Causal associations between chronic obstructive pulmonary disease and cardiovascular diseases: a Mendelian randomization study

doi:10.21203/rs.3.rs-3871875/v1

Background

A large number of studies have demonstrated links between chronic obstructive pulmonary disease (COPD) and cardiovascular diseases (CVDs). Due to the presence of numerous shared risk factors, it is currently unknown whether there is an independent or reverse causal relationship between the two.

Methods

Exposure and outcome data from the largest available genome-wide association studies were extracted for Mendelian randomization (MR) studies. Univariate MR analysis was performed using IVW as the primary analysis methods, and multiple sensitivity analyses were used to enhance the robustness of the results. This was further followed by mediation MR analysis of positive results after excluding confounding factors with multivariable MR analysis.

Results

The MR estimation based on IVW method indicated a strong association between genetically determined COPD and HF (OR = 1.12, 95%CI: 1.07–1.17, p = 3.15E-06), CHD (OR = 1.00, 95% CI: 1.00–1.01, p = 4.03E-05), EH (OR = 1.01, 95%CI: 1.01–1.01, p = 3.48E-06) as well as Stroke (OR = 1.00, 95% CI: 1.00–1.00, p = 3.14E-05). The results of multivariable MR analysis revealed that COPD is not significantly associated with CHD after adjusting for IL-6 or LDL or total cholesterol (p > 0.05). Our findings indicated that BMI, smoking initiation, smoking status, obesity and FEV1 played a role in the causal effect of COPD on HF, EH and Stroke.

Conclusions

We found positive causal relationship between COPD and HF, EH, Stroke essentially unaffected by other confounding factors. The causal relationship exhibited between COPD and CHD was influenced by confounding factors. BMI, obesity, initiation of smoking, smoking status and FEV1 were the mediators between COPD and CVDs.

Chronic obstructive pulmonary disease

Cardiovascular diseases

Mendelian randomization

Chronic obstructive pulmonary disease (COPD) and cardiovascular diseases (CVDs) are two clinically frequent diseases with high morbidity and mortality in the elderly. There is no complete cure for COPD, and most people can only rely on medication and oxygen therapy to relieve symptoms and stop the progression of the disease [1]. Moreover, COPD patients have a high prevalence of CVDs, which makes treatment more difficult and the prognosis worse. So, the best solution is to prevent the disease from occurring. Both diseases have many common risk factors such as smoking and age[2], clinically, the two types of diseases often co-occur in a single group of patients. And, although COPD is a lung disease, patients often die from cardiovascular disease or cancer. If patients with COPD take steps to prevent the onset of these common CVDs such as HF, CHD, EH, AF, and Stroke, or if patients with existing CVDs take steps to prevent the onset of COPD disease, the rate of disability and death from the disease can be greatly reduced.

There has been a great deal of investigation showing that COPD has a positive causal relationship with CVDs, and that patients with COPD have a greater likelihood of having a disease associated with CVDs. However, the two share many of the same risk factors, so it cannot be easily proven that COPD is an independent risk factor for CVDs. And there is still disagreement about the causal relationship between COPD and a specific cardiovascular disease and whether cardiovascular disease causes an increase in the prevalence of COPD patients. Mendelian randomization studies are able to avoid confounding factors and reverse causation and are suitable for application to the study of complex causal relationships between these two types of diseases.

Mendelian randomization studies are an approach that uses genetic variation, such as single nucleotide polymorphisms (SNPs), as instrumental variables to study causal relationships between risk factors and outcomes[3]. Because the two alleles on a chromosome are randomly assigned during meiosis, they are not subject to external confounding factors and essentially show whether the risk factor is causally related to the disease. Mendelian randomization studies are weaker than randomized controlled trial (RCT) and stronger than observational studies in terms of level of evidence. But they are easier to conduct than RCT, more economical in terms of financial resources and time [4]. In this study, we first executed a bidirectional two-sample Mendelian randomization study to determine the causal relationship between COPD and CVDs. A multivariate Mendelian randomization(MVMR) study of CVDs causally associated with COPD was then performed to calibrate potential confounding factors. Finally, we conducted a mediation Mendelian randomization study to determine whether these possible risk factors were mediators of COPD versus CVDs.

Study design

The bidirectional two-sample MR analysis was conducted to identify any potential causal association between COPD and 5 CVDs, including HF, CHD, EH, AF and Stroke (Fig. 1). And the methodology employed in this MR analysis was based on three fundamental assumptions. First, the extracted genetic variants should be strongly associated with the exposure factors(F > 10). Second, SNPs cannot be causally linked through relationships with confounders and outcomes. Third, SNPs cannot be directly related to the outcomes (Additional file 1: Fig. S1).

Genetic association datasets

Summary statistical data for COPD with European ancestry, comprising 13,530 cases and 454,945 controls, were extracted from the genome-wide association studies (GWAS) summary datasets [5]. To the best of our knowledge, it is the largest scale and latest GWAS study for COPD.

