Different genetic associations of chronic lung diseases and lung cancer risk in European and East Asian populations

To the Editor: Lung cancer (LC) is the leading cause of cancer-related deaths worldwide.^[1] Chronic lung diseases, such as chronic obstructive pulmonary disease (COPD), interstitial lung disease (ILD), pulmonary tuberculosis (PTB), bronchiectasis, and asthma, are non‑malignant disorders of the respiratory system.^[2] The associations between chronic lung diseases and LC risk reported by previous studies remain controversial, possibly due to geographical differences, genetic differences, and inherent biases of observational studies caused by confounding factors, small sample size, reverse causality, etc. Therefore, we conducted a bidirectional two-sample Mendelian randomization (MR) study [Supplementary Table 1, http://links.lww.com/CM9/D4] to analyze the causal relationship between chronic lung diseases and LC, aiming to enhance clinicians’ understanding of this association.^[3] In addition, as current epidemiological data vary across populations, this study separately analyzed the potential causal association between these diseases in European and East Asian populations and explored the possible genetic mechanisms involved in this relationship [Supplementary Figure 1, http://links.lww.com/CM9/D4].

The genome-wide association study summary statistics in the European population included COPD (13,530 cases and 454,945 controls), ILD (2267 cases and 467,560 controls), PTB (895 cases and 476,491 controls), bronchiectasis (2888 cases and 440,263 controls), asthma (9209 cases and 47,428 controls), and LC (3791 cases and 489,012 controls) cohorts. Genetic prediction data from Asian ancestry comprised COPD (4017 cases and 162,653 controls), ILD (1046 cases and 176,974 controls), PTB (7800 cases and 170,871 controls), bronchiectasis (241 cases and 161,803 controls), asthma (13,015 cases and 162,933 controls), and LC (4444 cases and 174,282 controls) cohorts.^[4,5] Detailed information regarding the data sources can be found in Supplementary Table 2, http://links.lww.com/CM9/D4.

Relevant single-nucleotide polymorphisms were selected as instrumental variables (IVs) by conducting quality control steps, which included P <5 × 10⁻⁶, clump distance >10,000 kb, r² <0.001, and F-statistic >10. The IVs associated with confounding factors, such as smoking, alcohol consumption, and body mass index, were removed. Inverse variance weighted (IVW), MR-Egger, and weighted median methods were mainly used for bidirectional two-sample MR. To evaluate the robustness of the results of MR and prevent the potential heterogeneity and pleiotropy, we conducted Cochran’s Q test, MR-Egger regression, leave-one-out, and MR pleiotropy residual sum and outlier (MR-PRESSO). All analyses were conducted using the “TwoSampleMR” and “MR-PRESSO” packages of R Version 4.4.0 (https://cran.r-project.org/mirrors.html).

In forward MR, we investigated the genetic effect of chronic lung diseases on the risk of LC. In the European population, the IVW method indicated that COPD and ILD were associated with a higher risk of LC. The presence of genetic factors from COPD increased the risk of LC by 31.9% (OR = 1.319, 95% CI = 1.183–1.471, P = 6.190 × 10⁻⁷). Compared with the control group, patients with ILD had significantly higher odds of LC (OR = 1.066, 95% CI = 1.007–1.129, P = 0.027). Notably, the COPD–LC association exhibited heterogeneity and pleiotropy, whereas the ILD–LC association did not. In addition, IVW analyses showed that PTB, bronchiectasis, and asthma had no causal effect on LC (PTB [OR = 1.007, 95% CI = 0.924–1.096, P = 0.881]), bronchiectasis (OR = 0.960, 95% CI = 0.865–1.064, P = 0.435), asthma (OR = 1.019, 95% CI = 0.954–1.089, P = 0.571). These results were confirmed by the MR-Egger and weighted median analyses. Sensitivity analysis showed no significant heterogeneity and pleiotropy [Table 1 and Supplementary Figure 2, http://links.lww.com/CM9/D4]. In the East Asian population, COPD and PTB were associated with a high risk of LC via IVW method. Genetic susceptibility to COPD increased the risk of LC by 20.0% (OR = 1.200, 95% CI = 1.038–1.388, P = 0.014), and PTB increased the risk of LC by 38.3% (OR = 1.383, 95% CI = 1.049–1.822, P = 0.021). However, ILD, bronchiectasis, and asthma exhibited no significant effect on the risk of LC (ILD: OR = 1.105, 95% CI = 0.985–1.240, P = 0.090; bronchiectasis: OR = 0.974, 95% CI = 0.943–1.005, P = 0.096; asthma: OR = 1.002, 95% CI = 0.930–1.080, P = 0.958). Sensitivity analyses revealed significant heterogeneity for COPD, ILD, and PTB, but not for bronchiectasis or asthma, whereas no horizontal pleiotropy was detected for any of the five diseases [Table 1 and Supplementary Figure 3, http://links.lww.com/CM9/D4].

