Causal association between myocardial infarction and atrial fibrillation: A bidirectional Mendelian randomization study

DOI: https://doi.org/10.21203/rs.3.rs-2823753/v1

Abstract

Background

At present, many studies have shown a significant correlation between myocardial infarction (MI) and atrial fibrillation (AF), but few focus on the bidirectional causal relationship between MI and AF. Therefore, this Mendelian randomization (MR) study was designed to examine the bidirectional causality between MI and AF.

Methods

We used the publicly available summary statistical dataset of MI from genome-wide analysis studies (GWAS; ebi-a-GCST011364; case = 14,825, control = 2,680). The summary statistical dataset of AF was obtained from a European population GWAS (finn-b-I9_AF_REIMB; case = 10,516, control = 116,926). A two-sample bidirectional MR analysis was performed using analysis methods including inverse-variance weighted (IVW), MR-Egger, and weighted median.

Results

In site-specific MI analyses, we screened 30 single nucleotide polymorphisms (SNPs) from GWAS as instrumental variables (IVs). Causal association between MI and AF can be supported by IVW (beta = 0.349, SE = 0.057, P < 0.001), MR-Egger (beta = 0.398, SE = 0.131, P = 0.005), and weighted median (beta = 0.352, SE = 0.068, P < 0.001). In the reverse MR analyses, we screened 20 SNPs as IVs and the casual effect of AF on MI was observed by IVW (beta = 0.047, SE = 0.022, P = 0.033).

Conclusions

Our MR study results showed a good causal effect of MI on AF. The MR analysis also found a causal effect of AF on MI.

1 Introduction

Myocardial infarction (MI) and atrial fibrillation (AF) are two of the most common cardiovascular diseases worldwide(1). MI is caused by the blockage of coronary arteries leading to the death of cardiac muscle cells, while AF is an arrhythmia characterized by irregular and rapid heartbeat(2, 3). These diseases have a significant impact on morbidity and mortality, with MI being the leading cause of death worldwide and AF increasing the risk of stroke and heart failure(4, 5).

Several risk factors have been identified for both MI and AF, including hypertension, diabetes, smoking, and obesity(6, 7). However, the causal relationship between the two diseases is still unclear, and it is challenging to establish causality due to the complexity and heterogeneity of cardiovascular diseases(8, 9).

Mendelian randomization (MR) is a method that utilizes genetic variants as instrumental variables (IVs) to infer causality between exposures and outcomes(10). This method has become increasingly popular in epidemiological studies because it can provide more reliable causal inference than traditional observational studies(11). MR can also be used to investigate the causal relationship between modifiable risk factors and disease outcomes, which could have important implications for prevention and treatment(12).

In recent years, MR studies have shown that genetic variants associated with modifiable risk factors, such as smoking being associated with MI and hypertension being associated with AF(13, 14). These findings suggest a potential causal relationship between these risk factors and the diseases. However, previous MR studies have mostly focused on the unidirectional causal relationship between the risk factors and the diseases. The bidirectional causal relationship between MI and AF has not been thoroughly investigated using MR(15, 16).

Therefore, in this study, we investigate the bidirectional causal relationship between MI and AF using MR. We will use genetic variants associated with MI or AF as IVs to estimate the causal effect of one disease on the other. This approach will enable us to investigate the potential causal relationship between MI and AF and provide insights into the underlying biological mechanisms.

Overall, our study aims to contribute to a better understanding of the causal relationship between MI and AF, which could have important implications for the prevention and treatment of these diseases. By identifying causality that contributes to the development of MI and AF, we may be able to develop more effective strategies for preventing and treating these conditions.

2 Methods

2.1 Overall study design

All data were obtained from published studies approved by the institutional review boards, and informed consent was obtained from the participants of the original study (17). Therefore, no further sanctions were required (18). The cause-and-effect relationship between MI and AF was analyzed using a two-sample MR study, and SNPs were defined as IVs. The use of SNPs to model randomized controlled trials (RCTs) can help identify causal relationships between exposure characteristics (i.e., MI) and outcome characteristics (i.e., AF).

