The genetic interaction between Atopic dermatitis and Celiac disease: A bidirectional two-sample Mendelian randomization study

doi:10.21203/rs.3.rs-4861154/v1

Background: Multiple previous observational studies have reported inconsistent results in investigating the association between atopic dermatitis (AD) and celiac disease (CD).Observational studies may be inadequate for elucidating potential causal relationships between AD and CD due to the issues of reverse causation and confounding variables. The existence of a bidirectional causal relationship between AD and CD remains unclear.

Objective: To investigate the potential bidirectional causal relationship between atopic dermatitis and celiac disease.

Methods: Utilizing the summary statistics from large-scale Genome-Wide Association Studies (GWASs) in European populations, we examined the potential bidirectional causal effects between AD and CD through a bidirectional Mendelian randomization analysis. The inverse-variance weighted (IVW) method was employed as the primary analytical approach. The MR-Egger, weighted median, simple mode, and weighted mode were utilized as complementary approaches.

Subsequently, a series of sensitivity analyses were conducted to evaluate the robustness and reliability of the MR findings.

Results: Our findings revealed that the IVW results indicated a significant increase in the prevalence of AD in individuals with genetically predicted CD (OR = 1.46, CI = 1.07-1.96, P_IVW = 0.014) and conversely (OR = 1.05, CI = 1.03-1.08, P_IVW = 4.80x10^-6). Furthermore, subsequent sensitivity analyses addressing heterogeneity, pleiotropy, and other potential biases substantiated these results.

Conclusions: We established a bidirectional causal relationship between atopic dermatitis and celiac disease through a two-sample bidirectional MR study. Therefore, early screening and preventive measures are recommended for patients with atopic dermatitis and celiac disease to mitigate long-term complications and reduce additional burdens.

Atopic dermatitis

Celiac disease

Causal association

genetic prediction

Mendelian randomization

Atopic dermatitis (AD) is a prevalent chronic inflammatory skin disorder. It is estimated to affect approximately 10% of adults and 20% of children in some developed countries¹. The underlying mechanisms for its occurrence are complex, encompassing epidermal barrier dysfunction, microbiome alterations, and systemic immune dysregulation², which are regulated by an interplay of genetic and environmental factors³. Previous studies have suggested that AD may be linked to various metabolic syndromes and several autoimmune diseases⁴. Celiac disease (CD) is an autoimmune disorder with a global prevalence of 1–2%, characterized by a genetic predisposition to gluten-induced immune responses⁵. In recent years, an increasing number of evidence has indicated that celiac disease impacts skin-related diseases⁶. Furthermore, compared to healthy populations, CD patients exhibit a higher susceptibility to suffer from other immune-mediated diseases, including inflammatory skin disorders^7,8.

Several prior observational studies have investigated the association between AD and CD, yielding inconsistent results^9–13. A valid causal relationship between AD and CD cannot be established due to variations in study methodologies, differences in prevalence time points, confounding factors, and the potential for reverse causation. Thus, the existence of a mutual causal relationship between AD and CD remains unclear.

Considering the limitations inherent in observational studies, we investigated the potential causal relationship between AD and CD by employing a Mendelian randomization (MR) study. This validated statistical methodology is based on the principle of the "random assignment of parental alleles to offspring", akin to randomized controlled trials (RCTs)^14–16. MR method utilizes genetic variants as instrumental variables (IVs) and relies on the principle of random and independent distribution. By leveraging genetic variants as IVs and adhering to the principle of random and independent distribution, MR research has a significant advantage that can effectively exclude the influence of confounding factors and reverse causality. Consequently, MR method can more effectively circumvent the limitations of traditional observational studies regarding causal inference. In this study, we acquired sufficiently large genome-wide association study (GWAS) data from relevant databases to conduct a two-sample bidirectional MR analysis, aiming to infer the potential causal effect between AD and CD.

2.1 Study design

To investigate the potential causal relationship between AD and CD, we integrated datasets from GWASs and conducted a two-sample bidirectional MR analysis. All original study data utilized were sourced from GWAS databases and contained no personal information. Consequently, no additional ethical approval or informed consent was necessary for this study. The primary design framework of this study is depicted in Fig. 1.

