To set the stage for the second day of the workshop, Anna Maria Siega-Riz of the University of Massachusetts Amherst and chair of the workshop planning committee indicated that many of the topics to be discussed were not included in the Nutrition During Pregnancy (IOM, 1990) or Nutrition During Lactation (IOM, 1991) reports. The sessions were intended to be broader in nature than the sessions during the first day.
Over the past several decades there have been advancements in our understanding of the relationships between nonnutritive factors associated with diet and maternal and fetal health during pregnancy, and maternal and infant health during lactation. Although there have been long-standing recommendations on caffeine intake during pregnancy and lactation, the landscape of caffeine-containing products has dramatically changed, which has given rise to new considerations. The microbiome has also come to the forefront, with evidence to suggest that maternal diet may play a key role in its formation. The interplay between maternal metabolism and dietary composition has also been explored, with evidence to suggest it has an influence on bioactive compounds in breast milk, which in turn can influence infant growth and development. The fifth session of the workshop, moderated by Deborah O’Connor, interim chair and professor in the Department of Nutritional Sciences at the University of Toronto, explored these select topics where new development and evidence has emerged. Highlights from the session presentations are presented in Box 6-1.
Janet Thorlton, clinical associate professor with the Department of Health Systems Science, Urbana Campus at the University of Illinois at Chicago College of Nursing, provided remarks on new developments related to caffeine. Her presentation covered current recommendations and resources for monitoring intake, changes in the landscape of products, regulations and labeling, caffeine metabolism, and adverse effects.
Caffeine Intake Recommendations and Monitoring
Caffeine is the most widely consumed psychostimulant worldwide. It is naturally found in coffee, tea, and chocolate and is added to a variety of
foods and beverages. Caffeine content varies, with a regular 12-ounce soft drink containing up to 40 mg and 8 oz of tea and coffee having up to 50 and 100 mg, respectively. The caffeine content of energy drinks can range from 40 to 250 mg per 8 oz. In addition to being in the food supply, caffeine is found in some medications, cosmetics, and supplements, noted Thorlton.
Caffeine has been classified as Generally Recognized as Safe for general consumption up to 400 mg/day. For the past 30 years, caffeine intake recommendations for pregnant and lactating women have been to consume less than 200 and 300 mg/day, respectively. Intake recommendations are further reduced for lactating women whose infants are newborn or preterm. Despite the long-standing recommendations, Thorlton indicated controversy still exists regarding the safety of caffeine exposure during pregnancy.
Thorlton highlighted some key resources that can be used to monitor caffeine consumption patterns. Related to safety, the Center for Food Safety and Applied Nutrition Adverse Event Reporting System tracks adverse event reports related to foods, cosmetics, or dietary supplements. Two nationally representative surveys collect information related to caffeine: Kantar Worldpanel (a market research firm monitoring worldwide beverage consumption) and the National Health and Nutrition Examination Survey (NHANES). Nevertheless, as the market is rapidly changing, the patterns of use of caffeine-added products are not well understood. Estimating exposure to caffeine from foods, beverages, and supplements poses challenges, said Thorlton. More evidence is needed to understand caffeine sensitivity, consumption practices, and prevalence of use given the potential for reporting biases and inaccuracies.
Changes in the Landscape of Caffeinated Products
The variety of caffeinated products that have been introduced to the market has dramatically changed the caffeine landscape. “In the past 20 years, caffeinated energy drinks, sports drinks, juices, and waters have been introduced,” said Thorlton. Popular caffeinated energy drinks may contain carbohydrates and a mixture of ingredients to promote a more powerful stimulant effect.
Thorlton presented a brief timeline of events to provide context for how the market for caffeinated products was able to expand. She began in 1994, with the introduction of the Dietary Supplement Health and Education Act (DSHEA), which she suggested allowed for the proliferation of new types of caffeinated products to enter the market beginning the mid-1990s. Around 2010, amid reports of adverse events, the U.S. Food and Drug Administration (FDA) began investigating the labeling of caffeinated products. The Institute of Medicine released a report on caffeine safety in 2014 (IOM, 2014). In 2015, FDA began warning consumers about pure synthetic caffeine pur-
chased from the Internet, which may be intended for cosmetic purposes rather than human consumption. For safety reasons, FDA has asked manufacturers to voluntarily stop making superconcentrated forms of caffeine available. To better understand caffeine metabolism, FDA has recently released guidance on the inclusion of pregnant women in clinical trials.
