Causal role of vitamin E in atopic dermatitis risk: A Mendelian randomization study

Abstract Prior studies suggested that vitamin E might be beneficial in alleviating atopic dermatitis, but confirming a causal link was hindered by limitations such as sample sizes and unaccounted confounders. The present study aimed to clarify this through Mendelian randomization (MR) analysis. GWAS summary statistics was obtained from public databases encompassing a study on vitamin E and two studies related to atopic dermatitis. Two sets of instrumental variables (IVs) were selected using lenient (p < 1e‐5) and strict (p < 5e‐6) thresholds for separate MR analyses. Inverse variance weighted (IVW) was used as the primary MR method, supplemented by six additional MR methods, and followed by a meta‐analysis to consolidate the impact of vitamin E on atopic dermatitis from two independent studies. Furthermore, various sensitivity tests were performed to assess the reliability of the MR results. A meta‐analysis of IVW analyses deriving from two different atopic dermatitis cohorts under lenient IV selection thresholds demonstrated that vitamin E exhibited a significant lowering risk of atopic dermatitis effect (OR = 0.817, 95% CI: 0.673–0.991, p = .041), which was validated under strict IV selection thresholds (OR = 0.822, 95% CI: 0.709–0.954, p = .010). In addition, six other MR methods remained parallel to IVW (OR > 1). Multiple sensitivity tests showed that MR analyses were not affected by heterogeneity and horizontal pleiotropy. Overall, this MR study supported vitamin E reducing the risk of atopic dermatitis. Consequently, maintaining an adequate intake of vitamin E could potentially serve as an effective preventive measure against atopic dermatitis.

Despite strides in comprehending its mechanisms, managing symptoms and averting flare-ups remain challenging, highlighting the urgent need for further research.Exploring modifiable risk factors, preventive strategies, and evidence-based dietary guidelines is crucial to mitigate the incidence and severity of this condition.
Vitamin E, a fat-soluble antioxidant, has long been acknowledged for its potential in promoting skin health, evident in its presence in various cosmetic products and its protective role against solar radiation damage (Keen & Hassan, 2016).Its antioxidant properties have led to suggestions about its potential in managing chronic inflammatory skin conditions, including atopic dermatitis (Liu et al., 2021).
Nevertheless, conflicting findings persist.Studies have shown inconclusive evidence, such as maternal intake of vitamin E during pregnancy not significantly affecting the risk of atopic dermatitis in offspring (Miyake et al., 2010), and no significant association found between vitamin E and atopic dermatitis in children (Nwaru et al., 2014).Consequently, there is insufficient evidence to support recommending vitamin E for atopic dermatitis treatment (Shi & Lio, 2019).Moreover, prior randomized controlled trials (RCTs) were conducted among individuals already diagnosed, leaving uncertainties about the impact of different levels of vitamin E supplementation on the incidence risk of atopic dermatitis.Therefore, the causal relationship between vitamin E and atopic dermatitis remains elusive.
Mendelian randomization (MR) is a statistical method that uses genetic variants as instrumental variables (IVs) to infer causal relationships between an exposure (in this case, vitamin E) and an outcome (atopic dermatitis) (Sanderson et al., 2022).This method has the advantage of minimizing confounding and reducing the possibility of reverse causality, as genetic variants are randomly allocated at conception and are thus unrelated to lifestyle factors (Swerdlow et al., 2016).Single nucleotide polymorphisms (SNPs) are the most utilized form of IVs for implementing MR studies (Sanderson et al., 2022).An important feature of MR studies is their capacity to evaluate the effects of varying levels of lifelong exposure on outcomes (Labrecque & Swanson, 2019).Particularly, in situations where conducting RCTs poses challenges due to cost or ethical constraints, MR studies offer a valuable alternative for assessing causal relationships.The abundance of genome-wide association study (GWAS) summary statistics has markedly facilitated MR studies, ensuring robust statistical analyses due to substantial sample sizes (Boehm & Zhou, 2022).The present study aimed to determine the causal effect of vitamin E on atopic dermatitis through MR analysis.

| Study design
In the present study, a two-sample MR analysis was conducted based on publicly available GWAS summary statistics using the "TwoSampleMR" package (https:// mrcieu.github.io/ TwoSa mpleMR/ ) in R software.This MR study was performed based on three MR core assumptions (Figure 1a): (i) IVs are strongly correlated with exposure; (ii) IVs are not associated with confounding factors; and (iii).IVs are not directly associated with outcome (Boehm & Zhou, 2022).Figure 1b presents the general flowchart of the present MR study.Specifically, one GWAS summary statistic for vitamin E and two GWAS summary statistics for atopic dermatitis were initially obtained from the publicly available databases.Subsequently, IVs representing vitamin E were screened based on three MR core assumptions.Next, the causal effect of vitamin E on atopic dermatitis (from two different outcome sources) was respective estimated and the results were integrated by meta-analysis with random effects models.Finally, diverse sensitivity tests were performed to evaluate the reliability of the MR results.Subsequent portions of this section will elucidate all the procedures in detail.

