Abstract
Importance
Previous reports have linked maternal prepregnancy obesity with low folate concentrations and child overweight or obesity (OWO) in separate studies. The role of maternal folate concentrations, alone or in combination with maternal OWO, in child metabolic health has not been examined in a prospective birth cohort.
Objective
We tested the hypotheses that maternal folate concentrations can significantly affect child metabolic health and that maternal sufficient folate concentrations can mitigate prepregnancy obesity-induced child metabolic risk.
Design
Prospective birth cohort study
Setting
The Boston Medical Center, MA, USA
Participants
This study included 1517 mother-child dyads recruited at birth from 1998–2012 and followed prospectively up to 9 years (median age: 6.2 years, range: 2–9 years).
Main Outcomes and Measures
Child BMI z-score calculated according to U.S. reference data, OWO defined as BMI≥85th percentile for age and gender, and metabolic biomarkers (leptin, insulin, and adiponectin).
Results
An “L-shaped” relationship between maternal folate concentrations and child OWO was observed: the risk of OWO was higher in the lowest quartile (Q1) as compared to Q2–Q4 with an odds ratio (OR) of 1.45 (95% confidence interval [CI], 1.13 to 1.87). The highest risk of child OWO was found among children of obese mothers with low folate concentrations (OR, 3.05, 95%CI, 1.91 to 4.86) compared to children of normal weight mothers with folate concentrations in Q2–Q4 after accounting for multiple covariables. Among children of obese mothers, their risk of OWO was associated with 43% reduction (OR, 0.57, 95%CI, 0.34–0.95) if their mothers had folate concentrations in Q2–Q4 compared to Q1. Similar patterns were observed for child metabolic biomarkers.
Conclusions and Relevance
In this urban low-income prospective birth cohort, we demonstrated an L-shaped relationship between maternal plasma folate concentrations and child OWO and the benefit of sufficient folate concentrations, especially among obese mothers. The “threshold” concentration identified in this study far exceeded the clinical definition of folate deficiency (<10.0 nmol/L), which was primarily based on the hematological effect of folate. Our findings underscore the need to establish “optimal” rather than minimal folate concentrations for preventing adverse metabolic outcomes in the offspring.
Introduction
In the U.S., more than 50% of reproductive age women are overweight or obese (OWO), with the prevalence at about 80% in non-Hispanic Black women.1 Maternal obesity has been linked to offspring obesity.2,3 This transgenerational transmission may originate in utero4 and amplify the obesity epidemic in current and future generations.5,6 However, questions remain regarding what early life factors can enhance or mitigate the adverse effects of maternal obesity on child obesity, and what underlying molecular mechanisms are involved.
Growing evidence suggests that maternal nutrition, through its impact on the fetal intrauterine environment, has a profound and life-long influence on child health.7 Among several specific nutrients that have been implicated, folate, an essential B vitamin involved in nucleic acid synthesis, DNA methylation, cellular growth,8 is particularly important. The demand for folate increases during pregnancy due to fetal and placental growth and uterus enlargement.9 Research has shown the benefits of sufficient folate on reproductive, neurodevelopmental and cardiovascular health,10–13 and, in particular, on preventing fetal neural tube defects (NTDs)12,13 and counteracting adverse reproductive effects of environmental endocrine disruptors.10 However, in the U.S., despite universal folic acid fortification of cereal grain products and the recommendation for women of reproductive age and especially pregnant women to take multivitamins,14 almost one quarter of these women have insufficient folate to prevent NTDs.15 Maternal obesity has been linked with low blood folate concentrations,16,17 and as such, obese mothers are at a higher risk of folate insufficiency compared to normal weight mothers. From this perspective, studying the individual and combined effects of maternal obesity and folate concentrations offers another possibility to counteract maternal obesity-induced metabolic risk via the optimization of maternal folate status, a relatively safe and inexpensive strategy.
Folate can improve insulin sensitivity18 and act as a superoxide scavenger.19 In animal models, methyl supplements during pregnancy can prevent transgenerational amplification of obesity.20 However, relevant data are lacking in humans. Two published observational studies based on retrospective recall of prenatal multivitamin intake did not find an association between prenatal vitamin intake and offspring adiposity,21,22 however, these studies did not measure maternal folate concentrations.
Using data on plasma folate concentrations from a well-established U.S. prospective birth cohort, we sought to test the hypotheses that maternal plasma folate concentrations can significantly affect child metabolic health, and that maternal sufficient folate concentrations can alleviate prepregnancy obesity- induced child metabolic risk as measured by child BMI, OWO and metabolic biomarkers (insulin, leptin, and adiponectin).
Methods
Study participants
This study included mother-infant pairs from the Boston Birth Cohort (BBC), which was initiated in 1998 with a rolling enrollment in Boston, MA, a predominantly urban, low-income minority population, as described previously.23 At enrollment, each mother completed a questionnaire interview that assessed prepregnancy weight, height, race/ethnicity, education, smoking, parity, perceived stress during pregnancy, prenatal multivitamin intake, and a blood draw within 48–72 hours after delivery. Since 2003, all children who were enrolled in the BBC and planned to receive primary care at the Boston Medical Center (BMC) were eligible for postnatal follow-up. A standardized questionnaire was used to assess postnatal demographic and environmental information.24
As illustrated in the flow chart (eFigure 1), of the 2891 mother-children pairs who were enrolled from 1998–2012 and followed prospectively up to 9 years, this study included 1517 mother-children pairs who completed at least one postnatal well child visit beyond age a years at the BMC; and have complete data on maternal BMI and plasma folate concentrations. Maternal demographic characteristics and birth outcomes were comparable between the samples included in and excluded from this study (eTable 1).
The study protocol was approved by the Institutional Review Boards of Boston University Medical Center, Ann & Robert H. Lurie Children’s Hospital of Chicago, and Johns Hopkins Bloomberg School of Public Health. Written informed consent was obtained from all of the study mothers.
Perinatal variables
Maternal prepregnancy weight and height were primarily based on a standard maternal questionnaire interview within 2–3 days of delivery. At that time, the study mothers were not aware of their folate status and did not know their child’s future growth trajectory. Maternal BMI was calculated as weight in kilograms divided by height in meters squared, and then categorized into three groups: normal weight (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), and obesity (≥30 kg/m2). Underweight mothers were removed from the analysis due to small sample size.
