Authors: Catherine P. Ward (aStanford Prevention Research Center, Department of Medicine, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Dalia Perelman (aStanford Prevention Research Center, Department of Medicine, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Lindsay R. Durand (aStanford Prevention Research Center, Department of Medicine, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Jennifer L. Robinson (aStanford Prevention Research Center, Department of Medicine, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Kristen M. Cunanan (bQuantitative Sciences Unit, Department of Medicine, Stanford University, Palo Alto, CA 94305, USA), Sailendharan Sudakaran (cDepartment of Microbiology and Immunology, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Roujheen Sabetan (aStanford Prevention Research Center, Department of Medicine, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Maggie J. Madrigal-Moeller (cDepartment of Microbiology and Immunology, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Christopher Dant (aStanford Prevention Research Center, Department of Medicine, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Erica D. Sonnenburg (cDepartment of Microbiology and Immunology, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Justin L. Sonnenburg (cDepartment of Microbiology and Immunology, School of Medicine, Stanford University, Palo Alto, CA 94305, USA), Christopher D. Gardner (aStanford Prevention Research Center, Department of Medicine, School of Medicine, Stanford University, Palo Alto, CA 94305, USA)
Categories: Article, Pregnancy, Microbiome, Fermented foods, High-fiber diet, Prenatal, Nutrition
Source: Contemporary clinical trials
Authors: Catherine P. Ward, Dalia Perelman, Lindsay R. Durand, Jennifer L. Robinson, Kristen M. Cunanan, Sailendharan Sudakaran, Roujheen Sabetan, Maggie J. Madrigal-Moeller, Christopher Dant, Erica D. Sonnenburg, Justin L. Sonnenburg, Christopher D. Gardner
Recent research underscores the crucial role of the gut microbiota in human health, particularly during states of altered homeostasis, including pregnancy. Additionally, it is not well understood how dietary changes during pregnancy affect the development of microbiomes of both mother and child.
Here, we describe the study design and methods for our randomized controlled trial, the fermented and fiber-rich foods on maternal and offspring microbiome study (FeFiFo-MOMS). We enrolled 135 women during early pregnancy, randomizing them to one of four diet increased fiber, increased fermented foods, increase in both, and no dietary intervention as a comparator arm. Samples were collected across pregnancy continuing to 18 months post-birth for clinical, microbiome, and immune marker analysis.
Our trial design intended to investigate the effects of dietary interventions—specifically, increased intake of high-fiber and fermented foods—on maternal gut microbiota diversity and its subsequent transmission to infants.
The FeFiFo-MOMS trial was designed to provide valuable insights into the modifiable dietary factors that could influence maternal and infant health through microbiota-mediated mechanisms and examine the broader implications of diet on pregnant mothers’ and infants’ health and disease.
Clinicaltrials.gov ID: NCT05123612
In the past decade, the significance of the gut microbiota in human health has become increasingly evident, with a focus on its association with various chronic diseases. The modern Western diet, characterized by its low fiber content, is believed to contribute to a shift in the gut microbiota population [1]. The deterioration of the gut microbiota has been implicated in numerous chronic diseases, including obesity, metabolic syndrome, inflammatory bowel disease, polycystic ovary syndrome, and others [2,3]. During pregnancy, women are at increased risk for developing metabolic conditions including gestational diabetes (GDM), with up to 15 % of women worldwide developing this condition [4]. Accordingly, there is a critical need to identify dietary strategies that can optimize human health and potentially prevent or reverse these metabolic conditions during and after pregnancy.
Diet plays a critical role in shifting gut microbiota communities [5–9]. Specifically, high-fiber and fermented foods have been shown to modulate the gut microbiome [10–12]. Studies comparing populations with different dietary fiber intakes have revealed increased abundance of Prevotella species, enhanced populations of butyrate-producing bacteria such as Faecalibacterium prausnitzii and Eubacterium rectale, and elevated levels of fiber-degrading Bifidobacterium species [13–16]. Central to the function of the gut microbiota is the digestion of microbiota-accessible carbohydrates (MACs), primarily found in dietary fibers. Through fermentation, these fibers are metabolized into a variety of small molecules, including short-chain fatty acids (SCFAs) which serve as a primary energy source for colon cells, and have been associated with anti-inflammatory properties [1,17,18]. Studies comparing high-versus low-fiber consumers have demonstrated not only taxonomic differences but also metabolomic signatures, with high-fiber diets associated with elevated fecal concentrations of butyrate, propionate, and acetate, as well as modified bile acid profiles [19–22]. Importantly, sustained low-fiber diets can lead to permanent loss of fiber-degrading taxa over generations, highlighting the long-term consequences of dietary fiber deficiency on microbiome composition [14,23,24].
