Effects of rheumatoid arthritis on ventricular structure and function: a Mendelian randomization study

doi:10.21203/rs.3.rs-4713343/v1

Download PDF

Article

Effects of rheumatoid arthritis on ventricular structure and function: a Mendelian randomization study

https://doi.org/10.21203/rs.3.rs-4713343/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Several observational studies have suggested an association between rheumatoid arthritis (RA) and abnormalities of ventricular structure and function, yet the causal relationship remains unclear. This study aimed to assess the causal association between seropositive and seronegative RA and ventricular structure and function via Mendelian randomization (MR). Genetic data were derived from a genome-wide association study (GWAS) of seropositive and seronegative RA conducted within the FinnGen study, and a GWAS of 16 biventricular functional and structural cardiac magnetic resonance (CMR) measurements involving subjects from the UK Biobank. Inverse variance weighting (IVW) served as the primary analytical strategy to determine the causal associations between seropositive RA, seronegative RA, and ventricular CMR parameters. The MR-Egger, weighted median (WME), weighted mode (WM), and sample mode (SM) methods were employed as supplementary analyses. Additionally, the direct effects of RA on ventricular structure and function were further explored via multivariate MR (MVMR). IVW methods demonstrated that genetically predicted seropositive RA was significantly associated with a reduced left ventricular (LV) stroke volume (adjusted without body surface area (BSA) and systolic blood pressure (SBP)) (OR = 0.635; 95%CI 0.482–0.837; P = 0.001). MVMR analysis indicated that seropositive RA could influence LV stroke volume independently of traditional cardiovascular disease risk factors (OR = 0.695; 95%CI 0.512–0.942; P = 0.019). Sensitivity analyses yielded robust and reliable results. Our genetic analyses revealed a significant causal relationship between the development of seropositive RA and reduced LV stroke volume, offering valuable insights into the prevention and treatment of adverse cardiovascular events in patients with RA.

Health sciences/Rheumatology/Rheumatic diseases/Rheumatoid arthritis

Health sciences/Cardiology/Cardiovascular biology/Cardiovascular diseases

Health sciences/Cardiology/Cardiovascular biology/Cardiovascular genetics

rheumatoid arthritis

cardiac magnetic resonance

causal relationship

mendelian randomization

Rheumatoid arthritis (RA) is an autoimmune disease characterized primarily by erosive arthritis, which can also manifest with pulmonary and cardiovascular complications [1]. The Global Burden of Disease (GBD) study indicates a global prevalence of RA of 0.27%, with both prevalence and incidence on the rise [2]. A prospective study has demonstrated that the relative risk of death from cardiovascular events in RA patients is two to five times higher than that in the general population [3]. In this context, the identification of cardiac morphological and functional abnormalities in RA patients is crucial for the diagnosis and treatment of early concomitant cardiovascular disease.

Subclinical inflammation in RA can manifest in multiple organs, including the heart; increasingly, a growing number of studies have identified abnormalities in RA ventricular structure and function, particularly during the early subclinical disease stage [4–7]. For instance, Ntusi et al., utilizing cardiac magnetic resonance (CMR), observed that RA predisposes individuals to left ventricular (LV) enlargement, deterioration in LV systolic function, myocardial swelling, myocardial fibrosis, and an increase in the cardiomyocyte's extracellular space volume [6]. Kobayashi et al. estimated that among RA patients, the prevalence of LV myocardial fibrosis foci ranges from 32–39%, while the incidences of active inflammation and perfusion defects in the LV myocardium stood at 12% and 11%, respectively [5, 8]. Despite observational studies establishing associations between RA and abnormal CMR characteristics, limitations including small sample sizes, varied clinical characteristics, and reverse causality could undermine the validity of these findings. Consequently, the causal relationship between RA and ventricular morphology and function remains uncertain.

Mendelian Randomization (MR) employs genetic variations as instrumental variables (IVs) to establish causal inferences between exposure factors and outcomes [9]. Compared to traditional observational methods, MR studies can substantially reduce the influence of confounding factors in causal analyses [10]. In this study, we utilized data from a genome-wide association study (GWAS) of 16 biventricular CMR traits measured in 36 548 United Kingdom Biobank (UKB) subjects, and investigated the genetic causality between RA and ventricular function and structure using a two-sample MR method. This analysis aims to contribute to a deeper understanding of the impact of RA on ventricular morphology and structure and provide insights into the diagnosis and treatment of RA-complicated cardiovascular diseases.