2.2 Data sources

2.1.1 Genetic instrument variants for exposure

MI data were obtained from the latest and largest published GWAS meta-analysis performed in 2021 by JA et al, including 14,825 MI cases and 2,680 control participants (19). The study was approved by the institutional review board, and informed consent was obtained from all participants in the original study. SNPs were selected based on the following criteria: i) SNPs strongly related to MI with genome-wide significance (p < 5×10− 8). ii) Independent of each other and to avoid bias owing to linkage disequilibrium (LD), the LD of SNPs related to MI had to fulfill r² <0.001, with a window size of 10,000 kb. iii) The correlation between IV and exposure factors is typically determined using the F-statistic of the SNPs. In general, IVs with an F-statistic greater than 10 are regarded as unbiased. F-statistic = (β/SE)2.

2.1.2 Study outcome: atrial fibrillation

Data on AF were obtained from the FinnGen project (FinnGen), which is available at https://gwas.mrcieu.ac.uk/datasets/finn-b-I9_AF_REIMB/ and includes 10,516 cases and116,926 controls of European ancestry.

2.3 Statistical analysis

It is essential to consider these hypotheses to provide a valid explanation for MR analysis (20). (i) It is well-established that IVs are strongly related to MI. (ii) AF is affected only by IVs due to MI defects. (iii) No confounding factors were present in the relationship between MI and AF according to IVs. The results can be affected by genetic variation through a single pathway rather than by separate exposure, namely horizontal pleiotropy, which contradicts the assumptions of MR and may bias the causal estimates. Three different analytical methods were used in the MR analysis to prevent this. Each analysis was based on a different horizontal multiplicity model. The benefit of comparing these three results is that the consistency of the three methods makes the results more credible. The main analysis was performed using an inverse variance weighting (IVW) approach, which provided the most accurate estimates but assumed that all SNPs were valid IVs (21). If one SNP does not meet the IVs assumption, the random-defect IVW will be used to generate a bias, which weighs each rate according to its standard error while considering possible heterogeneity. To satisfy the premise of a valid instrumental variable, the weighted median method requires at least 50% SNPs (22). After sorting the included SNPs based on the weights, we obtained the median of the corresponding distribution function according to the results of our experiments. Additionally, if the genetic instrument does not depend on pleiotropic effects, an effect estimate can be derived from MR-Egger regression (23). The pleiotropic effect was assessed using MR-Egger’s intercept. Furthermore, a directional multiplicative effect cannot be proven if MR-Egger’s intercept does not differ dramatically from zero (24).

2.4 Sensitivity analysis

Funnel plots can plot a single Wald ratio per SNP to display the directional level pleiotropy of the IVs. Nevertheless, the small number of IVs included makes it difficult to test for horizontal pleiotropy using funnel plots. The causal effect of the funnel plot was approximately symmetrical (Fig. 1). Leave-one-out analyses were performed to investigate whether estimates from IVW analyses were biased or dictated by individual SNPs, during meta-analyses that were conducted based on rerun IVW results for the remaining SNPs after omitting one SNP per succession. After removing each SNP, we performed MR analysis again systematically for the remaining SNPs. The results were consistent, indicating a significant causal relationship between the calculated results for all the SNPs (Fig. 2). In MR analysis, the second hypothesis is that SNPs inject results only by modifying the exposure of interest, without other confounding pathways. Directional multidirectionality was examined to obtain the intercept and p-value using MR-Egger regression. No horizontal pleiotropy was observed in the intercept of the MR-Egger regression (p > 0.05), further indicating that pleiotropy did not bias the causal effect. Furthermore, in the published GWAS, there was no evidence that the included MI associated SNPs were significantly associated with any phenotype except MI, which indicates that the assumptions of the third MR were not violated. Consequently, there was no evidence that the genetic instruments of the 30 MI-associated SNPs were significantly associated with any other phenotype on a genome-wide scale, supporting our third MR hypothesis, which is unlikely to be breached in our study. The “Two sample MR” (version 0.5.6) software package was applied for MR and sensitivity analysis in R (version 3.6.2) (25).

3 Results

3.1 Instrumental variables for MI

The SNPs’ signatures of the MI are shown in Supplementary Table S1. Finally, we selected 30 SNPs as the IVs. All genetic tools related to MI were at a genome-wide significance level (p < 5×10− 8, F > 10). Thus, none of the SNPs was susceptible to IVs. The causal effects of each genetic variant on AF are shown in forest (Figs. 3) and scatter plots (Fig. 4).

