n-3 PUFA are essential nutrients. Adequate consumption is vital in pregnancy, as n-3 PUFA, in particular long-chain DHA, contribute to offspring neurodevelopment and may improve pregnancy outcomes, including risk for preterm birth(Reference Middleton, Gomersall and Gould1). Fish and other seafood (hereafter ‘fish’) are the main dietary source of long-chain n-3 PUFA. Therefore, current guidance recommends intake of 8–12 ounces (224–336 g, or 2–3 servings) of fish per week during pregnancy(2,3) , with the goal of consuming an average of 200 mg/d of DHA(Reference Koletzko, Cetin and Brenna4).
Limited research suggests that few pregnant women consume the recommended amounts of fish or n-3 PUFA. The latest US Food and Drug Administration (FDA) assessment of dietary fish intake was conducted in 2014 and relied on data sources by then already decades old(5). In the 2004 Infant Feeding Practices Study II, the median intake of total fish by pregnant participants, excluding non-fish consumers, was 1·8 ounces/week (about 1 serving per month)(Reference Lando, Fein and Choinière6). In the 2013 National Health and Nutrition Examination Survey (NHANES), mean fish intake by pregnant women was 4·6 servings a month(Reference Thompson, Hein and Hanson7).
Additionally, studies suggest that fish and n-3 PUFA intake during pregnancy has been declining over past decades, likely in response to federal advisories about mercury in fish since 2001(Reference Oken, Kleinman and Berland8,Reference Shimshack and Ward9) . In the NHANES survey, mean DHA intake among women of childbearing age decreased from 56 mg/d in 2003–2004 to 42 mg/d in 2011–2012(Reference Nordgren, Lyden and Anderson-Berry10). Intake was highest in women who were non-Hispanic White and had higher education and income levels, similar to demographic patterns in non-population-based cohorts(Reference Donahue, Rifas-Shiman and Olsen11,Reference Oken, Østerdal and Gillman12) . Despite the importance of fish consumption during pregnancy, women consume substantially less than men and do not increase intake during pregnancy(Reference Thompson, Hein and Hanson7,Reference Nordgren, Lyden and Anderson-Berry10,Reference Murphy, Devarshi and Ekimura13) .
Most experts believe that fish consumption is the optimal way to meet recommendations for adequate n-3 PUFA intake(Reference Kris-Etherton and Innis14), in part because experimental evidence has not supported offspring developmental benefits of supplementation(Reference Nevins, Donovan and Snetselaar15,Reference Gould, Roberts and Makrides16) . For those who cannot or choose not to eat fish, n-3 PUFA supplements are recommended(Reference Adams, Kirby and Sorensen17). The extent to which pregnant women take n-3 PUFA supplements is not well described. In addition, it is unclear whether supplement use is more common in those with low fish intake. In the 2003–2012 NHANES surveys, only 9 % of pregnant women consumed an n-3 PUFA supplement, but this information was not presented according to year of pregnancy or by fish intake(Reference Nordgren, Lyden and Anderson-Berry10).
We examined data from the National Institutes of Health Environmental influences on Child Health Outcomes (ECHO) programme(Reference LeWinn, Caretta and Davis18,Reference Knapp, Kress and Parker19) to address our hypotheses that fish consumption would have declined over the past two decades and that supplement use would be more common among those who did not eat fish.
Methods
Study design, sample and measures
In October 2022, the ECHO data platform included information on more than 52 000 singleton pregnancies from sixty-nine cohorts across the USA and Puerto Rico (see online supplementary material, Supplementary Figure 1). We included data from twenty-three cohorts that collected information on fish intake during pregnancy and thirty-five that collected supplement intake during pregnancy. Within cohorts, pregnancies were included if they had information on either fish intake or supplement use.