Data sources

GWAS data for seropositive and seronegative RA were obtained from the FinnGen database (https://www.r10.finngen.fi/), comprising 4 596 seropositive RA patients and 172 834 controls, and 1 937 seronegative RA patients and the same number of controls. Case diagnoses were made with reference to M13 codes from the International Classification of Diseases, 10th edition (ICD-10). The GWAS regarding biventricular structure and function was derived from a study published by Schmidt et al. in 2023, which collected CMR images and genotyping data from 36 548 UKB subjects to explore genetic variants associated with biventricular CMR measurements [11]. CMR imaging is considered the gold standard for quantifying the function and morphology of the atria and ventricles, providing highly accurate morphological information capable of detecting changes at an early stage [12, 13]. Moreover, to minimize the effect of pre-existing diseases, subjects with prevalent conditions known to affect the characteristics of the left or right ventricle, such as myocardial infarction, heart failure, and congenital heart disease, were excluded. These studies received approval from respective ethical bodies, and informed consent was obtained from all participants.

Screening of instrumental variables

Single-nucleotide polymorphisms (SNPs) used as IVs must satisfy three assumptions [9]: i) IVs must be closely related to exposure factors; ii) IVs must not influence outcomes through pathways other than exposure; and iii) IVs must not be directly associated with outcomes unless such an association is established through exposure factors. The MR design flow is presented in Fig. 1. Based on the above assumptions, valid IVs were screened. To confirm the adequacy of valid IVs, SNPs closely associated with seropositive and seronegative RA were screened using significance thresholds of P < 5×10^− 8 and P < 5×10^− 6, respectively. Subsequently, SNPs with r² ≥ 0.001 and linkage disequilibrium (LD) distance ≤ 10 000 kb were excluded. F-statistic values were calculated using the formula [(r²×(n-2))/(1-r²)], selecting IVs with F-statistics greater than 10 to minimize bias from weak instruments [14]. The PhenoScanner platform (http://www.phenoscanner.medschl.cam.ac.uk/) was used to exclude SNPs directly linked with confounders such as obesity, smoking, diabetes, and hypertension. SNPs related to exposure factors and outcomes were pooled, and IVs with palindromic sequences were removed. The remaining SNPs after screening were ultimately utilized as IVs for the exposure factors in MR analysis.

Two-sample MR

The causal associations between seropositive and seronegative RA and biventricular CMR parameters were evaluated using a two-sample MR analysis. Inverse variance weighted (IVW) combines the Wald estimates for each SNP, considering each SNP as a natural experiment, and aims to simulate randomized controlled trials through the nature of random assignment. IVW enforces that the intercept of the slope of the regression is zero; when the genetic variants selected are all valid IVs, IVW is expected to provide the most accurate results [15, 16]. Consequently, the P-value of IVW was used as the main index for evaluation. MR-Egger, weighted median (WME), weighted mode (WM), and sample mode (SM) analyses were utilized to assist in the assessment of the MR effect value, whereby the direction of ORs across different methods being consistent with IVW indicated the robustness of the results [17]. The estimation framework of the MR-Egger method is based on the IVW method, however, unlike the IVW method, MR-Egger's regression analyses consider the presence of an intercept, relaxing the hypothesis criterion to satisfy the InSIDE assumption that the estimates of the causal effects remain uncorrelated with the pleiotropy of the IVs [18]. WME primarily assigns different weights to various genetic variants, thereby reducing the influence of extreme genetic variants on causal inference and enhancing the stability of results. Furthermore, the WME method boasts advantages such as increased estimation accuracy, broad applicability, and flexibility [17]. The WM method can derive a single causal effect estimate from multiple genetic variant IVs, exhibiting less bias than other methods in most instances and a lower Type I error rate [19]. The SM method categorizes SNPs into clusters according to similarity values and utilizes the cluster with the greatest number of SNPs to assess causal associations [20].

