Foodomics Analysis of a Mediterranean Diet Reveals Food-Specific Compounds That Are Detected in Human Plasma

Objectives

Foods are comprised of thousands of compounds that may be absorbed upon ingestion. Metabolomics offers the unprecedented capability of measuring these small molecules to discover biomarkers of intake and examine how dietary patterns may exert their effects on health. The objectives of this study were to: 1) perform metabolomics analyses of 100 foods provided as part of a controlled feeding Mediterranean-style eating pattern (MED) intervention to identify food-specific compounds (FSC); and 2) determine which FSC were observed in participants’ plasma samples.

Methods

Individuals (n = 41) completed a randomized, crossover, controlled feeding study with two five-week MED interventions separated by a four-week washout. Following extraction into lipid-rich and hydrophilic fractions, samples of 100 foods were analyzed in triplicate using liquid chromatography-mass spectrometry (LCMS). Participant plasma was collected pre- and post- each intervention period and similarly analyzed by LCMS. Food compounds were determined to be FSC if they were present in all replicates of a food but no other foods. Plasma data were assessed to determine presence of FSC. Analyses were completed using Mass Profiler Professional (MPP), and descriptive statistics were generated.

Results

A total of 1390 ± 1044 (mean ± SD, range 14–4908) compounds per food were detected. Ninety-nine foods contained FSC (mean 296 ± 393, range 1–2472) with 17% of all compounds unique to a given food. Feta cheese had the greatest number of FSC; pure maple syrup had the lowest. Ninety-six foods had at least one FSC detected in plasma (mean 28 ± 32, range 1–143) with 16% of all FSC found in plasma. Grape juice, apple juice, romaine lettuce, and peanuts each had one FSC in plasma, while salmon and beef had 143 and 76 FSC in plasma, respectively.

Conclusions

Metabolomics analyses of both foods and plasma confirmed that foods have compounds that are unique and that these unmetabolized compounds can be detected in plasma following consumption. This information can be used to identify food intake biomarkers and link specific components of eating patterns to health outcomes. Additional studies to examine reproducibility and determine dose-response will be vital to elucidating the relationship between food compounds within MED and health.

Funding Sources

NIH/NIDDK; Beef Checkoff (MED feeding study).

Unique-to-Salmon Compounds Increase in Plasma and Are Associated With Cardiovascular Health Following a Mediterranean Diet Intervention

Objectives

Salmon consumption may promote cardiovascular (CV) health through nutrients such as omega-3 fatty acids (n-3 FAs) or other yet unidentified compounds. Metabolomics allows for the comprehensive analysis of small molecules in foods and human samples. The objectives of this study were to: 1) identify compounds present in salmon that were increased in plasma after consumption of a salmon-containing Mediterranean-style eating pattern (MED); and 2) evaluate associations between salmon compounds detected in plasma and CV biomarkers.

Methods

Individuals (n = 41) completed a 16-week randomized, crossover, controlled feeding study of a MED containing 2 servings salmon/week. Assessments were completed pre- and post-intervention, where CV biomarkers (e.g., blood pressure, lipid profile) were measured and fasting plasma collected. Samples of participants’ plasma, salmon, and 99 other foods in the trial were analyzed in triplicate via liquid chromatography-mass spectrometry (LCMS). Compounds were determined to be unique-to-salmon (UTS) if detected in all replicates of salmon but none of the other 99 foods. For UTS compounds that were found in plasma, linear mixed effects models were used to assess change from pre- to post-intervention and the association between this change and changes in CV biomarkers. Significance was assessed after adjustment for false discovery rate.

Results

Relative to the other 99 foods, 508 compounds were determined to be UTS, and 143 of these UTS compounds were detected in plasma. Forty-eight UTS compounds significantly increased in plasma from pre- to post-intervention. Two compounds, a phosphatidylglycerolphosphate (40:4) and a triglyceride (58:11), were associated with improvement in ≥1 CV biomarker. The former, a cardiolipin precursor, was associated with improvements in total cholesterol, LDL cholesterol, and Apolipoprotein B.

Conclusions

Salmon compounds are detectable in human plasma after consumption within a complex diet and increase after intervention, suggesting potential as intake biomarkers. Further, bioactive compounds beyond nutrients such as n-3 FAs may be associated with CV benefits. Future studies should be conducted to quantify concentrations and define dose-response.

Funding Sources

NIH/NIDDK; Beef Checkoff (MED feeding study).

Role of a Preconception Maternal Nutrition Supplement and Pre-pregnancy BMI on Amnion DNA Methylation at Birth in Guatemalan Mother-Infant Dyads: The Women First Trial

Objectives

Maternal nutrition can alter the offspring epigenome at birth. We sought to examine epigenome-wide DNA methylation (DNAme) from a subset of Guatemalan mother-infant dyads from the Women First Preconception Maternal Nutrition Trial (WF). Women were randomized to either: Arm 1) women consumed a daily maternal nutrition supplement (MNS) ≥ 3 months prior to conception until delivery; Arm 2) women consumed the same MNS starting at 12 weeks gestation until delivery; or Arm 3) no MNS. We tested if infant DNAme from amnion tissue at birth (N = 99) was associated with: 1) timing of exposure to maternal MNS; 2) pre-pregnancy body mass index (ppBMI); and 3) the interaction of maternal MNS and ppBMI.

Methods

Bisulfite-converted DNAme libraries were constructed using Roche NimbleGen SeqCap Epi CpGiant probes and were sequenced via 2 × 150 paired end reads. We assessed the relationship between Arm, ppBMI, and Arm x ppBMI interaction on CpG methylation. All statistical models adjusted for multiple testing using false discovery rate (FDR) and controlled for maternal age, infant sex, exposure to smoke, infant genetics, and cellular heterogeneity. Gene set enrichment analyses were performed via Enrichr.

Results

We identified 480 CpGs associated with Arm, 4 CpGs associated with ppBMI, and 22 CpGs associated with the interaction of Arm x ppBMI (FDR < 0.05). Further, we found that DNAme was changed between Arms (1 vs 2, 1 vs 3). There were 300 CpGs that were different between Arms 1 and 2 and 159 CpGs that were different between Arms 1 and 3 that annotated to genes and passed FDR < 0.05. These results suggest preconception consumption of maternal MNS elicits different epigenetic responses as compared to MNS commencing during gestation or not at all. In addition, CpGs that annotated to genes were enriched in pathways associated with growth, development, and metabolism that included circadian rhythm, TCA cycle, Wnt signaling, and melatonin metabolism.

Conclusions

Our findings indicate that maternal MNS was robustly associated with amnion DNAme at birth. More specifically, preconception MNS resulted in DNAme changes that differed from the other Arms in biologically relevant pathways suggesting timing of maternal nutrition impacts the fetal epigenome. Future studies will examine DNAme associated with birth outcomes.

Funding Sources

Bill & Melinda Gates Foundation and NIH NICHD/ODS.

Family Care Indices and Linear Growth Predict INTER-NDA Scores for Child Development at Age 2 Years: Findings From the “Women First” Trial

Objectives

Nutrition during fetal and early postnatal life impacts brain development, however evidence from randomized trials from low-and-middle-income countries (LMIC) is limited. We evaluated effects of maternal nutrition supplementation before and during pregnancy on neurodevelopment and vision in children from 4 LMIC with high rates of stunting, using a simplified assessment tool.

Methods

Women First was an individually randomized trial comparing the impact of maternal nutrition supplementation initiated preconception vs at ∼12 wk gestation vs no supplement; interventions were discontinued at delivery in trial sites in Democratic Republic of the Congo, Guatemala, India, and Pakistan. Neurodevelopment and visual acuity and visual contrast sensitivity were assessed at 24 mo using the INTER-NDA and the Cardiff tests, respectively, in a random sub-set, representing ∼1/3 of infants with valid birth measurements from the WF trial. Anthropometry and Family Care Indicators (FCI) were also obtained at 24 mo.

Results

667 and 634 children (91% and 86% of sub-set) were included in INTER-NDA and vision analyses. Arm-proportionate contribution was 32.2–32.8%. Overall, we observed the following percentage of children with delays: 66.6% cognitive; 87.4% fine motor; 91.3% gross motor; and 11.7% language; 25.9% had positive behavior problems; and 26.0% and 21.0% had low acuity and contrast sensitivity scores, respectively. WF intervention arm was not significant for INTER-NDA scores, rates of mild-to-moderate or severe delay (F = 0.004–1.518, p = 0.28–0.99; X² = 3.81–5.92, p = 0.18–0.42; X² = 3.81–6.33, p = 0.18–0.43); or for low acuity and contrast sensitivity (X² = 0.91, p = 0.63 and X² = 1.64, p = 0.44). LAZ_24mo was significantly associated with cognitive, motor, language, and behavior scores; maternal education predicted cognitive, language, and vision; and FCI was significantly associated with fine motor, language, and behavior after adjusting for other covariates (R² 0.33 cognitive; 0.40 language; 0.12 motor; 0.05 positive and 0.11 negative behavior; and 0.31–0.34 for vision models).