Educational attainment was classified into high school and below vs. college and above; maternal smoking during pregnancy was classified into three groups: never smoker, intermittent, or continuous smoker;25 maternal race/ethnicity were classified as Black, Hispanic, or Other (which included White, Asian, Pacific Islander, and mothers who reported more than one race). Perceived stress during pregnancy was grouped into low vs. high.26 Maternal diabetes status was classified as nondiabetic or diabetic (either pre-existing or gestational diabetes).23 Infant breastfeeding history was collected in follow-up visits and 70.5% of the questionnaire interview was completed before the child reached 2 years of age. Breastfeeding was grouped into exclusively breast, exclusively formula, or mixed breast and formula.24
Gestational age was determined by the first day of the last menstrual period and early prenatal ultrasonographic results,25 and categorized into term (≥37wks), which was further grouped into late term (≥39wks) and early term (37–38 wks), and preterm (<37wks), which was further grouped into late preterm (34–36 wks) and early preterm (<34wks).23 Information regarding Cesarean section and birthweight was abstracted from the electronic medical record (EMR).
Definitions of overweight or obesity in childhood
Child weight and height were measured by medical staff during well-child visits as documented in the EMR. BMI z-scores and percentiles were calculated using U.S. national reference data.27 OWO was defined as BMI ≥85th percentile for age and sex.
Ascertainment of plasma folate and vitamin B12 concentrations and metabolic biomarkers
Plasma folate and vitamin B12 concentrations were measured by a commercial laboratory via chemiluminescent immunoassay using a MAGLUMI 2000 Analyzer (Snibe Co., Ltd, Shenzhen, China) with inter-assay coefficient of variation of <4%.11
We assessed child plasma insulin (a marker of insulin resistance),23 leptin (a marker of adiposity),28 and adiponectin/leptin ratio (a marker of insulin sensitivity)29 using established methods. Plasma insulin and leptin concentrations were determined using sandwich immunoassays based on flow metric xMAP technology on Luminex 200 machines (Luminex Corp., Austin, TX).23 Adiponectin was measured by ELISA with an inter-assay coefficient of variation of <5.8%.
Statistical Analysis
The primary outcomes of interest included BMI z-score (continuous variable) and OWO (binary) at the last well-child visit, as well as metabolic biomarkers in early childhood. The primary predictors were plasma folate concentrations that were evaluated both as continuous variables and as categorical variables in quartiles. At first, we examined the relationship of maternal plasma folate concentrations with child BMI z-score and OWO probability using smoothing plots (PROC LOESS). Given that there was no significant difference in the risk of OWO between the Q2–Q4 quartiles, we grouped maternal folate concentrations into low (Q1) vs. adequate (Q2–Q4) in the subsequent analyses as there is no standard, clinically meaningful cutpoint for low folate concentrations relevant to metabolic outcomes. Available clinical cutoffs are mainly for anemia and vary by definition.30 Furthermore, we estimated the combined effects of maternal folate concentrations and prepregnancy BMI categories on child BMI and metabolic biomarkers z-scores using linear regression models. As plasma insulin, leptin, and adiponectin/leptin ratio all had skewed distributions, these biomarkers were log transformed before calculating age- and gender-specific z-score. As a next step, we estimated the individual and combined effects of maternal folate concentrations and prepregnancy BMI categories on child OWO using logistic regression models. We tested the interaction of maternal prepregnancy BMI categories (as a categorical variable with three levels) and folate status (as a binary variable) on child BMI and metabolic biomarker z-scores and odds of OWO by including cross-product terms in models with indicator terms for BMI categories and folate status. The effect modifications were tested with the likelihood ratio test using an a priori alpha value of 0.05. To further assess the robustness of the findings, we conducted stratified analyses by child age group (2–5 years vs. 6–9 years), race (Black only, the major race group in the BBC), and preterm vs. term birth.
Covariables were selected based on the published literature and our previous studies in the Boston Birth Cohort. In addition, we adjusted for maternal BMI categories (overweight and obese vs. normal weight) to assess an individual effect of folate. To further address residual confounders, we performed propensity score–matched sensitivity analyses31 that compared maternal low folate (<20.37nmol/L) with adequate folate (20.37–185.5nmol/L). We further adjusted for child plasma folate concentrations and examined combined effects of child folate concentrations with maternal folate on the risk of child OWO. Child age and sex were not included in the regression model because they were already accounted for when we defined the outcome variables. All statistical analyses were performed using SAS (SAS Institute), version 9.4.
RESULTS
This study was composed of 1517 mother-child pairs, including 1019 (67.2%) Black pairs and 290 (19.1%) Hispanic pairs. In all, 443 (29.2%) and 381 (25.1%) mothers were overweight and obese in prepregnancy, respectively, and a total of 590 (38.9%) children were OWO at age 2–9 years. OWO children had higher birthweights, higher rates of formula feeding, and higher rates of maternal obesity and diabetes (Table 1).
Table 1.
