Background
Periconceptional diet quality is commonly suboptimal and sociodemographic disparities in diet quality exist. However, it is unknown whether individual periconceptional diet quality is associated with obstetric outcomes.
Objective
Our objective was to assess differences in maternal and neonatal outcomes according to maternal periconceptional diet quality.
Study Design
This is a secondary analysis of a large, multicenter prospective cohort study of 10,038 nulliparous women receiving obstetrical care at 8 United States centers. Women underwent 3 antenatal study visits and had detailed maternal and neonatal data abstracted by trained research personnel. In the first trimester (between 6 and 13 weeks), women completed the modified Block 2005 Food Frequency Questionnaire, a semiquantitative assessment of usual dietary intake for the 3 months around conception. Responses were scored using the Healthy Eating Index–2010, which assesses adherence to the 2010 Dietary Guidelines for Americans. Higher scores on the Healthy Eating Index represent better adherence. Healthy Eating Index scores were analyzed by quartile; quartile 4 represents the highest dietary quality. Bivariable and multivariable analyses were performed to assess associations between diet quality and outcomes. A sensitivity analysis in which markers of socioeconomic status were included in the multivariable Poisson regression models was performed.
Results
In the cohort of 8259 women with Healthy Eating Index data, the mean Healthy Eating Index score was 63 (±13) of 100. Women with the lowest quartile Healthy Eating Index scores were more likely to be younger, non-Hispanic black and Hispanic, publicly insured, low income, and tobacco users. They were more likely to have comorbidities (obesity, chronic hypertension, pregestational diabetes, mental health disorders), a higher prepregnancy body mass index, and less education. Women with lowest quartile scores experienced less frequent major perineal lacerations and more frequent postpartum hemorrhage requiring transfusion and hypertensive disorders of pregnancy, which persisted on multivariable analyses (controlling for age, body mass index, tobacco use, chronic hypertension, pregestational diabetes mellitus, and mental health disorders) comparing women in each quartile with quartile 4. Additionally, women in quartiles 1 and 2 experienced greater adjusted relative risk of cesarean delivery compared with women in quartile 4. Neonatal outcomes also differed by dietary quartile, with women in the lowest Healthy Eating Index quartile experiencing greater adjusted relative risk of preterm birth, neonatal intensive care unit admission, small for gestational age infant, and low birthweight, and lower risk of macrosomia; all neonatal findings also persisted in multivariable analyses. The sensitivity analysis with inclusion of markers of socioeconomic status (race/ethnicity, insurance status, marital status) in the multivariable models supported these findings.
Conclusion
Periconceptional diet quality among women in the United States is poor. Poorer periconceptional dietary quality is associated with adverse maternal and neonatal outcomes, even after controlling for potential comorbidities and body mass index, suggesting periconceptional diet may be an important social or biological determinant of health underlying existing health disparities.
Overall dietary quality is poor for most Americans. , Fewer than 3% of United States (US) adults have ideal diet scores, and ample public health data suggest poor dietary quality is associated with morbidity. Moreover, racial, ethnic, and socioeconomic disparities in dietary quality are substantial for nearly all measures, including diet scores, individual nutrient sources, and energy intake, and while overall dietary quality in the US may be improving, these disparities are widening. , , ,
Why was the study conducted?
Although disparities in periconceptional dietary quality exist, it is unknown whether individual periconceptional diet quality is associated with obstetric outcomes.
Key findings
Poor periconceptional dietary quality is associated with a greater relative risk of cesarean delivery, hypertensive disorders, postpartum hemorrhage, neonatal intensive care unit admission, preterm birth, and low birthweight, whereas it is associated with a lower risk of major perineal laceration and macrosomia.
What does this add to what is known?
Poor periconceptional dietary quality is associated with adverse perinatal outcomes, even after controlling for body mass index and potential comorbidities.
Reproductive-age women planning pregnancy have similarly poor diets , , despite potential fetal health implications. Multiple European-based studies show that women planning pregnancy are only marginally more likely to comply with dietary recommendations and that dietary patterns changed little from before pregnancy to early pregnancy. , , Thus, a woman’s periconceptional diet is highly reflective of her general nutritional patterns and dietary intake later in pregnancy.
In 2017, using data from a large cohort of US nulliparous women, Bodnar et al demonstrated both that periconceptional dietary quality is suboptimal in US women and that racial, ethnic, and sociodemographic disparities in dietary quality exist. In this analysis, non-Hispanic white women had the highest quality of periconceptional diet, whereas almost half of non-Hispanic black women had dietary quality in the lowest quintile. Furthermore, although the quality of diet increased with greater maternal education in all racial or ethnic groups, education was most strongly associated with diet quality for white women.
