Despite the paucity of evidence, recommendations exist internationally for pregnant women to increase their docosahexaenoic acid (DHA) intake to optimize fetal brain development. Our randomized controlled trial (RCT), in which pregnant women were allocated to 800 mg/d of DHA or matched placebo, showed that children’s mean cognitive, language, and motor scores did not differ between groups at 18 months, although fewer children in the DHA group had delayed development compared with controls.1 Surprisingly, girls in the DHA group had poorer language scores than girls in the control group.1 Herein we report neurodevelopmental outcomes at 4 years, which is when any subtle to moderate effects on development should have emerged and can be more reliably assessed.
The trial methods were previously published.1 All study procedures were conducted with written informed consent as approved by institutional review boards at each center. Children selected for assessment at 18 months of age, who had not died or withdrawn, were invited to attend an appointment when aged 4 years with a psychologist who was blinded to group allocation (June 18, 2010, to September 25, 2012). The primary outcome was the General Conceptual Ability (GCA) score of the Differential Ability Scales, second edition (DAS II; score range, 30-170; delayed, <85). Secondary outcomes included psychologist-assessed executive functioning and language, and parent-reported executive functioning and behavior (Table).
Information about children’s home environment, DHA intake, and medical conditions were collected. Statistical analyses were preplanned and conducted using SAS version 9.3 (SAS Institute Inc) and Stata release 12 (StataCorp). Analyses were conducted on an intention-to-treat basis for families who consented to follow-up, with missing outcomes imputed using chained equations; 100 imputed data sets were used to account for the sampling design and weights.2 Statistical significance was assessed at the 2-sided P < .05 level. Adjustments were made for center, parity, sex, and mother’s education and smoking status using linear or log binomial regression models. Sensitivity analyses conducted with available data only and with data imputed for the 726 children in the original sample produced similar results. A total of 536 children would provide 80% power (α = .05) to detect a 4-point difference in mean GCA between groups.
Of 726 children selected for the 18-month evaluation, 703 were eligible for the 4-year follow-up and 646 (91.9%; n = 313 in DHA group and n = 333 in control group) were included in the analysis (eFigure in Supplement). Mean GCA scores neither differed between groups (adjusted mean difference, 0.29 [95% CI, −1.35 to 1.93], P = .73; Table), nor did the percentage of children with delayed or advanced GCA scores. Other objective assessments of cognition, language, and executive functioning also did not differ between groups. However, the DHA group had poorer scores on some parentally reported scales of executive functioning and behavior. There was no evidence of effect modification by sex. Diagnoses of autism (2 DHA and 4 control) and hyperactivity disorders (0 cases) did not differ between groups.
Our data indicate that prenatal DHA supplementation does not influence objective assessments of cognition, language, and executive function at preschool age despite fewer preterm children in the DHA group, which was expected with the DHA intervention.1
Differences in secondary outcomes seen at 18 months (including cognitive delay and mean language scores) could no longer be detected and may have been diluted by other environmental factors or may have been chance findings. The majority of RCTs of DHA interventions during pregnancy have also reported null findings.3 However, few RCTs have attempted assessment beyond 2 years and those that have reported attrition rates of greater than 40%.3 Our trial has the advantages of good compliance with the intervention,1 a large sample, and high retention.
The subjective, parentally reported assessments indicated that children in the DHA group had poorer executive function and more behavioral difficulties than children in the control group, although the differences were small and unlikely to be of any clinical significance because all measures were within the normal range. These observations may be chance effects because of the high number of comparisons, or it is possible that women in the DHA group, who were more likely to correctly guess their group allocation at birth,4 had higher expectations of their children compared with controls.
Our data do not support prenatal DHA supplementation to enhance early childhood development.
Corresponding Author: Maria Makrides, BSc, BND, PhD, Women’s and Children’s Health Research Institute, 72 King William Rd, North Adelaide, SA, Australia 5006 (email@example.com)
Author Contributions: Dr Makrides had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Makrides, Yelland, Smithers, Anderson, Gibson.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Makrides, Gould, Yelland, Gibson.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Makrides, Yelland, Smithers.
Obtained funding: Makrides, Yelland, Smithers, Anderson, Gibson.
Administrative, technical, or material support: Makrides, Gould, Gawlik, Anderson, Gibson.
Study supervision: Makrides, Gould, Smithers, Anderson, Gibson.
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Makrides reported serving on scientific advisory boards for Nestle, Fonterra, and Nutricia. Dr Gibson reported serving on a scientific advisory board for Fonterra. The associated honoraria for Drs Makrides and Gibson are paid to their institutions to support conference travel and continuing education for postgraduate students and early career researchers. No other disclosures were reported.
Funding/Support: Both the original trial and the 4-year follow-up study were funded by National Health and Medical Research Council (NHMRC) grants 349301 (original trial) and 627174 (4-year follow-up). Treatment and control capsules were donated by Efamol (Surrey, England). Funding from the NMHRC grant was paid to Women’s and Children’s Health Research Institute and Data Management and Analysis Centre staff for contributions to the DOMInO 4-year follow-up trial.
Role of the Sponsors: The NHMRC and Efamol had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank Women’s and Children’s Health Research Institute (Adelaide) staff who assisted in the 4-year follow-up, particularly trial coordinator Helen Loudis, BHSc, who was compensated. We also thank the staff at the Data Management and Analysis Centre, School of Population Health, University of Adelaide. Beverly S. Muhlhausler, BSc(Hons), PhD (School of Agriculture, Food and Wine, University of Adelaide) assisted with setting up and implementing the project management system for the follow-up study and provided comment on the manuscript without compensation.
Trial Registration anzctr.org.au Identifiers: ACTRN12605000569606 and ACTRN12611001125910
et al. Effect of DHA supplementation during pregnancy on maternal depression and neurodevelopment of young children. JAMA
. 2010;304(15):1675-1683.PubMedGoogle ScholarCrossref
D. Multiple Imputation for Nonresponse in Surveys. New York, NY: Wiley; 1987.
M. The effect of maternal omega-3 (n-3) LCPUFA supplementation during pregnancy on early childhood cognitive and visual development. Am J Clin Nutr
. 2013;97(3):531-544.PubMedGoogle ScholarCrossref
M. Maternal supplementation with docosahexaenoic acid during pregnancy does not affect early visual development in the infant. Am J Clin Nutr
. 2011;93(6):1293-1299.PubMedGoogle ScholarCrossref