Growth and adiposity in newborns study (GAINS): The influence of prenatal DHA supplementation protocol

Abstract

Background:

Obesity and central fat mass (FM) accrual drive disease development and are related to greater morbidity and mortality. Excessive gestational weight gain (GWG) increases fetal fat accretion resulting in greater offspring FM across the lifespan. Studies associate greater maternal docosahexaenoic acid (DHA) levels with lower offspring FM and lower visceral adipose tissue during childhood, however, most U.S. pregnant women do not consume an adequate amount of DHA. We will determine if prenatal DHA supplementation is protective for body composition changes during infancy and toddlerhood in offspring exposed to excessive GWG.

Methods and design:

Infants born to women who participated in the Assessment of DHA on Reducing Early Preterm Birth randomized controlled trial (ADORE; NCT 02626299) will be invited to participate. Women were randomized to either a high 1000 mg or low 200 mg daily prenatal DHA supplement starting in the first trimester of pregnancy. Offspring body composition and adipose tissue distribution will be measured at 2 weeks, 6 months, 12 months, and 24 months using dual energy x-ray absorptiometry. Maternal GWG will be categorized as excessive or not excessive based on clinical guidelines.

Discussion:

Effective strategies to prevent obesity development are lacking. Exposures during the prenatal period are important in the establishment of the offspring phenotype. However, it is largely unknown which exposures can be successfully targeted to have a meaningful impact. This study will determine if prenatal DHA supplementation modifies the relationship between maternal weight gain and offspring FM and FM distribution at 24 months of age.

Ethics and dissemination:

The University of Kansas Medical Center Institutional Review Board (IRB) approved the study protocol (STUDY00140895). The results of the trial will be disseminated at conferences and in peer reviewed publications.

Trial registration:

ClinicalTrials.gov ID: NCT03310983

1. Introduction

Greater infant fat mass (FM) is directly linked to the development of overweight and obesity [1], a characteristic of 35.4% of US children and adolescents [2]. Childhood overweight, obesity, and central FM drive disease development [3], leading to the prediction of a rapid generational decline in life expectancy [4] and to a surge in obesity related medical care costs (~$150 billion/yr) [5]. It is critical to understand nutrients that favorably influence offspring FM accumulation and adipose tissue (AT) distribution as these are important drivers of obesity occurrence, disease risk, and severity of disease [6].

Appropriate prenatal nutrient intake and adequate gestational weight gain (GWG) are vital for optimal fetal development [7]. The first 1000 days post-conception are a critical exposure window that programs susceptibility for offspring disease development [8]. Studies have shown individuals exposed in utero to famine are at increased risk of cardiovascular disease [9], diabetes [10], obesity [11], and certain forms of cancer [12]. During the first 1000 days there is a rapid expansion in the number and size of adipocytes [13,14], such that FM represents 5–28% of neonate birth weight [15,16], however fat-free mass varies little. Thus, the variation in birth weight is due to an expansion in fat storage [17]. There are numerous factors that influence the expansion of adipocytes. The maternal diet [18-21], pre-pregnancy BMI [22-24], metabolic status¹⁸, and GWG [1,16,22,24] are related to the level of offspring adposity [25,26], whereas other factors may limit adipocyte expansion.

Optimal fetal development, and thus fat accretion, is disrupted by excessive GWG. Pregnancy is a particularly sensitive exposure window with long-term fetal programming ramifications. Excessive GWG is a known insult related to offspring obesity and disease development. Excessive GWG increases fetal fat accretion resulting in greater total FM and central FM in newborns [23,27], children [1,22] and in adults [24]. Research shows a strong relationship between excessive GWG and offspring obesity development [28,29], diabetes, and cardiovascular disease [30,31]. Recent studies have shown that increased maternal DHA (docosahexaenoic acid) levels are related to lower levels of offspring FM and central adiposity [32-35].