Top sources of energy, overall, in this study were foods rich in sugars and solid fats and included refined bread, soda, pasta, grain desserts, and alcohol.
Periconceptional dietary quality has been hypothesized to be an important determinant of maternal and fetal outcomes, , with suboptimal nutrition having a critical negative influence on fetal growth, placentation, inflammation, and maternal metabolic regulation, and possibly leading to differences in outcomes such as live birth rate or birthweight. Poor periconceptional dietary quality may affect pregnancy outcomes via mechanisms such as micronutrient deficiency or relationship with gestational weight gain. However, data to confirm this hypothesis are lacking, particularly in the US. Thus, our objective was to assess whether there is an association between periconceptional dietary quality and maternal and neonatal outcomes.
Materials and Methods
This is a secondary analysis of data from the Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-To-Be (nuMoM2b), which was a large, multicenter observational cohort study conducted at 8 US medical centers from 2010 to 2013. In this study, more than 10,000 nulliparous women with singleton pregnancies were enrolled for prospective study.
Recruitment was conducted at geographically diverse locations and was designed to sample a population reflective of the general US population. Women were eligible for enrollment if they had a live singleton pregnancy, had no previous pregnancy that progressed beyond 20 weeks of gestation, and were between 6 weeks 0 days and 13 weeks 6 days of gestation at recruitment.
Exclusion criteria included maternal age younger than 13 years, history of 3 or more spontaneous abortions, current pregnancy complicated by suspected fatal fetal malformation or known fetal aneuploidy, assisted reproduction with a donor oocyte, multifetal reduction, or plan to terminate the pregnancy. Data were collected via multiple sources, including in-person interviews, surveys completed by participants, and medical record review.
Participants completed 3 study visits with trained research personnel, with visit 1 occurring between 6 weeks 0 days and 13 weeks 6 days of gestation. At least 30 days after delivery, trained and certified chart abstractors reviewed the medical records of all participants and recorded final birth outcomes.
Full details of the study protocol previously have been published.
This analysis specifically addresses periconceptional dietary quality as the exposure of interest. At visit 1, women completed the modified Block 2005 Food Frequency Questionnaire, a semiquantitative assessment of usual dietary intake for the 3 months around conception. The Block questionnaire assesses 52 nutrients and 35 food groups from approximately 120 food and beverage items. The questionnaire includes serial adjustment items to estimate portion size, and the instrument has been validated in many populations. Details of the Block questionnaire have previously been reported by Bodnar et al.
Answers to the Block questionnaire were scored using the Healthy Eating Index 2010 (HEI-2010), or the HEI. , The HEI, which is a measure used to assess how well a set of foods aligns with key recommendations of the 2010 Dietary Guidelines for Americans, evaluates 12 key aspects of dietary quality, including adequacy of intake of specific food groups and moderation of intake of less nutritious foods. Higher scores represent better adherence to national guidelines, and an ideal score of 100 indicates that the reported food intake is consistent with the Dietary Guidelines recommendations.
The mean HEI-2010 score for adult Americans in 2007–2008 was 54.3 of 100, which indicated that the average diet of adult Americans did not align with dietary recommendations. This analysis is restricted to women with available HEI data.
We a priori selected 5 maternal and 5 neonatal outcomes of interest, each of which was chosen based on the plausible relationship of these outcomes with periconceptional food quality. , , Maternal outcomes included gestational diabetes mellitus (GDM), major perineal laceration (defined as third- or fourth-degree perineal laceration), cesarean delivery, postpartum hemorrhage requiring a blood transfusion, and hypertensive disorder of pregnancy.
A GDM diagnosis was based on clinical record review using each site’s local protocol for diagnosis. Postpartum hemorrhage was restricted to women who required a transfusion to assess associations with the most severe version of this outcome. Hypertensive disorder of pregnancy included antepartum gestational hypertension, or antepartum, intrapartum, or postpartum (up to 14 days) preeclampsia, eclampsia, or superimposed preeclampsia, as defined by the American College of Obstetricians and Gynecologists.
Neonatal outcomes of interest included preterm birth (<37 weeks of gestation), admission to the neonatal intensive care unit (NICU), small-for-gestational-age infant (defined as <10th percentile by criteria of Alexander et al ), low birthweight (defined as <2500 g), and macrosomia (defined as >4000 g).
Multiple maternal demographic and clinical characteristics were assessed as potentially confounding factors. Demographic factors included maternal age, insurance status (public vs non-public), marital status, household income (<200% or ≥200% of the poverty line), educational attainment (some college or greater vs no college), and self-reported race and ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, Asian, and other). Clinical factors included body mass index (BMI; kilograms per square meter) at visit 1, tobacco use currently or before pregnancy, chronic hypertension (regardless of medication status), pregestational diabetes mellitus, and any mental health disorder.