Our study focused primarily on investigating the causal relationship between COPD and different types of CVDs, including HF, CHD, EH, AF and Stroke. To gather the most extensive and up-to-date information on these outcomes within the European population, we selected the largest available GWAS studies for the primary MR analysis (study information outlined in Table 1). The summary statistics for HF (case/control: 47,309/930,014)[6], CHD (case/control: 10,157/351,037), EH (case/control: 54,358/408,652), AF (case/control:3,537/481,061)[7], and Stroke (case/control: 6,146/ 355,048) were extracted from large scale GWAS studies respectively. Exposure and outcome populations are drawn from as many different databases as possible to prevent false-positive results due to overlapping samples. For the validation analysis, we extracted summary statistics of UK biobank, FinnGen Biobank or other sources different from the primary analysis to check the consistence of the findings across different datasets.

Table 1

Detailed information for the GWAS data
Outcomes	GWAS ID	Data Source	Sample Size	Cases/Controls	PMID	Year	Population
HF	ebi-a-GCST009541	Shah S	977,323	47,309/930,014	31919418	2020	European
CHD	ukb-d-I9_CHD	Neale lab	361,194	10,157/351,037	NA	2018	-
EH	ukb-b-12493	Ben Elsworth	463,010	54,358/408,652	NA	2018	-
AF	ebi-a-GCST90038689	D < U + 00F6 > nerta < U + 015F > HM	484,598	3,537/481,061	33959723	2021	-
Stroke	ukb-d-C_STROKE	Neale lab	361,194	6,146/ 355,048	NA	2018	-
Exposure	GWAS ID	Data Source	Sample Size	Cases/Controls	PMID	Year
Smoking status	ebi-a-GCST90018807	Sakaue S	468,475	13,530/ 454,945	34594039	2021	-

Informed consent statement and ethics approval statement

As the individual studies that contributed to this MR study had already obtained the necessary consent and ethics approvals, no additional consent or ethics approval was specifically required for the present study.

IV selection

To identify the instrumental variables (IVs), we followed a two-step process. Firstly, we extracted single nucleotide polymorphisms (SNPs) robustly associated with the exposures (p < 5e-6). Secondly, we retained only the independent SNPs (kb = 10,000, r² < 0.001) based on the linkage disequilibrium (LD) structure of European populations. To assess the strength of the IVs, we employed the F-statistics of the remaining SNPs by the following formula: F = R² × (N–k–1)/((1–R²) × k), where R² = 2 × β² × (1–EAF) × EAF, N is the sample size of exposure factors, k is the number of SNPs, β is the estimate of genetic effect on exposure factors, and EAF is the frequency of the effect allele. When F > 10, SNPs were taken as having a strong correlation with exposure factors[8]. Whereas the SNPs selected in our study all had F values greater than 20, reinforcing robustness of results.

Statistical analysis

We used the "TwoSampleMR" software package to perform the MR principal analysis to obtain causality estimates as well as to perform various sensitivity analyses[9]. The MR estimates were represented by odds ratios (OR) with 95% confidence intervals (CIs). MR pleiotropy residual sum and outlier (MR-PRESSO) method was used to detect outlier SNPs, and adjust the horizontal pleiotropy by using the “MR-PRESSO” package[10]. The MVMR analysis was performed using the “MVMR” and “Mendelian Randomization” packages[11]. All statistics were calculated using R software 4.3.1 (The R Foundation for Statistical Computing).

Univariable MR analysis

The causal effects were estimated using the inverse variance weighted (IVW) method [12]. Since the IVW method requires all the IVs to be valid in order to obtain an unbiased estimation, we also performed the MR analyses using two alternative MR methods (weighted median and MR-Egger) to assess the robustness of the results[13]. Moreover, the potential horizontal pleiotropy was evaluated by the MR-Egger intercept. Cochran's Q statistic was used to calculate heterogeneity in the MR study. When p < 0.05 indicated the presence of heterogeneity, we used a random effects model to exclude or estimate SNPs. The causal effect was considered significant if the IVW p value was less than the Bonferroni-corrected threshold (p < 0.05/5 = 0.01) while 0.01 ≤ p < 0.05 indicated suggestive significance and the results from the weighted median and MR-Egger were consistent in the same direction.

Multivariable MR analysis

For the significant causal associations in the univariable MR analysis, the MVMR analysis was performed via using the MVMR-IVW method, aiming to adjust for potential confounding factors including BMI, CRP, IL-6, HDL, LDL, T1DM, T2DM, obesity, triglycerides, smoking initiation, Smoking status, Alcohol dependence and Total cholesterol[11].

Mediation MR analysis

To elucidate whether common risk factors for COPD and CVDs mediate a causal relationship between the two, we performed a mediation MR analysis using the two-step MR method[11]. In the first step, we calculated the causal effect of COPD on mediators (β1), and in the second step, we estimated the causal effect of mediators on CVDs (β2). The significance of the mediating effects (β1*β2) and the proportion of the mediation effect in the total effect were estimated using delta method.