Characteristics of mother-child dyads in the total sample and subgroups stratified by child overweight or obesity (OWO) status in the Boston Birth Cohort (BBC)a
| Variables | Total sample | Children without OWO | Children with OWO | p-value |
|---|---|---|---|---|
| No. | 1517 | 927 | 590 | |
| Maternal characteristics | ||||
| Maternal age, mean(SD), year | 28.6(6.5) | 28.1(6.5) | 29.2(6.5) | 0.001 |
| Race/ethnicity, no.(%) | 0.106 | |||
| Black | 1019(67.2) | 607(65.5) | 412(69.8) | |
| Hispanic | 290(19.1) | 180(19.4) | 110(18.7) | |
| Other | 208(13.7) | 140(15.1) | 68(11.5) | |
| Education, no.(%) | 0.456 | |||
| High school and lower | 988(65.1) | 597(64.4) | 391(66.3) | |
| College and higher | 529(34.9) | 330(35.6) | 199(33.7) | |
| Parity, no.(%) | 0.267 | |||
| Nulliparous | 618(40.7) | 388(41.9) | 230(39.0) | |
| Multiparous | 899(59.3) | 539(58.1) | 360(61.0) | |
| Smoking, no.(%) | 0.199 | |||
| Never | 1264(83.3) | 783(84.5) | 481(81.5) | |
| Quitter | 111(7.3) | 67(7.2) | 44(7.5) | |
| Continuous | 142(9.4) | 77(8.3) | 65(11.0) | |
| Perceived stress during pregnancy, no.(%) | 0.905 | |||
| Low | 1239(81.7) | 758(81.8) | 481(81.5) | |
| High | 278(18.3) | 169(18.2) | 109(18.5) | |
| Prepregnancy BMI category, no.(%) | <0.001 | |||
| 18.5–24.9 kg/m2 | 693(45.7) | 481(51.9) | 212(35.9) | |
| 25–29.9 kg/m2 | 443(29.2) | 258(27.8) | 185(31.4) | |
| ≥30kg/m2 | 381(25.1) | 188(20.3) | 193(32.7) | |
| Pre- or gestational diabetes, no.(%) | 0.008 | |||
| No | 1345(88.7) | 838(90.4) | 507(85.9) | |
| Yes | 172(11.3) | 89(9.6) | 83(14.1) | |
| Child characteristics | ||||
| Boy, no.(%) | 765(50.4) | 465(50.2) | 300(50.9) | 0.795 |
| Age, mean(SD), year | 6.15(2.41) | 5.79(2.44) | 6.71(2.23) | <0.001 |
| Birthweight, mean(SD), g | 2980(797) | 2866(803) | 3159(753) | <0.001 |
| Gestational age, mean(SD), wk | 37.9(3.3) | 37.7(3.6) | 38.3(2.9) | <0.001 |
| Preterm birth, no.(%) | 355(23.4) | 229(24.7) | 126(21.4) | 0.133 |
| Birthweight for gestational age, no.(%) | <0.001 | |||
| SGA | 175(11.5) | 136(14.7) | 39(6.6) | |
| AGA | 1181(77.9) | 725(78.2) | 456(77.3) | |
| LGA | 161(10.6) | 66(7.1) | 95(16.1) | |
| Breastfeeding, no.(%) | 0.036 | |||
| Exclusively formula | 371(24.5) | 212(22.9) | 159(26.5) | |
| Exclusively breastfed | 76(5.0) | 55(5.9) | 21(3.6) | |
| Both | 1070(70.5) | 660(71.2) | 410(69.5) |
The BBC uses a rolling enrollment; the study sample consists of children enrolled from 1998–2012 who have been followed from birth up to the last visit recorded by electronic medical record. Childhood OWO was defined as BMI ≥85th percentile for age and sex at the last visit. P values are for the comparison between children with OWO and without OWO.
The median (interquartile range [IQR]) for maternal plasma folate concentrations was 30.8 (20.4–44.4) nmol/L. Obese mothers had lower folate concentrations compared to normal weight mothers: Geometric means (95%CI) of folate concentrations in normal weight, overweight, and obese mothers were 30.7 (29.3–32.1), 30.3 (28.6–32.0), and 27.9 (26.5–29.5) nmol/L, respectively (p for trend=0.032). Rates of low (Q1) folate among normal weight, overweight, and obese mothers were 24.0%, 24.8%, and 27.0%, respectively.
Maternal folate concentrations, prepregnancy obesity, and child OWO
Maternal plasma folate concentrations were inversely associated with child BMI z-score and OWO, but the association was nonlinear (Wald chi-square test p<0.05). As shown in Figure 1A, there was a steep rise in BMI z-score for children whose mothers’ plasma folate concentrations were below the 25th percentile (i.e., 20.37 nmol/L). However, higher maternal folate concentrations beyond the median value did not confer additional benefits. Consistently, the increased risk of child OWO was mainly concentrated in the lowest folate quartile (Figure 1B): multivariate odds ratio (OR) was 1.50 (95%CI, 1.10–2.04), as compared with the higher quartiles (Table 2). Of note, very similar results were observed before vs. after the adjustment of birthweight.
Figure 1. Relationship between maternal plasma folate concentrations and offspring BMI z-score and probability of overweight or obesity during childhood in the Boston Birth Cohort (BBC)a.
Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); OWO, overweight or obesity.
SI conversion factor: To convert folate to ng/mL, divided by 2.266.
aThe BBC uses a rolling enrollment; the study sample consists of children enrolled from 1998–2012 who have been followed from birth up to the last visit recorded by electronic medical record.
Panel A displays the crude relationship between maternal plasma folate concentration and offspring BMI z-score. Due to a small sample size, the curve is truncated at 80 nmol/L (mean BMI z-score among children with maternal plasma folate concentration >80nmol/L was 0.60 [95%CI: 0.22 to 0.98], n=51). The curve (95%CI) was derived from smoothing plots (PROC LOESS).
Panel B displays the crude relationship between maternal plasma folate concentration and the probability of overweight or obesity. Due to a small sample size, the curve is truncated at 80 nmol/L (the proportion of OWO among children with maternal folate >80nmol/L was 37.3% [95%CI: 24.1 to 51.9], n=51). The curve (95%CI) was derived from smoothing plots (PROC LOESS).
Table 2.