Conclusions

The findings highlight the association between child development, linear growth, and family environment.

Funding Sources

Thrasher Research Fund, Bill & Melinda Gates Foundation, NICHD, ODS.

Associations Between Ambient Temperature and Circulating Maternal One-carbon and Amino Acid Metabolites in Pregnancy

Objectives

Excessive ambient temperature closely linked to climate change is an imminent threat to the health of mothers and children globally. Recent studies have shown detrimental associations between high temperatures during pregnancy and birth outcomes. The goal of this study was to examine relationships between circulating metabolites in the mother and environmental temperature in a resource-limited setting.

Methods

The study was a secondary analysis employing data from the Women First (WF) trial conduced in Thatta, Pakistan. As part of the trial, women of child-bearing age either consumed a comprehensive maternal nutritional supplement (MNS) prior to conception until delivery (Arm 1), from 12 wk gestation until delivery (Arm 2); or were not supplemented with MNS Arm 3. Daily maximum air temperatures were acquired from the closest surface observation systems (n = 131). Average daily maximum temperatures (Tmax) for a 90-day window representing the 1st trimester of pregnancy was calculated for each participant. The number of days in this period when the Tmax was &gt; 39°C was computed to assess ‘heat stress days’. Maternal blood was collected on dried blood spot (DBS) cards at 34 wk gestation. A targeted metabolomics assay measuring concentrations of 23 metabolites was performed on DBS samples using LC-MS/MS.

Results

Linear regression was employed to assess the associations between ambient temperature in the 1st trimester and metabolite levels (z-scores). Models were adjusted for MNS supplementation. Significant associations between Tmax in the 1st trimester and 7 metabolites were observed (FDR p-values &lt; 0.05). Of these, maternal choline concentrations were negatively correlated with maximal temperature in the 1st trimester (β = −0.063, p = 4.3e-10). Concentrations of glutamine, histidine, arginine, symmetrical-dimethylarginine, methionine and cysteine were positively associated (p &lt; 0.05) with daily maximal temperature.

Conclusions

Collectively, the findings indicate that exposure to excessive heat stress in the first trimester is associated with lower choline levels and other metabolites important in blood pressure regulation. While the findings are associative, our results point to underlying mechanistic alterations in maternal circulation in concert with environmental heat stress.

Funding Sources

Bill & Melinda Gates Foundation; NIH/NICHD/ODS.

Comparison of Methods for Estimating Discretionary Salt Intake in Field Settings

Objectives

Accurate and precise estimates of discretionary salt (DS) intake are critical for designing salt fortification interventions and counseling on salt intake reduction. This study compared four methods of estimating DS intake among non-pregnant women of reproductive age (NPWRA) in northern India to inform the design of a trial of multiply-fortified salt.

Methods

Participants were NPWRA (18–49 y) in Punjab, India. Weighed food records (WFR), same-day duplicate diet (DD) composites, and samples of household (HH) salt were collected simultaneously from 100 women and repeated on a subset of 40. Sodium (Na) and iodine contents of the DD composites were analyzed using Inductively Coupled Plasma (ICP)-Optical Emission Spectrometry and ICP-Mass Spectrometry. HH salt samples were also analyzed for iodine. Methods for estimating DS intake included: 1) WFR: DS consumed from recipes or added at time of consumption were weighed; 2) HH salt disappearance (HHSD): total DS used by HH on the observation day divided by number of HH members; 3) Sodium estimation (NaE): Na content of 40 replicate DD composites prepared without DS were subtracted from the Na content of the corresponding original DD and difference multiplied by the molar mass of NaCl; 4) Iodine method (IM): analyzed iodine content of milk and milk products and commercial snacks were subtracted from DD iodine content, and difference divided by the iodine content of the HH's salt sample. The relations between methods were explored using Pearson correlation and Bland Altman analyses.

Results

Mean ± SD intake of DS according to the WFR, HHSD, and NaE methods were 4.7 ± 1.8 g/d, 5.8 ± 3.3 g/d, and 4.1 ± 2.1 g/d, respectively. Results of IM are pending. Pearson correlation coefficients for DS intake estimates obtained from WFR vs. NaE and WFR vs. HHSD were 0.82 (p < 0.001) and 0.48 (p < 0.001), respectively. Mean ± SD bias (limits of agreement) were 0.68 ± 1.25 g/d (−1.77, 3.13) for WFR vs. NaE, and 1.8 ± 2.93 g/d (−4.56, 6.92) for HHSD vs. WFR methods.

Conclusions

Discretionary salt intake from WFR and NaE showed good agreement and are feasible to implement in field settings.

Funding Sources

Bill & Melinda Gates Foundation.

Comparison of Toddler Crown Rump Length and Leg Length in Four Low- and Middle-Income Research Sites: The Women First trial

Objectives

Length at 2y of age is a strong indicator of adult height and potential for other adverse effects in adulthood; leg length (LL) is suggested to be the more nutritionally responsive component of height. These analyses examine body proportions- measured as linear crown rump length (CRL), LL, and CRL: LL- among 24 mo old children from four low- and middle-income settings with high stunting rates (average across 4 sites = 65%) participating in the Women First Preconception Maternal Nutrition Trial (WF).

Methods

The WF study is a nutrition intervention trial of maternal-child pairs in four diverse research sites (Democratic Republic of the Congo [DRC], Guatemala, India, and Pakistan). At 24 mo of age, toddler length and CRL were obtained and subsequently used to calculate LL (length-CRL). ANOVAs with post hoc Tukey HSD for each parameter were performed in JMP Pro 16.0.0.

Results

No statistical differences were seen by maternal intervention arm, thus anthropometric data from 2,157 children (DRC n = 488, Guatemala n = 540, India n = 531, Pakistan n = 598) were explored by site. While mean (±SD) CRL did not significantly differ between Guatemala and India (48.9 ± 2.18, 48.9 ± 2.30cm), Pakistan and DRC (47.7 ± 2.52, 45.8 ± 2.22cm) had significantly lower mean CRL (p < 0.001), with DRC having the lowest value. Mean LL did not differ significantly between Guatemala and Pakistan (30.6 ± 1.74, 30.5 ± 2.21), nor between DRC and India (32.6 ± 2.28, 32.3 ± 2.01cm); however, DRC and India had significantly longer mean LL vs Guatemala and Pakistan (p < 0.0001). Statistically significant differences in mean CRL: LL (p < 0.0001) were observed among all four sites, with Guatemala exhibiting the largest mean (1.60 ± 0.09), attributable to shorter LL.

Conclusions

CRL and LL are indicative of proportionality, a measure of impaired growth. These data highlight that among the four research sites in the WF trial, Guatemala, which exhibits the highest rate of maternal stunting, has the least favorable body proportions (largest CRL: LL). The difference in body proportions between sites despite nutritional intervention warrants further examination of genomic and environmental (including nutritional) factors that may have an effect on anthropometrics such as LL, CRL, and CRL: LL.

Funding Sources

Bill & Melinda Gates Foundation, NIH NICHD/ODS.

The Association Between Maternal Placenta Growth Factor Levels and Small-for-Gestational Age Infants: Findings From the Multi-Country Women First Study

Objectives

Examine the association between Placenta Growth Factor (PlGF) levels during pregnancy, maternal nutrition supplementation (MNS) and the prevalence of small-for-gestational age (SGA) infants in the Women First (WF) Study.

Methods

WF is a RCT of MNS consisting of a lipid-based micronutrient supplement ± a protein-energy supplement (for low maternal BMI or weight gain), provided to women daily until delivery starting either ≥ 3 months pre-conception (Arm 1), at the end of the first trimester (Arm 2) or not at all (control group, Arm 3). Serum samples were obtained at 12 (Arms 1 and 2) and 34 weeks (Arms 1, 2 and 3) gestation from a subset of women in three of the WF sites (Guatemala [Guat] (n = 257); India [Ind] (n = 171, Arms 1 & 2 only); and Pakistan [Pak](n = 279)). PlGF was measured using an ELISA. An ultrasound late in the 1^st trimester determined whether infants were SGA at birth (weight-for-age centile < 10, INTERGROWTH-21^st standards). PlGF levels were examined for arm, site and time point differences using ANOVA and post-hoc Tukey testing. The association between SGA and PlGF was analyzed by logistic regression adjusting for maternal height, age and education (STATA software v. 17.0).