Univariable MR analysis

The MR estimation based on IVW method indicated a strong association between genetically determined COPD and HF (OR = 1.12, 95%CI: 1.07–1.17, p = 3.15E-06), CHD (OR = 1.00, 95% CI: 1.00–1.01, p = 4.03E-05), EH (OR = 1.01, 95%CI: 1.01–1.01, p = 3.48E-06) as well as Stroke (OR = 1.00, 95% CI: 1.00–1.00, p = 3.14E-05). However, no significant association was found between COPD and AF (OR = 1.00, 95%CI: 1.00–1.00, p = 0.99) (Fig. 2). Two alternative MR methods: weighted median and MR-Egger show similar results and the scatter plots also show consistent causal trends (Fig. 3A-D). The “Leave-one-out” plot identified that none of the SNPs dominate the estimated causal association between COPD and CVDs (Fig. 3E-H). Both p values of MR-Egger intercept and MR-PRESSO Global test > 0.05 reduce the potential for research results to be influenced by horizontal pleiotropy. The results of Cochran's Q statistic show weak heterogeneity exists in the MR study with outcome of HF so random effects models are used (Additional file 1: Table. S1).

The reverse MR analyses revealed no significant causal effect of genetic predisposition to 2 CVDs on the risk of COPD, including HF (OR = 1.06, 95% CI: 0.96–1.17, p = 0.22), Stroke (OR = 25.65, 95%CI: 0.29–2284.82 p = 0.16). Interestingly, there were strong causal inferences found CHD (OR = 10.227, 95% CI: 1.89–55.36, p = 0.007) and EH (OR = 3.82, 95% CI: 2.29–6.38, p = 2.88E-07). Causal relationship between AF and COPD is suggestive (OR = 36.91, 95% CI: 1.84–740.73, p = 0.02). (Additional file 1: Fig. S2).

To check whether the findings were consistent across different datasets, we replicated the MR analysis using data from different databases. We found similar significant causal association between COPD and 4 CVDs. However, there was no causal relationship shown between COPD and CHD when adjusted for horizontal pleiotropy (Additional file 1: Fig. S3).

MVMR analysis

The MVMR analysis was performed to evaluate the direct effect of COPD on CVD with adjustment of multiple other risk factors associated with CVD. The results obtained from the univariable MR analysis were partly consistent with the findings from the MVMR. However, COPD is not significantly associated with CHD after adjusting for IL-6 or LDL or total cholesterol (p > 0.05) (Table 2).

Table 2

Multivariable MR analysis outcomes
Potential confounding factors	Outcome	nsnp	Pval	OR	95%CI
BMI	HF	3	0.0023	1.077	1.027 ~ 1.129
	CHD	3	6.78E-05	1.004	1.002 ~ 1.006
	Stroke	3	0.0001	1.003	1.001 ~ 1.004
	EH	3	0.0002	1.008	1.004 ~ 1.013
CRP	HF	4	0.0002	1.169	1.076 ~ 1.271
	CHD	5	0.047	1.003	1.000 ~ 1.006
	Stroke	5	6.13E-08	1.004	1.002 ~ 1.005
	EH	5	0.00297	1.008	1.003 ~ 1.013
HDL	HF	3	6.66E-06	1.132	1.073 ~ 1.195
	CHD	4	0.0040	1.004	1.001 ~ 1.006
	Stroke	4	1.84E-07	1.004	1.003 ~ 1.006
	EH	4	4.20E-08	1.015	1.009 ~ 1.020
IL-6	HF	4	7.19E-05	1.171	1.083 ~ 1.266
	CHD	5	0.052	1.003	1.000 ~ 1.006
	Stroke	5	1.36E-09	1.004	1.003 ~ 1.005
	EH	5	0.0027	1.008	1.003 ~ 1.014
LDL	HF	5	1.63E-05	1.166	1.088 ~ 1.251
	CHD	6	0.141	1.003	0.999 ~ 1.006
	Stroke	6	0.0001	1.003	1.002 ~ 1.005
	EH	5	0.00019	1.013	1.006 ~ 1.020
T1DM	HF	6	0.00089	1.096	1.038 ~ 1.157
	CHD	7	0.0015	1.001	1.001 ~ 1.006
	Stroke	7	0.0028	1.002	1.001 ~ 1.004
	EH	6	0.00015	1.011	1.005 ~ 1.016
T2DM	HF	3	3.12E-05	1.125	1.064 ~ 1.189
	CHD	5	1.55E-05	1.005	1.003 ~ 1.007
	Stroke	5	0.0002	1.003	1.001 ~ 1.005
	EH	5	2.86E-06	1.013	1.007 ~ 1.018
obesity	HF	8	0.0095	1.078	1.018 ~ 1.141
	CHD	8	0.0263	1.002	1.000 ~ 1.004
	Stroke	8	0.0128	1.002	1.000 ~ 1.003
	EH	7	0.0253	1.005	1.001 ~ 1.010
triglycerides	HF	2	1.66E-08	1.187	1.118 ~ 1.259
	CHD	2	0.0091	1.004	1.001 ~ 1.007
	Stroke	2	4.35E-06	1.004	1.002 ~ 1.006
	EH	2	6.76E-06	1.015	1.008 ~ 1.022
Alcohol dependence	HF	10	1.35E-05	1.104	1.056 ~ 1.154
	CHD	12	0.0013	1.003	1.001 ~ 1.005
	Stroke	12	0.0010	1.002	1.001 ~ 1.003
	EH	11	8.17E-06	1.008	1.004 ~ 1.011
Smoking initiation	HF	9	0.0003	1.088	1.039 ~ 1.140
	CHD	10	0.0006	1.003	1.001 ~ 1.005
	Stroke	10	0.0009	1.002	1.001 ~ 1.004
	EH	9	3.85E-06	1.010	1.006 ~ 1.015
Smoking status	HF	9	0.0001	1.095	1.045 ~ 1.148
	CHD	8	3.49E-05	1.005	1.002 ~ 1.007
	Stroke	10	0.0055	1.002	1.001 ~ 1.003
	EH	9	6.57E-05	1.009	1.005 ~ 1.014
Total cholesterol	HF	2	6.81E-05	1.163	1.080 ~ 1.253
	CHD	4	0.3140	1.002	0.998 ~ 1.006
	Stroke	4	4.88E-05	1.004	1.002 ~ 1.005
	EH	4	0.0014	1.012	1.005 ~ 1.020