Sensitivity analyses

The heterogeneity of genetic variance estimates was evaluated using the IVW and the MR-Egger method via Cochran's Q test. A P-value less than 0.05 indicated heterogeneity among SNPs. In the presence of significant SNP heterogeneity, a random effects IVW model was employed; otherwise, a fixed effects model was utilized [10]. The MR-Egger method employs a weighted linear regression of gene outcome coefficients on gene exposure coefficients. Causal effect estimates were expressed as the slopes of linear regressions. Estimates of the average level of horizontal pleiotropy in genetic variants are represented by intercepts; a P-value greater than 0.05 implies absence of significant horizontal pleiotropy [21]. The MR-PRESSO test was applied to identify and adjust outliers for horizontal pleiotropy, assessing effects post-correction using an outlier-corrected method and by removing outliers among the IVs [22]. The Steiger test assesses the directionality of IVs, determining their influence on outcomes through exposure factors, thereby mitigating reverse causality.

Multivariate MR

We further employed a MVMR approach to demonstrate the direct effect of seropositive RA on ventricular structure and function by adjusting for confounders such as smoking, hypertensive heart disease, diabetes and other endocrine diseases. IVW methods, based on weighted linear regression, were employed to infer causality in MVMR analyses.

Statistical analyses

All MR analyses were conducted using the “TwoSampleMR” and “MR PRESSO” packages in R 4.0.2. The significance threshold was established at P < 0.05/32, corresponding to a Bonferroni correction for 32 outcomes. Results were deemed significant when P < 0.002, and suggestive when 0.002 < P < 0.05.

Selection of instrumental variables

Initial screening identified 13 IVs associated with seropositive RA and 22 IVs for seronegative RA. Using the PhenoScanner platform, we then excluded 3 SNPs related to confounding factors such as diabetes, hypothyroidism and systemic lupus erythematosus from the seropositive RA-associated genetic variants. These SNPs include rs34536443, rs6679677, and rs7731626. Three SNPs—rs2476601, rs10168266, and rs7731626—associated with body mass index, hypothyroidism, diabetes mellitus, and systemic lupus erythematosus, were excluded from the genetic variants linked to seronegative RA. All IVs had an F-statistic greater than 20, indicating the absence of weak IVs. Ultimately, 10 SNPs for seropositive RA and 19 SNPs for seronegative RA were included as IVs.

Causal relationship between seropositive RA and CMR measurements

IVW fixed-effects modeling demonstrated a significant association between seropositive RA and a reduction in LV stroke volume (adjusted without body surface area (BSA) and systolic blood pressure (SBP)) (OR = 0.635; 95%CI 0.482–0.837; P = 0.001). The study also provided suggestive evidence for an association between seropositive RA and a decrease in LV stroke volume (OR = 0.715; 95%CI 0.559–0.913; P = 0.007), as well as LV peak filling rate (adjusted without BSA and SBP) (OR = 0.110; 95%CI 0.016–0.782; P = 0.027) (Table 1). MR Egger regression showed no evidence of horizontal pleiotropy (P > 0.05), indicating that the IVs did not influence ventricular CMR parameters via pathways other than those associated with seropositive RA (Table 2). Heterogeneity tests revealed no significant heterogeneity among the IVs (P > 0.05). The MR-PRESSO analyses detected no outliers, further affirming the robustness of the findings (Table 3). The Steiger directionality test confirmed the accuracy of the causal direction from exposure to outcome, ruling out potential reverse causality.

