You are here

Article Text

Article menu
Coeliac disease
Investigating the role of iron status in the development of coeliac disease: a Mendelian randomisation study
  1. Isabel A Hujoel1,
  2. Margaux Louise Anna Hujoel2,3
  1. 1University of Washington, Seattle, Washington, USA
  2. 2Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts, USA
  3. 3Broad Institute of MIT and Harvard, Cambridge, Massachusetts, USA
  1. Correspondence to Dr Isabel A Hujoel; isabelh{at}uw.edu

Abstract

Objective The environmental trigger behind the increasing prevalence of coeliac disease is not known. One suggested cause is iron deficiency, which is common in coeliac disease. We aimed to evaluate this possible association with Mendelian randomisation (MR), which under certain assumptions can suggest a causal relationship.

Design We conducted a two-sample MR study examining the relationship between single nucleotide polymorphisms (SNPs) associated with iron status and the presence of coeliac disease. The SNPs were drawn from a meta-analysis of three genome-wide association studies (GWAS). The association between these SNPs and coeliac disease was assessed using GWAS summary statistics from the UK Biobank. This consists of 336 638 white British individuals, 1855 with coeliac disease. We performed an MR Egger test for pleiotropy and assessed the plausibility of the assumptions of MR to evaluate for possible causality.

Results There were four SNPs strongly associated with systemic iron status. These were not associated with known risk factors for coeliac disease. All four SNPs were available in the UK Biobank coeliac disease summary statistics. Harmonising exposure and outcome associations, we found that higher iron status was negatively associated with risk of coeliac disease (OR per 1 SD increase in serum iron: 0.65, 95% CI 0.47 to 0.91). Leave-one-out analyses had consistent results, and no single SNP drove the association. All three assumptions of MR appeared plausible.

Conclusion We found that genetically lower iron levels were associated with an increased risk of coeliac disease. Our findings highlight a potential opportunity for coeliac disease prevention.

  • celiac disease
  • iron deficiency
  • nutrition

Data availability statement

Data are available in a public, open access repository. Summary statistics for the association of SNPs with iron status: Supplementary Data 2 https://www.nature.com/articles/s42003-020-01575-z; https://www.decode.com/summarydata/. Summary statistics for the association of SNPs with celiac disease: UK-Biobank Single Variant Association Analysis Results; https://www.leelabsg.org/resources. Code available on request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/bmjgast-2023-001236

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • There are several postulated environmental triggers for coeliac disease, however evidence behind these is inconclusive, and at this time the reason coeliac disease is increasing in prevalence is not known.

    WHAT THIS STUDY ADDS

    • Using a study method that can suggest causality, we found that genetically low iron levels are associated with increased coeliac disease risk.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • These findings suggest that iron supplementation in select individuals may provide a potential protective effect against coeliac disease development.

    Introduction

    Coeliac disease is an immune-mediated condition, whose development depends on the presence of genetic factors, most notably the haplotypes DQ2 and DQ8, and environmental triggers, crucially gluten intake. Coeliac disease has a world-wide prevalence ranging from 0.7% to 1.8%.1 The incidence of coeliac disease has been increasing over the past several decades, with one systematic review and meta-analysis estimating a pooled average of a 7.5% increase per year over this time frame.2 While increased recognition and detection of the disease may be contributing to this rise, there also appears to be a true increase in disease prevalence. The reasons behind this increase are not known, but are thought to be secondary to environmental exposure.3 Identifying these exposures is paramount, as it offers the opportunity to potentially mitigate this rise.

    There have been many environmental triggers proposed for coeliac disease, including breastfeeding duration, timing and quantity of gluten exposure, perinatal factors, infections (including rotavirus), socioeconomic factors, geographical location, microbiome composition and medications.4–14 The evidence behind these triggers often is based on observational studies, has conflicting results and is not borne out in subsequent interventional studies aimed at coeliac disease prevention.

    Iron status has been proposed as an environmental trigger for coeliac disease due to an association between maternal iron supplementation during pregnancy and subsequent coeliac disease in the child, as well as increased coeliac disease in those with hemochromatosis.15–17 This has been suggested to be secondary to the impact of iron on the innate immune system and the microbiome.18 19 We aimed to evaluate the relationship between iron status and coeliac disease through the use of Mendelian randomisation (MR). MR is a statistical method that capitalises on the random allocation of single nucleotide polymorphisms (SNPs) at conception. Through this random allocation and the use of genetic variants which are inherently non-modifiable as instruments for modifiable exposures, MR can avoid confounding factors and can suggest causality as long as certain assumptions hold. These assumptions specify that the instrument must be associated with the exposure, that the instrument does not impact the outcome outside of the exposure of interest, and that the instrument does not influence other potential exposures which may have an impact on the outcome of interest. Crucially for this study, by relying on SNPs associated with iron status, and not on iron status itself, MR may avoid the confounding impact of coeliac disease causing iron deficiency.

