Effect of Prenatal Omega-3 Polyunsaturated Fatty Acid Supplementation on Childhood Eczema: A Systematic Review and Meta-Analysis

Eczema is a common allergic disease in children and has been increasing annually on a global scale over the past century [1‒3]. A clarification is made that atopic eczema may precede other allergic diseases such as asthma, allergic rhinitis, and food allergy in the atopic march, but this is not always the case [4‒6]. Notably, about 20% of all children suffered symptoms of eczema worldwide, and there is an increasing trend in prevalence rates in many countries [7], which has exerted a substantial, long-term psychosocial impact on children and their families and lowering children’s life satisfaction and self-esteem [8‒10]. It is highly unlikely that genetic alterations in the human population over the past centuries have caused the increased prevalence of allergic diseases. It is, however, likely that several changes to our lifestyles, including unhealthy diets, sedentary lifestyles, more diverse microbial exposures, etc., have altered our immune systems, which may contribute to the development of allergies [11, 12]. Indeed, diet and nutrition could have an impact on people’s health before birth through maternal metabolism [13].

Previous research suggested that the rising prevalence of allergic diseases in children in the last decades may be a result of modern dietary alterations, particularly the increment in the intake of omega-6 polyunsaturated fatty acids and relative omega-3 polyunsaturated fatty acid (ω-3 PUFAs) deficiency [14, 15]. Eicosapentaenoic acid and docosahexaenoic acid, present in high concentrations in fatty fish may decrease the inflammatory components of eczema by competing with n-6 fatty acids [16, 17]. ω-3 PUFAs have been reported to inhibit nuclear factor-κappa B (NF-κB) signaling pathway by inhibiting IκB phosphorylation, thereby downregulating the expression of inflammation-related genes and mediating host defense processes [18‒21]. Thus, there is some evidence showing that increased ω-3 PUFA supplementation may reduce the incidence of allergic disease and is a beneficial strategy to improve the health and life satisfaction of patients.

The fact that eczema is the earliest sign of allergic disease in infants shows the onset of an immune disorder existing early in life, as well as the potential impact of prenatal exposure to certain nutrients on the occurrence of eczema [22, 23]. The fetal period is known to be a crucial duration in view of the development of human immune activation and regulation, indicating that exposure to key dietary nutrients such as ω-3 PUFAs in the fetal stage plays a crucial role in preventing allergic diseases [24‒27]. The relationship between the development of eczema among children and maternal fish intake has been explored by several observational studies. Overall, increased risk of childhood eczema has been reported as the outcome of weekly maternal shellfish or fatty fish intake by some studies [28], whereas some other studies reported a reduced likelihood of infantile eczema due to fish intake in pregnancy [29, 30], or no notable relationship between eczema among youngsters and maternal fish intake [31, 32]. However, recent randomized control trials (RCTs) showed that prenatal ω-3 PUFA exposure did not have clinically meaningful eczema or IgE-associated-eczema protective effect [33‒38].

Two reviews of RCTs were conducted to examine the inter-relationship between the consumption of fish oil and the likelihood of developing an allergic disease [39, 40]. One review, which included three different RCTs, reported a pronounced decrease in the occurrence of atopic eczema at 12 months of age [39]; however, due to the small sample size of RCTs, the evidence is limited. In the case of the second review, the association between the development of childhood eczema and prenatal intake of ω-3 PUFAs was not significant [40]; however, the meta-analysis is not detailed enough because of the absence of subgroup analysis despite the inclusion of six RCTs.

In order to obtain better clinical evidence and to evaluate whether the intake of ω-3 PUFAs during pregnancy can decrease the prevalence of eczema or IgE-associated eczema in the offspring, we took an up-to-date and extensive systematic review of data on omega-3 PUFA intake commenced in pregnancy and the resulting likelihood of developing eczema in childhood.

The guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) were followed for conducting this systematic review of RCTs [41], and the PRISMA 2020 checklist was shown in online supplementary Material 1 (see www.karger.com/doi/10.1159/000526366 for all online suppl. material). A structured and comprehensive literature search was executed using the following databases from inception to March 31, 2021: PubMed, Web of Science, the Cochrane library, and Embase. For each database, the search strategy in looking up relevant publications was (n-3 PUFAs OR omega 3 fatty acid OR n-3 polyunsaturated fatty acid OR n-3 fatty acid OR docosahexaenoic acid OR eicosapentaenoic acid OR fish oil OR fish) AND (pregnant OR perinatal OR prenatal OR antenatal OR maternal OR gestational) AND (child OR offspring OR infant OR adolescent OR youth) AND (Immunoglobulin E-mediated hypersensitivity OR Type I hypersensitive OR atopic dermatitis OR atopic eczema OR allergic OR atopic). Detailed search strategy and syntax were shown in online supplementary Material 2.

