Association of asthma with the risk of cardiovascular disease: A Mendelian randomization study

A


Introduction
Currently, the potential impact of asthma on cardiovascular disease risk is of great concern.On one hand, asthma affects >260 million people globally, and most of the patients are first diagnosed during childhood or adolescence (Song et al., 2022).Then, the disease may intermittently or continuously haunt them for years or decades.Prolonged hypoxemia and increased pulmonary resistance induced by asthma might cause temporary or permanent harm to the cardiovascular system (Papi et al., 2018;Porsbjerg et al., 2023).On the other hand, many cardiovascular diseases, such as hypertension and coronary heart disease, are common in the elderly, and some middle-aged people are also affected by these diseases.Considering the close physiological correlation between the respiratory and cardiovascular systems, it is reasonable to speculate that asthma in early life has the potential to influence the risk of cardiovascular disease in older age.
In recent years, two meta-analyses included 40 studies and >4 million subjects, and reported that asthma significantly increased the morbidity and mortality of cardiovascular diseases (Xu et al., 2017;Zhang et al., 2022).A recent study adopted Framingham Offspring Cohort data, and also reported that asthma may be a risk factor for cardiovascular diseases (Pollevick et al., 2021).However, all these findings were reported by observational studies, and were susceptible to confounding factors.
In addition, the potential mechanism by which asthma promoted cardiovascular diseases had not been determined (Cazzola et al., 2023).Previous studies reported that circulating proteins may play a crucial role in the development of cardiovascular diseases (Feldreich et al., 2019;Piarulli et al., 2024).And, many of these proteins were used as diagnostic markers or therapeutic targets for such diseases (Folkersen et al., 2020).So, it was reasonable to assume that revealing the effect of asthma on peripheral cardiovascular proteins may certainly contribute to our understanding of these unknown mechanisms.
Therefore, the present study tried to explore the genetic effect of asthma on five common cardiovascular diseases and 90 peripheral cardiovascular proteins using a two-sample Mendelian randomization (MR) approach, and then to explore which peripheral proteins were involved in the mechanisms by which asthma promoted cardiovascular disease using a two-sample and multivariate MR approach.
The overview of the study was showed in Fig. 1.

Ethics requirements
This study was approved by the Ethics Committee of Qingdao Central Hospital (ID: FT-AEW-2022-13).All data for this study were obtained from anonymized public databases.

Summary data
In this study, asthma was the primary exposure factor.Five asthmarelated diseases were secondary exposure factors, including childhood asthma, obesity-related asthma, nonallergic asthma, asthma with chronic obstructive pulmonary disease (COPD), and asthma-related infection.Five cardiovascular diseases, including hypertension, atrial fibrillation, angina pectoris, myocardial infarction, and heart failure were outcome factors.And, 90 cardiovascular proteins were also outcome factors (Supplemental Table 1).
The data for asthma, and five asthma-related diseases were obtained from a FinnGen human genome-wide association study (GWAS) which provided the genetic insights in a well-phenotyped isolated population, and the sample sizes were 5865 to 42,163 cases and 202,399 to 335,114 controls (Kurki et al., 2023).All these subjects were European subjects (both sexes).All these cases were diagnosed by internationally recognized standards, classified by ICD-10 codes, and described in detail in the relevant literature (Kurki et al., 2023).
The data for hypertension was obtained from a MRC Integrative Epidemiology Unit (MRC-IEU) GWAS which included 54,358 cases and 408,652 controls.The data for atrial fibrillation was collected from a GWAS focusing on new insight into atrial fibrillation biology which included 60,620 cases and 970,216 controls (Nielsen et al., 2018).The data for angina pectoris was included from a GWAS drawing a crosspopulation atlas of genetic associations for 220 human phenotypes which included 30,025 cases and 440,906 controls (Sakaue et al., 2021).The data for myocardial infarction was obtained from a GWAS which focused on the common genetic associations between age-related diseases, and this study included 11,081 cases and 473,517 controls (Dönertas ¸et al., 2021).The data for heart failure was extracted from a genome-wide association and MR analysis providing insights into the pathogenesis of heart failure which included 47,309 cases and 930,014 controls (Shah et al., 2020).All these subjects were European subjects (both sexes), and all these cases were diagnosed by internationally recognized standards.
The data for 90 cardiovascular proteins were obtained from a GWAS study which adopted 21,758 European subjects (both sexes) to identify the protein quantitative trait loci (pQTL) and regulatory pathway maps of these proteins (Folkersen et al., 2020).
The characteristics of these summary data were showed in Table 1.

