Breastfeeding is recommended as the optimal feeding method for the first six months of life, followed by the introduction of solids and continued breastfeeding for a minimum of one year.18
These recommendations are largely based on evidence that breast milk supports normal growth and also has immunological properties that provide some early protection from infection, and is associated with creating a lower risk of infant morbidity and mortality.19
A growing body of literature also suggests that breastfeeding affords a small, yet consistent, protective effect against obesity. Specifically, Christopher Owen and colleagues conducted a systematic review of sixty-one studies, of which twenty-eight provided odds ratios to examine the influence of breastfeeding on obesity from infancy to adulthood. They found that breastfeeding was associated with a reduced risk of obesity among infants, young children, older children, and adults with an unadjusted odds ratio of 0.50, 0.90, 0.66 and 0.80, respectively.20
Moreover, Stephan Arenz and colleagues reviewed twenty-eight studies investigating the association between breastfeeding and childhood obesity that met the following inclusion criteria: relative risk had to be reported, age at last follow-up had to be between five and eighteen years, feeding mode had to be reported, and obesity had to be defined using BMI. Of these twenty-eight studies, nine studies comprising more than 69,000 children were eligible for the meta-analysis. They found a significant adjusted odds ratio (AOR) for “ever breastfed” of 0.78, 95% CI (0.71-0.85) in the fixed model.21
These odds ratios, which are significantly lower than 1.0, indicate a significantly lower risk for subsequent obesity among those who were breastfed, even when adjusting for other factors.
In one review of twenty-two high quality studies, fifteen found protective effects to be more consistently noted among school-aged children and adolescents than among younger children.22
One possible explanation is that the impact of breastfeeding on subsequent weight status may be an example of metabolic or behavioral programming, in which the impact of breastfeeding on weight status only emerges later in development, and in this case, may not be clearly manifested until adolescence or adulthood. However, at this point, the mechanism(s) by which breastfeeding exerts protective effects are not understood. Specifically, breastfeeding is the ideal feeding method for the human infant and influences the developing anatomy and physiology of the gastrointestinal tract in ways that differ from formula feeding, such that breast-fed and formula-fed individuals may differ in the absorption and utilization of nutrients later in life.23
In addition, there is some evidence for two complementary behavioral mechanisms that may explain the protective effects of breastfeeding. The first involves the effects of breastfeeding on food acceptance and the second involves the developing controls of energy intake.
The sensory properties of breast milk may facilitate the transition to the modified adult diet. Many flavors of the maternal diet appear in breast milk. For example, adult sensory panels can detect odors of garlic,24
in milk samples of lactating women who ingested those flavors prior to providing milk samples. Flavors in human milk influence infant consumption. For example, breast milk flavored with garlic27
increased infant sucking time at the breast compared to breast milk without garlic or vanilla flavor. Mennella and colleagues also tested the hypothesis that experience with flavor in breast milk modifies the infants' acceptance and enjoyment of those foods in a sample of forty-five mothers and their babies that were randomly assigned to one of three groups. The first group drank carrot juice during pregnancy and water during lactation; group two drank water during pregnancy and carrot juice during lactation, and the control drank water during both conditions.29
Results revealed that repeated postnatal exposure to carrot flavors increased acceptance and enjoyment of carrot flavor in infant cereal. These findings indicate that flavors in breast milk, which vary with the maternal diet, provide the infant with a changing flavor environment. This early flavor experience appears to facilitate the infant's acceptance of foods of the modified adult diet, especially those foods consumed by the mother during lactation.30
In contrast to the varied flavor experience provided by breastmilk, formula provides the infant with the same consistent flavor experience.
There is limited evidence that these early differences in flavor experience provided by the breast and formula feeding also influence infants' subsequent acceptance of solid foods, especially those foods that might not otherwise be readily accepted, such as vegetables. For example, Susan Sullivan and Leann Birch conducted a short term longitudinal study of nineteen breastfed and seventeen exclusively formula fed four- to six-month-old infants and their mothers to examine the influence of milk feeding regimen and repeated exposure on acceptance of their first pureed vegetable. Participants were randomly assigned to be repeatedly fed one vegetable, either pureed peas or green beans. Results revealed infant feeding regimen moderated the effects of repeated exposure; the initial intake of vegetables did not differ between breastfed and formula-fed infants, but breastfed infants increased their intake more rapidly over days than formula fed infants, and continued to consume significantly more vegetables after ten exposures.31
These findings are consistent with the view that breastfeeding can more easily facilitate the acceptance of solid foods compared to formula feeding.