Regulations and Labeling of Caffeinated Products
The dietary supplement market has expanded substantially over the past 30 years. Since the passage of the DSHEA, the industry has grown from a $4 billion market with approximately 4,000 products available to a $40 billion market representing more than 90,000 products, some of which contain caffeine, said Thorlton.
FDA regulates dietary supplements and medications differently. At present, dietary supplements are regulated under the DSHEA, although FDA is currently considering strengthening regulations and oversight, indicated Thorlton. Product safety and correct labeling is the responsibility of the dietary supplement manufacturers. For some caffeinated products, such as certain energy drinks, manufacturers have the option to market the product as a dietary supplement or as a food and beverage product. The decision has implications for regulations and labeling. Dietary supplements are regulated under the DSHEA and contain a Supplement Facts panel, whereas foods and some beverages are regulated under the Federal Food, Drug, and Cosmetic Act and bear Nutrition Facts panels. Thorlton noted that individuals using Supplemental Nutrition Assistance Program (SNAP) benefits, for instance, would be able to purchase an energy drink containing a Nutrition Facts panel, but would not be able to purchase an energy drink containing a Supplement Facts panel. She also cautioned that supplement labels can contain fine print with important details. In one example, a product included a footnote that provided the allergen information and recommended the product was not suitable for children or pregnant women. Concern has been raised about consumer misinterpretation of labels, especially for products in which an entire container is often consumed at one time despite it containing two servings. To address these concerns, manufacturers are now required to clearly label the serving size, calories, and nutrients for the entire container rather than per serving.
There are many factors that affect caffeine metabolism, clearance, and pharmacokinetics, said Thorlton. One such factor is genetic variability. This variability can affect the binding of caffeine to brain receptors, which in turn influences the experienced effect. People with the polymorphism
ADORA2A have been found to be caffeine sensitive. Genetic variability also influences the activity of CYP1A2, an enzyme responsible for metabolizing 95 percent of ingested caffeine. This polymorphism determines if a person is a fast or slow metabolizer of caffeine. In addition to genetics, alcohol intake, dietary factors, smoking, certain liver disease, and use of oral contraceptives can each affect caffeine clearance, said Thorlton.
Pregnancy affects caffeine clearance. Where in nonpregnant adults, caffeine has a half-life of approximately 3–4 hours, the half-life for caffeine in pregnant women increases across trimesters and can last 9–11 hours. Caffeine crosses the placental barrier, and the fetus relies on maternal clearance.
The complexities of lactational pharmacology make it difficult to discern the implications of caffeine intake during lactation. FDA has encouraged the inclusion of lactating women in clinical trials to better characterize effects in this group. Available evidence suggests that the iron content of breast milk is decreased among women who chronically consume coffee. While occasional caffeine intake is not contraindicated, chronic maternal intake has the potential to increase plasma levels among newborns.
Adverse Effects of Caffeine Intakes
Recent studies, reviews, and meta-analyses have explored the possible adverse health effects of caffeine intake during pregnancy and lactation. Sasaki et al. (2017), for instance, found that pregnant women who did not smoke, had caffeine intake above 300 mg/d, and had the genetic polymorphism for rapid caffeine metabolism were at increased risk of having an infant with decreased birth size. In their review of caffeine safety, Temple et al. (2017) identified pregnant women as a population group in which caffeine consumption could be harmful. A systematic review concluded that caffeine intake of 300 mg/day is generally not associated with adverse health effects among healthy pregnant women (Wikoff et al., 2017). For lactating women, McCreedy et al. (2018) concluded the available evidence on health effects of maternal caffeine intake on the breastfed child was limited and inconclusive. There have been suggestions that sugared beverages are a bigger threat to reproductive success than caffeinated beverages, said Thorlton.
MATERNAL DIET DURING PREGNANCY AND LACTATION AND THE INTERACTION WITH THE DEVELOPING INFANT MICROBIOME
“We know that the in utero environment shapes our metabolic hereditability in very unexpected and very creative ways,” said Kjersti Aagaard,
the Henry and Emma Meyer Chair in Obstetrics and Gynecology Professor at the Baylor College of Medicine and Texas Children’s Hospital. Over the course of her presentation, Aagaard provided an overview of interactions between the maternal diet and the developing infant microbiome, drawing on evidence in humans and nonhuman primates. She also discussed the relationships between maternal intake and human milk oligosaccharides (HMOs).