Table 1

MR estimates for the causal effect of RA on CMR parameters
Exposures	Outcomes	Methods	nSNPs	OR (95%CI)	P value
Seropositive RA	LV stroke volume	MR Egger	6	0.812 (0.530, 1.243)	0.392
		Weighted median	6	0.738 (0.571, 0.955)	0.021
		Inverse variance weighted	6	0.715 (0.559, 0.913)	0.007
		Simple mode	6	0.780 (0.494, 1.230)	0.333
		Weighted mode	6	0.752 (0.589, 0.960)	0.071
	LV peak filling rate (adjusted without BSA and SBP)	MR Egger	6	0.081 (0.002, 2.989)	0.244
		Weighted median	6	0.105 (0.022, 0.508)	0.005
		Inverse variance weighted	6	0.110 (0.016, 0.782)	0.027
		Simple mode	6	0.110 (0.005, 2.291)	0.213
		Weighted mode	6	0.116 (0.024, 0.565)	0.045
	LV stroke volume (adjusted without BSA and SBP)	MR Egger	6	0.665 (0.401, 1.105)	0.190
		Weighted median	6	0.636 (0.478, 0.847)	0.002
		Inverse variance weighted	6	0.635 (0.482, 0.837)	0.001
		Simple mode	6	0.739 (0.414, 1.321)	0.354
		Weighted mode	6	0.652 (0.473, 0.900)	0.048
Seronegative RA	RV ejection fraction	MR Egger	18	1.014 (0.874, 1.175)	0.862
		Weighted median	18	1.133 (0.975, 1.316)	0.104
		Inverse variance weighted	18	1.103 (1.005, 1.211)	0.039
		Simple mode	18	1.261 (0.962, 1.653)	0.112
		Weighted mode	18	1.192 (0.999, 1.423)	0.068
	LV ejection fraction (adjusted without BSA and SBP)	MR Egger	18	0.996 (0.821, 1.208)	0.967
		Weighted median	18	1.189 (1.012, 1.396)	0.035
		Inverse variance weighted	18	1.142 (1.007, 1.295)	0.038
		Simple mode	18	1.361 (1.004, 1.845)	0.063
		Weighted mode	18	1.230 (0.996, 1.520)	0.072
	RV ejection fraction (adjusted without BSA and SBP)	MR Egger	16	1.035 (0.871, 1.231)	0.701
		Weighted median	16	1.140 (0.984, 1.321)	0.081
		Inverse variance weighted	16	1.112 (1.000, 1.236)	0.049
		Simple mode	16	0.914 (0.682, 1.224)	0.554
		Weighted mode	16	1.195 (0.996, 1.435)	0.075
SNP: Single-nucleotide polymorphism; RA: rheumatoid arthritis; OR: odds ratio; CI: confidence interval; LV: left ventricular; RV: right ventricular; BSA: body surface area; SBP: systolic blood pressure.

Table 2

MR-Egger test for directional pleiotropy
Exposures	Outcomes	Intercept	SE	P value
Seropositive RA	LV stroke volume (adjusted without BSA and SBP)	-0.026	0.114	0.833
	LV stroke volume	-0.070	0.096	0.503
	LV peak filling rate (adjusted without BSA and SBP)	0.170	0.811	0.844
Seronegative RA	RV ejection fraction	0.041	0.029	0.175
	RV ejection fraction (adjusted without BSA and SBP)	0.037	0.037	0.326
	LV ejection fraction (adjusted without BSA and SBP)	0.066	0.037	0.098
SE: standard error; RA: rheumatoid arthritis; LV: left ventricular; RV: right ventricular; BSA: body surface area; SBP: systolic blood pressure.

Table 3

Heterogeneity test results
Exposures	Outcomes	MR Egger		Inverse variance weighted		MR-PRESSO
Exposures	Outcomes	Q	P value	Q	P value	Global test P value	Distortion test P value
Seropositive RA	LV stroke volume (adjusted without BSA and SBP)	6.030	0.197	6.106	0.296	0.312	NA
	LV stroke volume	4.973	0.290	5.646	0.342	0.385	NA
	LV peak filling rate (adjusted without BSA and SBP)	9.405	0.052	9.509	0.090	0.292	NA
Seronegative RA	RV ejection fraction	15.945	0.457	17.963	0.391	0.388	NA
	RV ejection fraction (adjusted without BSA and SBP)	17.183	0.247	18.457	0.239	0.239	NA
	LV ejection fraction (adjusted without BSA and SBP)	26.201	0.051	31.261	0.019	0.015	0.395
RA: rheumatoid arthritis; LV: left ventricular; RV: right ventricular; BSA: body surface area; SBP: systolic blood pressure.

Causal relationship between seronegative RA and CMR measurements

The IVW fixed-effects model identified a nominal causal effect linking seronegative RA to reductions in right ventricular ejection fraction (OR = 1.103; 95%CI 1.005–1.211; P = 0.039) and biventricular ejection fraction (adjusted without BSA and SBP) (left: OR = 1.142; 95%CI 1.007–1.295; P = 0.038; right: OR = 1.112; 95%CI 1.000–1.236; P = 0.049) (Table 1). Genetic correlation estimates for these cardiac phenotypes in seronegative RA indicated no horizontal pleiotropy (P > 0.05) (Table 2). The Steiger test confirmed the accuracy of the causal direction estimates.

