Article Text

Download PDFPDF

Cognitive assessment of children at age 2½ years after maternal fish oil supplementation in pregnancy: a randomised controlled trial
  1. J A Dunstan1,
  2. K Simmer1,
  3. G Dixon2,
  4. S L Prescott1
  1. 1
    School of Paediatrics and Child Health, University of Western Australia, Perth, Western Australia, Australia
  2. 2
    Telethon Institute for Child Health Research, Centre for Child Health Research, The University of Western Australia, Western Australia, Australia
  1. A/Professor S L Prescott, School of Paediatrics and Child Health, University of Western Australia, Princess Margaret Hospital, PO Box D184, Perth, WA 6840, Australia; susanp{at}


Objective: To assess the effects of antenatal omega 3 long-chain polyunsaturated fatty acid (n-3 LC PUFA) on cognitive development in a cohort of children whose mothers received high-dose fish oil in pregnancy.

Design: A double-blind randomised placebo-controlled trial.

Setting: Perth, Western Australia, Australia.

Patients: 98 pregnant women received the supplementation from 20 weeks’ gestation until delivery. Their infants (n = 72) were assessed at age 2½ years.

Interventions: Fish oil (2.2 g docosahexaenoic acid (DHA) and 1.1 g eicosapentaenoic acid (EPA)/day) or olive oil from 20 weeks’ gestation until delivery.

Outcome measures: Effects on infant growth and developmental quotients (Griffiths Mental Development Scales), receptive language (Peabody Picture Vocabulary Test) and behaviour (Child Behaviour Checklist).

Results: Children in the fish oil-supplemented group (n = 33) attained a significantly higher score for eye and hand coordination (mean ((SD) score 114 (10.2)) than those in the placebo group (n = 39, mean score 108 (SD 11.3); p = 0.021, adjusted p = 0.008). Eye and hand coordination scores correlated with n-3 PUFA levels in cord blood erythrocytes (EPA: r = 0.320, p = 0.007; DHA: r = 0.308, p = 0.009) and inversely correlated with n-6 PUFA (arachidonic acid 20:4n-6: r = −0.331, p = 0.005). Growth measurements in the two groups were similar at age 2½ years.

Conclusion: Maternal fish oil supplementation during pregnancy is safe for the fetus and infant, and may have potentially beneficial effects on the child’s eye and hand coordination. Further studies are needed to determine the significance of this finding.

View Full Text

Statistics from

During gestation, large amounts of docosahexaenoic acid (DHA; 22:6n-3) and arachidonic acid (20:4n-6) are deposited in the fetal retina and brain,1 and these fatty acids seem to be critical for normal neuronal and visual development.24 The composition of long-chain polyunsaturated fatty acids (LC PUFAs) influences membrane stability, fluidity and function of many cell types through its effects on gene expression and tissue differentiation.4 Deficiencies in LC PUFAs could have implications for wide-ranging effects on health and development. Despite selective transplacental transfer of LC PUFAs5 with higher concentrations in the neonatal circulation,68 fatty acid status in the fetus remains dependent on maternal fatty acid status and dietary intake. The recent decline in dietary DHA and other n-3 PUFAs with progressive westernisation has raised concerns about its potential effects on the developing fetus. The implications are even greater for preterm infants who are deprived of maternal supplies in the third trimester. This has led to growing interest in the role and indications for LC PUFA supplementation in early life, both in pregnancy and in the early postnatal period for preterm infants. With concerns about methylmercury contamination in certain types of fish, there is growing consumption of pharmaceutical-grade fish oil supplements, which have been shown to be safe.

So far, only three studies have assessed the effects of LC PUFAs in term pregnancies on cognitive function of children after maternal supplementation with cod liver oil,9 10 fish oil11 and DHA-enriched eggs.12 Although data are still limited, these studies have shown associations between DHA and several functional neurological outcomes including improved attention up to age 2 years12 and improved mental processing at age 4 years.10 Other studies have assessed the effects of supplementation on preterm infants during the equivalent (third trimester) period ex utero, and have shown some beneficial effects on neurodevelopmental parameters.13 Although any conclusive long-term benefit of PUFA-supplemented formula in preterm infants has yet to be proved,14 15 a meta-regression analysis of seven trials in term infants showed that the DHA dose in milk formula was positively related to visual acuity measurements at age 4 months.16

Although the benefits are unclear, theoretical concerns remain about the potential detrimental effects of selective n-3 PUFA supplementation, which could lead to a reduction in n-6 arachidonic acid status at high doses. Thus, although the available literature suggests more beneficial rather than harmful effects (if any) on development, clearly, more studies are needed to confirm this.