Assessment of fish and supplement intake
We performed a keyword search and form review for FFQ that assessed fish intake and any questionnaires that assessed supplement intake. For fish intake, we converted relevant questionnaire items and summed as appropriate to weekly total intake. We then constructed a four-level categorical variable: (1) never or less than once per month, (2) once per month to less than once per week, (3) one to two times per week and (4) more than twice per week. Additionally, we created a binary variable of never or less than once per month (which we summarise as ‘never’) v. more (‘ever’). For supplements, we created a binary variable to indicate any (v. no) use of supplements with fish oil or n-3 fatty acids.
Assessment of other characteristics
We captured other variables of interest from harmonised derived tables of maternal self-reported sociodemographic characteristics, including age, race, ethnicity, education, tobacco or nicotine use during pregnancy (Y/N), and pre-pregnancy BMI, each of which we categorised as in Table 1.
Characteristics of 10 800 ECHO-wide cohort participants with information on fish consumption during pregnancy
ECHO, Environmental influences on Child Health Outcomes; GED, General Educational Development; BA, Bachelor of Arts; BS, Bachelor of Science.
* Results from dataset with missing values imputed using multiple imputation for age (n 1049, 9·7 %), race/ethnicity (n 586, 5·4 %), education (n 1048, 9·7 %), pre-pregnancy BMI (n 2525, 23·4 %), and tobacco or nicotine use (n 2650, 24·5 %).
† Relative risk from regression model, including all covariates in the table, with missing values imputed using multiple imputation.
Statistical analysis
We performed multiple imputation (n 25 imputations) using SAS Proc MI to fill in missing data on covariates of interest. We then performed a log-binomial regression analysis, mutually adjusted for measured demographic characteristics (as in Table 1), smoking status and pre-pregnancy BMI. We included a random effect for cohort to account for the nested nature of the pooled data and conducted leave-one-cohort-out analyses to confirm that no cohort explained the overall associations. Additionally, we included the number of fish questions asked on the dietary questionnaire (1 or more than 1) given our prior research that found that asking more questions leads to a higher estimate of fish intake(Reference Oken, Guthrie and Bloomingdale20). Income was not significantly associated with either outcome and including it did not substantially alter any other estimates; therefore, we did not include it in our final models.
Results
Among 10 800 pregnant women with information on fish consumption, 24·6 % reported never consuming fish and 75·4 % reported ever consuming fish during pregnancy (Table 1): 40·1 % less than 1 serving per week, 22·1 % 1–2 servings per week and 13·2 % more than 2 servings per week (see online supplementary material, Supplemental Table 1). In the multivariable regression analyses with imputed missing covariates (Table 1), the likelihood of ever (v. never) consuming fish during pregnancy remained higher in those who were older (relative risk (RR) 1·14, 95 % CI 1·10, 1·18 for >40 v. <29 years), were other than non-Hispanic White (RR 1·13, 95 % CI 1·08, 1·18 for non-Hispanic Black; 1·05, 95 % CI 1·01, 1·10 for non-Hispanic Asian; 1·06, 95 % CI 1·02, 1·10 for Hispanic), had lower BMI (RR 0·97, 95 % CI 0·95, 1·0 for overweight v. normal BMI), and used tobacco or nicotine products (RR 1·04, 95 % CI 1·01, 1·08). After accounting for demographics and the number of fish questions included on the different dietary questionnaires, no differences in fish intake were observed by year of delivery.
Among 12 646 pregnant women with information on n-3 PUFA supplement intake, 16·2 % reported any supplement use. Supplement use was uncommon before 2005 (less than 0·05 %), but not substantially different afterwards. In multivariable regression analyses (Table 2), supplement use was more likely at an older age and a higher level of education: those over 40 years of age were about twice as likely to use supplements than those less than 29 years of age (RR 2·01, 95 % CI 1·35, 3·00) and those with a graduate degree were more likely to use supplements than those with less than a high school education (RR 1·71, 95 % CI 1·21, 2·41). Supplement use was less likely among non-Hispanic Black (RR 0·61, 95 % CI 0·48, 0·76) and Hispanic (RR 0·67, 95 % CI 0·55, 0·80) participants compared with non-Hispanic White participants, those who used tobacco or nicotine products (RR 0·81, 95 % CI 0·68, 0·98) and those with a higher BMI (RR 0·79, 95 % CI 0·68, 0·90 for BMI ≥ 30 v. <25 kg/m2). In contrast to the advice that those who do not consume fish should take an n-3 fatty acid supplement, supplement use was highest among those with greater fish consumption (Table 2 and Fig. 1).
