The function that the timing of introduction of solid foods may have in the development of child obesity has not been adequately explored, either as a potential confounder of the relationship between breastfeeding and child obesity, or as an independent modifiable risk factor.
To determine the association between infant feeding practices and child overweight/obesity.
Patients and Methods:
Six hundred and twenty subjects were recruited antenatally from 1990 to 1994. A total of 18 telephone interviews over the first 2 years of life recorded infant feeding practices. At mean age of 10 years, height and weight were measured for 307 subjects. Multiple logistic regression was used to determine whether infant feeding practices (duration of exclusive and any breastfeeding, and age at introduction of solid foods) were associated with odds of being overweight/obese (internationally age- and sex-standardized body mass index category) at age 10 years, after adjustment for confounders.
Delayed introduction of solid foods was associated with reduced odds of being overweight/obese at age 10 years, after controlling for socioeconomic status, parental smoking and childcare attendance (adjusted odds ratio (aOR)=0.903 per week, 95% CI=0.841–0.970, P=0.005). Antenatal parental smoking was associated with overweight/obesity at age 10 years (aOR=3.178, 95% CI=1.643–6.147, P=0.001). Duration of exclusive or any breastfeeding was not associated with the outcome.
Delayed introduction of solids is associated with reduced odds of child overweight/obesity. Wider promotion of current infant feeding guidelines could have a significant impact on the rates of child overweight and obesity.
Child obesity is now a major public health issue globally, with an established epidemic in developed countries, and similar trends evolving in many developing nations.1 It is associated with adverse health consequences in childhood,2 and is likely to be carried into adulthood.3, 4 Effective evidence-based treatments for child obesity are lacking;5 prevention needs to be an important strategy in tackling this problem.
There is longstanding debate over whether infant feeding practices might influence subsequent obesity.6 Reviews have found that crude associations between breastfeeding and reduced child obesity were significantly attenuated by adjustment for potential confounders.7, 8 A recent large randomized trial found increased duration and exclusivity of breastfeeding was not associated with reduced adiposity measures at age 6.5 years,9 suggesting that earlier published associations may be explained by residual confounding. One such unmeasured confounder may be the timing of introduction of solid foods, which has been less extensively studied in relation to child obesity than breastfeeding,10, 11, 12, 13 and has not yet been synthesized into a published review. Detailed prospective studies are needed to better understand these relationships and to provide an evidence base for the promotion of current infant feeding guidelines.14 Therefore, we performed an opportunistic analysis of data from the Melbourne Atopy Cohort Study (MACS), a longitudinal study conducted from birth to 10 years, to determine the association between infant feeding practices and child overweight/obesity.
Materials and methods
The MACS enrolled the children of 620 pregnant women at antenatal clinics in Melbourne, Australia, from 1990 to 1994. Subjects were recruited if they were at risk of developing atopy because of having a first-degree relative with asthma, atopic eczema, hayfever or food allergy. All women were recruited with informed consent, and the Mercy Hospital ethics committee provided ethics approval for the study.
Demographic information and parental smoking status was recorded at the time of recruitment. All mothers were encouraged to breastfeed exclusively for the first 6 months. Four-weekly telephone interviews were used to record detailed infant feeding practices up to the age of 64 weeks, with further interviews at 78 weeks and 2 years of age. The age in weeks at which breast milk and infant feeding formula were commenced and age in weeks at first ingestion of solid foods were recorded at these interviews. In 1997, parents were asked how many months their child had attended regular childcare during the first 6 months of life, the second 6 months of life and the second year of life. Data on weight and length at birth were retrospectively collected from health records in a subset of subjects. At a mean 10 years of age (s.d.: 1.9), height and weight were measured as part of lung function assessment in 307 subjects who were able to be followed up.
Exclusive breastfeeding was defined as no other oral intake apart from breast milk or water. Any breastfeeding was defined as oral intake of any quantity of breast milk, in combination with any other solids or liquids. Solid foods were defined as any food other than milk, including, for example, baby cereals and pureed foods. Duration of exclusive breastfeeding, any breastfeeding and age at first introduction of any solid diet were all measured in weeks since birth.
Fathers’ occupational status was classified according to the Australian National University 3_2 (ANU3_2) scale.15 Regular attendance at childcare was defined as attendance at childcare outside the home with at least two other children for 1 or more months.
Height and weight data collected at mean age of 10 years were used to calculate body mass index (BMI) SDS (z-scores), and these were further classified into three internationally age- and sex-standardized BMI categories—normal weight, overweight and obese.16 The overweight and obese categories were combined to create a binary outcome variable of ‘healthy’ or ‘above healthy’ BMI.