3.2 Mendelian randomization analyses for AF

We evaluated the causal relationship between MI levels and AF using IVW, MR-Egger, and weighted median regression (Table 1). Our findings suggest an increased risk of AF in patients with MI (OR = 1.41; 95% CI 1.26 − 1.58, p = 6×10− 10).

  
Table 1

MR results of the causal effect between MI and AF and heterogeneity Test

MR methods

nSnp

Pleiotropy

  

Heterogeneity Test

Beta

SE

P Value

Cochran’s Q Statistic

P Value

Causal effect of MI on AF

MR Egger

30

0.398

0.131

0.005

  

50.201

0.006

Weighted median

30

0.352

0.068

< 0.001

      

Inverse-variance weighted

30

0.349

0.057

< 0.001

  

50.506

0.008

Simple mode

30

0.365

0.122

0.006

      

Weighted mode

30

0.365

0.089

< 0.001

      

Causal effect of AF on MI

MR Egger

20

-0.025

0.045

0.580

  

29.164

0.046

Weighted median

20

0.017

0.024

0.485

      

Inverse-variance weighted

20

0.047

0.022

0.033

  

34.464

0.016

Simple mode

20

0.012

0.040

0.763

      

Weighted mode

20

0.015

0.029

0.603

      
MR, mendelian randomization; MI, myocardial infarction; AF, atrial fibrillation; Snp, single nucleotide polymorphism; SE, standard error.


3.3 Bidirectional Mendelian Randomization

We also sought to explore whether AF influenced MI. Therefore, we reversed the functions of the exposures and outcomes to perform a bidirectional MR analysis and determine the effects of a genetically increased risk of AF on MI. To that end, we selected SNPs that were significant genome-wide (p < 5E − 08) and independently inherited (r2 < 0.01) without LD for AF from FinnGen project (Supplementary Table S2). We then applied the same MR methods as above (Table 1). The statistical tests of the bidirectional MR analysis were two-sided, and the results of the MR analyses and sensitivity analyses regarding the causal effects of AF on MI were considered statistically significant. Our findings also suggest an increased risk of MI in patients with AF (OR = 1.04; 95% CI 1.00 − 1.09, p = 0.03). However, we found heterogeneity in the results, which may be relevant to the included population (p = 0.01). (Fig. 5)

4 Discussion

MR is a tool used in causal inference, and its three core assumptions are prerequisites for reliable causal inference. The genetic variant used as an instrument is strongly associated with the exposure, is independent of any confounders, and only affects the outcome through its effect on the exposure. In this study, by using SNPs as the IVs and excluding the bias caused by confounding factors, we elucidated the bidirectional causal relationship between MI and AF using the MR methods. We extracted data on MI and AF from database and ensured that all SNPs selected as IVs were statistically correlated with the exposure factor and not in LD, and we conducted horizontal pleiotropy analysis to exclude SNPs that might be related to confounding factors and eliminated weak IVs. We then harmonized the effect alleles in the GWAS data of the exposure and outcome factors and used five analysis methods to conduct MR, while also testing for heterogeneity and pleiotropy in the IVs. Finally, the study results showed a good causal effect of MI on AF, as evidenced by the consistency and monotonicity of the scatter plot, one-by-one analysis, and distribution of SNPs in the funnel plot. The IVW analysis also found a causal effect of AF on MI but the effect performance is not as good as that of MI on AF.

Our findings are consistent with previous observational studies that have shown a positive association between MI and AF(26). The reasons for MI leading to AF may be related to the following aspects. Experimental studies have found that when the perfusion of the right coronary artery and the left circumflex branch is significantly reduced, the excitability of the atrial myocardium is increased, and the conduction velocity is accelerated, which can cause re-entry and fibrillation(27). After MI occurs, the atrium is excessively stretched which increases the excitement of atrial myocytes and prolongs the length of the electrical conduction pathway, which is conducive to the formation of re-entry and fibrillation(28). In addition, MI-induced myocardial fibrosis also favors the formation of re-entry and fibrillation. Inflammation plays an important role in the occurrence and maintenance of AF. When MI occurs, a large amount of inflammatory factors are released. In the non-infarct and non-ischemic areas, the expression of inflammatory factors also increases, indicating that AF under acute MI conditions may be a marker of widespread inflammation(29). ln addition, the reduced parasympathetic tone, increased sympathetic nerve output, and hormones including B-type natriuretic peptide after MI may be related to the occurrence of AF(30, 31).