| IVs selection and data harmonization
IVs were screened based on three core assumptions (Figure 1a).To fulfill core assumption 1, all SNPs associated with vitamin E were first screened.Due to the lack of SNPs meeting the genome-wide significance threshold (p < 5e-8), for the feasibility of analysis, two additional sets of thresholds ([i] lenient threshold: p < 1e-5 and [ii] strict threshold: p < 5e-6) were utilized to screen out two sets of SNPs related to vitamin E. Subsequently, the SNPs in linkage disequilibrium were removed to ensure the independence of each SNP (based on the threshold of r 2 < 0.001 within 10,000 kb).To fulfill the second core assumption, traits associated with the remaining SNPs were comprehensively examined from the PhenoScanner V2 online platform (Kamat et al., 2019), and SNPs associated with confounders were eliminated.To meet the third core assumption, the association of SNPs with vitamin E should be stronger than their association with atopic dermatitis.Subsequently, the information of the selected SNPs in the GWAS summary statistics for atopic dermatitis was extracted and harmonized, thus ensuring that effect alleles and other alleles were consistent among the summary statistics for exposure and outcome.Furthermore, all palindromic SNPs will be excluded.After all the above screening procedures, the remaining SNPs were included as IVs.Finally, the F-statistics for each IV were computed to evaluate their strength, and only those with F-statistics >10 were able to mitigate the estimation bias arising from weak IVs (Zheng et al., 2017).

| Statistical analysis
In the present study, we conducted a series of MR analyses using two distinct IV selection thresholds (p < 1e-5 and p < 5e-6) to evaluate the effect of vitamin E on atopic dermatitis.For each threshold, the effects of vitamin E on atopic dermatitis were assessed using two independent datasets: the Budu-Aggrey et al. dataset and the FinnGen dataset.The primary MR method applied in this analysis was the inverse variance weighted (IVW) method (Pagoni et al., 2019).Additionally, several complementary MR methods were utilized to reinforce the IVW results, namely MR-Egger, MR-PRESSO, maximum likelihood, simple mode, weighted median, and weighted mode (Bowden et al., 2016;Burgess & Thompson, 2017;Hartwig et al., 2017;Milligan, 2003;Verbanck et al., 2018).

| Selection of IVs
After eliminating the SNPs in linkage disequilibrium, 31 and 12 independent SNPs associated with vitamin E were identified at the thresholds of p < 1e-5 and p < 5e-6, respectively (Table S1).Next, through the PhenoScanner V2 platform, rs12128707 was found to be associated with educational attainment, rs12165526 with cetraben emollient cream treatment, and rs979218 with epaderm ointment treatment was associated (Table S2).Since they could be potential confounding factors, these three SNPs were excluded from subsequent analyses.Tables S3 and S4 show the SNPs filtering process for obtaining IVs that were used for MR analysis.Ultimately, under the threshold of p < 1e-5, 27 and 26 IVs, respectively, were utilized to assess the causal impact of vitamin E on atopic dermatitis in the Budu-Aggrey et al. and FinnGen R9 cohorts (Table S5).
Meanwhile, under the threshold of p < 5e-6, 10 IVs were used to evaluate the causal impact of vitamin E on atopic dermatitis in both cohorts (Table S6).In addition, the F-statistics of all IVs were >10.
Similarly, based on a strict IVs selection threshold (p < 5e-6), a metaanalysis of IVW results indicated that vitamin E significantly reduced the risk of atopic dermatitis (OR = 0.822, 95% CI: 0.709-0.954,p = .010),while the results of the other six MR methods remained parallel to IVW (Figure 2b).Overall, MR analyses based on both lenient and strict IVs selection threshold demonstrated that vitamin E reduces atopic dermatitis risk.

| Results of sensitivity tests
Next, various sensitivity tests were performed to assess the reliability of the MR analysis.Cochran's Q test showed that there was no significant heterogeneity (p > .05)among the IVs screened based on both lenient and strict thresholds (Table 1).In addition, both the MR-Egger intercept test and the MR-PRESSO global test showed that MR analyses performed using IVs screened at either threshold were not affected by significant horizontal pleiotropy (p > .05)(Table 2).
Finally, the leave-one-out sensitivity test showed that there were no leading IVs that would significantly impact the overall MR results (Figure 3).Therefore, the sensitivity tests indicated that the MR results were robust with no heterogeneity or horizontal pleiotropy.trial (RCT) involving 96 subjects with atopic dermatitis, the intake of 400 IU/day of vitamin E was compared to a placebo, revealing a significant improvement in subjective symptoms and serum IgE levels (Tsoureli-Nikita et al., 2002).Another RCT conducted on 70 participants with mild-to-moderate atopic dermatitis, it was found that oral administration of 400 IU of vitamin E significantly improved symptoms and quality of life compared to the placebo group (Jaffary et al., 2015).

