Maternal Obesity, Breast Milk Composition, and Infant Growth

Overview

Today the majority of pregnant women in the United States are either overweight or obese at conception with their offspring having greater adiposity at birth, a 2-fold greater risk of later obesity, and neonatal insulin resistance. It was long thought that breast milk composition was fairly uniform among women, having been optimized through evolutionary time to provide adequate sole nutrition for the growing infant regardless of the environmental circumstances. However, recent evidence shows that breast milk is a highly complex fluid with significant inter-individual variation in hormonal and cytokine concentrations, human milk oligosaccharides (HMOs) and other features. Pervasive maternal obesity and gestational diabetes are evolutionarily novel conditions for the human species but little effort has yet been made to systematically examine how they are associated with breast milk adipose-tissue derived hormone and cytokine (adipocytokine) variation, HMOs, or other features, or whether that variation relates to infant metabolic status. The objective of this study is to comprehensively assess the "lactational programming" hypothesis, that is, whether or not recently documented variation in breast-milk composition is related to both maternal and infant metabolic status. The central hypothesis is that a graded, dose-response relationship between maternal adiposity and GDM exists with adipocytokine concentrations, HMOs, and other features in breast milk and that the milk concentrations of these features are associated with altered body composition in their exclusively breast-fed offspring. The results of the study will be used to design interventions to reduce maternal weight during pregnancy and lactation and to augment lactation education materials to focus on the needs of breastfeeding women with obesity and GDM.