Instrumental variables
Single nucleotide polymorphisms (SNPs) were the most common type of heritable variation in humans.In this study, the eligible SNPs were used as instrumental variables, and these SNPs were extracted from the summary data of the exposure factors according to the correlation, independence and exclusivity assumptions of MR (Burgess et al., 2017).
The extraction criteria were as follows: (1) The SNPs were relatively common, i.e., the minor allele frequency was >1 %. (2) The genomewide association P-value must be <5 × 10 − 8 , and the F-statistic should be >10.(3) Removing the linkage disequilibrium was conducted using an r2 of 0.001 and a distance of 10,000 kb (i.e., r2<0.001，distance= 10,000 kb).(4) The SNPs significantly associated with the confounding factors were also excluded using the public database "phenoscanner".The confounding factors were all common risk factors for cardiovascular diseases, such as obesity.
Through these steps, this study obtained a series of eligible SNPs that allowed for the following analyses.The data imputation was not conducted, and the proxy instrumental variables were also not adopted.The characteristics of these included SNPs were listed in Supplemental

Harmonization
These SNPs identified in the previous step were again extracted from the summary data of the outcome factors.Then, the study harmonized the allele orientations of the SNPs obtained from the two sources, and formed the datasets that may be used for the following MR analyses.

Preliminary analysis
The random-effect inverse variance weighted (IVW) was used to initially explore the potential impact of asthma on the peripheral levels of 90 cardiovascular proteins (P < 0.05).The power of IVW in detecting causality had been demonstrated, but this method was also susceptible to horizontal pleiotropy (Burgess et al., 2023).So, this method was ideally suited for screening for cardiovascular proteins that were potentially affected by asthma.

Mendelian randomization
Five cardiovascular diseases and a range of screened cardiovascular proteins were included.Four MR methods (i.e.IVW, weighted median, MR-Egger, and MR-PRESSO) were adopted to determine the causal association of asthma with these cardiovascular diseases and proteins (Burgess et al., 2023).Of these methods, the IVW had the most powerful causal inference ability, but it required that all SNPs were unaffected by horizontal pleiotropy; the weighted median allowed that about 50 % of SNPs were horizontally pleiotropic; the MR-Egger allowed about 100 % of SNPs to be horizontally pleiotropic and provided the most conservative results; and the MR-PRESSO can remove the outliers and produce the corrected results, whereas these outliers were implicated in horizontal pleiotropy.If the outcome was a continuous variable, these methods reported β values, 95 % confidence intervals (95%CIs), and Pvalues; If the outcome was a categorical variable, these methods reported odds ratios (ORs), 95%CIs, and P-values.
If the results of the four methods were inconsistent, the final result was considered significant if both of the following criteria were met: (1) the IVW result was significant, and (2) the results of the four methods were in the same direction.Due to multiple comparisons, the Bonferroni correction was used.Briefly, a P-value <0.05 suggested nominal statistical significance, while a P-value <0.002 (including asthma, five cardiovascular diseases, and 19 screened cardiovascular proteins) suggested significant statistical significance.
In the further analyses, the effect of five asthma-related diseases on cardiovascular diseases and cardiovascular proteins were evaluated.A P-value <0.05 suggested statistical significance.
There were three methods for the sensitivity analyses (Zhang and Ghosh, 2021).The Cochran's Q test was used to detect the heterogeneity.The MR-Egger intercept and MR-PRESSO tests were adopted to measure the horizontal pleiotropy.A P-value <0.05 suggested significant heterogeneity or horizontal pleiotropy.
In the mechanistic analyses, the interactions between cardiovascular diseases and cardiovascular proteins were evaluated.Then, the study used a multivariate MR approach to explore the mediating role of cardiovascular proteins in the promotion of cardiovascular diseases by asthma.A P-value <0.05 suggested statistical significance.