    Methods

    We conducted a two-sample MR study in which the associations between the genetic instruments (SNPs) and iron status and between the genetic instruments and coeliac disease were measured in different studies. There are three assumptions for the results from this MR study to be valid: (1) the genetic instruments must be associated with systemic iron status (relevance assumption), (2) the genetic instruments must influence coeliac disease only through their effect on systemic iron status (exclusion restriction) and (3) the genetic instruments must not be associated with any confounders, measured or unmeasured (independence assumption).20

    The exposure of interest was systemic iron status, for which no single biomarker exists but which can be assessed using serum iron, ferritin, transferrin and transferrin saturation biomarkers. As previous studies have done, selected genetic instruments were associated with all four of these biomarkers in a manner consistent with an effect on systemic iron status.21 22 A recent meta-analysis of three genome-wide association studies (GWASs; from Iceland, the UK and Denmark) identified four such SNPs: rs1800562 and rs1799945 in the hemochromatosis (HFE) gene, rs855791 in the transmembrane protease serine 6 (TMPRSS6) gene and rs57659670 predicted to affect the Dual Oxidase 2 (DUOX2) gene; the three former instruments were found in a previous iron GWAS whereas the latter variant is novel to this recent meta-analysis.23 24 We used publicly available GWAS summary statistics from the aforementioned meta-analysis (see the Data availability section). All biomarkers were rank-based inverse normal transformed to a standard normal distribution (on a sex-specific basis) and adjusted for age (within the UK cohort, biomarkers were also adjusted for menopausal status, ABO blood group, body mass index, smoking levels, alcohol levels and iron supplementation status). All associations in the meta-analysis had consistent effects directions across all cohorts. Further details on the meta-analysis methodology and heterogeneity between cohort-specific results are available in Bell et al.24

    The outcome of interest was coeliac disease; we used publicly available GWAS summary statistics from the UK Biobank to assess the association between the genetic instruments for systemic iron status and coeliac disease.25 These summary statistics result from an analysis performed on data from white individuals of British ancestry; the model was adjusted for genetic relatedness, sex, birth year, and the first four principal components (see the Data availability section). The UK Biobank enrolled roughly 500 000 individuals (aged 40–69 when recruited) between 2006 and 2010. The summary statistics used were restricted to 336 638 white British individuals, of whom 1855 had coeliac disease.26 The diagnosis of coeliac disease was based on PheCode 557.1 and the summary statistics were computed with SAIGE, which uses a saddlepoint approximation to control for this case–control imbalance, thereby providing accurate p values even when the prevalence of a trait is sufficiently low.27 PheCodes are a phenotyping tool based on International Classification of Diseases codes.

    Both the exposure and outcome studies were conducted in Northern Europe (Iceland, the UK, and Denmark and the UK, respectively). UK samples in the exposure study are from the INTERVAL trial whereas those from the outcome study are from UK Biobank; as both studies are from the UK, sample overlap could exist, however, if it does exist it is likely small and moreover, the INTERVAL study is only one of three studies included in the meta-analysis underlying the exposure data.

    To assess the plausibility of the assumption that the genetic instruments only influence coeliac disease through their effect on systemic iron status (exclusion restriction), we searched an online database of SNP-phenotype associations (PhenoScanner, http://www.phenoscanner.medschl.cam.ac.uk/phenoscanner) for secondary phenotypes associated with the four selected instruments and conducted an MR Egger test for pleiotropy.20 28

    To assess the plausibility of the assumption the genetic instruments are not associated with any confounders (independence assumption) we examined the association between our genetic instruments for iron status and known risk factors for coeliac development available in the UK Biobank, namely sex and age.29

    As noted previously, no single biomarker exists for systemic iron status and as such we used serum iron levels to quantify the genetic associations of the instruments with systemic iron status as was previously done.21 To combine information from multiple independent genetic variants into a single causal estimate, we primarily used an inverse-variance weighted (IVW) method; we used random effects when we were analysing at least four variants. An IVW estimate is a consistent estimator if all genetic variants satisfy the three MR assumptions.30

    We conducted four sensitivity analyses. First, we ran leave-one-out tests in which the overall association was recomputed when each SNP was left out in turn to ensure no single variant drove the association. Second, we obtained an MR estimate through a weighted median estimator (WME), which is a consistent estimator if at least 50% of the weight comes from variants that satisfy the assumptions.30 Third, we obtained an MR estimate through MR-Egger which, under a weaker assumption, provides a valid test of the null causal hypothesis and a consistent estimate even if all variants failed to satisfy the three MR assumptions.31 Fourth, we investigated using the other three biomarkers (other than serum iron levels) to quantify the genetic associations of the instruments with systemic iron status.

