Longitudinal Associations of Plasma Phospholipid Fatty Acids in Pregnancy with Neonatal Anthropometry: Results from the NICHD Fetal Growth Studies—Singleton Cohort

Despite increasing interest in the health effects of polyunsaturated FAs (PUFAs), their roles in fetal and neonatal growth remain understudied. Within the NICHD Fetal Growth Studies—Singleton Cohort, we prospectively investigated the associations of individual and subclasses of plasma phospholipid PUFAs at gestational weeks (GW) 10–14, 15–26, 23–31, and 33–39 with neonatal anthropometric measures as surrogates for fetal growth among 107 women with gestational diabetes mellitus (GDM) and 214 non-GDM controls. Multivariable weighted linear regression models estimated the associations between plasma phospholipid PUFAs and neonatal anthropometric measures. Adjusted beta coefficients for phospholipid docosahexaenoic acid (DHA) per standard deviation (SD) increase at GW 23–31 in association with birthweight z-score, neonatal length, and neonatal fat mass were 0.25 (95% CI: 0.08–0.41), 0.57 (0.11–1.03) cm, and 54.99 (23.57–86.42) g, respectively; all false discovery rates (FDRs) < 0.05. Estimated Δ5-desaturase activity per SD increase at GW 33–39 but not at other time points was positively associated with birthweight z-score: 0.29 (95% CI: 0.08–0.33); neonatal length: 0.61 (0.29–0.94) cm; and neonatal fat mass: 32.59 (8.21–56.96) g; all FDRs < 0.05. Longitudinal analysis showed consistent results. Our findings suggest that mid-to-late pregnancy presented as critical windows for primarily diet-derived DHA and Δ5-desaturase activity in relation to neonatal anthropometric measures.


Introduction
Optimal intake of nutrients during pregnancy is important not only for growth and development in utero but also in the prenatal period and later life [1][2][3]. Despite the fact that maternal glucose is the dominant fuel source for the fetus [4], it is not the sole driver of fetal development [5]. Other contributors, such as maternal lipids, are available to the fetus due to the presence of placental lipoprotein receptors, lipoprotein lipase, and fatty-acid-binding protein, suggesting the need to further evaluate the role of diverse maternal fuels in birth outcomes [6,7].
Despite increasing interest in the health benefits of polyunsaturated FAs (PUFAs) [8], the roles of the two classes of plasma phospholipid PUFAs, n-3 and n-6 PUFAs, in fetal growth and development remain understudied. In addition, emerging evidence indicates that individual phospholipid FAs may have distinct nutritional and physiologic roles in human metabolism [9][10][11][12]. Prenatal PUFA concentrations have been linked to the duration of pregnancy, neonatal respiratory distress syndrome, and childhood adiposity [13][14][15]. To date, data on the associations of maternal plasma phospholipid PUFA composition in pregnancy with neonatal birth size and adiposity have been limited and inconsistent [16,17]. Further, nascent evidence suggests that fetal growth may be influenced by maternal lipids to a different extent depending on the timing of exposure during gestation [18][19][20]. However, previous data on maternal lipid levels and fetal growth have been limited to cross-sectional studies with a one-point-in-time measurement of triglyceride levels, mostly in late pregnancy [21][22][23]. As such, longitudinal data on maternal individual phospholipid PUFA concentrations throughout pregnancy are needed to improve our understanding of the impact of intrauterine nonglycemic nutrient supplies on neonatal anthropometric measures as surrogate measures for intrauterine growth and fetal development.
To address these knowledge gaps, we aimed to investigate the longitudinal and prospective associations of individual and subclasses of plasma phospholipid PUFA levels throughout pregnancy with neonatal anthropometric measures and to explore the sensitive window of exposure to plasma phospholipid PUFAs in pregnancy in relation to neonatal size and adiposity.

Study Sample
This study used data from low-risk pregnancies within the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Fetal Growth Study-Singleton Cohort (approval code: 09-CH-N152) [24,25]. The study enrolled a total of 2802 pregnant women in gestational weeks 8-13 at 12 U.S. clinical centers from 2009 to 2013. The study recruited pregnant women aged 18-40 years who were free of preexisting diseases, such as hypertension, diabetes mellitus, cancer, human immunodeficiency virus, or acquired immunodeficiency syndrome. All participating clinical sites and the NICHD obtained approval from their respective institutional review boards. All participants provided written and informed consent.
This study used biomarker data measured in a nested case-control study, the primary aim of which was to assess the associations between biomarkers in early to mid-pregnancy and the subsequent risk of gestational diabetes (GDM) [12]. The secondary aim of the GDM case-control study was to assess associations between biomarkers and fetal growth, as assessed by neonatal anthropometric measures. This study leveraged biomarker data in 321 women (107 GDM cases and 214 non-GDM controls) from the GDM case-control study. The GDM cases were identified via medical record review according to the Carpenter and Coustan criteria [26]. Each GDM case was individually matched 1:2 to non-GDM controls (n = 214) by age (±2 years), race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, and Asian/Pacific Islander), and gestational age at blood sample collection (±2 weeks).