2.2 Data sources

AD related GWAS data to were sourced from a study by Sliz et al. ¹⁷(2022), publicly accessible in the GWAS database (https://gwas.mrcieu.ac.uk/) under the identifier: ebi-a-GCST90027161. The AD dataset encompasses approximately 16,121,213 single nucleotide polymorphisms (SNPs) and 796,661 individuals.CD associated GWAS data were obtained from a study by Trynka et al.¹⁸ (2011), publicly accessible in the GWAS database (https://gwas.mrcieu.ac.uk/). The diagnostic criteria for CD include clinical diagnostic guidelines, histopathology such as small bowel biopsy, and serologic testing. The participants in this dataset are from various countries, including the United Kingdom and Finland. The dataset identifier is ieu-a-1058. The CD dataset encompasses approximately 38,037 SNPs and 24,269 European individuals, including 12,041 cases and 12,228 controls.

2.2 IVs identification

Specific screening steps are illustrated in Fig. 1, ensuring that instrumental variables (IVs) effectively serve as proxies to investigate the causal relationship between the exposure and the outcome. Three core assumptions must be met to select robust IVs¹⁹: ①Relevance: the chosen IVs should exhibit a strong correlation with the exposure factor under investigation, ensuring they can adequately proxy the exposure of interest²⁰.② Independence: the effect of IVs on the outcome should be independent of other risk factors or confounders, thus excluding potential confounding influences²⁰. ③Exclusivity: the IVs should influence the outcome solely through the exposure, excluding any pleiotropic effects²¹. Appropriate instrumental variables were selected from two distinct GWAS pooled datasets.

For the selection of instrumental variables (IVs), the specific steps were as follows(Fig. 1). Initially, we identified single nucleotide polymorphisms (SNPs) strongly associated with CD using a screening criterion of P < 5×10^− 8. In order to obtain more SNPs for AD, we applied a less stringent criterion of P < 5×10^{− 6 22}. Subsequently, using European genome data, we applied parameters r² < 0.001 and kb = 10000 to eliminate linkage disequilibrium (LD). Additionally, we ensured a strong association between IVs and the exposure by calculating the SNP F-statistic using the formula F=(N-2)×R²/(1-R²)^23,24, where R² represents the proportion of variability in each IV, determined by R² = 2×beta2×EAF×(1-EAF)²⁴.If the F-statistic is less than 10, the IV is considered weak and may bias MR estimates²⁵.SNPs with F values below 10 were excluded, while those with F values above 10 were retained for subsequent MR analyses²⁰. To further minimize the influence of potential pleiotropic effects due to confounders, we excluded SNPs that might affect the outcome using the Phenoscanner database (http://www.phenoscanner.medschl.cam.ac.uk/)²⁶. Additionally, to further exclude the influence of horizontal pleiotropy on the outcome, we calculated the p-value for pleiotropy significance for each initially included SNP using the MR-PRESSO test. SNPs with p < 0.05 were considered pleiotropic and were excluded²⁷. Finally, to prevent alleles from confounding the results, palindromic SNPs were removed through palindromic sequence detection²⁸.

2.2. Statistical analysis

The study employed the IVW method as the primary analytical approach. A p-value of P < 0.05 indicates a statistically significant causal association between the exposure and the outcome²⁹. The MR-Egger, weighted median, simple mode, and weighted mode were utilized as complementary approaches¹⁴. If the p-value for the supplementary method was P < 0.05, the causal association indicated by the IVW method was further corroborated. Similarly, a causal association between exposure and outcome was supported if the p-value for the supplementary method was P > 0.05, provided the remaining methods yielded results consistent with the IVW (i.e., identical positive or negative OR values)¹⁴.

To evaluate the reliability and stability of the MR causal effect. Heterogeneity among IVs was assessed using Cochran's Q test. A P-value < 0.05 indicated significant heterogeneity among IVs. However, the presence of heterogeneity did not decisively impact the causal effect, thus it was considered acceptable³⁰. Horizontal pleiotropy was evaluated by calculating the intercept of MR-Egger regression. A P-value < 0.05 indicated the presence of pleiotropy in the causal analysis^27,31. When the intercept term of MR-Egger regression approaches zero or P > 0.05, it suggests the absence of horizontal pleiotropy³². Consequently, the interference from other confounders in this causal conclusion can be minimally excluded. The leave-one-out method was employed to assess the influence of individual SNPs on the causal effect²⁷. If the removal of any single SNP did not significantly alter the final result, it indicated that the MR results were not sensitive to any single SNP and were thus reliable and robust.

