The developing immune system is known to be very sensitive to environmental chemicals 1 2 and the antenatal and early childhood period is an important window of vulnerability 3 . Prenatal exposure to environmental contaminants may increase IgE levels in cord blood 4 5 and in early childhood 6 and may be associated with respiratory or allergic symptoms in preschool children (wheezing or food allergy) 7 8 . Some of these contaminants are present in seafood, including polychlorinated biphenyls (PCBs), dioxins, other polychlorinated compounds, perfluorinated chemicals (PFCs), and metals 9 10 . On the other hand, seafood is a major source of long-chain polyunsaturated fatty acids (PUFAs), especially n-3 PUFAs, gestational exposure to which is well known to have beneficial effects on children’s subsequent health 11 . Moreover, there is growing evidence that n-3 PUFAs have anti-inflammatory properties and may modulate immune responses 12 13 . The positive effect of these PUFAs may therefore counterbalance in part the deleterious impact of contaminants. The levels of n-3 PUFAs (higher in fish than shellfish) and contaminants (often high in shellfish) however, differ between fish and shellfish 9 10 . Accordingly, these categories of seafood should be assessed separately.

The PELAGIE cohort of pregnant women living in Brittany (the western part of France surrounded by the sea), is particularly suitable for such an assessment, as shellfish consumption is sufficiently high to allow fish and shellfish effects to be disentangled.

The objective of this study was to estimate the association between fish and shellfish intake during pregnancy on wheeze, eczema and food allergy in early childhood.

Compared with the 1500 participants, nonrespondents (n = 1496) were younger at the birth of the PELAGIE child (p < 0.001), less educated (p < 0.001), and more likely to smoke (p < 0.001). Shellfish consumption did not differ between respondents and nonrespondents but the latter were less likely to eat fish (p = 0.001) (Table 1).

*Chi square test.

The mean age of participating mothers was 30.4 (SD 4.2) years, and 67% had completed high school. Over 25% reported smoking at the beginning of pregnancy. Thirty percent of the children had a family history of asthma and/or allergy. Most were born at term, by vaginal delivery, with a normal birth weight for gestational age. The mean age of the children at follow-up was 26.6 (SD 2.2) months (Table 2).

*Chi square test.

Overall, 252 (17%) women reported eating fish either never or less than once a month, 815 (54%) 1 to 4 times a month, and 433 (29%) more than 4 times a month; 496 (33%) reported eating shellfish once a month or more. Fish and shellfish consumption were positively correlated (p < 0.001). The median age at introduction of fish to the child was 8 months (IQR: 7–12) and was not associated with prenatal seafood consumption (p-value of 0.67 and 0.58 for prenatal fish and shellfish intake respectively).

Wheeze status was assessed in 1491 children: 299 (20%) had certainly had at least one wheezing episode and another 656 (44%) probably had at least one. Eczema status was known for 1487 children, 475 (32%) of whom were classified positive. Among them 319 (67%) had been medically diagnosed with eczema. Food allergy status was known for 1487 children, 136 (9%) were classified with food allergy: 37 had a medical diagnosis of cow’s milk allergy, 41 a medical diagnosis of a food allergy, and 22 of both, while 36 children had apparent allergic reactions to food but no doctor’s diagnosis. Several children had several diagnoses, and the most important correlation was observed between eczema and food allergy (p < 0.001). The associations with potential risk factors are presented in Table 2.

The change in the ORs between the crude and adjusted estimates did not exceed 7%. Table 3 presents the adjusted estimates. A borderline significant association was observed between moderate pre-parturition fish consumption before pregnancy and a lower risk of “certain wheeze” during the first two years of life (aOR = 0.69 (0.45–1.05)). Including both types of seafood consumption in the model did not change the results. Maternal seafood consumption was not associated with eczema. Shellfish intake once a month or more was associated with a higher risk of a food allergy (aOR = 1.62 (1.11–2.37)). The association decreased slightly when both types of seafood were included. No interactions were observed.

*Adjusted for: mother’s age, maternal education, folic acid supplementation, familial history of asthma/allergy, child’s sex, small-for-gestational age, infant’s method of feeding, day care attendance, post-natal exposure to tobacco and, child’s age at follow-up.

**Both fish and shellfish consumption were included in the model with the adjustment variables above.

§ p-value < 0.10; ‡ p-value < 0.05.

After multiple imputation (n = 2981; that is, 2996 families eligible for follow-up minus those for whom seafood consumption was not known), the association between prenatal fish consumption and wheeze was no longer significant (aOR = 0.86 (063–1.17)). On the other hand, multiple imputation confirmed the results for eczema and food allergy observed in the complete-case analysis (Table 3).

In this large, population-based birth cohort study, moderate maternal pre-parturition fish consumption may be associated with a lower risk of wheeze before age two, and no significant association was observed with eczema and food allergy. Shellfish consumption at least once a month was associated with a higher risk of a food allergy.

Shellfish consumption was four times higher in the PELAGIE population than at the national level 14 20 . This allowed us to disentangle the effects of these different seafoods and makes this study, to the best of our knowledge, the first to examine the influence of maternal pre-parturition shellfish consumption on food allergy. Moreover, this study of this highly educated population provided high quality questionnaire data. Its prospective longitudinal nature enabled us to characterize exposure long before maternal reports of outcomes and to collect substantial data about factors known to be associated with wheeze, eczema, and food allergy at the age of two.

