Causal Association Between Genetically Determined Iron Status and Lung Cancer: A Two-Sample Mendelian Randomization Study



Background: observational studies have demonstrated that the causal relationship between iron status and lung cancer remains unclear causality. A two-sample mendelian randomization (MR) study was performed to identify the causal association between iron status traits and risk of lung cancer.

Methods: Single nucleotide polymorphisms (SNPs) strongly associated (P <5 × 10-8) with four biomarkers of systemic iron status were obtained from a genome-wide association study containing 48,972 subjects from the Genetics of Iron Status Consortium (GISC). Summary-level data for the genetic associations with lung cancer were extracted from International Lung Cancer Consortium (ILCCO) on 11,348 cases and 15,861 controls. We used a two-sample MR study to obtain causal estimates and assessed by sensitivity analyses.

Results: Genetically determined serum iron [OR, 0.86; 95% confidence interval (CI), 0.76-1.00; P=0.025)] had a negative effect on lung cancer, while no significant association with lung cancer was found in transferrin saturation (OR, 0.91; 95% CI, 0.81-1.02; P=0.110), log10 ferritin (OR, 0.76; 95% CI, 0.52-1.12; P=0.170), and transferrin (OR, 1.09; 95% CI, 0.86-1.38; P=0.470).

Conclusion: In conclusion, our results strongly supported a causal link between genetically determined lower serum iron levels and increased risk of lung cancer. These findings may provide a potential clinical prevention and treatment strategy.


Lung cancer is the leading cause of malignancy-related mortality worldwide, accounting for 18% of the total cancer deaths [1]. Although molecular mechanisms underlying initiation and progression of lung cancer remains unclear, it is already known that the major risk factors for lung cancer include cigarette smoking, air pollution, unhealthy lifestyles, occupational lung carcinogens, and genetic susceptibility [2]. The heritability of lung cancer is estimated at approximately 18%, which suggests that genetic factors play an important role in the etiology of lung cancer [3]. With the development of next generation sequencing and microarray technologies, GWAS has been widely used to identify genetic variations associated with complex diseases. Since the GWAS study of lung cancer was first carried out in 2008 [4], over 40 genetic loci have been identified with susceptibility of lung cancer in European and Asian populations [5, 6, 7]. These susceptibility loci identification will help predict high-risk for lung cancer in population, as well as develop new therapeutic targets. Advances in understanding genetic predisposing factors for lung cancer can improve our diagnostic, prognostic, and therapeutic classification systems.

Previous epidemiological studies have indicated that systemic iron dysregulation is closely associated with the incidence and development of lung cancer. As an essential trace element, iron plays a crucial role in a variety of cellular events including heme biosynthesis, mitochondrial respiration, and DNA repair [8]. Many experimental studies also supported a direct or indirect association between iron and lung cancer [9, 10]. Particularly, iron-dependent ferroptosis has recently been investigated in lung cancer, and may be a promising target for anti-cancer therapy. However, there are still some unsolved questions and controversies for iron status in lung cancer according to previous epidemiological data. Several studies showed that high dietary iron intake significantly increased the risk of lung cancer [11, 12, 13]. On the contrary, Muka and fellows reported that high intake of iron was associated with reduced risk of lung cancer in a prospective study [14]. Furthermore, a meta-analysis result revealed that serum iron levels had no significant association with lung cancer risk [15]. The link between iron status and the risk of lung cancer is therefore under much debate.

Mendelian randomization (MR) analysis is a genetic epidemiological method to investigate the causal association between exposures and disease outcomes by using genetic instrument variants (IVs), such as single-nucleotide polymorphisms (SNPs) [16]. Because of the destined germline genotypes, MR has the possibility of overcoming limitations of confounding and reverse causation bias in observational study [16]. Hence, we conducted a two-sample MR analysis to uncover the genetic association between serum iron status (serum iron, log10 ferritin, transferrin saturation, and transferrin) and lung cancer.

Materials And Methods

Study Design

We performed a two-sample mendelian randomization (TSMR) analysis based on the publicly available summary-level statistics of genome-wide association (GWAS) studies. The genetic variants used as IVs in this TSMR study are compliant with three basic assumptions (Fig. 1). First, the genetic variants selected as IVs are strongly associated with exposure (relevance assumption). Second, the selected IVs are not associated with any known confounders (independence assumption). Third, the used IVs can only affect the risk of outcome through the exposure, independent from pleiotropy (exclusion assumption).