However, DHA status is lower in US women than in other developed countries [36-38]. Data from recent Midwest based pregnancy cohorts report early pregnancy DHA red blood cell (RBC) levels ranged from 4.3 to 5.0% [37,39,40] compared to >6% reported by others [41-43]. US pregnant women and women of reproductive age consume ~60 mg DHA/day [44,45], and synthesize very little DHA (~40 mg) from α-linolenic acid (ALA; 18:3n-3) consumed in other foods [46,47]. This falls [48] well below the Dietary Reference Intake (DRI) of 200 mg/day of DHA during pregnancy. Only 1.5% of pregnant women reported taking a prenatal vitamin that included DHA, totaling 1.18 g/month for DHA+ eicosapentaenoic acid (EPA), well below the recommended 6–9 g/month [49].

Low maternal intake of DHA may have an impact on the offspring phenotype and body composition changes. It has been proposed that polyunsaturated fatty acids (PUFA) like DHA influence adipocyte development and fat accrual in two ways: influencing adipocyte differentiation and fatty acid metabolism regulation. DHA is a long chain PUFA, and a member of the n-3 fatty acid family. Animal and cell culture studies found exposure to n-3 fatty acids prevented adipocyte maturation [50-53]. The n-3 fatty acids also act like hormones, binding peroxisome proliferator-activated receptors (PPAR) that direct gene expression for fatty acid metabolism [54-56]. Exposure to n-6 fatty acids upregulated PPAR leading to increased adipogenesis [57-61]. However, the n-3 fatty acids suppressed genes involved in lipogenesis (fatty acid synthase, lipoprotein lipase, and stearoyl-CoA desaturase-1) and increased the expression of genes involved with β oxidation (acetyl Co-A oxidase) [62,63]. The overall net effect of increased exposure to n-3 fatty acids was a decrease in adipose tissue deposition. In fact, rats prenatally supplemented with n-3 PUFA had offspring with lower body weight and adipose tissue levels when compared to offspring exposed prenatally to a maternal diet high in n-6 PUFA or low in n-3 PUFA [50,51].

Therefore, the purpose of this study is to investigate if prenatal DHA supplementation can modify the effect of excessive GWG on offspring body composition and AT distribution during infancy and toddlerhood (birth to 24 months). To accomplish this aim, infants born to women who participated in the Assessment of DHA on Reducing Early Preterm Birth randomized controlled trial (ADORE; NCT 02626299) will be invited to enroll. Women in the ADORE trial were randomized to either a high 1000 mg or low 200 mg daily prenatal DHA supplement starting in the first trimester of pregnancy. It is important to identify interventions that protect against adverse programming effects on the offspring phenotype, and lead to a decrease in disease risk.

2. Methods

2.1. Study overview of the parent trial (ADORE)

The ADORE (R01 HD083292; NCT02626299) trial protocol [64] and results have been published [65]. An overview of the study and methods will be provided. The primary purpose of the parent trial was to determine if prenatal DHA supplements of 1000 mg/day vs. 200 mg/day can reduce early pre-term birth (ePTB; Specific Aim 1) and to conduct a secondary pregnancy efficacy analysis to determine if there is a subset of pregnancies most likely to benefit from DHA supplementation (Specific Aim 2). sRAGE is a marker of inflammation that may reflect chronic inflammation. To obtain information about the effects of DHA on inflammation, sRAGE was measured (Specific Aim 3). ADORE was a randomized, double-blind, controlled Phase III Clinical Trial of DHA supplementation during the last two trimesters of pregnancy. Subjects were randomized to consume 1000 mg/day DHA or 200 mg/day DHA. The primary outcome was to move early pre-term birth to late pre-term birth.

ADORE study enrollment occurred at three sites (Kansas City, KS, Columbus, OH, and Cincinnati, OH), however, the GAINS study will only follow offspring from the Kansas City, KS site. Listed in Table 1 is the inclusion and exclusion criteria for the ADORE trial. Women who were ≥18 years of age and were in their 12th to 20th week of gestation (based upon 2014 American Congress of Obstetricians and Gynecologists (ACOG) guidelines) were eligible for enrollment. Pregnant women were randomized to one of two arms (groups) with a maximum number of pregnant women n_max = 400 at the Kansas City site. Recruitment at the Kansas City site was outperforming the other two sites, therefore a total of n = 489 were enrolled, greater than initially estimated.