To explore differences in characteristics between the final study cohort and the group lost to follow-up, baseline demographic characteristics and infant feeding practices were compared between the two groups. In a subgroup for whom weight and length at birth and data on childcare attendance were available, these were also compared between the groups. Independent group t-tests, Mann–Whitney U-tests or χ2 tests were used as appropriate for the data.
Multiple logistic regression modelling was used to determine whether infant feeding practices are associated with the risk of above healthy BMI. Feeding practices were treated as continuous exposures. To test for non-linear relationships, Stata17 was used to fit 44 different fractional polynomial models. Retention of a linear exposure term was appropriate in all cases. The following potential confounders were considered: maternal education, type of residence, father's occupational status, parental smoking, regular childcare attendance and BMI at birth. They were included in the final model if:
there was a strong a priori reason for retaining them in the model (considered to be the case for markers of socioeconomic status) or
they were associated with the outcome in univariate analysis (odds ratio of >2 or P-value for odds ratio <0.1) or
they confounded the relationship between any of the feeding practices and risk of the above healthy BMI (change in odds ratio >15%).
Variables between which there was a biologically plausible rationale for effect modification were tested through the inclusion of interaction parameters in the models using the likelihood ratio test.
All analysis was performed using Stata 9.1.17
Comparison of demographic characteristics of the group followed to 10 years (final cohort) and those lost to follow-up show mothers of those lost to follow-up were on average younger, less likely to be married and of lower socioeconomic status than the final cohort (data not shown). They also breastfed for a shorter duration and introduced solids earlier, though very few subjects in either group breastfed exclusively for 6 months, or delayed the introduction of solids to this time (Table 1).
The final cohort was largely of Caucasian ethnicity (305/307). At mean age of 10 years follow-up, approximately 28% (n=85) of subjects were above healthy BMI for their age and sex: 25.1% (n=77) overweight and 2.6% (n=8) obese.
In univariate analysis, both duration of exclusive breastfeeding and age at introduction of solid foods were associated with increased risk of above healthy BMI (Table 2). Of the potential confounders examined (data not shown), only parental smoking was independently associated with the outcome. There were 17 (5.6%) mothers who smoked and 48 (15.7%) fathers who smoked. In multivariate analysis, neither duration of exclusive nor any breastfeeding was associated with above healthy BMI; however, the age at introduction of solid foods and parental smoking remained associated with the outcome. Separating parental antenatal smoking into maternal or paternal smoking gave adjusted odds ratios (aORs) of 1.647 (95% CI=0.531–5.105, P=0.387) and 2.336 (95% CI=1.151–4.739, P=0.019), respectively; all other associations remained unchanged (data not shown).
The mean age at introduction of solids was 20.49 (95% CI=19.96–21.02) weeks for the healthy BMI group, and 18.67 (95% CI=17.62–19.72) weeks for the above healthy BMI group.
There was neither evidence of interaction between the duration of any or exclusive breastfeeding and the age at introduction of solid foods, nor between any of the infant feeding variables and parental smoking, childcare attendance or birth BMI (P-value for interaction terms all >0.2). There was no evidence of non-linear associations between feeding practices and risk of above healthy BMI.
In a subset of 178 subjects for whom it was known, neither birth weight, nor ponderal index at birth was associated with above healthy BMI at age 10 years, and in a subanalysis of these subjects, all associations in the earlier multivariate model remained of similar magnitude and direction (data not shown).
This study found strong evidence that delaying the introduction of solid foods in infancy is associated with reduced odds of above healthy BMI at age 10 years, independent of the effects of breastfeeding, parental smoking, socioeconomic status, childcare attendance and birth weight. The duration of exclusive or any breastfeeding was not independently associated with above healthy BMI at age 10 years. This supports the lack of association between breastfeeding and child overweight found by a recent large randomized trial,9 and suggests that the function performed by the timing of introduction of solid foods in the development of child obesity requires greater attention.
An observational study such as this cannot show causality; however, its longitudinal nature confirms the temporality of the association. Further, the magnitude is of potential public health significance. An aOR of above healthy weight of 0.903 (95% CI=0.841–0.970) per week translates to an aOR of 0.657 (95% CI=0.491–0.881) per month.
The median age at introduction of solid foods in this cohort was 20 weeks.
