Atopic dermatitis (AD) is a chronic inflammatory skin disease, mainly presented serve itching and recurring eczematous lesions (1). The prevalence of AD is up to 10–20% among adults and children in developed countries (2). It is the most prevalent non-fatal skin disease, imposing a significant psychosocial expense on individuals and their families, and raising the risk of allergic diseases and other immune-mediated inflammations as well as mental health issues. The pathophysiology of AD is complex, involving a strong genetic susceptibility, epithelial barrier dysfunction, immune dysregulation, and abnormalities in the skin microbiome (3). Barrier dysfunction in AD is associated with lipid abnormalities in the stratum corneum, which may be caused by the disturbance between degradation proteases and lipid synthesis enzymes (4). Recent study showed the lack of sebum may develop AD via IL-33 signal pathway initiating skin inflammation and mentioned the role of Microbial dysbiosis in AD (5).

Fatty acids are a series of carboxylic acids with an aliphatic chain, saturated or unsaturated, including saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs) as well as polyunsaturated fatty acids (PUFAs). In either form, fatty acids are an essential energy source of the human diet and an important structural component of cells (6). It has been proved that dietary intake of fatty acids have a relationship with the development of inflammatory and immune system (7).

Essential fatty acids (EFAs) are a type of PUFAs that must be obtained through diet because it cannot be synthesized by human bodies (8). Dermatitis (scaling and dry skin) and increased TEWL are two typical clinical signs of EFAs deficiency (9, 10). EFAs may have beneficial effects in health and in the control of chronic diseases (11). The two types of EFAs include an omega-6 (n-6) fatty acid (linoleic acid) and an omega-3 (n-3) fatty acid (alpha-linolenic acid). EFAs deficiency in the skin is one suspected factor involved in AD (12). An altered stratum corneum lipid composition signifies the importance of ceramides and free fatty acids for the impaired skin barrier function in AD (13). Although studies showing altered PUFA levels in the development of AD, it cannot be ruled out that diets or other social factors result in the reverse causation.

Nutrient supplementation, including fatty acids, can prevent the development of AD or reduce the severity of AD in newborns to children younger than 3 years suggesting the role of fatty acids in AD (14). However, there is no evidence on whether PUFAs have causal effects on AD.

Nowadays many advanced methods to identify causal inference are given rise to use because randomized Controlled Trails (RCTs) are easily influenced by human factors and social environment, which can help to guide whether RCTs are theoretically required. Mendelian randomization (MR) can avoid reverse causation and the interference of weaken confounding factors to address causal interference regarding how exposures affect the different outcomes by using genetic instrumental variants (15). Three principal assumptions should be met. The first hypothesis is that genetic instrumental variants are closely related to exposure factors. The second hypothesis is that genetic instrumental variants should strictly not be related to any confounding factors. The third hypothesis is that genetic instrumental variants affect the outcome only through exposure factors and not through other pathways. The present study applied a two-sample MR approach to figure out whether fatty acids levels are related to AD and which PUFAs are most important.

In this study, MR was used to identify the casual link between each kind of fatty acids and AD. We observed that a higher level of the n-3 fatty acids closely associated to a decreased chance of developing AD, suggesting a protective effect of the n-3 fatty acids, including DHA, in AD. MVMR analyses also demonstrated that the independent causal relationship between n-3 fatty acids and AD, indicating that the effects of n-3 fatty acids are probably independent of (and without being mediated or confounded by) n-6 fatty acids, PUFAs or SFAs.

In the MVMR analyses, the p value of direct effects of n-3 fatty acids on AD were less significant, most likely because these conditional analyses had less power. Given that the conditioned estimate for n-3 fatty acids was relatively unchanged, it appears that the biosynthetic pathway for n-3 fatty acids may be significant in relation to the risk of AD.

The leave-one-out analyses showed that these effects for n-3 fatty acids were mainly driven by one SNP (rs174546) which is in the FADS gene cluster, thus we hypothesize that this relationship may be considerably mediated by desaturation steps occurring in PUFA biosynthesis. Our research offers meaningful information for identifying new biomarkers and comprehending the pathophysiological mechanisms underlying AD.

In real world, it is hard to ensure how fatty acids related to AD. For example, Fujii, M. et al. reported that deficiency of PUFAs is mainly responsible for AD (26). Higher prenatal n-6 fatty acids exposure may raise the odds of AD in a child with maternal inheritance (27). According to some studies, using black currant seed oil, which is with high amounts of EFAs, as a dietary supplement may help to prevent AD (28).