| DISCUSS ION
Nevertheless, these RCTs have been limited by small sample sizes, and could also be influenced by confounding factors that are difficult to eliminate.The present study demonstrated the preventive effect of vitamin E on atopic dermatitis by MR analysis based on large-sample GWAS summary statistics, thereby strengthening the prior findings.
Whether based on lenient IVs selection thresholds or strict IVs selection thresholds, MR analyses indicated compelling results for vitamin E reducing atopic dermatitis risk.Interestingly, employing a meta-analysis of IVW results as our primary analytical approach revealed similar findings across two distinct thresholds (OR = 0.817 at the lenient threshold vs. OR = 0.822 at the strict threshold).The result needs to be interpreted rationally.Firstly, vitamin E levels were estimated from participants' dietary questionnaire responses in the UK Biobank cohort, with a median intake of 9.366 mg and a standard deviation of 5.402 mg.Secondly, this MR study simulated the causal effect of a lifelong high level of vitamin E on the risk of atopic dermatitis.
Therefore, it can be understood that an increase of 5.402 mg in daily average vitamin E intake throughout life could potentially reduce the risk of developing atopic dermatitis by approximately 18%.Therefore, it is more essential for individuals with high skin disorder risk factors to focus on consuming adequate amounts of vitamin E.
The plausible effect of vitamin E against atopic dermatitis can be attributed to multiple mechanisms.Recent research on atopic dermatitis has highlighted the significance of oxidative stress in contributing to cellular damage and tissue inflammation (Ji & Li, 2016).Reactive oxidants induce lipid peroxidation in epidermal keratinocytes, leading to cellular damage and upregulation of pro-inflammatory cytokines through the nuclear factor-κΒ pathway (Kruk & Duchnik, 2014;Wullaert et al., 2011).In addition to its antioxidant properties, vitamin E has demonstrated the ability to mitigate inflammation and enhance the expression of keratinocyte differentiation markers (Hayashi et al., 2012;Kato & Takahashi, 2012).Furthermore, the water-soluble analog of vitamin E, Trolox, has exhibited the capacity to reduce lipid peroxidation and concurrently increase the activities of key enzymes in keratinocytes, suggesting its potential in combating atopic dermatitis (Gehin et al., 2006).
Overall, this study provides evidence for vitamin E intake for the prevention and treatment of atopic dermatitis.Sources of vitamin E in the diet include various common, specialty, and underutilized edible oils, as well as cereals, legume seeds, animal-derived products, wild plants, and edible flowers (Shahidi et al., 2021).Therefore, it is crucial to emphasize the consumption of these foods to maintain an adequate daily intake of vitamin E, as deficiencies in this nutrient can lead to inflammation (Roth-Walter et al., 2024).Several strengths characterize this study.First, the MR analysis is based on large-sample GWAS summary statistics, ensuring statistical power.In particular, when using a lenient threshold, a relatively large number of IVs resulted in close to 90% statistical power for IVW (Table S5).Second, the summary statistics for exposure and outcomes are derived from independent European cohorts, reducing the risk of a type I error.Third, two diverse IV selection thresholds, multiple MR methods, and meta-analysis integration of results for two sets of outcomes were employed, thereby increasing the reliability of the present study.
Nevertheless, this study possesses certain limitations that necessitate clarification.First, this study was conducted based on GWAS summary statistics derived from European individuals, therefore the generalizability of the results to other ethnical populations was uncertain.Second, since the MR study was performed using summarylevel statistics rather than individual-level statistics, the analysis was restricted to assessing the impact of vitamin E on the entire study population, and stratified analyses were not available.Third, since vitamin E intake was estimated based on participants' responses to a dietary questionnaire, it may be affected by memory bias and response bias.
Since atopic dermatitis is a binary variable, the results of Mendelian randomization (MR) are presented using odds ratios (OR) and 95% confidence intervals (CI).An OR > 1 indicates that vitamin E increases the risk of atopic dermatitis, while an OR < 1 indicates that vitamin E decreases the risk of atopic dermatitis.The MR results are considered statistically significant only when the meta-analysis result of the IVW is p < .05,and the OR values from all other supplementary MR methods are consistent with the IVW result.Subsequently, some sensitivity tests were performed.Cochran's Q test was used to assess heterogeneity.MR-Egger intercept test and MR-PRESSO global test were used to assess horizontal pleiotropy.Leave-one-out test was employed to assess stability.

F
Overview of the present MR study.(a) Three core assumptions of the MR study.(b) The general flow of the present MR study.
This study is the first to use MR analysis to elucidate the role of vitamin E in reducing the risk of atopic dermatitis, thus adding confidence to previous clinical studies suggesting that vitamin E represents a good anti-atopic dermatitis treatment.Numerous previous studies have indicated the potential benefits of vitamin E in managing atopic dermatitis.In a randomized controlled F I G U R E 2 Results of the MR analysis.(a) Results of seven MR analysis approaches based on a lenient IVs selection threshold (p < 1e-5).(b) Results of seven MR analysis approaches based on a strict IVs selection threshold (p < 5e-6).
Horizontal pleiotropy assessed by the MR-Egger intercept test and the MR-PRESSO global test.