Evidence on the Developing Microbiome in Humans
Maternal exposures have the potential to influence the neonatal and infant microbiome in different ways, such as through intrauterine colonization, immune education, and colonization resistance (the occupying of specific microbes in a niche space). Although some vertical transfer of live microbes from the pregnant women to the developing fetus likely happens, the data supporting this are limited, indicated Aagaard. She suggested that focus should be given to how tolerance to commensal microbes is acquired, and how and when immune differentiation between these commensal microbes and pathogenic microbes occurs. Equally important are investigations into understanding the process of how and when colonization resistance to pathogenic microbes occurs, and understanding the role of commensal microbes in occupying these niches and thereby limiting the capacity of pathogenic microbes in competing for nutrients and substrates in body niches, Aagaard said.
Aagaard’s earlier work explored the constituents and activity of vaginal microbiomes of pregnant women. Compared to nonpregnant women, pregnant women’s vaginal microbiomes had a significantly distinct clustering and were both less rich and less diverse (Aagaard et al., 2012). Although reduction in microbiome diversity has been found to lead to disease states for other conditions (e.g., inflammatory bowel disease), it appears to be normal for pregnancy and changes as women transition from early to late pregnancy, then again to term delivery and postpartum. Aagaard explained it happens through an ecological process of microbial coexclusion and co-occurrence.
Vaginal microbiomes during pregnancy are not the same as those found in the infant’s gut during the first week of life (Aagaard et al., 2012; Chu et al., 2017; Jost et al., 2012). For instance, Lactobacillus species, which are dominant in the vaginal microbiome, are not the dominant genera present in the neonatal gut microbiome. This finding led Aagaard’s group to explore other sources and modifiers of the neonatal microbiome, including the placenta and the maternal diet (Aagaard et al., 2014; Prince et al., 2016; Seferovic et al., 2019). They found that the neonates of mothers who consumed more than 30 percent of energy from fat (high-fat diet) had
different gut microbiomes at birth than infants of mothers who consumed lower fat diets, and these changes in the microbiome community appeared to persist over time (Chu et al., 2017). The infants in the study, however, were breastfed, which made it difficult to determine if the findings stemmed from gestation or lactation exposures.
To further explore the relationship between maternal dietary fat intake and infant gut microbiome, Aagaard’s team conducted a population-based prospective, longitudinal cohort (Chu et al., 2017). Maternal intake of a high-fat diet was found to decrease the amount of Bacteroides species in infant stool at 6 weeks of age, whereas ever being exposed to infant formula increased it. Aagaard reported that maternal high-fat diet and ever consuming infant formula were stronger drivers of differences in the amount of Bacteroides species in the infant gut than Cesarean delivery.
Evidence from Nonhuman Primate Models
Using the Japanese macaque as a nonhuman primate model, Aagaard’s group assessed the influence of the maternal diet on the infant’s microbiome and explored whether the effects persisted in the long term (Ma et al., 2014; Pace et al., 2018; Prince et al., 2019). Dams were fed either a control diet (13 percent energy from fat) or a “great American diet” (35 percent energy from fat). Over time, approximately one-third of the monkeys did not develop obesity despite being on the higher-fat diet (Harris et al., 2016; Suter et al., 2011, 2012), which allowed Aagaard’s team to tease apart effects of the high-fat diet as opposed to obesity. For the fetal studies, the monkeys were delivered at the equivalent of 32 weeks gestation for humans (Aagaard-Tillery et al., 2008; McCurdy et al., 2009). As pertained to the work on the effect of the maternal high-fat diet exposure during gestation, they compared the fetal microbiomes to that of their siblings and nonsiblings, which did not undergo fetal necropsy but rather underwent necropsy at 1 or 3 years of age (Ma et al., 2014; Pace et al., 2018; Prince et al., 2019). They found that the fetal microbiome was sparse and aligned closer to the microbiomes of the placenta, followed more distantly by the maternal sites of the oral, vaginal, and gut communities. The fetal, infant, and juvenile microbiomes of offspring whose mothers consumed the high-fat “great American diet” were distinct, even if the offspring consumed a control diet (Ma et al., 2014; Pace et al., 2018; Prince et al., 2019). This finding suggests that a maternal high-fat diet has a lasting imprint on the offspring’s microbiome, said Aagaard. She went on to note that such differences were not seen among offspring whose mothers were obese but had been reverted to a control diet prior to pregnancy.