Results

In participants with PlGF measurements, the prevalence of SGA infants is 23% in Guat, 44% in Ind and 34% in Pak. PlGF levels increased from 12 to 34 weeks gestation in Arms 1 and 2 in every site (Tukey's adjusted P < 0.0001). There are no differences in PlGF among sites at 12 weeks, whereas all sites are different at 34 weeks (Tukey's adjusted P < 0.0001). Regarding the association between SGA and PlGF (adjusting for maternal height, age and education), there are no differences by arm within each time point and site (arms were combined in subsequent analyses). PlGF is associated with SGA only in Guat at 34 weeks: odds of an infant being SGA decreases as PlGF increases (OR = 0.9989, 95% CI: 0.9981, 0.9998, unadjusted p = 0.015). For a 50 pg/mL increase in PlGF, the odds of the infant being SGA decreased by 0.948 (5.2%; OR 95% CI: 0.909, 0.990).

Conclusions

An association between PlGF and prevalence of SGA was found in Guatemala at 34 weeks: as PlGF increases, the odds of a SGA infant decreases. The MNS did not result in differences among arms.

Funding Sources

Bill & Melinda Gates Foundation, Eunice Kennedy Shriver NICHD and NIH ODS.

A Multi-country Association Analysis of Maternal Selenium (Se) Levels and Infant Birth Outcomes: Findings From the Women First Study

Objectives

Selenium (Se) soil content varies worldwide impacting population levels. Maternal Se deficiency may be related to preterm birth, small for gestational age (SGA), and low birth weight (LBW, < 2,500g) possibly through its roles in antioxidant and thyroid hormone function particularly during the first trimester. We examined associations of Se levels and birth outcomes in a multi-country nutritional intervention trial of pregnant women.

Methods

This is a secondary analysis of the Women First trial (NCT01883193), a randomized controlled study investigating timing of maternal multiple micronutrient (with 130 mcg Se) and protein-energy supplementation (MNS) on fetal growth. MNS was started 3 months prior to conception (Arm 1), at ∼12 wk of gestation (Arm 2), or not at all (Arm 3) in women in India, Guatemala, and Pakistan. Gestational age was determined by first trimester ultrasound. At 12 and 34 wk gestation, maternal serum Se was measured by ICP-MS (n = 143–325 per timepoint per site). TSH was measured by 1-step sandwich (Guatemala and Pakistan) or chemiluminescence immunoassay (India). We used t-test or ANOVA to test group differences and linear models to test association of Se with infant outcomes and TSH, adjusting for treatment arm and study site.

Results

Se levels were highest in Guatemalan women (mean ± SD, 12 wk: 102.1 ug/L ± 12.6; 34 wk: 103.9 ± 14.2) and lowest in Pakistani women (12 wk: 85.7 ± 14.5; 34 wk: 77.9 ± 14.8) at both timepoints (P < 0.0001). MNS supplementation before or during gestation increased serum Se levels at 34 wk in Guatemalan women only (p = 0.038). In Pakistan a negative association between maternal Se at 34 wk and LBW was seen in female (n = 153; p = 0.041) but not male (n = 143; p = 0.97) infants. Se levels were not associated with preterm birth or SGA status at any site. TSH levels did not differ by study arm but were lower in Pakistan at 12 and 34 wk compared to other sites. There was no association between TSH and Se levels at any timepoint at any site.

Conclusions

Maternal Se levels may impact fetal birth weight in vulnerable populations in a sexually dimorphic manner. Preconception Se supplementation as part of a comprehensive MNS did not increase Se levels in Pakistan where soil Se content is low. Although Se did not correlate with TSH levels in this study, future work will incorporate co-influence of iodine.

Funding Sources

Bill & Melinda Gates Foundation, NIH NICHD/ODS.

Neurodevelopment Scores at 24 Months Are Associated With Maternal Education, Home Environment, and Linear Growth in Offspring of the Women First Trial

Objectives

Healthy physical growth has been associated with better performance on neurodevelopmental testing in young children in low resource settings. We examined the relationship of growth indicators from 6 to 24 months of age to the sub-scales of the Bayley Scales of Infant Development-III (BSID-III) at 24 months in offspring of participants in the four-country preconception maternal nutrition trial, Women First (WF), which found the maternal intervention to benefit birth length, which was predictive of length and risk of stunting at 24 months.

Methods

With no post-delivery intervention, follow-up visits for anthropometry were obtained at 6-month intervals; BSID-III and Family Care Indicators (FCI) were completed at 24 months in a random sub-set of the WF offspring, representing 2/3 of infants with valid birth measurements. Multiple covariates (intervention arm, site, cluster within site, maternal education, age, SES, FCI subscales, LBW, and change in anthropometry Z-scores from 6 to 24 months, e.g., length-for-age, DLAZ6-24) were considered as confounders or predictors and were adjusted in a general linear model to predict adjusted mean differences (AMD) in BSID scores.

Results

1,386 infants (93% of those randomized to sub-set) were included in the analysis (n = 441,486, 459 for Arms 1, 2, 3, respectively). Four covariates were positively associated (p ≤ 0.01) with all 3 of the BSID subscales cognitive(C), motor(M), and social-emotional(SE) scales, respectively: maternal secondary education (+AMD 3.60(C), 3.35(M), 2.80(SE)); DLAZ6-24 (+AMD 1.98(C), 3.15(M), 2.06(SE)); BW &gt; 2500 g (+AMD 1.64(C), 2.20(M), 2.31(SE)); and FCI play materials (+AMD 1.46(C), 1.36(M), 1.72(SE)). No other anthropometric or maternal variables were consistently associated with BSID subscales.

Conclusions

The findings underscore the multiple critical components of nurturing care for early child development, including the caregiver, adequate birth weight, healthy linear growth, and opportunities for learning. Results suggest that the positive effects of maternal nutrition supplementation in the WF trial on birth length and postnatal linear growth may, in turn, have enduring downstream benefits for their children's neurodevelopment.

Funding Sources

The Bill & Melinda Gates Foundation, NICHD, ODS.

Infant Young Child Feeding Practices From 12 to 24 Months of Age of Offspring From the Women First Trial

Objectives

Evaluate infant young child feeding (IYCF) trends from 12 to 24 mo in four low middle-income countries: Democratic Republic of Congo (DRC); Guatemala; India; and Pakistan.

Methods

2413 children (570 DRC, 614 Guatemala, 589 India, and 640 Pakistan) born to women from the Women First Trial enrolled. Trained research health workers visited households and assessed children's feeding practices from 12 to 24 mo. Child feeding indicators including the prevalence and longitudinal trends of meeting minimum diet diversity (MDD), minimum meal frequency (MMF), and minimum acceptable diet (MAD) were analyzed using the 2018 IYCF definitions from 12 to 24 mo for all four sites individually and combined.

Results

MDD significantly increased from 16% at 12 mo to 29% at 24 mo for all sites combined (p &lt; 0.001); by site: 3 to 10 for DRC (p &lt; 0.001), 33 to 42 for Guatemala (p = 0.001), 28 to 63 for India (p &lt; 0.001), and 1 to 2 for Pakistan (not significant). MMF significantly increased from 49% at 12 mo to 68% at 24 mo for all sites (p &lt; 0.001); by site: 22 to 49 for DRC (p &lt; 0.001), 89 to 92 for Guatemala (not significant), 66 to 73 for India (0.012), and 18 to 58 for Pakistan (p &lt; 0.001). MAD significantly increased (p &lt; 0.001) for all sites, and by site for DRC and India only. Food groups least consumed at 24 mo for all sites were nuts/pulses (prevalence 20%), meats/fish/insects (21%), and vitamin A rich fruits & vegetables (38%); compared to grains/roots/tubers which were consumed by 98% of participants. For all sites, milk & dairy and eggs consumption significantly increased from 50 to 60% (p &lt; 0.001) and 26 to 50% (p &lt; 0.001), respectively. Breastfeeding decreased from 94% at 12 mo to 49% at 24 mo for all sites with the lowest prevalence of breastfeeding in Pakistan (23%), whereas the other sites were ∼50%.

Conclusions

For all sites, the prevalence of participants meeting IYCF recommendations for feeding indicators increased from 12 to 24 mo, but prevalence at 24 mo remained well below recommendations in all sites. Children with low IYCF scores are at risk for low intakes of vital nutrients, protein and vitamin A rich foods, which may negatively impact periods of rapid growth and development in these resource limited populations.

Funding Sources

Bill & Melinda Gates Foundation and Eunice Kennedy Shriver NICHD.

Heat Stress-Associated Growth Retardation in the First 1000 Days Is Mitigated by Preconception Nutritional Supplementation

Objectives

Understanding the interactions between nutrition and climate-driven health outcomes is increasingly necessary. Employing data from the Women First trial conducted in Thatta, Pakistan, we examined the interactions between maternal nutritional deficits and heat stress (HS) on birth outcomes.