Multivariate Mendelian randomization

MVMR analysis demonstrated a significant association between the development of seronegative RA and LV stroke volume (adjusted without BSA and SBP), even after adjusting for traditional cardiovascular risk factors (OR = 0.695; 95% CI 0.512–0.942; P = 0.019) (Fig. 2.).

This study represents the inaugural investigation into the causal relationship between RA and ventricular structural function using MR methods. Our study genetically established that the development of seropositive RA is significantly associated with a reduction in LV stroke volume (adjusted without BSA and SBP).

Cardiac complications associated with RA are typically insidious in presentation and frequently lack overt clinical symptoms. The patient's activity level is often limited due to pronounced symptoms of swelling and pain in the extremities and joints. As a result, symptoms of mild cardiac insufficiency may often be overlooked, potentially resulting in severe cardiac dysfunction in the later stages of the disease, a significantly shorter life expectancy, and a markedly higher mortality rate [23]. Observational studies indicate that comorbid cardiac morphological and functional abnormalities are prevalent in large populations of individuals with RA lacking a history of cardiac disease [24, 25]. Our study has established that the onset of seropositive RA is associated with a reduction in LV stroke volume (adjusted without BSA and SBP), indicative of deteriorating LV systolic function. Consistent with our study, Tański et al. observed that RA was associated with deterioration of LV systolic function, as measured by CMR, in subjects who exhibited no clinical signs of cardiac injury [24]. A cross-sectional study demonstrated that patients with RA who had no cardiovascular disease exhibited significantly lower global longitudinal strain (GLS) in the left ventricle, indicative of LV systolic dysfunction [25]. A prospective analysis revealed that combined circumferential and longitudinal LV systolic dysfunction was identified in approximately 25% of patients with asymptomatic RA [26]. Ntusi et al. reported a significantly impaired LV peak systolic and diastolic global circumferential strain (GCS) in patients with RA as compared to controls, despite exhibiting normal global LV systolic function [27]. Midtbø et al. discovered through echocardiographic evaluation of RA patients with varying levels of disease activity that, despite a normal ejection fraction, active RA correlated with reductions in LV systolic myocardial function and was unlinked to traditional cardiovascular risk factors such as diabetes and smoking [28], which suggests that RA could be an independent risk factor for deterioration of LV systolic function. The aforementioned studies initially revealed the impact of RA on LV systolic function, but did not account for confounding factors such as disease duration, time on treatment, and other comorbidities; we addressed these limitations with the MR approach. To further mitigate the impact of confounding factors in our MR analysis, adjustments were made for smoking, hypertensive heart disease, diabetes, and endocrine disorders, and the findings still indicated a significant causal relationship between seropositive RA and reduced LV stroke volume (adjusted without BSA and SBP), highlighting that seropositive RA itself might be a significant factor.

Inflammation in RA patients alters the applicability of traditional risk factors when assessing ventricular function. Recent studies demonstrate that the mechanisms underlying LV systolic dysfunction in RA patients encompass hemodynamic alterations, inflammation, immunity, and oxidative stress, all contributing to direct myocardial damage [29]. Multiple lines of evidence indicate that disease activity correlates with decreased LV systolic function among RA patients [28]. Reduced disease activity in seronegative RA, compared to seropositive RA [30], may account for the inability of MR analysis to replicate findings in seronegative RA cohorts. A prospective study also observed that patients with early RA, who had higher disease activity and were positive for RF and ACPA, exhibited prolonged T1 relaxation times in basal myocardial segments [31]. Furthermore, given the persistent systemic inflammation observed in RA, the outcomes of MR analyses underscore the need for additional research on disease mechanisms and prompt further exploration of inflammation and its impact on abnormal ventricular function, in light of these findings.

