This study demonstrates that breastfeeding duration is directly associated with food variety at two years of age, independent of maternal characteristics known to be associated with diet quality in young children [19
]. To the best of the authors’ knowledge this is the first study to investigate the association of breastfeeding duration and food variety in toddlers. However, ever having been breastfed has been positively associated with a healthier dietary pattern amongst older Australian children [17
A plausible biological mechanism for these associations can be explained as follows. Children’s food preferences are shaped by repeated exposure of food and flavours pre- and post-natally [16
], particularly during the weaning period [9
]. Flavours from the mother’s diet are transmitted via amniotic fluid [21
] and breast milk [15
]. Breastfed infants, unlike those who are solely fed formula, are exposed to a variety of flavours well before they are introduced to solid foods. Consequently, breastfeeding is linked with greater acceptability of new food and flavours during the weaning period, a theory which is supported by evidence from animal studies [23
] and experimental studies in humans [21
In this study breastfeeding duration was shown also to be associated with the intake of a greater number of fruits and vegetables. This finding is consistent with the results of a number of studies which have shown breastfed infants to be more accepting of novel fruits and vegetables during the weaning period than formula fed infants [16
]. However, any benefits associated with breastfeeding may be relatively short-lived. Hausner and colleagues [31
] demonstrated that while breastfed infants initially had a higher acceptance of a novel flavoured (caraway) puree than formula-fed infants this difference in acceptance disappeared after repeated flavour exposure. Nevertheless, it may take up to 10 repeated exposures to achieve this equalisation [9
] and it is unclear whether in non-experimental settings parents of infants will persevere to this extent when introducing new food items into their child’s diets. Breastfeeding may contribute to higher food variety by reducing food neophobia and facilitating the transition to solid foods with lower resistance on the part of the breastfed infant.
Apart from breastfeeding duration, a number of maternal and family characteristics were associated with children’s food variety. Maternal educational attainment was the strongest independent predictor of the CFVS and also an independent predictor of FVVS. This finding is consistent with the literature which suggests that maternal education is strongly directly associated with children’s diet quality [19
]. Studies have demonstrated that more educated mothers tend to follow adult healthy eating guidelines and to eat healthier than less educated mothers and that these differences are reflected in the diets of their children [20
]. Consistent with the literature, this study also found that food variety, both CFVS and FVVS, amongst 2 years old children was directly associated with maternal age [20
] with better food variety observed in children of older mothers.
Another strong determinant of food variety was the presence of older siblings. This study showed that a higher CFVS and FVVS were associated with the absence of older siblings. Robinson and colleagues [20
] showed that compared to infants with older siblings, firstborn infants were “more likely to be fed a diet that complied with feeding guidelines and less likely to have a diet characterised by energy-dense, low-micronutrient food”. Similarly, Koh and colleagues [33
] reported that infants with older siblings were more likely to have received non-core foods earlier than firstborn infants and a study of English 3 year-olds [19
] reported that children with older siblings were more likely to receive a diet comprised primarily of snacks and finger foods and less likely to consume a “traditional” British “meat and two veg” diet. North and colleagues [19
] provide two possible explanations for this association. First, mothers who have to attend to the needs of more than one child may not have the time to prepare a traditional family meal and second older siblings may bring “snack” and “junk” foods into the household through for instance, mothers capitulating to the pestering of older children to buy these foods which are highly advertised.
As with all studies this study has a number of limitations. A major limitation was that the CFVS and FVVS were derived from a single 24 h recall that was completed by each infant’s mother. The intake data were of varying quality and for most cases poorly quantified. As a result each variety score reflects exposure to a particular food and not whether an infant had consumed a nutritionally meaningful amount of this food. An assumption is made that food variety is directly related to diet quality but we were unable to validate this assumption in our study. Other studies however, have demonstrated that this assumption in general holds true [4
]. Furthermore, the analysis sample is not representative of all participants in the Raine study or the general population. Young and less educated women were significantly less likely to complete the 24 h recall than older and better educated women. Both maternal age and education as well as a variety of other socio-demographic variables have been associated with breastfeeding practices [35
] and it may be other factors not controlled for in this study that influence both a mother’s decision not to breastfeed and her choice of foods which she feeds her child. Given the modest effect size demonstrated, the possibility of residual confounding cannot be discounted. Another limitation of this study was that few women (<10%) chose not to breastfeed providing a relatively small sample of women to compare against in order to understand the association of breastfeeding and food variety. While the breastfeeding initiation and duration rates reported in this study are not dissimilar to national [36
] and local [37
] data of the day they are high relative to rates reported for the UK and the USA [38
]. It is recommended that the association between breastfeeding duration and food variety and/or diet quality be investigated further in populations with more diverse breastfeeding practices and socio-demographic characteristics.
Breastfeeding is recognised as the optimal way to nourish infants [39
] and compared to formula fed infants, breastfed infants have a lower risk of a variety of infections in infancy [39
], and obesity, hypertension and raised cholesterol in later life [41
]. Diet variety is a marker of diet quality [4
] which has been associated with obesity among adolescents [6
] and young adults [7
], and risk of chronic disease such as colorectal cancer in adults [8
]. Given our finding that breastfeeding duration is directly associated with diet variety at 2 years of age and that eating patterns have been shown by others to track into later life [10
], this study provides further justification for the promotion and support of breastfeeding.