The meta-analysis included studies that met the criteria as below: (1) design: RCT study; (2) participants: pregnant women and their offspring; (3) fish oil group: supplementation with capsules rich in ω-3 PUFAs or salmon starting in the gestational stage; (4) control group: supplementation with placebo (including olive oil, soya bean oil, vegetable oil, etc.); (5) outcomes: the incidence of “any eczema” or “IgE-associated eczema.” Two authors independently assessed the published studies for inclusion and reviewed the references of eligible studies for possible additional publications such as pieces of grey literature, dissertations, conferences, and completed but unpublished papers. Any possible discrepancy was discussed and resolved by those two researchers and, if necessary, by third-party arbitration.

For this review, the primary outcome variable was the prevalence of “any eczema” (eczema symptoms, characterized by redness, itching, and swollen and typical skin lesions, with the presence or absence of a positive skin prick test, whether or not the presence or absence of a positive skin prick test or IgE antibodies in blood). The secondary outcome variable was IgE-associated eczema, which was characterized as the existence of eczema symptoms with a positive skin prick test to ≥1 allergen extract (sensitization) or the presence of IgE antibodies in blood.

A standardized data extraction table was used. Extracted data included the first authors, study location, year of publication, participants, study design, family history of allergic disease, start and end time of supplementation, intervention, placebo, length of follow-up, and the results of the outcomes. Any disagreements on the interventions or outcomes during the process of data extraction were discussed and resolved by the review team. During the process, no blinding of the names of the authors, journals, outcomes of the trials, or institutions was employed.

Assessment of the quality of the studies was made by employing the updated tool (Rob2) provided by the Cochrane collaboration group [42]. A couple of authors extracted the data independently and evaluated the study quality. Any possible discrepancy was discussed and resolved by those two researchers and, if necessary, by third-party arbitration.

Relative risk (RR) with a 95% confidence interval (CI) was employed to evaluate the impact of supplementation with ω-3 PUFAs during pregnancy on the occurrence of allergic diseases in children. The hazard ratios and incidence rate ratios were directly considered as RR. The calculated crude RRs from events/total of the included studies were used under circumstances when the RRs were not published.

I² statistics were employed for quantifying the potential variability in the effect estimates among the included studies. When the heterogeneity was not obvious (I² ≤ 50%), a fixed-effect model was implemented to summarize the outcomes. In case of substantial heterogeneity (I² > 50%), the meta-analysis was carried out with a random-effects model [43]. Unpublished studies and ongoing RCT databases were strictly explored by researchers to minimize reporting bias. Egger’s test was carried out for assessing the potential publication bias. A funnel plot was not applied because the number of pooled studies was less than 10. The analyses and forest plots contained in this review were performed using Stata 15.1/SE, and the “risk of bias graph” was created using Review Manager 5.3.

Subgroup analyses were conducted with study location, dose, allergic risk of the offspring, gestational stage at commencing of supplementation, duration of supplementation, and age of offspring. An offspring with a high risk of allergic disease refers to those whose father, mother, or sibling has a medical history of a diagnosed allergic ailment, such as allergic rhinitis, asthma, eczema, or hay fever.

Sensitivity analysis was conducted by omitting one study in turn and recalculating the pooled RRs of the rest of the studies to examine how deletion can affect overall results.

From the database search, 2,310 publications were initially picked out in total. Following the deletion of repeated articles, 1,833 publications were left. Of these, 408 publications were removed after the titles and abstracts were read, resulting in 69 publications requiring reading in full. Sixty full-text publications were rejected for various reasons. Finally, nine publications were selected in qualitative synthesis; seven publications from 6 unique RCTs were selected in quantitative synthesis (meta-analysis) after removing the repetitive studies. The detailed selection process is shown in Figure 1.

In total, 1,926 pregnant women were included in this review, and 1,646 mother-infant pairs completed the RCTs. There were 912 people in the experimental category and 858 people in the control group. The participants were all recruited from clinics or hospitals. The studies contained in this review were performed in different continents of the world: four studies were carried out in Australia [33, 37, 44, 45], four in Europe [34, 35, 38, 46], and one in North America [36].

