Introduction

Childhood obesity is an increasing public health problem.^1,2,3 Recent cross-sectional studies^4,5,6,7 confirmed earlier findings by Kramer⁸ that children at school admission and adolescents who were breast-fed in infancy had a lower risk of being overweight or obese at school age, and that there was a negative dose–effect relationship between breast-feeding duration and the occurrence of obesity. As obesity is barely amenable to treatment,^9,10 any promising prospect for prevention should be scrutinized, and implemented where possible. However, cross-sectional data may have limitations with respect to recall of the mode of feeding in infancy. They usually do not contain a time course of obesity prevalence between infancy and school age, and not all of them include sufficient additional information on skin-fold thickness, parental BMI and smoking status, or relevant socioeconomic data. Butte,¹¹ reviewing several studies on the relationship between breast-feeding in infancy and later obesity, comes to the conclusion that the subject remains controversial.

The German Multicenter Atopy Study (MAS) offers the opportunity to analyse the data of a longitudinal study that recruited its participants at birth.¹² It includes nutritional and anthropometric data collected at 0, 3, 6, 12, 18 months and annually thereafter in a standardized manner, including parental BMI, and socioeconomic status, as well as parental smoking and maternal smoking during pregnancy. Thus, it offers the opportunity to test the findings of the cross-sectional studies in a longitudinal nonintervention study.

Subjects and methods

Subjects

The subjects of the MAS were enrolled at birth. Out of a total of 7609 singleton term infants born during 1990 in four urban (u) and two rural (r) catchment areas of six delivery units in Germany (Berlin (2u), Duesseldorf (u), Freiburg (r), Mainz (r) and Munich (u)), 1314 were enrolled in the project: a total of 499 subjects had two first-degree relatives (parents mostly) suffering from an allergy and/or having an elevated cord blood IgE level (>0.9 KU), and 815 infants were identified by a dynamic (longitudinal) randomization procedure from the remaining large majority (91%) of the term infants.¹² Excluding the infants with major malformations, and the ones admitted to a newborn care unit, they represent the normal live births of the catchment areas of the respective delivery departments during the whole year 1990.

Not all data were available on every child and age: cohort mortality, and missing data for parental BMI, smoking history, family income/social status, mode of feeding in infancy, height, length, weight, triceps, or subscapular skin fold, reduced the number of cases for statistical analysis.

To obtain groups sufficiently large for comparison, especially for multivariate statistical analysis of the 6 y data, subjects who had never been breast-fed (8.1%), and those breast-fed in part for 2 months only (23.4%) were combined into the category of bottle-fed infants (BO, 31.5%), and subjects breast-fed for 3 months or more into the category of breast-fed infants (BR, 68.5%).¹³ Mothers who had stopped breast-feeding by 3 months usually had started at least complementary bottle-feeding well before. The categories BO and BR are being applied in all comparisons.

In the description of the change in prevalence of overweight/obesity over time (Figure 1, Figure 2, Figure 3 and Figure 4), and for comparing results of uni- and multivariate analyses at 6 y of age, only the 480 cases were included who had complete data for BMI, triceps, and subscapular skin fold, as well as for all independent variables. The influence of loss to follow-up and missing values was tested by Cochran–Mantel–Haenzel statistics, and was shown to be insignificant with respect to the outcomes (BMI and skin-fold thickness).

Figure 1.

Prevalence of overweight, according to BMI. Proportion (%) of children exceeding the 90th percentile of the BMI reference values (15), depending on feeding mode in infancy. The 480 cases used were those for whom complete data were still available at the age of 6 y. ^**P<0.01; ^***P<0.001.

Full figure and legend (53K)

Figure 2.

Prevalence of obesity, according to BMI. Proportion (%) of children exceeding the 97th percentile of the BMI reference values (15), depending on feeding mode in infancy. The 480 cases used were those for whom complete data were still available at the age of 6 y. ^**P<0.01; ^***P<0.001.

Full figure and legend (58K)

Figure 3.

Prevalence of adiposity, according to triceps skin-fold thickness. Proportion (%) of children exceeding the 97th percentile of the reference values,¹⁶ depending on feeding mode in infancy. The 480 cases used were those for whom complete data were still available at the age of 6 y. ^*P<0.05; ^***P<0.001.

Full figure and legend (48K)

Figure 4.

Prevalence of adiposity, according to subscapular skin-fold thickness. Proportion (%) of children exceeding the 97th percentile of the reference values,¹⁶ depending on feeding mode in infancy. The 480 cases used were those for whom complete data were still available at the age of 6 y. ^*P<0.05; ^**P<0.001.

