Human milk composition differs by maternal BMI in the first 9 months postpartum

ABSTRACT

Background

Studies indicate that maternal weight status modulates human milk composition; however, results are conflicting.

Objectives

Our objective was to examine the relation between maternal body composition and human milk macronutrients and bioactive components and also their association with infant daily intakes and body composition.

Methods

Human milk samples were obtained from a longitudinal study (NCT 01131117) in normal weight (NW: 18.5–24.9 kg/m², n = 88) and overweight/obese (OW: 25–35 kg/m², n = 86) women between 0.5 and 9 mo postpartum. Macronutrient content was estimated using mid-infrared spectroscopy. Leptin, insulin, and C-reactive protein (CRP) were measured using electrochemiluminescence immunoassays. Infant body composition was obtained using quantitative MRI. Linear mixed models were adjusted for postpartum age and infant sex.

Results

Human milk in OW mothers was higher in fat and protein and lower in carbohydrate content at some time points compared with that in NW mothers. Human milk leptin, insulin, and CRP concentrations were higher in OW mothers compared with NW mothers, with infants of OW mothers exposed to 1.5–2.5 times higher concentrations of leptin and insulin compared with infants of NW mothers. Similar results were observed when concentrations were normalized to infant daily intake and body weight. The effect sizes of infant daily intakes associated with infant growth parameters were small for macronutrients [0.005–0.05 z-score units and 0.02–0.45 fat mass index (FMI) or fat-free mass index units per unit of change in composition, P < 0.05]. Larger effect sizes were seen with human milk insulin and leptin (0.24 z-score units and 0.37–1.15 FMI units per unit of change in composition, P < 0.05).

Conclusions

These findings demonstrate that infants of OW mothers are exposed to higher concentrations of insulin, leptin, and, to a lesser extent, CRP. The bioavailability of these 3 human milk bioactives and their mechanisms of action in the infant are unclear.

This trial was registered at clinicaltrials.gov as NCT01131117.

Introduction

Human milk is the primary source of nutrition for infants, and it is recommended that infants be exclusively breastfed for the first 6 mo of life (1), a sensitive window of development during which nutritional exposures can elicit long-term effects on the infant's health (2). Human milk is a complex matrix containing macronutrients and bioactive molecules to support infant growth and development (3). Numerous studies have linked breastfeeding to obesity prevention in the offspring (4–6); however, other reports fail to find this association (7–9).

Human milk macronutrient composition is variable across lactation and between mothers due to genetic, nutritional, and environmental factors (10). Higher maternal body fat mass (percentage fat mass) (11, 12) and BMI (in kg/m²) (13, 14) have been associated with higher protein and fat concentrations in human milk in some studies but not in others (15–17). Even less robust evidence exists on how maternal overweight and obesity modulate milk macronutrient content over time, especially from transitional to mature milk.

Maternal body composition also influences human milk nonnutritive bioactive components believed to impact infant growth, such as insulin, leptin, and cytokines (15, 18, 19). Several studies have measured the association between maternal BMI and breast milk insulin concentrations, but the results are also contrasting (18–21). Regardless of maternal BMI, increased human milk insulin concentrations have been negatively associated with infant body composition outcomes at age 1 mo (19) but not at age 6 mo (18). The concentration of human milk leptin has also been positively associated with maternal overweight and obesity (11, 17, 19), and infant body composition has been shown to be inversely associated with milk leptin concentrations up to age 6 mo (18, 19). Concentrations of C-reactive protein (CRP) in human milk increase in proportion to maternal BMI (22), but it is not known if this relates in any way with offspring growth or body composition.