    All statistical analyses were conducted using the R package MendelianRandomization and MR Base.

    Patient and public involvement

    Patient and public involvement was not present in this study.

    Results

    There were four independent and strongly associated SNPs for systemic iron status: rs1800562 and rs1799945 in the HFE gene, rs855791 in the TMPRSS6 gene and rs57659670 predicted to affect the DUOX2 gene.24 Associations between these instruments with the four relevant biomarkers are shown in online supplemental table 1. The two variants in HFE are in low linkage disequilibrium (LD) (r2<0.01) when considering European populations in 1000 genomes using the LDlink resource.32 We assessed the strength of the instruments through the F-statistic; all four instruments have high F-statistics (F>10; online supplemental table 2).

    The associations for all the genetic instruments for systemic iron status were available in the UK Biobank summary statistics; we extracted the summary statistics for these SNPs and harmonised exposure and outcome associations (table 1; online supplemental table 3).

    View this table:
    Table 1

    Harmonised exposure and outcome relationships

    Using PhenoScanner, we found that the genetic instruments were not significantly associated with known risk factors for coeliac disease (online supplemental table 4).28 There was no significant evidence of directional pleiotropy as the MR-Egger intercept did not differ from the null (p=0.46; table 2). None of the SNPs were associated with either sex or age (online supplemental table 5).

    View this table:
    Table 2

    Mendelian randomisation estimates of the association of systemic iron status (quantified by serum iron levels) with coeliac disease

    We found that higher iron status was negatively associated with risk of coeliac disease (OR per 1 SD increase in serum iron: 0.65, 95% CI 0.47 to 0.91, p=0.01; table 2, figure 1A). SNP-specific effect associations (ORs) of serum iron with coeliac disease are shown in figure 1A. There was little heterogeneity as assessed with the Cochran Q statistic (Q=4.8, p=0.18).

    Figure 1
    Figure 1

    Forest plot of (A) the genetic instrument-specific and summary estimate for the causal effect of systemic iron status (quantified by serum iron levels) on coeliac disease risk, and (B) the summary estimate for the causal effect of systemic iron status on coeliac disease risk leaving each instrument out one at a time. Dots denote point estimates, horizontal lines denote 95% CIs. MR, Mendelian randomisation.

    The results from leave-one-out analyses generally were consistent and there did not appear to be a single variant driving the association (figure 1B). We further note that when both HFE related SNPs are excluded results appear consistent (see online supplemental note). The MR estimate obtained through the WME approach was consistent with the primary IVW estimate (OR=0.71, 95% CI 0.52 to 0.98, p=0.035; table 2), while the estimate obtained with MR-Egger was non-significant (OR=0.95, 95% CI 0.34 to 2.7, p=0.92; table 2). However, it has been previously noted that MR-Egger is substantially less efficient than IVW or a WME.30 We primarily used serum iron levels to quantify the genetic associations of the instruments with systemic iron status; analyses using the other three biomarkers are shown in online supplemental table 6 however these analyses were limited by low power (see Supplementary Note).

    Discussion

    We found that genetically higher iron levels were associated with a decreased risk of having coeliac disease and that, conversely, genetically lower iron levels were associated with a higher risk of having coeliac disease. If the assumptions of MR hold, this association suggests a causal effect of iron deficiency on subsequent coeliac disease development. While iron deficiency is very unlikely to be the only environmental trigger for coeliac disease, iron deficiency is highly prevalent in coeliac disease, and there are intriguing temporal and demographic associations between the two conditions and potential pathophysiological mechanisms.

    Iron deficiency is the most common nutritional deficiency globally, and is common in both less-developed and developed countries. In the USA, in fact, iron deficiency has been increasing over the same time period that coeliac disease has.2 33 This increase in iron deficiency has been attributed to decreased iron intake secondary to both a decrease in iron concentration in food items, and a shift away from iron-rich beef and towards poultry. Iron deficiency is highest in those sex and age-groups where coeliac disease has highest prevalence, such as children and adolescent and premenopausal women.34 35