The individual and combined effect of maternal folate concentrations and prepregnancy BMI categories on child BMI z-score and overweight or obesity at age 2–9 years (n=1517) in the Boston Birth Cohort (BBC)a
| Maternal | N | Child BMI z-score
|
Child overweight or obesity
|
|||
|---|---|---|---|---|---|---|
| BMI | Folate | Mean (SD) | β (95%CI) | Case, n(%) | OR (95%CI) | |
| Quartileb | ||||||
| Q4 | 379 | 0.61 (1.13) | ref | 132(34.8) | 1.00 | |
| Q3 | 380 | 0.57 (1.33) | −0.10 (−0.27 to 0.07) | 147(38.7) | 1.11 (0.82 to 1.52) | |
| Q2 | 379 | 0.57 (1.27) | −0.07 (−0.24 to 0.09) | 135(35.6) | 0.99 (0.72 to 1.35) | |
| Q1 | 379 | 0.89 (1.30) | 0.21 (0.04 to 0.38) | 176(46.4) | 1.50 (1.10 to 2.04) | |
| Folate per quartile decrease | 0.06 (0.01 to 0.12) | 1.12 (1.01 to 1.23) | ||||
| Binaryb | ||||||
| Q2–Q4 | 1138 | 0.58 (1.24) | ref | 414(36.4) | 1.00 | |
| Q1 | 379 | 0.89 (1.30) | 0.27 (0.13 to 0.41) | 176(46.4) | 1.45 (1.13 to 1.87) | |
| NL | 693 | 0.41 (1.25) | ref | 212(30.6) | 1.00 | |
| OWc | 443 | 0.71 (1.20) | 0.23 (0.09 to 0.37) | 185(41.8) | 1.49 (1.15 to 1.93) | |
| OBc | 381 | 1.06 (1.26) | 0.54 (0.39 to 0.70) | 193(50.7) | 2.03 (1.53 to 2.68) | |
| P for trend | <0.001 | <0.001 | ||||
| Combined effect | ||||||
| NL | Q2–Q4 | 527 | 0.35 (1.22) | ref | 151(28.7) | 1.00 |
| Q1 | 166 | 0.59 (1.31) | 0.20 (−0.01 to 0.41) | 61(36.7) | 1.35 (0.91 to 1.98) | |
| OW | Q2–Q4 | 333 | 0.65 (1.17) | 0.22 (0.06 to 0.38) | 131(39.3) | 1.45 (1.07 to 1.96) |
| Q1 | 110 | 0.90 (1.27) | 0.46 (0.21 to 0.70) | 54(49.1) | 2.13 (1.37 to 3.30) | |
| OB | Q2–Q4 | 278 | 0.95 (1.27) | 0.48 (0.30 to 0.65) | 132(47.5) | 1.90 (1.38 to 2.62) |
| Q1 | 103 | 1.37 (1.19) | 0.89 (0.63 to 1.15) | 61(59.2) | 3.05 (1.91 to 4.86) | |
| Stratified by BMI categories | ||||||
| NL | Q1 | 166 | 0.59 (1.31) | ref | 61(36.8) | 1.00 |
| Q2–Q4 | 527 | 0.35 (1.22) | −0.17 (−0.39 to 0.05) | 151(28.7) | 0.79 (0.53 to 1.19) | |
| OW | Q1 | 110 | 0.90 (1.27) | ref | 54(49.1) | 1.00 |
| Q2–Q4 | 333 | 0.65 (1.17) | −0.23 (−0.48 to 0.01) | 131(39.3) | 0.69 (0.43 to 1.09) | |
| OB | Q1 | 103 | 1.37 (1.19) | ref | 61(59.2) | 1.00 |
| Q2–Q4 | 278 | 0.95 (1.27) | −0.48 (−0.75 to −0.21) | 132(47.5) | 0.57 (0.34 to 0.95) | |
Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); NL, normal weight; OW, overweight; OB, obesity.
SI conversion factor: To convert folate to ng/mL, divided by 2.266.
The BBC uses a rolling enrollment; the study sample consists of children enrolled from 1998–2012 who have been followed from birth up to the last visit recorded by electronic medical record. Q2–Q4 folate concentration range: 20.37–185.51nmol/L, Q1 folate concentration range: 6.64–20.36 nmol/L. Maternal prepregnancy BMI was categorized into three groups: NL: BMI 18.5–24.9 kg/m2, OW: 25–29.9kg/m2, and OB: ≥30kg/m2.
Adjusted for maternal age, race, education, smoking, parity, perceived stress during pregnancy, diabetes, plasma vitamin B12 concentration during pregnancy, infant’s gestational age category, birthweight, and breastfeeding. There was no significant interaction between maternal folate status and prepregnancy obesity (p>0.05).
Also adjusted for maternal prepregnancy BMI category.
Additional adjustment for plasma folate concentration
As shown in Table 2 and Figure 2, there was a significant combined effect of maternal folate concentrations and prepregnancy obesity on offspring BMI z-score and risk of OWO. Children of obese mothers with low folate had a 0.89 unit(95%CI, 0.63–1.15) increase in BMI z-score, and a 3.05-fold (95%CI, 1.91–4.86) increased risk of OWO compared to those whose mothers with adequate folate and normal weight. However, there was no significant interaction between maternal folate status and prepregnancy obesity (p>0.05). Given maternal obesity, children of mothers with adequate folate concentrations were associated with a 43% reduction in the risk of OWO compared to children of mothers with low folate concentrations (OR=0.57 [95%CI, 0.34–0.95], p=0.032).
Figure 2. The combined effect of maternal plasma folate status and maternal prepregnancy BMI categories on child’s BMI z-score and proportion of overweight or obesity in the Boston Birth Cohort (BBC)a.
Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); OWO, overweight or obesity; Q, quartile.
SI conversion factor: To convert folate to ng/mL, divided by 2.266.
aThe BBC uses a rolling enrollment; the study sample consists of children enrolled from 1998–2012 who have been followed from birth up to the last visit recorded by electronic medical record. Q2–Q4 folate concentration range: 20.37–185.51nmol/L, Q1 folate concentration range: 6.64–20.36 nmol/L. Maternal prepregnancy BMI was categorized into three groups: NL: BMI 18.5–24.9 kg/m2, OW: 25–29.9kg/m2, and OB: ≥30kg/m2.
In panel A, Y-axes is least square mean (95%CI) of child BMI z-score, estimated from Generalized Linear Model (GLM) with adjustment for maternal age, race, education, smoking, parity, perceived stress during pregnancy, diabetes, plasma vitamin B12 concentration during pregnancy, infant’s gestational age category, birthweight, and breastfeeding. There was no significant interaction between maternal folate status and prepregnancy obesity (p>0.05).
In panel B, Y-axes is adjusted probability (95%CI) of child overweight or obesity, estimated from logistic regression model with adjustment for above covariables. There was no significant interaction between maternal folate status and prepregnancy obesity (p>0.05).