Methods

Women of child-bearing age either consumed a small quantity lipid-based nutrition supplement (sqLNS) prior to conception (Arm 1), from 12 wk of gestation until delivery (Arm 2); or were not supplemented with sqLNS (Arm 3). At the Pakistan site, samples sizes with birth outcomes obtained within 48 h of delivery for Arms 1, 2 and 3 were 157,150 and 138, respectively. Daily maximal air temperatures were used to were compute the average daily maximal temperatures (Tmax) for each infant over four 90-day windows representing each

trimester of gestation and 3 mo preconception. The number of days in each period when the Tmax was &gt; 39°C was computed to assess ‘heat stress days’.

Results

Linear regression models assessing the associations between heat exposure and outcome variables (gestational age adjusted Z-scores of birth length (LGAZ), birth weight (WGAZ) and head circumference (HCGAZ) revealed a striking negative association of both ambient temperature and HS days in the first trimester on birth length (LGAZ) and head circumference (HCGAZ) (p &lt; 0.005). For each 5°C increase in the Tmax in the first trimester, LGAZ decreased by 0.15 z-scores and HCGAZ decreased by 0.11 z-scores. Heat exposure (Tmax) in the second trimester also associated with HCGAZ (p = 0.02 and z-score of −0.11 for every 5°C) but not with birth length or birth weight. RNA-sequencing of placenta indicated a robust association of 1st trimester HS with genes involved in protein targeting to ER, protein biosynthesis, and cytoplasmic translation. Finally, while non-intervention groups (Arms 2 and 3) showed a robust negative association of excessive HS with LGAZ (p &lt; 0.05), preconception intervention with sqLNS (Arm 1) significantly mitigated heat-associated decreases in birth length.

Conclusions

These findings indicate that excessive HS in the first trimester is associated with decreased linear growth in utero. Importantly, sqLNS mitigated growth restriction in newborns implicating nutritional status as a driving force in resilience to HS.

Funding Sources

Supported by The BMGF and NICHD.

Bell Peppers Provide Consistent β-cryptoxanthin Content Independent of Organic Status, Fresh, or Cooked, North American Country of Origin and Season

Objectives

The carotenoid β-cryptoxanthin is a natural pigment that is both an antioxidant and a precursor to retinol. Research supports that β-cryptoxanthin has greater bioavailability than β-carotene in humans. Red bell peppers have more than double the amount of β-cryptoxanthin than any of the top seven consumed vegetables, as ranked by the USDA. To determine if the amounts of β-cryptoxanthin in bell peppers are dependent upon the organic status, color, cooking, season or location that the fruit was grown within North America, β-cryptoxanthin was measured and compared in green, red and yellow bell peppers.

Methods

An assortment of bell peppers were purchased in the greater Denver, CO region. Green, red and yellow peppers; organic and non-organic; and peppers grown in Canada, the US and Mexico during two different seasons were selected for analysis. The effects of lightly sautéing compared to fresh peppers and season of growth were compared. Samples (100 mg/1 mL) were freeze-dried, then prepared by liquid-liquid extraction for untargeted liquid chromatography mass spectrometry (LC-MS)-based metabolomics analysis. An accurate mass and retention time (AMRT) database was used to identify and quantify β-cryptoxanthin. Linear regression was used to assess the relationship between β-cryptoxanthin and pepper qualities.

Results

β-cryptoxanthin concentration was significantly higher in red bell peppers compared to green (11.8-fold) and yellow peppers (7.1-fold) (P = 1.624e-11). β-cryptoxanthin concentration does not appear to be influenced by organic status, season or geographic location. Likewise, the cooked peppers were similar in β-cryptoxanthin content compared to their fresh counterparts.

Conclusions

Our results suggest that the consumption of bell peppers as a source of β-cryptoxanthin is consistent across organic status, fresh, cooked, season and the location in which they were grown. While β-cryptoxanthin concentration in significantly higher in red bell peppers, more research is necessary in order to determine whether these differences result in any altered health outcomes.

Funding Sources

National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK).

Retrospective Assessment of Selenium Status in High Risk Infants

Objectives

Selenium (Se) is an essential trace mineral important in neonatal development that contributes to oxidative stress and the inflammatory response. Se deficiency in preterm infants is associated with late onset sepsis, bronchopulmonary dysplasia, poor neurodevelopmental outcomes, and retinopathy of prematurity. Current Se dosing in many US Neonatal Intensive Care Units is based on the amount of Se in term breastmilk. However, it is unclear if this is sufficient and higher levels may be needed to prevent or treat Se deficiency. Our goal was to evaluate if the current practice supplementing total parenteral nutrition (TPN) with sodium selenite 2 mcg/kg/d is sufficient at preventing Se deficiency in high risk infants.

Methods

This is a retrospective chart review of Se status at a level IV Children's Hospital NICU from January 1, 2017 to August 30, 2019. Infants were included if born from 22–42 weeks gestation and received TPN for &gt;4 weeks. They were excluded if there was concern for active sepsis or bacterial illness at time of Se draw. Normal Se status was defined as 45–90 ng/mL for infants 0–2 months. Birth weight, IUGR status, gestational age, and % enteral feeds were evaluated. Se deficient infants received higher Se dosing at 5–7 mcg/kg/d. Repeat levels were evaluated after 4 weeks. Results are reported as mean ± SD.

Results

Se status was assessed for 39 infants. Average gestational age was 29.8 ± 5.36 weeks. Average birth weight was 1499 ± 837 g. At the time of first Se assessment, 78% of infants were Se deficient, with a mean Se level of 40.95 ± 12 ng/mL. Repeat Se levels on higher dosing was assessed in 23 infants. After &gt;4 weeks of higher Se dosing, 35% of infants remained Se deficient, with a mean Se level of 54.04 ± 14 ng/mL. By t-test, statistically fewer infants were Se deficient on higher Se dosing (P &lt; 0.0003).

Conclusions

Infants on prolonged TPN &gt;4 weeks are at high risk for Se deficiency. Se dosing at 2 mcg/kg/day is insufficient in preventing deficiency for a majority of these babies. Higher Se dosing improved the percentage of Se sufficient infants, but a third remained deficient. Future studies are needed to prospectively determine if higher Se in TPN prevents Se deficiency.

Funding Sources

University of Colorado, Section of Neonatal-Perinatal Medicine.

Astaxanthin Levels Are Higher in Fresh Salmon Compared to Canned and Pouch Varieties

Objectives

Astaxanthin, a predominately marine-source carotenoid, is the subject of a large number of studies for its antioxidant and anti-inflammatory properties. Astaxanthin is not generally a primary carotenoid in human plasma due to relatively low dietary intake. Salmon is the one of the few dietary sources of astaxanthin in typical American diets and the concentration may vary by the source of salmon foods. A study was performed to 1) Compare astaxanthin concentration in various sources of salmon; 2) Compare astaxanthin plasma concentrations before and after salmon consumption.

Methods

An assortment of salmon types and forms was purchased in the greater Denver, CO region: wild Pacific, farmed Atlantic, canned and pouch. Plasma samples were collected from five participants prior to and after a five week Mediterranean diet intervention study, which included two servings of salmon per week. Salmon samples were freeze-dried, then both salmon (in triplicate) and plasma samples were prepared by liquid-liquid extraction for untargeted liquid chromatography-mass spectrometry analysis. An accurate mass and retention time database was used to identify and quantify astaxanthin. ANOVA with Tukey multiple testing corrections was used to assess the relationship between astaxanthin and the different salmon products, and paired t-tests for astaxanthin in plasma.

Results

Astaxanthin concentration was significantly higher in fresh salmon compared to pouch packaged (23.0-fold; P = 1.70e-04) and canned (34.9-fold; P = 1.23e-08). Interestingly, astaxanthin levels were similar between fresh wild Pacific and fresh farmed Atlantic salmon (0.91-fold, P = 0.82) and by mode of cooking (i.e., fresh, cooked, frozen; P = 0.81). Astaxanthin concentration in plasma was significantly increased after farmed Atlantic salmon consumption (1.98-fold, P = 6.16e-09).

Conclusions

Our data suggest that astaxanthin concentration varies among different processed salmon products compared to wild and farmed salmon. After salmon consumption, plasma astaxanthin concentration increased and may have potential as a biomarker of salmon consumption.

Funding Sources

National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK).

Hepatic Postnatal Developmental Expression of Trace Mineral Associated Antioxidant Enzymes

Objectives

Oxidative stress is central to the etiology of many diseases of prematurity. Lower antioxidant defenses render premature infants vulnerable to oxidative damage secondary to infection and oxygen therapy. Antioxidant enzymes (AOE) increase perinatally in the blood and lungs. Many AOE require a micronutrient such as selenium (Se) or zinc (Zn) to function at maximum efficacy. These trace elements are low in neonates compared to adults.

The liver is an important immune surveillance organ where antioxidant defense is critical for host response. It also plays a major role in micronutrient processing. However, the developmental regulation and expression of AOE in the liver is incompletely described. We hypothesized the neonatal liver would have decreased trace mineral associated AOE.