Mediation MR analysis

The two-step MR was employed to perform mediation MR analysis. We aimed to investigate whether the causal relationship between COPD and CVDs could be mediated by BMI, CRP, HDL, LDL, T1DM, T2DM, obesity, triglycerides, alcohol dependence, total cholesterol, acute upper respiratory infections, acute lower respiratory infection and FEV1. Interestingly, our findings indicated that BMI, smoking initiation, smoking status and FEV1 played a role in the causal effect of COPD on HF, EH and Stroke. The proportions of mediation for BMI were 25.43% (95%CI: 10.31 ~ 40.54%), 33.92% (95%CI:11.94 ~ 55.89%) and 12.06% (95%CI:4.48 ~ 19.65%), respectively. The proportions of mediation for smoking initiation were 14.17% (95%CI: 4.16 ~ 24.18%), 11.70% (95%CI:1.57 ~ 21.83%) and 13.16% (95%CI:3.79 ~ 22.53%), respectively. The proportions of mediation for smoking status were 12.54% (95%CI:4.34 ~ 20.74%), 20.40% (95%CI: 7.04 ~ 33.76%) and 10.71% (95%CI: 3.12 ~ 18.30%), respectively. The proportions of mediation for FEV1 were 7.48% (95%CI:1.20 ~ 13.76%), 25.01% (95%CI:13.84 ~ 36.18%) and 8.68% (95%CI: 1.15 ~ 16.21%), respectively. And obesity played a role in the causal effect of COPD on HF and EH. The proportions of mediation were 30.98% (95%CI: 9.78 ~ 52.18%) and 35.59% (95%CI: 9.64 ~ 61.55%), respectively. (Table 3)

Table 3

Mediation MR analysis outcomes
Mediators	Outcomes	Mediation effect in total effect	95%CI
BMI	HF	25.43%	10.31 ~ 40.54%
	Stroke	12.06%	4.48 ~ 19.65%
	EH	33.92%	11.94 ~ 55.89%
CRP	HF	0.31%	-18.98 ~ 19.61%
	Stroke	-5.25%	-41.54 ~ 31.04%
	EH	3.31%	-22.35 ~ 28.97%
HDL	HF	2.21%	-2.50 ~ 6.93%
	Stroke	1.05%	-1.52 ~ 3.62%
	EH	3.60%	-7.11 ~ 14.31%
LDL	HF	-4.27%	-14.59 ~ 6.05%
	Stroke	-2.39%	-9.06 ~ 4.28%
	EH	-0.71%	-5.54 ~ 4.13%
T1DM	HF	0.23%	-6.33 ~ 6.79%
	Stroke	6.62%	-3.82 ~ 17.07%
	EH	10.40%	-5.03 ~ 25.82%
T2DM	HF	3.70%	-1.18 ~ 8.58%
	Stroke	1.59%	-1.12 ~ 4.31%
	EH	6.19%	-3.21 ~ 15.59%
Obesity	HF	30.98%	9.78 ~ 52.18%
	Stroke	-3.01%	-15.32 ~ 9.30%
	EH	35.59%	9.64 ~ 61.55%
Triglycerides	HF	0.02%	-4.28 ~ 4.24%
	Stroke	-0.06%	-1.46 ~ 1.33%
	EH	1.28%	-10.58 ~ 8.03%
Alcohol dependence	HF	2.35%	-3.24 ~ 7.94%
	Stroke	2.15%	-4.37 ~ 8.68%
	EH	1.74%	-3.47 ~ 6.96%
Smoking initiation	HF	14.17%	4.16 ~ 24.18%
	Stroke	13.16%	3.79 ~ 22.53%
	EH	11.70%	1.57 ~ 21.83%
Smoking status	HF	12.54%	4.34 ~ 20.74%
	Stroke	10.71%	3.12 ~ 18.30%
	EH	20.40%	7.04 ~ 33.76%
Total cholesterol	HF	-3.80%	-11.07 ~ 3.48%
	Stroke	-1.33%	-5.11 ~ 2.45%
	EH	0.15%	-3.57 ~ 3.88%
Acute upper respiratory infections	HF	9.09%	-1.26 ~ 19.45%
	Stroke	-3.19%	-11.80 ~ 5.43%
	EH	6.07%	-4.46 ~ 16.60%
Acute lower respiratory infection	HF	-4.20%	-12.63 ~ 4.24%
	Stroke	-0.41%	-13.46 ~ 12.64%
	EH	7.53%	-4.46 ~ 19.53%
FEV1	HF	7.48%	1.20 ~ 13.76%
	Stroke	8.68%	1.15 ~ 16.21%
	EH	25.01%	13.84 ~ 36.18%