Maternal folate concentrations, prepregnancy obesity, and child metabolic biomarkers
As compared with maternal adequate folate concentrations, low folate concentrations were associated with an increase in the concentrations of insulin and leptin, and a reduction in the adiponectin/leptin ratio in offspring. Children of obese mothers with low folate concentrations had a 0.39 unit (95%CI: 0.09–0.69) increase in plasma insulin z-scores compared to the group with adequate maternal folate concentrations and normal maternal weight. A similar pattern was seen for leptin. In contrast, the children of obese mothers with low folate concentrations were associated with a 0.43 unit (0.17–0.70) decrease in their adiponectin/leptin ratio z-scores. Given maternal obesity, children of mothers with adequate folate concentrations had a 0.27 unit (95%CI: −0.61 to 0.07) decrease in plasma insulin, a 0.32 unit (95%CI: 0.03–0.62) decrease in leptin; and a 0.43 unit (95%CI: 0.14–0.72) increase in adiponectin/leptin ratio z-scores compared to children of mothers with low folate concentrations (Table 3, eFigure 2). These associations were attenuated after further adjustment for child adiposity (eTable 2–4).
Table 3.
The individual and combined effect of maternal folate concentrations and prepregnancy BMI categories on child metabolic biomarkers in the Boston Birth Cohort (BBC)a
| Maternal | Insulin z-score (n=757) |
Leptin z-score (n=1009) |
Adiponectin/leptin ratio z-score (n=985) |
||||
|---|---|---|---|---|---|---|---|
| BMI | Folate | Mean (SD) | β (95%CI) | Mean (SD) | β (95%CI) | Mean (SD) | β (95%CI) |
| Quartileb | |||||||
| Q4 | −0.13(0.97) | ref | −0.02(0.95) | ref | −0.03(0.97) | ref | |
| Q3 | −0.05(1.10) | 0.09 (−0.12 to 0.30) | −0.08(0.95) | −0.07 (−0.24 to 0.10) | 0.17(0.95) | 0.21(0.03 to 0.38) | |
| Q2 | −0.01(1.04) | 0.11 (−0.08 to 0.31) | 0.01(0.96) | 0.03 (−0.14 to 0.20) | −0.02(0.93) | 0.00(−0.17 to 0.18) | |
| Q1 | 0.09(0.99) | 0.19 (−0.01 to 0.39) | 0.14(1.06) | 0.16 (−0.02 to 0.33) | −0.14(1.10) | −0.10(−0.27 to 0.08) | |
| Folate per quartile decrease | 0.06 (0.00 to 0.12) | 0.06 (0.00 to 0.11) | −0.05(−0.11 to 0.01) | ||||
| Binaryb | |||||||
| Q2–Q4 | −0.06(1.03) | ref | −0.03(0.95) | ref | 0.04(0.95) | ref | |
| Q1 | 0.09(0.99) | 0.12(−0.04 to 0.29) | 0.14(1.06) | 0.17(0.03 to 0.31) | −0.14(1.10) | −0.17(−0.31 to −0.02) | |
| NL | −0.10(0.99) | ref | −0.06(0.96) | ref | 0.03(0.99) | ref | |
| OWc | 0.01(1.01) | 0.16(−0.01 to 0.33) | −0.04(0.92) | −0.01(−0.18 to 0.14) | 0.01(0.95) | 0.00(−0.15 to 0.15) | |
| OBc | 0.08(1.09) | 0.20(0.01 to 0.39) | 0.20(1.08) | 0.24 (0.08 to 0.40) | −0.12(1.03) | −0.14(−0.30 to 0.02) | |
| P for trend | |||||||
| Combined effect | |||||||
| NL | Q2–Q4 | −0.13(0.98) | ref | −0.10(0.92) | ref | 0.06(0.95) | ref |
| Q1 | −0.02(1.02) | 0.04(−0.21 to 0.28) | 0.05(1.05) | 0.12(−0.09 to 0.33) | −0.04(1.11) | −0.05(−0.26 to 0.17) | |
| OW | Q2–Q4 | −0.02(1.03) | 0.13(−0.07 to 0.32) | −0.06(0.93) | −0.01(−0.18 to 0.16) | 0.04(0.92) | 0.03(−0.14 to 0.20) |
| Q1 | 0.12(0.91) | 0.28(−0.03 to 0.58) | 0.05(0.92) | 0.12(−0.13 to 0.37) | −0.08(1.01) | −0.11(−0.37 to 0.14) | |
| OB | Q2–Q4 | 0.01(1.13) | 0.15(−0.06 to 0.37) | 0.12(1.02) | 0.19(0.01 to 0.37) | −0.01(0.98) | −0.04(−0.22 to 0.14) |
| Q1 | 0.24(1.00) | 0.39(0.09 to 0.69) | 0.39(1.20) | 0.49(0.23 to 0.75) | −0.39(1.13) | −0.43(−0.70 to −0.17) | |
| Stratified by BMI categories | |||||||
| NL | Q1 | −0.02(1.02) | ref | 0.05(1.05) | ref | −0.04(1.11) | ref |
| Q2–Q4 | −0.13(0.98) | −0.04(−0.28 to 0.20) | −0.10(0.92) | −0.13(−0.34 to 0.07) | 0.06(0.95) | 0.06(−0.15 to 0.27) | |
| OW | Q1 | 0.12(0.91) | ref | 0.05(0.92) | ref | −0.08(1.01) | ref |
| Q2–Q4 | −0.02(1.03) | −0.16(−0.47 to 0.15) | −0.06(0.93) | −0.16(−0.39 to 0.08) | 0.04(0.92) | 0.14(−0.10 to 0.39) | |
| OB | Q1 | 0.24(1.00) | ref | 0.39(1.20) | ref | −0.39(1.13) | ref |
| Q2–Q4 | 0.01(1.13) | −0.27(−0.61 to 0.07) | 0.12(1.02) | −0.32(−0.62 to −0.03) | −0.01(0.98) | 0.43(0.14 to 0.72) | |
Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); NL, normal weight; OW, overweight; OB, obesity.
SI conversion factor: To convert folate to ng/mL, divided by 2.266.