Effect of asthma on five cardiovascular diseases
In Table 2, the IVW reported that asthma nominally increased the risk of hypertension and atrial fibrillation (OR = 1.009, 95%CI = 1.003-1.016,P = 0.004; OR = 1.074, 95%CI = 1.024-1.127,P = 0.003).The results of weighted median, MR-Egger and MR-PRESSO supported the results of IVW.The forest plots of these results were showed in Fig. 2. In addition, the IVW reported that asthma did not affect the risk of angina pectoris, myocardial infarction and heart failure (P>0.05).In Supplemental Table 3, the Cochran's Q, MR-Egger intercept, and MR-PRESSO tests did not report any heterogeneity or horizontal pleiotropy (P>0.05).
In Supplemental Table 4, hypertension and atrial fibrillation were included in further analyses.The IVW reported that obesity-related asthma, asthma with COPD, and asthma-related infection increased the risk of hypertension (P = 0.004, P = 0.010, P = 0.010), and childhood asthma increased the risk of atrial fibrillation (P = 3.159e-05).The results of weighted median, MR-Egger and MR-PRESSO also supported the results of IVW.In Supplemental Table 5, the sensitivity analyses did not report any heterogeneity or horizontal pleiotropy (P>0.05).
In Supplemental Table 9, the IVW reported that childhood asthma, obesity-related asthma, non-allergic asthma, and asthma with COPD were associated with the increased level of TNFSF14 (P = 0.026, P = 0.049, P = 0.026, P = 0.006), and childhood asthma, obesity-related asthma, non-allergic asthma, asthma with COPD, and asthma-related infection were associated with the increased level of C -C motif chemokine 4 (P = 0.034, P = 0.039, P = 0.033, P = 0.023, P = 0.001).The other three methods supported these IVW results.In supplemental Table 10, the sensitivity analyses did not find any heterogeneity and horizontal pleiotropy (P>0.05).

Mechanism analysis
In Supplemental Table 11, the interactions between two cardiovascular diseases (i.e., hypertension and atrial fibrillation) and two cardiovascular proteins (i.e., TNFSF14 and C -C motif chemokine 4) were evaluated.The IVW reported that C-C motif chemokine 4 increased the risk of hypertension (OR = 1.003, 95%CI = 1.001-1.005,P = 0.043).The other three methods supported this IVW result.In addition, no other interactions between these diseases and proteins were found in this study (P>0.05).In Supplemental Table 12, the sensitivity analyses did not find any heterogeneity and horizontal pleiotropy (P>0.05).
Thus, asthma, C-C motif chemokine 4, and hypertension were included in the following analysis.The multivariate MR showed that after adjusting for C-C motif chemokine 4, asthma still increased the risk of hypertension, but its P value changed from 0.004 to 0.011 (OR = 1.007, 95%CI = 1.002-1.013,P = 0.011).