Generally, alterations in LV function and structure are linked to an elevated risk of cardiovascular death [32]. Our findings could prove clinically useful in identifying a subgroup of RA patients at very high risk for adverse cardiovascular events. Current therapeutic strategies for RA remain reliant on aggressive antirheumatic therapy, thus targeting a reduction in disease activity to attenuate or stabilize myocardial involvement progression. For example, Kobayashi et al. demonstrated that tocilizumab treatment elevated LV ejection fraction and reduced LV mass index in RA patients, correlated with a reduction in RA disease activity [33]. A prospective study observed a decrease in disease activity scores and an improvement in ventricular function in patients with active early RA over one year of treatment with disease-modifying antirheumatic drugs (DMARDs) [34]. Lehmonen et al. noted an improvement in myocardial function and a decrease in inflammatory activity, as detected by CMR markers, in patients with active early RA after one year of antirapeutic therapy [35]. Yokoe et al. demonstrated that biological antirheumatic agents could be linked to reduced disease activity and improved LV function in RA patients [36]. Thus, our findings offer new insights into the prevention and treatment of cardiovascular events in reversible-stage RA and underscore the importance of addressing LV dysfunction in the early clinical management of RA. The prevalence and predictors of abnormal LV systolic function in RA patients remain unknown; therefore, it is advised that these patients undergo cardiac systemic examination as soon as possible, and that early detection of myocardial changes is crucial to ensuring early therapeutic intervention. Emphasis should also be placed on the early treatment of active RA, as it not only prevents joint destruction but also mitigates concomitant myocardial damage.

Although this study suggests a causal effect between seropositive RA and reduced LV stroke volume (adjusted without BSA and SBP), several limitations must be acknowledged. First, this study was conducted with a European population, and its findings may not be directly applicable to other populations. Secondly, additional MR analyses with larger sample sizes or randomized controlled trials are essential to enhance the accuracy of the results. Third, MR analyses did not extend to exploring the biological mechanisms underlying the reduction of LV stroke volume in seropositive RA. Finally, given that this study relied upon observed correlation data, drawing absolute causal conclusions is not possible.

In conclusion, our study establishes a significant causal relationship between the development of seropositive RA and a reduction in LV stroke volume (adjusted without BSA and SBP). This finding emphasizes the importance of routine cardiac screening and aggressive antirheumatic therapy for patients with RA, which are essential for the prevention and treatment of adverse cardiovascular events.

Acknowledgments

The authors appreciate the public datasets provided by the research or consortiums cited and included in this study.

Author contributions

Funding acquisition: D.L.; conceptualization: H.L., D.L.; investigation: H.L., D.Y.; formal analysis: H.L.; writing – original draft: H.L.; writing – review & editing:D.L., W.Y., J.L. All authors reviewed the manuscript.

Funding

This research was funded by the Shandong Provincial Natural Science Foundation Joint Special Fund Project (grant number ZR2021LZY002) and the 2020 Qilu Outstanding Young Talent Cultivation Project for Health and Wellness (grant number rc2021002-23).

Data Availability Statement

The datasets generated during and analyzed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest

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

Ethics approval and consent to participate

All data were sourced from published studies or publicly available GWAS summary data for which ethical approval and informed consent had been obtained.

