Assortative weight gain in mother–daughter and father–son pairs: an emerging source of childhood obesity. Longitudinal study of trios (EarlyBird 43)



To look for same-sex (gender assortative) association of body mass index (BMI) in healthy trios (mother, father and child) from a contemporary birth cohort, which might imply shared environment rather than shared genes because selective mother–daughter and father–son gene transmission is not a common Mendelian trait.


Prospective (longitudinal) cohort study with four annual time points, from 5 to 8 years.


226 healthy trios from a 1995 to 1996 birth cohort randomly selected in the city of Plymouth, UK.


Average BMI of the two parents and maternal/paternal BMI separately related to the BMI-SDS (standard deviation score) of all offspring and to the BMI-SDS of the sons and the daughters separately.


There were big differences in BMI-SDS among the daughters grouped according to mothers' category of BMI (effect size 1.37 SDS), but not their sons (effect size 0.16 SDS, gender interaction P<0.004), and among the sons grouped according to their fathers' BMI (effect size 1.28 SDS), but not their daughters (effect size 0.17, gender interaction P=0.02). Children whose same-sex parents were of normal weight, weighed either close to (girls+0.20 BMI-SDS) or less than (boys,−0.34 BMI-SDS) children of 20 years ago, and did not change from 5 to 8 years. In contrast, the risks of obesity at 8 years were 10-fold greater (girls 41%, P<0.001) or sixfold greater (boys 18%, P<0.05) if the same-sex parent was obese. Longitudinal linear mixed effects (multilevel) modelling showed a marked influence of maternal and paternal BMI on the rate of weight gain, which was unaffected by birth weight of the child. We report perhaps the largest effect sizes so far recorded in childhood obesity.


Childhood obesity today seems to be largely confined to those whose same-sex parents are obese, and the link does not seem to be genetic. Parental obesity, like smoking, might be targeted in the interests of the child.


Obesity is an important threat to health, and is believed to underlie the insulin resistance associated with metabolic disorders such as diabetes, hypertension and heart disease.1 Children have shown the most rapid rise in obesity, but it is not clear why.2 Obesity is multifactorial,3 but it remains uncertain how much of childhood obesity can be attributed to genetic susceptibility and how much to environmental risk.4

Gestation can confound estimates of genetic transmission between mother and child,5 and attempts have been made to assess the influence of gestation on childhood obesity among family trios by comparing the mother–child correlation of body mass (in which gestation could play a role) with the father–child correlation (in which it could not). Results were for a long time conflicting,6, 7, 8, 9 but a recent cross-sectional study of over 4000 such trios from the ALSPAC (Avon Longitudinal Study of Parents and Children) reported no evidence for an independent effect of gestation on childhood body mass index (BMI) at 7 years.10

Autosomal genetic transmission is usually gender unselective, and Mendelian inheritance does not commonly admit selective mother–daughter or father–son transmission of sex-linked traits. Accordingly, gender-assortative weight gain on a population scale, where obesity in the offspring shows linkage between mother–daughter or father–son, but not mother–son or father–daughter, would be more likely to have an environmental than a genetic basis.

In this report, we have sought to separate environmental from genetic factors in childhood obesity by extending the analysis of BMI relationships to gender-assorted pairings of mother–daughter and father–son, comparing them with mother–son and father–daughter. In doing so, we have explored the influence of gestation, birth weight and parental BMI on childhood BMI, and have used linear mixed effects (multilevel) modelling to examine the impact of parental obesity on the rate of weight gain in their offspring over time.

Materials and methods


The EarlyBird study incorporates a 1995–1996 birth cohort recruited in 2000–2001 when the children were 5 years, and has been described in detail elsewhere.11 Local research ethics committee approval was obtained in 1999. Important to this report was the care taken in randomisation of the cohort and quality assurance of the measurements.

With the parents' written consent and children's assent, a total of 307 children (137 girls and 170 boys) were originally recruited to the EarlyBird study at 5 years. Fifty-six were excluded from this analysis because the father was either absent, or known not to be the biological father, and one because the mother was absent. A further two fathers (one under treatment for congestive cardiac failure and the other taking anabolic steroids) and four mothers (all pregnant) were not measured. Of the remaining 244 children, 18 left the study after the baseline visit (5 years) and were omitted from the analysis for lack of follow up. Accordingly, the final analysis was carried out on 226 family trios (125 sons and 101 daughters) for which measures of BMI were available on both biological parents when the children were 5 years, and on the child at a minimum of two annual time points. BMI in both parents and children was derived from direct measurement of height (Harpenden stadiometer, Holtain Ltd., Crosswell, Crymych, Dyfed, UK) and weight (Tanita Solar 1632W electronic scales, West Drayton, UK), performed in blind duplicate and averaged. The analysis was taken to 8 years to avoid the variable impact of puberty.