Participants in four RCTs [33‒35, 37] had a family history of allergic disease; one RCT [38] involved normal pregnant women, and one RCT focused on women with depression [36]. The time of follow-up was different from study to study and ranged from 6 months to 6 years.

Of the 7 publications included in the quantitative synthesis, the study of Palmer et al. [45] and Best et al. [37] are from the same RCT with different follow-up time, resulting in 6 unique RCTs included in this meta-analysis. The salient features and outcomes of these publications are elaborated in Table 1.

Participants in five RCTs were divided into two groups, respectively, an experimental group (supplemented with salmon or fish oil capsules) and a control group [33‒35, 37, 38] (supplemented with olive oil, soya bean oil, vegetable oil, or no supplement). One RCT studied three groups, a docosahexaenoic acid-rich fish oil group, an eicosapentaenoic acid-rich fish oil group, and a soya oil placebo group [36], and we combined the data for the two fish oil groups into a single intervention group for the analysis. The amount of daily ω-3 PUFAs supplemented ranged from 900 mg to 3,700 mg.

The supplementation commenced in the pregnancy term, ranging from 12- to 24-week gestational age, and it was stopped at the time of delivery in six studies [33, 35‒37, 44, 45] and after delivery in three studies [34, 38, 46]. The duration of the intervention ranged from 10 weeks to 29 weeks after statistical calculation.

The primary outcome of “any eczema” was diagnosed by a physician, pediatrician, or nurse in 4 RCTs [33‒35, 38], reported by a parent in 1 RCT [37], and the remainder used validated National Health Interview Survey questions [47] in 1 RCT [36]. The secondary outcome of “IgE-associated eczema” reported in 5 publications (from 2 unique RCTs) was assessed by experienced researchers, with skin prick test performed by pediatric nurses [34, 37, 44‒46]. Detailed diagnostic definition methods of the outcomes in each publication are presented in Table 2.

None of the publications relayed fish oil supplementation to have a protective effect on “any eczema,” and one study conducted by Berman in 2016 even reported a substantially increased risk of childhood eczema in fish oil supplementation group [36]. None of the studies reported benefits of intake of fish oil in pregnant women and “IgE-associated eczema” in childhood except one conducted by Furuhjelm in 2009 [46].

We performed a risk of bias assessment, such as randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, selection of the reported result, and overall bias, employing the updated tool (Rob2) provided by the Cochrane collaboration group for RCTs [42]. Individual item was assessed as low-risk, high-risk, or some concerns. The assessed risk of bias for each is article shown in Figure 2.

A total of 7 publications (from 6 unique RCTs) were eventually included in the meta-analysis, one of which reported not only the risk of “any eczema” but also the incidence of “IgE-associated eczema” in children [34].

Six publications (from 6 unique RCTs) documented an interrelation between intake of ω-3 PUFAs and the incidence of “any eczema” [33‒38]. The pooled data of six RCTs suggested that in comparison with the placebo group, the occurrence of “any eczema” in the offspring was not reduced in the experimental group (RR: 1.09, 95% CI: 0.82~1.46, p = 0.54, I² = 55.4%) (shown in Fig. 3a).

Two unique RCTs with the longest follow-up assessed the incidence of “IgE-associated eczema” [34, 37], but neither trial described a decreased risk in the offspring in experimental group. The pooled data revealed that the consumption of ω-3 PUFAs in pregnancy had no significant protective effect on this outcome in comparison with the placebo (RR: 0.67, 95% CI: 0.29~1.57, p = 0.34, I² = 63.0%) (shown in Fig. 3b).

In the pooled data, there existed no notable differences among the experimental group and control group in studies carried out in any subgroups of “any eczema”. The detailed results of the “any eczema” subgroup analysis are presented in Table 3.

In the pooled data, the risk of “IgE-associated eczema” was pronouncedly decreased in the children ≤3 years of age in the experimental group (RR: 0.70, 95% CI: 0.50~0.96, p = 0.026, I² = 38.1%) in the subgroup analysis. Subgroup analyses for the secondary outcome were only stratified by “the age of offspring” because of the small number of RCTs included. The detailed results of the “IgE-associated eczema” subgroup analysis are presented in Table 4.

The sensitivity analysis demonstrated no significant changes in the outcome of “any eczema” whose RRs ranged from 1 to 1.22 (RR: 1.00, 95% CI: 0.81–1.25, RR 1.22, 95% CI: 0.82–1.81) (shown in online suppl. Material 3-1). The sensitivity analysis of ω-3 PUFA supplementation during pregnancy on the incidence of “IgE-associated eczema” was shown in online supplementary Material 3-2.