Full figure and legend (57K)

Measurements

Weights were measured on calibrated beam balances with the children completely undressed, and the parents (2-1/2 y post partum) wearing underwear, to the nearest 10 g in infants and 100 g in older children and adults. Prone length was determined during the first 2 y of life, and standing height thereafter in the children, and in their parents, on Harpenden measuring boards or stadiometers, respectively. Readings were recorded to the nearest 0.2 millimetre. After calibration, inter- and intra-observer accuracy was within 0.5 cm. Triceps and subscapular skin-fold thicknesses were determined using Harpenden calipers. Readings after 3 s were to the nearest millimetre. All measurements were taken applying highly standardized procedures¹⁴ with the technique of measurement of the observers calibrated by one of the authors (RLB), whose techniques had been calibrated by Whitehouse and Tanner, London. Until the age of 6 y, the children were measured on eight occasions (Figure 1 and Figure 2) within 1 week of the specified month of age during the first year of life and within 2 weeks thereafter. While weighing and measuring was accepted at most visits, some children were afraid of the caliper, and refused to have their skin-fold thickness taken. This had to be and was respected according to the ethical requirements to be applied in experiments with persons unable to give formal consent.

At every follow-up visit for physical examination, anthropometry, allergy-specific sampling, and information on feeding were collected by protocol, by questionnaire, and by a 24-h recall interview.¹³

Definitions

The BMI was compared to the age- and gender-specific BMIs of Rolland-Cachera et al,¹⁵ with values exceeding the 90th percentile considered as overweight and the ones exceeding the 97th percentile being defined as obese. Normative, age- and gender-specific values by Prader et al¹⁶ served as the standards for skin-fold thickness. Again, the 90th and the 97th percentiles were taken as the cut-offs, respectively. For newborn infants, and for the first 3 months of life, there are no criteria for obesity. Therefore, in the youngest age groups, means of BMI and skin folds are compared. Parents with a BMI exceeding the 90th percentile of the study participants were classified as overweight. In the mothers, this corresponded to a BMI of 27 kg/m².

Maternal smoking status during pregnancy was represented most reliably by dichotomizing it into 'yes' or 'no'. Social indexes of the family ((1) educational, (2) occupational level of the parent contributing the highest share to the family income (in 90% of the families, the father), (3) family income itself were all assigned a score between 1 and 4 points, and then summed, resulting in scores ranging from 3 to 12. The sum was divided by 4, resulting in 3 SES classes: low, middle, and high.^17,18 In the analysis of the MAS-data, this classification separated best.¹⁹

Statistics

All data were entered into an SIR data bank (Scientific Information Retrieval data bank software) with plausibility controls, and with double entry for major parts of the data set. The statistical software package SAS 6.12 was mainly used. The prevalence of overweight and obesity was computed gender specifically for every visit from the age of 6 months on. Univariate odds ratios (ORs) as well as multivariate adjusted ORs were computed applying logistic regression analysis. Candidate determinants included prematurity, birth weight in three classes (<2500 g, 2500–3999 g, 4000 g), breast-feeding, beikost (food other than milk or formula²⁰) before 6 months of age, number of older siblings, overweight parents, smoking status of the mother, maternal smoking during pregnancy, family social status, age, and educational status of the mother. Gender was generally included by applying gender-specific anthropometric reference data. The multivariate model included factors that were associated with the exposure as well as with the outcome variable (P<0.05), the inclusion of which into the model changed the crude OR for breast-feeding by at least 10%. Independent of an association with any exposure or outcome variables, factors were included that changed the crude OR for breast-feeding by at least 20%. The ORs were computed with their 95% confidence intervals (CIs). Statistical significance for continuous variables in univariate comparisons was tested by a two-tailed t-test, for categorical variables by ² test, in the analysis of variance of continuous variables by f-test, and in Mantel–Haenzel equations by chi-square-test. Bias as a result of loss to follow-up was tested by Cochran–Mantel–Haenzel statistics. Finally, where indicated, we applied Kendall's , and for inhomogeneous subsets Duncan's test. SAS (and SPSS) offers these options.

The MAS was approved by the ethical committee of the Charité–Virchow–Klinikum, Humboldt University, Berlin. Informed written consent for participating with their children in the study was obtained by the parents. Any change in the preparedness to participate in individual tests, as well as the will of the participating children, was respected. The study was reviewed and approved by the Berlin office for data protection.

Results

Tables 1a and 1b present means and standard deviations as well as proportions (%) of the relevant outcome, and influential variables of all bottle- and breast-fed infants for whom data were available at any specific point of time. Because of loss to follow-up, and missing values, n varies between ages and variables. To facilitate comparisons, Tables 2a and 2b present the same type of data only for those 480 children for whom complete data were still available at the age of 6 y. At birth, mean BMI was nearly identical for the BR and BO groups. At the age of 3 months, the BMI and triceps skin-fold thickness are already significantly higher in BO.