To our knowledge, no longitudinal studies have examined the impact of maternal BMI on human milk macronutrients and other bioactive components as well as their association with infant body composition, while considering milk composition from transitional to mature milk and milk intake. The objective of this study was to examine the association of maternal BMI on infant human milk energy intake from 2 wk to 9 mo postpartum as a primary outcome. The concentrations of human milk macronutrients (protein, fat, and carbohydrate) and nonnutritive factors (leptin, insulin, CRP, IL-6, IL-8, and TNF-α) were considered as secondary outcomes. Relations between human milk composition and infant body composition outcomes were also investigated. We hypothesized that human milk from mothers with overweight and obesity would have greater caloric, fat, protein, insulin, leptin, and CRP content compared with milk from mothers who are normal weight. We further hypothesized these alterations in human milk composition from mothers with overweight and obesity would be significantly associated with infant fat mass index (FMI).

Methods

Participants and study procedures

Participants were mother–child dyads enrolled in the longitudinal Glowing study (NCT01131117). Women were eligible if they had a BMI between 18.5 and 35 (Figure 1). Women were excluded from the study if they had pre-existing or ongoing medical conditions (e.g., diabetes mellitus and hypertension), complications during pregnancy, used medications during pregnancy that are known to influence fetal growth, smoked, drank alcohol, or were not second parity. Only healthy, full-term (≥37 weeks of gestation) children were included in the analyses. Participants were grouped by maternal BMI measured at first trimester as either normal weight (NW; BMI: 18.5–24.9) or overweight/obese (OW; BMI: 25–35). Women were enrolled between 2010 and 2014, and infants were followed until age 2 y (last study visit completed August 2017). Data from a total of 174 mother–child pairs were available for human milk analyses. The study procedures were approved by the Institutional Review Board of the University of Arkansas for Medical Sciences.

FIGURE 1.

Cohort flow diagram. LTFU, lost to follow-up.

Human milk

Participants were asked to collect a human milk sample at the second feeding of the day (or before 09:00) by expressing fully 1 breast at postnatal age 0.5, 1, 2, 3, 4, 5, 6, and 9 mo using either a manual or an electric pump. The samples were stored at −70°C until analysis. Macronutrients (fat, protein, and carbohydrates) were measured using a Miris Human Milk Analyzer (Miris) according to manufacturer's instructions, and caloric content was derived. Leptin, insulin, CRP, IL-6, IL-8, and TNF-α concentrations were measured using multispot assay kits (Meso Scale Diagnostics). Leptin, insulin, and CRP values were log-transformed for infant growth analyses.

Maternal body composition and gestational weight gain

Maternal anthropometrics and body composition were measured before or at 10 weeks of gestation. Maternal weight and height were measured with a standing digital scale (Tanita) and a wall-mounted stadiometer (Perspective Enterprises), respectively. Maternal fat mass was measured using air displacement plethysmography (BodPod; Cosmed). Gestational weight gain was calculated based on the weight gain between the participant's first study visit and week 36 of gestation. Using the Institute of Medicine gestational weight gain guidelines for NW and OW women, weight gain during pregnancy was categorized as inadequate, adequate, or excessive weight, while adjusting for gestational age (23).

Child body composition

At each postnatal research visit (0.5, 1, 2, 3, 4, 5, 6, and 9 mo postpartum), infant weight and length were measured using a tared scale (Seca) and a length board with a sliding foot piece (Perspective Enterprises), respectively. Length-for-age (LFA), weight-for-age (WFA), and weight-for-length (WFL) z scores were calculated based on the WHO Child Growth Standards (24). Infant fat and fat-free mass were obtained using quantitative NMR (EchoMRI-AH; Echo Medical Systems) as previously described (25). FMI and fat-free mass index (FFMI) were calculated by dividing fat mass or fat-free mass (in kilograms) by the participant's length squared (in meters).

Child dietary intake

Infant daily human milk intake was assessed by measuring the infant's weight before and after a nursing session combined with 3-d weighed food records. Infants were categorized as either exclusively breastfed or mixed fed based on their formula intake. If at any visit infants received >100 mL of formula per day, they were then considered a mixed feeder from that visit forward. Data from visits that lacked dietary intake data were not used in the daily intake analyses, but human milk composition was still considered. Infant formula intake was assessed using the 3-d weighed food records.

