Protective effect of breastfeeding with regard to children’s behavioral and cognitive problems

Studies suggest that there is an association between breastfeeding and a wide range of positive health outcomes in children, including a lower risk of acute ear infections, respiratory tract infections, asthma, obesity, diabetes mellitus, and leukemia [1]. In addition, breastfeeding has been associated with greater cognitive development in childhood [2‐6]. In the Port Pirie Cohort Study, Wigg et al. [7] reported that breastfed children had a 1.2 (-2.0 to 4.4) and 0.8 (-1.9 to 3.5) point advantage on the Wechsler Full-Scale IQ test at 7 years of age and 11 to 13 years of age. Breastfeeding has also been associated with attention-deficit hyperactivity disorder (ADHD) [8, 9] and other externalizing behavioral problems [10] and internalizing behavioral problems [10, 11]. However, due to these behavioral outcomes being strongly related to the child’s IQ [12, 13], which is highly associated with breastfeeding [4, 7], it is uncertain whether the association between breastfeeding and the child’s behavioral development is mediated or confounded by the child’s IQ. By contrast, it is possible that inattention problem of children reduce IQ scores by interfering with the child’s capacity to learn in school. For example, Friedman et al. [14] investigated the relation between attention problem at several time points during childhood and IQ scores at age 16 in a large sample, and found longitudinal correlations between -0.21 and -0.27. Similarly, Polderman et al. [15] reported longitudinal negative correlations between attention problem measured by teachers and parents at age 5, and IQ scores at age 12 (r = -0.30 and -0.31, respectively).

In addition, due to the mother’s IQ being highly predictive of both breastfeeding status and the child’s IQ, the mother’s IQ should be considered when examining the association between breastfeeding, the child’s IQ, and the child’s behavioral outcomes [4, 16]. For example, in the 1979 US national longitudinal survey of youth, Der et al. [4] reported that before adjustment for maternal intelligence, breastfeeding was associated with an increase of around 4 points in the child’s IQ, but adjustment for mother's IQ reduced this advantage by 71% to 75%. The results of meta-analysis of eight studies that controlled for maternal IQ also shows that maternal IQ explains most of the effect of breastfeeding, but not all [4].

In this investigation, we attempted to further clarify the complex relationships between breastfeeding, behavioral problems, and the mother’s and child’s IQs among school-aged children in Korea. We tested the following hypotheses: 1) breastfeeding is independently associated with behavioral problems and ADHD, even after controlling for child’s IQ and maternal IQ and 2) breastfeeding is independently associated with child’s IQ, even after controlling for the diagnosis of ADHD and maternal IQ.

Of the initially recruited 1,089 children, 215 study subjects were excluded due to their responses being incomplete, leaving a total of 874 subjects (80.3%; 509 boys, 365 girls; mean age 9.05 ± 0.71 years) in the analysis. Among these 874 participants, 522 (59.7%) children were breastfed and 352 (40.3%) were not breastfed during infancy. There were no significant differences in socio-demographic characteristics between breastfeeding and non-breastfeeding groups. However, both the mother’s and the child’s IQs were significantly higher in the breastfeeding group than the non-breastfeeding group (Table 1).

Figure 1 shows the prevalence estimates of ADHD and the morbidity of internalizing, externalizing, or overall behavioral problems according to the child’s IQ. All behavioral problems were the most common in children with IQ <100 and the least common in children with IQ >115, although only the prevalence of internalizing problems was significantly different between IQ groups (X² = 9.33, p = 0.009).

After adjusting for gender, age, residential area, and yearly family income (Model 1), a lack of breastfeeding was associated with increased risk of morbidity of internalizing (AOR = 2.45, 95% CIs = 1.30-4.60), externalizing (AOR = 2.20, 95% CIs = 1.03-4.70) and overall (AOR = 3.36, 95% CIs = 1.53-7.34) behavioral problems as well as the diagnosis of ADHD (AOR = 2.01, 95% CIs = 1.07-3.79). After adjusting for the child’s IQ (Model 2), associations between a lack of breastfeeding and the morbidities of internalizing or overall behavioral problems weakened but remained significant (AOR = 2.30, 95% CIs = 1.22-4.31and AOR = 2.88, 95% CIs = 1.38-6.62, respectively). Significant associations between a lack of breastfeeding and the morbidities of externalizing problems or ADHD were eliminated in this model. These trends preserved even after adjusting for maternal IQ in addition to the child’s IQ and other covariates (Model 3) (Table 2).

Logistic regression analyses revealed that after adjusting for gender, age, residential area, and yearly family income (Model 1), a lack of breastfeeding was associated with increased risk of morbidity of internalizing (AOR = 2.45, 95% CIs = 1.30-4.60), externalizing (AOR = 2.20, 95% CIs = 1.03-4.70) and overall (AOR = 3.36, 95% CIs = 1.53-7.34) behavioral problems as well as the diagnosis of ADHD (AOR = 2.01, 95% CIs = 1.07-3.79). After adjusting for the child’s IQ (Model 2), associations between a lack of breastfeeding and the morbidities of internalizing or overall behavioral problems weakened but remained significant (AOR = 2.30, 95% CIs = 1.22-4.31and AOR = 2.88, 95% CIs = 1.38-6.62, respectively). Significant associations between a lack of breastfeeding and the morbidities of externalizing problems or ADHD were eliminated in this model. These trends preserved even after adjusting for maternal IQ in addition to the child’s IQ and other covariates (Model 3) (Table 2).