The shift in dietary patterns over the last century, especially in Western societies, has led to a decline in dietary fiber intake. “Westernization” of microbiota has been observed in US immigrants, with loss of microbial taxa and functions accompanied by increased inflammatory markers [24–26]. The average population in the United States consumes significantly less fiber than recommended, with only about 5 % of the US population consuming adequate intake of fiber of approximately 14 g per 1000 kcal of intake per day, the amount recommended by the Institute of Medicine [27]. This decrease in fiber intake has been linked to an increased risk of various chronic diseases, including type 2 diabetes, obesity, and inflammatory bowel syndrome [28,29].
In addition to the decline in fiber intake, there has been a marked reduction in the consumption of fermented foods in Western diets compared to the historical norm [30]. Fermented foods, which were once a regular part of diets across various cultures, have drastically diminished in modern times due to the rise of industrialized food processing and refrigeration [31]. Common fermented foods include yogurt, kefir, kimchi, sauerkraut, and kombucha, each harboring distinct microbial communities. For example, traditionally fermented dairy products often contain different microbes compared to fermented vegetables [32–35]. Recent interventional studies have shown that regular consumption of fermented foods increases microbiota diversity and reduces inflammatory markers [10]. Specifically, a 10-week fermented food intervention conducted by our groups significantly increased microbiota diversity, including increased abundance of families Lachnospiraceae, Ruminococcaceae, and Streptococcaceae [10]. Reduced intake of these foods may lead to decreased microbial diversity and resilience, which are associated with a heightened risk of metabolic and inflammatory diseases [10,36–38]. Incorporating fermented foods into the diet, along with high-fiber foods, could help restore gut microbial diversity and improve overall health outcomes, particularly during vulnerable periods such as pregnancy.
To address the effects of high-fiber and fermented foods on maternal gut microbiota diversity and its subsequent transmission to infants, we designed a randomized clinical trial to explore different dietary approaches available to the general population, including increased high-fiber and fermented foods, evaluating their effects on the maternal gut microbiota diversity during pregnancy, and subsequent microbial transmission to infants after birth. By exposing pregnant women to these dietary components during pregnancy, we aimed to positively modify their microbiota to pass on beneficial microbes to their newborns. Additionally, we hypothesize metabolite microbial byproducts may cross the intestinal lining during pregnancy and impact both fetus and maternal immune system [10]. Long-term adherence to these dietary habits may also help prevent chronic diseases post-pregnancy; however, this hypothesis was not tested in our trial.
Historically, pregnant women are an understudied population [39]. We therefore aimed to examine the potential benefits of increased intake of high-fiber and fermented foods during pregnancy, which, to our knowledge, has not been studied before in a dietary intervention trial. There is a lack of data from controlled human intervention studies on fiber during pregnancy and their impact on microbiota diversity. Observational studies suggest that increased intake of fermented foods during pregnancy may desirably modulate the gut microbiota; however, the data gathered from our intervention trial is better designed to identify these changes [40]. We broadly hypothesize that dietary interventions aimed at increasing fiber and fermented foods intake during pregnancy will lead to an increase in gut microbiota diversity and subsequent transmission to infants and positively affect metabolic and immune function in both mothers and infants.
Participants were enrolled on a rolling basis between January 12, 2022 to August 22, 2023. Participants were assigned a health educator following baseline sample collection and surveys [10,41–45]. The intervention involved one-on-one remote sessions every two weeks, providing guidance on diet targets and tailored strategies for incorporating high-fiber and/or fermented foods into daily intake. Participants were encouraged to attend virtual group sessions every three months. See Table 2 for Health Education topics.
Participants were recruited using several methods, Study Pages (Facebook or Instagram advertisement), Stanford Nutrition Studies Research Group email notification, physician referral, flyer in doctor’s office or clinic, flyer in Stanford Health Care Tri-Valley, flyer in community, Stanford WELL registry, Spotify, Baby Center website/blog, and Perinatal Diagnostic Center (PDC) call.