Table 1a - BMI at birth (BMI_B), 3 months (BMI_3mo) and 6 y (BMI_6 y) of age as well as the skin-fold thickness (3 months, 6 y), parental BMI, and age of the mother at delivery according to feeding mode (all data available for any single variable are included in this table).

Full table

Table 1b - Categorical variables associated with feeding mode (and BMI) (all data available for any single variable are included in this table; ²-test, See also Table 1a).

Full table

Table 2a - BMI at birth (BMI_B), 3 months (BMI_3mo) and 6 y of age as well as the skin-fold thickness (3 months, 6 y), parental BMI, and age of the mother at delivery according to feeding mode (see Table 1) (only the data for the 480 cases are included for whom complete data for every outcome and independent variable was available at 6 y of age).

Full table

Table 2b - Categorical variables associated with feeding mode (and BMI) (only the data for the 480 cases are included for whom complete data for every outcome and independent variable was available at 6 y of age).

Full table

Applying the BMI reference data of Rolland-Cachera et al,²² Figure 1 and Figure 2 show the age-specific prevalence of overweight and obesity in the 480 BO and BR children having still complete data at the age of 6 y. In both groups, the prevalence of overweight (Figure 1) and obesity (Figure 2) decreases from 6 to 18 months, and increases thereafter. The nadir age appears to be 18 months. At any age between 6 months and 6 y, a somewhat greater proportion was overweight or obese in BO than in BR. Between the ages of 4 y, and of 5 and 6 y, in BO the prevalence of obesity nearly doubled and tripled, respectively. With only minor changes of obesity prevalence in BR, the difference of BMI between groups became statistically significant.

Figure 3 and Figure 4 visualize the changing prevalence of adipose values for the triceps and subscapular skin-fold thickness: compared to the 97th percentile values of the Prader et al¹⁶ reference data, a nadir age for the triceps skin-fold thickness again appears to be at 18 months; the differences between BO and BR become larger and statistically significant from 48 months on. For the subscapular skin-fold thickness, the nadir age appears less clear-cut: while the prevalence of adipose values looks relatively stable in the BR, it increases in BO from 12 months on more than three-fold, the differences to BR being statistically significant at the ages of 24, 48, and 72 months.

Before estimating the magnitude of the influence of breast-feeding in infancy on overweight and obesity at the age of 6 y by unadjusted and (multivariate) adjusted ORs, the potential role of all available independent variables as defined in the Methods section on these outcomes was computed (Table 3). Of 11 factors, only four remained statistically significant in multivariate logistic regression analysis. Table 4 and Table 5 present the ORs and 95% CIs of the effect of breast-feeding. When the mode of infant feeding was considered alone (unadjusted ORs), breast-feeding lowered the risk for overweight and obesity significantly by about two-thirds. After the adjustments, breast-feeding for 3 months or more as well as high family social status lowered whereas overweight of the mothers and their smoking during pregnancy significantly increased the risk for overweight and obesity at the age of 6 y.

Table 3 - Independent variables for weight status at 6 y of age of all cases available.

Full table

Table 4 - Risk of being overweight (BMI, skin folds P 90) at 6 y of age for children breast-fed vs bottle-fed.

Full table

Table 5 - Risk of obesity (BMI, skin folds P97) at 6 y of age for children breast-fed vs bottle-fed.

Full table

Discussion

Several studies have addressed the relation of breast-feeding to childhood obesity with differing results. As Butte¹¹ points out most of them have limitations because of problems with respect to definitions, missing information on birth weight, confounders, or on time course. The studies collected their data under different conditions, fox example, early post-war vs recent years, Eastern or Western European countries, or at different outcome ages. This may explain differences in findings.

In this study, the differences in BMI were not present at birth. At the age of 1 month, the breast-fed infants on average were even somewhat fatter than those exclusively or partially bottle-fed. However, between the second and the third month for the first time, the bottle-fed developed a higher mean BMI and thicker skin-folds than the breast-fed infants (Table 1 and Table 2). Subsequently, the BO consistently had a higher prevalence of overweight and obesity (Figure 1 and Figure 2). The nadir age of overweight and obesity appears to be at about 18 months, when the prevalence starts to increase slowly in BR and much steeper in BO, the differences in BMI becoming statistically significant from the age of 60 months on. The skin-fold thickness values behave similarly (Figure 3 and Figure 4). The findings are robust: they remain stable after controlling for the variables that proved to be relevant in the multivariate analysis, that is, overweight of the mothers, their smoking during pregnancy, and the social status of the family (Table 4 and Table 5). The effect can also be seen in models that include all available cases for each specific outcome variable irrespective of missing values of the other variables (data not shown). The long-term effect of the mode of feeding on skin-fold thickness appears relevant in particular, because skin-fold thickness is a different, and possibly a more specific, indicator of adiposity than BMI.²¹

Our regular annual observations, for the first time, clearly show that the adiposity rebound is being brought forward by bottle-feeding in infancy (Figure 1 and Figure 2). This may have long-term consequences.²² The body shape is affected by many factors on its way through life, and influences that are still visible years later should be scrutinized for their usefulness in prevention.