Characteristics of 12 646 ECHO-wide cohort participants with information on n-3 fatty acid supplement consumption during pregnancy
ECHO, Environmental influences on Child Health Outcomes; GED, General Educational Development; BA, Bachelor of Arts; BS, Bachelor of Science.
* Results from dataset with missing values imputed using multiple imputation for age (n 966, 7·6 %), race/ethnicity (n 338, 2·7 %), education (n 468, 3·7 %), pre-pregnancy BMI (n 2351, 18·6 %), and tobacco or nicotine use (n 1229, 9·7 %).
† Relative risk from regression model, including all covariates in the table, with missing values imputed using multiple imputation.
Likelihood of n-3 polyunsaturated supplement use in pregnancy according to fish consumption during pregnancy within the Environmental influences on Child Health Outcomes (ECHO) cohort
Discussion
A large number of observational studies and some randomised trials have examined associations of total prenatal fish or n-3 PUFA intake with a range of outcomes(Reference Oken21). Expert opinion has coalesced around the benefits of regular fish or supplement consumption to achieve an intake of at least 500 mg/d of long-chain n-3 PUFA (including 200 mg/d of DHA)(Reference Makrides22). Using data from over 10 000 pregnancies across the USA occurring from 1999 through 2020, we observed that almost a quarter of women reported never consuming fish, and only 16 % consumed any n-3 fatty acid supplements. Additionally, fish intake correlated with demographic and health characteristics, albeit somewhat less strongly and not entirely similar to supplement use. Similar to supplement use, fish intake was higher in those who were older and had a higher income and education, but different from supplements, fish intake was higher in those with racial/ethnic identities other than non-Hispanic White and in those who used tobacco and nicotine products. Supplement intake tracked even more strongly with demographics, with the highest likelihood of intake among those who were older, had a higher education and income, and were non-Hispanic White or Asian. Additionally, supplement use was less common among those at higher risk for adverse pregnancy outcomes as a function of using tobacco or nicotine products or having a higher BMI.
Limited recent information is available about fish intake in US pregnancies, yet such estimates are essential for efforts to model health risks and benefits from nutrients and contaminants commonly found in fish. For example, in the most recent FDA assessment of the net effects of eating commercial fish on fetal neurodevelopment, which was conducted in 2014, the FDA estimated the types and amount of fish that people eat(5). This estimation was based on three sources of data that were collected about 15, about 20 and about 30 years ago: National Marine Fisheries Service market share data on consumable commercial fish from 2007, NHANES data from 1999 to 2000, and US Department of Agriculture’s Continuing Survey of Food Intake by Individuals (CSFII) data from 1989 to 1991. With these data, the FDA estimated that women of childbearing age in the USA consumed a mean of 3·7 and a median of 1·9 ounces of fish per week. It is notable that this model was based entirely on dietary data from non-pregnant persons and assumed that pregnant women eat a similar amount, despite evidence that women consume less fish during pregnancy(Reference Lando, Fein and Choinière6,Reference Bloomingdale, Guthrie and Price23) . Further, fish consumption in women of childbearing age may have decreased since these data were collected(Reference Oken, Kleinman and Berland8,Reference Shimshack and Ward9) .