3.1 Selected IVs

According to the previous selection criteria for validated IVs, all final SNPs included had F-values greater than 10, indicating a strong correlation between AD and CD and the chosen instrumental variables. 4 SNPs were ultimately selected for the MR analysis of the causal risk of AD to CD. During the exploration of the potential causal risk relationship between AD and CD, rs13119723 and rs9296009 were identified as outliers by the MR-PRESSO method and were subsequently excluded. Consequently, 13 SNPs were ultimately selected for the MR analysis of the causal risk of AD to CD.

3.2 The casual effect AD on CD

In our study, we initially conducted Mendelian Randomization (MR) analysis with the AD as the exposure and CD as the outcome. The results of the inverse variance weighting (IVW) method indicated (Table 1, Fig. 2A) that genetically predicted AD was significantly associated with an increased incidence of CD (OR = 1.46, CI = 1.07–1.96, P_IVW = 0.014). Although the p-values for MR-Egger, weighted median, simple mode, and weighted mode were greater than 0.05, the consistent direction of OR for IVW, MR-Egger, and WM supported the reliability of the findings. This suggests that AD may be a potential risk factor for the prevalence of CD. Subsequently, we conducted a series of sensitivity analyses to further validate the reliability and robustness of the findings (Table 2, Fig. 2). Cochran's Q test indicated Q_pval = 0.350 (P > 0.05), suggesting no heterogeneity in the effect of AD on CD. The MR-Egger intercept test showed pval = 0.918 (P > 0.05), indicating no horizontal pleiotropy of AD on CD. The leave-one-out test demonstrated that the causal risk relationship of AD on CD was not significantly influenced by the exclusion of any single SNP (Fig. 3). The various sensitivity analyses confirmed the reliability of the MR study findings.

Table 2

Heterogeneity of Wald ratios and MR-Egger test for directional pleiotropy. df, degree of freedom; MR, Mendelian randomization; Q, heterogeneity statistic Q.
Exposure	Heterogeneity
Exposure	Outcome	Q	df	p-value
Atopic dermatitis	Celiac Disease	2.101	2.00	0.350
Celiac Disease	Atopic dermatitis	11.520	11	0.401
Exposure	MR-Egger test for directional pleiotropy
Exposure	Outcome	Intercept	SE	p-value
Atopic dermatitis	Celiac Disease	0.004	0.038	0.918
Celiac Disease	Atopic dermatitis	-0.002	0.005	0.731

3.3 The casual effect CD on AD

We conducted an inverse Mendelian Randomization (MR) analysis with the CD as the exposure and AD as the outcome. IVW method results indicated (Table 1, Fig. 2) that genetically predicted CD was significantly associated with an increased prevalence of AD. (OR = 1.05, CI = 1.03–1.08, P_IVW = 4.80 × 10^− 6). Although the p-values for MR-Egger, weighted median, simple mode, and weighted mode were all less than 0.05, and the OR was in the same direction (Table 1), which further supports the hypothesis that CD may be a significant potential risk factor for AD incidence. This further supports the result that CD may be a significant potential risk factor for AD incidence. Subsequently, several sensitivity analyses were conducted to further verify the reliability and robustness of the findings (Table 2, Fig. 2). Cochran's Q-test indicated Q_pval = 0.401 (P > 0.05), suggesting no heterogeneity in the effect of CD on AD. The MR-Egger intercept test indicated pval = 0.731 (P > 0.05), suggesting no horizontal pleiotropy of CD on AD. The leave-one-out test demonstrated (Fig. 3) that the causal risk relationship between CD and AD was not significantly influenced by the exclusion of any single SNP. The sensitivity analyses described above showed that the results of the MR study were reliable.