Risk factors

Obesity of the mother,²³ and social status^24,25 being strong determinants of childhood obesity, also in the present study (Table 3), made controlling for them mandatory. There appears to be consensus that genetic trait as well as intrauterine influences, and cultural transmission determine major proportions of the emergence of adiposity.^{21,26,27,28,29,30,31,32,33} The influence of the mother's BMI on that of her child in the present study is in line with these observations and theories.

Obese and lower social class mothers tend to have less success in breast-feeding.^34,35 In fact, the reduced breast-feeding success of overweight and lower social status mothers could be even the real causal factor for the increased risk of their offspring becoming obese.

High as well as low birth weight may increase the risk of later adiposity.^{28,30,36,37,38,39} Birth weight (in three classes, Table 3) had an effect only in univariate comparisons: while the lowest class did not seem to influence later overweight, the highest increased the risk. However, birth weight was excluded from the final model. It may have a more complex U-shaped relationship to later adiposity.^28,37 However, this aspect could not be tested further because, in our sample, there were very few cases having a birth weight of less than 2.500 kg. Admission to a neonatal-care ward had been an exclusion criterion of the MAS study.

Mothers suffering from an atopic disease tend to breast-feed their infants longer than nonatopic mothers,³⁵ and there appears to be a tendency for a higher prevalence of asthma in overweight children and adults; other atopies do not seem to be related to BMI.^40,41 If obesity were a potential risk factor for asthma, it would not interfere with our observations and interpretation, as the tendency of atopic mothers to breast-feed for longer, and yet a possibly higher prevalence of overweight in asthmatic children, could only have increased the prevalence of overweight or obesity in the breast-fed group, thereby diminishing the difference between groups.

Study design aspects

Cross-sectional studies on early determinants of later adiposity that date back many years, like breast-feeding,^4,5,6,7,8 usually discuss recall bias as one of their potential limitations, despite the fact that the recall for breast feeding may correlate well with that documented in health clinic charts in some studies.⁴² In this context, our prospective birth cohort study may have several advantages: (1) The data were collected close to their origination, ensuring a high degree of validity and precision of any early influences as well as later outcomes. Recall bias did not constitute a problem. (2) The time course of the development of overweight is made visible, and it can be shown at what age the process begins. (3) The BMI differences between feeding groups were not present at the beginning. The longitudinal observation shows them to be related to the mode of feeding.

Despite the fact that our study identified the time course of BMI differences, and thus the sequence of potential causes and effects, the determinants identified here, including breast-feeding and smoking, could not be assigned at random to different groups. Therefore, causal inferences can still only be made with some caution.

A potential drawback of longitudinal studies is related to repeated interaction between the subjects and the professional observers, which may have some unintentional interventive effect.^19,35 In the present study, it could have generally decreased the occurrence of overweight in the children at risk, thereby diminishing the observed effect of breast-feeding.

The group of formerly bottle-fed children includes some that were at least partially breast-fed for less than 3 months. Their assignment to the bottle-fed group could, to some extent, have weakened the observed protective effect of breast-feeding.

Considering the loss to follow-up in the present study, significantly more bottle-fed children dropped out (43.7 vs 25.7%). This could have produced a selection bias if, for example, of the BO more slim subjects had dropped out than overweight and obese ones. The potential effects of loss to follow-up were tested by Cochran–Mantel–Haenzel statistics. Loss to follow-up did not produce any significant bias with respect to our outcome variables (BMI, skin-fold thickness). Focusing on the 480 cases for whom the data were still complete for every variable at the age of 6 y does not produce different conclusions than if all data available at any age are included despite the results differing somewhat in magnitude (data not shown).

Differing findings in other studies

Our observations are in line with the results of several cross-sectional studies,^4,5,6,7,8 and with the cohort study by Yoneyama et al⁴³ who found that bottle-fed infants at the age of 20 months were heavier than the breast-fed ones.