Self-reported outcomes

At 2 weeks postpartum, mothers reported their infants’ race, sex, birth weight, and birth length, as well as their delivery mode. Gestational age was calculated using the mother's last menstrual period and the child's date of birth.

Statistical analysis

Descriptive statistics (mean and SE of the mean or counts and percentage) were calculated for maternal demographic data, body composition, gestational weight gain, and gestational age, as well as for infant birth weight and length. To compare the NW and OW groups, t tests were performed for continuous data, and chi-square tests were used for categorical data (Table 1). Significance was set at P < 0.05. Linear mixed-effect models with Bonferroni-corrected post hoc pairwise comparisons were constructed to investigate any differences between the human milk composition of NW and OW mothers and the intake of infants of NW and OW mothers (Figures 2 and 3) while adjusting for postpartum age, infant feeding mode (exclusive compared with mixed), and infant sex (26, 27). Linear mixed-effect models were also constructed to determine the influence of human milk composition on infant growth parameters (LFA, WFA, WFL, FMI, and FFMI), while adjusting for infant feeding mode (exclusive compared with mixed) and infant sex (Tables 2 and 3). Statistical analysis was performed using R version 3.6.1 (R Foundation for Statistical Computing) (28, 29). Modeling was performed using lme4 version 1.1.21 using residual maximum likelihood estimates (26), lmerTest version 3.1.1 (30), and emmeans version 1.4.3.01 (27) R packages (R Foundation for Statistical Computing).

TABLE 1.

Maternal and infant characteristics¹

	All	NW (BMI <25 kg/m2)	OW (BMI 25–35 kg/m2)	P value2 (NW vs. OW)
n (%)	174	88 (50.6)	86 (49.4)
Maternal characteristics
Age at birth, y	30.43 ± 0.26	30.41 ± 0.36	30.46 ± 0.39	0.93
Race				0.87
Caucasian	150 (86.2)	75 (85.2)	75 (87.2)
Non-Caucasian	24 (13.8)	13 (14.8)	11 (12.8)
BMI, kg/m2	25.58 ± 0.31	22.27 ± 0.18	28.96 ± 0.32	<0.0001
Fat mass, %	25.00 ± 0.69	18.26 ± 0.49	31.81 ± 0.78	<0.0001
Gestational weight gain, kg	11.89 ± 0.30	12.63 ± 0.30	11.14 ± 0.51	0.013
Gestational weight gain, IOM category3				<0.0001
Inadequate	38 (21.8)	27 (30.7)	11 (12.8)
Adequate	82 (47.1)	52 (59.1)	30 (34.9)
Excessive	54 (31.0)	9 (10.2)	45 (52.3)
Gestational age at delivery, wk	39.26 ± 0.07	39.14 ± 0.11	39.37 ± 0.09	0.10
Delivery method				0.041
Vaginal	117 (67.2)	66 (75.0)	51 (59.3)
Cesarean section	57 (32.8)	22 (25.0)	35 (40.7)
Infant characteristics
Sex				0.78
Female	72 (41.4)	35 (39.8)	37 (43.0)
Male	102 (58.6)	55 (60.2)	49 (57.0)
Birth weight, kg	3.53 ± 0.04	3.50 ± 0.05	3.56 ± 0.05	0.39
Birth length, cm	51.10 ± 0.20	50.92 ± 0.32	51.29 ± 0.24	0.35
Sample size3
0.5 mo	157 [117]	77 [58]	80 [59]
1 mo	147 [115]	73 [55]	74 [60]
2 mo	136 [108]	67 [52]	69 [56]
3 mo	131 [104]	63 [51]	68 [53]
4 mo	121 [83]	60 [41]	61 [42]
5 mo	107 [72]	53 [32]	54 [40]
6 mo	107 [92]	55 [51]	52 [41]
9 mo	84 [41]	46 [18]	38 [23]

¹Data presented as counts (%) or means ± SEMs. IOM: Institute of Medicine; NW, normal weight; OW, overweight.