Table1: Mendelian randomization analyses causal relationships between Atopic dermatitis and Celiac disease using genetic variations. CI, confidence interval; IVW, inverse variance-weighted;

Exoposure	Outcome	Snp	Method	OR	95% CI	pval
Atopic dermatitis	Celiac Disease	4	IVW	1.46	1.08–1.97	0.014
		4	MR Egger	1.37	0.50–3.80	0.602
		4	Weighted median	1.28	0.87–1.88	0.215
		4	Simple mode	1.20	0.71–2.06	0.544
		4	Weighted mode	1.20	0.70–2.08	0.556
Celiac Disease	Atopic dermatitis	13	IVW	1.05	1.03–1.07	< 0.001
		13	MR Egger	1.05	1.02–1.08	0.007
		13	Weighted median	1.049	1.02–1.08	< 0.001
		13	Simple mode	1.06	1.02–1.10	0.010
		13	Weighted mode	1.05	1.02–1.08	< 0.003

Utilizing pooled statistics from large-scale GWASs in European populations, we investigated the potential causal relationship between AD and CD through a bidirectional Mendelian randomization study. Our findings indicated that genetically predicted AD was positively associated with CD, and conversely, genetically predicted CD was positively associated with AD. Subsequent sensitivity analyses further confirmed the reliability and robustness of these results. These results have significant implications for the diagnosis and treatment of AD in conjunction with skin-related inflammatory diseases in clinical practice. Moreover, these findings provide a foundation for further investigation into the reciprocal genetic mechanisms underlying AD and CD.

In recent years, several observational studies have been consistent with our findings. A cross-sectional study of 116,816 patients by Shalom et al. ¹²(2020) demonstrated that atopic dermatitis was significantly associated with a higher prevalence of celiac disease across the entire study population and within each subgroup (OR = 1.609, 95% CI = 1.42–1.82, p < 0.001). Lu et al. ¹¹ (2021) conducted a systematic review and meta-analysis, revealing a higher prevalence of celiac disease in patients with AD compared to controls (OR = 1.98, 95% CI = 1.51–2.60), although with significant heterogeneity (I² = 86.7%, p = 0.000). Krishna et al. ³³(2021) conducted a population-based cohort study, which showed an increased prevalence of celiac disease in patients with AD (RR = 1.41,95% CI = 1.32–1.50). Andersen et al. ³⁴(2021) conducted a Danish hospital-based cohort study, demonstrating that AD significantly increased the incidence of CD (OR = 5.19 95% CI, 2.93–9.20). Ciacci et al. ⁹(2021) conducted an observational study, demonstrating that the incidence of atopic dermatitis in patients with celiac disease was approximately three times greater than that of their spouses without AD.

The causal relationship between AD and CD remains inconclusive due to varying study methodologies, different prevalence time points, selection and recall biases, and the possibility of reverse causation. Multiple previous observational studies have yielded conflicting results regarding the reciprocal causal relationship between AD and CD. In 2020, de Lusignan et al. ¹⁰conducted a cohort study on a population of 173,709 individuals, finding that people with AD did not have an increased risk of developing CD. An investigative study by Turjanmaa et al. ¹³in 2023 demonstrated that, compared to healthy populations, individuals with celiac disease did not have an increased risk of developing atopic dermatitis. Furthermore, a British study in 2022 involving 173,709 children, adults, and matched controls revealed that the prevalence of celiac disease was higher in adults diagnosed with atopic dermatitis, but not in children¹⁰. However, a study by Ress et al. ³⁵found that the prevalence of celiac disease was four times higher in 351 children with active atopic dermatitis compared to a randomly selected group of school-aged children, as determined by serologic testing (OR = 4.18, 95% CI = 1.12–15.64).

The biological mechanisms underpinning the association between AD and CD remain unclear. Several published studies provide insights into the potential biological relationship between AD and CD. Polymorphisms in the CTLA4 gene (cytotoxic T-lymphocyte-associated protein 4) have been linked to both AD and CD^36,37. Additionally, IgA deficiency has been correlated with an increased prevalence of autoimmune disorders³⁵, including AD, suggesting a shared genetic background between AD and CD The intestinal mucosa and associated immune system are primary sites for allergen exposure and the induction of immune responses. Local antigenic immune responses induce mucosal damage, leading to inflammation and increased permeability³⁸. This exposure allows lymphoid tissues to encounter a large number of macromolecules, triggering allergic reactions in distant organs. Additionally, the increase and decrease of FOXP3⁺ Tregs are closely associated with AD and CD, respectively³⁵. Most studies on the correlation mechanisms between AD and CD are still at the epigenetic hypothesis stage, and the mechanisms underlying the occurrence and development of these diseases require further investigation.