While the critical review of Butte¹¹ cannot be repeated here, we would like to comment on two studies with findings appearing in conflict with ours at first glance. Fomon et al,⁴⁴ re-examining the subjects of their study on growth and serum chemical values of breast-fed infants⁴⁵ published in 1978, found no difference in the prevalence of overweight in 8-y-old children whether breast-fed during the first 112 days or not. From their 1978 paper, it is evident that their breast-fed infants had received additional daily bottle-feedings as a supplement during the early weeks, and strained food from the fifth week on. Therefore, the parents or caretakers had some influence on the food intake of their infants. This may explain why they did not find differences between breast- and bottle-fed infants. In contrast, the breast-feeding mothers of the present study received advice about breast-feeding⁴⁶ at the maternity ward, including time ad libitum feeding, and at their request utilizing our telephone hotline as well as when coming to the study's appointments. No additional bottle-feeding or strained food was recommended before the age of 4 months. This produced more exclusive breast-feeders in the MAS study.¹³

In the study by Baranowski et al²³ a favourable effect of breast feeding duration on weight and BMI at the age of 3–4 y was not significant. This corresponds to our results, as we find statistically not significant favourable effects on BMI of former breast-feeding at the ages of 3 and 4 y, but larger and significant ones at the ages of 5 and 6 y. Possibly, the authors would have come up with similar results as ours had they observed their subjects 1 or 2 y longer.

Possible mechanisms

1. Breast-feeding could be part of a lifestyle running in families and counteracting the development of adiposity. The impact of breast-feeding as a proxy for lifestyle could be more obvious in a less favourable environment, as in the study by Elliott et al.⁴⁷
2. There could be substances in breast milk that imprint the neuro-hormonal system controlling food intake, as there were lasting differences between formerly bottle- and breast-fed infants in the responses of gut endocrine substances to a meal.^48,49
3. Breast-fed infants regulate their food intake quite precisely according to their needs for growth and maintenance, and in addition they control the breast milk production of their mother.^50,51,52,53 In contrast, satiety can easily be challenged by bottle-feeding, when the saturated infant is encouraged to empty the bottle, and by feeding formulas more concentrated in energy and nutrients than breast milk. If this occurs early and goes on, a reliable control over food intake may not develop.^54,55,56 In breast-feeding, establishing a precise and dependable point of satiety is guided by internal physiologic rather than by external social cues.⁵⁷ We favour this hypothesis to explain the protective effect of breast-feeding against later obesity. The results of Ong et al³⁸ support this view, in that the association of catch-up growth in the first 2 y of life with obesity at 5 y in the Avon longitudinal study seems to indicate a lasting effect of excessive growth rate (and thus overfeeding) in infancy. The large prospective study by Stettler et al⁵⁸ shows that rapid growth during the first 4 months of life is a risk factor for overweight at the age of 7 y, and seems to confirm our results and conclusions. The observations of Wilson et al,⁵⁹ showing an increased percentage of body fat and weight at 7 y of age if solid food was introduced early, would also fit this hypothesis, as in feeding solids early the mother or caretaker has an influence on food intake of the infant, thereby disrupting the infant's own control over food ntake.
4. The effect of smoking during pregnancy confirms the find-ings of Blake et al.⁶⁰ (1) It may act through its effect on prenatal growth with subsequent catch-up growth.³⁸ (2) It could be an indicator of a high risk for dependency, including overeating, transmitted from the mother to the child. (3) Also, it could be speculated that smoking acts by means of metabolic imprinting on the system controlling food intake and satiety.

The present study does not contain the variables nor does it have a sample size sufficiently large to test these hypotheses further. The 'real world being multivariate',^61,62 all of the potential pathways leading to adiposity could exist simultaneously, with their strengths varying according to different conditions.

As a next step, the global effects of breast-feeding and smoking during pregnancy on BMI could be tested in controlled intervention trials. Since the random assignment to breast- or bottle-feeding, and to smoking or nonsmoking during pregnancy is neither desirable nor possible, it is suggested that advice about smoking cessation and breast-feeding should be randomized by cluster assignment of catchment areas of pregnant women to different consultation groups.⁶³ Else, sequential cohort design could be applied, which avoids contamination, and ethical conflicts because of withholding important information from the controls. While recruiting the intervention group (smoking cessation during pregnancy, breast-feeding promotion), the control group could already be examined (at the outcome age), so it could not have participated in the intervention. The catchment area, the selection and recruiting procedures, outcome age, instruments for data collection, and other relevant methodologies have to be kept comparable, and remaining differences have to be controlled for.