General Study Design A prospective observational pregnancy cohort of 600 women who exclusively breastfeed their infants to at least 1 month post-partum, will be recruited in the second trimester of pregnancy, across a wide range of maternal pre-pregnancy BMI, with equal numbers of normal weight, overweight, and obese women, and with and without Gestational Diabetes Mellitis. The women will be recruited at two study centers (University of Oklahoma Health Sciences Center in Oklahoma City, OK and Health Partners Research Foundation in Saint Paul/Minneapolis, MN), with maternal and infant diet, body composition and metabolic health followed prospectively at 1, 3 and 6-months post-partum in Specific Aims 1 and 2. In Specific Aim 3, a subset of 60 of these mother/infant dyads "Mothers in Focus" (30 OB and 30 NW mothers) will be recruited for additional biospecimen collection, including milk samples at 2 and 6-months, maternal body composition changes using DXA, infant DXA, metabolomic comparison of serial milk samples, and follow-up to 12 rather than 6-months. Data Collection Elements: Maternal Adiposity. Maternal pre-pregnancy BMI status (NW, OW, OB) is the primary exposure used to power the study. Pre-pregnancy BMI will be determined using first (baseline) clinic weights from the electronic medical records (EMR) (no more than 90 d. after last menstrual period), and stature measured by research staff at the mother's first visit to our research centers. BMI will be categorized as underweight (BMI \<18.5), normal weight (BMI 18.5 - \<25), overweight (BMI 25 - \<30) and obese (BMI ≥30 kg/m2). Gestational weight gain (GWG) and postpartum weight loss/retention:Total GWG will be calculated as the difference between delivery weight and baseline weight (as documented above). If weight at delivery is not available, the last prenatal weight in the EMR, ≤14 days from birth, will be used. Trimester specific GWG will be calculated from the EMR. GWG will be used as a continuous and as a categorical exposure (excessive, appropriate, and insufficient, according to the IOM 2009 guidelines. Early PP weight loss (from delivery to 1-month postpartum), and later PP weight loss from 1-month to 3-months PP will be calculated. Concurrent changes in maternal weight and milk adipocytokine levels from 0 - 1 month and 1 - 3 months will be examined, hypothesizing that greater PP weight loss (lower weight retention) is associated with lower levels of milk adipocytokines. Gestational diabetes will be examined as a predictor of milk composition variation. Milk Collection: Changes in milk composition will be assessed from 1 to 3 months. Definition of exclusive breast feeding: Study staff will confirm that all infants have been exclusively breast milk fed (i.e. no formula, no liquid other than water and no solids) at least 14 days prior to the 1-month visit and plan to continue exclusively breastfeed to 3 months. Exclusive breast milk feeding will include any combination of expressed breast milk and breast milk fed via the breast. Breast Milk Collection Protocol: Mother and child will report to the study site at 1 and 3-months (± 5 days) postpartum, between 8:00-10:00 am, at least 1½ hours since the last feeding and while the mother has fasted at least 1 hour. A pre-feeding infant weight will be obtained using a high sensitivity scale, the mother will be encouraged to feed the child ad libitum from both breasts until volitionally satisfied, and a postfeeding infant weight will be obtained (difference = milk output). Questionnaire information, maternal anthropometry, and infant body composition data will be collected, and 2 hours later, the mother will provide a single breast expression sample using an electric breast pump (Medela, Inc., provided by the study team). Breast milk will be mixed, aliquoted and stored at -80°C. Milk Adipocytokine Assays: Milk samples will be thawed in the refrigerator and vortexed. Milk fat will be separated from the aqueous phase by centrifugation at 3,000g for 10 minutes at 4°C. The resulting skimmed milk will be assayed using commercially available immunoassay kits for insulin, IGF-1/BP, high molecular weight adiponectin, visfatin, leptin, IL-1β, IL-6, IL-10, and TNF-α. Glucose will be measured by the glucose oxidase method. To examine additional cytokine and growth factor concentrations (e.g., interleukins, interferons, TGF- β) in a manner that conserves sample, a multiplex assay (EMD Millipore, Billerica, MA) and the Luminex 200 Multiplex analyzer will be employed. Data Analysis for Specific Aim 1: For each of the milk adipocytokines (continuous dependent variables), separate mixed effects regression models (to account for serial milk data) will be constructed, first testing crude associations of milk IL-6 at both 1 and 3-months with each maternal adiposity variable (excessive GWG, pre-pregnancy BMI status, total GWG, trimester-specific GWG, early and later PP weight loss), and then minimally adjusted models (also adjusting for infant sex and gestational age, study center (OK or MN), maternal age, parity, ethnicity, and maternal socioeconomic status) and finally fully-adjusted regression models including those variables determined to be potential confounders. In secondary analyses, possible effect modification of the maternal adiposity associations will be assessed by adding interaction terms. If the interaction term is significant, a stratified analyses will be conducted. Data Collection for Specific Aim 2 Infant Body Composition. Total body composition will be measured using air displacement plethysmography (ADP) i.e. Pea Pod (v 3.1.1), a rapid, valid, highly reliable (% fat CV%\< 2.5% in our hands) method of total body volume estimation from which total body fat (FM) and fat-free mass (FFM) is determined using prediction equations. Both the University of Minnesota and the University of Oklahoma have the same ADP (same manufacturer, model and software versions). Infant insulin secretion and resistance at 6 months: C-peptide, or connecting peptide, as a marker of insulin secretion will be collected. Serum C-peptide is secreted in equimolar amounts by the pancreas as insulin, but is a more stable measure because it is not prone to degradation following hemolysis. C-peptide was used as a marker of fetal beta cell function in the HAPO (Hyperglycemia and Adverse Pregnancy Out-come) study to assess the effects of maternal obesity and gestational diabetes on adverse pregnancy and infant outcomes. Insulin resistance will be estimated using homeostasis model assessment (HOMA-IR). Blood will be obtained via heel-stick in the infant at the time of the 6 month visit to the study center, as is routine in our laboratories, which yields \~ 1.0 cc of blood. Whole blood will be kept on ice until centrifugation in our laboratories, and 0.5 mL of serum aliquoted at stored at -80 C until shipped to Dr. Fields' laboratory for analysis using ELISA (Millipore, Billerica, MA; Intra-assay CV = 5%; Inter-assay CV = 6%). Interpretation of C-peptide requires blood glucose level which will be measured in the same blood sample using the glucose oxidase method. Milk Macronutrient Content: Total fat content will be measured by the Mojonnier method Fatty acid composition will be assessed by gas chromatography using the International Dairy Federation (IDF) methods. Other standard methods will quantitate total nitrogen, ash, moisture, and carbohydrates. Infant Appetite, Satiety, and Diet: Shortened versions of the infant diet questionnaires used in the Infant Feeding Practices Study II at 1 month, 3-months, and 6-months will assess maternal report of infant dietary intake. Maternal perception of infant appetite will be measured using the Baby Eating Behavior Questionnaire. Infant Sleep: The Sadeh Brief Infant Sleep Questionnaire has been validated against actigraphy and will be used to assess sleeping behavior at each visit. Additional Potential Confounders: See Specific Aim 1. Data Analysis(Specific Aim 2) Statistical Analysis: The over-arching primary hypothesis is that milk adipocytokine concentrations will be associated with altered body composition from ages 1 to 6-months, and with elevated C-peptide or HOMA-IR at 6 months, independent of potential confounders. Example Hypothesis 2.a1: Higher levels of the appetite suppressing hormone leptin at 1 and/or 3-months will be associated with lower infant weight gain, reflecting specifically lower FFM gain from 1 - 6 months. We will use mixed effects linear and logistic regression as the primary analysis approach to address this and related hypotheses. Growth and body composition data will be serial (0 \[for weight\], 1, 3, and 6-months), which will be examined using mixed effects models which models the covariance structure of the repeated measures within subject and allows for subject-specific intercept and slopes to be modeled. Milk adipocytokines are the independent variables of interest. Because their individual relationships to infant growth and other outcomes may not be linear, quintiles or quartiles of milk adipocytokines as well as continuously distributed variables will be examined. Multiple milk adipocytokines will be included simultaneously in final models, to test their independent effects on infant outcomes and to assess their aggregate effects (variance explained).