Table 1.

Inclusion and exclusion criteria for the ADORE trial.

Inclusion criteria
1.	Pregnant females 18.0 years and older who are 12 to 20 weeks gestation at study entry
2.	Agree to consume study capsules and a typical prenatal supplement of 200 mg DHA
3.	Available by telephone
Exclusion criteria
1.	<18 years of age at enrollment
2.	Expecting multiple infants
3.	Gestational age at baseline <12 weeks or > 20 weeks
4.	Unable or unwilling to agree to consume capsules until delivery
5.	Unwilling to discontinue use of another prenatal supplement with DHA
6.	Women with allergy to any component of DHA product (including algae), soybean oil or corn oil

2.2. Fatty acid analysis

Maternal blood taken at enrollment and birth, and cord blood were all analyzed for red blood cell (RBC) fatty acid content. RBCs were separated from plasma and buffy coat by centrifugation (3000 ×g, 10 min; 4 °C), frozen, and stored under nitrogen at −80 °C until analysis. Phospholipids from erythrocytes were isolated according to a modified Folch method [37], and fractionated by thin-layer chromatography, transmethylated with boron trifluoride-methanol, and the resulting fatty acid methyl esters (FAME) separated and quantified using a Varian 3900 gas chromatograph with an SP-2560 capillary column (100 m, Sigma Aldrich) and a Star 6.41 Chromatography Workstation for peak integration and analysis [37]. Individual peaks were identified by comparison with qualitative standards (PUFA 1 and PUFA 2, Sigma Aldrich) and a weighed standard mixture (Supelco 37 Component FAME mix, Sigma Aldrich) were used to adjust fatty acids for area/weight to calculate a final weight percent of total fatty acids.

2.3. Gestational weight gain

GWG was calculated by subtracting the self-reported pre-pregnancy body weight from the last body weight measured in the outpatient clinic (from the electronic medical record) prior to delivery. In the clinic, body weight is taken in light clothing with shoes removed. To accurately categorize weight gain status, we accounted for gestational age at last recorded clinic visit. Using the recommended rate of weight gain range for the third trimester [7], we calculated a personalized range for each participant. The personalized range was used to classify GWG according to 2009 IOM GWG guidelines as excessive or not excessive [7].

2.4. Maternal dietary inake

Maternal dietary intake was measured at enrollment (12–20 wks in gestation). The National Cancer institute (NCI) Diet History Questionnaire II (DHQ-II) food frequency questionnaire was completed by each English-speaking non-Latina woman. Data collected from the DHQ-II was analyzed using Diet*Calc Analysis software to generate dietary intake of nutrients (including added sugars and fructose) and food groups. Spanish-speaking and English-speaking Latina women did not complete the DHQ-II and instead completed three 24-h dietary recalls, two during the weekdays and one on the weekend. The DHQ-II was not developed and has not been validated in a population of Latina adults. Therefore, to accurately assess dietary intake in our Latina population, 24-h dietary recalls were completed. To investigate if having different dietary intake methods introduces error, we will explore the mean intake for dietary variables to determine if differences are found. For variables with differences, we will explore the literature to understand if this an expected cultural difference in intake or likely representing error created by the dietary intake method used. A trained research staff fluent in Spanish collected the recall information using the multiple pass methods. The recalls were entered into the Nutrition Data System for Research (NDS-R, version 2017, Minneapolis, MN) for macro- and micronutrient analysis. Completion of recalls vs. the DHQ-II was done to be culturally sensitive and appropriately monitor the intake of Hispanic women who may not conform to a traditional US-diet that is represented by the DHQ-II.