A follow-up study sought to identify interventions that could correct the influence of the maternal high-fat diet. Offspring given probiotics in a
symbiotic preparation showed improvements in total cholesterol (Pace et al., 2018). However, discontinuation of the probiotics or being challenged with a high-fat diet reversed changes seen with the probiotics. The probiotic intervention, like the control diet, had no lasting effect on the offspring’s microbiome, further suggesting that the maternal imprint does not appear to be reversible (Ma et al., 2014; Pace et al., 2018; Prince et al., 2019).
Maternal Diet and Human Milk Oligosaccharides
Aagaard’s team has also explored whether there are certain exposures in the early-life environment that help to shape the neonatal and infant microbiome via macromolecule substrates, and whether those macromolecules are similarly affected by maternal diet. As background, Aagaard introduced the notion that while human milk is known to be the optimal nutrition source for infants, the “how and why” this is true remains elusive. She noted that other macromolecules that are found in human milk include oligosaccharides. Specifically, HMOs are favorable macromolecules that are indigestible by the infant but serve as substrates by microbiota. Her team conducted a randomized crossover trial with two cohorts of lactating women (Mohammad et al., 2011; Munch et al., 2013) to investigate the effects of different dietary exposures on breast milk composition, including both HMOs and the microbiome. One cohort was tested with different carbohydrate sources (glucose versus galactose), while the other was tested with changes to the energy source (high carbohydrate versus high fat). Maternal intake of different carbohydrate sources did not change the macronutrient or energy content of the breast milk. In contrast, high fat intake increased the fat and energy content of breast milk. Both cohorts revealed differences in HMO speciation content. The HMO-bound fucose was significantly different between the glucose and galactose exposures, whereas HMO-bound sialydase was different between the high-carbohydrate and high-fat exposures. Of note, Aagaard’s group found significant associations between the concentration of HMO-bound fucose and the abundance of fucosidase (a bacterial gene that digests fucose moieties) harbored by milk microbiota. These studies collectively reveal a successive mechanism by which the maternal diet during lactation alters milk HMO composition, which in turn shapes the functional milk microbiome prior to infant ingestion.
Amniotic fluid samples collected mid gestation also appear to contain certain HMOs, including trisaccharide-2′-fucosyllactose. Aagaard explained that there are women who do and women who do not secrete fucosyllated HMOs (i.e., secretors and nonsecretors), which can be predicted from blood group antigens. The distribution of fucosyllated HMOs in amniotic fluid was found to follow the population-based estimate, relative to the
proportion of predicted secretors and nonsecretors. A second, more detailed analysis found the HMO 3′-sialyllactose in all 270 amniotic fluid samples evaluated. The amniotic fluid concentrations of HMOs were lower than concentrations found in breast milk but higher than concentrations found in urine. The profile of microbes found in amniotic fluid share similarities and overlap with microbes found in the neonate’s gut at delivery (Chu et al., 2017). The presence of HMOs in both amniotic fluid and breast milk suggest they are fundamentally important in potentially shaping what functions these community members play, said Aagaard. She said HMOs could play a role in helping low biomass commensal communities colonize, with potential for colonization resistance of pathogenic microbes, and in developing gut integrity early in fetal development.
Caveats and Considerations
Aagaard concluded her remarks by noting some key considerations. She indicated that the metagenetic evidence she presented only reflects the presence of the microbes—the analyses do not differentiate between live and dead microbes. Emphasizing the importance of colonization resistance, Aagaard acknowledged it is unclear at this time the extent to which the bacteria themselves facilitate this process versus the role of macromolecules and small molecule intermediates, including HMOs or other metabolites. She also suggested that mechanistic evidence is needed to understand how macronutrients exert different effects. Finally, Aagaard underscored the importance of further exploration into the metagenome—“how it is acquired, how it changes, how resilient it is, how resilient it is not, will help us understand what are the lasting footprints of the mom [on her infant’s microbiome],” she said.