We screened individual pregnant mothers through a self-reported questionnaire. Recruitment occurred during the first trimester up to 20 weeks of pregnancy, with baseline samples collected between 12 and 22 weeks. See Table 1 for inclusion and exclusion criteria.
Women were randomly assigned to one of four (1) high-fiber, (2) high-fermented foods, (3) both, or (4) as comparator, no intervention—the usual care diet for pregnancy and postpartum periods.
The study was conducted as an open-label, single-blinded trial. The participants as well as the health education team were aware of which diet the pregnant women were assigned to.
To mitigate the burden and costs associated with clinic visits, participants received a total of 25 for the first visit, 50 for the third visit (which marks the last visit of pregnancy), and $75 for the fourth visit at 5–6 months postpartum, when participants were accompanied by their infants. These gift cards served as non-coercive incentives, compensating participants for their time and effort in commuting for clinic visits and accommodating the study requirements within their schedules.
At the start of the protocol, health educators provided an introduction of diet targets for each of the three intervention 1) increased fiber intake by at least 20 g per day from baseline levels (high-fiber), increased fermented foods (high fermented food), or the addition of both (high fiber and fermented foods). Health educators provided guidance about the types of foods in each category as well as tailored strategies for an initial ramp-up to the daily targets over the first four weeks. An increase in fermented foods was defined and communicated to participants as including at least 2 servings per day during the first week, and, as tolerated, building to reach 6 servings per day. Health educators stressed the importance of sustainability and finding ways to continue to include these fermented foods consistently at an achievable level for each participant. Detailed guidance was provided to encourage pregnant women to incorporate into their diet a variety of fiber sources (legumes, seeds, whole grains, nuts, vegetables, and fruits) and/or fermented foods (fermented dairy products, fermented vegetables, fermented non-alcoholic drinks). Participants were advised to choose foods high in fiber vs foods supplemented with isolated fibers or fiber supplements. Similarly, participants were advised to consume fermented food rather than probiotic supplements.
The intervention involved one-on-one remote sessions every two weeks for the three dietary change arms that included an orientation to the randomized diet and continued dietary guidance, followed by supportive sessions until the participant’s due date. Women were instructed to use a food tracking app (Cronometer) to log their food intake for at least several days before each session and to share these logs with the health educator to enable personalized support. Follow-up sessions included education to help build participants’ confidence and skills and address any barriers to maintaining the study diet. Some specific topics shopping for and preparing high-fiber and or fermented foods, meal planning, developing healthy habits, and sustaining change. Participants in the comparator diet arm met with a health educator three times during pregnancy to discuss healthy eating.
In addition to individual sessions, all participants were invited to informal, remote group sessions every three months. These sessions provided a space for participants to meet others in the same arm of the study, ask questions, and share tips and experiences. Secure Google groups were established for each arm to further encourage participant interaction, with health educators sharing tips, recipes, and other resources on a bimonthly basis.
During the postpartum phase of the study, women were no longer required to follow the prescribed diet, however, they were encouraged to keep following their dietary assignments to whatever extent they felt comfortable. Women included in the three intervention arms were minimally contacted before each data collection period by email, phone call, or, if requested, remote session. Outreach emails encouraged women to continue incorporating foods from the study diet and provided infants’ milestone education such as introducing solid foods, introducing allergens to offspring, and/or practical self-care tips.
Participants in the comparator arm were contacted every 6 months during the postpartum period to answer general nutrition questions. Group meetings were offered every three months for all participants, including the comparator arm, in which nutrition education was presented and questions could be answered. All participants were sent health questionnaires at 3, 6, 9, 12, and 18 months.
The assessment protocol included collecting blood, stool, vaginal swabs, and breast milk (Fig. 1).