We analysed BMI relationships within this cohort of family trios at several levels, and repeated measurements on the children annually on four consecutive occasions from 5 to 8 years. We sought to establish the mother–father (pairing) relationship, the parent–offspring relationship according to whether none, one or both of the parents was/were overweight, the mother–offspring and father–offspring relationship and, finally, the mother–daughter and father–son (same sex) and mother–son and father–daughter (opposite sex) relationships, using parental BMI as a category (World Health Organisation definitions of normal weight, overweight and obesity at <25, 25 and 30 kg/m2, respectively) and as a continuous variable. BMI-SDS in the children were calculated from the British 1990 standards.12

Data imputation

Imputation is justified in longitudinal studies where sufficient data exist to establish both general and individual trends. Its value lies in retaining subjects, who may otherwise be lost for lack of just one data point, at a small price. BMI was available at each of the four time points in 212 children, at three in 11 and at two in 4. The missing BMI measurements were imputed using a longitudinal method (linear mixed effects modelling). Accordingly, of the 904 (226 × 4) measurements of BMI used in the analysis, 885 (98%) were real data values and only 19 (2%) imputed.

Statistical analysis

Both cross-sectional and longitudinal analyses were carried out by SPSS version 15 (SPSS Inc., Chicago, IL, USA).

Linear regression analysis was performed to determine the strength of association between parental age and parental BMI, parental BMI and both offspring birth weight standard deviation score (BWt SDS) and offspring BMI-SDS during pre-puberty, maternal BMI and paternal BMI. Analysis of variance (ANOVA) was used to compare means, and the χ2-test for frequencies.

Linear mixed effects (multilevel) modelling was performed to determine whether the BMI-SDS of the offspring at age of 5 years and the subsequent change in BMI-SDS from 5 to 8 years were related to the BMI of the parent. The dependent variable was offspring BMI-SDS measured at four annual time points, and the fixed effects included age of offspring, BMI of parent, the ‘age of offspring × BMI of parent’ interaction, and confounders established by linear regression. Random intercepts and age-related slope estimates were included in the model. The analysis was carried out at three levels, relating:

  1. 1)

    The average BMI of the two parents to the BMI-SDS of all offspring (controlling for the age of each parent and gender of the offspring).

  2. 2)

    Maternal BMI and paternal BMI separately to the BMI-SDS of all offspring (controlling for the age of the relevant parent, BMI of the other parent and gender of the offspring).

  3. 3)

    Maternal BMI and paternal BMI separately to the BMI-SDS of the sons and the daughters separately (controlling for the age of the relevant parent and BMI of the other parent).

The analyses at all three levels were repeated with further adjustment for offspring BWt SDS, the only confounder that was significantly related to offspring BMI-SDS. To enable cross-sectional comparisons, partial correlations were performed between parental BMI and the BMI-SDS of offspring at age 5, 6, 7 and 8 years for the three levels described above.

Finally, the proportion of obese offspring within each category of parental BMI was calculated for same-sex parent–offspring pairs at each time point. The numbers of obese children born to normal weight and obese parents were compared using the Pearson's χ2-test.



The characteristics and anthropometric measures of the mothers and fathers when the children were 5 years are summarised in Table 1. The fathers were taller and of greater BMI than the mothers, and fewer than half of either were of normal weight. There was no relationship between age and BMI (mothers r=0.02, P=0.82; fathers r=0.03, P=0.69), but the age range was narrow. There was only a weak association (r=0.12, P=0.08) between maternal and paternal BMI (assortative pairing).