Sanghavi N, Ingrassia JP, Korem S, Ash J, Pan S, Wasserman A. Cardiovascular Manifestations in Rheumatoid Arthritis. Cardiol Rev. 2024; 32(2): 146–152.
Finckh A, Gilbert B, Hodkinson B, Bae SC, Thomas R, Deane KD, et al. Global epidemiology of rheumatoid arthritis. Nat Rev Rheumatol. 2022; 18(10): 591–602.
Riise T, Jacobsen BK, Gran JT, Haga HJ, Arnesen E. Total mortality is increased in rheumatoid arthritis. A 17-year prospective study. Clin Rheumatol. 2001; 20(2): 123–127.
Mavrogeni S, Dimitroulas T, Sfikakis PP, Kitas GD. Heart involvement in rheumatoid arthritis: multimodality imaging and the emerging role of cardiac magnetic resonance. Semin Arthritis Rheum. 2013; 43(3): 314–324.
Kobayashi Y, Giles JT, Hirano M, Yokoe I, Nakajima Y, Bathon JM, et al. Assessment of myocardial abnormalities in rheumatoid arthritis using a comprehensive cardiac magnetic resonance approach: a pilot study. Arthritis Res Ther. 2010; 12(5): R171.
Ntusi NAB, Piechnik SK, Francis JM, Ferreira VM, Matthews PM, Robson MD, et al. Diffuse Myocardial Fibrosis and Inflammation in Rheumatoid Arthritis: Insights From CMR T1 Mapping. JACC Cardiovasc Imaging. 2015; 8(5): 526–536.
Holmström M, Koivuniemi R, Korpi K, Kaasalainen T, Laine M, Kuuliala A, et al. Cardiac magnetic resonance imaging reveals frequent myocardial involvement and dysfunction in active rheumatoid arthritis. Clin Exp Rheumatol. 2016; 34(3): 416–423.
Kobayashi H, Kobayashi Y, Yokoe I, Akashi Y, Takei M, Giles JT. Magnetic Resonance Imaging-Detected Myocardial Inflammation and Fibrosis in Rheumatoid Arthritis: Associations With Disease Characteristics and N-Terminal Pro-Brain Natriuretic Peptide Levels. Arthritis Care Res (Hoboken). 2017; 69(9): 1304–1311.
Birney E. Mendelian Randomization. Cold Spring Harb Perspect Med. 2022; 12(4).
Taschler B, Smith SM, Nichols TE. Causal inference on neuroimaging data with Mendelian randomization. Neuroimage. 2022; 258: 119385.
Schmidt AF, Bourfiss M, Alasiri A, Puyol-Anton E, Chopade S, van Vugt M, et al. Druggable proteins influencing cardiac structure and function: Implications for heart failure therapies and cancer cardiotoxicity. Sci Adv. 2023; 9(17): eadd4984.
Schmidt A, Finan C, Bourfiss M, Velthuis B, Puyol-Antón E, Alasiri A, et al. Cardiac MRI to guide heart failure and atrial fibrillation drug discovery: a Mendelian randomization analysis. Res Sq. 2023.
Ricci F, Aung N, Gallina S, Zemrak F, Fung K, Bisaccia G, et al. Cardiovascular magnetic resonance reference values of mitral and tricuspid annular dimensions: the UK Biobank cohort. J Cardiovasc Magn Reson. 2020; 23(1): 5.
Pierce BL, Ahsan H, Vanderweele TJ. Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants. Int J Epidemiol. 2011; 40(3): 740–752.
Liu B, Ye D, Yang H, Song J, Sun X, Mao Y, et al. Two-Sample Mendelian Randomization Analysis Investigates Causal Associations Between Gut Microbial Genera and Inflammatory Bowel Disease, and Specificity Causal Associations in Ulcerative Colitis or Crohn's Disease. Front Immunol. 2022; 13: 921546.
Yin KJ, Huang JX, Wang P, Yang XK, Tao SS, Li HM, et al. No Genetic Causal Association Between Periodontitis and Arthritis: A Bidirectional Two-Sample Mendelian Randomization Analysis. Front Immunol. 2022; 13: 808832.
Hartley AE, Power GM, Sanderson E, Smith GD. A Guide for Understanding and Designing Mendelian Randomization Studies in the Musculoskeletal Field. JBMR Plus. 2022; 6(10): e10675.
Burgess S, Thompson SG. Interpreting findings from Mendelian randomization using the MR-Egger method. Eur J Epidemiol. 2017; 32(5): 377–389.
Hartwig FP, Davey Smith G, Bowden J. Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. Int J Epidemiol. 2017; 46(6): 1985–1998.
Dobrijevic E, Van Zwieten A, Kiryluk K, Grant AJ, Wong G, Teixeira-PInto A. Mendelian randomization for nephrologists. Kidney Int. 2023; 104(6): 1113–1123.
Boehm FJ, Zhou X. Statistical methods for Mendelian randomization in genome-wide association studies: A review. Comput Struct Biotechnol J. 2022; 20: 2338–2351.
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(5): 693–698.
Park E, Bathon J. Cardiovascular complications of rheumatoid arthritis. Curr Opin Rheumatol. 2024; 36(3): 209–216.