Compared to traditional observational studies, our bidirectional MR analysis offers several advantages, enhancing the credibility and reliability of the results. To the best of our knowledge, this is the first study to investigate the potential causal relationship between AD and CD using bidirectional MR analysis. This method effectively circumvents errors from reverse causality and confounding factors. Multiple sensitivity analyses further supported the reliability and robustness of our findings. Our study population consisted entirely of individuals of European ancestry, minimizing the effect of population stratification. However, this study has several limitations. We only investigated the overall reciprocal causality between AD and CD, without examining the causal effects across different gender and age subgroups, or the association strength between the severity of AD and CD. Future studies should address these gaps to further expand the applicability of our findings. Additionally, the current MR study's population comprised only individuals of European ancestry, which may limit the generalizability of the findings to other populations due to genetic differences between ethnic groups.

In conclusion, we have identified a bidirectional causal relationship between AD and CD among individuals of European ancestry. Therefore, our findings underscore the importance of early and regular screening for CD in patients with AD, which could facilitate timely diagnosis and prevent long-term complications. Additionally, early interventions for AD in patients with CD should be prioritized to minimize the emotional and financial burden associated with skin symptoms.

Ethics approval and consent to participate

All original study data utilized were sourced from GWAS databases and contained no personal information. All studies included were approved by relevant ethics committee.

Consent for publication:

Not applicable.

Conflicts of interest:

All authors disclose no conflict of interest.

Funding:

This work is supported by the Natural Science Foundation of Gansu Province (20JR10RA002).

Author Contribution

Jian Rong and Peng Chang wrote the main manuscript text; Yao yao Li prepared figures 1-2 and table1;Xin Wang acquired the data and analysis;Mei Song designed of the work and have drafted the work or substantively revised it.

Acknowledgement

All authors disclose no conflict of interest.

