Differences in Human Milk Lactose and Citrate Concentrations Based on Gestational Diabetes Status

Abstract

Background:

Exclusive breastfeeding is the optimal manner of early infant nutrition but women with gestational diabetes mellitus (GDM) often experience challenges with lactation in the early postpartum period. Increases in the colostral metabolites of lactose and citrate have been found to indicate increased milk production.

Materials and Methods:

A follow-up study of 133 postpartum women with and without GDM was conducted to examine differences in specific colostral metabolite levels using enzymatic methods to determine transition to lactogenesis II during the first week postpartum. We used linear mixed models for repeated measures over time to examine the effect of GDM on colostral metabolite levels at baseline and follow-up with fixed effects of GDM status, time, covariates, and interactions between time and GDM, between time and time, and between time, time and GDM into the model allowing quadratic trends over time.

Results:

Over time, lactose and citrate levels increased for all mothers (p < 0.001 and p < 0.001, respectively), although mothers with GDM had consistently lower lactose and citrate levels compared with nondiabetic mothers (p = 0.004 and p = 0.014, respectively). Age, prepregnancy body mass index, mode of birth, and parity did not independently influence colostral concentrations of lactose and citrate.

Conclusions:

Findings suggest that the rate of change overtime in lactose and citrate concentrations differ by GDM status. Further research examining the trajectory of colostral metabolite levels by GDM status is warranted.

Introduction

Gestational diabetes mellitus (GDM) is one of the most common conditions in pregnancy in the United States with an estimated prevalence rate of 8.7% according to population-based data.¹ Among women with type 1 diabetes, type 2 diabetes, and GDM, women with GDM had the higher rates of any breastfeeding in early postpartum,² although when compared with nondiabetic women, women with GDM have similar rates of any breastfeeding in early postpartum.^3,4 However, GDM has been associated with lower exclusive breastfeeding rates,^5,6 which can negatively influence maternal and infant health.⁷ Furthermore, among mothers who had GDM, 33% reported that they experienced a delay in their onset of lactation.⁸ Reduced duration of exclusive and any breastfeeding has been associated with delayed transition to lactogenesis II in the early postpartum period.⁹

Researchers have identified specific colostral metabolites that indicate the shift to lactogenesis II including decreasing sodium, chloride, and total nitrogen levels and increasing lactose and citrate levels.^10–12 A recent pilot study found that women with GDM have a significantly different colostral lactose and citrate concentrations compared with nondiabetic women, although there was only a one-time sample collection within 72 hours postpartum,¹³ potentially missing delayed peak expression of the colostral metabolites. The current study aimed to examine the changes in colostral metabolites of lactose and citrate levels through the first week postpartum among mothers with and without GDM.

Objective

The primary objective of the study was to determine if colostral lactose and citrate concentrations were lower in mothers with GDM compared with nondiabetic mothers, comparing the dynamic concentrations of the colostral metabolites in the first week postpartum.

Materials and Methods

A prospective observational study of 133 postpartum women, 13 (9.8%) with GDM and 120 (90.2%) without GDM, was conducted to examine differences in the selected colostral metabolite concentrations of lactose and citrate. Participants included otherwise healthy primiparous and multiparous mothers with and without GDM who delivered a singleton, term (>37 weeks gestation) infant between 2016 and 2018 in rural Appalachia, where diabetes and GDM rates are relatively high compared with overall national rates.¹⁴ The primary question of interest was whether there were differences in the changes in lactose and citrate concentrations based on diabetes status (GDM versus non-GDM), while controlling for potential covariates.

With guidance by a lactation consultant, mothers collected <3 mL colostrum samples daily by hand expression or pumping on days 2 through 7 postpartum. The samples were stored in the woman's home freezer for up to 2 weeks when they were transported to a −80°C freezer in the university until further transport to the laboratory for analysis. Milk samples were stored in the laboratory freezer at −80°C, then were thawed on ice, and centrifuged at 15,000g for 20 minutes at 4°C to separate the fatty layers from the aqueous fractions, which were transferred into a washed Amicon filter with 3,000 dalton molecular weight cutoff (Millipore, Burlington, MA) and then further centrifuged at 13,000g for 45 minutes at 4°C.