Methods

C57BL/6mice were sacrificed at P0, P7, P21 and 8–12 weeks. mRNA and protein expression of key AOE (SOD1, SOD2, SOD3, Gpx1, Gpx4, Msrb1, TrxR1) and factors for Se processing (Sephs2/Sps2, Scly, Pstk) were measured by qPCR and Western blot.

Results

Hepatic mRNA for selenoenzymes Gpx1 and Msrb1 were developmentally regulated, low at P0 and increased by adult (P &lt; 0.05, n = 5–6). Gpx1 protein increased 7–8-fold and Msrb1 protein increased 6-fold from P0 to adult (P &lt; 0.0001, n = 4). Gene expression of Zn related SOD1 and Mn SOD2 increased postnatally, low at P0 and increased in adult (P &lt; 0.01 n = 5–6). Protein expression for each increased 1.5 and 3-fold from P0 to adult respectively (P &lt; 0.001, n = 4) The mRNA and protein expression for Gpx4, TrxR1 and SOD3 remained constant postnatally.

As the greatest increase was observed in selenoenzymes, factors for Se processing were evaluated. Sephs2, Scly and Pstk mRNA increased from P0 compared to P21 and adult mice (P &lt; 0.05, n = 4–6). Protein expression for Pstk and Scly was highest at P21 and protein for Sps2 increased postnatally (P &lt; 0.01, n = 4).

Conclusions

The liver experiences a postnatal increase in essential trace mineral associated AOE. Additionally, the hepatic machinery for Se processing is low in neonatal mice. We speculate that the neonatal liver is vulnerable to oxidative stress secondary to low AOE defense. We also speculate states that decreased neonatal micronutrient status may further impair the hepatic redox state.

Funding Sources

CCTSI Child Maternal Health Mentored Grant (L.S).

Correction to: International prospective observational cohort study of Zika in infants and pregnancy (ZIP study): study protocol

Following publication of the original article [1], the author mentioned that two additional NIH staff were involved in the development of the protocol who did not receive recognition in the Acknowledgments section in their published article.

International prospective observational cohort study of Zika in infants and pregnancy (ZIP study): study protocol

BACKGROUND

Until recently, Zika virus (ZIKV) infections were considered mild and self-limiting. Since 2015, they have been associated with an increase in microcephaly and other birth defects in newborns. While this association has been observed in case reports and epidemiological studies, the nature and extent of the relationship between ZIKV and adverse pregnancy and pediatric health outcomes is not well understood. With the unique opportunity to prospectively explore the full spectrum of issues related to ZIKV exposure during pregnancy, we undertook a multi-country, prospective cohort study to evaluate the association between ZIKV and pregnancy, neonatal, and infant outcomes.

METHODS

At research sites in ZIKV endemic regions of Brazil (4 sites), Colombia, Guatemala, Nicaragua, Puerto Rico (2 sites), and Peru, up to 10,000 pregnant women will be recruited and consented in the first and early second trimesters of pregnancy and then followed through delivery up to 6 weeks post-partum; their infants will be followed until at least 1 year of age. Pregnant women with symptomatic ZIKV infection confirmed by presence of ZIKV RNA and/or IgM for ZIKV will also be enrolled, regardless of gestational age. Participants will be tested monthly for ZIKV infection; additional demographic, physical, laboratory and environmental data will be collected to assess the potential interaction of these variables with ZIKV infection. Delivery outcomes and detailed infant assessments, including physical and neurological outcomes, will be obtained.

DISCUSSION

With the emergence of ZIKV in the Americas and its association with adverse pregnancy outcomes in this region, a much better understanding of the spectrum of clinical outcomes associated with exposure to ZIKV during pregnancy is needed. This cohort study will provide information about maternal, fetal, and infant outcomes related to ZIKV infection, including congenital ZIKV syndrome, and manifestations that are not detectable at birth but may appear during the first year of life. In addition, the flexibility of the study design has provided an opportunity to modify study parameters in real time to provide rigorous research data to answer the most critical questions about the impact of congenital ZIKV exposure.

TRIAL REGISTRATION

NCT02856984 . Registered August 5, 2016. Retrospectively registered.

Polyhydramnios among women in a cluster-randomized trial of ultrasound during prenatal care within five low and low-middle income countries: a secondary analysis of the first look study

Background

In many low and low-middle income countries, the incidence of polyhydramnios is unknown, in part because ultrasound technology is not routinely used. Our objective was to report the incidence of polyhydramnios in five low and low-middle income countries, to determine maternal characteristics associated with polyhydramnios, and report pregnancy and neonatal outcomes.

Methods

We performed a secondary analysis of the First Look Study, a multi-national, cluster-randomized trial of ultrasound during prenatal care. We evaluated all women enrolled from Guatemala, Pakistan, Zambia, Kenya and the Democratic Republic of Congo (DRC) who received an examination by prenatal ultrasound. We used pairwise site comparisons with Tukey-Kramer adjustment and multivariable logistic models with general estimating equations to control for cluster-level effects. The diagnosis of polyhydramnios was confrimed by an U.S. based radiologist in a majority of cases (62%).

Results

We identified 305/18,640 (1.6%) cases of polyhydramnios. 229 (75%) cases were from the DRC, with an incidence of 10%. A higher percentage of women with polyhydramnios experienced obstructed labor (7% vs 4%) and fetal malposition (4% vs 2%). Neonatal death was more common when polyhydramnios was present (OR 2.43; CI 1.15, 5.13).

Conclusions

Polyhydramnios occured in these low and low-middle income countries at a rate similar to high-income contries except in the DRC where the incidence was 10%. Polyhydramnios was associated with obstructed labor, fetal malposition, and neonatal death.

Trial registration

NCT01990625, November 21, 2013.

Protein Intake During Early Complementary Feeding Affects the Gut Microbiota in U.S. Formula-fed Infants (FS04-03-19)

Objectives

Distinctive growth patterns were observed in formula-fed infants consuming a meat- or dairy-based complementary diet. The gut microbiota and its metabolites are characterized and compared in infants consuming these two common protein-rich foods.

Methods

Healthy, term, formula-fed infants were recruited from metro Denver area, matched by sex and race/ethnicity, and randomized to a meat or a dairy complementary food group from 5 to 12 months of age. Meat- and dairy-based complementary foods and the same infant formula were provided. Total protein intake during the 7-month intervention was ∼3 g/kg/d for both groups. Intakes of infant formula, cereal, fruits, and vegetables were ad libitum. 16S rRNA sequencing and fecal metabolomics were conducted on stool samples collected at 5, 10 and 12 months.

Results

Stool samples were collected from 59 infants (Meat n = 29; Dairy n = 30). Alpha diversity increased over time in the meat group (P = 0.007), not dairy (P = 0.17). Firmicutes was the most abundant phylum for both groups at all time points and Proteobacteria decreased abundance from 5 (10 ± 4%) to 12 months (5 ± 3%) in both groups. At Genus level, Bifidobacterium significantly decreased (21% to 16%, P = 0.001) and Faecalibacteriumsignificantly increased (0.5 to 3.5%, P = 0.0006) over time in both groups, as expected for this age range and with the transition to complementary feeding. A significant group-by-time interaction was observed for Ruminococcus (P = 0.001) and Roseburia (P = 0.002), with increased abundances only in the meat group, not dairy. The abundance of Ruminococcus at 12 months was positively associated with length-for-age Z scores in the dairy group (P = 0.007, R2 = 31%). Both Ruminococcus and Roseburia are potential short-chain fatty acid producers and we saw a significant increase of butyrate and acetate productions in the meat group over time.

Conclusions

Types of protein-rich foods during complementary feeding were associated with the gut microbial composition and metabolites in formula-fed infants. These changes of the gut microbiota may be associated with the different infant growth patterns.

Funding Sources

NIH (NIDDK), NIH/NCATS Colorado CTSA and (alphabetically) Abbott Nutrition, the American Heart Association, the Beef Checkoff through the National Cattlemen's Beef Association, Leprino foods, the National Pork Board.

Differential DNA Methylation of Human Metastable Epialleles in Guatemalan Infants at Birth Due to Timing of a Maternal Lipid-Based Nutrition Supplement and Pre-Pregnancy BMI (P11-139-19)

Objectives

Human metastable epialleles (MEs) are specific regions that are systemically methylated (lack tissue specificity), stable over time, and variably expressed between individuals. DNA methylation (DNAme) of MEs have been previously reported to be altered by maternal nutritional status at the time of conception. Here, we examined DNAme of MEs from a subset of Guatemalan mother/infant dyads as part of the Women First study, an international RCT that tested whether timing of a daily lipid-based nutrition supplement (LNS) improved birth length. The 3 study arms are: women consumed LNS ≥ 3 months prior to conception until delivery (Arm 1, N = 45); women consumed the same LNS commencing at 12 weeks gestation until delivery (Arm 2, N = 45); or no LNS (Arm 3, N = 44). The purpose of this study was to test if the timing of maternal LNS and pre-pregnancy BMI (ppBMI, BMI range = 20.1 – 38.4 kg/m2) led to differential DNAme of MEs in infants at birth.