²To compare groups, t tests were performed for continuous data, and chi-square tests were used for categorical data.

³Sample sizes presented as number of milk samples available for human milk composition [number of participants with valid daily volume intake data].

FIGURE 2.

Estimated marginal means of human milk macronutrient concentration and intakes during the first 9 mo postpartum. Predicted macronutrient concentrations by maternal BMI (mean ± SEM) are based on linear mixed-effect models adjusting for postpartum age and infant sex. Infant daily intake models were also adjusted for infant feeding mode (exclusive compared with mixed). The solid lines indicate maternal NW (18.5–24.9 kg/m²), and the dotted lines indicate maternal OW (25–35 kg/m²). A significant interaction between postpartum age and maternal group was observed in the energy, fat, and protein intake models. For NW participants in composition models, n = 77, 73, 67, 63, 60, 53, 55, and 46 at 0.5, 1, 2, 3, 4, 5, 6, and 9 mo, respectively. For OW participants in composition models, n = 80, 74, 69, 68, 61, 54, 52, and 38 at 0.5, 1, 2, 3, 4, 5, 6, and 9 mo, respectively. For NW participants in intake models, n = 58, 55, 52, 51, 41, 32, 51, and 18 at 0.5, 1, 2, 3, 4, 5, 6, and 9 mo, respectively. For OW participants in intake models, n = 59, 60, 56, 53, 42, 40, 41, and 23 at 0.5, 1, 2, 3, 4, 5, 6, and 9 mo, respectively. At each time point, *P < 0.05, **P < 0.01. NW, normal weight; OW, overweight.

FIGURE 3.

Estimated marginal means of human milk leptin, insulin, and CRP concentrations and intakes during the first 9 mo postpartum. Predicted leptin, insulin, and CRP concentrations by maternal BMI (mean ± SEM) are based on linear mixed-effect models adjusting for postpartum age and infant sex. Infant daily intakes models were also adjusted for infant feeding mode (exclusive compared with mixed). Solid lines indicate maternal NW (18.5–24.9 kg/m²), and dotted lines indicate maternal OW (25–35 kg/m²). A significant interaction between postpartum age and maternal group was observed in the leptin intake model. For NW participants in composition models, n = 77, 73, 67, 63, 60, 53, 55, and 46 at 0.5, 1, 2, 3, 4, 5, 6, and 9 mo, respectively. For OW participants in composition models, n = 80, 74, 69, 68, 61, 54, 52, and 38 at 0.5, 1, 2, 3, 4, 5, 6, and 9 mo, respectively. For NW participants in intake models, n = 58, 55, 52, 51, 41, 32, 51, and 18 at 0.5, 1, 2, 3, 4, 5, 6, and 9 mo, respectively. For OW participants in intake models, n = 59, 60, 56, 53, 42, 40, 41, and 23 at 0.5, 1, 2, 3, 4, 5, 6, and 9 mo, respectively. At each time point, *P < 0.05,^**P < 0.01. CRP, C-reactive protein; NW, normal weight; OW, overweight.

TABLE 2.

Associations between infant daily intakes of macronutrients, leptin, insulin, and C-reactive protein and length-for-age, weight-for-age, and weight-for-length z scores from 0.5 to 9 mo postpartum¹

	Length-for-age		Weight-for-age		Weight-for-length
	β ± SE	P value	β ± SE	P value	β ± SE	P value
Fat, g/d	−0.003 ± 0.003	0.21	−0.005 ± 0.002	0.040	−0.001 ± 0.002	0.58
Carbohydrates, g/d	−0.009 ± 0.003	0.002	−0.006 ± 0.002	0.026	0.01 ± 0.002	0.003
Protein, g/d	0.05 ± 0.02	0.012	0.05 ± 0.02	0.013	−0.02 ± 0.02	0.12
Leptin, log[pg]/d	0.24 ± 0.12	0.047	0.19 ± 0.11	0.09	−0.03 ± 0.09	0.72
Insulin, log[pg]/d	−0.09 ± 0.13	0.50	−0.01 ± 0.11	0.94	−0.07 ± 0.09	0.43
C-reactive protein, log[ng]/d	0.05 ± 0.07	0.47	0.08 ± 0.06	0.20	0.05 ± 0.05	0.32

¹Linear mixed-effect models adjusted for infant feeding mode (exclusive compared with mixed) and infant sex.