    Iron deficiency is highly prevalent in coeliac disease, described in up to roughly 80% of cases, and there has been increasing data to suggest that the relationship between these two conditions is complex.36 Classically, iron deficiency in coeliac disease has been attributed to poor iron absorption secondary to the mucosal atrophy in the duodenum, which is the primary site of iron absorption. However, only 10% of dietary iron needs to be absorbed by the duodenum in order to fulfil the body’s daily iron needs. Therefore, unless someone’s diet is very poor in iron, reduced mucosal surface is likely not to be the only explanation for deficiency. Several studies have also found increased iron deficiency in potential coeliac disease—individuals with elevated coeliac serology but no mucosal damage.37 After 2 years on a gluten-free diet, another study found that nearly one-half of those with iron deficiency at diagnosis remained iron deficient. In children, those with anaemia had more severe duodenal damage and higher serologic markers, however they were not more likely to have diarrhoea and failure to thrive, and despite a year of a gluten-free diet, they continued to have significantly lower haemoglobin levels than non-anaemic controls.38 39 Although gastrointestinal bleeding has been proposed as an explanation for these inconsistencies between duodenal damage and iron deficiency, data in support of this theory is controversial.40 41 This has prompted several studies into potential alterations of iron metabolism regulatory proteins in coeliac disease.42–46

    One regulatory protein, the transferrin receptor 1, may provide a direct pathogenic role of iron deficiency in coeliac disease development. The transferrin receptor is used to import iron into cells, and in the setting of iron deficiency, it is upregulated, and expression on enterocytes is increased. Studies have demonstrated this increased expression in both treated and untreated coeliac disease patients with iron deficiency.43 47 Matysiak-Budnik et al demonstrated that this receptor also acts to transport un-degraded gliadin peptides across enterocytes. Intestinal transport of intact gliadin peptides to the subepithelial region is a key step in the pathophysiology of coeliac disease.48 49 The authors note that epithelial cell proliferation (as may be seen following epithelial damage from intestinal infection) also promotes increased expression of transferrin receptor 1, providing an explanation for a possible pathogenic role for gastroenteritis as well as iron deficiency. Soluble transferrin receptor levels are markers of the expression of transferrin receptor in tissue, and are increased in the setting of iron deficiency. Interestingly, one study found that the levels, and thus the expression of transferrin receptor in tissue, was significantly higher in those with coeliac disease and iron deficiency as compared with those with iron deficiency and no coeliac disease. This was thought to be due to the proliferation of epithelial cells in untreated coeliac disease.50 Iron has also been proposed to contribute to coeliac disease development through its impact on the innate immune system, infection risk and microbiome.15 18 51 52 Hepcidin and lactoferrin, both proteins involved in iron metabolism, are involved in innate immunity, and iron status has been demonstrated to impact the microbiome, which has been implicated as having a role in the pathogenesis of coeliac disease. Recently, iron has been identified as playing a critical part in regulating the immune system in the intestine, and one recent study suggests that because of this, iron deficiency may play a role in the development of inflammatory bowel disease and other immunological disorders.53 Coeliac disease may be one of these disorders.

    The role of iron status in coeliac disease development has been previously implicated by the increased risk of coeliac disease in those with hereditary hemochromatosis, and in children whose mothers took iron supplementation during pregnancy. Hereditary hemochromatosis is a genetic condition that leads to iron overload. The increased risk for coeliac disease extends beyond 5 years after hemochromatosis diagnosis, when theoretically treatment has been started to decrease iron stores.17 Indeed, several studies have suggested that the association between the two conditions may be genetic in origin, and that HFE gene mutations (which are common in coeliac disease) may provide a survival advantage in coeliac disease by reducing iron deficiency (Supplementary Note).54 Notably, in hereditary hemochromatosis, the body does not appear to recognise true iron stores, and soluble transferrin receptor levels are elevated up to a transferrin saturation of 50% (whereas in controls they are elevated only to around 30%).55 This suggests increased transferrin receptor 1 levels even at higher iron levels in hereditary hemochromatosis and may suggest an additional explanation for the increased coeliac disease risk in the condition. The mechanism behind the increased risk for coeliac disease in children whose mothers took iron supplementation in pregnancy is similarly unclear. This increased risk was not seen in mothers with anaemia or who ate iron rich foods, but was seen in children who took iron supplementation before 18 months. Prior studies have found that maternal iron deficiency leads to increased transferrin receptor expression in the placenta but not soluble transferrin receptor levels in the cord blood.56 57 It is unknown if transferrin receptors in the placenta can transport intact gliadin peptides across the membrane, similar to enterocytes, theoretically delivering these peptides to the fetus.

    This study has several strengths. One strength of this MR study is that it is less likely to be affected by confounding compared with observational studies and, under certain assumptions, suggests a causal relationship. Another strength is that rather than using one genetic instrument, the evidence from multiple instruments were combined, thus providing more power. Moreover, by using a recent meta-analysis of iron GWAS the number of instruments for systemic iron status could be increased from 3 to 4.