References

1. Hanley AJG, Harris SB, Gittelsohn J, Wolever TMS, Saksvig B, Zinman B. Overweight among children and adolescents in a Native Canadian community: prevalence and associated factors. Am J Clin Nutr 2000; 71: 693–700. | PubMed | ChemPort |
2. Goran MI. Metabolic precursors and effects of obesity in children: a decade of progress. Am J Clin Nutr 2001; 73: 158–171. | PubMed | ChemPort |
3. Chinn S, Rona RJ. Prevalence and trends in overweight and obesity in three cross sectional studies of British children, 1974–94. BMJ 2001; 322: 24–26. | Article | PubMed | ChemPort |
4. Von Kries R, Koletzko B, Sauerwald T, von Mutius E, Barnert D, Grunert V, von Voss H. Breast feeding and obesity: cross sectional study. BMJ 1999; 319: 147–150. | PubMed | ChemPort |
5. Hediger ML, Overpeck MD, Kuczmarski RJ, Ruan WJ. Association between infant breastfeeding and overweight in young children. JAMA 2001; 285: 2453–2460. | Article | PubMed | ISI | ChemPort |
6. Gillman MW, Rifas-Shiman SL, Camargo CA, Berkey CS, Frazier AL, Rockett HRH, Field AE, Colditz GA. Risk of overweight among adolescents who were breastfed as infants. JAMA 2001; 285: 2461–2467. | Article | PubMed | ISI | ChemPort |
7. Liese AD, Hirsch T, von Mutius E, Keil U, Leupold W, Weiland SK. Inverse association of overweight and breast feeding in 9–10-y-old children in Germany. Int J Obes Relat Metab Disord 2001; 25: 1644–1650. | Article | PubMed | ChemPort |
8. Kramer MS. Do breast-feeding and delayed introduction of solid foods protect against subsequent obesity? J Pediatr 1981; 98: 883–887. | PubMed | ChemPort |
9. Sokol RJ. The chronic disease of childhood obesity: the sleeping giant has awakened. Editorial. J Pediatr 2000; 136: 711–713. | Article |
10. Bergmann KE, Bergmann RL. Therapie der Adipositas im Kindesalter. Pädiatr Grenzgeb 1997; 35: 409–424.
11. Butte NF. The role of breast feeding in obesity. Pediatr Clin North Am 2001; 48: 189–198. | Article | PubMed | ISI | ChemPort |
12. Bergmann RL, Edenharter G, Bergmann KE, Forster J, Bauer CP, Wahn V, Zepp F, Wahn U. Atopic dermatitis in early infancy predicts allergic airway disease at 5 years. Clin Exp Allergy 1998; 28: 965–970. | Article | PubMed | ISI | ChemPort |
13. Bergmann RL, Dudenhausen JW, Bergmann KE, Schmidt E, Wahn U. Wie werden Säuglinge in Deutschland ernährt. Ergebnisse aus der multizentrischen Allergiestudie am Geburtsjahrgang 1990. Monatsschr Kinderheilkd. 1994; 142: 412–417.
14. Cameron N. The methods of auxological anthropometry. In: Falkner F, Tanner JM (eds). Human growth. 2nd edn. Plenum Press: New York and London; 1986. pp 3–46.
15. Rolland-Cachera MF, Cole TJ, Sempé M, Tichet J, Rossignol C, Charraud A. Body mass index variations: centiles from birth to 87 years. Eur J Clin Nutr 1991; 45: 13–21. | PubMed | ChemPort |
16. Prader A, Largo RH, Molinari L, Issler C. Physical growth of Swiss children from birth to 20 years of age. Helv Paediatr Acta 1989; 52(Suppl): 82–85.
17. Winkler J. Die Messung des sozialen Status mit Hilfe eines Index in den Gesundheitssurveys der DHP (Scaling social status in the German National Health Surveys). In: Ahrens W, Bellach B, Jökel KH (eds). Messung soziodemographischer Merkmale in der Epidemiologie. RKI-Schriften: MMV München; 1998. pp 69–74.
18. Bergmann KE, Mensink GBM. Körpermae und Übergewicht. In: Bundesgesundheits-Survey. Gesundheitswesen. Sonderheft 1999; 2: S115–S120.
19. Bergmann RL, Edenharter G, Bergmann KE, Lau S, Wahn U the MAS-Research Group Socioeconomic status is a risk factor for allergy in parents but not in their children. Clin Exp Allergy 2000; 30: 1740–1745.
20. Anderson TA, Fomon SJ. Beikost. In: Fomon SJ (ed). Infant nutrition. Saunders: Philadelphia; 1974. pp 408–434.
21. Bouchard C, Pérusse L. Genetics of obesity. Annu Rev Nutr 1993; 13: 337–354. | Article | PubMed | ISI | ChemPort |
22. Rolland-Cachera M-F, Deheeger M, Bellisle F, Sempé M, Guilloud-Bataille M, Patois E. Adiposity rebound in children: a simple indicator for predicting obesity. Am J Clin Nutr 1984; 39: 129–135. | PubMed | ChemPort |