TABLE 3.

Associations between infant daily intakes of macronutrients, leptin, insulin, and C-reactive protein and fat mass index and fat-free mass index from 0.5 to 9 mo postpartum¹

	Fat mass index2		Fat-free mass index3
	β ± SE	P value	β ± SE	P value
Fat, g/d	−0.0005 ± 0.003	0.88	0.0004 ± 0.002	0.88
Carbohydrates, g/d	0.06 ± 0.004	<0.001	0.02 ± 0.003	<0.001
Protein, g/d	−0.45 ± 0.03	<0.001	−0.08 ± 0.02	<0.001
Leptin, log[pg]/d	−1.15 ± 0.16	<0.001	−0.10 ± 0.11	0.40
Insulin, log[pg]/d	0.37 ± 0.17	0.028	−0.03 ± 0.12	0.79
C-reactive protein, log[ng]/d	0.10 ± 0.09	0.29	−0.11 ± 0.07	0.10

¹Linear mixed-effect models adjusted for infant feeding mode (exclusive compared with mixed) and infant sex.

²Fat mass calculated as fat mass (kg)/length squared (m²).

³Fat-free mass calculated as fat free mass (kg)/length squared (m²).

Results

Participant characteristics

Of the 320 women evaluated for eligibility, 284 were enrolled in the Glowing study and 224 delivered term infants following an uncomplicated pregnancy (Figure 1). Among these, 38 women did not breastfeed and 12 were lost to follow-up or did not provide a human milk sample. Thus, 174 women were included in the analyses. Women were mostly Caucasian (86%), 88 (51%) had normal weight (BMI: 18.5–24.9), and 86 (49%) had overweight or obesity (BMI: 25–35; Table 1). By study design, there were significant differences in BMI (∼6.7 points, P < 0.001) and fat mass (∼14%, P < 0.001) between NW and OW women. Gestational weight gain was significantly lower in OW women compared with NW women (−1.5 kg, P = 0.013); however, there was a significant association between maternal group and gestational weight gain outside the Institute of Medicine guidelines (χ² = 36.6, P < 0.001). The incidence of cesarean section delivery was higher in OW (41%) compared with NW (25%) women. There were no differences in maternal race, age, gestational age, sex distribution, and birth weight or birth length between the 2 groups. All children were full term at birth. Of all the participants who initiated lactation, 84 (54%) were still breastfeeding at 9 mo postpartum. The prevalence of women breastfeeding over time was similar in NW and OW women. The women who were no longer breastfeeding at 9 mo had a higher BMI (P = 0.027) and their infants had higher WFA (P = 0.007) and WFL z scores (P < 0.001; Supplemental Table 1).