    This study also has potential limitations. Those with coeliac disease were identified by PheCode, which is derived from international classification of disease codes, and therefore this could have led to misclassification. Additionally, while MR provides some protection against biases such as reverse causation, it is not completely invulnerable.58 However, we note the mechanism linking the genetic instruments to serum iron levels (and iron status) is well-understood for most instruments, thereby mitigating the risk of reverse causation. In more detail, if the mechanisms linking the variants to iron status were unclear, it would be possible these variants could influence risk of coeliac disease and cause a downstream effect on iron status, however most of our instruments affect genes known to influence iron status (HFE and TMPRSS6). Another limitation is that by restricting to instruments that associate with all four iron status markers (with association directions consistent with an effect on systemic iron status), we have the potential to lose power by excluding variants associating with any iron marker.

    This study suggests that iron deficiency may play a role in coeliac disease development. A causal relationship should ideally be confirmed with clinical trials, and if affirmative, should be followed by prevention studies. More broadly, this study raises questions of the potential ties between iron deficiency and gluten-sensitivity, and the potential implications of dietary restrictions and avoidant restrictive food intake disorders which are common in coeliac disease, inflammatory bowel disease, and irritable bowel syndrome.

    Data availability statement

    Data are available in a public, open access repository. Summary statistics for the association of SNPs with iron status: Supplementary Data 2 https://www.nature.com/articles/s42003-020-01575-z; https://www.decode.com/summarydata/. Summary statistics for the association of SNPs with celiac disease: UK-Biobank Single Variant Association Analysis Results; https://www.leelabsg.org/resources. Code available on request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Ethics approval was not obtained as publicly available data was used.

    Acknowledgments

    The findings of this paper were presented at the American College of Gastroenterology 2022 Annual Meeting and were published in abstract form 59.