The prevalence of above healthy weight in the group who started solids at or before 20 weeks was 59/170=34.7%. The prevalence of above healthy weight in those who had delayed the introduction of solids by 4 weeks, commencing at 24 weeks or later, was 13/67=19.4% (relative risk=0.6). This represents an absolute risk reduction for above healthy weight of 15.3%, and, if causal, a number needed to treat 6.5, that is for every 6.5 children who delay the commencement of solids from 20 to 24 weeks of age, 1 could benefit through maintaining a healthy weight at age 10 years. This effect would clearly have a significant impact at the population level.
Few other published studies have examined the function of solids in overweight/obesity and measured outcomes after 5 years of age, and none have recorded the age at commencement of solids with the precision of a continuous variable measured in weeks. In a large cohort study, Wilson et al.10 found that those who started solids before 15 weeks had higher percentage of body fat and weight (but not BMI) at age 7 years than those who started at or after 15 weeks. In contrast, Burdette et al.,11 using dual-energy X-ray absorptiometry, found neither breastfeeding nor the timing of introduction of solid foods (before or after 4 months of age) had an effect on adiposity at age 5 years. Reilly et al.12 found no relationship between the timing of introduction of ‘complementary feeding’ (not defined) in 2-month increments and BMI aged 7 years. Total dietary energy intake at 4 months predicts continued postnatal weight gain and childhood BMI.18
Strengths of our study include frequent prospective data collection of feeding practices, information on many important potential confounders and measurement of outcomes using a standardized tool with clinical relevance, at an age beyond early childhood, and, therefore, after the time of adiposity rebound. We have also assessed dose–response relationships, including the possibility of non-linear associations, between the duration of feeding practices and risk of above healthy BMI.
The main limitation of this study is the inability to control for maternal or paternal BMI, which is known to be significantly associated with child (and adult) obesity.19 However, Burdette et al.11 found that obese mothers were not less likely to follow American Academy of Pediatrics infant feeding recommendations in relation to solids than non-obese mothers. This suggests that it is relatively unlikely (although still possible) that the important observed association is due to confounding. Our study is relatively small; however, the associations, which have been shown, are likely to persist and be replicated with even greater precision in larger studies.
The cohort we examined was recruited by virtue of being at risk of atopy. The generalizability of these results is, therefore, unknown, but we see no obvious reason for the observation applying specifically to those with atopic risk. Further, we have no information as to why solids were introduced when they were; perceived satiety may well have been a factor, with early introduction of solids being a marker for ‘hungry’ babies whose appetites continue to be bigger as children. This would still be important information to confirm through further studies. Ultimately, given the size of the effect, we describe potentially of having significant public health benefits; randomized controlled trials using promotion of infant feeding guidelines as the intervention are needed to examine causality.
Finally, attrition bias is a potential concern in this study, which had approximately 50% loss to 10-year follow-up. However, the group lost to follow-up commenced solids on average 1 week earlier than those who were followed. Hence, for the observed results to be spurious, the lost group would need to have had significantly lower rates of above healthy BMI at age 10 years, which seems unlikely, given their sociodemographic characteristics. On the contrary, it is highly possible that the observed effect may have been larger if the full cohort had been followed.
The mechanisms through which the timing of introduction of solid foods could influence child overweight and obesity are uncertain. The effect is probably not mediated through infant growth.20 Randomized trials in both formula-fed and breastfed infants have shown no significant difference in infant growth parameters up to 6 months of age in those who start solids earlier (3–4 months) compared with later (6 months),21, 22 as complementary solid foods displace breast milk or formula so that net caloric intake remains similar. However, without any calorie increase, it is plausible that the introduction of specific dietary components in solid foods or resultant alterations in gut flora may result in epigenetic modification of metabolic programming, as shown in rodent models.23, 24, 25, 26 These early feeding-related changes may have lifelong detrimental effects. Other hypotheses relate to specific nutrients, for example excess protein intake in infancy may increase insulin-like growth factor-1 levels, leading to increased differentiation of preadipocytes into adipocytes, and leading to earlier age at adiposity rebound.27 Specific alterations in the secretion of gastrointestinal hormones may also be involved. For example, ghrelin is a peptide, which stimulates food intake and increases adiposity. Fak et al.24 have shown that gastric expression and plasma levels of ghrelin increase in rats at the time of introduction of solid food into their diet, and that delayed weaning is associated with reduced grehlin levels. Prenatal nutritional status and interactions with the child's early environment may have profound and prolonged effects on obesity, weight distribution and associated metabolic disease; these changes may well contribute to the array of emerging problems in developing and indigenous communities. Animal models of alterations to early life feeding clearly show sustained changes in appetite setting and defended weight trajectory.28, 29 Early life may provide a time period for truly effective preventative interventions.30
Antenatal parental smoking was strongly associated with the above healthy BMI in childhood in this cohort, independent of socioeconomic factors and feeding practices. Parental smoking is likely to cluster with other unhealthy behaviours, which may explain some of this association; however, there is also an emerging biological rationale by which smoking itself may affect child overweight and obesity.31 Most human studies have examined the function of maternal smoking, and generally found a positive association;32, 33 however, our study found that paternal smoking was a strong risk factor. A recent study has also described an independent positive association between paternal smoking and obesity in early school age children.34 The strong message regarding smoking and early childhood health should be targeted to both parents.