FRUCTOSE AND OLIGOSACCHARIDES IN HUMAN BREAST MILK AND THEIR EFFECTS ON INFANT BODY COMPOSITION AND COGNITIVE OUTCOMES
Michael Goran, professor of pediatrics at Children’s Hospital Los Angeles, which is affiliated with the Keck School of Medicine of the University of Southern California, explained his interest in understanding the effects of breast milk sugar content stemmed from an exploration of modifiable determinants of obesity in children. Beginning with evidence on the relationships between sugar-sweetened beverages, breastfeeding, and childhood obesity, Goran’s remarks also provided an overview of HMOs, fructose exposure and metabolism during infancy, and low-calorie sweeteners.
Relationships Between Sugar-Sweetened Beverages, Breastfeeding, and Childhood Obesity
In 2012, Goran’s group published a journal article that explored the interplay between early introduction of sugar-sweetened beverages, breastfeeding, and childhood obesity among approximately 1,500 children enrolled in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) (Davis et al., 2012). The study revealed that breastfeeding a child for more than 12 months was protective against obesity risk associated with sugar-sweetened beverage consumption at 3–4 years of age. Goran noted that Project Viva had found that maternal intake of sugar-sweetened beverages in the second trimester of pregnancy was associated with markers of obesity in later childhood (Gillman et al., 2017); others have reported similar findings, he said. “There is certainly evidence now for a transmission effect of sugars during pregnancy, and data show that the earlier the introduction during childhood, the bigger the effect on obesity risk,” said Goran.
Overview of HMOs
Lactose, a disaccharide of glucose and galactose, is the primary sugar found in human milk at a concentration of 65 g/L. Oligosaccharides consist of different combinations of three or more sugars and are abundant in human milk, with concentrations similar to that of protein (5–15 versus 12 g/L, respectively). By contrast, the oligosaccharide content of cow milk is significantly lower than its protein content (0.05 versus 35 g/L, respectively). Echoing remarks made by Aagaard, Goran noted that HMOs are not absorbed, but rather serve as prebiotics that are digested by the microbes of the infant gut.
Studies have explored differences in HMO concentrations between and within lactating women. Analyses of data from the Canadian CHILD Cohort found that women who secreted 2′-fucosyllactose in their breast milk had different HMO profiles than those who did not secrete it (Azad et al., 2018). Goran added that 2′-fucosyllactose secretion is genetically determined. HMO composition appears to be highly variable between women. Although HMO concentrations for a given woman appears stable over the course of 1 day, there is evidence that their concentrations can dynamically change over time. For instance, concentrations of 3′-sialyllactose appear to increase over time, whereas concentrations of 6′-sialyllactose decrease over time. These two HMOs have similar structures and play a role in delivering sialic acid for infant brain development. Goran noted that a project is currently under way investigating HMO compositional changes that occur with breastfeeding beyond 12 months.
Fructose Metabolism and Exposure During Infancy
Despite having the same chemical formula, fructose and glucose have different metabolic fates. Glucose is used for energy throughout the body. In contrast, almost all of the fructose consumed is taken up by the liver in an insulin-independent manner and converted to triglycerides. The triglycerides can either be packaged in very low-density lipoproteins and exported out of the liver or remain in the liver. The intake of fruit juice or beverages containing high-fructose syrup increases the amount of fructose in the liver, which can lead to the accumulation of fat in the liver and for more lipids to be in circulation.
An infant’s natural feeding environment does not contain fructose. Rather, the infant consumes lactose and uses the glucose for energy and the galactose for brain development. Infants are not born with GLUT-5 transporters needed to absorb fructose. During this time, the fructose will remain in the infant’s gut and serve as a prebiotic to gut bacteria. The machinery to absorb fructose develops with sufficient fructose intake.
In a small pilot study of 27 lactating women, detectable levels of fructose were found in breast milk, although the concentrations were variable. Fructose content of breast milk was positively associated with measures of infant body composition, including body weight, lean mass, fat mass, and bone mineral content (Goran et al., 2017).
An acute crossover feeding trial assessed breast milk composition in the 6 hours after lactating women were challenged with a beverage containing high-fructose corn syrup and a beverage with a noncaloric sweetener (Berger et al., 2018). Breast milk glucose concentrations rose after consumption of the sugar-sweetened beverage, but was only significantly different at the 2-hour time point after intake. The rise in breast milk fructose concentrations were much faster and stayed elevated longer when the women consumed the high-fructose syrup beverage.