Data Sources

In this TSMR study, iron status traits (serum iron, log10 ferritin, transferrin saturation, and transferrin) were used as the exposure. Summary data of iron status were obtained from the Genetics of Iron Status Consortium (GISC), which contained 23,986 subjects of European ancestry from 11 discovery and 24,986 subjects of European ancestry 8 replication cohorts [17]. For the integrated population from 11 cohorts, 55% were women and the average age was 47. Summary statistical data on the association of IVs with lung cancer were extracted from International Lung Cancer Consortium (ILCCO) on 11,348 cases and 15,861 controls [18]. All data used in this study are available in the online public database. Ethical approval and patient consent can be retrieved in the primary GWAS studies.

Genetic Instrumental Variables

The single nucleotide polymorphisms (SNPs) used as IVs in this study were identified with a strong association (p < 5 × 10− 8) and independent inheritance (R2 < 0.001, 10Mb threshold window) without any linkage disequilibrium (LD) based on the 1,000 genomes reference panel. Detailed information for each SNP is presented in Table 1. These SNPs have been used as IVs for iron status (serum iron, log10 ferritin, transferrin saturation, and transferrin) in previous MR studies and were not considered as weak instrumental variables (F statistic > 10).

Table 1

The genetic instrumental variables proxied for serum iron status


Iron, µmol/L


Transferrin Saturation, %


Log10 Ferritin, µg/L


Transferrin, g/L






Beta (SE)



Beta (SE)



Beta (SE)



Beta (SE)







0.328 (0.016)



0.577 (0.016)



0.204 (0.016)



-0.479 (0.016)







0.189 (0.010)



0.231 (0.010)



0.065 (0.010)



-0.114 (0.010)







0.181 (0.007)



0.190 (0.008)



0.055 (0.007)



-0.044 (0.007)


EA/OA, effect allele/ other allele; EAF, effect allele frequency; SE, standard error.

Statistical Analyses

We used inverse variance weighted (IVW) to assess the causal associations between iron status and the risk of lung cancer. The effect measures were the odds ratio (OR) of the risk of lung cancer, which was normalized to one SD increment in each iron status biomarkers. Weighted median, MR Egger, simple mode, and weighted mode methods were used for statistical sensitivity analyses to ensure the robustness of selected IVs. All MR analyses were performed by using the “TwoSampleMR (0.5.6)” package for R version 4.1.3.


Three SNPs (rs1800562, rs1799945 and rs855791) were conservative genetic variables strongly associated with serum iron, transferrin saturation, ferritin, and transferrin, which were selected as genetic IVs proxied for systemic iron status in this TSMR study (Table 1). The SNP-lung cancer associations were obtained from GWAS data of the ILCCO Consortium (Table 2). The pooled MR analyses for the effect of the four iron status markers on the risk of lung cancer were shown in Table 3. The causal estimates demonstrated a protective effect on lung cancer risk for serum iron (OR, 0.86; 95% CI, 0.76-1.00; P = 0.025) by the inverse variance weighted method, whereas no significance was found in transferrin saturation (OR, 0.91; 95% CI, 0.81–1.02; P = 0.110), log10 ferritin (OR, 0.76; 95% CI, 0.52–1.12; P = 0.170), and transferrin (OR, 1.09; 95% CI, 0.86–1.38; P = 0.470). Furthermore, the association with lung cancer across all three SNPs was estimated by sensitivity analyses including the weighted median, single mode-based, weighted mode-based and MR-Egger methods (Fig. 2), and no heterogeneity and pleiotropy were observed indicating the above results were consistent and robust. Cochran’s Q tests also confirmed that no heterogeneity was detected in inverse variance weighted method and MR-Egger regression for serum iron, transferrin saturation, ferritin, and transferrin. Individual MR estimates showed that SNP rs855791 (TMPRSS6) associated with a significant effect of serum iron, transferrin saturation and ferritin on the reduced risk of lung cancer, but exerted an opposite effect of transferrin on lung cancer (Fig. 3). These results indicated that SNP rs855791 could drive a significant effect of serum iron on lung cancer.

Table 2

SNP-lung cancer associations obtained from GWAS data of the ILCCO Consortium


Lung cancer







Beta (SE)








-0.021 (0.038)








-0.018 (0.025)








-0.046 (0.019)


Chr, chromosome; EA/OA, effect allele/ other allele; EAF, effect allele frequency; SE, standard error.

Table 3

Mendelian randomization estimates for the causal effect of iron status on lung cancer risk.