Neonatal Anthropometric Measures
Birthweight and gestational age at delivery were abstracted and calculated from neonatal medical records. Neonatal measurements were taken 12-24 h after delivery and obtained in duplicate. Infants born very preterm (≤32 weeks) and moderately preterm (33-36 weeks) were measured at 32 completed weeks of gestation-corrected age and when stabilized, respectively. Birthweight z-score was calculated based on the sex-and gestational-age-specific U.S. national reference percentiles of birthweight [32]. Neonatal length in cm, the distance from the infant's feet to the top of the head, was measured 12-24 h after delivery by verified study personnel using an infantometer. Skinfold measurements in cm were taken using a Lange skinfold caliper on the right side of the body at the abdominal flank, anterior thigh, subscapular, and triceps. The measurements were summed as an indicator for neonatal adiposity [33]. Fat mass was estimated using the method developed by Catalano et al.: fat mass = 0.39055 × birthweight + 0.0453 × flank skinfold − 0.03237 × birth length + 0.54657 [34].

Covariates
Maternal demographic, lifestyle, and clinical factor data were collected from medical records and structured questionnaires. Covariates were selected a priori: maternal age (years), race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, and Asian), education (high school or less, some college/associate degree, and four-year college degree Nutrients 2022, 14, 592 4 of 14 or higher), nulliparity (yes/no), and pre-pregnancy body mass index (underweight: <25.0; normal: 25.0-29.9; overweight: 30.0-34.9; obese: 35.0-44.9 kg/m 2 ), gestational weight gain up to the respective visit (continuous), gestational age at blood collection (weeks), gestational age at delivery (weeks), and postnatal days at neonatal assessment (continuous; for all models except birthweight). In this low-risk population, women without obesity who smoked during the 6 months preceding the index pregnancy were ineligible, and only five women with obesity reported smoking during the 6 months before pregnancy. Thus, smoking was not included as a covariate.

Statistical Methods
For all analyses, the matched case-control sample was reweighted to represent the entire original cohort. The sampling weights were created via an inverse likelihood approach [35]. GDM subjects had a sampling probability of 1. The sampling probability of each selected non-GDM control was calculated using logistic regression among all non-GDM women in the entire cohort, including matching factors for control selection (age, race/ethnicity, and gestational week at blood collection). We used bootstrapping with 200 replicates to confirm the variance of our weighted models and observed similar findings.
For the primary analyses, we treated sex-and gestational-age-specific birthweight z-score, neonatal length, and neonatal fat mass as continuous variables. Individual PUFAs and PUFA ratios were each analyzed as a continuous variable per standard deviation (SD). Multivariable linear regression models were fitted to assess the temporal associations of individual PUFAs, subclasses of PUFAs (i.e., n-3 and n-6 PUFAs), and PUFA ratios at each of the four visits (gestational weeks 10-14, 15-26, 23-31, and 33-39 weeks) with neonatal anthropometric measures, adjusting for the aforementioned covariates. Further, the longitudinal trends of plasma phospholipid PUFAs and ratios during pregnancy were assessed by fitting generalized linear mixed models with participant-specific random intercepts, an autoregressive covariance structure, a random effect for the matched case-control pairs, and an interaction term of a cross product between concentrations of plasma phospholipid PUFAs and sample collection time. The Benjamini-Hochberg false discovery rate (FDR)-controlling method was used as the post hoc adjustment for multiple comparisons.
We further explored whether pre-pregnancy obesity status modified the associations of individual plasma phospholipid PUFA and PUFA ratios with neonatal anthropometric measures. P for interaction was obtained using the likelihood ratio test. An interaction effect was considered significant if the p value was <0.10, recognizing the limitations in statistical power when testing for interactions in multiplicative models [36][37][38][39]. Models stratified by pre-pregnancy obesity status were adjusted for the aforementioned covariates, including pre-pregnancy BMI as a continuous variable to account for any residual confounding within each stratum. We conducted a sensitivity analysis by excluding the five women with obesity who reported smoking during the 6 months before pregnancy. All analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA) and R version 3.0.2 (Vienna, Austria).