Table 1.

The results of forward MR results obtained using IVW, MR Egger, and weighted median methods in European and East Asian populations.

Exposure	Number of SNPs	MR methods	OR (95% CI)	P value	P for heterogeneity	Intercept	P for pleiotropy
European
COPD	41	IVW	1.319 (1.183–1.471)	6.190E–7	0.004	−0.025	0.028
		MR-Egger	1.790 (1.369–2.340)	9.910E–5	0.017
		WM	1.290 (1.129–1.474)	1.806E–4	–
ILD	14	IVW	1.066 (1.007–1.129)	0.027	0.261	−0.003	0.824
		MR-Egger	1.075 (0.981–1.177)	0.146	0.204
		WM	1.083 (1.014–1.157)	0.018	–
PTB	11	IVW	1.007 (0.924–1.096)	0.881	0.135	0.013	0.336
		MR-Egger	0.960 (0.847–1.088)	0.535	0.146
		WM	0.961 (0.873–1.057)	0.412	–
Bronchiectasis	8	IVW	0.960 (0.865–1.064)	0.435	0.628	−0.023	0.410
		MR-Egger	1.068 (0.825–1.383)	0.635	0.612
		WM	0.963 (0.843–1.100)	0.576	–
Asthma	15	IVW	1.019 (0.954–1.089)	0.571	0.481	0.013	0.362
		MR-Egger	0.959 (0.830–1.107)	0.575	0.472
		WM	1.034 (0.941–1.135)	0.488	–
East Asian
COPD	13	IVW	1.200 (1.038–1.388)	0.014	0.002	0.010	0.807
		MR-Egger	1.133 (0.701–1.830)	0.621	0.001
		WM	1.080 (0.927–1.257)	0.323	–
ILD	6	IVW	1.105 (0.985–1.240)	0.090	0.036	−0.004	0.920
		MR-Egger	1.117 (0.882–1.414)	0.410	0.018
		WM	1.122 (1.009–1.248)	0.328	–
PTB	5	IVW	1.383 (1.049–1.822)	0.021	0.034	−0.009	0.894
		MR-Egger	1.475 (0.584–3.729)	0.471	0.016
		WM	1.463 (1.178–1.818)	0.001	–
Bronchiectasis	5	IVW	0.974 (0.943–1.005)	0.096	0.753	−0.015	0.434
		MR-Egger	0.986 (0.945–1.028)	0.552	0.777
		WM	0.979 (0.941–1.019)	0.297	–
Asthma	55	IVW	1.002 (0.930–1.080)	0.958	0.101	0.011	0.290
		MR-Egger	0.911 (0.753–1.102)	0.341	0.106
		WM	1.007 (0.902–1.125)	0.895	–

CI: Confidence interval; COPD: Chronic obstructive pulmonary disease; ILD: Interstitial lung disease; IVW: Inverse variance weighted; LC: Lung cancer; MR: Mendelian randomization; OR: Odds ratio; P for heterogeneity: P value from Cochran’s Q test; P for pleiotropy: P value of the MR-Egger intercept test for horizontal pleiotropy; PTB: Pulmonary tuberculosis; SNP: Single nucleotide polymorphism; WM: Weighted median; –: Not available.