Laughter MR, Maymone MBC, Mashayekhi S et al. The global burden of atopic dermatitis: lessons from the Global Burden of Disease Study 1990–2017. Brit j dermatol 2021; 184: 304–9.
Budu-Aggrey A, Kilanowski A, Sobczyk MK et al. European and multi-ancestry genome-wide association meta-analysis of atopic dermatitis highlights importance of systemic immune regulation. Nat Commun 2023; 14: 6172.
Weidinger S, Novak N. Atopic dermatitis. Lancet 2016; 387: 1109–22.
Zhou W, Cai J, Li Z, Lin Y. Association of atopic dermatitis with autoimmune diseases: A bidirectional and multivariable two-sample mendelian randomization study. Front Immunol 2023; 14: 1132719.
Gnodi E, Meneveri R, Barisani D. Celiac disease: From genetics to epigenetics. World j gastroentero 2022; 28: 449–63.
Lebwohl B, Söderling J, Roelstraete B et al. Risk of skin disorders in patients with celiac disease: A population-based cohort study. J am acad dermatol 2021; 85: 1456–64.
Elli L, Bonura A, Garavaglia D et al. Immunological comorbity in coeliac disease: associations, risk factors and clinical implications. J clin immunol 2012; 32: 984–90.
Rodrigo L, Beteta-Gorriti V, Alvarez N et al. Cutaneous and Mucosal Manifestations Associated with Celiac Disease. Nutrients 2018; 10: null.
Ciacci C, Cavallaro R, Iovino P et al. Allergy prevalence in adult celiac disease. Gastroenterology 2003; 124: A657.
de Lusignan S, Alexander H, Broderick C et al. Atopic dermatitis and risk of autoimmune conditions: Population-based cohort study. J allergy clin immun 2022; 150: 709–13.
Lu Z, Zeng N, Cheng Y et al. Atopic dermatitis and risk of autoimmune diseases: a systematic review and meta-analysis. Allergy Asthma Clin Immunol 2021; 17: 96.
Shalom G, Kridin K, Raviv KO et al. Atopic Dermatitis and Celiac Disease: A Cross-Sectional Study of 116,816 Patients. Am j clin dermatol 2020; 21: 133–8.
Turjanmaa E, Hervonen K, Huhtala H et al. Patient-reported burden of skin disorders in coeliac disease. Scand j gastroentero 2023; 58: 1391–7.
Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet epidemiol 2016; 40: 304–14.
Lawlor DA, Harbord RM, Sterne JA et al. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat med 2008; 27: 1133–63.
Weith M, Beyer A. The next step in Mendelian randomization. Elife 2023; 12: null.
Sliz E, Huilaja L, Pasanen A et al. Uniting biobank resources reveals novel genetic pathways modulating susceptibility for atopic dermatitis. J allergy clin immun 2022; 149: 1105-12.e9.
Trynka G, Hunt KA, Bockett NA et al. Dense genotyping identifies and localizes multiple common and rare variant association signals in celiac disease. Nat genet 2011; 43: 1193–201.
Zuber V, Grinberg NF, Gill D et al. Combining evidence from Mendelian randomization and colocalization: Review and comparison of approaches. Am j hum genet 2022; 109: 767–82.
Lee K, Lim CY. Mendelian Randomization Analysis in Observational Epidemiology. J Lipid Atheroscler 2019; 8: 67–77.
Davies NM, Holmes MV, Davey Smith G. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. Bmj 2018; 362: k601.
O'Hagan R, Caldas SA, Correa da Rosa JM et al. The impact of atopic dermatitis on alopecia areata: A 2-sample Mendelian randomization study. J am acad dermatol 2023; 89: 600–2.
Chen Z, Chen Z, Jin X. Mendelian randomization supports causality between overweight status and accelerated aging. Aging cell 2023; 22: e13899.
Papadimitriou N, Dimou N, Tsilidis KK et al. Physical activity and risks of breast and colorectal cancer: a Mendelian randomisation analysis. Nat Commun 2020; 11: 597.
Pierce BL, Burgess S. Efficient design for Mendelian randomization studies: subsample and 2-sample instrumental variable estimators. Am j epidemiol 2013; 178: 1177–84.
Kamat MA, Blackshaw JA, Young R et al. PhenoScanner V2: an expanded tool for searching human genotype-phenotype associations. Bioinformatics 2019; 35: 4851–3.
Verbanck M, Chen CY, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat genet 2018; 50: 693–8.
Zhang L, Zhang C, Zhang J et al. A Bidirectional Mendelian Randomization Study of Sarcopenia-Related Traits and Knee Osteoarthritis. Clin Interv Aging 2023; 18: 1577–86.
Bowden J, Del Greco M F, Minelli C et al. Improving the accuracy of two-sample summary-data Mendelian randomization: moving beyond the NOME assumption. Int j epidemiol 2019; 48: 728–42.
Luo M, Zheng Y, Zhuo Q et al. The causal effects of atopic dermatitis on the risk of skin cancers: A two-sample Mendelian randomization study. J eur acad dermatol 2024; 38: 703–9.
Bowden J, Del Greco M F, Minelli C et al. A framework for the investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat med 2017; 36: 1783–802.
Burgess S, Thompson SG. Interpreting findings from Mendelian randomization using the MR-Egger method. Eur j epidemiol 2017; 32: 377–89.
Krishna MT, Subramanian A, Adderley NJ et al. Allergic diseases and long-term risk of autoimmune disorders: longitudinal cohort study and cluster analysis. Eur respir j 2019; 54: null.
Andersen YM, Egeberg A, Gislason GH et al. Autoimmune diseases in adults with atopic dermatitis. J am acad dermatol 2017; 76: 274 – 80.e1.
Ress K, Annus T, Putnik U et al. Celiac disease in children with atopic dermatitis. Pediatr dermatol 2014; 31: 483–8.
Jones G, Wu S, Jang N et al. Polymorphisms within the CTLA4 gene are associated with infant atopic dermatitis. Brit j dermatol 2006; 154: 467–71.
King AL, Moodie SJ, Fraser JS et al. Coeliac disease: investigation of proposed causal variants in the CTLA4 gene region. Eur j immunogenet 2003; 30: 427–32.
Bischoff SC, Mayer JH, Manns MP. Allergy and the gut. Int arch allergy imm 2000; 121: 270–83.

No competing interests reported.

The genetic interaction between Atopic dermatitis and Celiac disease: A bidirectional two-sample Mendelian randomization study

Status:

Version 1

Abstract

Figures

1 Introduction

2 Material and methods