Linear regression analyses revealed that after adjusting for gender, age, residential area, and yearly family income, a lack of breastfeeding was associated with lower child’s IQ (unstandardized coefficient B = -4.17, 95% CIs = -6.07 to -2.28). This association slightly weakened but remained significantly after adjusting for maternal IQ (B = -3.82, 95% CIs = -5.23 to -1.68) and even after adjusting for the diagnosis of ADHD in addition to maternal IQ and other covariates (B = -3.32, 95% CIs = -5.10 to -1.55).

In this study, we found that a lack of breastfeeding was associated with increased morbidity of ADHD and internalizing and externalizing behavioral problems and low intelligence in childhood. Protective effect of breastfeeding on these behavioral problems might be partially mediated by the child’s IQ, whereas positive effect of breastfeeding on intelligence might be partially mediated by attention problem of children.

Existing research on the effect on breastfeeding on behavioral problems tends to focus on infancy and early childhood. Positive behavioral outcomes in breastfed infants include a greater degree of engagement and emotional regulation [22], fewer abnormal reflexes, signs of depression and withdrawal [23], and more alertness during social interactions [24]. In contrast, the results from a large, cluster-randomized trial did not find significant differences in behavioral outcomes at age 6 for those infants whose mothers were encouraged to breastfeed exclusively and for longer durations compared to infants whose mothers were not encouraged to do this [6]. In a pregnancy cohort study followed for 14 years, a shorter duration of breastfeeding was associated with increased morbidity of externalizing, internalizing, and overall behavioral problems, as measured by the CBCL [10]. The cited study controlled for several confounding variables, including family, social, economic, birth, and psychological factors in early life, but the study did not investigate the confounding effect of the mother’s or child’s IQ. Consistent with previous studies, we found a significant association between a lack of breastfeeding and the increased morbidities of externalizing, internalizing, and overall behavioral problems. The mother’s and child’s IQs explained some of the effect of breastfeeding, but not all. In particular, a lack of breastfeeding has sizable impact on internalizing or overall behavioral problems, independent of its effect on the child’s IQ.

ADHD is a common externalizing behavioral disorder that affects 8% to 12% of school-aged children [25]. ADHD is characterized by symptoms of inattention and/or hyperactivity/impulsivity. With an estimated heritability of approximately 75%, ADHD is generally regarded as having a genetic basis [26]. However, the remaining phenotypic variance (25%) in ADHD has been largely attributed to environmental factors, and several modifiable environmental risk factors have been suggested, such as tobacco smoke or alcohol exposure in pregnancy [27]. Previously, two studies examined the association between breastfeeding and ADHD. In one study, authors reported that the mean duration of breastfeeding was shorter in children with ADHD than that in controls (0.45 years vs. 0.55 years) [9]. In another study, authors found a significant association between ADHD and lack of breastfeeding at 3 months of age (OR = 3.08, 95% CI 1.46-6.50) [8]. However, neither study considered the confounding effect of the mother’s or child’s IQs.

Several possible bidirectional or interactive relationships could link breastfeeding with the presence of ADHD and the child’s IQ. One possibility is that lack of breastfeeding decreases intelligence [4, 7] and low child’s IQ increases the likelihood of ADHD [13]. Other possibility is that lack of breastfeeding increases the likelihood of ADHD [8, 9] and attention problem of children decreases the child’s IQ [14, 15]. Consistent with previous studies [8, 9], we found a significant association between lack of breastfeeding and ADHD, but this association was not statistically significant after controlling for the child’s IQ. A recent review of research on attention problems suggested that such research should control for IQ performance and, thereby, control for the potential influence of cognitive competence on ADHD symptoms [12]. We also found a significant association between lack of breastfeeding and low intelligence, in line with previous studies [4, 7]. This association weakened, but remained significant even after controlling for the presence of ADHD, suggesting sizable impact of breastfeeding on the child’s IQ, independent of its effect on the child’s attention problem.

This study has several limitations. First, due to the information on breastfeeding being based on the mothers’ recollection of the children, the respondents’ reports may be characterized by inaccuracies. Second, the small sample size of the morbidity group did not provide sufficient statistical power to detect modest differences. Third, we did not obtain information on the duration of breastfeeding, and we classified infant feeding only as ever being breastfed or never being breastfed. Failure to account for breastfeeding duration could lead to an underestimation of the true effect of breastfeeding on a child’s behavior. Finally, we did not obtain information on maternal education, home environment, maternal affective state, or mother-child attachment, which may have confounded the association between a lack of breastfeeding and child’s internalizing problems. Therefore, future prospective studies with detailed data regarding breastfeeding exposureand the number of additional confoundersare required to confirm our findings.

Despite these limitations, this is the first investigation of the effect of breastfeeding on behavioral and cognitive problems using an adequate control for maternal IQ and the child’s IQ or attention problem. Our findings suggest that there is a protective effect of breastfeeding on childhood behavioral outcomes with a partial mediation of this effect by the child’s IQ, and there is a positive effect of breastfeeding on childhood intelligence with a partial mediation of this effect by the child’s attention problem.