2.5. Study overview of the GAINS proposal

The Growth and adiposity in newborns study (GAINS; NCT 03310983; DK118220) takes a multidisciplinary approach, combining nutrition during pregnancy, a critical period of early life exposure, to evaluate the association between prenatal DHA supplementation and GWG status on offspring fat accrual and AT distribution at 24 months. To achieve the study aims, we will recruit offspring born to women who participated in the ADORE trial. At enrollment of the ADORE trial, all participants will be asked if they can be contacted for other related studies. Their decision is recorded in the consent and a database. A total of 97% of the parent RCT participants have agreed to be contacted. Late in pregnancy, women enrolled in the ADORE trial will be mailed a letter introducing them to the follow-up study. A follow-up visit will occur at delivery, where the women will be approached and asked if they are interested in the study.

The primary aim of GAINS is to determine if the prenatal dose of DHA (1000 mg/day vs. 200 mg/day) interacts with GWG (excessive vs. non-excessive) to influence total infant FM at 24 months. Over half (56%) of all pregnant women, regardless of pre-pregnancy BMI, experience excessive GWG. In the ADORE trial Kansas City participants only, the excessive GWG rate was 53.7%. All participants will have equal opportunity to be consented into the proposed study. The primary analysis tests the high- vs. low-dose DHA effect in excessive GWG verses the high- vs. low-dose DHA effect in non-excessive GWG. Therefore, since all participants (normal, overweight, and obese) have an opportunity to have excessive or non-excessive GWG, we will include women from all pre-pregnancy BMI groups. Since we hypothesize the effect will be found in all mothers who gain excessive gestational weight, regardless of the BMI group, inclusion of all BMI groups will not dilute the effect or decrease power. The secondary aim will determine if the prenatal dose of DHA (1000 mg/day vs. 200 mg/day) interacts with sex (male vs. female) to influence infant central FM at 24 months. Differences by sex in total FM and AT distribution have been noted across the lifespan [66-68]. Further, the location of AT is an important determinant of metabolic and cardiovascular disease risk where centrally located FM is related to greater incidence of disease [69,70].

Table 2 provides an overview of the study visits and timing of assessments. Infant body composition and AT distribution will be measured at 2 weeks and 6, 12, and 24 months using dual energy x-ray absorptiometry (DXA). The PI and study staff will remain blinded to group assignment until all infants complete the 24-month follow-up visit. Upon completion of the 24-month visit, the PI will have access to study assignment (1000 mg/day or 200 mg/day) to complete analyses for the aims of the GAINS study.

Table 2.

An outline of study visits and procedures for the GAINS study.

Study procedure	36 wks- delivery	2 wks	6 months	12 months	24 months
Consent	•
DXA scan & Anthropometry		•	•	•	•
Diet recall & feeding questionnaire		•	•	•	•

2.6. Anthropometrics

Body weight will be assessed on the same calibrated scale throughout the study duration (Detetco Scales, Webb City, MO). Length will be measured using an infant length board (Shorr Productions) and at 18 months old, standing height will be measured using a wall mounted stadiometer (Accu-Hite, Seca Corp, Hanover, MD). Subjects will remove shoes and be centered on the stadiometer. Height will be recorded to the nearest 0.1 cm. Two measurements will be taken, and the average will be recorded.

2.7. Medical records

Medical records of the child will be collected and coded. Using a prior study coding system [71], the study team will categorize and code any medical events from birth to 24 months. Examples include, diagnosis of allergies, a broken bone, feeding issues, respiratory virus, etc.,