We examined the causal relationship between COPD and CVDs using GWAS summary datasets and concluded that: (1) There were positive associations between COPD and HF, CHD, EH, and Stroke, but the causality of COPD with CHD was influenced by confounding factors such as IL-6, LDL, and total cholesterol levels. COPD did not show causality with AF. (2) HF and Stroke were not causally associated with COPD, CHD and EH were significantly causally associated with COPD, and AF had a suggestive causal association with COPD. (3) BMI, smoking initiation, smoking status, and FEV1 were identified as mediators of COPD associated with HF, EH, and stroke as well. Additionally, obesity was found to be a mediator of COPD associated with HF and EH.

COPD is a pulmonary syndrome characterized by irreversible airflow limitation due to obstruction of small airways and decreased alveolar elasticity. The main mechanism of interaction between COPD and CVDs is the increase of apoptosis, anti-endothelial cell antibodies, endothelial cell dysfunction, and systemic inflammatory response and oxidative stress. Patients with COPD are prone to tissue hypoxia, which triggers their own inflammatory response and oxidative stress, increased LDL oxidation, increased macrophage phagocytosis, and increased transformed foam cells. LDL oxidation is increased, enhancing phagocytosis by macrophages and increasing the number of transformed foam cells[14]. Upregulated expression of cell adhesion molecules, intercellular adhesion molecules, and selectins in endothelial cells promotes deposition of lipid plaques and thrombus [15]. Relevant studies have shown that inflammation is an important factor mediating COPD complicating CVDs [16]. Inflammatory reactions in the lungs can develop into systemic inflammation through the blood circulation, and damage to blood vessels from inflammation plays an important role in this. Damage to the arterial endothelium by inflammatory factors induces atherosclerosis and thrombosis, leading to CVDs, and Mills et al. demonstrated that the high incidence of CVDs in patients with COPD is associated with atherosclerosis, which in turn is associated with vascular dysfunction due to inflammation[17]. Reactive oxygen species released by inflammatory cells oxidize lipids attached to the sidewalls of blood vessels, and increased levels of oxidized lipoproteins promote the progression of atherosclerosis [18]. Dysregulation of proteases and antiproteases is an important mechanism in the development of COPD, of which the main ones associated with CVDs are matrix metalloproteinases (MMPs), such as MMP-2, MMP-9, and MMP-12. at the onset of the disease, they are present in large quantities in atherosclerotic plaques, in the circulation, in lung tissues, and in patient sputum [19–22]. MMP, oxidative stress and inflammatory cell production interact to promote plaque progression.

The results of the present study are in agreement with most of the previous ones. In a cohort study from a Danish hospital, it was shown that the percentage of HF occurrence was significantly increased in patients with COPD compared to those without COPD (13.3% vs. 4.0%)[23]. And, HF had the largest OR of the three positive results, implying the highest propensity to develop the disease in common CVDs. Curkendall et al discovered that, after adjusting for cardiovascular risk factors, heart failure (OR = 3.84, CI: 3.56–4.14) had a higher likelihood of occurring compared to angina (OR = 1.61, CI:1.47–1.76), acute myocardial infarction (OR = 1.61, CI:1.43–1.81), and stroke (OR = 1.11, CI: 1.02–1.21), with a propensity 2–3 times greater[24]. This outcome was expected due to the pulmonary hypertension and right heart failure of pulmonary origin resulting from the late progression of COPD. In several observational studies, COPD has shown a clear causal relationship with CHD. COPD did not show a causal relationship with CHD after adjusting for confounders IL-6, LDL, and total cholesterol levels, respectively. Similarly, no causal relationship between COPD and coronary artery disease was shown in an MR study of COPD and cardiac traits (causal estimate = 0.004, p = 0.40)[25]. In another more comprehensive bidirectional MR study of COPD and cardiac traits, Cross-trait LD score regression showed that FEV1, FEV, and FEV1/FVC correlated with coronary artery disease, but there was no causal relationship in either MR study [26]. Considering that the MR study examined the relationship at the genetic level, the influence of confounding factors was excluded. Therefore, it is hypothesized that the previous observational study may have been the result of a strong association between the two due to the presence of too many common risk factors. Statin of traditional drugs that improve the prognosis of cardiovascular disease have anti-inflammatory and antioxidant, lipid-lowering, and immunomodulatory effects to improve endothelial cell function. A meta-analysis shows statin improve exercise tolerance, quality of life, and lung function in patients with COPD, and greater benefits are seen in patients with COPD combined with CVDs, systemic inflammation, and hyperlipidemia [27]. However, in another RCT, simvastatin had no benefit on acute exacerbation rates or time to first acute exacerbation in COPD patients who did not have heart disease or diabetes [28]. Therefore, in conjunction with our study, we hypothesize that the benefits of statin in COPD patients may be primarily derived from lipid-lowering, antioxidant therapy for heart disease. Early clinical use of statin in patients with COPD may be able to delay the atherosclerotic process, reduce the risk of cardiovascular disease, and prevent the onset of CHD. Multivariate MR analysis of a previous study showed that very severe COPD was associated with a higher risk of developing hypertension (OR = 1.6, 95% CI: 1.4–1.9)[29]. But the specific mechanisms involved need further study. The present study demonstrates a causal relationship between COPD and Stroke even after adjusting for multiple confounders, similar to the results of most previous studies [30, 31]. A recent cross-sectional study by Chen et al showed a higher incidence of stroke in patients with established COPD compared to patients without COPD (14.1% vs. 3.9%), but there was no causal relationship between COPD and stroke after adjusting for confounders[32]. So there is still a need for further research on the independent causal relationship between the two.