The BBC uses a rolling enrollment; the study sample consists of children enrolled from 1998–2012 who have been followed from birth up to the last visit recorded by electronic medical record. Q2–Q4 folate concentration range: 20.37–185.51nmol/L, Q1 folate concentration range: 6.64–20.36 nmol/L. Maternal prepregnancy BMI was categorized into three groups: NL: BMI 18.5–24.9 kg/m2, OW: 25–29.9kg/m2, and OB: ≥30kg/m2.
Adjusted for maternal age, race, education, smoking, parity, perceived stress during pregnancy, diabetes, plasma vitamin B12 concentration during pregnancy, infant’s gestational age category, birthweight, and breastfeeding. There was no significant interaction between maternal folate status and prepregnancy obesity (p>0.05).
Additional adjustment for maternal prepregnancy BMI category.
Additional adjustment for plasma folate concentration.
Sensitivity analyses to assess robustness of the findings
The associations described above did not differ appreciably across the following strata defined by child age (2–5 years vs. 6–9 years, eTables 5–6), race (Black only, eTable 7), and gestation (preterm vs. term birth, eTable 8), or with additional controlling for other prenatal and perinatal factors including whether the index child was a planned pregnancy or not, and whether the index child was delivered by Cesarean section (eTable 9–10). A propensity-score–matched analysis showed that low maternal folate concentrations (<20.37nmol/L), as compared with adequate concentrations (20.37–185.5nmol/L), were associated with an increased risk of childhood OWO (OR, 1.37; 95% CI, 1.03–1.84) (eTable 11). When child metabolic biomarkers were limited within the first 2 years of life, the association between maternal folate concentrations and child metabolic biomarkers were not appreciably changed (eTable 12).
Among 803 study children with postnatal folate concentrations, children had higher plasma folate concentrations than their mothers, geometric mean (95%CI): 37.4(36.4–38.5) vs. 29.8(29.0–30.8) (p<0.001). If the threshold for low folate was defined as <20.37 (maternal Q1 folate cutpoint), 72(9%) of children had low folate concentrations. The association between maternal folate and offspring BMI z-score and the risk of OWO remained after further adjustment for child folate status (eTables 13–14).
Discussion
To our knowledge, this is the first prospective birth cohort study to evaluate the individual and combined effects of maternal prepregnancy BMI and plasma folate concentrations on offspring metabolic outcomes including BMI, OWO status, and metabolic biomarkers. Our findings lend further support that maternal prenatal nutritional status may play an important role in child metabolic disorders.2,3
Our data revealed a nonlinear “L-shaped” relationship: maternal plasma folate concentration in the lowest quartile was associated with increased the risk of offspring OWO. Above this threshold, higher folate concentrations did not confer additional benefit, also suggesting a “ceiling effect” of folate. The “threshold” concentration identified in this study far exceeded the clinical definition of folate deficiency (<10.0 nmol/L), which was primarily based on the hematological effect of folate.32 Our findings are in agreement with a recent paper that used a similar approach to estimate an optimal population red blood cell folate concentration for the prevention of neural tube defects.33 Taken together, these findings underscore the need to establish “optimal” rather than merely minimal folate concentrations for preventing adverse metabolic outcomes in offspring.
Our study suggests that adequate maternal folate concentrations could mitigate the adverse effects of maternal obesity on child metabolic risk. We demonstrated that, given maternal obesity, the risk of child OWO was associated with a 43% reduction if the mothers had adequate folate concentrations compared to low folate concentrations.
Our data on metabolic biomarkers lends further support to the biological plausibility of our findings. Previous mechanistic studies indicate that factors that create an unfavorable cardio-metabolic intrauterine environment such as increased insulin resistance, elevated glucose concentrations, and increased oxidative stress in obese mothers may lead to OWO in offspring.4,16,34,35 Consistently, we found that children with obese mothers who had lower folate concentrations had the most unfavorable metabolic profiles (increased insulin and leptin and decreased adiponectin/leptin ratio in childhood). These associations appear to be mediated by child adiposity. Our findings are in agreement with a previous antenatal micronutrient randomized trial study showing that in-utero exposure to a folic acid supplement reduced the risk of metabolic syndrome in childhood.18 In contrast, an Indian study (of an observational cohort) reported a positive association between maternal folate concentration and offspring homeostatic model assessment-insulin resistance (HOMA-IR).36 Notably, this study assumed a linear relationship between maternal folate concentration and offspring HOMA-IR, and did not analyze the combined effect of maternal folate and BMI on child metabolic risk. Our data suggest that inadequate maternal folate concentrations may have deleterious long-term metabolic effects beyond fetal overgrowth and child OWO, while adequate maternal folate concentrations can counterbalance the detrimental effects of maternal obesity on the offspring.
Several limitations should also be acknowledged. We used maternal plasma folate concentrations taken at 1–3 days after delivery, which is at best a proxy of folate nutrition during the 3rd trimester of pregnancy. While peri-conception and first trimester folate concentrations are important for neural tube and brain development, fetal weight gain mostly occurs in the 3rd trimester.37 Our findings support that maternal folate concentration during the 3rd trimester is important for child metabolic outcomes. Although we did not measure fetal folate concentrations, a previous study suggested a high degree of transplacental passage of maternal folate to the fetus.38 Maternal prepregnancy BMI was primarily based on self-reported height and weight, thus it may be subject to reporting bias. Nevertheless, in a subset of the study population (N=672), self-reported BMIs were compared with those taken from medical records and showed a high degree of agreement (r=0.89, p<0.001). In addition, exclusion of 1374 children for a variety of reasons may have resulted in selection bias, though the demographic characteristics were comparable with those of the included participants. Lastly, child plasma insulin, leptin and adiponectin concentrations were measured in non-fasting samples. As we discussed in our previous report,23 the timing of blood sampling occurred randomly (any time during the clinical hours), which may have introduced background noise and thus biased our results towards null.
In conclusion, we demonstrated an “L-shaped” relationship between maternal folate concentrations and child metabolic risk, suggesting both a threshold and a ceiling effect of folate. We found that low maternal folate concentrations during late pregnancy can increase child metabolic risk as measured by BMI z-score, OWO and metabolic biomarkers; and conversely, sufficient maternal folate concentrations can mitigate the detrimental effects of maternal obesity on the offspring. Our findings underscore the need to establish and ensure “optimal” rather than minimal maternal folate concentrations for preventing offspring adverse metabolic outcomes, especially among obese mothers.