Methods

Bisulfite-converted DNAme libraries were constructed using NimbleGen SeqCap Epi CpGiant probes from amnion tissue collected at birth. Individuals were sequenced via 2 × 150 paired end reads using the Illumina NovaSeq sequencer. Subjects that passed quality control (131/142 subjects) were used in subsequent statistical analyses. A linear model was used to test for the interaction between maternal LNS and ppBMI on infant DNAme for each base pair site within 296 previously identified candidate ME regions. The number of methylated base pairs per region ranged from 1 – 737. A significance level adjusting for the 296 regions was set at P ≤ 0.000169.

Results

We identified 6 ME regions with significant interactions, demonstrating differential ME DNAme due to intervention arm was dependent on ppBMI values. Our analyses also identified 3 CpGs associated with ppBMI regardless of LNS status and 1 CpG associated with LNS regardless of ppBMI suggesting a role for independent effects of maternal LNS and ppBMI on ME DNAme.

Conclusions

Our findings indicate that timing of maternal LNS and ppBMI contribute to DNAme of candidate MEs in infants at birth, suggesting epigenetic influences due to in utero exposures. Future analyses will identify genes associated with changes in ME DNAme and the role of DNAme on infant growth.

Funding Sources

Bill & Melinda Gates Foundation and National Institutes of Health.

Maternal Characteristics Affect Fetal Growth Response to Maternal Supplements in the Women First Preconception Trial (WF) (P10-017-19)

Objectives

Determine if maternal characteristics modified newborn anthropometric outcomes in the WF trial (ClinicalTrials.gov NCT01883193).

Methods

Secondary analysis included combined data for all 1465 maternal infant dyads in WF sites in Guatemala, India, and Pakistan who had 1st trimester ultrasounds and newborn anthropometry with the three WF arms maintained: Arm 1 commenced a comprehensive nutrition supplement ≥3 months prior to conception; Arm 2 commenced the same supplement in the 1st trimester, and Arm 3 received no trial supplements. Maternal characteristics included were: baseline, BMI, hemoglobin, age, education, SES, and parity plus newborn sex. Newborn outcomes were Z-scores for length (LAZ), weight (WAZ), and weight-to-length ratio (WLRZ). Mixed effect regression models were fit for each outcome, including treatment arm, effect modifier, and treatment arm x effect modifier interaction as predictors and controlling for study site, maternal characteristics, and newborn sex.

Results

Parity, anemia and newborn sex were significant effect modifiers favoring para 0 vs para ≥1, anemia vs non anemia, and newborn male vs female. Effect of Arm 1 vs 3 was significantly larger for para 0 vs ≥1 women on length and weight (Table). Arm 2 vs 3 was not associated with improvements for para 0 in weight (P = 0.273) or WLRZ (P = 0.710). Arms 1 and 2 (vs 3) were associated with significantly higher length, weight, and WLRZ for anemic women.

For parity and anemia, effect sizes for Arm 1 were greater than for Arm 2 for WAZ and WLZ (P < 0.05), but not LAZ. Arm 1 and 2 were associated with significantly higher weight and WLRZ for male vs female newborn.

Conclusions

In diverse low resource populations, impaired fetal growth (weight and length) is substantially improved in nulliparous and in anemic women but minimally or not at all in parous and in non-anemic women. Correction of weight decrements is most pronounced with improvement in maternal nutrition commencing prior to conception.

Funding Sources

Bill & Melinda Gates Foundation; NIH, NICHD and ODS.

Supporting Tables, Images and/or Graphs

Differences in Birth Size Associated with Preconception Maternal Nutrition Intervention Persist in Postnatal Growth Through 6 Months (OR10-05-19)

Objectives

Newborn size, including length, has been associated with later growth. Preconception nutrition supplementation in Women First (WF) trial (clinicaltrials.gov NCT01883193) was associated with higher newborn length and lower rates of stunting. The objective was to examine postnatal growth outcomes through 6 mo in the WF offspring.

Methods

Subjects were 2450 infants from the WF sites in Democratic Republic of Congo (DRC), Guatemala, India, and Pakistan with the 3 WF arms maintained: Arm 1 started a lipid-based nutrition supplement ≥ 3 mo prior to conception; Arm 2 started same supplement at ∼11 wk gestation; and Arm 3 received no trial supplements. Maternal supplementation was discontinued at delivery. Anthropometric measurements were obtained at birth, 0.5, 1, 3, and 6 mo; Z-scores were calculated from WHO Child Growth Standards. Primary and secondary outcomes were analyzed using longitudinal generalized estimating equations accounting for the correlation of repeated measures over time and study cluster. Intervention effects were evaluated within each site as well as overall, adjusting for site and baseline maternal covariates that differed by arm. Statistical interaction between arm and infant sex was evaluated and included if significant.

Results

Longitudinal growth (length and weight) curves from 0–6 mo demonstrated differences by intervention arm, with Arms 1 and 2 more favorable compared to Arm 3 (Table + LAZ in Figure). Relative risks (RR) for stunting were significantly lower for Arms 1 and 2 (vs. Arm 3) (Table). Site-specific analyses revealed some site differences warranting further exploration.

Conclusions

Without any postnatal intervention, differences in fetal growth evident at birth persisted at 6 months.

Funding Sources

Bill & Melinda Gates Foundation; NIH, NICHD & ODS.

Supporting Tables, Images and/or Graphs

The Time Course of Human Milk Insulin and Glucose Response to an Oral Glucose Challenge - A Case Study (P11-045-19)

Objectives

It is often cited that insulin in human milk (HM) increases postprandially along with maternal serum insulin. However, this response has never been documented in humans, and the characteristics of this increase remain unstudied.

Methods

Two healthy lactating women emptied their breasts after a fast (> 8 hours) using an electric breast pump. After consuming a 50g glucose water, each woman emptied a single breast at 15, 30, 60, 90, and 120 minutes. Skim milk was generated via centrifugation and HM glucose and insulin were measured via hexokinase assay and chemiluminescent immunoassay (Beckman Coulter). At each of these time points maternal blood was collected via finger prick and capillary glucose measured via handheld glucometer. Additional blood was spotted onto a dried blood spot (DBS) card, and maternal insulin was measured from the DBS cards via Ultrasensitive ELISA (Mercodia).

Results

Both insulin and glucose concentrations rose in HM after the glucose load (Figure A, B). The amplitudes of both HM insulin and glucose were lower than that of maternal circulation. Neither HM insulin nor glucose were correlated with concentrations in maternal blood. However insulin concentrations were tightly correlated with glucose concentrations in both HM (P < 0.0001, R2 = 0.84) and maternal blood (P < 0.001, R2 = 0.72). At 2 hours post-glucose challenge, both maternal blood insulin and glucose had returned to near fasting levels (insulin: 15.1 ± 7.8 µU/mL; glucose: 107 ± 4 mg/dL). However, HM insulin and glucose concentrations remained elevated (Figure A, B). At 2 hours, HM insulin remained 13 times higher than fasting concentrations and HM glucose remained 3.9 times higher than fasting concentrations.

Conclusions

To our knowledge, these are the first data in humans to characterize the time course of HM insulin response to an oral glucose challenge. These data will inform the design of HM composition studies when free-living HM samples are collected. The impact of variation in these components over the day on the recipient infant deserves further research.

Funding Sources

Internally Funded.

Supporting Tables, Images and/or Graphs

Different Gut Microbial Profiles in African and South Asian Women of Childbearing Age in the Women First (WF) Trial (FS07-05-19)

Objectives

To characterize and compare the gut microbial structures in women of childbearing age from Africa (the Democratic Republic of the Congo, DRC) and South Asia (India)

Methods

Women of childbearing age were recruited from rural DRC and India as part of the WF preconception maternal nutrition trial. Findings presented include the gut microbiota (16S rRNA sequencing) of women at the time of randomization prior to conception in the WF trial and represented participants from two of the WF sites, with distinctive ethnicity, diet, culture and geographical locations. Women were recruited from 12 villages in rural DRC and 9 villages from rural India. 24-h dietary recalls were conducted on half of the participants during the first trimester of pregnancy.

