Gestational Diabetes Mellitus,Human Milk Composition,and Infant Growth

Abstract

Background:

Gestational diabetes mellitus (GDM) is known to affect human milk composition. Aims of this study were to compare macronutrient and energy content of human milk of women with (GDM⁺) and without GDM (GDM⁻), to assess the association between maternal health and human milk macronutrient and energy content and association between human milk macronutrient and energy content and infant growth.

Study Design and Methods:

Two months after delivery, hindmilk samples were collected. Triglyceride (TG), lactose, and protein content of human milk were measured. An oral glucose tolerance test was performed. Infant weight and length at birth and 2 months were collected. Weight-for-age (WAZ) and weight-for-length z-scores were calculated.

Results:

Twenty-four GDM⁺ and 29 GDM⁻ women were included. Protein, lactose, and energy content of human milk were similar between groups. TG concentration was higher in GDM⁺ than in GDM⁻ women (6.3 ± 2.0 versus 5.3 ± 1.2, p = 0.04). This difference was no longer significant after adjustment for maternal age and infant sex (p = 0.23). Maternal age was associated with TG (r = 0.28, p = 0.04) and lactose (r = −0.30, p = 0.03), while fasting glucose was associated with proteins (r = 0.30, p = 0.03) and tended to be associated with TG (r = 0.27, p = 0.05) and energy (r = 0.24, p = 0.08). TG levels in human milk were associated with weight (β: 0.26, 95% confidence interval [CI]: 0.02 to 0.50) and WAZ (β: 0.40, 95% CI: 0.05 to 0.75) at 2 months among children unexposed (GDM⁻) to GDM, but not among children exposed (GDM⁺)

Conclusions:

In conclusion, GDM status, maternal age, and fasting glucose level were associated with human milk composition. Finally, TG in human milk was associated with infant growth among GDM⁻ children but not among GDM⁺ children. ClinicalTrials.gov NCT02872402.

Background

Gestational diabetes mellitus (GDM), defined as hyperglycemia with first onset or recognition during pregnancy, is a complication of pregnancy that is widely prevalent, affecting one in sixth birth worldwide.^1,2 GDM is associated with later consequences for both the mother and her offspring.^3,4 Among others, mothers and children are at risk of developing type 2 diabetes.^5–7 The fetus exposed in utero exposure to GDM is also associated with an increased risk of obesity.⁸

Human milk has been associated with a reduced risk of obesity compared to formula feeding.⁹ Human milk feeding has been shown to be associated with obesity prevention among children exposed to GDM in utero (GDM⁺), although a longer duration of human milk feeding seems necessary to achieve the protective effect of human milk in this population.¹⁰ Moreover, some studies rather showed that human milk feeding was associated with increased obesity related outcomes among GDM⁺ children, and this could be due to an altered composition of human milk in the context of diabetes.^11–13

A recent literature review showed that diabetes is associated with human milk composition and that different type of diabetes would result in different composition of the human milk.¹⁴ Among GDM⁺ women specifically, human milk is altered in different components such as insulin, ghrelin, or leptin, as well as in macronutrient.¹⁴ However, among the few studies that investigated the association between GDM status and macronutrient content of mature human milk, results are conflicting.^15–17 Given that these components of human milk are known to influence infant growth, studying the association between human milk macronutrient content in the context of GDM and infant growth is relevant to find mechanisms linking the association between breastfeeding and growth in this population at risk of obesity.¹⁸

In addition, the role of maternal postnatal anthropometric and glycemic profiles in the association between GDM status and human milk composition has not been previously investigated. It is of major interest to investigate the role of maternal health on human milk composition at the time of the human milk collection, particularly among GDM⁺ women given that glycemia may affect breast milk composition, in addition to weight retention and/or advanced maternal age, two risk factors of GDM onset.^4,16

Aims of this study were: (1) to compare human milk content in triglycerides (TGs), lactose, protein, and energy of GDM⁺ women with women without GDM (GDM⁻ women), (2) to evaluate the association between maternal anthropometric and glycemic profiles and macronutrient and energy content of human milk, and (3) to evaluate the association between human milk composition in macronutrient and energy and infant's growth from birth to 2 months of age. We hypothesized that human milk composition is different between groups, that maternal health profile is associated with the human milk composition, and that human milk composition is associated with infant's growth, regardless of the GDM status of the mother.

Materials and Methods

Participants

GDM⁺ women participated in an intervention study for 18 months, starting at 2 months postpartum. This study aimed to assess, within a pilot randomized controlled trial, the feasibility and acceptability of a lifestyle intervention initiated early after childbirth and continuing until 18 months postpartum. This intervention was designed to prevent type 2 diabetes development among women with a history of GDM. Women with a pregnancy complicated by GDM and followed by endocrinologists and dieticians at the two main hospitals with a neonatal care unit in Quebec City, Canada, were invited to participate in this clinical trial starting 2 months after delivery (clinical trial NCT02872402). Invitations to participate in this study were also sent by emails to the community of Université Laval. The recruitment was conducted between January 2017 and September 2019. This study included GDM⁺ women, aged ≥18 years, with a body mass index (BMI) ≥18.5 kg/m². Exclusion criteria were multiple pregnancy, preterm delivery (<37 weeks), a history of type 1 or type 2 diabetes mellitus, a history of bariatric surgery, or planning a pregnancy for the next year. GDM was diagnosed according to Diabetes Canada criteria.¹⁹

In addition, women of 2-month-old infants but without a history of GDM were recruited from March to September 2020 using emails from the Université Laval community, as well as posts shared on social medias. These women were recruited as a control group of this study (GDM⁻ women). The same inclusion and exclusion criteria than GDM⁺ women were applied, except that these women did not have any history of GDM. A total of 24 GDM⁺ and 29 GDM⁻ women were included in this study. Written consent was provided by all participants. This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects/patients were approved by the Centre Hospitalier Universitaire de Québec Ethics Committee (2017-3225 and 2020-5075).