23. Baranowski T, Rassin DK, Harrison JA, Henske JC. Ethnicity, infant-feeding practices, and childhood adiposity. J Dev Behav Pediatr 1990; 11: 234–239. | PubMed | ISI | ChemPort |
24. Dietz W. Factors associated with childhood obesity. Nutrition 1991; 7: 290–291.
25. Thefeld W, Bergmann KE, Hermann-Kunz E, Junge B, Knopf H, Melchert H-U, Mensink G, Stolzenberg H, Tietze K. Wohnortgröe und Gesundheit Stadt-/Land- Unterschiede. In: Belach B-M (ed). Die Gesundheit der Deutschen, Chapter 1. RKI-Hefte 15/96. Robert Koch – Institut: Berlin; 1996.
26. Stunkard AJ, Harris JR, Pedersen RL, Mc Clearn GE. The body mass index of twins who have been reared apart. New Engl J Med 1990; 322: 1483–1487. | PubMed | ISI | ChemPort |
27. Stunkard AJ, Berkowitz RI, Stallings VA, Schoeller DA. Energy, intake not energy output, is a determinant of body size in infants. Am J Clin Nutr 1999; 69: 524–530. | PubMed | ISI | ChemPort |
28. McCance DR, Pettitt DJ, Hanson RL, Jacobsson LT, Knowler WC, Benett PH. Birth weight and non-insulin dependent diabetes: thrifty genotype, thrifty phenotype, or surviving small baby genotype? BMJ 1994; 308: 942–945. | PubMed | ChemPort |
29. Friedman JM. Molecular genetic approaches to complex traits. In: Bouchard C, Bray GA (eds). Regulation of body weight: Biological and Behavioral Mechanisms. Wiley & Sons: Chichester; 1996. pp 251–269.
30. Bergmann KE, Bergmann RL, Eisenberg A. Offspring of diabetic mothers have a higher risk for childhood overweight than offspring of diabetic fathers. Nutr Res 1984; 4: 545–552.
31. Pettitt DJ. Long-term effects of the intrauterine environment, birth weight, and breast-feeding in Pima Indians. Diabetes Care 1998; 21: B138–B141.
32. Vohr BR, McGarvey ST, Tucker R. Effects of maternal gestational diabetes on offspring adiposity at 4–7 years of age. Diabetes Care 1999; 22: 1284–1291. | PubMed | ISI | ChemPort |
33. Whitaker RC, Dietz WH. Role of the prenatal environment in the development of obesity. J Pediatr 1998; 132: 768–776. | PubMed | ISI | ChemPort |
34. Hilson JA, Rasmussen KM, Kjolhede CL. Maternal obesity and breast-feeding success in a rural population of white women. Am J Clin Nutr 1997; 66: 1371–1378.
35. Bergmann RL, Diepgen TL, Kuss O, Bergmann KE, Sommerfeld C, Kujat J, Wahn U. Breastfeeding duration is a risk factor for atopic eczema. Clin Exp Allergy 2002; 32: 205–209. | Article | PubMed |
36. Ravelli GP, Stein ZA, Susser MW. Obesity in young men after famine exposure in utero and early infancy. N Engl J Med 1976; 295: 349–353. | PubMed | ISI | ChemPort |
37. Dörner G, Plagemann A. Perinatal hyperinsulinism as a possible predisposing factor for diabetes mellitus, obesity, and enhanced cardiovascular risk in later life. Horm Metab Res 1994; 26: 213–221. | PubMed |
38. Ong KK, Ahmed ML, Emmett PM, Preece MA, Dunger DB. Association between postnatal catch-up growth and obesity in childhood: prospective cohort study. BMJ 2000; 320: 967–971. | Article | PubMed | ChemPort |
39. Eriksson JG, Forsén T, Tuomilchto J, Osmond C, Barker DJP. Early growth and coronary heart disease in later life: longitudinal study. BMJ 2001; 822: 949–953.
40. von Kries R, Hermann M, Grunert VP, von Mutius E. Is obesity a risk factor for childhood asthma? Allergy 2001; 56: 318–322. | Article | PubMed | ISI | ChemPort |
41. Young SY, Gunzenhauser JD, Malone KE, McTiernan A. Body mass index and asthma in the military population of the northwestern United States. Arch Int Med 2001; 161: 1605–1611.
42. Kark JD, Troya G, Friedlander Y, Slater PE, Stein Y. Validity of maternal reporting of breast feeding history and the association with blood lipids in 17 year olds in Jerusalem. J Epidemiol Community Health 1984; 38: 218–225.
43. Yoneyama K, Nagata H, Asano H. Growth of Japanese breast-fed infants from birth to 20 months. Ann Hum Biol 1994; 21: 597–608.
44. Fomon SJ, Rogers R, Ziegler EE, Nelson SE, Thomas L. Indices of fatness and serum cholesterol at age 8 years in relation to feeding and growth during early infancy. Pediatr Res 1984; 18: 1233–1238.