Goran acknowledged that concentrations of breast milk glucose are in milligrams per milliliter, whereas concentrations of fructose are in micrograms per milliliter, which raises the question as to whether such a relatively low exposure has an effect on the infant. Investigators at the University of California, Los Angeles, have explored this concept, assessing the effects of different fructose concentrations on preadipocyte gene expression (Shepherd et al., 1993). GLUT-4 expression was increased for all fructose exposures relative to the control condition without fructose. Goran noted that the increase in GLUT-4 expression could have important implications for cellular metabolism, adipogenesis, and glucose intolerance.
The Mothers Milk Study, currently in progress, seeks to build and expand on the findings about breast milk composition and the implications for children’s health. The study recruited 240 Hispanic mother–infant
dyads and has been collecting maternal milk samples, infant stool samples, and measures of maternal and infant diet and obesity every 6 months over the children’s first 2 years of life. The study seeks to find maternal dietary factors and/or breast milk compounds associated with healthy gut bacteria colonization, obesity prevention, and cognitive development promotion. Goran presented some preliminary findings assessing infants in the first 6 months. Of the 17 most abundant HMOs identified in breast milk samples, lacto-N-fucopentaose II (LNFPII) is the only one that appeared to have an inverse relationship with infant weight change in the first 6 months.
The Mothers Milk Study is also investigating different mechanisms and gut bacteria products that may affect metabolism, brain development, energy regulation, and infant feeding behaviors. One such analysis under way is exploring whether HMO content is associated with food responsiveness among infants at 1 month of age. Goran admitted interpretation of the data is challenging, as food responsiveness at such an early age may very well be beneficial, despite it being a risk factor for obesity later in childhood. Another analysis is looking for a relationship between HMOs and brain development. Animal studies have found significant relationships between 2′-fucosyllactose and learning, memory, and attention in rodents, which has led some companies to add it to infant formulas. However, he indicated that no human studies have established such links. The Mothers Milk Study data were used to evaluate longitudinal exposure to various HMOs and their relationships with cognitive, language, and motor outcomes in the first 2 years of life. The results suggest a positive relationship between the number of milk feedings and children’s cognitive outcomes at 24 months (Berger et al., 2020). Much of the variation seen was explained by breast milk concentration of 2′-fucosyllactose, which was significantly associated with more favorable cognitive test outcomes at 1 month of age.
Goran concluded his presentation by discussing low-calorie sweetener intakes. An analysis of NHANES data found positive relationships between low-calorie sweetener intake and both energy and sugar intake among children. Children who consumed both low-calorie sweetened beverages and sugar-sweetened beverages consumed more than 400 kcal/day more than children who only consumed water (Sylvetsky et al., 2019). “Low-calorie sweeteners are not doing the job they are supposed to do,” said Goran.
Evidence suggests consumption of low-calorie sweeteners is increasing among pregnant women. Nearly one-quarter of pregnant women in the United States consume low-calorie sweeteners on a daily basis, primarily from beverages (Sylvetsky et al., 2019). Infants of women who consumed more low-calorie sweeteners during pregnancy were at a two-fold increased
risk of obesity by 2 years of age (Azad et al., 2016). Goran noted that this finding, which has been shown in other studies, underscores that maternal diet during pregnancy can have long-lasting effects on their children. He also remarked that low-calorie sweeteners are also transferred to infants via breast milk (Rother et al., 2018).
Following Goran’s presentation, he, Aagaard, and Thorlton participated in a session discussion, moderated by O’Connor. Questions raised by audience members touched on topics related to maternal intake, breast milk and infant formula, and fructose.
Several of the questions pertained to concepts related to maternal intake. Siega-Riz wondered if women whose diets are suboptimal entering pregnancy could alter fetal imprinting if they took probiotics early and throughout pregnancy. Indicating that pregnant women are generally motivated to comply with dietary recommendations when they understand the benefits, Aagaard suggested a more prudent approach would be to improve dietary intake. From the audience, Patrick Catalano of the Tufts University School of Medicine added that probiotic trials do not typically find benefits, and those that do generally report reduced risk of gestational diabetes. Aagaard responded that many variables could affect the effect of a probiotic (e.g., dose, strain). Given this complexity, she underscored the importance of focusing on improving intakes through food, which can feed the commensal microbiota.