    References

    1. ↵
      1. Singh P,
      2. Arora A,
      3. Strand TA, et al
      . Global prevalence of celiac disease: systematic review and meta-analysis. Clin Gastroenterol Hepatol 2018;16:823–36. doi:10.1016/j.cgh.2017.06.037
      OpenUrlCrossRefPubMed
    2. ↵
      1. King JA,
      2. Jeong J,
      3. Underwood FE, et al
      . Incidence of celiac disease is increasing over time: A systematic review and meta-analysis. Am J Gastroenterol 2020;115:507–25. doi:10.14309/ajg.0000000000000523
      OpenUrlCrossRefPubMed
    3. ↵
      1. Lebwohl B,
      2. Rubio-Tapia A
      . Epidemiology, presentation, and diagnosis of celiac disease. Gastroenterology 2021;160:63–75. doi:10.1053/j.gastro.2020.06.098
      OpenUrlCrossRef
    4. ↵
      1. Caminero A,
      2. McCarville JL,
      3. Galipeau HJ, et al
      . Duodenal bacterial proteolytic activity determines sensitivity to dietary antigen through protease-activated Receptor-2. Nat Commun 2019;10:1198. doi:10.1038/s41467-019-09037-9
    5. ↵
      1. Norris JM,
      2. Barriga K,
      3. Hoffenberg EJ, et al
      . Risk of celiac disease Autoimmunity and timing of gluten introduction in the diet of infants at increased risk of disease. JAMA 2005;293:2343–51. doi:10.1001/jama.293.19.2343
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Mårild K,
      2. Ye W,
      3. Lebwohl B, et al
      . Antibiotic exposure and the development of Coeliac disease: a nationwide case-control study. BMC Gastroenterol 2013;13:109. doi:10.1186/1471-230X-13-109
    7. ↵
      1. Canova C,
      2. Zabeo V,
      3. Pitter G, et al
      . Association of maternal education, early infections, and antibiotic use with celiac disease: a population-based birth cohort study in northeastern Italy. Am J Epidemiol 2014;180:76–85. doi:10.1093/aje/kwu101
      OpenUrlCrossRefPubMed
    8. ↵
      1. Lebwohl B,
      2. Spechler SJ,
      3. Wang TC, et al
      . Use of proton pump inhibitors and subsequent risk of celiac disease. Dig Liver Dis 2014;46:36–40. doi:10.1016/j.dld.2013.08.128
      OpenUrlCrossRefPubMed
    9. ↵
      1. Stene LC,
      2. Honeyman MC,
      3. Hoffenberg EJ, et al
      . Rotavirus infection frequency and risk of celiac disease Autoimmunity in early childhood: a longitudinal study. Am J Gastroenterol 2006;101:2333–40. doi:10.1111/j.1572-0241.2006.00741.x
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Koletzko S,
      2. Lee H-S,
      3. Beyerlein A, et al
      . Cesarean section on the risk of celiac disease in the offspring: the Teddy study. J Pediatr Gastroenterol Nutr 2018;66:417–24. doi:10.1097/MPG.0000000000001682
      OpenUrlPubMed
    11. ↵
      1. Mårild K,
      2. Stephansson O,
      3. Montgomery S, et al
      . Pregnancy outcome and risk of celiac disease in offspring: a nationwide case-control study. Gastroenterology 2012;142:39–45. doi:10.1053/j.gastro.2011.09.047
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Ludvigsson JF,
      2. Lebwohl B
      . Three papers indicate that amount of gluten play a role for celiac disease - but only a minor role. Acta Paediatr 2020;109:8–10. doi:10.1111/apa.15057
      OpenUrl
    13. ↵
      1. Celdir MG,
      2. Jansson-Knodell CL,
      3. Hujoel IA, et al
      . Latitude and celiac disease prevalence: A meta-analysis and meta-regression. Clin Gastroenterol Hepatol 2022;20:e1231–9. doi:10.1016/j.cgh.2020.09.052
      OpenUrl
    14. ↵
      1. Vriezinga SL,
      2. Auricchio R,
      3. Bravi E, et al
      . Randomized feeding intervention in infants at high risk for celiac disease. N Engl J Med 2014;371:1304–15. doi:10.1056/NEJMoa1404172
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Lebwohl B,
      2. Ludvigsson JF,
      3. Green PHR
      . Editorial: the unfolding story of celiac disease risk factors. Clin Gastroenterol Hepatol 2014;12:632–5. doi:10.1016/j.cgh.2013.10.031
      OpenUrlCrossRefPubMed
    16. ↵
      1. Størdal K,
      2. Haugen M,
      3. Brantsæter AL, et al
      . Association between maternal iron supplementation during pregnancy and risk of celiac disease in children. Clin Gastroenterol Hepatol 2014;12:624–31. doi:10.1016/j.cgh.2013.09.061
      OpenUrl
    17. ↵
      1. Ludvigsson JF,
      2. Murray JA,
      3. Adams PC, et al
      . Does Hemochromatosis Predispose to celiac disease? A study of 29,096 celiac disease patients. Scand J Gastroenterol 2013;48:176–82. doi:10.3109/00365521.2012.749511
      OpenUrl
    18. ↵
      1. Nairz M,
      2. Schroll A,
      3. Sonnweber T, et al
      . The struggle for iron - a metal at the host-pathogen interface. Cell Microbiol 2010;12:1691–702. doi:10.1111/j.1462-5822.2010.01529.x
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Zimmermann MB,
      2. Chassard C,
      3. Rohner F, et al
      . The effects of iron Fortification on the gut Microbiota in African children: a randomized controlled trial in Cote D’Ivoire. Am J Clin Nutr 2010;92:1406–15. doi:10.3945/ajcn.110.004564