Serum iron

Transferrin Saturation



Inverse variance weighted


OR (95% CI)

0.86 (0.76, 0.98)

0.91 (0.81, 1.02)

0.76 (0.52, 1.12)

1.09 (0.86, 1.38)

Q statistic (P-value)

1.69 (0.43)

2.95 (0.23)

3.46 (0.18)

5.33 (0.07)



OR (95% CI)

1.12 (0.64, 1.96)

1.05 (0.85, 1.30)

1.13 (0.65, 1.96)

0.95 (0.78, 1.15)

Q statistic (P-value)

0.77 (0.38)

0.71 (0.40)

0.75 (0.39)

0.65 (0.42)

Intercept (P-value)

-0.06 (0.51)

-0.05 (0.38)

-0.043 (0.35)

0.040 (0.28)

Weighted median


OR (95% CI)

0.89 (0.77, 1.04)

0.94 (0.84, 1.05)

0.84 (0.60, 1.18)

1.06 (0.91, 1.24)

Simple mode


OR (95% CI)

0.92 (0.75, 1.14)

0.94 (0.80, 1.11)

0.83 (0.47, 1.45)

1.11 (0.86, 1.43)

Weighted mode


OR (95% CI)

0.93 (0.74, 1.16)

0.96 (0.84, 1.08)

0.88 (0.62, 1.25)

1.06 (0.90, 1.24)

OR, odds ratio; CI, confidence interval


In this TSMR study, we estimated the causal link between serum iron status and the risk of lung cancer based on genetic summary data from two previously GWASs. According to the estimates of MR analyses, we revealed that genetic predisposition to lower levels of serum iron was causally associated with a higher risk of lung cancer, while no significant causal effect was observed in transferrin saturation, ferritin, and transferrin. The association was driven by rs855791, which had the largest impact on iron status among the IVs. These findings might provide novel insights into the association of serum iron status with lung tumorigenesis.

To figure out the genetic association between systemic iron status and lung cancer, we used only the three SNPs proxied for all four iron status markers at genome-wide significance. These SNPs have a concordant direction of effect on serum iron, transferrin saturation, ferritin, and transferrin, thus to be valid instruments. Several sensitivity analyses were performed to support the direct causal relationship between serum iron and lung cancer. However, the causal estimates for transferrin saturation, ferritin, and transferrin showed an insignificant trend on the risk of lung cancer. This inconsistence might be attributed to the selected genetic IVs and the different populations derived from GWAS summary data. Future studies based on more robust GWASs are warranted to address the inconsistence.

Iron is an essential trace element required for cellular activities. Both iron overload and iron deficiency are related to significant abnormalities in cellular function [19]. Given that, experimental studies have demonstrated a dual role of iron in tumor growth and metastasis [19]. Epidemiological studies have suggested an association between excess iron and increased cancer incidence and risk [21]. However, there are still controversies for iron status in cancer initiation and progression. Particularly, epidemiological studies on the role of iron status in lung cancer are sparse and inconsistent. Moreover, a recent TSMR study reported the associations of iron status with overall cancer and 22 site-specific cancers, and provide no evidence in support of a causal association between iron status and lung cancer [22]. To be noted, the summary data for lung cancer used in the study contained only 2,838 individuals from UK Biobank, which might have limited potency to obtain meaningful results. Herein, we took advantage of the summary data from ILCCO composed of 11,348 cases and 15,861 controls, and indicated the causal association of serum iron with lung cancer.

However, there are several limitations in our study. First, our results were based on GWAS summary data from European populations, which impeded the generalization of findings to other populations. Second, linear association of iron status biomarkers with lung cancer was obtained by MR analyses, but U-shaped relationship or threshold effect may be more fitting and proper in practice. Third, this study did not explore the causal effect of iron status on different histologic subtypes of lung cancer including adenocarcinoma, squamous carcinoma, and non-small cell lung cancer.

In summary, our work suggested the causal association of reduced serum iron with an increased risk of lung cancer, while the levels of transferrin saturation, ferritin, and transferrin were not significantly associated with lung cancer. Future studies are required to explore the exact mechanism and relationship of iron status with lung cancer, which may provide a clinical value for prevention and treatment strategy.



All data generated or analyzed during this study are included in this article.


We are extremely grateful to the GIS consortium and the ILCCO consortium for providing relevant publicly available public data, which have been mentioned in the data sources section in this study. 


All authors contributed to research design, data analyses and writing of the paper. 


