Milky ways: effects of maternal obesity on human milk composition and childhood obesity risk

See corresponding article on page 1009.

Breastfeeding is promoted by the American Academy of Pediatrics, Institute of Medicine, WHO, and other leading medical organizations as a way to curb childhood obesity (1–3). Although breastfeeding has numerous benefits for maternal and infant health, the evidence on obesity prevention is contradictory at best. Several large studies have documented protective effects (4), but meta-analyses have found evidence of publication bias (5, 6) and mixed effects on obesity (7). Potential benefits of breastfeeding on offspring BMI may be blunted in women with obesity or gestational diabetes mellitus (8). When confounders such as maternal BMI, education, household income, and smoking are taken into account, associations between breastfeeding and childhood obesity are attenuated or eliminated (9). Further, a large cluster-randomized trial of breastfeeding promotion in Belarus even suggested that prolonged breastfeeding was associated with increased BMI in adolescence (10).

There are several potential reasons for these inconsistent effects. For example, mothers who breastfeed may differ from those who opt for formula in ways that cannot be fully measured or adjusted for. It is also difficult, in human studies, to disentangle effects of breastfeeding from effects of prenatal exposures, or to rule out effects of inherited genetic risk factors for obesity. Moreover, breastfeeding may have different effects on obesity than consumption of pumped breast milk, but studies frequently do not distinguish between the two. Finally, it may be challenging to dissect effects of breastfeeding from effects of early compared with late introduction of complementary foods, which are also risk factors for childhood obesity.

Another potential reason for inconsistent associations between breastfeeding and childhood obesity is that human milk composition may vary. Over the past several years, interindividual variability in human milk composition has been an area of active research. It has been hypothesized that milk from obese mothers contains differential amounts of “obesogenic” compared with “protective” components. Indeed, breast milk contains not only nutrients, but also >1000 bioactive substances—including antibodies, hormones, cytokines, metabolites, nucleotides, microbiota, exosomes, etc.—with the potential to influence infant physiology and development. Understanding the degree to which maternal factors such as BMI, gestational weight gain, age, diabetes, and exercise may influence breast-milk composition, and how this may relate to early childhood weight gain, is an area with broad importance to maternal and child health. Such studies are essential to developing well-informed public health recommendations around breastfeeding for different subgroups of women. One might envision that, if the benefits of breastfeeding were more pronounced in certain subpopulations, targeted efforts to promote and support breastfeeding would be warranted.

Two publications (11, 12) in recent issues of The American Journal of Clinical Nutrition significantly advance our understanding of maternal obesity as an important determinant of human milk composition. The article by Daniel et al. (11) is a meta-analysis distilling the results of 66 studies, representing 4764 women, on the effects of maternal BMI (in kg/m²) on human milk macronutrient and energy content. The article by Saben et al. (12) moves beyond macronutrients to analyze the effects of maternal obesity on the human milk metabolome, and the associations of milk metabolites with infant body composition. The latter represents the largest study to date on the associations of human milk metabolites with maternal and infant adiposity. Together, these 2 studies support the importance of maternal obesity in shaping human milk composition and point to new links between human milk composition and early childhood obesity risk.

Daniel et al. systematically analyzed 69 articles reporting on 66 studies that included data on maternal BMI (or weight and height), together with quantitative data on breast-milk energy, fat, or protein content, from 1 to 6 mo postpartum. The meta-regression included 40 data points on the association of maternal BMI and breast-milk energy content, 63 data points for the effect on breast-milk fat content, and 40 data points for the effect on total protein content. Their results reveal robust positive associations between maternal BMI and breast-milk fat content: for each 1-kg/m² increase in maternal prepregnancy or postpartum BMI, breast-milk fat increased by 0.56 g/L. This translates to an estimated 16.5% increase in fat in human milk of a mother with a BMI of 30, compared with a mother with a BMI of ≤18.5. There were no observed associations between maternal BMI and breast-milk total protein or energy content. However, the authors reported that there was substantial heterogeneity between studies, resulting from differences in sample collection, assay methodology, and variation in exclusivity of breastfeeding.