The Atherosclerosis Risk in Communities study found that AF patients have a 63% increased risk of acute MI(32). MI The reasons for AF leading to MI may be related to the following aspects. AF is associated with systemic signs of inflammation that could promote a pro-thrombotic state and eventually MI(33). AF and its risk factors, such as hypertension, diabetes, and dyslipidemia, can lead to platelet activation, which is a critical step in the process of MI(34, 35). When the heart rate of AF patients is too fast, it can lead to increased myocardial oxygen consumption and decreased coronary blood flow, which can easily cause type 2 MI due to an imbalance between oxygen supply and demand(36). The traction of the atrial muscle can significantly enhance sympathetic nerve activity, stimulate the release of catecholamines in the heart, and activate adrenergic receptors to cause vasoconstriction(37). However, most of studies are based on laboratory data or observational results. Our study provides stronger evidence for a causal relationship between the two conditions using a MR approach. Additionally, our study is the first to demonstrate a bidirectional causal relationship between MI and AF. The mechanisms underlying the causal relationship between MI and AF remains unclear. Further studies are needed to elucidate the exact mechanisms underlying the causal relationship between MI and AF.

This study has several advantages. MR is used to elucidate the bidirectional causal relationship between MI and AF, providing important clinical implications(38). Clinicians should be aware of the increased risk of AF in patients with a history of MI and consider appropriate management strategies to reduce the risk of AF. Conversely, patients with AF should be screened for underlying cardiovascular diseases, including MI. Additionally, our study reminds the importance of primary prevention of both MI and AF. Moreover, the reliability of causal association evidence based on MR studies is between that of observational epidemiological studies and experimental epidemiological studies(12). In other words, the evidence level of MR studies is higher than that of cohort studies and only slightly lower than that of RCTs studies. However, the high cost of RCTs makes them less feasible. MR can provide powerful evidence when RCTs cannot be implemented, for example, due to ethical reasons. MR studies can effectively overcome biases caused by confounding and reverse causality issues, providing reliable evidence for inferring the causal relationship between exposure factors and outcomes(39). In the MR analysis process, selecting SNPs as IVs can directly infer a causal relationship between exposure and outcome, as SNPs are randomly distributed to individuals through genetic inheritance and are not affected by external environmental and other confounding factors(40).

There are several limitations to our study. First, our study was conducted in individuals of European descent, and the generalizability of our findings to other populations is unclear(41). Second, MR assumes that genetic variants are not associated with confounding factors, and our results could be biased if this assumption is violated. Third, our study only examined the relationship between MI and AF and did not investigate the effect of treatment on either condition. Finally, AF and MI have specific subtypes, such as paroxysmal atrial fibrillation, persistent atrial fibrillation, type 1 MI, and type 2 MI. Due to limitations in the data source, causal relationship analysis cannot be performed on these subtypes.

In future studies, it will be important to investigate the mechanism underlying the causal relationship between MI and AF and to explore potential therapeutic strategies for reducing the risk of both conditions. Additionally, studies in diverse populations are needed to determine the generalizability of our findings. Finally, investigating the effect of treatment on both MI and AF could provide important insights into the management of these conditions.

5 Conclusion

In conclusion, the causal effect of MI on AF was found by our MR analyses, which indicated that the occurrence of MI may increase the risk of AF in European population. In the reverse MR analyses, the causal effect of AF on MI was also observed.

Declarations

Conflict of Interest

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

Author Contributions

JR and ZW contributed to conception and design of the study. JR and YH organized collection and assembly of data. JR and PW performed the statistical analysis. JR wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

Funding

No funding was involved in this study.

Acknowledgments

This manuscript has been released as a pre-print at Research Square (doi:).

Data Availability Statement

Publicly available datasets (ID: ebi-a-GCST011364 and finn-b-I9_AF_REIMB) were achieved from the website (https://gwas.mrcieu.ac.uk/).