Breastfeeding is important for many reasons and rightly deserves promotion on public health grounds, but based on the findings of this study, it is unlikely to impact significantly on the obesity epidemic. Parental smoking should similarly be discouraged for a myriad of well-described reasons. Pre-eminently, however, our important finding suggests that promotion of the simple and no-cost act of delaying the introduction of solid foods to the age of 6 months may have a significant effect on the present epidemic of child overweight and obesity.
WHO Consultation on Obesity. Obesity: Preventing and Managing the Global Epidemic: Report of A WHO Consultation WHO Technical Report Series. World Health Organization: Geneva, 2000.
Dietz WH . Health consequences of obesity in youth: childhood predictors of adult disease. Pediatrics 1998; 101: 518–525.
Guo S, Roche A, Cameron C, Gardner J, Siervogel R . The predictive value of childhood body mass index values for overweight at age 35y. Am J Clin Nutr 1994; 59: 810–819.
Whitaker R, Wright J, Pepe M, Seidel K, Dietz W . Predicting obesity in young adulthood from childhood and parental obesity. New Engl J Med 1997; 337: 869–873.
Summerbell C, Ashton V, Campbell K, Edmunds L, Kelly S, Waters E . Interventions for treating obesity in children. Cochrane Database Syst Rev 2003; CD001872.
Kramer M . Do breast-feeding and delayed introduction of solid foods protect against subsequent obesity? J Pediatr 1981; 98: 883–887.
Arenz S, Rückerl R, Koletzko B, von Kries R . Breast-feeding and childhood obesity—a systematic review. Int J Obesity 2004; 28: 1247–1256.
Owen G, Martin R, Whinchup P, Smith G, Cook D . Effect of infant feeding on the risk of obesity across the life course: a quantitative review of published evidence. Pediatrics 2005; 115: 1367–1377.
Kramer MS, Matush L, Vanilovich I, Platt RW, Bogdanovich N, Sevkovskaya Z et al. Effects of prolonged and exclusive breastfeeding on child height, weight, adiposity, and blood pressure at age 6.5 y: evidence from a large randomized trial. Am J Clin Nutr 2007; 86: 1717–1721.
Wilson A, Forsyth J, Greene S, Irvine L, Hau C, Howie P . 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.
Burdette HL, Whitaker RC, Hall WC, Daniels SR . Breastfeeding, introduction of complementary foods, and adiposity at 5 y of age. Am J Clin Nutr 2006; 83: 550–558.
Reilly J, Armstrong J, Dorosty A, Emmett P, Ness A, Rogers I et al. Early life risk factors for obesity in childhood: cohort study. BMJ 2005; 330: 1357–1363.
Hediger ML, Overpeck M, Kuczmarski R, Ruan W . Association between infant breastfeeding and overweight in young children. JAMA 2001; 285: 2453–2460.
Dewey KG . Guiding Principles for Complementary Feeding of the Breastfed Child. Pan American Health Organisation/World Health Organisation: Washington, DC, 2003.
McMillan J, Jones F . The ANU3_2 scale: a revised occupational status scale for Australia. J Sociol 2000; 36: 64–80.
Cole T, Bellizzi M, Flegal K, Dietz W . Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 2000; 320: 1240–1245.
StataCorp. Stata Statistical Software: Release 9.1 9.1 edn. College Station: Texas, 2006.
Ong KK, Emmett PM, Noble S, Ness A, Dunger DB . Dietary energy intake at the age of 4 months predicts postnatal weight gain and childhood body mass index. Pediatrics 2006; 117: e503–e508.
Parsons T, Power C, Logan S, Summerbell C . Childhood predictors of adult obesity: a systematic review. Int J Obesity 1999; 23: S1–S107.