This study was funded by the key research and development planned project of Hunan province (grant number 2022SK2013). 


The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as potential conflicts of interest.


  1. H. Sung, J. Ferlay, R.L. Siegel, et al., Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries, CA Cancer J Clin 71 (2021)209–249.
  2. J. Malhotra, M. Malvezzi, E. Negri, et al., Risk factors for lung cancer worldwide, EUR RESPIR J 48 (2016)889–902.
  3. L.A. Mucci, J.B. Hjelmborg, J.R. Harris, et al., Familial Risk and Heritability of Cancer Among Twins in Nordic Countries, JAMA 315 (2016)68–76.
  4. C.I. Amos, X. Wu, P. Broderick, et al., Genome-wide association scan of tag SNPs identifies a susceptibility locus for lung cancer at 15q25.1, NAT GENET 40 (2008)616–22.
  5. M.N. Timofeeva, R.J. Hung, T. Rafnar, et al., Influence of common genetic variation on lung cancer risk: meta-analysis of 14 900 cases and 29 485 controls, HUM MOL GENET 21 (2012)4980–95.
  6. Z. Hu, C. Wu, Y. Shi, et al., A genome-wide association study identifies two new lung cancer susceptibility loci at 13q12.12 and 22q12.2 in Han Chinese, NAT GENET 43 (2011)792–6.
  7. J.D. McKay, R.J. Hung, Y. Han, et al., Large-scale association analysis identifies new lung cancer susceptibility loci and heterogeneity in genetic susceptibility across histological subtypes, NAT GENET 49 (2017)1126–1132.
  8. S.V. Torti, F.M. Torti, Iron: The cancer connection, MOL ASPECTS MED 75 (2020)100860.
  9. W. Xiong, L. Wang, F. Yu, Regulation of cellular iron metabolism and its implications in lung cancer progression, MED ONCOL 31 (2014)28.
  10. Y. Fan, B. Liu, F. Chen, et al., Hepcidin Upregulation in Lung Cancer: A Potential Therapeutic Target Associated With Immune Infiltration, FRONT IMMUNOL 12 (2021)612144.
  11. A. Fonseca-Nunes, P. Jakszyn, A. Agudo, Iron and cancer risk–a systematic review and meta-analysis of the epidemiological evidence, Cancer Epidemiol Biomarkers Prev 23 (2014)12–31.
  12. G.M. Sukiennicki, W. Marciniak, M. Muszyńska, et al., Iron levels, genes involved in iron metabolism and antioxidative processes and lung cancer incidence, PLOS ONE 14 (2019)e0208610.
  13. H.A. Ward, J. Whitman, D.C. Muller, et al., Haem iron intake and risk of lung cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort, EUR J CLIN NUTR 73 (2019)1122–1132.
  14. T. Muka, B. Kraja, R. Ruiter, et al., Dietary mineral intake and lung cancer risk: the Rotterdam Study, EUR J NUTR 56 (2017)1637–1646.
  15. H.F. Chen, L.X. Wu, X.F. Li, et al., A meta-analysis of association between serum iron levels and lung cancer risk, Cell Mol Biol (Noisy-le-grand) 64 (2018)33–37.
  16. C.A. Emdin, A.V. Khera, S. Kathiresan, Mendelian Randomization, JAMA 318 (2017)1925–1926.
  17. B. Benyamin, T. Esko, J.S. Ried, et al., Novel loci affecting iron homeostasis and their effects in individuals at risk for hemochromatosis, NAT COMMUN 5 (2014)4926.
  18. Y. Wang, J.D. McKay, T. Rafnar, et al., Rare variants of large effect in BRCA2 and CHEK2 affect risk of lung cancer, NAT GENET 46 (2014)736–41.
  19. K. Roemhild, F. von Maltzahn, R. Weiskirchen, et al., Iron metabolism: pathophysiology and pharmacology, TRENDS PHARMACOL SCI 42 (2021)640–656.
  20. S.V. Torti, F.M. Torti, Iron and cancer: more ore to be mined, NAT REV CANCER 13 (2013)342 – 55.
  21. S.V. Torti, D.H. Manz, B.T. Paul, et al., Iron and Cancer, ANNU REV NUTR 38 (2018)97–125.
  22. S. Yuan, P. Carter, M. Vithayathil, et al., Iron Status and Cancer Risk in UK Biobank: A Two-Sample Mendelian Randomization Study, NUTRIENTS 12 (2020).