Table 1 Summary characteristics of parents and their offspring at 5 years


The characteristics and anthropometric measures of the children at entry into the study are shown alongside their parents in Table 1. The boys were somewhat taller than the girls (not shown), but the girls were of greater BMI. The median, −2 s.d. and +2 s.d. BMI of the boys and girls are plotted from 5 to 8 years on their respective 1990 standard curves in Figure 1. The median BMI's of the EarlyBird boys at all time points were the same as those of boys in 1990, whereas those of the girls were only slightly higher. Again, the BMI values corresponding to −2.0 s.d. BMI were much the same as they were in 1990, but those corresponding to +2.0 s.d. were substantially higher, particularly in the girls. The correlations between birth weight SDS and BMI-SDS at 5 years and at 8 years were weak or weak–moderate, with high statistical probability (5 years: r=0.23, P<0.001 adjusted for gender; 8 years: r=0.27, P<0.001 adjusted for gender). There were no statistically significant differences in the BMI's of children or parents grouped by birth order of the child.

Figure 1

Median BMI (linear scale)±2 s.d. of sons and daughters from 5 to 8 years plotted on 1990 UK BMI standard curves.12


There was a moderate relationship between parental BMI (averaged) and the BMI-SDS of their children at 5 years (r=0.29, P<0.001), which strengthened by 8 years (r=0.38, P<0.001) (Table 2a), such that 3, 17 and 29%, respectively, of the children were overweight/obese when neither, one or both of the parents was/were overweight/obese. The effect size of parental BMI on their offspring's BMI-SDS is shown in Table 3a. A substantial effect was already present at 5 years, where for every 10 kg/m2 difference in parental BMI the offspring BMI differed by almost one BMI-SDS. This effect increased further by one-third by 8 years. There was little evidence of any relationship between parental BMI and the children's birth weights (r=0.08, P=0.26).

Table 2 Correlation between parent BMI and offspring BMI-SDS among same-sex and opposite-sex pairings from age 5 to 8 years (a) Controlling for age of each parent and offspring gender and birth weight (b) Controlling for age of parent, BMI of other parent and offspring gender and birth weight (c) Controlling for age of parent, BMI of other parent and offspring birth weight
Table 3a Effect size (from linear mixed effects modelling) on offspring BMI-SDS at 5 years and on change in BMI-SDS from 5 to 8 years (Δ5–8 years), of unit difference in parent BMI (1 kg m−2) modelled for all parents–all offspring

Mothers–children, fathers–children

The correlation between the BMI-SDS of the children and BMI of their mothers was weak-to-moderate at each annual time point, and not different from the corresponding correlations with the fathers (Table 2b). The effect of mother's BMI was independent of the effect of the father's BMI and vice versa. Again, adjustment for birth weight had no effect, as there was little evidence of any association between the child's birth weight SDS and his/her mother's or father's BMI (r=0.09, P=0.18; r=0.02, P=0.77, respectively). The effect sizes of the mothers' and fathers' BMI separately on that of their offspring, irrespective of gender, are shown in Table 3b. As with the all parents-all offspring analysis, the effects on the offspring BMI reached at 5 years, and on its rate of change thereafter, were statistically significant.

Table 3b Effect size (from linear mixed effects modelling) on offspring BMI-SDS at 5 years and on change in BMI-SDS from 5 to 8 years (Δ5–8 years), of unit difference in parent BMI (1 kg/m2) modelled for mothers–all offspring, fathers–all offspring

Mothers–daughters, mothers–sons and fathers–sons, fathers–daughters

The correlations at each time point between maternal BMI and BMI-SDS of their daughters and sons separately, and between paternal BMI and BMI-SDS of their sons and daughters separately, are shown in Table 2c. The same-sex correlations between parent and child were universally stronger than the opposite-sex correlations, and by 8 years gained moderate strength at r=0.40 in mother–daughters and r=0.36 in father–sons. Adjustment for birth weight SDS did not influence these correlations.

The mean BMI-SDS of boys and girls, according to whether their mothers or fathers were of normal weight, overweight or obese, are plotted at 5, 6, 7 and 8 years in Figure 2, the key figure in this report. There were no differences in the children's BMI-SDS related to category of parental BMI between opposite-sex pairs, but striking differences between same-sex (gender-assorted) pairs, which were statistically highly significant. The differences among same-sex pairs according to BMI category of the parent were not present at birth (data not shown), but were significant at 5 years and diverged further from 5 to 8 years.

Figure 2

Mean BMI-SDS (±1 s.e.m.) of sons and daughters at 5, 6, 7 and 8 years by category of parental BMI (normal weight, overweight or obese) sons and mothers (a), daughters and mothers (b), sons and fathers (c), daughters and fathers (d).