Maternal health and sociodemographic characteristics

For the current study, only data collected during the baseline visit (at 2 months postpartum) were used for the GDM⁺ group. GDM⁻ women were submitted to the same baseline visit than GDM⁺ women. During this visit, all participating women completed a self-administered sociodemographic questionnaire and questionnaires on infant feeding practices and antenatal data.

During the 2-month postpartum visit, weight of all women was measured on a calibrated balance to the nearest 0.1 kg, height was measured using a stadiometer, and BMI was calculated (kg/m²). Body composition was measured with a dual energy X-ray absorptiometry scanner (GE Lunar Prodigy Bone Densitometer; GE Healthcare Lunar, Madison, WI) by trained professionals using the Lunar enCORE software version 14.1, and body fat distribution (fat mass percentage and fat mass percentage in android and gynoid regions) was analyzed as previously described.²⁰ Fasting blood sample was collected for all women, and a 75 g 2-hour oral glucose tolerance test was performed. Plasma glucose and insulin at 0 and 2 hours were measured enzymatically and using radioimmunoassay, respectively.^21,22 The homeostasis model assessment for insulin resistance (HOMA-IR) index was calculated as follows: [fasting insulinemia (μU/L) × fasting glycemia (mmol/L)/22.5].²³

Infant growth

Weight and length at birth and at 2 months of age were collected using the child health record. Weight-for-age (WAZ) and weight-for-length (WLZ) sex-specific z-scores were calculated using the growth standard charts from the World Health Organization.²⁴ The difference in growth parameters between 2 months and birth was calculated and presented as delta (Δ) values.

Human milk collection and composition

Women were asked to provide 30–60 mL of hindmilk during the visit or at home at the closest moment from the 2-month visit. Human milk was either pumped or expressed manually at the end of a feeding and was collected in a sterile cup. The timing of the human milk collection was documented by the mother, and samples were categorized based on the timing of collection, that is, morning (6:00 to 11:59 am), afternoon (12:00 to 17:59 pm), evening (18:00 to 23:59 pm), and night (12:00 to 5:59 am). Women that expressed their breast milk at home were asked to store the sample in their freezer before bringing it back to the research center in a transport bag containing an ice pack. All human milk samples were stored in freezers at −20°C for 1 month after collection and at −80°C afterward. Frozen samples were defrosted at 0°C to 4°C, and whole human milk samples were vortexed at high speed for 30 seconds to ensure sample homogeneity immediately before aliquoting in 2 mL samples. Another freezing cycle at −80°C was performed for all samples prior to laboratory analysis.

TG content of human milk was measured by colorimetry (Single Reagent, GPO PAP Method, 2018). Given that TG represents 98% of lipids in human milk, it was used as a surrogate for total lipid content of human milk.²⁵ Similarly, lactose represents the main carbohydrate of human milk and was then used as a surrogate for carbohydrates.^26,27

Lactose was measured by high-performance liquid chromatography with an Agilent 1100 chromatograph (Agilent Technologies, Santa Clara, CA), equipped with an Agilent 1260 refractive index detector. An ICSep ICE-ION-300 column (Transgenomic, Omaha, NE) was used with 0.0065 N of H₂SO₄ as the mobile phase, at a flow rate of 0.4 mL/min and the run time is 45 minutes. The column temperature was kept constant at 40°C. Samples were diluted by two in acetonitrile, and the supernatant was then diluted by fivefold dilution with ultra pure water and filtered (nylon 0.45 μm; Chromatographic Specialties, Brockville, ON, Canada) before injection (15 μL). A mixture of standards: lactose, glucose, and galactose (all from Sigma-Aldrich) was used as an external standard. Total solids were analyzed according to the AOAC international method (990.20, 2000).

Total protein content of dried human milk was determined by the Dumas combustion method using a Rapid Micro N Cube (Elementar, Francfort-sur-le-Main, Germany). Nitrogen content was converted into a protein concentration by applying a nitrogen-to-protein conversion factor of 6.25.²⁸ All samples were analyzed in duplicate, and concentrations were converted in g/100 mL of human milk. Energy content of human milk was calculated using Atwater conversion where 9, 4, and 4 kilocalories (kcal) are associated to every gram of lipid, carbohydrate, and protein, respectively.²⁹

Statistical analysis

Participant's characteristics were compared using Student's t-test, Chi-square test (χ²), or Fisher exact test. Human milk composition between GDM⁺ and GDM⁻ women was compared using Student's t-test and ANOVA adjusted for maternal age and infant sex. Assumptions for statistical analysis were tested, and data transformation according to Box-Cox procedure was computed when needed. Pearson's coefficients of correlation were calculated to assess the association between maternal characteristics, that is, maternal age, glycemic and anthropometric profiles, and human milk composition in the entire cohort and separately among GDM⁺ and GDM⁻ groups. Additional adjustment for timing of milk sample collection was made.