Filtered samples were mixed with NMR buffer for nuclear magnetic resonance (NMR) spectroscopy analysis, as previously described.^15,16 One-dimensional proton NMR spectra were collected on a Bruker Avance III 850 MHz NMR spectrometer equipped with TXI 850 MHz S3 5 mm probe with z-gradient at 298K 5 mm samples. Metabolite concentrations were determined by comparing the integrated peak areas of non or minimally overlapped peaks from lactose and citrate with calibration curves generated from the integrated areas measured from NMR spectra collected on pure solutions of lactose and citrate measured at concentrations of physiologically relevant ranges for each of the metabolites at 850 MHz using the identical pulse sequences and pulse sequence parameters as used for the milk sample data collection.

Statistical analysis

Mean and standard deviations were calculated to describe continuous variables, whereas frequencies and proportions were used to describe categorical variables. Independent samples t-test, Mann–Whitney test, and Pearson's correlation coefficient were used to examine the relationship between maternal characteristics and lactose and citrate levels. Spaghetti plots and plots of means over time stratified by GDM status were used to visually inspect the profiles of lactose and citrate levels over time. We used SAS 9.4 (SAS Institute, Inc., Cary, NC) to perform a linear mixed effects analysis using PROC MIXED procedures to determine the relationship between GDM status, time, and levels of lactose and citrate. Statistical significance was defined as a p-value <0.05.

As fixed effects, we entered GDM status, time, covariates, and three interaction terms time*GDM, time*time, time*time*GDM into the model allowing quadratic trends over time. As random effects, we had intercepts for subjects as well as by-subject random slopes for the effect of time. p-Values were obtained by likelihood ratio tests of the full model with the effect in question. We used the Akaike information criterion to assist with selecting the appropriate statistical model.

A post hoc power analysis was conducted using the software package GPower.¹⁷ The sample size of 133 was used for the statistical power analyses with an alpha level of p < 0.05, and repeated measures and within–between interactions equation were used as a baseline with effect sizes defined according to Cohen's¹⁸ guidelines. The post hoc analyses revealed that there was more than adequate power at the moderate-to-large effect size level as the statistical power exceeded 0.99 for the detection of a moderate-to-large effect size. The study was approved by the investigators' academic Institutional Review Board.

Results

Among the 133 postpartum participants, a greater proportion of the sample were non-GDM (90.2%) and were predominantly multipara (65.4%). Baseline characteristics and differences between the women based on diabetes status are presented in Table 1. Diabetes management among women with GDM included diet only (n = 4, 40.0%); diet and exercise (n = 2, 20.0%); diet, exercise, and oral medication (n = 1, 10.0%); diet, exercise, and insulin (n = 2, 20.0%); and insulin only (n = 1, 10.0%). In bivariate analyses, mothers with GDM compared with nondiabetic mothers and primiparous compared with multiparous mothers had significantly lower metabolite baseline concentration of lactose (p = 0.027 and p = 0.013, respectively), whereas there were no significant differences in citrate levels. There were no statistically significant correlations detected between age, BMI, and baseline metabolite concentrations (Table 2).

Table 1.

Baseline Characteristics of the Study Sample (n = 133)