By the age of 8 years, there was a significant gender interaction in the relationship of offspring BMI-SDS with the BMI of both mothers (interaction P=0.004) and fathers (interaction P=0.02). Thus, the difference in BMI between those with normal-weight mothers and those with obese mothers was much greater for their daughters (1.37 SDS) than for their sons (0.16 SDS). On the other hand, the difference in offspring BMI between those with normal-weight fathers and those with obese fathers was much greater for their sons (1.28 SDS) than for their daughters (0.17 SDS). It is these offspring gender interactions that render the weight gain gender assortative.

Although the daughters of obese mothers and sons of obese fathers continued to gain excess weight year by year, the children of normal-weight same-sex parents did not, remaining parallel to the 1990 BMI standards throughout. The mean BMI of the daughters of normal-weight mothers lay between +0.10 and 0.20 SDS above the median for 1990, and the mean BMI of the sons of normal-weight fathers −0.34 SDS below the median of 1990. The same-sex offspring of overweight parents occupied an intermediate position. There were no associations of meaningful size between BMI of the parent and birth weight SDS of the same sex or opposite sex offspring (mother–son/mother–daughter: r=0.06/0.13, P=0.52/0.21 father–son/father–daughter: r=0.07/0.05, P=0.45/0.61).

The differences in Figure 2 were confirmed by linear mixed effects (multilevel) modelling, which analysed parental BMI as a continuous variable rather than by categories (Table 3c). The association between parental BMI and the offspring BMI-SDS at 5 years, and BMI-SDS change from 5 to 8 years, were both statistically significant for same-sex pairs, but not for opposite-sex pairs.

Table 3c Effect size (from linear mixed effects modelling) on offspring BMI-SDS at 5 years and on change in BMI-SDS from 5to 8 years (Δ5–8 years), of unit difference in parent BMI (1 kg/m2) modelled for mothers–sons, mothers–daughters, fathers–sons, fathers–daughters

The prevalence of obesity at each time point in the sons and daughters according to the BMI category of the same-sex parent is shown in Figure 3. At 8 years, only 3% (1/35) of the sons of normal-weight fathers and 4% (2/45) of the daughters of normal-weight mothers were obese, little different from the population as a whole on whom the standards were based a generation earlier. On the other hand, 18% (5/28) of the sons of obese fathers and 41% (7/17) of the daughters of obese mothers were obese.

Figure 3

The prevalence of obesity (>ninety-eighth centile1990 UK BMI standards) among children grouped according to BMI category of same-sex parent (normal weight <25, overweight 25–30, obese 30 kg/m2).


Our data suggest that nowadays the relationship between parental BMI and weight gain in their children is gender assortative. The BMI of the daughter was associated with that of her mother, and of the son with that of his father. The corresponding relationships in mother–son and father–daughter pairs were weak or absent. We believe the effect size may be the largest yet recorded in childhood obesity. The data do not challenge the existing literature, but extend the analysis of parent–child relationships to gender-assorted pairings. The rise in childhood obesity over recent time seems to be largely confined to those whose same-sex parents are overweight or obese. Same-sex assortment of parent-child obesity has not been previously recorded.

Assortment among mother–daughter and father–son pairs offers a means of distinguishing environmental from genetic causality that analysis of gender-unmatched pairs cannot. The selective transmission of traits from father to son or mother to daughter is unusual for several reasons. The Y-chromosome carries few genes, and Y-linked traits are uncommon. They also tend to be associated with male infertility. X-linked traits from the mother are commoner, but the chances of a daughter being affected are the square root of those of the son (for example, 1:10 000 against 1:100). Same-sex transmission is therefore atypical, and the gender-assortative associations we describe point rather to the environment as the dominant link between parent and offspring BMI. We cannot exclude an unusual pattern of genetic transfer or gestational imprinting (epigenetics),13 but a pattern of linkage that involves father–son as much as mother–daughter, and same-sex but not opposite-sex pairings seems more in line with an environmental, possibly behavioural, influence—the mother acting as role-model for her daughter, and the father for his son.