Finally, multivariable regression models were computed to estimate the potential association between macronutrient and energy content of human milk and infant anthropometric profile at birth, 2 months, and between birth and 2 months. Interactions with GDM status were tested in these models to verify whether human milk composition association with growth varies according to in utero exposure to GDM. Models were adjusted for current maternal BMI, gestational age, and infant sex. Models assessing association between lactose, protein, or lipid content of human milk and infant growth were also adjusted for energy content of human milk. Additional adjustment for breastfeeding exclusivity status at the time of human milk collection was also tested. The statistical software SAS version 9.4 was used to compute analyses. Post hoc power calculation using G*power software (version 3.1.9.7, 2020) was computed using current sample size and parameter estimates derived from the comparison of TG levels between groups. A statistical power of 69% was obtained.

Results

Participant's characteristics are presented in Table 1. Briefly, GDM⁺ women were older than the control group, had higher prepregnancy and current BMI, total fat mass percentage and fat mass percentage in gynoid and android regions, as well as higher values of plasma glucose and insulin (Table 1). None of the participating women, either from GDM⁺ or GDM⁻ groups, was diabetic at the time of the visit.

Table 1.

Participant's Characteristics

Characteristics	Total sample, n = 53	GDM⁺, n = 24	GDM⁻, n = 29	p ^a
Demographic and antenatal data
Age (years)	31.6 ± 3.7	33.5 ± 3.6	30.0 ± 3.1	<0.001
Parity (number of children)
1	28 (2.8)	12 (50.0)	16 (55.2)	0.71
2 or more	25 (47.2)	12 (50.0)	13 (44.8)
Gestational age at birth (weeks)	39.1 ± 1.1	38.6 ± 1.0	39.4 ± 1.1	0.01
Infant sex
Girls	23 (44.2)	6 (26.1)	17 (58.6)	0.02
Boys	29 (55.8)	17 (73.9)	12 (41.4)
Education levels
High school or less	3 (5.7)	1 (4.2)	2 (6.9)	0.91
College	9 (17.0)	4 (16.7)	5 (17.2)
University	41 (77.3)	19 (79.2)	22 (75.9)
Family income (CAD$/year)
0–39,000	1 (1.9)	0 (0.0)	1 (3.5)	0.39
40,000–79,000	11 (20.8)	7 (29.2)	4 (13.8)
80,000–99,999	12 (22.6)	4 (16.7)	8 (27.6)
>100,000	29 (54.7)	13 (54.2)	16 (55.2)
Exclusive breastfeeding at 2 months (yes)	46 (86.8)	17 (70.8)	29 (100.0)	0.002
GDM treatment
Diet only	—	10 (41.7)	—	—
Insulin or oral hypoglycemic agent	—	14 (58.3)	—	—
Metabolic
Prepregnancy BMI (kg/m²)	28.6 ± 7.4	30.9 ± 7.7	26.8 ± 6.8	0.05
Current BMI (kg/m²)	29.2 ± 6.6	31.0 ± 7.2	27.7 ± 5.9	0.07
Fat mass (%)	40.2 ± 8.1	42.7 ± 6.8	38.0 ± 8.6	0.03
Fat mass android (%)	42.6 ± 11.3	46.5 ± 8.6	39.2 ± 12.4	0.02
Fat mass gynoid (%)	45.2 ± 7.9	47.5 ± 7.1	43.3 ± 8.2	0.05
Fasting glucose (mmol/mL)	4.9 ± 0.4	5.0 ± 0.4	4.8 ± 0.3	0.05
2-Hour glucose (mmol/L)	5.6 ± 1.3	6.2 ± 1.5	5.1 ± 0.9	0.002
Fasting insulin (pmol/L)	50.3 ± 29.6	49.4 ± 20.5	51.0 ± 35.4	0.70
2-Hour insulin (pmol/L)	253.1 ± 142.4	302.0 ± 171.9	215.1 ± 102.5	0.08
HOMA-IR	1.60 ± 0.98	1.60 ± 0.75	1.60 ± 1.15	0.58
Birth weight (kg)^b	3.35 ± 0.37	3.38 ± 0.34	3.34 ± 0.39	0.29^c
Infant weight at 2 months (kg)^d	5.33 ± 0.64	5.64 ± 0.52	5.13 ± 0.63	0.003^c

Results are mean ± standard deviation or n (%).

Student's t-test, chi square test, or Fisher exact test between GDM⁺ and GDM⁻ groups.

n = 18 for GDM⁺ and 28 for GDM⁻.

ANOVA adjusted for gestational age at birth.

n = 17 for GDM⁺ and 26 for GDM⁻.

BMI, body mass index; GDM⁺, women with gestational diabetes mellitus; GDM⁻, women without gestational diabetes mellitus; HOMA-IR, homeostasis model assessment for insulin resistance.

Human milk content in protein, lactose, and energy was similar between GDM⁺ and GDM⁻ women, while the TG concentration in human milk of GDM⁺ women was higher than in the control group (6.28 ± 2.03 g/100 mL compared to 5.29 ± 1.18 g/100 mL, p = 0.04, respectively, Table 2). This difference was no longer significant after adjustments for maternal age and infant sex (Table 2).

Table 2.