	Overall	Non-GDM	GDM	p^a
	n (%)	n (%)	n (%)	p^a
Marital status				1.000
Never married	9 (6.8)	8 (88.9)	1 (11.1)
Married/living together	124 (93.2)	112 (90.3)	12 (9.7)
Race/ethnicity				0.032
Non-Hispanic white	119 (89.5)	110 (92.4)	9 (7.6)
Other	14 (10.5)	10 (71.4)	4 (28.6)
Education				0.377
High school	6 (4.5)	6 (100.0)	0 (0.0)
Some college	32 (24.1)	27 (84.4)	5 (15.6)
College graduate	95 (71.4)	87 (91.6)	8 (8.4)
Employed				1.000
No	32 (24.1)	29 (90.6)	3 (9.4)
Yes	101 (75.9)	91 (90.1)	10 (9.9)
Insurance				0.243
Private/employer-based	105 (79.0)	97 (92.4)	8 (7.6)
Medicaid/Medicare	26 (19.5)	21 (80.8)	5 (19.2)
Other	2 (1.5)	2 (100.0)	0 (0.0)
WIC enrollment				0.126
No	112 (84.2)	103 (92.0)	9 (8.0)
Yes	21 (15.8)	17 (81.0)	4 (19.0)
Currently smoking				1.000
No	126 (94.7)	113 (89.7)	13 (10.3)
Yes	7 (5.3)	7 (100.0)	0 (0.0)
Mode of birth				0.755
Vaginal	96 (72.2)	87 (90.6)	9 (9.4)
Cesarean	37 (27.8)	33 (89.2)	4 (10.8)
Parity				0.137
Primipara	46 (34.6)	39 (84.8)	7 (15.2)
Multipara	87 (65.4)	81 (93.1)	6 (6.9)
Prepregnancy BMI				0.365
<30 kg/m²	87 (65.9)	80 (91.1)	7 (8.1)
≥30 kg/m²	45 (34.1)	39 (86.7)	6 (13.3)
Initiate breastfeeding within 1 hour				0.481
No	29 (22.0)	25 (86.2)	4 (13.8)
Yes	103 (78.0)	94 (91.3)	9 (8.7)
Formula				0.005
No	113 (85.0)	106 (93.8)	7 (6.2)
Yes	20 (15.0)	14 (70.0)	6 (30.0)
Skin-to-skin practice				1.000
No	5 (3.9)	5 (100.0)	0 (0.0)
Yes	124 (96.1)	111 (89.5)	13 (10.5)
	Mean (SD)	Mean (SD)	Mean (SD)	p^b
Age (years)	30.6 (4.40)	30.5 (4.49)	31.5 (3.41)	0.472
Prepregnancy BMI (kg/m²)	28.2 (7.04)	27.9 (7.08)	31.1 (6.11)	0.118
Gestational age (weeks)	39.4 (1.19)	39.4 (1.20)	39.1 (1.15)	0.332
Birth weight (kg)	3.6 (0.45)	3.6 (0.45)	3.4 (0.44)	0.222
Lactose (mM)	119.0 (38.73)	121.6 (36.96)	91.9 (48.23)	0.027
Citrate (mM)	1.0 (0.81)	1.1 (0.83)	0.7 (0.56)	0.163
Glucose (mM)	0.6 (0.31)	0.6 (0.31)	0.5 (0.31)	0.419

Missing data: prepregnancy BMI (n = 1), initiate breastfeeding within 1 hour (n = 1), skin-to-skin practice (n = 4).

Frequencies, and percentages are reported with respect to the row attributes.

Fisher's exact test p-values.

Independent sample t-test p-values.

BMI, body mass index; GDM, gestational diabetes mellitus; M, mean; n, frequencies; SD, standard deviation; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children.

Table 2.

Descriptive Statistics of Baseline Metabolite Levels by Maternal Characteristics

	n	Lactose (mM)		Citrate (mM)
	n	Mean (SD)	p	Mean (SD)	p
Mode of delivery			0.691		0.525
Vaginal	78	119.87 (38.55)		0.99 (0.74)
Cesarean	24	116.24 (40.03)		1.13 (1.03)
GDM status			0.027		0.163
Non-GDM	93	121.63 (36.96)		1.06 (0.83)
GDM	9	91.94 (48.23)		0.66 (0.56)
Parity			0.013		0.147
Primipara	28	103.66 (44.65)		0.83 (0.63)
Multipara	74	124.82 (34.84)		1.09 (0.87)
Age (in years)	102	0.06	0.580	0.17	0.091
Prepregnancy BMI (kg/m²)	102	−0.16	0.098	−0.09	0.341

BMI, body mass index; GDM, gestational diabetes mellitus; n, frequencies; SD, standard deviation.

Table 3 displays fixed effect estimates from the full model. There were statistically significant differences over time for lactose (p < 0.001) and citrate levels (p < 0.001) for all mothers. Controlling for other predictors, among mothers with GDM, there was an average decrease of 47.43 and 1.03 mM in lactose and citrate levels, respectively, compared with nondiabetic mothers. Overall, both lactose and citrate levels show significant linear and quadratic effects over time. Additionally, there were statistically significant differences in the linear and quadratic trends of lactose levels by GDM. However, quadratic trends for citrate did not differ by GDM.

Table 3.

Fixed Effect Estimates for Model Predicting Lactose and Citrate Concentrations

Effect	Lactose		Citrate
Effect	Estimate (SE)	p	Estimate (SE)	p
Intercept	106.65 (17.70)	<0.001	0.10 (0.46)	0.819
Age (in years)	−0.46 (0.46)	0.321	0.01 (0.01)	0.602
Prepregnancy BMI (kg/m²)	0.45 (0.28)	0.117	0.01 (0.01)	0.266
Mode of delivery
Cesarean (ref = vaginal)	−2.18 (4.40)	0.621	0.05 (0.11)	0.633
Parity
Primiparous (ref = multiparous)	−5.31 (4.20)	0.208	0.19 (0.11)	0.079
GDM
GDM (ref = non-GDM)	−47.43 (16.22)	0.004	−1.03 (0.42)	0.014
Time (in days)	24.90 (2.45)	<0.001	0.57 (0.07)	<0.001
Time^*GDM	21.69 (8.10)	0.008	0.52 (0.23)	0.028
Time^*time	−2.49 (0.33)	<0.001	−0.06 (0.01)	<0.001
Time^time^GDM	−2.29 (1.06)	0.032	−0.05 (0.03)	0.108

BMI, body mass index; GDM, gestational diabetes mellitus; n, frequencies; SD, standard deviation.