45. Fomon SJ, Filer LJ, Thomas LN, Rogers RR. Growth and serum chemical values of normal breastfed infants. Acta Pediatr Scand 1978; 273: 1–29.
46. World Health Organization/United Nations Children's Fund. Ten steps to successful breastfeeding in protection, promotion and support of breastfeeding: the special role of maternity services. World Health Organization/United Nations Children's Fund: Geneva Switzerland; 1989.
47. Elliott KG, Kjolhede CL, Gournis E, Rasmussen KM. Duration of breastfeeding associated with obesity during adolescence. Obesity Res 1997; 5: 538–541.
48. Lucas A, Boyes S, Bloom R, Aynsley-Green A. Metabolic and endocrine responses to a milk feed in six-day-old term infants: differences between breast and cow's milk formula feeding. Acta Paediatr Scand 1981; 70: 195–200. | PubMed | ChemPort |
49. Salmenperä L, Perheentupa J, Siimes MA, Adrian TE, Bloom SR, Aynsley-Green A. Effects of feeding regimen on blood glucose levels and plasma concentrations of pancreatic hormones and gut regulatory peptides at 9 months of age: comparison between infants fed with milk formula and infants exclusively breast-fed from birth. J Pediatr Gastroenterol Nutr 1988; 7: 651–656.
50. Dewey KG. Nutrition, growth, and complementary feeding of the breastfed infant. Pediatr Clin North Am 2001; 48: 87–104. | Article | PubMed | ISI | ChemPort |
51. Kent JC, Mitoulas L, Cox DB, Owens RA, Hartmann PE. Breast volume and milk production during extended lactation in women. Exp Physiol 1999; 84: 435–447.
52. Matheny RJ, Birch LL, Picciano MF. Control of intake by human-milk-fed infants: relationships between feeding size and interval. Developmental Psychobiology 1990; 23: 511–518.
53. Bergmann RL, Bergmann KE. Nutrition and growth infancy. In: Falkner F, Tanner JM (eds). Human growth. A comprehensive treatise, 2nd edn, Vol III: Methodology, ecological genetic and nutritional effects on growth. Plenum Press: New York; 1986, pp 389–413.
54. Widdowson ME, Mc Cance RA. A review: new thoughts on growth. Pediatr Res 1975; 9: 154–156. | Article | PubMed | ISI | ChemPort |
55. Birch LL, Fisher JO. Development of eating behaviors among children and adolescents. Pediatrics 1998; 101: 539–549. | PubMed | ISI | ChemPort |
56. Levin BE. Metabolic imprinting on genetically predisposed neural circuits perpetuates obesity. Nutrition 2000; 16: 909–915.
57. Ounsted M, Sleigh G. The infant's self-regulation of food intake and weight gain. Difference in metabolic balance after growth constraint or acceleration in utero. Lancet 1975; 7922: 1393–1397.
58. Stettler N, Zemel BS, Kumanyika S, Stallings VA. Infant weight gain and childhood overweight status in a multicenter cohort study. Pediatrics 2002;109: 194–199. | Article | PubMed | ISI |
59. Wilson AC, Forsyth JS, Greene SA, Irvine L, Hau C, Howie PW. Relation of infant diet to childhood health: seven year follow up of cohort of children in Dundee infant feeding study. BMJ 1998; 316: 21–25. | PubMed | ChemPort |
60. Blake KV, Gurrin LC, Evans SF, Beilin LJ, Landau LI, Stanley FJ, Newnham JP. Maternal cigarette smoking during pregnancy, low birth weight and subsequent blood pressure in early childhood. Early Hum Development 2000; 57: 137–147.
61. Bergmann KE, Bergmann RL, Schulz J, Gra T, Wahn U. Prediction of atopic disease in the newborn: methodological aspects. Clin Experim Allergy 1990; 203(Suppl): 21–26.
62. Rothman KJ, Greenland S. Modern epidemiology. 2nd edn. Lippincott-Raven: Philadelphia; 1998.
63. Kramer MS, Chalmers B, Hodnett ED, Sevkovskaya Z, Dzikovich I, Shapiro S, Collet J-P, Vanilovich I, Mezen I, Ducruet T, Shishko G, Zubovich V, Mknuik D, Gluchanina E, Dombrovskiy V, Ustinovitch A, Kot T, Bogdanovich N, Ovchinikova L, Helsing E. Promotion of Breastfeeding Intervention Trial (PROBIT). A randomized trial in the Republic of Belarus. JAMA 2001; 285: 413–420. | Article | PubMed | ISI | ChemPort |

Acknowledgements

Supported by a grant of the German Federal Ministry of Research and Technology BMBF # 01 EE 9406/8