Discussion
This longitudinal study of data from the prospective NICHD Fetal Growth Studies-Singleton Cohort provided insights into 11 individual plasma phospholipid PUFAs and 3 PUFA product-to-precursor ratios throughout gestation in relation to neonatal anthropometric measures. We found both time-specific and prospective longitudinal associations of plasma phospholipid DHA and estimated ∆5-desaturase activity with neonatal anthropometric measures. Specifically, the primarily diet-derived plasma phospholipid DHA at gestational weeks 23-31 was positively associated with birthweight z-score, neonatal length, and neonatal fat mass; however, other primarily diet-derived plasma phospholipid PUFAs, including n-3 EPA and n-6 LA, were not significantly associated with any of the neonatal anthropometric measures. Among the PUFA ratios, only the estimated ∆5-desaturase activity at gestational weeks 33-39 was positively associated with birthweight z-score, neonatal length, and neonatal fat mass.

Comparison with Studies on DHA and Neonatal Anthropometry
The results of previous studies on the plasma level of DHA have shown inconclusive findings regarding its associations with neonatal anthropometric measures. Rump et al. reported that lower maternal plasma phospholipid DHA levels at gestational weeks 16 weeks or earlier and umbilical cord plasma concentrations of DHA were both associated with heavier newborns in a sample of primarily Caucasian mother-infant dyads in the Netherlands [40]. In contrast, Bernard et al. found no association between maternal plasma DHA levels at gestational weeks 26-28 and birthweight among Asians [41]. We found that plasma phospholipid DHA levels at gestational weeks 23-31 were positively associated with birthweight z-score, neonatal length, and neonatal fat mass, suggesting that higher maternal plasma phospholipid DHA levels in late pregnancy may stimulate fetal growth among this multiracial/ethnic population of pregnant women with low-risk obstetrical profiles. The inconsistent findings could be attributable to differences in population characteristics, time window of exposure measurement, and varied degrees of covariate adjustment.
When evaluated longitudinally throughout pregnancy, plasma phospholipid DHA levels were no longer associated with neonatal fat mass, which may reflect an averaged-out effect due to the lack of associations between DHA at other periods of pregnancy (gestational weeks 10-14, 15-26, and 33-39) and neonatal fat mass. Given that fetal adipose depot tends to accumulate particularly in late pregnancy [42], the time-specific and longitudinal approaches suggest that 23-31 weeks of gestation could be a sensitive time window of exposure to DHA in utero in relation to fetal growth.

Comparison with Studies on Estimated ∆5-Desaturase Activity and Neonatal Anthropometry
We observed positive associations, from both time-specific and longitudinal analyses, between estimated ∆5-desaturase activity levels in late pregnancy (gestational weeks [33][34][35][36][37][38][39] and higher birthweight z-score, neonatal length, and neonatal fat mass. A recent study in Japan reported that the maternal ∆5-desaturase index at gestational weeks 24-30 did not vary among adequate-, small-, and large-for-gestational-age infants, whereas the cord blood erythrocyte ∆5-desaturase index was higher among small-versus adequate-forgestational-age infants [43]. The different findings could be attributed to differences in study population characteristics, with a much lower median pre-pregnancy BMI in Japanese women compared to our study sample (20.5-21.9 vs. 24.6 kg/m 2 ). Further, given that fetal sources of DHA are mainly through placenta-mediated transfer, these findings suggest that small-for-gestational-age infants may be aggressively synthesizing DHA based on the enhanced ∆5-desaturase activity to meet developmental needs, highlighting the important role of DHA and ∆5-desaturase in fetal growth and development.
Although some studies reported that concentrations of n-6 PUFA were inversely associated with birthweight [44,45], we did not find any significant associations of plasma phospholipid n-6 PUFA levels with neonatal growth in our time-specific analysis. Furthermore, after stratifying by obesity status, more n-6 PUFAs showed significant associations with neonatal anthropometric measures, highlighting that pre-pregnancy obesity status may be a potential effect modifier for the role of maternal n-6 PUFA levels in fetal growth. Monthé-Drèze et al. showed that women with pre-pregnancy obesity had higher concentrations of n-6 PUFAs and an attenuated response to n-3 PUFA supplementation [46]. The Dutch Generation R cohort also reported higher plasma n-6 PUFA levels around 20.5 weeks of gestation among women with pre-pregnancy obesity compared to women without pre-pregnancy obesity [47]. Notably, n-6 PUFAs are considered more proinflammatory compared to n-3 PUFAs [48,49]. The optimal balance of n-3 and n-6 PUFAs has been shown to reduce obesity-related inflammation [50,51]. Future investigations on the potential interplay among plasma phospholipid PUFAs, maternal obesity status, and inflammatory markers are warranted to improve our understanding of the obesity-specific associations between PUFAs and neonatal anthropometric measures.