Women participated in 24-h dietary recalls over the phone administered by a diet assessor using Nutrition Data System for Research (NDSR) for data analysis. Participants were able to refer to their logged data in Cronometer during dietary recall weeks to aid in reporting their intake to a diet assessor. Three dietary recalls were taken per time point at baseline (12–22 weeks of pregnancy), 26–30 weeks of pregnancy, and 34–38 weeks of pregnancy. Using collected NDSR data, servings of fermented food and fiber will be calculated for each participant across the intervention to assess adherence to assigned dietary intervention. NDSR provides grams of fiber per day which can be used to generate both baseline data and adherence over the intervention. Our team tracked fermented food intake using an approved list of fermented foods, provided to participants and used during patient education, to track servings of fermented foods at baseline and across the intervention (Supplementary Table 1). These foods were noted and could be cross referenced with amount consumed as tracked in NDSR to calculate servings of fermented foods per day. We recommended participants aim to adhere on a continuum, thus participants were coached to consume as many servings as they could of fermented foods per day with 6 servings given as an upper goal.
Blood samples were collected at Stanford Clinical Labs, Stanford Clinical and Translational Research Unit (CTRU), Quest Diagnostics, and at participants’ homes using a micro-sampling kit from the Tasso-M20 device. The original study plan was to include only Bay Area participants able to travel the CTRU. However, due to slower than anticipated recruitment, the catchment area was expanded to all of California, and required adding the option of having blood samples taken and analyzed through Quest Diagnostics.
Anthropometric measurements such as height and weight were measured at CTRU visits for all CTRU participants. Weight and height were not measured for participants who went to Quest Diagnostics.
Stool and vaginal swabs were collected at home using provided kits (DNA Genoteck) with instructions for storage and mailing back to Stanford. Breast milk was collected from patients using CTRU as their clinic location, but not from those using the Quest facilities.
All questionnaires were conducted through RedCap, a secure online data capture system that is used by Stanford to collect and manage data. See Table 3 for questionnaire details.
All clinical blood samples collected at CTRU were sent to the Core Lab at Washington University, St. Louis, for laboratory analysis conducted via standard clinical protocols. All clinical blood samples collected at Quest Labs had laboratory analysis conducted on-site at Quest Labs via standard clinical protocols.
The UC San Diego Microbiome Core performed nucleic acid extractions utilizing previously published protocols [46]. Briefly, samples were purified using the MagMAX Microbiome Ultra Nucleic Acid Isolation Kit (Thermo Fisher Scientific, USA) and automated on KingFisher Flex robots (Thermo Fisher Scientific, USA). Blank controls and mock communities (Zymo Research Corporation, USA) were included and carried through all downstream processing steps. DNA was quantified using a PicoGreen fluorescence assay ((Thermo Fisher Scientific, USA) and metagenomic libraries were prepared with the KAPA Hyper-Plus kit (Roche Diagnostics, USA) and automated on EpMotion automated liquid handlers (Eppendorf, Germany). Sequencing was performed on the Illumina NovaSeq 6000 sequencing platform with paired-end 150bpcycles at the Institute for Genomic Medicine (IGM), UC San Diego.
16S rRNA gene amplification was performed according to the Earth Microbiome Project protocol [47]. Briefly, Illumina primers with unique forward primer barcodes were used to amplify the V4 region of the 16S rRNA gene (515fB-806r) [48]. Amplification was performed with single reactions per sample [49], and equal volumes of each amplicon were pooled for sequencing. 16S libraries were sequenced at the UC San Diego Institute for Genomic Medicine on the Illumina MiSeq sequencing platform with paired-end 250bpcycles.
Blood samples were collected using Tasso devices sent to participants, which were then mailed back to Stanford University. Samples were analyzed at The Human Immune Monitoring Center (HIMC) at Stanford University and quantified using the NULISA platform for immune protein analysis (Inflammation 250-plex).
Microbiome data was analyzed using Quantitative Insights into Microbial Ecology (QIIME2). Amplicon sequencing reads were denoised and quality filtered using the program DADA2. The resultant sequence variants, equivalent to operational taxonomic units, were aligned and masked using MAFFT and the phylogenetic tree of the amplicon sequence variants was created using FastTree. Taxonomy was assigned using a Bayesian classifier – Scikit-learn based on a pretrained Silva database (version 138) curated to the exact 16S amplicon region.