In the study by Saben et al., the human milk metabolome was analyzed at 0.5 mo (n = 159), 2 mo (n = 131), and 6 mo (n = 94) postpartum, from women with normal (BMI < 25.0) compared with obese (BMI ≥ 30.0) prepregnancy BMI. Using an untargeted approach (GC-MS), they identified 115 known metabolite peaks and 240 unknown metabolites in human milk. Of these metabolites, 111 were associated with maternal BMI (P < 0.05). Milk from obese mothers was enriched in monosaccharides and sugar alcohols. The use of repeated milk samples allowed the group to discover a time-dependent effect of maternal BMI on milk composition, with a significant interaction of maternal BMI and time on metabolites related to the glucose–alanine cycle, amino acid metabolism (e.g., glutamate, glycine, serine), and the urea cycle. This suggests that maternal BMI may influence the maturation of human milk from early so-called “transitional” breast milk to “mature” milk. Intriguingly, a proportion of the metabolites that differed according to maternal weight status (e.g., mannose, lyxitol, shikimic acid) also predicted higher infant adiposity over the first 6 mo of life. These results suggest that certain milk metabolites might act as mediators of mother–child transmission of obesity risk, although this possibility requires further mechanistic studies in animal models and/or cell lines.

One of the methodological strengths of this study was the assessment of maternal adiposity using air displacement plethysmography in addition to BMI, given that there can be substantial variation in fat mass for a given BMI (14). Although the use of fat mass as the measure of maternal obesity yielded similar associations with milk metabolites as BMI, the adjusted correlation coefficients tended to be stronger with maternal fat mass. Another strength was that infant body composition was analyzed using MRI, which provided precise estimates of adipose tissue lipid content and fat-free mass. On the other hand, limitations of this study included attrition in the sample size from 0.5 (n = 159) to 6 mo (n = 94), which raises the possibility of selection bias, and the inclusion of both exclusively breastfeeding and mixed-feeding mothers.

Together, these 2 studies add to the growing evidence that maternal obesity is an important driver of interindividual variation in milk composition. Maternal BMI has previously been linked to differences in lipid composition, notably a higher proportion of SFAs and a lower ω-3:ω-6 PUFA ratio in milk (15). Maternal obesity has also been linked to increased milk content of insulin, leptin, C-reactive protein, IL-6, and TNF-α; lower adiponectin; and differences in oligosaccharide composition (16–18). Despite these findings, there has not been an obvious “smoking gun” linking milk composition to childhood obesity risk. For example, milk from obese mothers is higher in insulin, but insulin is inversely related to infant weight gain, suggesting that it may be a protective factor (16). Similarly, other milk components associated with maternal obesity, including leptin, adiponectin, and IL-6, show inverse associations with infant weight gain (19, 20). As such, the identification of maternal obesity–associated differences in milk composition that are in turn linked to infant adiposity lends new support to the hypothesis that some “obesogenic” components are more abundant in milk from obese mothers.

The results of both highlighted studies should be interpreted with caution, because they are based on associations, and cannot be used to infer causation. It also cannot be emphasized enough that any identified differences in milk composition between subgroups of women cannot and should not be used to discourage certain groups of women from breastfeeding. Indeed, breastfeeding has multiple beneficial effects on infant development, cognition, gut health, and immune function, as well as on maternal obesity, diabetes risk, and breast cancer risk. An important area for future studies will be to test whether targeted interventions in mothers, e.g., exercise, attenuation of gestational weight gain, or nutritional supplements, may allow mothers to optimize their milk composition to maximize benefits to their offspring.

Notes

Supported by National Institute of Child Health and Human Development grant HD091974 (to EI), the Harold Hamm Diabetes Foundation, and National Institute of Diabetes and Digestive and Kidney Diseases grant P30DK036836.