Human milk composition and intake

Human milk macronutrient concentrations were not significantly different between NW and OW women at 0.5, 1, 2, 3, or 9 mo of life after adjusting for postpartum age and sex (Figure 2, models summarized in Supplemental Tables 2–7). Fat concentration was higher in human milk from OW women at 6 mo (P = 0.002), whereas protein concentration was higher at 5 and 6 mo (P < 0.03) postpartum compared with that of NW women. Human milk carbohydrate concentration was lower in OW women at 4 and 6 mo postpartum than in NW women (P < 0.04). Because human milk volume intake can be regulated by the infant (31), it was hypothesized that infants may self-regulate their intake to counter differences in human milk composition seen between NW and OW women. Thus, all human milk composition data were normalized by the infant's daily intake of human milk volume and formula volume (when applicable) to yield infant daily intakes of macronutrients and other metabolites. Infants of OW women had significantly higher daily intakes of fat and protein at age 1 and 6 mo after adjusting for postpartum age, infant feeding mode, and sex (Figure 2, P < 0.02) compared with infants of NW women. Human milk energy content was higher in human milk from OW women at 6 mo, and infants of OW women had higher daily energy intake at 1 and 6 mo (Figure 2, P < 0.02). Human milk leptin concentration was significantly higher in OW women compared with NW women at all visits after adjusting for postpartum age and sex (Figure 3, P < 0.001; models summarized in Supplemental Table 2). Human milk insulin concentration was significantly higher in OW women compared with NW women at 1, 2, 3, 4, 6, and 9 mo (P < 0.04). Human milk CRP concentration was significantly higher in OW women compared with NW women at 1, 2, 3, 4, and 9 mo (P < 0.02). Infant daily intake of leptin was significantly higher from ages 0.5 to 6 mo (P < 0.01), whereas infant daily intake of insulin was higher at ages 1, 2, 3, 5, 6, and 9 mo in infants of OW women than in infant from NW women (P < 0.04). Infant daily intake of CRP was significantly higher at ages 1, 3, and 5 mo in infants of OW women than in infants of NW women (P < 0.04). There were no significant differences between NW and OW women in IL-6, IL-8, and TNF-α concentrations between ages 0.5 and 9 mo after adjusting for postpartum age and sex (Supplemental Figure 1). Infant sex was not associated with human milk composition or nutrient daily intake in this cohort (data not shown). Considering maternal BMI at 1 or 12 mo postpartum or delivery mode did not modify the overall findings and conclusions drawn from calculating BMI at early pregnancy (Supplemental Figures 2 and 3).

When normalized by body weight, similar results were obtained for macronutrient concentration with higher intake of fat at 1 and 6 mo (Supplemental Figure 4, P < 0.004; models summarized in Supplemental Table 2). Daily intake of leptin remained higher until age 5 mo (Supplemental Figure 4, P < 0.02), whereas daily intakes of insulin and CRP were only significantly higher in infants of OW women compared with NW women at 1 (P = 0.004) and 2 mo (P = 0.046) and at 1 mo (P = 0.003), respectively (Supplemental Figure 4).

Associations with infant growth

Daily intakes of select milk factors for infants of NW and OW women were associated with their growth parameters during the first 9 mo of life (Tables 2 and 3). Results associating daily intakes to z scores were disparate (Table 2). For instance, daily intake of carbohydrates was negatively associated with LFA z scores (β = −0.009, P = 0.002), whereas infant's daily intakes of protein and leptin were positively associated with LFA z scores (β = 0.05, P = 0.012 and β = 0.24, P = 0.047, respectively). Daily intakes of fat and carbohydrates were negatively associated with WFA z scores (β = −0.005, P = 0.040 and β = −0.006, P = 0.026, respectively), whereas infant's daily intake of protein was positively associated with WFA z scores (β = 0.05, P = 0.013). Finally, only daily intake of carbohydrates was positively associated with WFL z scores (β = 0.01, P = 0.003). Results associating daily intakes during the first 9 mo of life with FMI are presented in Table 3. Infant daily intakes of carbohydrates and insulin positively associated with FMI (β = 0.06, P < 0.001 and β = 0.37, P = 0.028, respectively), whereas infant daily intakes of protein and leptin negatively associated with FMI (β = −0.45, P < 0.001 and β = −1.15, P < 0.001, respectively). Infant daily intakes of carbohydrates positively associated with FFMI (β = 0.02, P < 0.001), whereas daily intakes of protein negatively associated with FFMI (β = −0.08, P < 0.001). To investigate potential differences in growth between NW and OW groups, as suggested by Young et al. (32), exploratory analyses of human milk composition on growth outcomes were performed on both groups separately. The effects of carbohydrates and protein on LFA and WFA were driven by the OW group (no effect was found in the NW group; Supplemental Table 8). Similarly, the association between daily intakes of insulin and FMI was driven by the OW group (Supplemental Table 9). Of interest, daily intake of CRP was associated with FMI in the OW group and not in the NW group, whereas the opposite was observed for the association of daily intake of CRP and FFMI (Supplemental Table 9).