Human Milk Composition in Proteins, Lactose, Triglycerides, and Energy According to Gestational Diabetes Mellitus Status

Components	Total, n = 53	GDM⁺, n = 24	GDM⁻, n = 29	p ^a	p ^b
Proteins (g/100 mL)	1.2 ± 0.3 (0.7–2.1)	1.3 ± 0.3 (0.8–2.1)	1.2 ± 0.2 (0.7–1.8)	0.36	0.75
TG (g/100 mL)	5.7 ± 1.7 (2.2–10.7)	6.3 ± 2.0 (2.2–10.7)	5.3 ± 1.2 (3.3–7.0)	0.04	0.23
Lactose (g/100 mL)	11.9 ± 1.7 (8.8–16.1)	11.5 ± 1.5 (8.8–14.8)	12.2 ± 1.9 (9.7–16.1)	0.15	0.90
Energy (kcal/100 mL)	104 ± 15 (76–143)	108 ± 18 (76–143)	101 ± 13 (77–129)	0.13	0.27

Results are expressed as crude mean ± standard deviation with (minimum–maximum values).

Comparison between groups using a Student's t-test.

Comparison between groups using ANOVA adjusted for maternal age and infant sex.

GDM⁺, women with gestational diabetes mellitus; GDM⁻, women without gestational diabetes mellitus; TG, triglyceride.

As shown in Table 3, maternal fasting glucose was positively associated with concentrations of proteins, TG, and energy, while fasting insulin was associated with energy content of human milk in the entire cohort. In contrast, when analyses were conducted separately, fasting insulin levels and HOMA-IR were positively associated with protein, TG, and energy content of human milk among GDM⁺ women only. No association was observed between maternal 2-hour glucose or 2-hour insulin levels and any component of the human milk, neither in the entire cohort nor among GDM⁺ or GDM⁻ women (p > 0.10, results not shown). Finally, maternal age was negatively associated with lactose concentrations in human milk in the entire cohort only and positively associated with TG levels in human milk of all women and among GDM⁻ women. Additional adjustment for the timing of human milk collection and infant sex did not change results observed (results not shown).

Table 3.

Association Between Maternal Glycemic and Anthropometric Profiles and Human Milk Composition in Proteins, Lactose, Triglycerides, and Energy According to Gestational Diabetes Mellitus Status

Maternal characteristics	Entire sample, n = 53		GDM⁺, n = 24		GDM⁻, n = 29
Maternal characteristics	r	p	r	p	r	p
Proteins
Age	0.20	0.15	0.16	0.46	0.17	0.38
Current BMI	0.13	0.34	0.16	0.45	0.04	0.84
Fat mass (%)	0.16	0.25	0.16	0.44	0.11	0.58
Android fat mass (%)	0.11	0.44	0.14	0.51	0.03	0.90
Gynoid fat mass (%)	0.18	0.19	0.17	0.44	0.15	0.44
Fasting glucose	0.30	0.03	0.33	0.12	0.22	0.25
Fasting insulin	0.20	0.16	0.50	0.02	−0.02	0.94
HOMA-IR	0.15	0.31	0.54	0.01	−0.11	0.59
Lactose
Age	−0.30	0.03	−0.17	0.42	−0.30	0.11
Current BMI	−0.04	0.78	−0.05	0.83	0.05	0.79
Fat mass (%)	−0.03	0.85	−0.04	0.85	0.08	0.67
Android fat mass (%)	−0.04	0.79	−0.0006	0.98	0.05	0.78
Gynoid fat mass (%)	−0.03	0.85	−0.06	0.78	0.09	0.66
Fasting glucose	−0.10	0.48	−0.31	0.14	0.19	0.33
Fasting insulin	0.16	0.26	0.25	0.26	0.14	0.47
HOMA-IR	0.06	0.69	0.08	0.72	0.05	0.79
TG
Age	0.28	0.04	0.03	0.90	0.40	0.03
Current BMI	0.18	0.19	0.14	0.51	0.09	0.65
Fat mass (%)	0.14	0.32	0.05	0.82	0.07	0.72
Android fat mass (%)	0.18	0.21	0.06	0.78	0.13	0.50
Gynoid fat mass (%)	0.07	0.64	−0.04	0.87	0.004	0.98
Fasting glucose	0.27	0.05	0.28	0.19	0.09	0.65
Fasting insulin	0.23	0.11	0.37	0.09	0.12	0.53
HOMA-IR	0.17	0.23	0.40	0.07	0.03	0.86
Energy
Age	0.15	0.27	−0.02	0.93	0.17	0.37
Current BMI	0.17	0.23	0.14	0.52	0.11	0.59
Fat mass (%)	0.14	0.34	0.05	0.82	0.12	0.56
Android fat mass (%)	0.16	0.25	0.07	0.75	0.14	0.47
Gynoid fat mass (%)	0.07	0.64	−0.05	0.83	0.07	0.74
Fasting glucose	0.24	0.08	0.20	0.35	0.20	0.30
Fasting insulin	0.31	0.03	0.50	0.02	0.18	0.34
HOMA-IR	0.20	0.15	0.47	0.03	0.05	0.80

Results are Pearson coefficient of correlation (r).

BMI, body mass index; GDM⁺, women with gestational diabetes mellitus; GDM⁻, women without gestational diabetes mellitus; HOMA-IR, homeostasis model assessment for insulin resistance; TG, triglyceride.