We were able to deal with these issues in a cohort of children whose mothers received n-3 PUFA fish oil in pregnancy. This cohort was part of a randomised controlled trial designed to assess the effects of n-3 PUFA supplementation during pregnancy on neonatal immune responses, ultimately with a view to reducing the risk of allergic disease.17 High doses of fish oil (4 g/day) resulted in markedly higher levels of n-3 PUFA but considerably lower levels of n-3 arachidonic acid in cell membranes of the neonates from the intervention group.18 We used this opportunity to test the hypothesis that exposure to high-dose fish oil in utero would not have a deleterious effect on subsequent growth and development (up to age 2½ years). These data were collected to provide further information to assess the safety of (high-dose) fatty acid supplementation in the preterm period (both in utero and ex utero).



Ninety eight pregnant Caucasian women with allergic disease were recruited before 20 weeks of pregnancy between January 2000 and September 2001 to a randomised double-blind controlled trial conducted in Perth, Western Australia, Australia. Women were ineligible for the study if they smoked, had medical problems, a complicated pregnancy, seafood allergy, or if their normal dietary intake exceeded two meals of fish per week. Children were excluded from the study if they were born before 36 weeks’ gestation or with major disease (to avoid the confounding effects on immune response) or if cord blood was not collected. Anthropometric data were collected from hospital records at the time of delivery. The sample size for this study is small, as the study was designed to assess immunological effects rather than neurodevelopmental outcome and growth.17 Assuming a level of 80% statistical power and a significance level of 0.05, the sample size allowed us to detect a difference in height of 3.3 cm and in weight of 1.4 kg at age 2–3 years.

Study design and intervention

Women were block randomised according to parity, age, body mass index before pregnancy and maternal allergy to minimise the potential effects of these confounding factors. Women received four (1 g) fish oil capsules (fish oil group, n = 52; Ocean Nutrition, Halifax, Nova Scotia, Canada) providing 1.1 g eicosapentaenoic acid (EPA; 20:5n-3) and 2.2 g DHA/day, or four (1 g) capsules of olive oil per day (olive oil or placebo group, n = 46; providing 2.7 g n-9 oleic acid per day; Pan Laboratories, Moorebank, New South Wales, Australia) from 20 weeks’ gestation until birth. This dose was chosen as approximately equivalent to one fatty fish meal per day.19 Both active and placebo capsules contained 3–4 mg/g oil α-tocopherol (vitamin E) as an antioxidant. Randomisation and allocation of capsules was carried out in a blinded manner, and capsules in the two groups were image matched. Mothers and research staff remained blinded until completion of the cognitive testing. Compliance was confirmed by measuring fatty acid levels in erythrocytes in the mothers at 30 and 37 weeks’ gestation and 6 weeks postnatally.18

Follow-up of children at age 21/2 years

All eligible children whose mothers completed supplementation (n = 83) were invited to attend follow-up visits. Height, weight and head circumference were measured. As secondary outcome measures, children were assessed for development (Griffiths Mental Development Scales (GMDS)),20 receptive language (Peabody Picture Vocabulary Test) (PPVT)) IIIA21 and behaviour (Child Behavior Checklist (CBCL) 1½–5 years).22 These tests were chosen because they have valid and reliable psychometric properties for assessing children aged 2½ years. The GMDS comprise six subscales of development (locomotor, personal social, speech and hearing, eye and hand coordination, performance, and practical reasoning). One quotient was calculated for each subscale, as well as a general quotient that was derived as the mean of the subscale quotients. The GMDS20 version used has been updated since the commencement of the study; however, the original version still provides robust comparative data. The PPVT21 is a test of listening comprehension to measure receptive vocabulary attainment for standard English. A standard score in the PPVT IIIA is calculated from the raw score. The average range for receptive language is defined as a mean standard score of 100 (standard deviation (SD) 15). The CBCL 1½–5 years is an instrument designed to measure parental perception of child competencies and behaviours in children aged 1½–5 years, and includes a language development survey based on length of phrases and number of words in the child’s vocabulary. Raw scores for the CBCL are used to calculate internalising, externalising and total problem behaviour T scores. A T score of ⩾64 is in the clinical range for behaviour problems.

Fatty acid analyses

Phospholipid fatty acid analyses were carried out as described previously.18 The fatty acids were quantified as a percentage of the weight of the total fatty acids measured. The results were expressed as the total sum of n-3 PUFAs (20:5n-3, 22:5n-3 and 22:6n-3) and n-6 PUFAs (18:2n-6, 20:3n-6, 20:4n-6, 22:3n-6 and 22:4n-6), and the ratio of n-3 to n-6 fatty acids.