Supplementary Material
Acknowledgments
We wish to thank all of the study participants and the Boston Medical Center Labor and Delivery Nursing Staff for their support and help with the study. We are also grateful for the dedication and hard work of the field team at the Department of Pediatrics, Boston University School of Medicine.
Source of Funding: The Boston Birth Cohort (the parent study) is supported in part by the March of Dimes PERI grants (20-FY02-56, #21-FY07-605), and the National Institutes of Health (NIH) grants (R21ES011666, R21HD066471, and R01HD041702). The follow-up study is supported in part by the NIH grants (U01AI090727, R21AI079872, and R01HD086013), and the Maternal and Child Health Bureau grant (R40MC27443). Dr. Cuilin Zhang is supported by the intramural research program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH.
Role of the Sponsor: The sponsors had no role in the design and/or conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, and approval of the manuscript.
Footnotes
Author contribution:
Dr. X Wang is the principal investigator of the Boston Birth Cohort (the parent study), and has full access to all of the study data and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: G Wang, X Wang,
Acquisition of data: Pearson, Chen, Y Ji, Zuckerman, G Wang, X Wang
Analysis and interpretation of data: G Wang, Hong, Mistry, Paige, Cheng, Z Ji, H Ji, Hu, Zhang, Ren, Hou, X Wang
Drafting of the manuscript: G Wang, Cheng, Bartell, X Wang
Critical revision of the manuscript for important intellectual content: Hong, Pearson, Caruso, Paige, Z Ji, Zuckerman, Cheng, Mistry, Ren, Hou, Hu, Zhang, Bartell, X Wang
Data management and statistical analysis: Hong, Z Ji, H Ji, Caruso, Pearson, G Wang, X Wang
Administrative, technical, and/or material support: Chen, Y Ji, Caruso, Pearson
Study supervision: X Wang
Funding: X Wang
Conflicts of interests: The authors have declared that no conflicting interests exist.
References
- 1.Ogden C, Carroll M, Kit B, Flegal K. Prevalence of Childhood and Adult Obesity in the United States, 2011–2012. Jama. 2014;11(8):806–814. doi: 10.1001/jama.2014.732. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hochner H, Friedlander Y, Calderon-Margalit R, et al. Associations of maternal prepregnancy body mass index and gestational weight gain with adult offspring cardiometabolic risk factors: the Jerusalem Perinatal Family Follow-up Study. Circulation. 2012;125(11):1381–1389. doi: 10.1161/CIRCULATIONAHA.111.070060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lawlor DA, Fraser A, Lindsay RS, et al. Association of existing diabetes, gestational diabetes and glycosuria in pregnancy with macrosomia and offspring body mass index, waist and fat mass in later childhood: findings from a prospective pregnancy cohort. Diabetologia. 2010;53(1):89–97. doi: 10.1007/s00125-009-1560-z. [DOI] [PubMed] [Google Scholar]
- 4.Lawlor DA. The Society for Social Medicine John Pemberton Lecture 2011. Developmental overnutrition–an old hypothesis with new importance? Int J Epidemiol. 2013;42(1):7–29. doi: 10.1093/ije/dys209. [DOI] [PubMed] [Google Scholar]
- 5.Dabelea D, Crume T. Maternal environment and the transgenerational cycle of obesity and diabetes. Diabetes. 2011;60(7):1849–1855. doi: 10.2337/db11-0400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Frias AE, Grove KL. Obesity: a transgenerational problem linked to nutrition during pregnancy. Seminars in reproductive medicine. 2012;30(6):472–478. doi: 10.1055/s-0032-1328875. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Barker DJ, Gluckman PD, Godfrey KM, Harding JE, Owens JA, Robinson JS. Fetal nutrition and cardiovascular disease in adult life. Lancet. 1993;341(8850):938–941. doi: 10.1016/0140-6736(93)91224-a. [DOI] [PubMed] [Google Scholar]
- 8.Relton CL, Pearce MS, Parker L. The influence of erythrocyte folate and serum vitamin B12 status on birth weight. Br J Nutr. 2005;93(5):593–599. doi: 10.1079/bjn20041395. [DOI] [PubMed] [Google Scholar]
- 9.Greenberg JA, Bell SJ, Guan Y, Yu YH. Folic Acid supplementation and pregnancy: more than just neural tube defect prevention. Rev Obstet Gynecol. 2011;4(2):52–59. [PMC free article] [PubMed] [Google Scholar]
- 10.Ouyang F, Longnecker MP, Venners SA, et al. Preconception serum 1,1,1-trichloro-2,2,bis(p-chlorophenyl)ethane and B-vitamin status: independent and joint effects on women’s reproductive outcomes. Am J Clin Nutr. 2014;100(6):1470–1478. doi: 10.3945/ajcn.114.088377. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Huo Y, Li J, Qin X, et al. Efficacy of folic acid therapy in primary prevention of stroke among adults with hypertension in China: the CSPPT randomized clinical trial. Jama. 2015;313(13):1325–1335. doi: 10.1001/jama.2015.2274. [DOI] [PubMed] [Google Scholar]
- 12.Prevention of neural tube defects: results of the Medical Research Council Vitamin Study. MRC Vitamin Study Research Group. Lancet. 1991;338(8760):131–137. [PubMed] [Google Scholar]
- 13.Czeizel AE, Dudas I. Prevention of the first occurrence of neural-tube defects by periconceptional vitamin supplementation. N Engl J Med. 1992;327(26):1832–1835. doi: 10.1056/NEJM199212243272602. [DOI] [PubMed] [Google Scholar]
- 14.Crider KS, Bailey LB, Berry RJ. Folic acid food fortification-its history, effect, concerns, and future directions. Nutrients. 2011;3(3):370–384. doi: 10.3390/nu3030370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.News From the Centers for Disease Control and Prevention: Higher Folic Acid Intake Needed. Jama. 2015;313(11):1094. [Google Scholar]