Finally, associations between human milk composition and infant growth are presented in Table 4. In the entire cohort, every increase of 1 g of lactose per 100 mL of human milk was associated with an increase in 0.39 unit of change in WLZ between birth and 2 months of age. Additional adjustment for exclusivity of breastfeeding at the time of human milk collection did not influence these results (results not shown).

Table 4.

Association Between Human Milk Composition in Proteins, Lactose, Triglycerides, and Energy and Infant Growth

Growth parameters	Proteins^a			Lactose^a			TG^a			Energy
Growth parameters	β	95% CI	p-interaction with GDM status	β	95% CI	p-interaction with GDM status	β	95% CI	p-interaction with GDM status	β	95% CI	p-interaction with GDM status
Weight at 2 months^b	−0.37	−1.09 to 0.35	0.02	−0.04	−0.14 to 0.06	0.14	0.10	−0.12 to 0.32	0.01	0.007	−0.004 to 0.02	0.16
Δ Weight 0–2 months^b	−0.36	−1.00 to 0.29	0.12	0.03	−0.06 to 0.12	0.25	−0.06	−0.25 to 0.14	0.07	0.003	−0.006 to 0.01	0.16
WAZ at 2 months^b	−0.55	−1.62 to 0.53	0.02	−0.06	−0.21 to 0.91	0.14	0.15	−0.18 to 0.47	0.007	0.01	−0.005 to 0.03	0.20
Δ WAZ 0–2 months^b	−0.53	−1.57 to 0.52	0.31	0.09	−0.05 to 0.24	0.35	−0.18	−0.49 to 0.14	0.15	0.002	−0.01 to 0.02	0.23
WLZ at 2 months^c	−0.29	−1.67 to 1.08	0.004	0.12	−0.05 to 0.30	0.67	−0.26	−0.65 to 0.12	0.009	0.02	0.0003 to 0.04	0.001
Δ WLZ 0–2 months^d	0.03	−2.14 to 2.20	0.37	0.39	0.14 to 0.63	0.71	−0.85	−1.40 to −0.29	0.90	0.007	−0.02 to 0.04	0.34

Results are β estimates from regression analyses with CI of 95% conducted in the entire cohort. Adjustments for gestational age, infant sex, and maternal BMI were performed.

Additional adjustment for energy content of human milk. Interaction for GDM status was tested.

n = 17 for GDM⁺ and n = 26 for GDM⁻.

n = 17 for GDM⁺ and n = 22 for GDM⁻.

n = 16 for GDM⁺ and n = 21 for GDM⁺.

CI, confidence interval; GDM, gestational diabetes mellitus; TG, triglyceride; WAZ, weight-for-age; WLZ, weight-for-length.

Some human milk components were associated with growth parameters at 2 months differently according to children in utero exposure to GDM, as shown by a test for interaction between human milk composition and GDM status that revealed some incompatibility with the null hypothesis of no interaction (p < 0.05, Table 4). Indeed, proteins and TG levels in human milk were associated with weight, WAZ, and WLZ at 2 months differently among GDM⁺ and children unexposed to GDM (GDM⁻) (Table 5). However, only TG in human milk was associated positively with weight and WAZ at 2 months among GDM⁻ children, while all other associations were not statistically significant according to 95% confidence interval (CI) (Table 5). Finally, energy level in human milk was associated differently with WLZ at 2 months among GDM⁺ and GDM⁻ children (p for interaction of 0.001, Table 4). Energy in human milk was positively associated with WLZ among GDM⁻ children and negatively associated among GDM⁺ children (β: 0.05, 95% CI: 0.02 to 0.08 and −0.007, 95% CI: 0.12 to 0.05, respectively).

Table 5.

Association Between Human Milk Composition in Proteins, Lactose, and Triglycerides and Infant Growth According to In Utero Exposure to Gestational Diabetes Mellitus

Growth parameters	Proteins				TG
	GDM⁺		GDM⁻		GDM⁺		GDM⁻
	β	95% CI	β	95% CI	β	95% CI	β	95% CI
Weight at 2 months^a	−1.20	−3.23 to 0.83	0.26	−0.56 to 1.08	0.002	−0.43 to 0.44	0.26	0.02 to 0.50
WAZ at 2 months^a	−1.79	−4.82 to 1.23	0.40	−0.82 to 1.62	−0.003	−0.65 to 0.64	0.40	0.05 to 0.75
WLZ at 2 months^b	−1.85	−5.53 to 1.83	1.20	−0.39 to 2.80	−0.38	−1.17 to 0.41	0.09	−0.35 to 0.53

Results are β estimates from regression analyses with CI of 95%. Adjustments for gestational age, infant sex, maternal BMI, and energy content of human milk were performed.

n = 17 for GDM⁺ and n = 26 for GDM⁻.

n = 17 for GDM⁺ and n = 22 for GDM⁻.

CI, confidence interval; GDM⁺, children exposed to gestational diabetes mellitus in utero. GDM⁻, children unexposed to gestational diabetes mellitus in utero; TG, triglyceride; WAZ, weight-for-age; WLZ, weight-for-length.

Discussion

Results of this study showed that GDM⁺ women had higher levels of TG in their mature human milk compared to the control group. However, this association may be partly explained by their more advanced age. Other factors such as fasting glucose and insulin levels were also associated with human milk composition, highlighting the importance to monitor glycemia during the postpartum period. Finally, TG levels in human milk were positively associated with infant's weight and WAZ at 2 months in GDM⁻ but not in GDM⁺ children.