References

  1. Mahmood SS, Levy D, Vasan RS, Wang TJ. The Framingham Heart Study and the Epidemiology of Cardiovascular Disease: A Historical Perspective. Lancet (London, England) (2014) 383(9921):999-1008. Epub 2013/10/03. doi: 10.1016/S0140-6736(13)61752-3.
  2. Murphy A, Goldberg S. Mechanical Complications of Myocardial Infarction. The American journal of medicine (2022) 135(12):1401-9. Epub 2022/09/09. doi: 10.1016/j.amjmed.2022.08.017.
  3. Sagris M, Vardas EP, Theofilis P, Antonopoulos AS, Oikonomou E, Tousoulis D. Atrial Fibrillation: Pathogenesis, Predisposing Factors, and Genetics. International journal of molecular sciences (2021) 23(1). Epub 2022/01/12. doi: 10.3390/ijms23010006.
  4. Baman JR, Passman RS. Atrial Fibrillation. Jama (2021) 325(21):2218. Epub 2021/06/02. doi: 10.1001/jama.2020.23700.
  5. Lu L, Liu M, Sun R, Zheng Y, Zhang P. Myocardial Infarction: Symptoms and Treatments. Cell biochemistry and biophysics (2015) 72(3):865-7. Epub 2015/02/02. doi: 10.1007/s12013-015-0553-4.
  6. Jortveit J, Pripp AH, Langorgen J, Halvorsen S. Incidence, Risk Factors and Outcome of Young Patients with Myocardial Infarction. Heart (British Cardiac Society) (2020) 106(18):1420-6. Epub 2020/03/01. doi: 10.1136/heartjnl-2019-316067.
  7. Hendriks JM, Gallagher C, Middeldorp ME, Lau DH, Sanders P. Risk Factor Management and Atrial Fibrillation. Europace : European pacing, arrhythmias, and cardiac electrophysiology : journal of the working groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology of the European Society of Cardiology (2021) 23(23 Suppl 2):ii52-ii60. Epub 2021/04/11. doi: 10.1093/europace/euaa346.
  8. Belkouche A, Yao H, Putot A, Chague F, Rochette L, Danchin N, et al. The Multifaceted Interplay between Atrial Fibrillation and Myocardial Infarction: A Review. Journal of clinical medicine (2021) 10(2). Epub 2021/01/13. doi: 10.3390/jcm10020198.
  9. Lubitz SA, Magnani JW, Ellinor PT, Benjamin EJ. Atrial Fibrillation and Death after Myocardial Infarction: Risk Marker or Causal Mediator? Circulation (2011) 123(19):2063-5. Epub 2011/05/04. doi: 10.1161/CIRCULATIONAHA.111.030171.
  10. Birney E. Mendelian Randomization. Cold Spring Harbor perspectives in medicine (2022) 12(4). Epub 2021/12/08. doi: 10.1101/cshperspect.a041302.
  11. Zheng J, Baird D, Borges MC, Bowden J, Hemani G, Haycock P, et al. Recent Developments in Mendelian Randomization Studies. Current epidemiology reports (2017) 4(4):330-45. Epub 2017/12/12. doi: 10.1007/s40471-017-0128-6.
  12. Davey Smith G, Hemani G. Mendelian Randomization: Genetic Anchors for Causal Inference in Epidemiological Studies. Human molecular genetics (2014) 23(R1):R89-98. Epub 2014/07/30. doi: 10.1093/hmg/ddu328.
  13. Lu Y, Wang Z, Zheng L. Association of Smoking with Coronary Artery Disease and Myocardial Infarction: A Mendelian Randomization Study. European journal of preventive cardiology (2021) 28(12):e11-e2. Epub 2020/02/25. doi: 10.1177/2047487320907747.
  14. Liao LZ, Wen XY, Zhang SZ, Li WD, Zhuang XD. Hypertension and Atrial Fibrillation: A Study on Epidemiology and Mendelian Randomization Causality. Frontiers in cardiovascular medicine (2021) 8:644405. Epub 2021/04/10. doi: 10.3389/fcvm.2021.644405.
  15. Mekhael M, Marrouche N, Hajjar AHE, Donnellan E. The Relationship between Atrial Fibrillation and Coronary Artery Disease: Understanding Common Denominators. Trends in cardiovascular medicine (2022). Epub 2022/10/02. doi: 10.1016/j.tcm.2022.09.006.