Why have these gender-assortative relationships not been remarked on before? We think possibly because they are new. It has long been recognised that obesity runs in families,14 and studies from the 1980's of twins reared apart,15, 16 and of adopted children,17, 18 concluded that the influence of parents on their offspring's weight was more genetic than environmental. But such studies were undertaken before the current rise in childhood obesity.19 Two studies in particular, that looked for but found no same-sex relationship in BMI of the kind we report here, were based on birth cohorts dating back to the 1950's and 1970's.8, 20 Had the gender-assortative link in BMI existed then, we would not be in a position to link it to the subsequent rise in childhood obesity now.

A recent analysis of contemporary twins suggests that the heritability of BMI remains high in children generally,21 but the rise in mean BMI without a change in the median suggests that the past generation has been witness to change in a specific group of children, not to change in the population as a whole. The median BMI of contemporary children seems to be much the same as it was in the 1980's (1990 UK standards), indicating that the BMI of the average child has altered little during the past 20 years. In contrast, the BMI corresponding to +2 s.d. is now considerably higher than the 1990 standard for both boys and girls, suggesting that society now incorporates children of higher body weight who were not around in 1990. It is by implication these children—and these alone—who account for the rise in childhood obesity and, given insufficient time in a single generation for genetic change to have occurred, it is likely they signify environmental factors that were earlier either not present or much weaker. These factors seem to be embedded within mother–daughter and father–son relationships, and understanding the relationships may be important to preventing childhood obesity. Arguments for causal relationships in general are helped by evidence of changing outcome in response to varying dose. Grouping the daughters according to their mothers' category of BMI, and the sons according to their fathers', produced dose-dependent differences in offspring BMI which were sizeable.

At what age does the association between parental and offspring BMI emerge? The wide divergence in offspring BMI-SDS according to category of parental BMI already present at 5 years, but lack of difference at birth, suggests an early post-natal (but not gestational) influence, and we have documented elsewhere that over 90% of the weight centiles crossed before puberty by contemporary girls (>70% in boys) are crossed before the age of 5 years.22 Others have suggested that the first few months of life may be crucial.23 The implications may be important, where so much emphasis is placed on school physical education, school dinners, school transport and so on. Such factors do not operate before the age of 5 years, and are in any case often beyond parental influence.

Multilevel modelling using BMI as a continuous variable further suggested that the assortative effects were dynamic, insofar as the rate of weight gain during the period from 5 to 8 years was also a function of same-sex parental BMI. Importantly, the offspring of parents of normal weight, girls or boys, gained no excess weight over the course of the study. In contrast to reports on children born earlier in the twentieth century, birth weight and gestation had no influence on these relationships. The difference may reflect dominance of the (greater) weight acquired after birth by contemporary children.

The similarity of association between offspring BMI and maternal or paternal BMI again implies that gestational factors, which are unique to the mother, make little contribution to obesity in contemporary children. A similar conclusion was drawn by the Cardiovascular Risk in Young Finns Study group and in the ALSPAC cohort,20, 10 where the correlations were closely similar to our own.

Strengths and limitations

This study has strengths and limitations. It is relatively small for epidemiological purposes, but numbers do not need to exceed the requirements of statistical power, and the differences/correlations that we report are statistically highly significant. Furthermore, EarlyBird is longitudinal and involves a well-defined population of uniform age, which is crucial to the resolution of age-related events. Importantly, the heights and weights of EarlyBird parents were based on direct measurement rather than self-report. This may be crucial to such analyses, because women typically under-report their weight, and men typically over-report their height. Parental BMI was nevertheless measured when their offspring were 5 years, so that we relied on correlation with BMI in earlier years to draw conclusions about assortative pairing and impact of gestation on the offspring's later BMI. The effect sizes, and their consistency over time and between the genders, nevertheless suggest that the conclusions are robust, and that reliance on single BMI measurements in the parents was justified. BMI has limitations as a surrogate for adiposity, because it does not distinguish fat from lean. There are no 1990 UK reference standards for body fat by which to judge how much of the recent rise in BMI can attributed to fat, but the % body fat (DEXA) of EarlyBird children above the 1990 UK ninety-eighth centile for BMI at 8 years was substantially higher than those below it, while the percentage lean was not (data not shown). Longitudinal analysis using multilevel modelling makes the most of data, and we believe this is the first study to have analysed the rate of weight gain during childhood in relation to parental BMI by this method. We have not been able to fully exclude non-paternity in this study but, where step-relationships were declared, the trio was excluded. Although the relationships we report seem statistically robust, they may not be generalisable to other populations, and they cannot be used to imply direction of causality. Finally, the BMI differences we have adduced are only as accurate as the BMI standards on which they were based. The 1990 UK standards are not true trends, but the smoothed data from several cross-sectional sets.24