Studies related to reverse causality and mechanisms between COPD and CVDs are inadequate compared to studies of forward causality. Reverse MR analysis in the current investigation showed that EH was able to promote the development of COPD, which may be due in part to the use of anti-hypertensive medications, such as beta-blockers [33]. The use of medications for COPD and CVDs has always been conflicting and a point of concern for clinicians, and further research is needed in this area in the future. Several studies have shown a relationship between COPD and AF, but the direction is not conclusive, and there are still studies validating the increased incidence of COPD in patients with AF [34–41]. The results of this study show reverse causation of COPD with AF and CHD, but both of them are influenced by common cardiopulmonary risk factors as well as cardiopulmonary diseases, and the exact mechanisms need to be further explored.

Mediating factors were explored in that the presence of these mediating factors in patients with COPD increased the incidence and exacerbation of CVDs in the presence of pre-existing COPD or these risk factors. FEV1 is an independent risk factor for CVDs, and the results of the mediator MR study support the findings of previous studies [42]. Similar findings were obtained in a cross-sectional study, where the prevalence of hypertension and CVDs increased progressively with increasing severity of COPD (decrease in FEV1) [29]. Obesity or increased BMI may bring about a variety of metabolic disorders in the body, such as high blood glucose, high blood lipids, and increased cardiac burden due to water and sodium retention. Nicotine in cigarettes damages the vascular endothelial and causes an increase in the oxygen free radicals within organism. Weight loss and smoking cessation have been the most important preventive and therapeutic tools in patients with COPD combined with CVDs.

In order to obtain robust results, we used multiple methods to exclude the effects of horizontal pleiotropy and confounding factors. We used multiple datasets to validate our positive results. And, we derived the mediator and mediator proportions of positive outcomes through a mediation Mendelian randomization study. However, some deficiencies exist in our study. First of all, our research population is Europeans so findings of the survey cannot apply their findings to other populations. Second, study bias is unavoidable. The relationship between COPD and CVDs is complex, and despite our best efforts to minimize pleiotropy, it is possible that unknown intrinsic and extrinsic associations between other exposures and outcomes may lead to pleiotropy. It is not practical to study all possible confounders and mediators in MVMR and mediator MR in a completely exhaustive manner. Finally, there was no classification of COPD into different types (chronic bronchitis and emphysema), acute and calm phases, and different levels of severity. Currently, no study has been able to clarify in which type of patients with impaired lung function CVDs can be more prevalent, and it is also difficult to explore the specific relationship between COPD subtypes and individual CVDs [43, 44], as the MR study is an emerging research methodology and the available research data are limited, more detailed MR studies need to be done in the future.

We found a positive causal relationship between COPD and HF, EH, Stroke essentially unaffected by other confounders. The causal relationship exhibited between COPD and CHD was affected by confounders IL-6, LDL, total cholesterol levels. There was an inverse causal relationship between COPD and CHD, EH.BMI, obesity, initiation of smoking, smoking status and FEV1 were the mediators between COPD and CVDs. Intervention of intermediate factors as well as the treatment of COPD will reduce the incidence and exacerbation of CVDs and improve the prognosis of patients.

COPD	Chronic obstructive pulmonary disease
CVDs	Cardiovascular diseases
HF	Heart failure
CHD	Coronary heart disease
EH	Essential hypertension
AF	Atrial fibrillation
SNPs	Single nucleotide polymorphisms
RCT	Randomized controlled trial
GWAS	Genome-wide association studies
IVs	instrumental variables
LD	linkage disequilibrium
OR	odds ratios
CIs	confidence intervals
MR-PRESSO	MR pleiotropy residual sum and outlier
MVMR	Multivariable MR
IVW	Inverse variance weighted
BMI	Body mass index
CRP	C-reactive protein
HDL	High density lipoprotein
IL-6	Interleukin-6
LDL	Low density lipoprotein
T1DM	Type 1 diabetes mellitus
T2DM	Type 2 diabetes mellitus
FEV1	Forced expiratory volume in the first second
MMP	Matrix metalloproteinase

Acknowledgements

We would like to thank the researchers and study participants for their contributions.

Funding

This work was supported by The National Natural Science Foundation of China (81970300).

Data Availability

All of the datasets used in this study came from publicly accessible resources found at (https://gwas.mrcieu.ac.uk/).