Human milk concentrations in lactose, proteins, and energy were similar between GDM⁺ and GDM⁻ women, while concentration in TG was higher among GDM⁺ women. However, this difference was no longer significant after adjustment for maternal age and infant sex. The study conducted by Dritsakou et al showed no difference in lipid concentration between human milk of GDM⁺ and GDM⁻ women.¹⁶ Nevertheless, they found that human milk of GDM⁺ women contained higher energy levels, and they hypothesized that this could be due to a higher lipid content in human milk of GDM⁺ women that was not statistically perceived in their results.¹⁶ In contrast, results of the present study are not in accordance with the study conducted by Shapira et al that instead observed lower lipid concentration in human milk of GDM⁺ compared to GDM⁻ women.¹⁵ However, this latter study collected human milk samples only 2 weeks after delivery, compared to 2 months in the current study, which could explain difference in results.¹⁵ Furthermore, adjustments for important confounding factors, such as maternal age, were not performed in these previous two studies.^15,16

In the present study, maternal age was associated with lactose and TG content of human milk in all women. This is not in accordance with a study conducted by Lubetzky et al that did not observe differences in mature human milk fat content between ≥35 years women and younger women and found greater content of lactose among older women.³⁰

In contrast, results of the present study are in accordance with results from Dritsakou et al,¹⁶ and with other results from transitory milk or colostrum samples.^30–32 Even if results between studies are hardly comparable given the different methods used to analyze human milk sample and characteristics of the studied population, it is possible to hypothesize that results observed in the present study may be explained by the fact that older women would produce less amount of milk, given that a lower density of mammary glands and hormonal changes occur with age, resulting in a different human milk composition.^32,33 However, it has also been proposed that human milk lipid production is greater among older women compared to younger women.³⁰ Further investigations should be conducted to better understand why the association between maternal age and lipid content of human milk was only seen among GDM⁻ women.

This study also showed that fasting glucose levels were associated with protein, lipid, and energy content of human milk in all women, while fasting insulin was positively associated with energy. Furthermore, among GDM⁺ women only, fasting insulin and HOMA-IR were associated with protein, TG, and energy content of human milk. To our knowledge, no other study has evaluated the association between the glycemic and insulinemic profile of GDM⁺ women at the time of human milk collection and the nutritional composition. However, other studies conducted among women with type 1 diabetes suggested that the underlying hypothesis behind differences in human milk composition of women with and without diabetes is partly based on glucose dysregulation associated with this pathology. Accordingly, results conducted among type 1 diabetic lactating women showed that when diabetes is tightly controlled, macronutrient content in human milk is not different than in the general population.³⁴

This study also showed that human milk composition in TG and proteins were associated differently with weight, WAZ, and WLZ at 2 months according to in utero exposure to GDM, and these results were independent of maternal BMI and energy content of human milk. Our results may prompt that the presence of certain macronutrient in human milk would be associated differently with infant's growth depending on GDM status of the mother, particularly the TG concentration of human milk, which is in turn associated with the mother's higher insulin.

In contrast, the positive association between the TG content of human milk and weight parameters among GDM⁻ children is in contradiction with results from Prentice et al showing a negative association between TG in human milk and infant growth.³⁵ However, this negative association was observed between 3 and 12 months, while there was no association observed at 3 months of age.³⁵ Furthermore, the current study did not evaluate the lipid composition of human milk, that is, the proportion of different lipid components, which may also affect infant growth and could possibly explain the differences observed with results from other studies.¹⁸

To our knowledge, no study investigated the association between macronutrient content in human milk and infant growth among GDM⁺ children specifically. Nonetheless, it is possible to speculate that the association between human milk feeding and infant growth among GDM⁺ children may be of lower magnitude than among GDM⁻ children, given their predisposition to obesity.¹⁰ It is also possible that the absence of association observed among GDM⁺ children is a result of a lack of statistical power. It may also explain why no significant association between protein content of human milk and weight parameters in GDM⁺ and GDM⁻ children was observed when analyses were stratified according to GDM status. In contrast, opposite results observed in both groups, that is, proteins associated negatively with weight, WAZ, and WLZ at 2 months among GDM⁺ children and associated positively among GDM⁻ children are intriguing and need to be further investigated.

Finally, the positive association observed between lactose and infant growth in the entire sample is in accordance with other studies conducted in the general population.^35,36 Prentice et al hypothesized that greater intakes of carbohydrates would result in greater storage of glycogen and fat, explaining the positive association with infant weight measurements.³⁵

This study presented several strengths. First, in addition to assessing the association between GDM status and human milk composition in macronutrient, the link between human milk composition and infant growth has been investigated, answering a gap of the current literature. Second, anthropometric and glycemic profiles of women were assessed at the same moment that human milk samples were collected, allowing the investigation of the association between maternal health and human milk composition in the context of GDM. Finally, human milk samples were collected and analyzed using a standardized protocol to ensure comparability of samples between groups.