No statistically significant publication bias was suggested by the outcome of Egger’s test (p > 0.05) and Begg’s (p > 0.05) test for each outcome in this meta-analysis.

Generally speaking, the current meta-analysis showed that prenatal ω-3 PUFA supplementation was ineffective in reducing infantile “any eczema” or “IgE-associated eczema,” except in IgE-associated eczema in children under 3 years of age. Subgroup analyses showed that fish oil supplementation during pregnancy had no protective effect on “any eczema,” which is consistent with the previous two reviews [39, 40]. Significantly, the subgroup analysis showed a notable inverse exposure-response relationship among infantile IgE-associated eczema in the first 3 years of life and maternal intake of ω-3 PUFAs in pregnancy. Since only one study was conducted on IgE-associated eczema in children over 3 years old, this subgroup cannot be analyzed systematically [37].

It is worth noting that there is no unified standard for the classification and definition of eczema because of the variable locations and times of onset [52‒54]. Objective detection methods, for instance, a positive skin prick test and/or raised IgE, are requested for the clinical diagnosis or classification of atopic eczema [51, 52]. The subgroup analysis showed that prenatal ω-3 PUFA supplementation is beneficial to prevent IgE-associated eczema in the initial 3 years, which, in addition to the embryonic stage, is a critical age for immune development and growth [53, 54]. Therefore, interventions of ω-3-PUFAs during pregnancy could have immunoregulatory effects and influence the occurrence of IgE-associated eczema.

The incidence of eczema is subjected to the difference in geographical locations, living environment, and race [55, 56] as well as to prenatal stress and the emotional state of mothers [57]. A study has shown that female babies are more likely to develop eczema [58]. An observational study has shown that children with parental allergic diseases were at increased risk of eczema [59]. A survey based on the prevalence of eczema in the USA revealed that young adulthood is the age of the peak onset of eczema and that 54% of adults with active eczema reported onset in adulthood [51]. The significance of the meta-analysis is to some extent affected by the fact that these potential confounding variables cannot be unified in a meta-analysis. Significant heterogeneity was observed in this meta-analysis, which can be explained by various regimens, doses, study location, duration, follow-up time, center settings, populations enrolled, etc.

Advantages of this review should be noted. We used a detailed search strategy for identifying articles, aiming for comprehensive coverage of all the pertinent literature, without limitations on publication status or language. The meta-analyses also included the longest available follow-up data for each outcome. However, there are potential limitations of the review that require more consideration. First, there are limited studies assessing the conclusion they obtained, resulting in a lack of sufficient supporting evidence. Second, there were many confounding factors that could not be discounted for our analysis, such as the quality of the air environment at home or outside, sensitization to milk, and whether there were house pets at home [60]. Thus, the results of sensitivity analysis showed high heterogeneity and low stability. In addition, the relatively small number of studies included in the paper may reduce the generalizability and universality of the conclusion. We will keep on observing the effects of the ω-3 PUFA supplement during pregnancy on eczema in children.

What should be mentioned is that ω-3 PUFAs seem to have anti-inflammatory properties and reduce oxidative stress, circulating inflammatory markers [61]. Required for the maintenance of normal brain function, ω-3 PUFAs can also improve working efficiency and long-term memories [16]. What should also be mentioned is the limited number of RCTs that seem to drive the results the most, with long duration of ω-3 PUFA supplementation or large sample size [27, 37]. More well-designed RCTs with large sample sizes were recommended to be carried out to better comprehend the relationship between supplementation with ω-3 PUFAs in pregnancy and the risk of eczema or other relevant diseases in children.

The available evidence showed that fish oil supplementation during pregnancy does not reduce “any eczema”/“IgE-associated eczema” in all children, and there may be a subgroup-specific effect under 3 years of life in reducing the incidence of IgE-associated eczema. However, the findings from subgroup analysis need further and careful consideration because of significant heterogeneity and inevitable sources of bias among studies.

The authors have no conflicts of interest to declare.

No funding was obtained for this study.

Huili Wang and Haili Jiang: conceived the systematic review and conducted searches. Yin Jia and Yafang Huang: assessed inclusion and extracted data, assessed risk of bias, and performed the meta-analyses. Yin Jia: writing the review and editing the manuscript. All authors critically reviewed the manuscript for content and approved the final manuscript as submitted and gave consent for publication of this article.

Y.J. and Y.H. both contributed equally to this work.Edited by: H.-U. Simon, Bern.