- 16.Tinker SC, Hamner HC, Berry RJ, Bailey LB, Pfeiffer CM. Does obesity modify the association of supplemental folic acid with folate status among nonpregnant women of childbearing age in the United States? Birth Defects Res A Clin Mol Teratol. 2012;94(10):749–755. doi: 10.1002/bdra.23024. [DOI] [PubMed] [Google Scholar]
- 17.da Silva VR, Hausman DB, Kauwell GP, et al. Obesity affects short-term folate pharmacokinetics in women of childbearing age. Int J Obes (Lond) 2013;37(12):1608–1610. doi: 10.1038/ijo.2013.41. [DOI] [PubMed] [Google Scholar]
- 18.Stewart CP, Christian P, Schulze KJ, Leclerq SC, West KP, Jr, Khatry SK. Antenatal micronutrient supplementation reduces metabolic syndrome in 6- to 8-year-old children in rural Nepal. J Nutr. 2009;139(8):1575–1581. doi: 10.3945/jn.109.106666. [DOI] [PubMed] [Google Scholar]
- 19.Doshi SN, McDowell IF, Moat SJ, et al. Folate improves endothelial function in coronary artery disease: an effect mediated by reduction of intracellular superoxide? Arterioscler Thromb Vasc Biol. 2001;21(7):1196–1202. doi: 10.1161/hq0701.092000. [DOI] [PubMed] [Google Scholar]
- 20.Waterland RA, Travisano M, Tahiliani KG, Rached MT, Mirza S. Methyl donor supplementation prevents transgenerational amplification of obesity. Int J Obes (Lond) 2008;32(9):1373–1379. doi: 10.1038/ijo.2008.100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Dougan MM, Willett WC, Michels KB. Prenatal vitamin intake during pregnancy and offspring obesity. Int J Obes (Lond) 2015;39(1):69–74. doi: 10.1038/ijo.2014.107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lewis SJ, Leary S, Davey Smith G, Ness A. Body composition at age 9 years, maternal folate intake during pregnancy and methyltetrahydrofolate reductase (MTHFR) C677T genotype. Br J Nutr. 2009;102(4):493–496. doi: 10.1017/S0007114509231746. [DOI] [PubMed] [Google Scholar]
- 23.Wang G, Divall S, Radovick S, et al. Preterm birth and random plasma insulin levels at birth and in early childhood. Jama. 2014;311(6):587–596. doi: 10.1001/jama.2014.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hong X, Wang G, Liu X, et al. Gene polymorphisms, breast-feeding, and development of food sensitization in early childhood. J Allergy Clin Immunol. 2011;128(2):374–381 e372. doi: 10.1016/j.jaci.2011.05.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wang X, Zuckerman B, Pearson C, et al. Maternal cigarette smoking, metabolic gene polymorphism, and infant birth weight. Jama. 2002;287(2):195–202. doi: 10.1001/jama.287.2.195. [DOI] [PubMed] [Google Scholar]
- 26.Yu Y, Zhang S, Wang G, et al. The combined association of psychosocial stress and chronic hypertension with preeclampsia. Am J Obstet Gynecol. 2013;209(5):438 e431–438 e412. doi: 10.1016/j.ajog.2013.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.National Center for Health Statistics. CDC growth charts, United States. 2000 http://www.cdc.gov/growthcharts/. Accessed November 26, 2013.
- 28.Zhang S, Liu X, Brickman WJ, et al. Association of plasma leptin concentrations with adiposity measurements in rural Chinese adolescents. J Clin Endocrinol Metab. 2009;94(9):3497–3504. doi: 10.1210/jc.2009-1060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Inoue M, Maehata E, Yano M, Taniyama M, Suzuki S. Correlation between the adiponectin-leptin ratio and parameters of insulin resistance in patients with type 2 diabetes. Metabolism. 2005;54(3):281–286. doi: 10.1016/j.metabol.2004.09.006. [DOI] [PubMed] [Google Scholar]
- 30.Serum and red blood cell folate concentrations for assessing folate status in populations. World Health Organization (WHO); website http://www.who.int/nutrition/publications/micronutrients/indicators_serum_RBC_folate/en/. Accessed March 9, 2016. [Google Scholar]
- 31.Austin PC, Stuart EA. Estimating the effect of treatment on binary outcomes using full matching on the propensity score. Statistical methods in medical research Sep 1 pii: 0962280215601134. [Epub ahead of print] 2015 doi: 10.1177/0962280215601134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Obeid R, Koletzko B, Pietrzik K. Critical evaluation of lowering the recommended dietary intake of folate. Clin Nutr. 2014;33(2):252–259. doi: 10.1016/j.clnu.2013.12.013. [DOI] [PubMed] [Google Scholar]
- 33.Crider KS, Devine O, Hao L, et al. Population red blood cell folate concentrations for prevention of neural tube defects: Bayesian model. Bmj. 2014;349:g4554. doi: 10.1136/bmj.g4554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Fraser A, Tilling K, Macdonald-Wallis C, et al. Association of maternal weight gain in pregnancy with offspring obesity and metabolic and vascular traits in childhood. Circulation. 2010;121(23):2557–2564. doi: 10.1161/CIRCULATIONAHA.109.906081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Oken E, Rifas-Shiman SL, Field AE, Frazier AL, Gillman MW. Maternal gestational weight gain and offspring weight in adolescence. Obstet Gynecol. 2008;112(5):999–1006. doi: 10.1097/AOG.0b013e31818a5d50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Krishnaveni GV, Veena SR, Karat SC, Yajnik CS, Fall CH. Association between maternal folate concentrations during pregnancy and insulin resistance in Indian children. Diabetologia. 2014;57(1):110–121. doi: 10.1007/s00125-013-3086-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Strauss RS. Effects of the intrauterine environment on childhood growth. Br Med Bull. 1997;53(1):81–95. doi: 10.1093/oxfordjournals.bmb.a011608. [DOI] [PubMed] [Google Scholar]
- 38.Jacquemyn Y, Ajaji M, Karepouan N, Jacquemyn N, Van Sande H. Vitamin B12 and folic acid status of term pregnant women and newborns in the Antwerp region, Belgium. Clin Exp Obstet Gynecol. 2014;41(2):141–143. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.