WHO Working Group on the Growth Reference Protocol and the WHO Task Force on Methods for the Natural Regulation of Fertility. Growth of healthy infants and the timing, type, and frequency of complementary foods. Am J Clin Nutr 2002; 76: 620–627.
Mehta K, Specker BL, Bartholmey S, Giddens J, Ho M . Trial on timing of introduction to solids and food type on infant growth. Pediatrics 1998; 102: 569–573.
Cohen RJ, Brown KH, Canahuati J, Rivera LL, Dewey KG . Effects of age of introduction of complementary foods on infant breast milk intake, total energy intake, and growth: a randomised intervention study in Honduras. Lancet 1994; 344: 288–293.
Levin BE, Keesey RE . Defense of differfing body weight set points in diet-induced obese and resistant rats. Am J Physiol Regul Integr Comp Physiol 1998; 274: R412–R419.
Fak F, Friis-Hansen L, Westrom B, Wierup N . Gastric ghrelin cell development is hampered and plasma ghrelin is reduced by delayed weaning in rats. J Endocrinol 2007; 192: 345–352.
De Giorgio MR, Yoshioka M, St-Amand J . Feeding induced changes in the hypothalamic transcriptome. Clin Chim Acta 2009; 406: 103–107.
Haemer MA, Huang TT, Daniels SR . The effect of neurohormonal factors, epigenetic factors, and gut microbiota on risk of obesity. Prev Chronic Dis 2009; 6: A95.
Rolland-Cachera M, Deheeger M, Bellisle F . Increasing prevalence of obesity among 18-year-old males in Sweden: evidence for early determinants. Acta Paediatrica 1999; 88: 365–367.
Chen H, Simar D, Lambert K, Mercier J, Morris MJ . Maternal and postnatal overnutrition differentially impact appetite regulators and fuel metabolism. Endocrinology 2008; 149: 5348–5356.
Levin BE, Keesey RE . Defense of differing body weight set points in diet-induced obese and resistant rats. Am J Physiol 1998; 274: R412–R419.
Gluckman PD, Hanson MA . Developmental and epigenetic pathways to obesity: an evolutionary-developmental perspective. Int J Obes (Lond) 2008; 32 (Suppl 7): S62–S71.
von Kries R, Toschke A, Koletzko B, Slikker WJ . Maternal smoking during pregnancy and childhood obesity. Am J Epidemiol 2002; 156: 954–961.
Ino T . A meta-analysis of association between maternal smoking during pregnancy and offspring obesity. Pediatr Int 2010; 52: 94–99.
Koupil I, Toivanen P . Social and early-life determinants of overweight and obesity in 18-year-old Swedish men. Int J Obes (Lond) 2008; 32: 73–81.
von Kries R, Bolte G, Baghi L, Toschke AM . Parental smoking and childhood obesity—is maternal smoking in pregnancy the critical exposure? Int J Epidemiol 2008; 37: 210–216.
We thank MACS chief investigators Dr David Hill and Dr Cliff Hosking; Christine Axelrad for recruitment and data collection; Dr John Thorburn for assistance in patient recruitment and administrative assistance; the Mercy Hospital Department of Obstetrics for participant recruitment; Anne Balloch for data management and all of the MACS children and parents for their participation and ongoing support for this study.
Initial development of MACS was supported by Nestlé Australia. The Asthma Foundation of Victoria supported the 10-year follow-up. The sponsors of the study had no role in the study design, data collection, data analysis, data interpretation or writing of the report. The corresponding author had full access to all data and final responsibility for the decision to submit for publication.
About this article
Cite this article
Seach, K., Dharmage, S., Lowe, A. et al. Delayed introduction of solid feeding reduces child overweight and obesity at 10 years. Int J Obes 34, 1475–1479 (2010). https://doi.org/10.1038/ijo.2010.101
- educational diet
- complementary feeding
- body weight
Minerva Medica (2020)
International Journal of Qualitative Studies on Health and Well-being (2019)
Extensive literature search as preparatory work for a systematic review on health outcomes related to the age of introduction of complementary food for the scientific assessment of the appropriate age of introduction of complementary feeding into an infant's diet
EFSA Supporting Publications (2019)
Maternal smoking during pregnancy and offspring body mass index and overweight: a systematic review and meta-analysis
Cadernos de Saúde Pública (2019)
The relationship between parental locus of control and adolescent obesity: a longitudinal pre-birth cohort
International Journal of Obesity (2019)