Results

Stool samples were collected from n = 217 women (DRC n = 117; India n = 100). Alpha diversity of the gut microbiota was higher in DRC than in India using Chao1 (91 ± 11 vs. 82 ± 12, P < 0.001). At Phylum level, the three most abundant phyla were Bacteroidetes (51%), Firmicutes (33%) and Proteobacteria (7%) with no difference between sites. At Genus level, Prevotella remained the only strain that was over 10% abundance (DRC: 27 ± 11%; India: 29 ± 13%; P = 0.32). Some fiber fermenting strains Succinivibrio (DRC: 3.2 ± 0.8%; India: 3.2 ± 0.6%; P = 0.11) and Roseburia (DRC: 2.0 ± 0.3%; India: 2.1 ± 0.4%; P = 0.31) remained at relatively high abundance without differences between sites. Bifidobacterium was higher in India (4.95 ± 1.0%) than DRC (0.3 ± 0.1%; P = 1.4E-28), and so was Lactobacillus (DRC: 0.2 ± 0.0%; India: 1.2 ± 0.1%; P = 3.0E-14) and Faeclibacterium (DRC: 6.0 ± 1.7%; India: 8.4 ± 2.9%; P = 2E-7). Ruminococcus was higher in DRC (2.3 ± 0.7%) than in India (1.8 ± 0.4%; P = 1.3E-5). These observations were consistent with dietary intakes of high plant-based foods for both countries; India being primarily vegetarian with regular fermented dairy consumption and DRC consumes animal-based foods regularly. The gut microbiota was not affected by any demographical variables.

Conclusions

Consumption of plant-based foods, animal-flesh foods and fermented dairy foods all had independent effects on the gut microbiota, and the effects of diet apparently overrides other potential factors such as socioeconomic status.

Funding Sources

Bill & Melinda Gates Foundation; NIH, NICHD & ODS.

The Effect of a Preconception Nutrition Supplement on One Carbon Metabolites (P24-028-19)

Objectives

Studies in both animals and humans have shown that maternal dietary restriction and supplementation of one carbon (1C) metabolites (methyl donors), such as methionine and choline, can impact offspring growth, insulin resistance, and DNA methylation. However, there has been limited longitudinal research of 1C metabolite concentrations over the reproduction cycle of human pregnancy. The purpose of this study was to investigate if 1C metabolite concentrations change prior to and during pregnancy and if a preconception lipid-based nutrition supplement (LNS) influences such changes.

Methods

This study was a secondary analysis as part of the Women First study (clinicaltrials.gov, NCT01883193), a large, randomized controlled trial investigating whether the timing of maternal LNS initiation would impact fetal growth and development. The study arms were supplementation at least 3 months prior to conception (Arm 1), supplementation at ∼12 weeks of gestation (Arm 2), or no supplementation (Arm 3). Dried blood spot (DBS) cards were collected at study enrollment prior to conception, and at 12 and 34 weeks gestation. A targeted 1C metabolite assay (27 metabolites) was performed on a subset of DBS samples from Guatemalan women (n = 134) at each time point using liquid chromatography/tandem mass spectrometry. Longitudinal analyses were performed using linear mixed modeling to investigate the influence of time and LNS on these metabolites.

Results

The concentrations of two metabolites were changed by intervention status: asymmetric dimethylarginine (ADMA) and symmetric dimethylarginine (SDMA). Twenty-one of 27 metabolites significantly changed from preconception and across gestation after correcting for multiple testing using the Bonferroni correction (P < 0.00185).

Conclusions

Preconception LNS significantly decreased the level of ADMA, a metabolite which has been implicated in intrauterine growth restriction and preeclampsia. More work is needed to determine whether this intervention could influence development of these conditions.

Funding Sources

Bill & Melinda Gates Foundation, National Institutes of Health.

Term Offspring of Nulliparous Women Have Lower Weight-for-age Z-scores Than Multiparous Women in Chimaltenango, Guatemala (P11-065-19)

Objectives

Women´s parity may be associated with their current offspring growth outcomes. We aimed to determine the association between parity and birthweight among Guatemalan women.

Methods

We analyzed 50,987 term births that occurred between 2014 and 2016 in the Chimaltenango, Guatemala site of the Global Network for Women and Children's Health Maternal Neonatal Health Registry. Gestational age was determined by ultrasound or last menstrual period. Birthweight was obtained within the first 48 hours of life with up to 10 grams of accuracy and converted to z-scores using INTERGROWTH-21st with quality control. The means of birthweight z-scores were computed by ANOVA; adjusting for maternal age and height did not change the results. P-values < 0.05 were considered significant.

Results

Nulliparous women had offspring of lower birthweight than women with parity 1–3 and 4 or more.

Conclusions

These results may have implications for the targeting of nutritional interventions aimed to improve pregnancy outcomes.

Funding Sources

Grant from National Institute of Child Health and Human Development.

Supporting Tables, Images and/or Graphs

Exchangeable Zinc Pool (EZP) Size in Bangladeshi Toddlers at Risk of Environmental Enteric Dysfunction (EED) Is Not Influenced by Inflammation (OR07-03-19)

Objectives

The effect of infection and inflammation on EZP size, a putative biomarker of Zn status, is unknown. We sought to evaluate the relationship of EZP to systemic inflammation in a group of toddlers with high risk of Zn deficiency and EED.

Methods

Subjects were 112 children from an urban slum area in Dhaka, Bangladesh, aged 18–24 months, at risk of EED and who participated in stable isotope studies of Zn homeostasis while on habitual diets. All children underwent screening for EED, using L: M ratio, and multiple markers of systemic and intestinal inflammation were obtained. EZP size was measured from intravenously administered stable isotope of Zn, with urine collections obtained over 4 days, starting 3 days after isotope administration. The size of EZP was calculated by dividing the mass (mg) of intravenous isotope dose (67 Zn or 68 Zn) by the enrichment value at the y intercept of the linear regression: EZP (mg) = IV dose/y-intercept. The intercept is estimated from linear regression of a semi-log plot of urine enrichment data. Multiple regression analysis was applied to examine relationships among child size, serum Zn, markers of systemic and intestinal inflammation and EZP. EZP was also compared to previously published evaluation of EZP from multiple studies (Miller LV, J Nutr, 2016).

Results

Mean (± SD) age of subjects was 19 ± 2 months; mean weight 9.1 ± 1.0 kg; mean dietary Zn was 2.7 mg/d. Mean EZP was 3.7 ± 0.5 mg/kg, compared to an expected value of 4.1 mg/kg based on the previous published analysis of all child data; all data points were within the 90% prediction interval from those data (Figure). Compared to existing child data, these children weighed less and had smaller EZP. EZP/kg was directly associated with serum Zn concentration, and inversely associated with hemoglobin and dietary Zn (best model R2 = 0.24). The analyses indicated no evidence of an association of EZP/kg with any biomarkers of inflammation.

Conclusions

The apparent absence of an effect of inflammation on EZP size may provide some advantage over serum Zn. The inverse relationship with dietary Zn suggests redistribution of Zn pools in chronic marginal Zn status. Evaluation of utility of EZP as a biomarker of Zn status will require measurement during controlled interventions.

Funding Sources

Bill & Melinda Gates Foundation.

Supporting Tables, Images and/or Graphs

The use of Zinc Stable Isotopes in the Study of Iron-Zinc Interactions in Chilean Women

The objective of this study was to compare the fractional zinc absorption (FAZ) and the size of the rapidly exchangeable zinc pool (EZP) after three months of iron supplementation in women consuming ferrous sulfate between meals. Twenty-one non-anemic apparently healthy women received on average 55.1 ± 18.5 mg elemental iron per day as ferrous sulfate, and five received no supplemental iron. Fractional absorption of zinc was determined before and three days after finishing the third month of iron supplementation by using an extrinsic labeling with zinc stable isotopes and a dual isotope enrichment method in urine. EZP was determined from urine enrichment following intravenous administration of 70Zn. Results of selected zinc-related variables in the iron supplemented women were (before vs. after iron supplementation): FAZ with meal 0.22 vs. 0.24, p = .23; FAZ in fasting state 0.58 vs. 0.69, p = .005; EZP 177 mg vs. 160 mg, p = .058; plasma zinc 90.6 vs. 86.1 μg/dl, p = .065. The control group remained unchanged. The capacity to absorb zinc was increased three days after terminating a period of iron supplementation as compared with the pre-iron period. This may be attributable to impairment of zinc status by the iron supplements as evidenced by a trend for lower plasma zinc and EZP.

The gut microbiota in infants of obese mothers increases inflammation and susceptibility to NAFLD

Maternal obesity is associated with increased risk for offspring obesity and non-alcoholic fatty liver disease (NAFLD), but the causal drivers of this association are unclear. Early colonization of the infant gut by microbes plays a critical role in establishing immunity and metabolic function. Here, we compare germ-free mice colonized with stool microbes (MB) from 2-week-old infants born to obese (Inf-ObMB) or normal-weight (Inf-NWMB) mothers. Inf-ObMB-colonized mice demonstrate increased hepatic gene expression for endoplasmic reticulum stress and innate immunity together with histological signs of periportal inflammation, a histological pattern more commonly reported in pediatric cases of NAFLD. Inf-ObMB mice show increased intestinal permeability, reduced macrophage phagocytosis, and dampened cytokine production suggestive of impaired macrophage function. Furthermore, exposure to a Western-style diet in Inf-ObMB mice promotes excess weight gain and accelerates NAFLD. Overall, these results provide functional evidence supporting a causative role of maternal obesity-associated infant dysbiosis in childhood obesity and NAFLD.

