Since the 1950s, we have witnessed a dramatic increase in the prevalence of allergy-related diseases, such as eczema (atopic dermatitis), asthma, allergic rhinoconjunctivitis (ARC, hay fever) and food allergies [1
]. At the same time, there exists a large variability in the prevalence and severity of these diseases, not just between countries, but also across regions within the same country. This variability suggests environmental characteristics, such as eating habits, may influence the local prevalence [2
]. Changes in dietary habits have also been flagged as one of the potential drivers of the increasing prevalence of allergy-related diseases. In particular, higher consumption of omega-6 polyunsaturated fatty acids (n-
6 PUFAs) and lower consumption of omega-3 fatty acids (n-
3 PUFAs) has been highlighted as one of the major changes in the Western diets which has occurred over the same time period as allergy-related disease prevalence has risen [3
]. Plausible biologically mechanisms can explain a causal relationship between increased dietary n-
6 PUFAS and allergy-related diseases [4
3 PUFAs, on the other hand, are thought to have beneficial health effects [5
] and both randomized controlled trials (RCTs) and epidemiological studies have sought to determine if higher dietary n-
3 PUFAs prevents the development of allergy-related diseases by regulating the perinatal immune system [6
]. The two largest RCTs assessing prenatal n-
3 PUFAs supplementation have produced inconsistent results [8
]. Bisgaard et al. observed a protective effect against persistent wheeze or asthma and lower respiratory tract infections up to five years [9
], whilst Best et al. found protective effects only for atopic eczema and allergic sensitization up to six years [8
]. A systematic review and meta-analysis of both randomized and quasi-randomized interventions concluded that maternal n-
3 PUFA supplementation may be beneficial in the prevention of egg sensitization, but not other allergic outcomes [10
]. Similarly, systematic reviews which have assessed trials of infant supplementation have failed to identify consistent preventive effects [11
Oily fish is a major dietary source of n-3
PUFAs, and several epidemiological studies have shown protective effects from eating fish during both pregnancy [14
] and infancy [18
] on the development of allergy-related diseases in childhood, however the results are conflicting. A small RCT, “The Salmon Pregnancy Study”, demonstrated salmon consumption in pregnancy modified neonatal immune responses, but had no effect on markers of infant atopy at six months of age [21
]. Looking at observational studies, a systematic review and meta-analysis found no association between fish intake during pregnancy and allergy in the offspring, however fish consumption in the first year of life was associated with reduced risk of eczema and allergic rhinitis [22
]. Another meta-analysis concluded that introduction of fish early in life (six to nine months) and intake of fish at least once a week reduces asthma and wheeze in children up to 4.5 years of age [23
]. Few publications have considered maternal and infant fish or cod liver oil consumption within the same study and, due correlations between maternal and infant dietary habits, there is a need for studies considering both in order to disentangle their effects.
We have previously published results from the Prevention of Allergy among Children in Trondheim (PACT) study describing a 30% reduction in doctor-diagnosed asthma at two years of age after a community-based lifestyle intervention designed to increase fish and cod liver oil intake, reduce tobacco exposure and reduce indoor dampness during pregnancy and the first two years of life [24
]. From the same study, we also found a negative association between eating fish once a week at one year and eczema at two years [25
The aim of the current study was to extend our previous work on diet and allergy-related disease in infancy and investigate associations between maternal and infant consumption of fish and cod liver oil on allergy-related diseases at six years of age. Using the PACT study, we sought to estimate the effect of four exposures (consumption of any, oily or lean fish, or cod liver oil), at four time points (pregnancy and breastfeeding for mothers and one- and two-years for infants) on four allergy-related diseases/symptom at six years of age (current eczema, asthma and wheeze, and cumulative incidence of ARC).
Fish consumption at least once a week at one year of age was found to be associated with a 28%, 40% and 34% reduction in the odds of current eczema, asthma, and wheeze at six years of age, respectively. These findings are consistent with our previous analyses indicating that infant fish consumption at one year reduced the risk of eczema at two years of age [25
]. The protective effect of fish consumption at one year appears to be relevant for both boys and girls, is not conclusively associated with a particular fish type, is independent of maternal consumption during pregnancy and breastfeeding, and does not appear to be the result of reverse causality or confounding-by-indication. We also found that cod liver oil consumption at least four times per week at one year of age tended to be associated with a lower risk of allergy-related outcomes at six years. Otherwise, we found no other consistent associations between allergy-related outcomes and fish or cod liver oil consumption by mothers during pregnancy and breastfeeding, or infants at two years of age. Nor could we exclude that the apparent protective effect of early fish introduction is primarily a result of reverse causality or confounding-by-indication.
These findings are in line with recent meta-analyses which found that infant consumption of fish, but not maternal consumption during pregnancy, was associated with a reduced risk of eczema and ARC and possibly asthma [22
]. Similarly, a pooled analysis of multiple birth cohorts found that the risk of wheeze, asthma, and ARC in preschool and school-age children was not associated with maternal consumption of fish during pregnancy [30
]. Together with the findings presented in this paper, these studies suggest that any preventive effect of fish consumption is best achieved by increasing fish in the infants’ diet. Indeed, our findings suggest that maternal fish consumption alone may tend to increase the risk of asthma, wheeze, and ARC.
In an attempt to tease out the effect of maternal and infant fish consumption, we applied an alternative adjusted logistic regression model to each comparison, which included maternal or infant dietary consumption at other time points. As one would expect, we observed that maternal and infant consumption are correlated. Assuming that maternal consumption partially determines infant consumption at one year, when the association of maternal fish consumption during pregnancy and allergic outcomes is adjusted for infant’s consumption at one year of age we are essentially conditioning on a mediator. In doing so, we are no longer estimating the total effect of consumption during pregnancy but are essentially estimating the natural direct effect of fish consumption in pregnancy on allergic outcomes at six years. The estimates produced by model 2 suggesting that maternal fish consumption during pregnancy may increase the risk of asthma, wheeze and ARC are, therefore, not directly comparable to previous studies, and do not represent conflicting results. However, these findings need to be investigated in other large studies and considered in light of other health benefits of fish consumption for pregnant women.