Evidence to support routine n-3 supplementation in pregnancy has not been entirely consistent(Reference Middleton, Gomersall and Gould1,Reference Saccone, Saccone and Berghella24) . Nevertheless, those with low baseline fish intake or n-3 PUFA status(Reference Bisgaard, Stokholm and Chawes25,Reference Simmonds, Sullivan and Skubisz26) or a BMI that indicates obesity(Reference Rasool, Alvarado-Flores and O’Tierney-Ginn27,Reference Satokar, Cutfield and Cameron-Smith28) may particularly benefit from supplementation. We did not observe that supplement use was more common in those with low fish intake or a high BMI, but rather the opposite. In another study of ECHO participants, more than 99 % of the sample reported use of vitamins and minerals containing supplements during pregnancy, but that analysis did not include n-3 PUFA supplements(Reference Sauder, Harte and Ringham29). In contrast to vitamin and mineral supplement use, which abated most nutrient risk disparities from diet alone, we found that n-3 PUFA supplement use was less common among those who did not eat fish.
The large sample size is a strength, and we included data from over the past two decades up to 2020. In contrast, published NHANES analyses include about 1,000 pregnancies and a decade of data extending to 2015. Limitations include our inability to assess specific fish types, given the varied dietary assessment instruments used across cohorts, or to assign intake by trimester. However, most prior studies have examined total fish intake, and current US guidelines recommend total fish intake rather than specific subtypes such as ‘fatty’ fish(2). Also, although different studies administered different questionnaires, we accounted for the number of questions asked about fish. Additionally, we do not have information on supplement dose. Both fish and supplement intake were self-reported, as is typical in studies of usual diet, and reporting may have been biased.
The ECHO population is nationwide but not necessarily nationally representative(Reference Knapp, Kress and Parker19), as it draws upon individuals who elected to enrol in cohorts and who may be more health conscious than the general population. The very low fish and n-3 PUFA supplement intake we observed may overestimate actual use in all US pregnancies, as more health-conscious persons may consume more fish and supplements, or alternatively, it may be that more health-conscious persons try to avoid mercury exposure from fish. Our results are especially timely given that both the WHO and US National Academies are currently evaluating the evidence on fish intake in pregnancy(30,31) . Ongoing effective public health advice and resources to support clinicians(Reference Judge, Lucas and Kuzoian32,Reference Ziegenfuss, Dinh and McCann33) are needed to encourage consumption of low-mercury fish during pregnancy and intake of n-3 supplements among those who do not consume fish.
Acknowledgements
The authors would like to acknowledge the unique contributions of Zhumin Zhang, MS, PhD, University of Wisconsin-Madison, to this study, which included investigation, resources, supervision and funding acquisition.
The authors wish to thank our ECHO colleagues; the medical, nursing and programme staff; and the children and families participating in the ECHO cohorts. We also acknowledge the contribution of the following ECHO programme collaborators:
ECHO components
Coordinating Center: Duke Clinical Research Institute, Durham, North Carolina: Smith PB, Newby KL; Data Analysis Center: Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland: Jacobson LP; Research Triangle Institute, Durham, North Carolina: Catellier DJ; Person-Reported Outcomes Core: Northwestern University, Evanston, Illinois: Gershon R, Cella D.
ECHO awardees and cohorts
Northeastern University, Boston, Massachusetts: Alshawabkeh, AN; Icahn School of Medicine at Mount Sinai, New York, NY: Teitelbaum SL; Stroustrup A; Cohen Children’s Medical Center, Northwell Health: Stroustrup A; Memorial Hospital of Rhode Island, Providence RI: Koinis Mitchell D, Deoni S; New York State Psychiatric Institute, New York, NY: Duarte C; University of Puerto Rico, San Juan, PR: Canino G; Avera Health Rapid City, Rapid City, sd: Elliott A; Kaiser Permanente Northern California Division of Research, Oakland, CA: Ferrara A; University of Wisconsin, Madison WI: Gern J; Henry Ford Health System:, Detroit, MI: Zoratti E; Washington University in St Louis, St Louis, MO: Rivera-Spoljaric K; University of Wisconsin, Madison, WI: Singh A; Vanderbilt University, Nashville, TN: Hartert T; University of Southern California, Los Angeles, CA: Breton C; Farzan S; Habre R; University of California Davis Mind Institute, Sacramento, CA: Hertz-Picciotto I; University of Washington, Department of Environmental and Occupational Health Sciences, Seattle, WA: Karr C; University of Tennessee Health Science Center, Memphis, TN: Mason A; Brigham and Women’s Hospital, Boston, MA: Weiss S; Boston University Medical Center, Boston, MA: O’Connor G; Kaiser Permanente, Southern California, San Diego, CA: Zeiger R; Washington University of St. Louis, St Louis, MO: Bacharier L; University of California, UC Davis Medical Center Mind Institute, Davis, CA: Schmidt R; Johns Hopkins Bloomberg School of Public Health: Volk H; Kaiser Permanente Northern California Division of Research: Croen L; Columbia University Medical Center, New York, NY: Herbstman J; University of Illinois, Beckman Institute, Urbana, IL: Schantz S; University of California, San Francisco:, San Francisco, CA: Woodruff T; New York School of Medicine, New York, NY: Trasande L; Icahn School of Medicine at Mount Sinai, New York, NY: Wright R; Boston Children’s Hospital, Boston MA: Bosquet-Enlow M.