2.8. Dual energy x-ray absorptiometry to measure total body fat and AT distribution

To assess body composition and distribution changes from birth to two years old, we selected a validated technique that was feasible to use in a newborn, infant, and toddler population. There are limited techniques that can be used across these ages. For the Pea Pod, an infant outgrows the system around six months of age, while other systems cannot be used until the infant is between two to five years old (Tod Pod and Bod Pod). Therefore, to address our sample and study needs and avoid introduction of error when switching to a different technique during data collection, we chose to use dual energy x-ray absorptiometry (DXA; Prodigy, Madison, WI, encore software version 13.60) to measure body composition and regional AT distribution. Using specific anatomic landmarks as previously described [72] and depicted in Fig. 1, regions including the arms, legs, and trunk will be demarcated (arm fat, leg fat, and trunk fat). To identify the arms, the landmarks including the soft tissue extending from the center of the arm socket to the phalange tips will be used. To identify the legs, the landmarks using the soft tissue extending from a computer-generated line drawn through and perpendicular to the axis of the femoral neck and angled with the pelvic rim to the phalange tips will be used. The trunk region will be identified by the region below the chin and above the shoulders, and interior to the regions identified for the arms and legs. Infants were placed in an immobilizer to prevent movement and ensure body positioning to allow for identification of landmarks and regions. One rater analyzed all scans and determined if movement was detected that would have invalidated the results. Those scans were marked as unusable. Calculations will be completed for FM comprising the central (trunk) and peripheral (arms plus legs) regions. Therefore, AT distribution will be defined as central or peripheral based on this analysis.

Fig. 1.

Overview for the body regions measured by dual energy x-ray absorptiometry (DXA).

2.9. Dietary intake of the child

One multiple-pass 24-h dietary recall will be administered to the child's caregiver and collected by trained research staff at each visit to characterize energy and nutrient intake. 24-h recalls accurately estimate dietary intake [73,74] and contain less reporting bias than diet records [73,75]. The recalls will be entered into NDS-R (version 2017, Minneapolis, MN) for macro- and micronutrient analysis. Specific details on current method of infant feeding will be collected (e.g., breastfeeding, formula) and introduction of solids will be assessed at each study visit.

2.10. Power calculation

The primary aim is to determine if the effect of GWG status (excessive vs non-excessive) is modified by the prenatal dose of DHA (1000 mg/day vs 200 mg/day) to influence total infant FM at 24 months of age. We performed a power analysis and found a total of n = 120 would be required to answer the primary aim. Of the total n = 120, we anticipate n = 54 will experience non-excessive GWG (45%) and n = 66 will experience excessive GWG (56%) [76] with each approximately equally allocated to low and high DHA supplementation groups. For the given effect size (using population mean differences of −168 vs 304 g and a standard deviation within group of 442 g), the power is 0.815. This means that 81.5% of studies would be expected to yield a significant effect to detect an interaction, rejecting the null hypothesis that the two population means are equal. The test is two-sided. Allowing for 20% attrition, the sample size increases to n = 150.

2.11. Statistical analyses

Using a two-way analysis of variance (ANOVA), we will determine if prenatal DHA supplementation level (200 mg/day vs. 1000 mg/day) modifies the relationship between maternal weight gain status (excessive vs. non-excessive) and offspring FM at 24 months of age (dependent variable). We will test for a statistical interaction between GWG group and DHA dose. The specific interaction from this ANOVA will test the contrast to investigate if the supplement of 1000 mg DHA/day compared to 200 mg DHA/day during pregnancy can reduce infant FM at 24 months and specifically if this reduction is greater in offspring exposed to excessive GWG versus non-excessive GWG. A two-sided t-test statistic (alpha = 0.05) is calculated from the appropriate contrast from the two-way ANOVA (GWG and DHA).

Using a one-way ANOVA, we will determine if DHA supplementation level (200 mg/day vs. 1000 mg/day) impacts offspring central FM at 24 months. In addition, we will determine if sex modifies the relationship between DHA supplementation level (200 vs. 1000 mg) and offspring central FM at 24 months of age. A two-way ANOVA will be used to test for a statistical interaction between DHA dose and offspring sex on 24-month central FM (dependent variable). We will test for a statistical interaction between sex and DHA dose. The specific interaction from this ANOVA will test the contrast to investigate if the supplement of 1000 mg DHA/day compared to 200 mg DHA/day during pregnancy can reduce infant FM at 24 months and specifically if this reduction is greater in female offspring exposed prenatally to a high dose of DHA versus female offspring exposed to a low dose of DHA. A two-sided t-test statistic (alpha = 0.05) is calculated from the non-excessive contrast from the two-way ANOVA (gender and DHA).

