eating behaviour. Perusse et al27 indicated that familial aggregation was present in energy intake. The home environmental effects may partly account for the family resemblance in energy intake. In another cross-sectional study, Nguyen et al28 suggested that mothers might contribute to the development of obesity in children by influencing their dietary fat intake. Another environmental factor that parents and children could share is physical activity and sedentary behaviour. There are few studies on that topic, but Simonen et al29 in the Quebec Family Study showed that physical activity level was indeed characterized by a significant degree of familial resemblance. The pattern of familial correlations they obtained suggested that shared familial environmental factors were as important as genetic factors in accounting for the familial resemblance in physical activity. Some data suggest that the parent–child fatness relationships are stronger with mothers than fathers, and that the mother–offspring relationship strengthens as the child gets older.12,22 Indeed, our results suggested a specific effect of mothers in the familial resemblance in subcutaneous adiposity, especially at the end of puberty. Larger correlations with mothers or fathers later in life are suggestive of environmental factors predominantly shared with one or the other parent, and are implicated, respectively, in the subcutaneous fat deposit or troncular distribution of adiposity. However, except for BMI, we did not observe any correlation in the change in adiposity between mothers and children, which could have confirmed this hypothesis. Nevertheless, a remaining effect of prenatal environment could participate in the predominant effect of the global adiposity parameters of mothers.30 Our results also suggest a specific role of fathers in the troncular International Journal of Obesity

distribution of adiposity, but these results remain difficult to interpret.

Conclusion In conclusion, our results suggest that an effect of maternal adiposity acts early in life on the adiposity of the child. This maternal effect seems to persist across childhood and to strengthen with puberty, especially for subcutaneous adiposity. Maternal and paternal adiposity seem to have quite distinct effects at two key periods of child’s adiposity development such as the prenatal and pubertal periods.31 References 1 Gortmaker SL, Must A, Perrin JM, Sobol AM, Dietz WH. Social and economic consequences of overweight in adolescence and young adulthood. N Engl J Med 1993; 329: 1008–1012. 2 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 1935. N Engl J Med 1992; 327: 1350–1355. 3 Garn SM, Sullivan TV, Hawthorne VM. Fatness and obesity of the parents of obese individuals. Am J Clin Nutr 1989; 50: 1308–1313. 4 Bouchard C, Pe´russe L, Rice T, Rao DC. The genetics of human obesity. In: Bray GA, Bouchard C, James WPT (eds). Handbook of obesity. Marcel Dekker Inc.: New York; 1998. pp 157–190. 5 Stunkard AJ, Berkowitz RI, Stallings VA, Cater JR. Weights of parents and infants: is there a relationship? Int J Obes Relat Metab Disord 1999; 23: 159–162. 6 Bayley N. Some increasing parent–child similarities during the growth of children. J Educ Psychol 1954; 45: 1–21. 7 Hewitt D, Stewart A. The Oxford Child Health Survey: a study of the influence of social and genetic factors on infant weight. Hum Biol 1952; 24: 309–319. 8 Garn SM, Pesick SD. Relationship between various maternal body mass measures and size of the newborn. Am J Clin Nutr 1982; 36: 664–668. 9 Whitaker RC, Deeks CM, Baughcum AE, Specker BL. The relationship of childhood adiposity to parent body mass index and eating behavior. Obes Res 2000; 8: 234–240. 10 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: 1532–1536. 11 Maynard LM, Wisemandle W, Roche AF, Chumlea WC, Guo SS, Siervogel RM. Childhood body composition in relation to body mass index. Pediatrics 2001; 107: 344–350. 12 Kaplowitz HJ, Wild KA, Mueller WH, Decker M, Tanner JM. Serial and parent–child changes in components of body fat distribution and fatness in children from the London Longitudinal Growth Study, ages two to eighteen years. Hum Biol 1988; 60: 739–758. 13 Maillard G, Charles MA, Lafay L, Thibult N, Vray M, Borys JM, Basdevant A, Eschwege E, Romon M. Macronutrient energy intake and adiposity in non obese prepubertal children aged 5– 11 y (the Fleurbaix Laventie Ville Sante Study). Int J Obes Relat Metab Disord 2000; 24: 1608–1617. 14 Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 2000; 320: 1240–1243. 15 Donner A, Koval JJ. A multivariate analysis of family data. Am J Epidemiol 1981; 114: 149–154. 16 Frison L, Pocock SJ. Repeated measures in clinical trials: analysis using mean summary statistics and its implications for design. Stat Med 1992; 11: 1685–1704. 17 Vickers AJ. The use of percentage change from baseline as an outcome in a controlled trial is statistically inefficient: a simulation study. BMC Med Res Methodol 2001; 1: 6.


1229 18 Lalouel J. GEMINI: a computer program for optimization of general nonlinear functions, Technical Report no. 14 Department of Medical Biophysics and Computing, University of Utah, Salt Lake City; 1979. 19 Thomson J. Height and weight at three years. Health Bull 1955; 13: 16–17. 20 Hewitt D. Some familial correlations in height, weight and skeletal maturity. Ann Hum Genet 1957; 22: 26–35. 21 Ounsted M, Moar V, Scott A. Growth in the first four years: IV. Correlations with parental measures in small-for-dates and largefor-dates babies. Early Hum Dev 1982; 7: 357–366. 22 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: 376–381. 23 Ferrari S, Rizzoli R, Slosman D, Bonjour JP. Familial resemblance for bone mineral mass is expressed before puberty. J Clin Endocrinol Metab 1998; 83: 358–361. 24 Treuth MS, Butte NF, Ellis KJ, Martin LJ, Comuzzie AG. Familial resemblance of body composition in prepubertal girls and their biological parents. Am J Clin Nutr 2001; 74: 529–533.

25 Allison DB, Neale MC, Kezis MI, Alfonso VC, Heshka S, Heymsfield SB. Assortative mating for relative weight: genetic implications. Behav Genet 1996; 26: 103–111. 26 Hunt MS, Katzmarzyk PT, Perusse L, Rice T, Rao DC, Bouchard C. Familial resemblance of 7-year changes in body mass and adiposity. Obes Res 2002; 10: 507–517. 27 Perusse L, Tremblay A, Leblanc C, Cloninger CR, Reich T, Rice J, Bouchard C. Familial resemblance in energy intake: contribution of genetic and environmental factors. Am J Clin Nutr 1988; 47: 629–635. 28 Nguyen VT, Larson DE, Johnson RK, Goran MI. Fat intake and adiposity in children of lean and obese parents. Am J Clin Nutr 1996; 63: 507–513. 29 Simonen RL, Perusse L, Rankinen T, Rice T, Rao DC, Bouchard C. Familial aggregation of physical activity levels in the Quebec Family Study. Med Sci Sports Exerc 2002; 34: 1137–1142. 30 Prentice AM. Intrauterine factors, adiposity, and hyperinsulinaemia. BMJ 2003; 327: 880–881. 31 Dietz WH. Critical periods in childhood for the development of obesity. Am J Clin Nutr 1994; 59: 955–959.