Limits of this study included the fact that human milk samples were collected only once. Indeed, pooled milk samples collected during 24 hours would have been more representative of the real composition of human milk ingested by the breastfed infant. Moreover, this study did not document the quantity of human milk consumed by the infant. Furthermore, multiple cycles of freezing may have affected stability of human milk samples in this study although as mentioned above, the same method was used for all samples to ensure that this did not affect comparisons between groups. However, these results should not be compared to other studies conducted under different protocols.

In addition, information on history of acute mastitis or plastic surgery of the breast, which may have influenced the human milk composition, was not included in the exclusion criteria and was not available. Although history of chronic disease (other than diabetes) was not an exclusion criterion, none of the participating women had any history of cardiovascular disease, dyslipidemia, or cancer. Finally, results of this study could not be generalized to other populations than Caucasian women with a high degree of education. In addition, the mean BMI of participants in both groups was relatively high, and GDM⁺ women had normal glycemic profile with none of them having type 2 diabetes at the time of the visit.

Conclusions

To conclude, this study suggested that some risk factors for GDM, such as advanced maternal age, may explain the association between GDM status and higher levels of TG in human milk. Other factors such as glycemia and insulinemia were also associated with human milk composition, highlighting the importance to control glycemia during the postpartum period among GDM⁺ women. This study also suggested that lipids in human milk were associated with infant weight parameters at 2 months of age in a different manner between GDM⁺ and GDM⁻ children. While further studies are needed in a larger sample of women, this study suggests that maternal characteristics such as GDM status, advanced age, or fasting glycemia and insulinemia may affect the composition of human milk and, therefore, affect infant growth. This strengthened the need to better understand the impact of maternal health, including GDM status, on human milk composition, as well as on the future health of infants.

Footnotes

Acknowledgments

The authors thank Pierre Gagnon (NUTRISS, INAF, Université Laval) for its assistance with statistical analyses.

Authors' Contributions

A.D., A.V., V.D.M., C.D., and J.R. participated in the conception and the design of the study. C.D. and J.P. made a substantial contribution to data acquisition. V.P., V.R., N.L., and G.S.-A. conducted laboratory analyses. C.D., L.L., and J.R. participated in data analysis and interpretation. C.D. wrote the first draft of the article. All authors revised it critically for important intellectual content and approved the final version. J.R. is responsible of the integrity of the study.

Disclosure Statement

No competing financial interests exist.

Funding Information

This project was funded by Diabète Québec, the Canadian Foundation for Dietetic Research (CFDR) and the Institute of Nutrition and Functional Foods (INAF). Camille Dugas received a scholarship from the Fonds de recherche du Québec–Santé (FRQS).

References

International diabetes federation. Gestational diabetes. 2019. Available from: https://www.idf.org/our-activities/care-prevention/gdm [Last accessed: April 9, 2022].

Coustan

. Gestational diabetes mellitus. Clin Chem, 2013; 59(9):1310–1321; doi: 10.1373/clinchem.2013.203331

Tranidou

, Dagklis

, Tsakiridis

, et al. Risk of developing metabolic syndrome after gestational diabetes mellitus—A systematic review and meta-analysis. J Endocrinol Invest, 2021; 44(6):1139–1149; doi: 10.1007/s40618-020-01464-6

Baz

, Riveline

, Gautier

. Endocrinology of pregnancy: Gestational diabetes mellitus: Definition, aetiological and clinical aspects. Eur J Endocrinol, 2016; 174(2):R43–R51; doi: 10.1530/eje-15-0378

Bellamy

, Casas

, Hingorani

, et al. Type 2 diabetes mellitus after gestational diabetes: A systematic review and meta-analysis. Lancet, 2009; 373(9677):1773–1779; doi: 10.1016/s0140-6736(09)60731-5

Bush

, Chandler-Laney

, Rouse

, et al. Higher maternal gestational glucose concentration is associated with lower offspring insulin sensitivity and altered beta-cell function. J Clin Endocrinol Metab, 2011; 96(5):E803–E809; doi: 10.1210/jc.2010-2902

Clausen

, Mathiesen

, Hansen

, et al. High prevalence of type 2 diabetes and pre-diabetes in adult offspring of women with gestational diabetes mellitus or type 1 diabetes: The role of intrauterine hyperglycemia. Diabetes Care, 2008; 31(2):340–346; doi: 10.2337/dc07-1596

Chandler-Laney

, Bush

, Granger

, et al. Overweight status and intrauterine exposure to gestational diabetes are associated with children's metabolic health. Pediatr Obes, 2012; 7(1):44–52; doi: 10.1111/j.2047-6310.2011.00009.x

Yan

, Liu

, Zhu

, et al. The association between breastfeeding and childhood obesity: A meta-analysis. BMC Public Health, 2014; 14:1267; doi: 10.1186/1471-2458-14-1267

10.

Dugas

, Perron

, Kearney

, et al. Postnatal prevention of childhood obesity in offspring prenatally exposed to gestational diabetes mellitus: Where are we now. Obes Facts, 2017; 10(4):396–406; doi: 10.1159/000477407

11.

Plagemann

, Harder

, Franke

, et al. Long-term impact of neonatal breast-feeding on body weight and glucose tolerance in children of diabetic mothers. Diabetes Care, 2002; 25(1):16–22; doi: 10.2337/diacare.25.1.16

12.

Aris

, Soh

, Tint

, et al. Associations of infant milk feed type on early postnatal growth of offspring exposed and unexposed to gestational diabetes in utero. Eur J Nutr, 2017; 56(1):55–64.

13.