Author contributions

Conception and design: Guangzan Yu, Lulu Liu; Acquisition, statistical analysis or interpretation of the data: Guangzan Yu, Qian Ma; Drafting of the manuscript: Guangzan Yu, Fusheng Han, Hua He . All authors have readand agreed to the published version of the manuscript.

Ethics approval and consent to participate

The data used in this paper are publicly available, ethically approved.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.

Lareau SC, Fahy B, Meek P, Wang A. Chronic Obstructive Pulmonary Disease (COPD). Am J Respir Crit Care Med. 2019;199:P1–P2.
Dacha S, Chaiwong W, Tajarernmuang P. Association of cardiovascular disease with COPD: cardiac function and structure evaluation. Jornal Brasileiro de Pneumologia: Publicacao Oficial Da Sociedade Brasileira de Pneumologia E Tisilogia. 2022;48:e20220388.
Levin MG, Burgess S. Mendelian Randomization as a Tool for Cardiovascular Research: A Review. JAMA Cardiol 2023;:.
Emdin CA, Khera AV, Kathiresan S, Mendelian Randomization. JAMA. 2017;318:1925–6.
Sakaue S, Kanai M, Tanigawa Y, Karjalainen J, Kurki M, Koshiba S, et al. A cross-population atlas of genetic associations for 220 human phenotypes. Nat Genet. 2021;53:1415–24.
Shah S, Henry A, Roselli C, Lin H, Sveinbjörnsson G, Fatemifar G, et al. Genome-wide association and Mendelian randomisation analysis provide insights into the pathogenesis of heart failure. Nat Commun. 2020;11:163.
Dönertaş HM, Fabian DK, Valenzuela MF, Partridge L, Thornton JM. Common genetic associations between age-related diseases. Nat Aging. 2021;1:400–12.
Pierce BL, Ahsan H, Vanderweele TJ. Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants. Int J Epidemiol. 2011;40:740–52.
Hemani G, Zheng J, Elsworth B, Wade KH, Haberland V, Baird D et al. The MR-Base platform supports systematic causal inference across the human phenome. Elife 2018; 7.
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.
Sanderson E. Multivariable Mendelian Randomization and Mediation. Cold Spring. Harb Perspect Med 2021; 11.
Burgess S, Butterworth A, Thompson SG. Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol. 2013;37:658–65.
Davies NM, Holmes MV, Davey Smith G. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ (Clinical Research ed.). 2018; 362: k601.
Lattimore JD, Wilcox I, Nakhla S, Langenfeld M, Jessup W, Celermajer DS. Repetitive hypoxia increases lipid loading in human macrophages-a potentially atherogenic effect. Atherosclerosis. 2005;179:255–9.
Ichikawa H, Flores S, Kvietys PR, Wolf RE, Yoshikawa T, Granger DN, et al. Molecular mechanisms of anoxia/reoxygenation-induced neutrophil adherence to cultured endothelial cells. Circ Res. 1997;81:922–31.
Gan WQ, Man SF, Senthilselvan A, Sin DD. Association between chronic obstructive pulmonary disease and systemic inflammation: a systematic review and a meta-analysis. Thorax. 2004;59:574–80.
Mills NL, Miller JJ, Anand A, Robinson SD, Frazer GA, Anderson D, et al. Increased arterial stiffness in patients with chronic obstructive pulmonary disease: a mechanism for increased cardiovascular risk. Thorax. 2008;63:306–11.
Kirkham P, Rahman I. Oxidative stress in asthma and COPD: antioxidants as a therapeutic strategy. Pharmacol Ther. 2006;111:476–94.
Baraldo S, Bazzan E, Zanin ME, Turato G, Garbisa S, Maestrelli P, et al. Matrix metalloproteinase-2 protein in lung periphery is related to COPD progression. Chest. 2007;132:1733–40.
Bolton CE, Stone MD, Edwards PH, Duckers JM, Evans WD, Shale DJ. Circulating matrix metalloproteinase-9 and osteoporosis in patients with chronic obstructive pulmonary disease. Chron Respir Dis. 2009;6:81–7.
Yasmin, McEniery CM, Wallace S, Dakham Z, Pulsalkar P, Maki-Petaja K, et al. Matrix metalloproteinase-9 (MMP-9), MMP-2, and serum elastase activity are associated with systolic hypertension and arterial stiffness. Arterioscler Thromb Vasc Biol. 2005;25:372.
Maclay JD, MacNee W. Cardiovascular disease in COPD: mechanisms. Chest. 2013;143:798–807.
Skajaa N, Laugesen K, Horváth-Puhó E, Sørensen HT. Comorbidities and mortality among patients with chronic obstructive pulmonary disease. BMJ Open Respir Res 2023; 10.
Curkendall SM, DeLuise C, Jones JK, Lanes S, Stang MR, Goehring E, et al. Cardiovascular disease in patients with chronic obstructive pulmonary disease, Saskatchewan Canada cardiovascular disease in COPD patients. Ann Epidemiol. 2006;16:63–70.
Zhu Z, Wang X, Li X, Lin Y, Shen S, Liu CL, et al. Genetic overlap of chronic obstructive pulmonary disease and cardiovascular disease-related traits: a large-scale genome-wide cross-trait analysis. Respir Res. 2019;20:64.