Disclaimer
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Financial support
Research reported in this publication was supported by the Environmental influences on Child Health Outcomes (ECHO) programme, Office of the Director, National Institutes of Health, under Award Numbers U2COD023375 (Coordinating Center), U24OD023382 (Data Analysis Center), U24OD023319 with co-funding from the Office of Behavioral and Social Science Research (PRO Core), UH3OD023251 (Alshawabkeh), UH3OD023320 (Aschner), UH3OD023253 (Camargo), UH3OD023248 (Dabelea), UH3OD023313 (Koinis-Mitchell), UH3OD023328 (Duarte), UH3OD023318 (Dunlop), UH3OD023279 (Elliott), UH3OD023289 (Ferrara), UH3OD023282 (Gern), UH3OD023287 (Breton), UH3OD023365 (Hertz-Picciotto), UH3OD023244 (Hipwell), UH3OD023275 (Karagas), UH3OD023271 (Karr), UH3OD023347 (Lester), UH3OD023268 (Weiss), UH3OD023342 (Lyall), UH3OD023349 (O’Connor), UH3OD023286 (Oken), UH3OD023285 (Kerver), UH3OD023290 (Herbstman), UH3OD023272 (Schantz), UH3OD023305 (Trasande), UH3OD023337 (Wright).
Conflict of interest
None of the authors reports a conflict of interest.
Authorship
Conceptualisation: Oken, Lyall, Musci and Westlake. Methodology: Oken, Lyall, Musci, Westlake and Gachigi. Software: Westlake and Gachigi. Validation: Musci, Westlake and Gachigi. Formal analysis: Musci, Westlake and Gachigi. Investigation: Oken, Aschner, Barnes, Bastain, Buss, Camargo, Cordero, Dabelea, Dunlop, Ghassabian, Hipwell, Hockett, Karagas, Lugo-Candelas, Margolis, O’Connor, Shuster, Straughen and Lyall. Resources: Oken, Aschner, Barnes, Bastain, Buss, Camargo, Cordero, Dabelea, Dunlop, Ghassabian, Hipwell, Hockett, Karagas, Lugo-Candelas, Margolis, O’Connor, Shuster, Straughen and Lyall. Data Curation: Westlake. Writing, Original Draft: Oken. Writing, Review and Editing: All. Visualisation: Oken. Supervision: Oken, Aschner, Barnes, Bastain, Buss, Camargo, Cordero, Dabelea, Dunlop, Ghassabian, Hipwell, Hockett, Karagas, Lugo-Candelas, Margolis, O’Connor, Shuster, Straughen and Lyall. Project Administration: Oken and Lyall. Funding Acquisition: Oken, Aschner, Barnes, Bastain, Buss, Camargo, Cordero, Dabelea, Dunlop, Ghassabian, Hipwell, Hockett, Karagas, Lugo-Candelas, Margolis, O’Connor, Shuster, Straughen, and Lyall.
Ethics of human subject participation
This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and participant activities were overseen and approved by both central and site-specific Institutional Review Boards. All participants provided written informed consent.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S136898002400051X.
Open access