Statistical analyses

The independent samples t test using SPSS V.11.5 was used to compare the equality of means (equal variances not assumed) between the fish oil-supplemented group and the control group. Differences between the groups for dichotomous data were determined by the χ2 test. The groups were also combined to determine associations between individual fatty acid proportions and subscales of development in the GMDS using Pearson’s correlation.

Potential confounding factors including length of gestation, maternal age and education, parity, sex of the infant and breast feeding were determined by Pearson’s correlation. A linear regression (single-step) model was used to examine relationships between continuous variables, after adjusting for confounding factors; p<0.05 was considered significant for all analyses. Owing to the exploratory nature of this study, we did not wish to exclude any important relationships by using stringent correction factors for multiple analyses, and we recognised the potential for a type 1 error.

The ethics committees of St John of God Hospital, Subiaco, Australia, and Princess Margaret Hospital, Perth, Australia, approved the study, and all women gave written consent.


Study demographics

Figure 1 shows the flow of participants through the trial. In all, 83 women and their healthy babies completed the study at birth (40 in the fish oil group and 43 in the placebo group) as reported previously.23 Seven children were lost to follow-up and 33 children were assessed at age 2½ years (mean age 34.0 (SD 5.3) months) in the fish oil group. In the olive oil group, four children were lost to follow-up and 39 were assessed at age 2½ years (mean age 34.7 (SD 5.7) months, p = 0.599). Although mothers in the fish oil group were slightly younger than those in the olive oil group, we found no other differences between the mothers and children who completed the follow-up at 2.5 years (table 1).

Figure 1 Progress of participants through the trial. *Fish oil or placebo received from 20 weeks’ gestation to delivery.
Table 1 Characteristics of the two study populations

Fatty acid composition at birth

The content of n-3 PUFAs (EPA and DHA) in the phospholipid fraction of erythrocyte membranes in the umbilical cord blood was significantly increased and that of the n-6 PUFA, arachidonic acid, was significantly decreased in infants from the fish oil group compared with controls (p<0.001; table 2).

Table 2 Phospholipid fatty acids from erythrocytes in cord blood of neonates who completed the cognitive studies

Growth measurements

We found no significant differences between the fish oil group (n = 28) and the olive oil group (n = 36). The mean age for both groups was 30 months; the mean (SD) height was 93.8 (3.8) cm for the fish oil group v 93.3 (4.6) cm for the olive oil group (p = 0.642); the mean (SD) weight was 14.5 (2.0) v 14.1 (2.0) kg, respectively (p = 0.456); and the head circumference was 49.4 (1.6) v 49.8 (1.7) cm, respectively (p = 0.304).

GMDS results

Table 3 gives the results of the Griffiths assessments. Children from the fish oil group attained a significantly higher score for eye and hand coordination (p = 0.021). When tested in a linear regression model with eye and hand coordination as the dependent variable and potential confounding factors (maternal age, maternal education and duration of breast feeding) as independent variables, the effect of supplementation remained significant (β = −7.213, p = 0.008). We found no significant difference between the groups for the mean general quotient or for mean quotients for other subscales of development (table 3).

Table 3 Mean subscales quotients and general quotient scores of Griffiths assessments for both the study populations

Receptive language results

Results from the PPVT indicate no significant difference (p = 0.110) between the mean (SD) standard score obtained in the fish oil group (n = 31; 101.3 (9.9)) and in the olive oil group (n = 39; 97.4 (9.7)).

CBCL results

Results from the CBCL 1½–5 years indicated no significant differences between the mean (SD) T scores of the fish oil and olive oil groups for internalising (44 (8.7) v 45 (9.2); p = 0.576), externalising (49 (9) v 48 (9); p = 0.706) and total problem behaviour scales (35 (5.2) v 36 (5.2); p = 0.548). We found no significant difference between the groups for mean length of phrases (mean (SD) centile score 73.2 (15.8) v 67.8 (20.2), respectively; p = 0.300) and vocabulary centile score (mean (SD) 57.6 (16.9) v 55.3 (20.2), respectively; p = 0.650).

Relationship between LC PUFA levels in erythrocyte membranes at birth and neurodevelopmental outcomes in children at age 34 months

These relationships were examined for the study population as a whole (table 4). We found a significant positive correlation between the eye and hand coordination score at age 34 months and n-3 PUFA composition of cord blood erythrocytes (EPA, p = 0.007; DHA, p = 0.009). Conversely, the eye and hand coordination and performance scores were inversely correlated with arachidonic acid (p = 0.005 and 0.044, respectively). The significant associations were independent of potential confounding factors including maternal education and the duration of breast feeding. We found no other significant associations of LC PUFA levels in cell membranes at birth with developmental, language or behaviour outcomes at age 2½–3 years.