It is important not to over-interpret these findings, and to note that they relate here only to pre-pubertal children. However, they may be important because they imply that the recent rise in childhood obesity could have an uncomplicated explanation, and they might explain why behavioural intervention at the level of the child has little impact on BMI.25 The lack of gestational effect on the offspring's weight, dominant influence of the same-sex parent and evidence for an ‘epidemic’ confined selectively to the children of overweight/obese parents suggest that childhood obesity may be less a ‘feed-forward’ effect of parental genes and gestational programming, as has been suggested in the past,26, 27 and more a ‘feed-backward’ effect of parental behaviour. Childhood obesity in the 21st century may owe more to obese parents than to obesity genes. Genes are difficult to modify, but parental obesity, like parental smoking, could readily be targeted in the interests of the child.

Conflict of interest

The authors declare no conflict of interest.


  1. 1

    Lean ME . Pathophysiology of obesity. Proc Nutr Soc 2000; 59: 331–336.

    CAS  Article  Google Scholar 

  2. 2

    Ebbeling CB, Pawlak DB, Ludwig DS . Childhood obesity: public-health crisis, common sense cure. Lancet 2002; 360: 473–482.

    Article  Google Scholar 

  3. 3

    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–S107.

    Google Scholar 

  4. 4

    Barness LA, Opitz JM, Gilbert-Barness E . Obesity: genetic, molecular, and environmental aspects. Am J Med Genet A 2007; 143A: 3016–3034.

    CAS  Article  Google Scholar 

  5. 5

    Hillier TA, Pedula KL, Schmidt MM, Mullen JA, Charles MA, Pettitt DJ . Childhood obesity and metabolic imprinting: the ongoing effects of maternal hyperglycemia. Diabetes Care 2007; 30: 2287–2292.

    Article  Google Scholar 

  6. 6

    Garn SM, Clark DC . Trends in fatness and the origins of obesity ad hoc committee to review the ten-state nutrition survey. Pediatrics 1976; 57: 443–456.

    CAS  PubMed  Google Scholar 

  7. 7

    Keiller SM, Colley JR, Carpenter RG . Obesity in school children and their parents. Ann Hum Biol 1979; 6: 443–455.

    CAS  Article  Google Scholar 

  8. 8

    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.

    CAS  Article  Google Scholar 

  9. 9

    Sorensen TI, Holst C, Stunkard AJ, Skovgaard LT . Correlations of body mass index of adult adoptees and their biological and adoptive relatives. Int J Obes Relat Metab Disord 1992; 16: 227–236.

    CAS  PubMed  Google Scholar 

  10. 10

    Davey SG, Steer C, Leary S, Ness A . Is there an intrauterine influence on obesity? Evidence from parent child associations in the Avon Longitudinal Study of Parents and Children (ALSPAC). Arch Dis Child 2007; 92: 876–880.

    Article  Google Scholar 

  11. 11

    Voss LD, Kirkby J, Metcalf BS, Jeffery AN, O'Riordan C, Murphy MJ et al. Preventable factors in childhood that lead to insulin resistance, diabetes mellitus and the metabolic syndrome: the EarlyBird diabetes study 1. J Pediatr Endocrinol Metab 2003; 16: 1211–1224.

    Article  Google Scholar 

  12. 12

    Cole TJ, Freeman JV, Preece MA . Body mass index reference curves for the UK, 1990. Arch Dis Child 1995; 73: 25–29.

    CAS  Article  Google Scholar 

  13. 13

    Waterland RA, Travisano M, Tahiliani KG, Rached MT, Mirza S . Methyl donor supplementation prevents transgenerational amplification of obesity. Int J Obes (Lond) 2008; 32: 1373–1379.

    CAS  Article  Google Scholar 

  14. 14

    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.

    Article  Google Scholar 

  15. 15

    Allison DB, Kaprio J, Korkeila M, Koskenvuo M, Neale MC, Hayakawa K . The heritability of body mass index among an international sample of monozygotic twins reared apart. Int J Obes Relat Metab Disord 1996; 20: 501–506.