In a study published by Goksör et al. [31
], the introduction of fish into the infants’ diet before nine months of age was associated with a reduced risk of wheeze at 4.5 years after adjusting for a number of potential confounders, including eczema and food allergy during the first year of life, and for frequency of fish consumption at one year of age. In contrast, our own findings suggest that the apparent association between early introduction to fish and lower risk of the allergic outcomes was probably influenced by delayed introduction of fish after the development of eczema, asthma or ARC. The discrepancy between our findings and those of Goksör et al. [31
] findings may be due to differences in the measure of “frequent” fish consumption (at least once per month vs. per week), the choice of potential confounding factors, or the inclusion of fish consumption at one year as a covariate since this may partially mediate the effect of early introduction of fish on wheeze.
Early introduction of allergenic foods, including fish, has more recently been investigated in the primary prevention of food allergy. A recent review found evidence that the immunoregulatory and anti-inflammatory properties of allergenic foods can provoke oral tolerance if introduced early to both low-risk and high-risk infants [32
], and a number of RCTs have produced promising results for the early introduction of peanut and egg in the prevention of allergy to these foods in high-risk infants [33
]. The benefits of early fish introduction as a means of preventing fish allergy are unknown. To date, only the EAT study has included fish in the early introduction regime [34
], however, the incidence of fish allergy in the study participants was too low to be able to confirm whether the early introduction can prevent fish allergy. We did not have sufficient information about food allergy among the children in the PACT study to be able to investigate this possibility in the present study.
In terms of cod liver oil consumption, our findings are in contrast with the results from the RCT by Bisgaard et al. [9
] which found a reduced risk of recurrent wheeze and asthma up to five years of age after n-3
PUFA supplementation during pregnancy. The supplement recommended in the PACT study was equivalent to 1.2 g n-3
PUFA per day, half the dose used their RCT, so it is unclear if the contrasting effects are due to the dosage, residual confounding or misclassification in our study. It is also worth noting that the preventive effect reported by Bisgaard et al. [9
] was not observed in the other large RCT [8
] or meta-analyses [10
] and needs to be confirmed in future studies. Another interesting observation made by Bisgaard et al. was that the preventive effect of n-3
PUFA supplementation was most pronounced for infants of women with a low baseline eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) levels. The PACT study design does not allow us to investigate this further, although it would suggest that encouraging n
-PUFA intake in populations with normally high oily fish and cod liver oil consumption may only provide a small preventive benefit. In addition to baseline levels of PUFAs, other factors may influence the effect of n-3
PUFAs on allergic outcome, such as vitamin D status or concomitant vitamin D supplementation. Results from the Vitamin D Antenatal Asthma Reduction Trial (VDAART) suggest that infants with high cord blood 25 (OH)D, in combination with high plasma or dietary PUFA at three years of age, was associated with the lowest odds of allergic disease [35
]. This would suggest that antenatal vitamin D status may influence the magnitude of the preventive effect of postnatal PUFA consumption. Interaction between maternal vitamin D and PUFA supplementation was not observed by Bisgaard et al. [9
]. In a separate analysis from the VDAART study, prenatal fish oil supplementation was associated with a reduced risk of asthma/recurrent wheeze in offspring, however, baseline vitamin D in mothers did not affect this preventive effect [36
The strengths of this study are the large size of the study and that it comprises exposure both during pregnancy, lactation and in the child’s two years of life. Information on diet at each of these time points allows us to distinguish between the influence of the mother’s and child’s dietary consumption. The prospective design reduces the risk of recall bias, and we were able to consider the influence of reverse causality and confounding-by-indication by using parental reports of the age of disease or symptom onset. At the same time, limitations of this study include the risk of selection bias, the high drop-out rate, and the lack of objectively confirmed and repeated dietary intake information and allergic outcomes. The PACT study aimed to recruit a general population. To investigate the degree of self-selection bias, a non-participant study was conducted using a short questionnaire. Information regarding maternal age, allergic disease in the family, smoking and socioeconomic status was completed for 391 consecutive parents visiting maternal postnatal care centers. These non-participants were found to be similar to participants in these baseline characteristics [4
]. For the current study, comparing participants from drop-outs indicated that there were some minor differences, however, we do not believe these have substantially influenced the results. Lastly, the lack of objectively confirmed dietary intake and parentally reported allergic outcomes may introduce a degree of misclassification of these measures, although we consider it most likely that any misclassification is non-differential. Only one questionnaire was completed during pregnancy and we, therefore, do not have the possibility to determine how representative the reported dietary habits are of the entire pregnancy period, nor can we use this data to investigate the importance of the timing of fish or cod liver oil consumption during pregnancy. Analyses from the VDAART study suggest that consumption during both the first and third trimester is of the greatest preventive effect against asthma/recurrent wheeze [36
], however, this is in a population with comparatively very low fish oil consumption. Allergic outcomes were parentally reported at two and six years. While there was no specific follow-up of health outcomes among PACT participants before two years of age, we believe that most children with allergic symptoms would have been identified through routine follow-ups at community health centers. Throughout Norway, children attend free-of-charge routine follow-ups with a child health nurse at six weeks, and 3, 4, 5, 6, 8, 10, 12, 15, 18, and 24 months. Children are additionally assessed by a doctor at six weeks, and 6, 12, and 24 months. Nonetheless, the age of disease or symptom debut recorded in the PACT study may still be subject to recall bias.