Biological Plausibility and Implications
The biological mechanisms underlying the positive associations of maternal DHA and estimated ∆5-desaturase activity with neonatal anthropometric measurements among a sample of low-risk pregnant women remain to be elucidated. Plasma phospholipid DHA is primarily derived from dietary sources, such as oily fish [52]. In animal models, DHA supplementation can ameliorate the altered lung development due to intrauterine growth restriction [53]. In clinical trials, DHA supplementation has been positively linked to longer gestation duration and larger infant size [54], consistent with our observational data on newborn size. ∆5-Desaturase is one of the key enzymes that regulate the metabolism of PUFAs in humans and is necessary for DHA synthesis [52]. Our findings suggest that the enhanced activity of ∆5-desaturase (DGLA to AA) may stimulate fetal growth among our study population of low-risk pregnant women. Our results also indicate that late pregnancy may be a critical time window for the role of plasma phospholipid DHA and ∆5-desaturase in fetal growth. Indeed, fetal weight increases substantially and fetal brain growth accelerates in late pregnancy [55,56].

Strengths and Limitations
Our study has notable strengths. First, the prospective and longitudinal data analysis at four timepoints allowed examination of the temporal relationships between plasma phospholipid PUFA levels and enzyme activity in early to late pregnancy and neonatal anthropometric measures. Second, the objective measurement of plasma phospholipid PUFA levels allowed us to assess the associations of individual and subclasses of circulating PUFAs with measures of neonatal anthropometry. Therefore, our findings may shed light on previous inconsistent inferences concerning dietary intakes (including both foods and supplementation) of PUFAs in relation to neonatal size, which has inevitably been subject to measurement errors of dietary assessment via subjective reports. Last, our study sample was drawn from a socio-demographically diverse cohort of women across the United States, increasing the generalizability of our findings.
Several potential limitations merit discussion. Despite the significant associations observed between individual PUFAs and neonatal anthropometric measures, we cannot exclude the possibility of the relatively modest sample size causing underestimation of the significance of true associations due to statistical power. A larger sample size in future studies is needed to validate our findings. The generalization of our findings to women with higher-risk obstetrical profiles remains to be established; however, inclusion of overall healthy women may minimize reverse causality and residual confounding due to preexisting complications and unhealthy behaviors. Given that concentrations of plasma phospholipid PUFAs are a function of exogenous and endogenous sources, we cannot disentangle the specific contributions of exogenous (e.g., dietary intakes including both foods and supplementation) versus endogenous factors (e.g., genetics and biochemistry). However, the primary goal of the present study was to examine the overall circulating plasma PUFA status in association with neonatal anthropometrics. Future research examining the specific contributions of various sources to circulating levels of individual PUFAs is warranted. Finally, our findings are based on an observational study within a prospective cohort of pregnant women; further intervention studies are warranted to confirm our findings and the causal relationship.

Conclusions
Our findings suggest potentially important roles of plasma phospholipid DHA and ∆5-desaturase activity in fetal growth and development. Our findings also provide new insights into the specific time window during which PUFAs may particularly impact fetal growth, with neonatal anthropometrics as surrogate measures. More significant associations of plasma phospholipid PUFAs with neonatal anthropometric measures were particularly evident in late pregnancy (gestational weeks 23-31 and 33-39) compared to earlier gestational windows, suggesting that noncarbohydrate fuels of lipids in this sensitive time window may be particularly critical for fetal growth and development. Future research in other populations is needed to confirm our findings.
Supplementary Materials: The following are available online at: https://www.mdpi.com/article/ 10.3390/nu14030592/s1, Table S1: Study sample characteristics according to weighting in the NICHD Fetal Growth Studies-Singleton Cohort; Table S2: Adjusted beta coefficients for neonatal birth size and body composition in association with individual and subclasses of maternal plasma phospholipid PUFAs and PUFA ratios per standard deviation increase during pregnancy; Table S3: Adjusted beta coefficients for neonatal birth size and body composition in association with longitudinal analysis of repeated measures of individual and subclasses of maternal plasma phospholipid PUFAs and PUFA ratios per standard deviation increase across pregnancy (n = 333); Table S4: Adjusted beta coefficients (95% CI) for neonatal birth size and body composition in association with individual and subclasses of maternal plasma phospholipid PUFAs and PUFA ratios per standard deviation during gestational weeks of 23-31 stratified by pre-pregnancy obesity status.  Data Availability Statement: Data described in the manuscript, code book, and analytic code will be available upon request pending application and approval of a data-sharing agreement.