    CAS  PubMed  Google Scholar 

  16. 16

    Stunkard AJ, Harris JR, Pedersen NL, McClearn GE . The body-mass index of twins who have been reared apart. N Engl J Med 1990; 322: 1483–1487.

    CAS  Article  Google Scholar 

  17. 17

    Sorensen TI, Holst C, Stunkard AJ . Childhood body mass index—genetic and familial environmental influences assessed in a longitudinal adoption study. Int J Obes Relat Metab Disord 1992; 16: 705–714.

    CAS  PubMed  Google Scholar 

  18. 18

    Stunkard AJ, Sorensen TI, Hanis C, Teasdale TW, Chakraborty R, Schull WJ et al. An adoption study of human obesity. N Engl J Med 1986; 314: 193–198.

    CAS  Article  Google Scholar 

  19. 19

    Hughes JM, Li L, Chinn S, Rona RJ . Trends in growth in England and Scotland, 1972–1994. Arch Dis Child 1997; 76: 182–189.

    CAS  Article  Google Scholar 

  20. 20

    Kivimaki M, Lawlor DA, Smith GD, Elovainio M, Jokela M, Keltikangas-Jarvinen L et al. Substantial intergenerational increases in body mass index are not explained by the fetal overnutrition hypothesis: the Cardiovascular Risk in Young Finns Study. Am J Clin Nutr 2007; 86: 1509–1514.

    CAS  Article  Google Scholar 

  21. 21

    Wardle J, Carnell S, Haworth CM, Plomin R . Evidence for a strong genetic influence on childhood adiposity despite the force of the obesogenic environment. Am J Clin Nutr 2008; 87: 398–404.

    CAS  Article  Google Scholar 

  22. 22

    Gardner DS, Metcalf BS, Hosking J, Jeffery AN, Voss LD, Wilkin TJ . Contribution of early weight gain to childhood overweight and metabolic health: a longitudinal study. (EarlyBird 36). Pediatrics 2009; 123: e67–e73.

    Article  Google Scholar 

  23. 23

    Gillman MW . The first months of life: a critical period for development of obesity. Am J Clin Nutr 2008; 87: 1587–1589.

    CAS  Article  Google Scholar 

  24. 24

    Freeman JV, Cole TJ, Chinn S, Jones PR, White EM, Preece MA . Cross sectional stature and weight reference curves for the UK, 1990. Arch Dis Child 1995; 73: 17–24.

    CAS  Article  Google Scholar 

  25. 25

    Kamath CC, Vickers KS, Ehrlich A, McGovern L, Johnson J, Singhal V et al. Behavioral interventions to prevent childhood obesity. A systematic review and meta-analyses of randomized trials. J Clin Endocrinol Metab 2008; 93: 4606–4615.

    CAS  Article  Google Scholar 

  26. 26

    Levin BE . Metabolic imprinting on genetically predisposed neural circuits perpetuates obesity. Nutrition 2000; 16: 909–915.

    CAS  Article  Google Scholar 

  27. 27

    Whitaker RC, Dietz WH . Role of the prenatal environment in the development of obesity. J Pediatr 1998; 132: 768–776.

    CAS  Article  Google Scholar 

Download references


We gratefully acknowledge funding from the following organizations: Bright Futures Trust, Smith's Charity, Diabetes UK, NHS R&D, DoH, Child Growth Foundation, Diabetes Foundation and EarlyBird Diabetes Trust.

Author information



Corresponding author

Correspondence to T J Wilkin.

Additional information

Author contributions

I declare that I participated in the contributions made to the study and that I have seen and approved the final version. EMP-P contributed to the analysis and writing of the paper; BSM provided statistical advice; JH helped with data management and statistical advice; ANJ helped in data collection; LDV acted as the EarlyBird Study Co-ordinator; TJW developed concept, helped in writing the paper and also acted as EarlyBird Study Director.

The funding organisations had no role in the execution of this study.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Perez-Pastor, E., Metcalf, B., Hosking, J. et al. Assortative weight gain in mother–daughter and father–son pairs: an emerging source of childhood obesity. Longitudinal study of trios (EarlyBird 43). Int J Obes 33, 727–735 (2009).

Download citation


  • childhood obesity
  • parental BMI
  • gender assortment
  • pathogenesis
  • environment

Further reading