Discussion
Cardiovascular diseases are serious threats to public health, and asthma may be an important risk factor for these diseases.However, the correlation between asthma and cardiovascular diseases has not been fully elucidated, and the underlying mechanisms remain unknown.In this study, we used a MR approach to explore the effect of asthma on five common cardiovascular diseases and 90 peripheral cardiovascular proteins, and further explored whether these cardiovascular proteins played a mediating role in the promotion of cardiovascular disease by asthma.The findings obtained had the potential to help us unravel the above queries, and had relatively significant clinical and scientific implications.
Based on the results, the study found that asthma can genetically increase the risk of hypertension and atrial fibrillation by about 0.9 % to 7 %, and childhood asthma, obesity-related asthma, asthma with COPD, and asthma-related infection also affected the risk of these cardiovascular diseases to varying degrees.In these results, the OR value for hypertension deviated from value one by a very small amount, meaning that asthma caused only a very small increase in the risk of the disease.However, hypertension were caused by a combination of many etiological factors.If a single factor can increase its risk by 0.9 %, we thought that this factor cannot be ignored.More importantly, these correlations were present at the genetic level, which implied that asthma may be deeply involved in the pathogenesis of hypertension.
These findings were supported by previous observational studies.For example, Nasreen et al. included over 10,000 subjects, and found that the prevalence of hypertension was significantly higher in asthmatics than in non-asthmatic healthy individuals (Nasreen et al., 2018).Christiansen et al. included approximately 110,000 subjects, and showed a correlation between asthma severity and the development of hypertension (Christiansen et al., 2016).A recent meta-analysis reported by Nogueira-Garcia et al. included seven observational studies and 1.4 million subjects, and showed that asthma was associated with a higher risk of atrial fibrillation (Nogueira-Garcia et al., 2024a).
This study also found that asthma can elevate the peripheral levels of two proteins (i.e., TNFSF14 and C -C motif chemokine 4) by 10 % to 15 % of the standard deviation.Considering the diversity of asthma, this study still explored the effect of multiple types of asthma (e.g., childhood asthma, obesity-related asthma, etc.) on the peripheral levels of these two proteins, and the results obtained were broadly consistent with the original results.More importantly, the following analyses revealed that C -C motif chemokine 4 can increase the risk of hypertension, and adjusting for this protein obviously reduced the strength of the asthma-hypertension correlation.These were suggesting that C -C motif chemokine 4 may play a potential mediating role in the correlation between asthma and hypertension.
The C-C motif chemokine 4 protein had only received attention in recent years, and was considered a cardiovascular protein (Folkersen et al., 2020).In addition, this protein had been closely associated with glucose regulation and inflammatory responses, which were risk factors or causative factors for cardiovascular diseases (Luo et al., 2023).However, the relevance of this protein to asthma had not been reported.A recent study confirmed that C-C motif chemokine 4 was a risk factor for COVID-19 and could be involved in assessing the severity of the disease (Gan et al., 2022).These findings at least confirmed the importance of this protein in cardiovascular and respiratory diseases, and indirectly supported the results of the present study.
The TNFSF14 protein was an immune checkpoint molecule that played an integral role in immune response.On one hand, TNFSF14 was involved in the development of atherosclerosis, and elevated peripheral TNFSF14 levels were independently associated with the occurrence of cardiovascular events in patients with stable coronary artery disease (Hsu et al., 2019).On the other hand, TNFSF14 expression was significantly elevated in patients with severe asthma, and TNFRSF14 expression in mast cells promoted multiple pathophysiological processes of asthma, such as airway inflammation (Frøssing et al., 2022;Sibilano et al., 2016).The findings obtained by the present study further added to our knowledge of this protein.
In addition, the present study found that asthma can nominally affect the levels of 14 other cardiovascular proteins, such as interleukin-18.Because of their limited statistical strength, these findings obtained by the present study were preliminary.So, the exact effect of asthma on these proteins and the role of these proteins in the promotion of Note: MR = Mendelian randomization, IVW = Inverse variance weighted, CI = Confidence interval, TNFSF14 = Tumor necrosis factor ligand superfamily member 14.
S. Han et al. cardiovascular diseases by asthma needed to be explored in further studies.