Wielscher M, Amaral AFS, van der Plaat D, Wain LV, Sebert S, Mosen-Ansorena D, et al. Genetic correlation and causal relationships between cardio-metabolic traits and lung function impairment. Genome Med. 2021;13:104.
Zhang W, Zhang Y, Li CW, Jones P, Wang C, Fan Y. Effect of Statins on COPD: A Meta-Analysis of Randomized Controlled Trials. Chest. 2017;152:1159–68.
Criner GJ, Connett JE, Aaron SD, Albert RK, Bailey WC, Casaburi R, et al. Simvastatin for the prevention of exacerbations in moderate-to-severe COPD. N Engl J Med. 2014;370:2201–10.
Mahishale V, Angadi N, Metgudmath V, Eti A, Lolly M, Khan S. Prevalence and impact of diabetes, hypertension, and cardiovascular diseases in chronic obstructive pulmonary diseases: A hospital-based cross-section study. J Transl Int Med. 2015;3:155–60.
Ding C, Wang R, Gong X, Yuan Y. Stroke risk of COPD patients and death risk of COPD patients following a stroke: A systematic review and meta-analysis. Med (Baltim). 2023;102:e35502.
Morgan AD, Rothnie KJ, Bhaskaran K, Smeeth L, Quint JK. Chronic obstructive pulmonary disease and the risk of 12 cardiovascular diseases: a population-based study using UK primary care data. Thorax. 2018;73:877–9.
Chen H, Luo X, Du Y, He C, Lu Y, Shi Z, et al. Association between chronic obstructive pulmonary disease and cardiovascular disease in adults aged 40 years and above: data from NHANES 2013–2018. BMC Pulm Med. 2023;23:318.
Schnabel E, Karrasch S, Schulz H, Gläser S, Meisinger C, Heier M, et al. High blood pressure, antihypertensive medication and lung function in a general adult population. Respir Res. 2011;12:50.
Cerit L, Duygu H, Gulsen K, Kemal HS, Ozcem B, Balcioglu O, et al. Is SYNTAX Score Predictive of Atrial Fibrillation after On-Pump Coronary Artery Bypass Graft Surgery? Korean Circ J. 2016;46:798–803.
Chahal H, Heckbert SR, Barr RG, Bluemke DA, Jain A, Habibi M, et al. Ability of Reduced Lung Function to Predict Development of Atrial Fibrillation in Persons Aged 45 to 84 Years (from the Multi-Ethnic Study of Atherosclerosis-Lung Study). Am J Cardiol. 2015;115:1700–4.
Konecny T, Park JY, Somers KR, Konecny D, Orban M, Soucek F, et al. Relation of chronic obstructive pulmonary disease to atrial and ventricular arrhythmias. Am J Cardiol. 2014;114:272–7.
Johnson LS, Juhlin T, Engström G, Nilsson PM. Reduced forced expiratory volume is associated with increased incidence of atrial fibrillation: the Malmo Preventive Project. Europace. 2014;16:182–8.
Geçmen Ç, Babür Güler G, Erdoğan E, Hatipoğlu S, Güler E, Yılmaz F, et al. SYNTAX score predicts postoperative atrial fibrillation in patients undergoing on-pump isolated coronary artery bypass grafting surgery. Anatol J Cardiol. 2016;16:655–61.
Terzano C, Romani S, Conti V, Paone G, Oriolo F, Vitarelli A. Atrial fibrillation in the acute, hypercapnic exacerbations of COPD. Eur Rev Med Pharmacol Sci. 2014;18:2908–17.
Shibata Y, Watanabe T, Osaka D, Abe S, Inoue S, Tokairin Y, et al. Impairment of pulmonary function is an independent risk factor for atrial fibrillation: the Takahata study. Int J Med Sci. 2011;8:514–22.
Li J, Agarwal SK, Alonso A, Blecker S, Chamberlain AM, London SJ, et al. Airflow obstruction, lung function, and incidence of atrial fibrillation: the Atherosclerosis Risk in Communities (ARIC) study. Circulation. 2014;129:971–80.
Sin DD, Wu L, Man SF. The relationship between reduced lung function and cardiovascular mortality: a population-based study and a systematic review of the literature. Chest. 2005;127:1952–9.
Vanfleteren LEGW, Spruit MA, Groenen M, Gaffron S, van Empel VPM, Bruijnzeel PLB, et al. Clusters of comorbidities based on validated objective measurements and systemic inflammation in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2013;187:728–35.
Pinto LM, Alghamdi M, Benedetti A, Zaihra T, Landry T, Bourbeau J. Derivation and validation of clinical phenotypes for COPD: a systematic review. Respir Res. 2015;16:50.

No competing interests reported.

Causal associations between chronic obstructive pulmonary disease and cardiovascular diseases: a Mendelian randomization study

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Introduction

Materials and methods

Study design

Genetic association datasets

IV selection

Statistical analysis

Univariable MR analysis

Multivariable MR analysis

Mediation MR analysis

Results

Univariable MR analysis

MVMR analysis

Mediation MR analysis

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1