A second hypothesis regarding the protective effect of breastfeeding on later risk of overweight is that breastfeeding provides the infant with greater opportunity for self-regulation of intake. A limited body of evidence suggests that infants have some ability to self-regulate caloric intake by adjusting the volume of milk consumed,32
although this can be influenced by maternal feeding practices. In bottle feeding, the infant can obtain milk with less effort than from the breast, so the formula-fed infant is more passive in the feeding process and has fewer opportunities to control the amount consumed, making it easy to over-feed the infant. In contrast, the breastfed infant must take an active role in order to transfer milk from the breast. The higher levels of maternal control that are possible with bottle feeding reduce infants' opportunities to control the amount consumed at a feeding.33
Limited evidence indicates that bottle-fed infants consume more milk and gain weight more rapidly than breastfed infants, increasing their risk for childhood obesity.34
Moreover, research suggests the difference in milk intake between breastfed and formula-fed infants becomes greater with age.35
In short, while evidence is limited, breast feeding and formula feeding provide very different opportunities for early self-regulation of energy intake, and additional research is needed to assess how these differing feeding methods influence the developing controls of energy intake, weight gain, and risk for childhood obesity.
Whether and how infants exert control during feeding to regulate energy intake are not new questions. Clara Davis conducted seminal research in the late 1920s and 1930s, providing the first evidence of an unlearned ability to self-regulate energy intake in infancy. In Davis' studies, infants and toddlers grew well and had few illnesses when given the opportunity to select and consume a variety of simply prepared foods at each meal.36
As previously mentioned, Samuel Fomon and colleagues revisited the issue of self-regulation of energy intake by systematically varying the energy density of infant formula.37
By six weeks of age, full-term infants who were fed a concentrated formula (100 kcal/mL) consumed smaller volumes than did those infants who were fed a diluted formula (54 kcal/mL), such that total daily energy intake did not differ between the two groups. In 1977, observational data from Sharon Pearcey and John De Castro complemented these experimental findings, revealing that individual variability in energy consumed at meals among twelve-month-old infants was close to forty-seven percent, while variability in daily energy intake was seventeen percent.38
Similarly, Roberta Cohen and colleagues39
found no difference in daily energy intake among infants four to six months of age who were fed only breast milk versus those who were fed breast milk along with complementary foods, suggesting that infants were adjusting their intake of breast milk in response to the addition of solid foods.
The ability to regulate energy intake has also been described in preschool-age children. Children responded to covert manipulations in the energy content of foods served as first courses by adjusting their subsequent intake, such that their total energy intake for the meal and energy consumed over a thirty-hour period40
was maintained across conditions in which low- or high-energy foods were provided as a first course. Differences among preschool-age children in their ability to self-regulate energy intake have been associated with differences in weight status. For example, Susan Johnson and Leann Birch examined the influence of weight status on regulation of energy intake in seventy-seven three- to five-year-old children. Each child participated in two treatments, differing only in whether children received a low- or high-calorie preload of fruit flavored drinks of equal volume before lunch. After twenty minutes, children self-selected intake from a familiar lunch menus (i.e., turkey hot dogs, American cheese, unsweetened applesauce, carrot sticks, fruit bars, and 2% milk) to assess their ability to adjust food intake in response to changes in energy density of the preload drinks. They found that children who showed little evidence of adjusting their lunch intake in response to the energy differences in the preloads were significantly heavier.41
Leann Birch and Jennifer Fisher used a similar protocol to investigate the association between weight status and children's caloric compensation in a sample of 197 non-Hispanic white five-year-old girls. Data were used from two separate lunches which differed in whether a low- or high-energy preload drink was consumed prior to lunch. Again, after a brief delay, participants ate a self-selected lunch (i.e., sandwich, carrots, applesauce, cookies, and milk) ad libitum
. Results indicated substantial individual differences in the extent to which girls adjusted their energy intake at lunch in response to the differences in preload energy content. On average the girls only compensated for about half of the energy in the preloads. In this case, greater maternal restriction in feeding was associated with poorer compensation and higher weight status in daughters.42
While infants show a predisposition to respond to differences in energy density early in life, the child's early experience, including child feeding practices, shape the development of individual differences in self-regulation abilities.43
That infants and young children are capable of self-regulating energy intake under laboratory conditions, in the absence of adult intervention, and in the presence of only simply prepared healthy foods, does not speak to the extent to which this ability can be exercised in current family environments.