  16. Ruddox V, Sandven I, Munkhaugen J, Skattebu J, Edvardsen T, Otterstad JE. Atrial Fibrillation and the Risk for Myocardial Infarction, All-Cause Mortality and Heart Failure: A Systematic Review and Meta-Analysis. European journal of preventive cardiology (2017) 24(14):1555-66. Epub 2017/06/16. doi: 10.1177/2047487317715769.
  17. Lawlor DA. Commentary: Two-Sample Mendelian Randomization: Opportunities and Challenges. International journal of epidemiology (2016) 45(3):908-15. Epub 2016/07/19. doi: 10.1093/ije/dyw127.
  18. Tan JS, Liu NN, Guo TT, Hu S, Hua L. Genetically Predicted Obesity and Risk of Deep Vein Thrombosis. Thrombosis research (2021) 207:16-24. Epub 2021/09/11. doi: 10.1016/j.thromres.2021.08.026.
  19. Hartiala JA, Han Y, Jia Q, Hilser JR, Huang P, Gukasyan J, et al. Genome-Wide Analysis Identifies Novel Susceptibility Loci for Myocardial Infarction. European heart journal (2021) 42(9):919-33. Epub 2021/02/04. doi: 10.1093/eurheartj/ehaa1040.
  20. VanderWeele TJ, Tchetgen Tchetgen EJ, Cornelis M, Kraft P. Methodological Challenges in Mendelian Randomization. Epidemiology (Cambridge, Mass) (2014) 25(3):427-35. Epub 2014/04/01. doi: 10.1097/EDE.0000000000000081.
  21. Burgess S, Butterworth A, Thompson SG. Mendelian Randomization Analysis with Multiple Genetic Variants Using Summarized Data. Genetic epidemiology (2013) 37(7):658-65. Epub 2013/10/12. doi: 10.1002/gepi.21758.
  22. Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genetic epidemiology (2016) 40(4):304-14. Epub 2016/04/12. doi: 10.1002/gepi.21965.
  23. Bowden J, Davey Smith G, Burgess S. Mendelian Randomization with Invalid Instruments: Effect Estimation and Bias Detection through Egger Regression. International journal of epidemiology (2015) 44(2):512-25. Epub 2015/06/08. doi: 10.1093/ije/dyv080.
  24. Zhang Y, Liu Z, Choudhury T, Cornelis MC, Liu W. Habitual Coffee Intake and Risk for Nonalcoholic Fatty Liver Disease: A Two-Sample Mendelian Randomization Study. European journal of nutrition (2021) 60(4):1761-7. Epub 2020/08/29. doi: 10.1007/s00394-020-02369-z.
  25. Broadbent JR, Foley CN, Grant AJ, Mason AM, Staley JR, Burgess S. Mendelianrandomization V0.5.0: Updates to an R Package for Performing Mendelian Randomization Analyses Using Summarized Data. Wellcome open research (2020) 5:252. Epub 2021/01/02. doi: 10.12688/wellcomeopenres.16374.2.
  26. Mori H, Fukui K, Maeda A, Akashi Y, Ako J, Ikari Y, et al. Impact of Atrial Fibrillation and the Clinical Outcomes in Patients with Acute Myocardial Infarction from the K-Active Registry. Journal of cardiology (2022) 79(6):768-75. Epub 2022/02/27. doi: 10.1016/j.jjcc.2022.02.007.
  27. van Diepen S, Siha H, Fu Y, Westerhout CM, Lopes RD, Granger CB, et al. Do Baseline Atrial Electrocardiographic and Infarction Patterns Predict New-Onset Atrial Fibrillation after St-Elevation Myocardial Infarction? Insights from the Assessment of Pexelizumab in Acute Myocardial Infarction Trial. Journal of electrocardiology (2010) 43(4):351-8. Epub 2010/05/07. doi: 10.1016/j.jelectrocard.2010.04.001.
  28. Wang J, Yang YM, Zhu J. Mechanisms of New-Onset Atrial Fibrillation Complicating Acute Coronary Syndrome. Herz (2015) 40 Suppl 1:18-26. Epub 2014/10/30. doi: 10.1007/s00059-014-4149-3.