In the absence of trials, it is therefore difficult to determine if changes at any stage along the fatty acids manufacturing pathway would lead to decreased levels of n-3 fatty acids or increasing amounts of particular fatty acids. For instance, dysfunction of essential enzymes that mediate the synthesis of fatty acids will result in the unstable ratio of these fatty acids. Therefore, a chronic pro-inflammatory state could result from relative disordered lipid metabolism (29). N-6 and n-3 fatty acid have a relationship of competitive inhibition, so it is hard to say which has the great influence on AD. Recent studies reported that low-ratio n-6/n-3 fatty acids supplementation significantly decreased the levels of certain inflammation cytokines (30), indicating the importance of alteration in fatty acids synthesis pathway. Meanwhile, our results reveal the importance of rs174583 in fatty acids desaturase 2 (FADS2) gene for the regulation of fatty acids in AD. Genetic variation in the FADS locus was strongly associated with biologically important lipids (31). It is known that FADS gene polymorphisms are related with the amounts of fatty acids in human tissues, which may modify how fatty acid-related atopic diseases exist symptoms (32). While the mechanics of actions of n-3 fatty acids on AD are probably complex, identifying independent impacts of the FADS gene could illuminate mechanistic pathways in AD etiology.

MR studies are also useful for informing which supplements may have the greatest risk-related effects. As indicated by our study, protective effects of PUFA were demonstrated, with significant indication for the n-3 fatty acids, particularly DHA. Since the exploration for PUFA and AD are continually developing, MR research could provide more insight on which PUFA should be noticed.

The estimated effects of n-3 fatty acids on AD using MR method should be explanatory. Actually, effects discovered using MR mainly resulting from genetic inheritance. Therefore, in traditional traits, the adulthood gained fatty acids supplements over a relatively short period of time are unlikely to have any significant effect. Although our study provides evidence that a higher level of n-3 fatty acids may reduce the chance of developing AD, A placebo-controlled study found no effect of EFAs supplements in AD (33). Kong WS et al. identified the long-chain saturated fatty acids milk diet would develop eczema, accompanied by increased gut type 3 innate lymphoid cells (34). Another cohort study showed that the lower n-3 fatty acid levels in maternal plasma were associated with the prevalence of atopic disorder (35).

Meanwhile, although tissue and blood fatty acids have been used as biomarkers for decades (36), the test methods are developing (37, 38). Further studies regarding the normal range of EFAs are required to assist clinical researchers in the relationship of fatty acids and diseases.

Our studies have several strengths. First, MR analysis is used that could not be influenced by reverse causation and confounding in comparison to ordinary traits, and then we used the recently largest GWAS datasets both chosen from European ancestry, thereby ruling out the confounding variable of population. Meanwhile, using F statistics to identify the genetic instrumental variants are strong enough for MR analysis. Finally, there are different methodological approaches have been applied to prove the estimate effects, including univariable Mendelian randomization, multivariable Mendelian randomization and single-SNP analyses.

However, several limitations need to be listed. Firstly, although the F-statistics of the instruments for our univariable analyses were not weak, weak instrument bias may exist in our MVMR analyses. Second, comprehensive analysis of fatty acids and AD was constrained because we can only obtain now available GWAS data. For example, MVMR analyses can be used to estimate the direct effects of each fatty acids independent of others, but there are no datasets could be able to estimate the effect of PUFA: SFA ratio which may be critical underlying AD pathogenesis. And although the largest GWAS of AD up to now was obtained, the clarified information about the patients was not provided. Therefore, there is no design groups about AD’s stage and severity. Third, The GWAS effect sizes were calculated using fatty acids plasma levels, as a result the fatty acids levels in epidermal layer might not be accurately reflected that might more conspicuously influence chance of developing AD. Finally, based on previous studies, the relationship between diet and dermatologic diseases is not clear (39), further studies are necessary to determine whether PUFAs have effects on AD in the real world.

In this study, we applied MR to estimate the effect of fatty acids on AD. Our results show a genetically protective role of n-3 fatty acids in the development of AD. These results also imply that patients with AD are more likely to have genetic variation in the FADS locus. Thus, it is important for researchers and dermatologists to pay more attention on omega-3 fatty acid.

Publicly available datasets were analyzed in this study. This data can be found at: https://gwas.mrcieu.ac.uk/

J-YL wrote the draft manuscript and prepared the data. L-JM revised the manuscript. J-PY and PY designed the manuscript. B-XB designed and revised the manuscript. All authors have contributed to the article and approved the submitted version.

This project was supported by the National Natural Science Foundation of China (NSFC) grant number 81872513.

The authors declare that the study was performed without any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.