Discussion

To the best of our knowledge, this is the first published research study demonstrating dynamic concentration level differences in lactose and citrate using colostral samples collected daily over the course of the first week postpartum from mothers with GDM and mothers without diabetes. Researchers have found that with the shift to lactogenesis II in early postpartum, colostral lactose and citrate levels significantly increased over time¹¹ with delayed shifts among women with type 1 diabetes.^10,19,20 Findings of our study show that mothers with GDM have a sustained delayed shift in colostral lactose levels compared with women without diabetes. These findings are further supported by a previous study of maternal serum biomarkers, lower insulin, and adiponectin levels were associated with later onset of lactation among mothers who had comparatively poor prenatal glucose tolerance.²¹

While primiparity negatively influenced lactose levels, the differences were not significant when time was included in the fixed effects model. Other studies have found primiparity to be associated with delayed lactogenesis II, but they used different methods of data collection, including using solely maternal report and not examining colostral metabolite levels. For example, researchers found a perceived delay in lactogenesis II reported by primipara mothers who were overweight or obese, 30 years of age or older, infant birth weight over 3,600 g, and experienced breastfeeding issues.²²

In a study of mothers with GDM, reports of delayed lactogenesis II were associated with obesity, older age, insulin treatment for GDM, and suboptimal breastfeeding in the hospital, but biomarkers were not examined.⁸ In our study, BMI and maternal age were not factors associated with colostral lactose and citrate levels, leaving diabetes status to be the primary influencing factor when comparing mothers with and without GDM. This finding suggests that mothers with GDM experience a lag in the dramatic increase of colostral lactose concentrations, potentially delaying the transition to lactogenesis II in the early postpartum period.

Women who sense delayed lactogenesis II may be concerned about insufficient milk supply and result in premature termination of exclusive breastfeeding and early formula supplementation.⁹ Moreover, a previous study found that women with GDM were more likely to supplement with formula.¹³ The current study findings along with previous research support further exploration to understand the clinical implications of slower increments of lactose concentrations in mothers with GDM.

A limitation of the study is that some women were unable to collect data every day during the first week postpartum for various reasons, including lack of time, having forgotten, return home from the hospital, and inability to collect sufficient milk sample. We, therefore, excluded women's samples where there were fewer than three distinct days of colostrum samples.

While only 9.8% of the participants had GDM, the post hoc analysis confirmed that there was sufficient number of participants in the study to detect moderate and large effect sizes, not for detecting a small effect. It is noteworthy that the prevalence of GDM in the current study reflects representative estimate of GDM for the general pregnant population in the United States.^14,23 Additionally, the coefficient of variation of metabolites measured using NMR differ by the concentration level of metabolites, ranging from 5% to 10% for high-concentration metabolites such as lactose to 15–30% for low-concentration metabolites such as citrate.²⁴ The potential implication of this variation should be considered when interpreting results from NMR-based metabolite measurements.

Conclusion

In summary, our primary objective of this follow-up study was to determine if colostral lactose and citrate concentrations were lower in mothers with GDM compared with nondiabetic mothers in the first week postpartum. We found that while lactose levels dramatically increase for all women in the first week postpartum, women with GDM have consistently lower lactose levels compared with nondiabetic women. We believe this information can be helpful as a benchmark for establishment of a pathway to lactogenesis II in the subgroup of mothers with GDM.

Footnotes

Acknowledgment

The authors thank Ohio University College of Health Sciences and Professions, School of Nursing, and Baker Fund for providing financial support of the study.

Disclosure Statement

No competing financial interests exist.

Funding Information

I.R.A.C. received support for the study from Ohio University's College of Health Sciences and Professions and the School of Nursing; I.R.A.C. and Z.T.H. received support from the Ohio University Baker Fund.

References

DeSisto

, Kim

, Sharma

. Prevalence estimates of gestational diabetes mellitus in the United States, Pregnancy Risk Assessment Monitoring System (PRAMS), 2007-2010. Prev Chronic Dis, 2014; 11:e104.