One MR study in 2022 had explored the impact of asthma on cardiovascular disease, and reported that asthma can affect the risk of atrial fibrillation and heart failure, which was not entirely consistent with the present study (Chen et al., 2022).It was important to note that the present study was an updated analysis for this topic.The relevant explanations were listed as follows: (1) Though the same data for atrial fibrillation and heart failure were used in both studies, the summary data on the main exposure factor (i.e., asthma) in the present study was published in 2023, which had a larger sample size and provided 41 suitable instrumental variables.And, the previous MR study only collected 9 instrumental variables for predicting asthma.So, the present study had better test efficacy and more reliable results.(2) The present study conducted several subgroup analyses, and explored the effect of different types of asthma on the risk of cardiovascular diseases.(3) The present study also explored the effect of asthma on 90 cardiovascular proteins, and conducted further mechanism analyses.Therefore, we believed that the present study provided more valuable information on this topic.
Atrial fibrillation was a cardiac arrhythmia, which was not classified as a typical cardiovascular event (such as myocardial infarction).However, several observational studies had reported that asthma can increase the risk of atrial fibrillation at the phenotypic level (Nogueira-Garcia et al., 2024b).So, it was necessary to explore this issue at the genetic level.At present, previous studies had not focused on the mechanisms involved.Based on existing knowledge, we speculated that the mechanism by which asthma increased the risk of atrial fibrillation may be implicated in the inflammatory response.Because asthma was an inflammatory disease, while previous studies had also revealed that inflammation can alter atrial electrophysiological process and anatomy through multiple pathways to induce atrial fibrillation (Hu et al., 2015).In addition, the mechanism by which asthma increased the risk of other cardiovascular diseases may be related to the atherosclerosis because previous animal experiments had reported that asthma can exacerbate atherosclerosis in a variety of ways (Gao et al., 2019;Wang et al., 2014).However, the atherosclerosis was also related to inflammation, so there might be a complex relationship between the multiple mechanisms by which asthma promoted cardiovascular diseases.
In the study, the Bonferroni method was used to correct the P-values due to multiple comparisons.This approach was very rigorous, and can minimize the false positive rates of the results, but it was also possible to reject some potential true positive results (Curran-Everett, 2000).In contrast, false discovery rate (FDR) may be a more gentle and balanced correction method (Curran-Everett, 2000).However, the aim of this study was to investigate the pathogenic effect of asthma on several cardiovascular diseases, and to further screen out the proteins most likely to be involved in the above mechanisms from up to 90 cardiovascular proteins.Based on these aims, we were more concerned with being able to screen for the most meaningful cardiovascular proteins rather than the possibility of negating some potentially positive results.Therefore, the Bonferroni method may be more helpful in achieving the objectives of this study.Meanwhile, the strategy of this study did not really lose those potential positive results which did not reach the Bonferroni corrected P-value, as they were also reported by the study, and were defined as nominal positive results.From these aspects, the strategy of this study not only achieved the main purpose of the study, but also provided comprehensive reference information for future studies.
The main strength of this study was the MR design.Compared to observational studies, MR studies were epidemiological analyses at the genetic level, which greatly avoided the influence of confounding factors and made the results more reliable.Another advantage of this study was its comprehensiveness.Because as many as five cardiovascular diseases and 90 cardiovascular proteins were studied in this study, and especially because these proteins were summarized based on a large number of results from current research on cardiovascular diseases, they can cover most aspects of the pathogenesis of cardiovascular diseases.The third strength of this study was the mechanistic analyses, which revealed the potential mediating role of the C -C motif chemokine 4 mentioned above.The main limitation of this study was that the analysis was conducted only in European populations, but not in Asian or African populations.The second limitation of this study was that the potential effect of asthma medications on the risk of cardiovascular diseases was not explored.The explanation for these limitations was insufficient data.So, the present study planned to conduct further analyses to make up for these limitations after the data had been updated in the future.
In conclusion, the study confirmed that asthma was a risk factor for hypertension and atrial fibrillation at the genetic level, and asthma also genetically increased the peripheral levels of TNFSF14 and C -C motif chemokine 4.More importantly, the study reported that C-C motif chemokine 4 might play a mediating role in the mechanism by which asthma promoted hypertension.These findings helped us to fully understand the health risks of asthma and helped us to better prevent and control cardiovascular diseases.

Table 1
Characteristics of summary data in the study.

Table 2
Mendelian randomization analyses for asthma affecting on several cardiovascular diseases.

Table 3
Mendelian randomization analyses for asthma affecting on peripheral cardiovascular protein levels.