Table 4 Association between cognitive outcomes and fatty acids in cord blood erythrocyte membranes


We found that cord erythrocyte phospholipid n-3 LC PUFA levels positively correlated and n-6 levels negatively correlated with eye and hand coordination, a finding that, to the best of our knowledge, has not been reported previously.

Our finding of enhanced eye and hand coordination with fish oil supplementation is plausible and consistent with previously reported benefits on visual function after postnatal n-3 PUFA supplementation in both preterm14 24 and term15 25 infants. Although the underlying mechanism is not understood, DHA is known to facilitate rapid phototransduction in the retinal membrane,26 and deficiencies are associated with reduced retinal function in infant primates.2 Furthermore, effects on visual evoked potential could indicate that DHA may also have an effect on the development of the visual cortex.27 Finally, improved stereoacuity in infants has been associated with LC PUFA formula supplementation28 and fish intake of lactating mothers.29

To our knowledge, only one other study has assessed the effects of supplementation with high-dose fish oil in pregnancy on cognitive development of the offspring. A randomised clinical trial by Helland et al9 involved 590 pregnant women who received fish oil at half the dose we used in this study, from 18 weeks’ gestation until 3 months post partum. No differences in development were observed in the 269 infants tested at 6 and 9 months; however, fish oil supplementation was associated with increased mental processing in children at age 4 years. Additionally, mental processing scores were significantly correlated with maternal intake of DHA in pregnancy after adjusting for potential confounding factors10; this is consistent with observed correlations of DHA (and EPA) intake with eye and hand coordination in this study.

Other studies have found positive relationships between n-3 PUFAs at birth (principally DHA) and aspects of visual and neurological development, in either observational studies3032 or intervention studies using much lower levels of supplementation.11 12 33 Our findings suggest that detection of the potentially beneficial effects of DHA in pregnancy may require larger doses. Further, although it is difficult to directly extrapolate the pregnancy dosage to supplementation of the preterm infant, the doses in our study resulted in similar increases in cord blood levels of DHA to those achieved with the higher doses trialled in preterm infants.34

What is already known on this topic

  • Long-chain fatty acids are essential for normal neuronal and visual development.

  • Fetal fatty acid status is dependent on maternal fatty acid status and dietary intake.

  • Previous data are limited, but indicate potential benefits of relatively low-dose n-3 polyunsaturated fatty acid (PUFA) supplementation in pregnancy on functional neurological outcomes. However, no data are available for the effect of high-dose n-3 PUFA supplementation in pregnancy on neurodevelopment.

What this study adds

  • Children of mothers supplemented with 3.7 g n-3 long-chain polyunsaturated fatty acid (LC PUFA) in pregnancy had improved hand and eye coordination at age 2½ years.

  • High-dose n-3 LC PUFA taken in the second half of pregnancy does not seem to have any deleterious effects on neurodevelopment or growth.

The small size of the sample is a limitation of our study. Although we cannot exclude a chance finding, there seem to be no adverse effects on measures of development, receptive language, behaviour and growth. Children who received prenatal fish oil supplementation tended to perform better in all subscales of development and had higher scores for receptive language, average phrase length and vocabulary scores when assessed at an average age of 34 months. Both the control group and the fish oil group scored higher than the mean (100) in all GMDS subscales, with the fish oil group scoring higher than the control group in each subscale. However, in eye and hand coordination on the GMDS, the difference was statistically significant when the fish oil group was compared with controls. We acknowledge the potential for a type 1 error due to multiple comparisons; however, the significance was strengthened when we used a multi-regression model to adjust for potential confounding factors. We are confident in these results as we have robust comparative data from a population-based sample of Western Australian infants assessed at a similar age, with the GMDS yielding similar mean subscale and general quotient scores.35

These preliminary data indicate that supplementation with a relatively high-dose fish oil during the last 20 weeks of pregnancy is not only safe but also seems to have potential beneficial effects that need to be explored further. Given the scarcity of data to support the efficacy of fish oil supplementation during pregnancy, our data have a potentially important role in informing on the effects of fish oil supplementation on early postnatal infant development. In conclusion, our findings are important in tackling concerns that a relatively selective supplementation (with n-3 PUFA but not n-6 PUFA) could have detrimental effects by displacing other essential fatty acids.


We thank the staff and patients who helped in this study. We particularly thank Dr Trevor Mori and Dr Ann Barden, who were responsible for the fatty acid analyses.


View Abstract


  • Funding: This study was funded by a grant from the Raine Medical Research Foundation of Western Australia and a grant from the National Health and Medical Research Foundation of Australia. Dr Janet Dunstan is supported by the Child Health Research Foundation of Western Australia.

  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Perspective
    C King M Fewtrell