Rodekamp

, Harder

, Kohlhoff

, et al. Long-term impact of breast-feeding on body weight and glucose tolerance in children of diabetic mothers: Role of the late neonatal period and early infancy. Diabetes Care, 2005; 28(6):1457–1462.

14.

Peila

, Gazzolo

, Bertino

, et al. Influence of diabetes during pregnancy on human milk composition. Nutrients, 2020; 12(1):185; doi: 10.3390/nu12010185

15.

Shapira

, Mandel

, Mimouni

, et al. The effect of gestational diabetes mellitus on human milk macronutrients content. J Perinatol, 2019; 39(6):820–823; doi: 10.1038/s41372-019-0362-5

16.

Dritsakou

, Liosis

, Valsami

, et al. The impact of maternal- and neonatal-associated factors on human milk's macronutrients and energy. J Matern Fetal Neonatal Med, 2017; 30(11):1302–1308; doi: 10.1080/14767058.2016.1212329

17.

Amaral

, Silva

, Soares

, et al. What are the maternal factors that potentially intervenes in the nutritional composition of human milk?. Nutrients, 2021; 13(5):1587; doi: 10.3390/nu13051587

18.

Eriksen

, Christensen

, Lind

, et al. Human milk composition and infant growth. Curr Opin Clin Nutr Metab Care, 2018; 21(3):200–206; doi: 10.1097/mco.0000000000000466

19.

Canadian Diabetes Association Clinical Practice Guidelines Expert Committee. Canadian Diabetes Association 2013 Clinical Practice Guidelines for the Prevention and Management of Diabetes in Canada. Can J Diabetes, 2013; 37(Suppl 1):S1–S212.

20.

Kearney

, Perron

, Marc

, et al. Association of prenatal exposure to gestational diabetes with offspring body composition and regional body fat distribution. Clin Obes, 2018; 8(2):81–87; doi: 10.1111/cob.12237

21.

Richterich

, Dauwalder

. Determination of plasma glucose by hexokinase-glucose-6-phosphate dehydrogenase method [in German]. Schweiz Med Wochenschr, 1971; 101(17):615–618.

22.

Desbuquois

, Aurbach

. Use of polyethylene glycol to separate free and antibody-bound peptide hormones in radioimmunoassays. J Clin Endocrinol Metab, 1971; 33(5):732–738; doi: 10.1210/jcem-33-5-732

23.

Matthews

, Hosker

, Rudenski

, et al. Homeostasis model assessment: Insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia, 1985; 28(7):412–419; doi: 10.1007/bf00280883

24.

World Health Organization. WHO Child Growth Standards based on length/height, weight and age. Acta Paediatr Suppl, 2006; 450:76–85; doi: 10.1111/j.1651–2227.2006.tb02378.x.

25.

Koletzko

. Human Milk lipids. Ann Nutr Metab, 2016; 69(Suppl 2):28–40; doi: 10.1159/000452819

26.

Ballard

, Morrow

. Human milk composition: Nutrients and bioactive factors. Pediatr Clin North Am, 2013; 60(1):49–74; doi: 10.1016/j.pcl.2012.10.002

27.

Coppa

, Gabrielli

, Pierani

, et al. Changes in carbohydrate composition in human milk over 4months of lactation. Pediatrics, 1993; 91(3):637–641.

28.

Lönnerdal

, Forsum

, Hambraeus

. A longitudinal study of the protein, nitrogen, and lactose contents of human milk from Swedish well-nourished mothers. Am J Clin Nutr, 1976; 29(10):1127–1133; doi: 10.1093/ajcn/29.10.1127

29.

Merrill

, Watt

. Energy Value of Foods, Basis and Derivation. Human Nutrition Research Branch, Agricultural Research Service, U.S. Department of Agriculture: Washington, DC; 1955.

30.

Lubetzky

, Sever

, Mimouni

, et al. Human milk macronutrients content: Effect of advanced maternal age. Breastfeed Med, 2015; 10(9):433–436; doi: 10.1089/bfm.2015.0072

31.

Argov-Argaman

, Mandel

, Lubetzky

, et al. Human milk fatty acids composition is affected by maternal age. J Matern Fetal Neonatal Med, 2017; 30(1):34–37; doi: 10.3109/14767058.2016.1140142

32.

Hausman Kedem

, Mandel

, Domani

, et al. The effect of advanced maternal age upon human milk fat content. Breastfeed Med, 2013; 8(1):116–119; doi: 10.1089/bfm.2012.0035

33.

Stone

, Warren

, Pinney

, et al. Determinants of percentage and area measures of mammographic density. Am J Epidemiol, 2009; 170(12):1571–1578; doi: 10.1093/aje/kwp313

34.

van Beusekom

, Zeegers

, Martini

, et al. Milk of patients with tightly controlled insulin-dependent diabetes mellitus has normal macronutrient and fatty acid composition. Am J Clin Nutr, 1993; 57(6):938–943; doi: 10.1093/ajcn/57.6.938

35.

Prentice

, Ong

, Schoemaker

, et al. Breast milk nutrient content and infancy growth. Acta Paediatr, 2016; 105(6):641–647.

36.

Gridneva

, Rea

, Tie

, et al. Carbohydrates in human milk and body composition of term infants during the first 12 months of lactation. Nutrients, 2019; 11(7):1472; doi: 10.3390/nu11071472