A limitation of the study is the attrition at follow-up. Since nonrespondents at follow-up both ate fish less often and shared characteristics of families with a higher prevalence of child respiratory or allergic problems 21 , their absence is likely to induce bias in the estimation of the association between maternal fish consumption during pregnancy and wheeze. Sensitivity analyses with multiple imputation appear to confirm this hypothesis. On the other hand, there was no follow-up selection according to shellfish intake, and multiple imputation endorsed the association between maternal shellfish consumption and food allergy. Seafood consumption was evaluated before pregnancy, and women may have modified their habits at the beginning of pregnancy, especially for mollusk intake (because of the infectious risks). However, the biological half-life of both fatty acids and most of the contaminants in seafood is long, and reports of usual diet (i.e., before pregnancy) are likely to represent exposure during pregnancy. Measures of wheeze and eczema were adapted from the ISAAC questionnaires. Although this method has been validated for assessments of children from the age of 7 years 17 , it is widely used in epidemiological studies of younger children 22 23 24 25 . Reports of outcomes were not confirmed by clinical examination or biological markers for atopy, these unsupervised methods might have induced non-differential misclassification. We did not take childhood intake of fish into account for two principal reasons. First, the age at introduction of fish was not associated with prenatal seafood consumption in our study and therefore should not confound the association between prenatal seafood consumption and outcomes. Second, although some previous studies have observed that later introduction of fish may be associated with an increased risk of eczema or sensitization to food allergens 26 27 , others have used detailed prospective data to argue that these associations may in fact result from reverse causation: parental suspicions of allergic reactions were likely to influence the age at introduction of fish 28 . These two arguments justify our decision not to take childhood intake of fish into account in studying this association.

Some studies have examined the effect of maternal fish intake during pregnancy on either eczema 22 23 29 30 31 or asthma/wheezing 22 23 24 29 32 in children. Four of the five studies that focused on eczema showed a decreased risk with a high consumption of fish (>1/week) during pregnancy. The study 22 that did not observe any association was from the cohort with the lowest average fish consumption. As in that study, the percentage of high fish consumers is low in our population; leading to a lack of statistical power that may explain the absence of significant findings for this association. In our study we did not observe a consistent association between maternal fish consumption during pregnancy and childhood wheeze. Two previous studies have observed a decreased risk of asthma/wheezing 29 32 in children of women who consume high levels of fish. Of the five studies on this subject, those two had the populations of pregnant women with the highest average fish consumption and asthma/wheezing evaluation at the oldest age (6 and 7 years), when asthma is easier to diagnose. It is hypothesized that eating fish may reduce the risk of respiratory and allergic outcomes because it is a major source of long-chain n-3 PUFAs and micronutrients such as selenium 11 33 that may have anti-inflammatory properties 12 13 34 . On the other hand, fish is contaminated by marine pollutants, mainly PCBs, dioxins and other polychlorinated compounds, and methylmercury 9 10 . Previous study has reported no link between prenatal methylmercury exposure and asthma or atopic dermatitis at age 7 6 . However, two of the three epidemiological studies 7 8 35 that evaluated effects of prenatal exposure to PCBs, dioxins, or other polychlorinated compounds at age 2 noted an increased risk of wheezes 8 or food allergy 7 at higher levels of exposure, after adjustment for major confounding factors. To our knowledge, two published studies have suggested a link between maternal fish consumption and increased risk of allergies 22 36 . One focused on food allergy 36 . In their population, with a level of fish consumption similar to that in ours, they assessed food allergen sensitization by skin prick tests and observed that higher consumption of fish during pregnancy was associated with a higher risk of such sensitization. Their results must, however, be interpreted with caution as they were obtained in a cross-sectional study with a retrospective assessment of prenatal diet. The second of these studies focused on wheezing and eczema and was based on a cohort design. The population of this study, with its very low fish consumption, may not benefit of simultaneous n-3 PUFAs intake 11 22 .

This study (n = 2796) also observed an increased risk of wheeze and eczema in children of women who ate shellfish during pregnancy 22 . Our findings were not significant for these associations, perhaps because of the smaller sample size; they nonetheless suggest an increased risk of food allergy. French national surveys 9 10 have showed that on average, shellfish is less contaminated by PCBs, dioxins or other polychlorinated compounds, or methylmercury than fish is. Other contaminants have been detected in shellfish at higher concentrations than in fish, in particular, metals such as lead, cadmium, arsenic (including a small proportion of inorganic and methylated forms considered toxic 37 ), and organic compounds such as PFCs 9 10 38 . French national surveys 10 38 have also shown non-compliance with the regulatory shellfish threshold for cadmium or high exposure levels for cadmium and arsenic. Both metals have demonstrated their potential developmental immunotoxicity 39 40 41 42 . These contaminations, which were associated in shellfish with low concentrations of n-3 PUFAs, may therefore explain those findings.

Shellfish intake was associated with a higher risk of food allergy. In line with our previous work in the PELAGIE cohort showing that shellfish consumption is associated with decreased fecundability 43 and decreased fetal growth 14 , we suggest yet another potential adverse effect. Further large-scale epidemiological studies are needed to corroborate these results, identify the contaminants or components of shellfish responsible for the effects observed, determine the persistence of the associations seen at age 2, and investigate potential associations with health effects observable at later ages such as allergic asthma.

aOR: Adjusted odds ratio; CI: Confidence interval; IQR: Interquartile range; ISAAC: International study of asthma and allergies in childhood; OR: Odds ratio; PCBs: Polychlorinated biphenyls; PFCs: Perfluorinated chemicals; PUFAs: Polyunsaturated fatty acids; SD: Standard deviation; SGA: Small-for-gestational age.

The authors declare that they have no competing interests.

FP conducted the statistical analyses and drafted the initial manuscript. EB and HG took part in the statistical analyses. CM and FR were responsible for the field study including questionnaires and medical data. LM and JFV reviewed and revised the manuscript. SC designed the study and supervised the overall project. She critically reviewed the manuscript. All authors read and approved the final manuscript as submitted.