Soltani

, Arden

. Factors associated with breastfeeding up to 6 months postpartum among mothers with diabetes. J Obstet Gynecol Neonatal Nurs, 2009; 38:586–594.

Oza-Frank

, Chertok

, Bartley

. Differences in breast-feeding initiation and continuation by maternal diabetes status. Public Health Nutr, 2013; 18:727–735.

Oza-Frank

, Gunderson

. In-hospital breastfeeding experiences among women with gestational diabetes. Breastfeed Med, 2017; 12:261–268.

Haile

, Oza-Frank

, Azulay Chertok

, et al. Association between history of gestational diabetes and exclusive breastfeeding at hospital discharge. J Hum Lact, 2016; 32:NP36–NP43.

Verd

, de Sotto

, Fernandez

, et al. The effects of mild gestational hyperglycemia on exclusive breastfeeding cessation. Nutrients, 2016; 8:e742.

Bartick

, Schwarz

, Green

, et al. Suboptimal breastfeeding in the United States: Maternal and pediatric health outcomes and costs. Matern Child Nutr, 2017; 13:e12366.

Matias

, Dewey

, Quesenberry

, et al. Maternal prepregnancy obesity and insulin treatment during pregnancy are independently associated with delayed lactogenesis in women with recent gestational diabetes mellitus. Am J Clin Nutr, 2014; 99:115–121.

Brownell

, Howard

, Lawrence

, et al. Delayed onset lactogenesis II predicts the cessation of any or exclusive breastfeeding. J Pediatr, 2012; 161:608–614.

10.

Arthur

, Smith

, Hartmann

. Milk lactose, citrate, and glucose as markers of lactogenesis in normal and diabetic women. J Pediatr Gastroenterol Nutr, 1989; 9:488–496.

11.

Neville

, Morton

. Physiology and endocrine changes underlying human lactogenesis II. J Nutr, 2001; 131:3005S–3008S.

12.

Pang

, Hartmann

. Initiation of human lactation: Secretory differentiation and secretory activation. J Mammary Gland Biol Neoplasia, 2007; 12:211–221.

13.

Chertok

IRA

, Haile

, Silanikove

, et al. Colostral metabolites as indicators of lactogenesis II in women with gestational diabetes mellitus. Ann Pregn Care, 2017; 1:1001.

14.

Deputy

, Kim

, Conrey

, et al. Prevalence and changes in preexisting diabetes and gestational diabetes among women who had a life birth—United States, 2012-2016. MMWR Morb Mortal Wkly Rep, 2018; 67:1201–1207.

15.

Chihanga

, Ma

, Nicholson

, et al. NMR spectroscopy and electron microscopy identification of metabolic and ultrastructural changes to the kidney following ischemia-reperfusion injury. Am J Physiol Renal Physiol, 2018; 314:F154–F166.

16.

Schmahl

, Regan

, Rivers

, et al. NMR-based metabolic profiling of urine, serum, fecal, and pancreatic tissue samples from the Ptf1a-Cre; LSL-KrasG12D transgenic mouse model of pancreatic cancer. PLoS One, 2018; 13:e0200658.

17.

Faul

, Erdfelder

, Lang

, et al. G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods, 2007; 39:175–191.

18.

Cohen

JE.

Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc., 1988.

19.

Neubauer

, Ferris

, Chase

, et al. Physiology and endocrine changes underlying human lactogenesis II. J Nutr, 2001; 131:3005S–3008S.

20.

Murtaugh

, Ferris

, Capacchione

, et al. Energy intake and glycemia in lactating women with type 1 diabetes. J Am Diet Assoc, 1998; 98:642–648.

21.

Nommsen-Rivers

, Chantry

, Peerson

, et al. Delayed onset of lactogenesis among first-time mothers is related to maternal obesity and factors associated with ineffective breastfeeding. Am J Clin Nutr, 2010; 92:574–584.

22.

Nommsen-Rivers

, Dolan

, Huang

. Timing of stage II lactogenesis is predicted by antenatal metabolic health in a cohort of primiparas. Breastfeed Med, 2012; 7:43–49.

23.

Centers for Disease Control and Prevention. Gestational diabetes. 2019. Available at https://www.cdc.gov/diabetes/basics/gestational.html (accessed on November 1, 2019).

24.

Wang

, Goodpaster

, Kennedy

. Coefficient of variation, signal-to-noise ratio, and effects of normalization in validation of biomarkers from NMR-based metabonomics studies. Chemometr Intell Lab Syst, 2013; 128:9–16.