Trends and determinants of stillbirth in developing countries: results from the Global Network's Population-Based Birth Registry

Background

Stillbirth rates remain high, especially in low and middle-income countries, where rates are 25 per 1000, ten-fold higher than in high-income countries. The United Nations’ Every Newborn Action Plan has set a goal of 12 stillbirths per 1000 births by 2030 for all countries.

Methods

From a population-based pregnancy outcome registry, including data from 2010 to 2016 from two sites each in Africa (Zambia and Kenya) and India (Nagpur and Belagavi), as well as sites in Pakistan and Guatemala, we evaluated the stillbirth rates and rates of annual decline as well as risk factors for 427,111 births of which 12,181 were stillbirths.

Results

The mean stillbirth rates for the sites were 21.3 per 1000 births for Africa, 25.3 per 1000 births for India, 56.9 per 1000 births for Pakistan and 19.9 per 1000 births for Guatemala. From 2010 to 2016, across all sites, the mean stillbirth rate declined from 31.7 per 1000 births to 26.4 per 1000 births for an average annual decline of 3.0%. Risk factors for stillbirth were similar across the sites and included maternal age < 20 years and age > 35 years. Compared to parity 1–2, zero parity and parity > 3 were both associated with increased stillbirth risk and compared to women with any prenatal care, women with no prenatal care had significantly increased risk of stillbirth in all sites.

Conclusions

At the current rates of decline, stillbirth rates in these sites will not reach the Every Newborn Action Plan goal of 12 per 1000 births by 2030. More attention to the risk factors and treating the causes of stillbirths will be required to reach the Every Newborn Action Plan goal of stillbirth reduction.

Trial registration

NCT01073475.

Markers of Oxidative Stress in Human Milk do not Differ by Maternal BMI But are Related to Infant Growth Trajectories

Objective Obesity in adults is associated with inflammation and oxidative stress. Whether or not this phenotype is reflected in human milk (HM) composition, or may impact infant growth remains unknown. We investigated whether HM from overweight/obese (OW/Ob) mothers exhibited higher concentrations of inflammatory cytokines and markers of oxidative stress. We also correlated these bioactive components with infant growth patterns. Methods This was an observational cohort of 56 breastfeeding mothers and their infants [33 normal weight (NW) and 23 OW/Ob]. Infants were followed until 6 months of age and HM collected at 2-weeks and 4-months. Results Markers of oxidative stress, 8-hydroxy-deoxyguanosine (8OHdG) and 4-hydroxynonenol (HNE), decreased in HM over time (p < 0.001) and did not differ between NW and OW/Ob women. Concentrations of inflammatory cytokines, IL-6, IL-8, and TNF-α, were all inter-correlated (p < 0.001) but did not differ between NW and OW/Ob women. HM fat, protein, lactose, and total calories did not differ between NW and OW/Ob women. Infant growth patterns did not differ by group. In a model of infant weight-for-length-Z score trajectory, there was a significant interaction between both lactose and 8OHdG with maternal group: HM lactose and 8OHdG concentrations were both positively associated with increases in WLZ trajectory only among infants breastfed by OW/Ob mothers. Conclusions for Practice HM composition was relatively stable between NW and OW/Ob women. In exclusively breastfed infants, HM concentrations of lactose and 8OHdG, a marker of oxidative stress, may contribute to regulation of infant weight gain, especially among infants of OW/Ob women.

International summit on the nutrition of adolescent girls and young women: consensus statement

An international summit focusing on the difficult challenge of providing adequate nutrition for adolescent girls and young women in low- and middle-income countries was held in Portland, Oregon in 2015. Sixty-seven delegates from 17 countries agreed on a series of recommendations that would make progress toward improving the nutritional status of girls and young women in countries where their access to nutrition is compromised. Delegate recommendations include: (1) elevate the urgency of nutrition for girls and young women to a high international priority, (2) raise the social status of girls and young women in all regions of the world, (3) identify major knowledge gaps in the biology of adolescence that could be filled by robust research efforts, (4) and improve access to nutrient-rich foods for girls and young women. Attention to these recommendations would improve the health of young women in all nations of the world.

Maternal Beliefs Surrounding Infant Feeding, but Not Maternal BMI or Hospital Experience, Predict Breastfeeding Exclusivity and Behavior

BACKGROUND

Overweight/Obese (OW/Ob) women are at risk for breastfeeding failure. How maternal BMI affects lactation support received in-hospital, or maternal attitudes and beliefs surrounding infant feeding remains poorly understood. We investigated 1) the impact of in-hospital lactation support and maternal attitudes and behaviors regarding infant feeding on breastfeeding exclusivity, and 2) whether these potentially modifiable attitudes and behaviors differed between normal weights (NW) versus OW/Ob women.

METHODS

NW (n=18) and OW/Ob (n=20) women and their infants were followed from birth to 4-months postpartum. In-hospital experiences, problems and help received regarding breastfeeding were documented. Six maternal attitudes and behaviors surrounding infant feeding were assessed at 2-weeks and 4-months. These factors were compared between NW and OW/OB women, and in relation to breastfeeding exclusivity.

RESULTS

In-hospital experiences, assistance received regarding breastfeeding difficulties, and infant breastfeeding exposure did not differ between NW and OW/Ob women. At 4-months OW/Ob women were more likely to feed their infant on a schedule (p<0.03); this was the only difference in attitudes/behaviors between BMI-groups. Feeding the infant on a schedule was predictive of lower total breastfeeding exposure (p<0.05). Maternal concern about infant under-eating/becoming underweight was associated with several negative feeding behaviors, including reduced breastfeeding exposure (p<0.02), pressuring feeding style (p<0.01), and feeding to calm fussiness (p<0.01).

CONCLUSION

This hospital setting provided equitable breastfeeding support to OW/Ob and NW women. Maternal concern over infant under-eating/under-gaining and encouragement to feed on-demand are prime interventional targets to improve breastfeeding outcomes; the latter may be especially relevant to OW/Ob mothers.

HELIOS--A laboratory based on high-order harmonic generation of extreme ultraviolet photons for time-resolved spectroscopy

In this paper, we present the HELIOS (High Energy Laser Induced Overtone Source) laboratory, an in-house high-order harmonic generation facility which generates extreme ultraviolet (XUV) photon pulses in the range of 15-70 eV with monochromatized XUV pulse lengths below 35 fs. HELIOS is a source for time-resolved pump-probe/two-color spectroscopy in the sub-50 fs range, which can be operated at 5 kHz or 10 kHz. An optical parametric amplifier is available for pump-probe experiments with wavelengths ranging from 240 nm to 20,000 nm. The produced XUV radiation is monochromatized by a grating in the so-called off-plane mount. Together with overall design parameters, first monochromatized spectra are shown with an intensity of 2 ⋅ 10(10) photons/s (at 5 kHz) in the 29th harmonic, after the monochromator. The XUV pulse duration is measured to be <25 fs after monochromatization.

Redistribution of tissue zinc pools during lactation and dyshomeostasis during marginal zinc deficiency in mice

Zinc (Zn) requirements are increased during lactation. Increased demand is partially met through increased Zn absorption from the diet. It is estimated that 60-80% of women of reproductive age are at risk for Zn deficiency due to low intake of bioavailable Zn and increased demands during pregnancy and lactation. How Zn is redistributed within the body to meet the demands of lactation, and how Zn deficiency affects this process, is not understood. Female C57bl/6J mice were fed a control (ZA; 30mg Zn/kg) or a marginally Zn deficient (ZD; 15mg Zn/kg) diet for 30 days prior to mating through mid-lactation and compared with nulliparous mice fed the same diets. While stomach and plasma Zn concentration increased during lactation in mice fed ZA, mice fed ZD had lower stomach Zn concentration and abrogated plasma Zn levels during lactation. Additionally, femur Zn decreased during lactation in mice fed ZA, while mice fed ZD did not experience this decrease. Furthermore, red blood cell, pancreas, muscle and mammary gland Zn concentration increased, and liver and adrenal gland Zn decreased during lactation, independent of diet, while kidney Zn concentration increased only in mice fed ZD. Finally, maternal Zn deficiency significantly increased the liver Zn concentration in offspring but decreased weight gain and survival. This study provides novel insight into how Zn is redistributed to meet the increased metabolic demands of lactation and how marginal Zn deficiency interferes with these homeostatic adjustments.