We will investigate the impact of confounders in the statistical analyses proposed for aim 1 and aim 2. We propose the following to be confounders of the association between GWG status and DHA exposure on offspring FM at 24 months: maternal age, self-reported race/ethnicity, maternal education, maternal occupation, maternal prepregnancy BMI, GWG, smoking status, marital status, parity, prenatal diet, prenatal maternal omega-6 RBC fatty acid levels, maternal DHA levels at study entry and at delivery, infant diet, length of time breastfeeding, child age at assessment, infant gestational age, and child sex.

Lastly, we will perform a linear mixed model with multiple confounders as predictors to explore differences between the high vs. low supplement groups for infant FM and distribution. The dependent variable will be infant FM or infant central FM and the independent variable will be DHA group (fixed effect) and time (4 visits: 2 wks, and 6, 12, and 24 months). The model will include the fixed effects of treatment, time, and the treatment-time interaction, with the subject effect treated as random to account for the dependence among repeated observations. This analysis allows for unbalanced data (allows for missing data). We will use exploratory analyses to investigate confounders already discussed in the prior paragraph. The longitudinal model can accommodate certain types of missing data patterns that are ignorable, however, when the missing data pattern is non-ignorable, we will perform multiple imputation via the Bayesian model which uses Markov chain Monte Carlo that uses subject demographics. Significance will be set at p ≤ 0.05.

3. Discussion

Observational data have shown a beneficial impact of prenatal DHA intake on offspring body composition. Higher maternal prenatal n-3 status has been related to lower offspring skinfold thickness [33], central FM³², and increased fat-free mass (FFM) at 4 and 6 years old [34]. Vidakovic et al. [35] measured maternal prenatal DHA status and used DXA to measure offspring body composition and distribution at 6 years old. They associated a higher maternal prenatal DHA status with lower offspring FM and a more favorable AT distribution. Even though relationships were found with FM, no relationships were found between maternal DHA and offspring BMI, suggesting BMI was not a sensitive marker of fat accrual in children.

There have been RCT completed where prenatal or postnatal DHA was supplemented, but body composition was not the primary outcome. Two systematic reviews [77,78] and one meta-analysis [78] have been published summarizing these data. All reported RCT prenatally supplementing DHA have been completed outside of the U.S. and over a decade ago [59,60]. Six studies provided prenatal DHA supplementation and continued supplementation into the postnatal period [33,79-83] and four trials provided prenatal supplementation only (400 mg/day-920 mg/day DHA supplementation) [84-87]. Of the six trials, two studies found DHA supplementation reduced offspring BMI [33,82], one found DHA supplementation increased offspring BMI [83], and three found no effect [33,80,81]. In the trials providing prenatal DHA supplementation only, three found no difference in offspring adiposity measures between groups [86-88]. Of note, in studies finding no difference, adiposity was measured by BMI, weight/length, or skinfold thickness. The fourth trial found differences at 18 months [84] but not at 5 years [85].

There are potential reasons for these disparate findings. First, body composition was not measured using sophisticated techniques [88]. While skinfold measures are attractive to use due to ease in measurement, they are best suited to estimate regional fatness rather than relative fatness [89]. Second, analyses did not always control for maternal GWG, because it was not collected in the primary study [90]. Maternal GWG, especially when excessive, is strongly related to offspring FM in infancy [16,91], at 3 years old, 5 years old [1], 10 years old [22] and in adulthood [24]. Third, some study populations already had an adequate DHA status from consumption of nutritional supplements that provided DHA and data were lacking for maternal baseline DHA status or the change in maternal DHA when taking the supplement [88,90]. If the population already had adequate DHA status or were consuming DHA supplements, an effect of DHA on offspring body composition would not have been expected.