  29. Brundel B, Ai X, Hills MT, Kuipers MF, Lip GYH, de Groot NMS. Atrial Fibrillation. Nature reviews Disease primers (2022) 8(1):21. Epub 2022/04/09. doi: 10.1038/s41572-022-00347-9.
  30. Li Z, Liu Q, Liu F, Hidru TH, Yang Y, Wang S, et al. Atrial Cardiomyopathy Markers and New-Onset Atrial Fibrillation Risk in Patients with Acute Myocardial Infarction. European journal of internal medicine (2022) 102:72-9. Epub 2022/05/07. doi: 10.1016/j.ejim.2022.04.019.
  31. Chen PS, Chen LS, Fishbein MC, Lin SF, Nattel S. Role of the Autonomic Nervous System in Atrial Fibrillation: Pathophysiology and Therapy. Circ Res (2014) 114(9):1500-15. Epub 2014/04/26. doi: 10.1161/CIRCRESAHA.114.303772.
  32. Soliman EZ, Lopez F, O'Neal WT, Chen LY, Bengtson L, Zhang ZM, et al. Atrial Fibrillation and Risk of St-Segment-Elevation Versus Non-St-Segment-Elevation Myocardial Infarction: The Atherosclerosis Risk in Communities (Aric) Study. Circulation (2015) 131(21):1843-50. Epub 2015/04/29. doi: 10.1161/CIRCULATIONAHA.114.014145.
  33. Guo Y, Lip GY, Apostolakis S. Inflammation in Atrial Fibrillation. Journal of the American College of Cardiology (2012) 60(22):2263-70. Epub 2012/12/01. doi: 10.1016/j.jacc.2012.04.063.
  34. Davi G, Gresele P, Violi F, Basili S, Catalano M, Giammarresi C, et al. Diabetes Mellitus, Hypercholesterolemia, and Hypertension but Not Vascular Disease Per Se Are Associated with Persistent Platelet Activation in Vivo. Evidence Derived from the Study of Peripheral Arterial Disease. Circulation (1997) 96(1):69-75. Epub 1997/07/01. doi: 10.1161/01.cir.96.1.69.
  35. Pignatelli P, Pastori D, Carnevale R, Farcomeni A, Cangemi R, Nocella C, et al. Serum Nox2 and Urinary Isoprostanes Predict Vascular Events in Patients with Atrial Fibrillation. Thrombosis and haemostasis (2015) 113(3):617-24. Epub 2014/11/14. doi: 10.1160/TH14-07-0571.
  36. Camici PG, Gloekler S, Levy BI, Skalidis E, Tagliamonte E, Vardas P, et al. Ivabradine in Chronic Stable Angina: Effects by and Beyond Heart Rate Reduction. International journal of cardiology (2016) 215:1-6. Epub 2016/04/23. doi: 10.1016/j.ijcard.2016.04.001.
  37. Millar PJ, Murai H, Morris BL, Floras JS. Microneurographic Evidence in Healthy Middle-Aged Humans for a Sympathoexcitatory Reflex Activated by Atrial Pressure. American journal of physiology Heart and circulatory physiology (2013) 305(6):H931-8. Epub 2013/07/16. doi: 10.1152/ajpheart.00375.2013.
  38. Burgess S, Thompson SG. Interpreting Findings from Mendelian Randomization Using the Mr-Egger Method. European journal of epidemiology (2017) 32(5):377-89. Epub 2017/05/21. doi: 10.1007/s10654-017-0255-x.
  39. Sekula P, Del Greco MF, Pattaro C, Kottgen A. Mendelian Randomization as an Approach to Assess Causality Using Observational Data. Journal of the American Society of Nephrology : JASN (2016) 27(11):3253-65. Epub 2016/11/02. doi: 10.1681/ASN.2016010098.
  40. Burgess S, Thompson SG, Collaboration CCG. Avoiding Bias from Weak Instruments in Mendelian Randomization Studies. International journal of epidemiology (2011) 40(3):755-64. Epub 2011/03/19. doi: 10.1093/ije/dyr036.
  41. Wang K, Shi X, Zhu Z, Hao X, Chen L, Cheng S, et al. Mendelian Randomization Analysis of 37 Clinical Factors and Coronary Artery Disease in East Asian and European Populations. Genome medicine (2022) 14(1):63. Epub 2022/06/14. doi: 10.1186/s13073-022-01067-1.