Metagenomics sequencing reads were trimmed and deduplicated using FastP and the processed reads assembled using metaSpades and binned using MetaBat2. The quality of the binned genomes were analyzed using checkM. The assembled genomes - MAGs were dereplicated using dRep. We characterized the gut microbial diversity, composition, and functions and performed statistical and machine learning approaches to define associations between features of the microbiome, study arm, primary outcomes, secondary outcomes, and other metadata. Taxonomic assignments to the MAGs were performed using GTDB-tk and the gene-level functional profiles constructed using inStrain; genes for each genome were assigned using Prodigal and functional annotation against (1) Kyoto Encyclopedia of Genes and Genomes (KEGG) Orthologies (KOs) using kofam_scan, (2) carbohydrate-active enzymes (CAZymes) using dbCAN2, and (3) Comprehensive Antibiotic Resistance Database (CARD) Anti-Microbial Resistance (AMR) database using Diamond with the command “diamond blastp.
Our primary outcome for the study, registered at clinicaltrials.gov, was the total number of species detected in the infant stool as a measure of infant microbiota diversity. Specifically, we compared the difference in total number of species (Amplicon Sequence Variants [ASVs]) detected in stool between the high-fiber group, high-fermented foods group, high fiber plus fermented foods group, and comparator group (hereafter referred to as the four arms) at one month postpartum.
Six secondary outcomes were registered at clinicalgrials.gov for our 1) change in differences in total number of species (ASVs) detected in maternal stool from baseline to 36 weeks of pregnancy among the four arms; 2) change from baseline to 36 weeks of pregnancy in the differences in 19 individual inflammatory markers (β-NGF, LIF-R, IL-12B, IL10, CASP-8, LAP TGF-β-1, CD6, CD5, MCP-2, IL6, CCL20, IL18, VEGFA, MMP-10, MCP-4, CCL4, CXCL10, CCL19, FGF-21) detected in blood samples among the four arms; 3) mother’s differences in the amount of 19 individual inflammatory markers (β-NGF, LIF-R, IL-12B, IL10, CASP-8, LAP TGF-β-1, CD6, CD5, MCP-2, IL6, CCL20, IL18, VEGFA, MMP-10, MCP-4, CCL4, CXCL10, CCL19, FGF-21) at 6 months postpartum detected in heel blood among the four arms; 4) differences in infant weight-for-length growth chart percentiles among the four arms at 18 months postpartum; 5) differences in the proportion of women who gained weight within the pregnancy weight gain recommendations among the four arms measured at 36 weeks; and 6) change from baseline to 36 weeks of pregnancy in mothers’ maternal systolic blood pressure, diastolic blood pressure, LDL cholesterol, HDL cholesterol, triglycerides, glucose, and fasting insulin among the four arms. We analyzed all clinical data using linear mixed models programmed using R software.
Alpha diversity was measured using Observed Otus, Shannon and Simpson indices were calculated for all samples, with a rarefaction upper limit of median depth/sample count and the alpha diversity between different treatments was compared using Wilcoxon rank sum test. Samples were removed from further characterization if they did not contain sufficient reads. Beta-diversity was calculated, and ordination plots were generated using Bray–Curtis and Jaccard indices (non-phylogenetic) and weighted and unweighted UniFrac (phylogenetic) on amplicon sequence variant data leveled, according to the lowest sample depth. Permutation multivariate analysis of variance (PERMANOVA) testing was used for statistical analysis of beta-diversity measurements. Comparative and statistical analysis were conducted using R with packages such as phyloseq, tidyverse and ggplot2.
Regarding missing data, we used all data available in the study analysis, including for participants who dropped out of the study, up to their latest time point.
Comparative and statistical analysis were conducted using R software.
The results of this study will be published in multiple publications as data becomes available.
Here, we have described the methodology, study design, and analysis plan of our trial, FeFiFo-MOMS. Our intervention trial aimed to elucidate the effects of increased consumption of high-fiber and fermented foods during pregnancy on the diversity and composition of maternal gut microbiota, to enhance both maternal and neonatal health outcomes. The study was structured as a randomized controlled trial with dietary guidance and support using a combination of direct dietary assessments and biomarker analyses to monitor the effects of these dietary changes, including blood, stool, vaginal swabs, and breast milk. Once complete, our findings could significantly impact public health strategies, particularly in the prenatal-care sector. Further, our findings will help us better understand how dietary modulation during pregnancy impacts not only maternal gut microbiota and inflammatory status, but also the gut microbiome development and immune system of the offspring. Overall, the anticipated data from this trial will help us better understand the dietary influences on microbiota and their broader implications for the health of both pregnant mothers and their offspring.