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Davies NM,
      2. Holmes MV,
      3. Davey Smith G
      . Reading Mendelian Randomisation Sudies: a guide, glossary, and checklist for Clinicians. BMJ 2018;362:k601. doi:10.1136/bmj.k601
    21. ↵
      1. Gill D,
      2. Benyamin B,
      3. Moore LSP, et al
      . Associations of genetically determined iron status across the Phenome: A Mendelian randomization study. PLoS Med 2019;16:e1002833. doi:10.1371/journal.pmed.1002833
    22. ↵
      1. Gill D,
      2. Del Greco M F,
      3. Walker AP, et al
      . The effect of iron status on risk of coronary artery disease: A Mendelian randomization study-brief report. Arterioscler Thromb Vasc Biol 2017;37:1788–92. doi:10.1161/ATVBAHA.117.309757
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Benyamin B,
      2. Esko T,
      3. Ried JS, et al
      . Novel Loci affecting iron homeostasis and their effects in individuals at risk for Hemochromatosis. Nat Commun 2014;5:4926. doi:10.1038/ncomms5926
    24. ↵
      1. Bell S,
      2. Rigas AS,
      3. Magnusson MK, et al
      . A genome-side meta-analysis yields 46 new Loci associating with biomarkers of iron homeostasis. Commun Biol 2021;4:156. doi:10.1038/s42003-020-01575-z
    25. ↵
      1. Bycroft C,
      2. Freeman C,
      3. Petkova D, et al
      . The UK Biobank resource with deep Phenotyping and Genomic data. Nature 2018;562:203–9. doi:10.1038/s41586-018-0579-z
      OpenUrlCrossRefPubMed
    26. ↵
      1. Lee lab
      . UK-Biobank single variant association analysis results. 2018. Available: https://www.leelabsg.org/resources
    27. ↵
      1. Zhou W,
      2. Nielsen JB,
      3. Fritsche LG, et al
      . Efficiently controlling for case-control imbalance and sample relatedness in large-scale genetic Association studies. Nat Genet 2018;50:1335–41. doi:10.1038/s41588-018-0184-y
      OpenUrlCrossRefPubMed
    28. ↵
      1. Staley JR,
      2. Blackshaw J,
      3. Kamat MA, et al
      . Phenoscanner: a database of human genotype-phenotype associations. Bioinformatics 2016;32:3207–9. doi:10.1093/bioinformatics/btw373
      OpenUrlCrossRefPubMed
    29. ↵
      1. GWAS Results
      . Neale lab. n.d. Available: http://www.nealelab.is/uk-biobank
    30. ↵
      1. Bowden J,
      2. Davey Smith G,
      3. Haycock PC, et al
      . Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median Estimator. Genet Epidemiol 2016;40:304–14. doi:10.1002/gepi.21965
      OpenUrlCrossRefPubMed
    31. ↵
      1. Bowden J,
      2. Davey Smith G,
      3. Burgess S
      . Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol 2015;44:512–25. doi:10.1093/ije/dyv080
      OpenUrlCrossRefPubMed
    32. ↵
      1. Machiela MJ,
      2. Chanock SJ
      . Ldlink: a web-based application for exploring population-specific haplotype structure and linking correlated Alleles of possible functional variants. Bioinformatics 2015;31:3555–7. doi:10.1093/bioinformatics/btv402
      OpenUrlCrossRefPubMed
    33. ↵
      1. Sun H,
      2. Weaver CM
      . Decreased iron intake parallels rising iron deficiency anemia and rleated mortality rates in the US population. J Nutr 2021;151:1947–55. doi:10.1093/jn/nxab064
      OpenUrl
    34. ↵
      1. Martín-Masot R,
      2. Nestares MT,
      3. Diaz-Castro J, et al
      . Multifactorial etiology of anemia in celiac disease and effect of gluten-free diet: A comprehensive review. Nutrients 2019;11:2557. doi:10.3390/nu11112557
    35. ↵
      1. Turawa E,
      2. Awotiwon O,
      3. Dhansay MA, et al
      . n.d. Prevalence of anaemia, iron deficiency, and iron deficiency anaemia in women of reproductive age and children under 5 years of age in South Africa (1997-2021): A systematic review. IJERPH;18:12799. doi:10.3390/ijerph182312799
    36. ↵
      1. Berry N,
      2. Basha J,
      3. Varma N, et al
      . Anemia in celiac disease is Multifactorial in etiology: A prospective study from India. JGH Open 2018;2:196–200. doi:10.1002/jgh3.12073
      OpenUrl
    37. ↵
      1. Repo M,
      2. Lindfors K,
      3. Mäki M, et al
      . Anemia and iron deficiency in children with potential celiac disease. J Pediatr Gastroenterol Nutr 2017;64:56–62. doi:10.1097/MPG.0000000000001234
      OpenUrl
    38. ↵
      1. Rajalahti T,
      2. Repo M,
      3. Kivelä L, et al
      . Anemia in pediatric celiac disease: association with clinical and histological features and response to gluten-free diet. J Pediatr Gastroenterol Nutr 2017;64:e1–6. doi:10.1097/MPG.0000000000001221
      OpenUrl
    39. ↵
      1. Grisolano SW,
      2. Oxentenko AS,
      3. Murray JA, et al
      . The usefulness of routine small bowel biopsies in evaluation of iron deficiency anemia. J Clin Gastroenterol 2004;38:756–60. doi:10.1097/01.mcg.0000139034.38568.51
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      1. Mant MJ,
      2. Bain VG,
      3. Maguire CG, et al
      . Prevalence of occult gastrointestinal bleeding in celiac disease. Clin Gastroenterol Hepatol 2006;4:451–4. doi:10.1016/j.cgh.2005.12.010