Only one prenatally supplemented RCT has been reported where the primary outcome was offspring body composition [79]. The design was an open label trial, supplementing 1020 g/day DHA + 180 g/day EPA (n = 104) vs. a control diet (n = 104) taking place in Germany between the years of 2006 to 2009. No direct measure of body composition was used, instead an indirect measure, sum of 4 skinfolds, was used to assess total adiposity and abdominal ultrasound was used to assess differences in AT distribution. No differences were found at birth or through 1 year old. A subset (n = 110) returned for follow-up at 5 years old and no differences were found in any measures [92].

There are several potential reasons for null findings. First, skinfold measures were used to assess changes in adiposity and ultrasound was used to assess regional AT depots. The acceptability of ultrasound to assess regional AT depots remains to be proven and accepted [93]. Second, the sample had a low pre-pregnancy BMI of 22.4 ± 3.0 kg/m². Women with a normal weight BMI are less likely to gain excessive weight during pregnancy [76] and their offspring have less fat accrual from birth to 6 years old [16,94], therefore, they are at a lower risk of obesity development. Third, GWG was 15.1 ± 4.8 kg, with no report on percentage of women who gained excessively, and no control for GWG in the statistical analyses. This is important because GWG is strongly related to offspring fat accrual [1,16,22,24,91]. Fourth, the sample was relatively well educated, likely reflecting a more health-conscious sample and not likely to have offspring at risk for overweight or obesity. Fifth, at the 5-year follow-up visit, there was significant attrition (45%; original cohort n = 208; returning at 5 yr n = 114) and there is likely inadequate power to report group differences. In a subset of those returning (n = 42), they used MRI to measure AT depots with no difference detected. But with this even smaller sample, the ability to detect group differences was likely very limited as was their ability to consider effect modifiers.

Our study has both strengths and limitations. A strength of the study is leveraging maternal data that was vetted and collected as part of a prior NIH funded trial. Second, we are collecting longitudinal body composition data during a window of time (birth to two years) where limited data are available. Many datasets contain body composition measured during the first six months of life but not again until later in childhood, due to difficulties and limited techniques validated to measure body composition across these early ages. Our study also has limitations. The study was not designed to collect maternal postnatal diet or breastmilk samples to assess the impact of the postnatal period on our outcomes. We are collecting dietary recalls in the infants and infant feeding status (exclusively breastfed, mixed fed, or exclusively formula fed) and duration of breastfeeding and these variables will be controlled for in the data analyses. In calculation of maternal pre-pregnancy BMI and GWG, we are using maternal self-reported weight prior to pregnancy. Self-report could introduce bias, however, good agreement has been reported between maternal recall of pre-pregnancy weight and medical records for pre-pregnancy weight such that recall of maternal pre-pregnancy weight is considered a satisfactory substitute when chart extraction is not possible [95,96].

4. Conclusion

Effective strategies to prevent the development of obesity are lacking. Exposures during the prenatal period are important in the establishment of the offspring phenotype. However, it is largely unknown which exposures can be successfully targeted to have a meaningful impact. There is conflicting evidence for the effect of prenatal DHA supplementation on offspring adiposity accrual. Published data has come from trials where offspring body composition was not the primary outcome, conducted in populations not at high risk for obesity development, or simple measures of body composition were used, potentially missing an effect of DHA supplementation. Data coming from well-run clinical trials in pregnancy cohorts with longitudinal follow up are needed to understand the long-term impact of manipulation of prenatal DHA supplementation on the offspring phenotype. Therefore, this study will determine if prenatal DHA supplementation modifies the relationship between maternal weight gain and offspring FM and FM distribution at 24 months of age.

Data availability

Regarding sharing materials and managing of intellectual property, we will adhere to the NIH Grant Policy on Sharing of Unique Research Resources including the Sharing of biomedical Research Resources Principles and Guidelines for Recipients of the NIH Grants and Contracts.