Discussion

In this cohort, human milk composition from OW mothers was higher in fat and protein and lower in carbohydrate content at only a few time points throughout the first 9 mo of life compared with that from NW mothers. Human milk leptin content was higher in OW mothers throughout lactation and insulin and CRP content were higher at most time points than in NW mothers. Similar results were observed when the concentrations were normalized to daily intake of human milk and infant body weight. The differences in infant's exposure to macronutrients and other bioactives were associated with changes in infant growth parameters through 9 mo postpartum, although the effect sizes were small for macronutrients (0.005–0.05 z-score units and 0.02–0.45 FMI or FFMI units per unit of change in composition). A larger effect was seen with human milk insulin and leptin, where infants of OW mothers were exposed to 1.5–2.5 times higher concentrations compared with infants of NW mothers.

Maternal BMI and human milk macronutrients

Evidence of an effect of maternal adiposity on human milk macronutrients has been disparate. The findings from Nommsen et al. (33) demonstrating positive correlations between maternal percentage body fat and milk lipid concentration and energy content have been followed by a mix of contradictory findings in a range of cohorts from different countries (11, 12, 14, 34, 35). Higher human milk fat content has been reported in OW women compared with NW women (12, 33), even with preterm deliveries (36). To date, results on human milk protein content are inconsistent (11, 14). These disparate findings are likely due to differences in study designs, including nonstandardized human milk collection methods, different stages of lactation, or follow-up periods examined. However, it has been established that maternal obesity does not impact human milk carbohydrate content (17, 37). The presented results offer a unique view of the macronutrient changes during the first 9 mo of lactation in a large cohort of NW and OW mothers. Although absolute results show the trend identified by others, the significant differences were few, mostly in mature milk. Of interest, the higher human milk fat content was not associated with increased WFL z scores or FMI during the first 9 mo of life, suggesting that exposure to higher human milk fat content may not lead to growth differences previously reported in another cohort (38). Although this other cohort was larger, the human milk collection was from only 1 time period (4–8 wk postpartum), only assessed infant growth at 3 time points (birth and 3 and 12 mo), and did not consider differences in maternal BMI (38).

Maternal BMI and human milk nonnutritive factors

In accordance with previous reports (14, 19, 22), human milk IL-6, IL-8, and TNF-α concentrations were not different between NW and OW mothers. Differences in leptin and insulin content between human milk from OW women and that from NW women have also been noted in several other studies (18, 22, 35, 39, 40). To our knowledge, only 1 study reported an association between maternal BMI and human milk CRP concentrations (22), which is consistent with our findings.

Human milk nonnutritive factors and offspring growth

Negative associations between insulin and leptin human milk content and WLZ at 4 and 12 mo (19, 39), body weight at 4 and 6 mo (18, 41), lean body mass at 4 mo (41), and fat mass at 6 mo (18) have been previously reported. Interestingly, human milk insulin concentrations have been negatively associated with WFL trajectories of infants born to NW mothers but not OW mothers. These results are consistent with the negative association observed between infant's exposure to human milk leptin and FMI in this cohort. In our hands, infant's exposure to human milk insulin was not associated with LFA, WFA, or WFL z scores or FFMI during the first 9 mo of life. In fact, daily intake of insulin was positively associated with FMI during the first 9 mo of life, and daily intake of leptin was positively associated with LFA z scores.