      OpenUrlCrossRefPubMedWeb of Science
    41. ↵
      1. Logan RFA,
      2. Howarth GF,
      3. West J, et al
      . How often is a positive Faecal occult blood test the result of Coeliac disease Eur J Gastroenterol Hepatol 2003;15:1097–100. doi:10.1097/00042737-200310000-00006
      OpenUrlCrossRefPubMed
    42. ↵
      1. Repo M,
      2. Hannula M,
      3. Taavela J, et al
      . Iron transporter protein expressions in children with celiac disease. Nutrients 2021;13:776. doi:10.3390/nu13030776
    43. ↵
      1. Barisani D,
      2. Parafioriti A,
      3. Bardella MT, et al
      . Adaptive changes of Duodenal iron transport proteins in celiac disease. Physiol Genomics 2004;17:316–25. doi:10.1152/physiolgenomics.00211.2003
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      1. Sharma N,
      2. Begum J,
      3. Eksteen B, et al
      . Differential Ferritin expression is associated with iron deficiency in coeliac disease. Eur J Gastroenterol Hepatol 2009;21:794–804. doi:10.1097/MEG.0b013e328308676b
      OpenUrlPubMed
    45. ↵
      1. De Falco L,
      2. Tortora R,
      3. Imperatore N, et al
      . The role of Tmprss6 and HFE variants in iron deficiency anemia in celiac disease. Am J Hematol 2018;93:383–93. doi:10.1002/ajh.24991
      OpenUrl
    46. ↵
      1. Tolone C,
      2. Bellini G,
      3. Punzo F, et al
      . The Dmt1Ivs4+44C>A polymorphism and the risk of iron deficiency anemia in children with celiac disease. PLoS One 2017;12:e0185822. doi:10.1371/journal.pone.0185822
    47. ↵
      1. McDevitt J,
      2. OʼFarrelly C,
      3. Weir DG, et al
      . Proliferation-associated markers in coeliac duodenum. Europ J Gastroenterol Hepatol 1994;6:223–8. doi:10.1097/00042737-199403000-00006
      OpenUrl
    48. ↵
      1. Kagnoff MF
      . Overview and pathogenesis of celiac disease. Gastroenterology 2005;128:S10–8. doi:10.1053/j.gastro.2005.02.008
      OpenUrlCrossRefPubMedWeb of Science
    49. ↵
      1. Matysiak-Budnik T,
      2. Moura IC,
      3. Arcos-Fajardo M, et al
      . Secretory IgA mediates Retrotranscytosis of intact Gliadin peptides via the Transferrin receptor in celiac disease. J Exp Med 2008;205:143–54. doi:10.1084/jem.20071204
      OpenUrlAbstract/FREE Full Text
    50. ↵
      1. De Caterina M,
      2. Grimaldi E,
      3. Di Pascale G, et al
      . The soluble Transferrin receptor (sTfR)-Ferritin index is a potential Predictor of celiac disease in children with refractory iron deficiency anemia. Clin Chem Lab Med 2005;43:38–42. doi:10.1515/CCLM.2005.005
      OpenUrlCrossRefPubMed
    51. ↵
      1. Johnson EE,
      2. Wessling-Resnick M
      . Iron metabolism and the innate immune response to infection. Microbes Infect 2012;14:207–16. doi:10.1016/j.micinf.2011.10.001
      OpenUrlCrossRefPubMed
    52. ↵
      1. Rusu IG,
      2. Suharoschi R,
      3. Vodnar DC, et al
      . Iron supplementation influence on the gut Microbiota and Probiotic intake effect in iron deficiency - A literature-based review. Nutrients 2020;12:1993. doi:10.3390/nu12071993
    53. ↵
      1. Xiong L,
      2. Helm EY,
      3. Dean JW, et al
      . Nutrition impact on Ilc3 maintenance and function centers on a cell-intrinsic Cd71-iron axis. Nat Immunol 2023;24:1671–84. doi:10.1038/s41590-023-01612-z
      OpenUrlCrossRef
    54. ↵
      1. Butterworth JR,
      2. Cooper BT,
      3. Rosenberg WMC, et al
      . The role of Hemochromatosis susceptibility gene mutations in protecting against iron deficiency in celiac disease. Gastroenterology 2002;123:444–9. doi:10.1053/gast.2002.34778
      OpenUrlCrossRefPubMedWeb of Science
    55. ↵
      1. Brandão M,
      2. Oliveira JC,
      3. Bravo F, et al
      . The soluble Transferrin receptor as a marker of iron homeostasis in normal subjects and in HFE-related Hemochromatosis. Haematologica 2005;90:31–7.
      OpenUrlAbstract/FREE Full Text
    56. ↵
      1. Sweet DG,
      2. Savage G,
      3. Tubman TR, et al
      . Study of maternal influences on fetal iron status at term using cord blood Transferrin receptors. Arch Dis Child Fetal Neonatal Ed 2001;84:F40–3. doi:10.1136/fn.84.1.f40
      OpenUrlAbstract/FREE Full Text
    57. ↵
      1. Sangkhae V,
      2. Fisher AL,
      3. Wong S, et al
      . Effects of maternal iron status on Placental and fetal iron homeostasis. J Clin Invest 2020;130:625–40. doi:10.1172/JCI127341
      OpenUrl
    58. ↵
      1. Burgess S,
      2. Swanson SA,
      3. Labrecque JA
      . Are Mendelian randomization investigations immune from bias due to reverse causation Eur J Epidemiol 2021;36:253–7. doi:10.1007/s10654-021-00726-8
      OpenUrlPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Twitter @IsabelHujoel

    • Contributors Guarantor: IAH. IAH: conceptualisation; writing- original draft. MLAH: formal analysis; data curation; methodology; writing – review and editing.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.