In the reverse MR, we analyzed the genetic association between LC and the risk of chronic lung diseases. The results showed that LC genetic factors reduced the risk of bronchiectasis by 13.8% (OR = 0.862, 95% CI = 0.754–0.985, P = 0.029) in Europeans. In addition, LC did not significantly affect the risk of COPD (OR = 1.007, 95% CI = 0.914–1.109, P = 0.890), ILD (OR = 1.094, 95% CI = 0.927–1.291, P = 0.288), PTB (OR = 1.076, 95% CI = 0.985–1.174, P = 0.103), and asthma (OR = 1.090, 95% CI = 0.950–1.252, P = 0.220). Except for the COPD results, there was no significant heterogeneity and pleiotropy in other results [Supplementary Table 3 and Supplementary Figure 4, http://links.lww.com/CM9/D4]. In East Asians, genetic susceptibility to LC increased the risk of bronchiectasis by 85.0% (OR = 1.850, 95% CI = 1.238–2.764, P = 0.003). However, LC did no significantly affect the risk of COPD (OR = 1.015, 95% CI = 0.878–1.174, P = 0.839), ILD (OR = 0.952, 95% CI = 0.739–1.225, P = 0.700), PTB (OR = 1.004, 95% CI = 0.929–1.084, p = 0.929), and asthma (OR = 0.953, 95% CI = 0.895–1.015, P = 0.137). Except for COPD, none of the above results exhibited significant heterogeneity and pleiotropy [Supplementary Table 3 and Supplementary Figure 5, http://links.lww.com/CM9/D4].

Our study provided genetic evidence showing that the presence of COPD increased the risk of LC in both European and Asian populations, without reverse causation. In European populations, ILD genetically increased the risk of LC, whereas LC decreased the risk of bronchiectasis. In Asian populations, the presence of PTB increased the risk of LC, and LC increased the risk of bronchiectasis. The remaining MR results showed no significant causal association. Finally, we performed multivariable MR including genetic instruments for smoking together with those for each chronic lung disease to estimate disease-specific effects on lung cancer independent of smoking [Supplementary Table 4, http://links.lww.com/CM9/D4]. In addition, the detailed information about final IVs in different populations is listed in Supplementary Tables 5–14, http://links.lww.com/CM9/D4, respectively. Scatter plots, funnel plots, and leave-one-out analyses of forward and reverse MR analyses are shown in Supplementary Figures 6–15, http://links.lww.com/CM9/D4.

This study had several strengths. The study took into account the genetic differences between different ethnic groups and analyzed data from European and East Asian populations respectively. Furthermore, the study possessed the largest sample size and the most SNP loci in genome-wide association study to more comprehensively investigate the relationship between five chronic lung diseases and LC. We minimized the confounding factors leading to bias in observational studies and avoided reverse causality.

However, as limitations, some results were heterogeneous, and the heterogeneity could not be completely eliminated by the available testing methods. Fortunately, the MR-Egger regression intercept and MR-PRESSO test revealed that the results did not have horizontal pleiotropy. MR analysis revealed a reverse causal relationship between LC and bronchiectasis. This might be due to the small sample size of bronchiectasis in East Asia (n = 241 cases), resulting in insufficient statistical power and necessitating careful interpretation. More clinical or basic studies are needed to confirm the results. In addition, the MR results were not completely consistent in Europeans and East Asians; therefore, the possible mechanisms behind this discrepancy need further study.

Our MR study provided genetic evidence for the interrelationship between five chronic lung diseases and LC, and found differences between European and Asian populations. The findings can help clinicians more closely monitor patients with chronic lung diseases who have not yet developed LC. Moreover, these findings can help carefully follow and apply personalized treatments to patients who have already developed LC. This article also helps researchers understand the biological relationship between chronic lung diseases and LC.

Availability of data and materials

All data generated or analyzed during this study are included in this published article (Supplementary materials, http://links.lww.com/CM9/D4).

Acknowledgements

We thank the genome-wide association study consortia who made their summary statistics publicly available for this study, without whom this effort would not be possible, and thank the study staff and participants for their significant contributions.

Funding

This work was supported by The Science and Education Cultivation Fund of the National Cancer and Regional Medical Center of Shanxi Provincial Cancer Hospital (No. TD2023002), and Key Research Program of Department of Science and Technology of Heilongjiang Province (No. 2022ZXJ03C0).

Conflicts of interest

None.

Supplementary Material

cm9-139-2991-s001.docx ^{(2.9MB, docx)}