Regarding Aagaard’s suggesting an imprinting of maternal high-fat diets on the offspring microbiome, Carol Dreibelbis of 1,000 Days asked if the type of fat affected the infant’s microbiome. Aagaard explained that quality of fat was embedded, to an extent, in the groupings; women who consumed fat only from plant sources all fell in the low-fat group. She acknowledged that it was difficult to tease apart different effects of saturated and unsaturated fat.
On the topic of imprinting, Erica Gunderson of the Kaiser Permanente Northern California Division of Research asked if there were human data on the effects of changing maternal diet during lactation. Aagaard admitted that, in the nonhuman primate data, it is difficult to differentiate between effects of diet during gestation versus lactation owing to the inability to cross-foster in the Japanese macaque. Aagaard questioned the clinical relevance and practical reality of trying to implement specific dietary changes during lactation, noting that longitudinal studies that have investigated this
raise questions about which dietary changes actually result in beneficial effects related to changes to the infant microbiome. Gunderson followed up by asking about the implications for women who had gestational diabetes who then have high sugar-sweetened beverage intake postpartum. Aagaard underscored the importance of refocusing women on healthful nutrition during this life stage. Goran added that there is evidence that high-sugar diets are associated with greater postpartum weight, while high-fiber diets are protective. He acknowledged, however, that additional evidence on these relationships is needed.
An unidentified webcast audience member asked the session speakers if they would recommend that a woman with obesity still breastfeed even if she was a heavy consumer of sugar-sweetened beverages or caffeine. Thorlton, referring to caffeine intake guidelines she described in her presentation, noted that scaling back on breastfeeding might be warranted when the infant is a newborn or preterm. With respect to sugar-sweetened beverages, Goran indicated the mother should reduce her intake and not substitute with low-calorie sweetened products. Aagaard said that the recommendation would not be to breastfeed less. She challenged the concept of focusing on women with obesity, and suggested that greater attention needed to be given to nutrition across the life span, including during lactation. The evidence suggesting that maternal diet affects HMO composition in breast milk does not speak to the directionality of this effect, “whether that may in part have some adaptive influence on the baby, we do not know. But just because it is different does not make it bad,” said Aagaard.
Breast Milk and Infant Formula
Becky Banks, a midwife in private practice, asked the speakers if any of the breast milk findings they presented differ by whether the milk was consumed directly from the breast or was expressed. Goran stated that the Canadian Healthy Infant Longitudinal Development (CHILD) Study found a difference in HMOs between breast milk from the breast versus expressed breast milk consumed from a bottle. Aagaard added that there are likely maternal effects. “The source matters, and it probably matters not just for the baby, but for the mom,” said Aagaard.
An unidentified audience member asked Goran to expand on his comment about HMOs being added to infant formula, despite a lack of human data on their effects. Admitting he was not well versed in the full regulations, Goran noted that 2′-fucosyllactose is approved by FDA as Generally Recognized as Safe and 3′-sialyllactose has also been recently reviewed for such approval. Natural breastfeeding cohorts have demonstrated the benefits of these compounds, noted Aagaard. However, she raised questions about the effect of 2′-fucosyllactose being consumed by infants of non-
secretors. O’Connor added that there are studies that have been conducted in Europe that have explored the addition of HMOs to infant formulas, but mostly these have looked at safety and effects on growth and not cognitive outcomes.
Cindy Turner-Maffei of the Healthy Children Project wondered if milk from mothers who consume low-calorie sweeteners would have a taste difference. Goran indicated there are anecdotes, but he was not aware of published literature on the topic.
Referring to Goran’s comment about GLUT-5 transporters, Banks wanted to know when an infant starts absorbing and metabolizing fructose. Goran stated that the available evidence comes from animal studies, which suggest GLUT-5 is not active until fructose is introduced in the diet. He suggested that human-specific data are lacking.
A webcast audience member asked if any fructose is converted to glucose in the liver. Goran said that in mouse models, low fructose concentration can lead the gut to convert fructose to glucose, but he said that in humans, fructose reaching the liver will be used for de novo lipogenesis.
This page intentionally left blank.