We examined differences between maternal baseline demographic and clinical characteristics by HEI quartile using χ 2 and analysis of variance tests (ANOVA), as appropriate. We then assessed differences between maternal and neonatal outcomes by HEI quartile using χ 2 tests. HEI scores were analyzed by quartile because such groupings best reflect clinically relevant categories of dietary quality and are most consistent with existing literature.
Analyses for the outcome of GDM excluded women with pregestational diabetes mellitus. Using multivariable Poisson regression models, adjusted relative risks were constructed to estimate the independent associations of HEI quartile with each outcome, with HEI quartile 4 (highest level of food quality) as the referent, and each HEI quartile individually compared with the referent.
The multivariable model included potentially confounding variables that were associated with HEI quartile on bivariable models with a value of P < .05. Although markers of socioeconomic status differed by HEI quartile, these factors were (a priori) not used in multivariable models because of likely collider bias related to the potential causal relationship between socioeconomic factors, periconceptional dietary quality, and maternal and neonatal outcomes. Thus, final models did not include race/ethnicity, insurance status, marital status, and educational attainment. However, to confirm the primary findings, we performed a sensitivity analysis in which race/ethnicity, insurance status, and marital status were included in the multivariable Poisson models.
All analyses were carried out in STATA release 15.0 (StataCorp, College Station, TX). All statistical tests were 2 tailed and considered significant at the P < .05 level. Each site’s local governing institutional review board approved the study, and all women provided written informed consent prior to participation.
Results
The nuMoM2b cohort included 10,038 women, of whom 82% (n = 8259) were eligible for inclusion in this analysis. The mean HEI score was 63, with a standard deviation of 13 ( Figure ). Women in the lowest quartile had scores less than 53.7, whereas quartile 2 included 53.8–63.7, quartile 3 included 63.8–72.7, and quartile 4 included women with scores 72.8 and greater.
Women in the lowest HEI quartile, representing poorest dietary quality, were younger and more likely to be non-Hispanic black or Hispanic, have public insurance, use tobacco, and have a lower household income ( Table 1 ). They were less likely to be married and have at least some college education. Women in the lowest HEI quartile additionally had a higher mean prepregnancy BMI and were more likely to have comorbidities, including chronic hypertension, pregestational diabetes, and mental health disorders.
| Variables | HEI quartile 1 (n = 2065) | HEI quartile 2 (n = 2065) | HEI quartile 3 (n = 2065) | HEI quartile 4 (n = 2064) | P value a |
|---|---|---|---|---|---|
| Maternal age, years | 23.9 (±5.2) | 26.6 (±5.5) | 28.7 (±5.1) | 29.9 (±4.5) | < .001 |
| Race/ethnicity | < .001 | ||||
| Non-Hispanic white | 987 (47.8) | 1198 (58.1) | 1472 (71.3) | 1536 (74.4) | |
| Non-Hispanic black | 496 (24.0) | 277 (13.4) | 113 (5.5) | 58 (2.8) | |
| Hispanic | 420 (20.3) | 421 (20.4) | 287 (13.9) | 246 (11.9) | |
| Asian | 31 (1.5) | 68 (3.3) | 107 (5.2) | 142 (6.9) | |
| Other | 131 (6.3) | 99 (4.8) | 85 (4.1) | 82 (4.0) | |
| Public insurance | 1037 (50.7) | 604 (29.5) | 313 (15.2) | 174 (8.4) | < .001 |
| Household income <200% poverty line | 782 (55.7) | 567 (33.8) | 341 (18.5) | 241 (12.4) | < .001 |
| Married | 630 (30.5) | 1201 (58.2) | 1571 (76.2) | 1795 (87.0) | < .001 |
| Some college education or greater | 1581 (82.0) | 1532 (90.7) | 1384 (96.5) | 1182 (98.8) | < .001 |
| Body mass index, kg/m 2 | 27.1 (±7.3) | 26.9 (±6.6) | 25.9 (±5.6) | 24.9 (±4.9) | < .001 |
| Ever used tobacco | 1047 (50.7) | 864 (41.9) | 788 (38.3) | 756 (36.6) | < .001 |
| Chronic hypertension | 64 (3.3) | 60 (3.0) | 43 (2.2) | 24 (1.2) | < .001 |
| Pregestational diabetes mellitus | 39 (2.0) | 33 (1.7) | 29 (1.5) | 16 (0.8) | .018 |
| Mental health disorder | 433 (22.0) | 356 (17.9) | 339 (17.0) | 289 (14.6) | < .001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