Maternal adiposity drives the main effects

As suggested previously by Young et al. (32), it appears that the growth and body composition associations with human milk composition are dependent on maternal BMI. Indeed, all of the effects that were identified in this cohort were driven by the OW group and not present in the NW group. Although previous research demonstrated an effect of maternal OW on human milk composition, many of these studies were cross-sectional in design, did not evaluate associations with infant body composition outcomes, or failed to account for infant milk intake, thereby making it difficult to fully understand the interaction between human milk composition and its long-term impact on the infant's growth. Only 2 studies have investigated daily macronutrient and nonnutritive bioactive intakes rather than human milk concentrations—Heinig et al. (42) (daily protein intake in breastfed and formula-fed infants) and Grote et al. (43) (macronutrient intake at 3, 6, and 12 mo)—but neither combined these investigations with maternal BMI or adiposity.

Limitations and strengths

Although we strived to use the best methodologies for this study, there are still limitations. First, the daily intakes were estimated using weight prior to and after nursing and a 3-d weighed food record, which is not as accurate as the use of doubly labeled water to estimate energy intake. Furthermore, the samples were not collected over a 24-h period, and mothers were not fasted, which may have provided more variability for insulin results. Another limitation of this study is the absence of maternal dietary intake data during the lactation period. A larger and more diverse cohort may have also helped dissect interindividual differences and variability observed.

The results presented clearly demonstrate that infants of OW mothers are exposed to higher concentrations of insulin, leptin, and, to a lesser extent, CRP. Nevertheless, the bioavailability of these 3 human milk bioactives is still unclear, and their mechanisms of action in the infant are yet to be elucidated (18, 44, 45). Leptin and insulin receptors are expressed in the infant gut, suggesting a potential role of these hormones locally (46, 47). However, it is still unclear how human milk insulin and leptin survive trypsin activity, which has been shown to rapidly digest insulin in adults (48). Human milk has protease inhibitors (α₁-antitrypsin and α₁-antichymotrypsin) that may protect insulin from degradation prior to reaching the duodenum, and it is also possible that milk fat globules (49, 50) are involved in transporting and protecting these bioactives up to the intestinal tract, where they can be biologically active (51).

This study is strengthened by the longitudinal human milk composition data reported on a large cohort. Furthermore, the adjustment for human milk daily intake and body weight provides a more comprehensive analysis of the infant's exposure. The sample collection was standardized for the time of day and consisted of the full expression of 1 breast, which procured a quality sample at each study visit. Finally, the longitudinal modeling of the associations between daily macronutrient and bioactive intakes and growth parameters is a unique strength among the existing literature. Other large longitudinal studies using state-of-the-art methodologies are critically needed to assess the role of nonnutritive bioactives on infant growth and body composition. The combined analyses of human milk micronutrients, fatty acids (52), oligosaccharides, untargeted metabolomics (53), and microbiome (35) outcomes would increase the understanding of the complex matrix of human milk composition and its effect on infant growth and development.

Conclusions

This study is one of the first to examine the impact of maternal BMI on human milk composition throughout the first 9 mo of lactation while considering infant human milk daily volume intakes and relating these intakes to infant growth and body composition. By combining these elements, the study allows for a more comprehensive understanding of maternal BMI effects on human milk composition and its association with infant growth. The concentrations of leptin and insulin in human milk from OW mothers were higher than those in human milk from NW mothers, resulting in higher daily intakes by their infants and modulation of FMI from 0.5 to 9 mo postpartum. This study demonstrates the value of considering both human milk composition and infant daily intake when assessing infant growth.

Notes

This work was supported by USDA ARS 6026-51000-012-06S and NIH grant R01 DK107516.

Data described in the manuscript will be made available upon request.

Supplemental Tables 1–9 and Supplemental Figures 1–4 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/ajcn/.

Abbreviations used: CRP, C-reactive protein; FFMI, fat-free mass index; FMI, fat mass index; LFA, length-for-age; NW, normal weight; OW, overweight/obese; WFA, weight-for-age; WFL, weight-for-length.