Tański W, Gać P, Chachaj A, Mazur G, Poręba R, Szuba A. Left ventricular myocardial strain assessed by cardiac magnetic resonance feature tracking in patients with rheumatoid arthritis. Insights Imaging. 2021; 12(1): 5.
Hanvivadhanakul P, Buakhamsri A. Disease activity is associated with LV dysfunction in rheumatoid arthritis patients without clinical cardiovascular disease. Adv Rheumatol. 2019; 59(1): 56.
Cioffi G, Viapiana O, Ognibeni F, Dalbeni A, Gatti D, Mazzone C, et al. Combined Circumferential and Longitudinal Left Ventricular Systolic Dysfunction in Patients with Rheumatoid Arthritis without Overt Cardiac Disease. J Am Soc Echocardiogr. 2016; 29(7): 689–698.
Ntusi NAB, Francis JM, Gumedze F, Karvounis H, Matthews PM, Wordsworth PB, et al. Cardiovascular magnetic resonance characterization of myocardial and vascular function in rheumatoid arthritis patients. Hellenic J Cardiol. 2019; 60(1): 28–35.
Midtbø H, Semb AG, Matre K, Kvien TK, Gerdts E. Disease activity is associated with reduced left ventricular systolic myocardial function in patients with rheumatoid arthritis. Ann Rheum Dis. 2017; 76(2): 371–376.
Zhang M, Wang M, Tai Y, Tao J, Zhou W, Han Y, et al. Triggers of Cardiovascular Diseases in Rheumatoid Arthritis. Curr Probl Cardiol. 2022; 47(6): 100853.
Reed E, Hedström AK, Hansson M, Mathsson-Alm L, Brynedal B, Saevarsdottir S, et al. Presence of autoantibodies in "seronegative" rheumatoid arthritis associates with classical risk factors and high disease activity. Arthritis Res Ther. 2020; 22(1): 170.
Federico JA, Syväranta SA, Tuohinen SS, Holmström MM, Peltomaa RL, Koivuniemi RP, et al. Myocardial T1 mapping by cardiac magnetic resonance imaging shows early myocardial changes in treatment-naive patients with active rheumatoid arthritis and positive autoantibodies. Clin Exp Rheumatol. 2024.
Corrao S, Messina S, Pistone G, Calvo L, Scaglione R, Licata G. Heart involvement in rheumatoid arthritis: systematic review and meta-analysis. Int J Cardiol. 2013; 167(5): 2031–2038.
Kobayashi Y, Kobayashi H, Hirano M, Giles JT. Left ventricular regional dysfunction using cardiac magnetic resonance imaging in rheumatoid arthritis patients without cardiac symptoms: comparison between methotrexate and biologics treatment groups. J Rheumatol. 2014; 41(7): 1560–1562.
Koivuniemi R, Kuuliala A, Kivistö S, Holmström M, Hämäläinen M, Moilanen E, et al. Induction of remission in female rheumatoid arthritis patients is associated with stabilization of myocardial abnormalities: a prospective cardiac magnetic resonance follow-up study. Scand J Rheumatol. 2021; 50(2): 104–112.
Lehmonen L, Vuorinen AM, Koivuniemi R, Leirisalo-Repo M, Holmström M, Kivistö S, et al. One-Year Follow-up Study Detects Myocardial Changes with Cardiovascular Magnetic Resonance Tagging in Active Rheumatoid Arthritis. Acad Radiol. 2018; 25(4): 476–485.
Yokoe I, Kobayashi H, Kobayashi Y, Nishiwaki A, Sugiyama K, Nagasawa Y, et al. Impact of biological treatment on left ventricular dysfunction determined by global circumferential, longitudinal and radial strain values using cardiac magnetic resonance imaging in patients with rheumatoid arthritis. Int J Rheum Dis. 2020; 23(10): 1363–1371.

No competing interests reported.

SupplementarymaterialTable.xlsx
Table S1 Seropositive rheumatoid arthritis instrumental variables Table S2 Seronegative rheumatoid arthritis instrumental variables Table S3 MR estimates for the causal effect of RA on CMR parameters

Download PDF

Editor invited by journal
05 Aug, 2024
Submission checks completed at journal
29 Jul, 2024
First submitted to journal
09 Jul, 2024

You are reading this latest preprint version

Effects of rheumatoid arthritis on ventricular structure and function: a Mendelian randomization study

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Data sources

Screening of instrumental variables

Two-sample MR

Sensitivity analyses

Multivariate MR

Statistical analyses

Results

Selection of instrumental variables

Causal relationship between seropositive RA and CMR measurements

Causal relationship between seronegative RA and CMR measurements

Multivariate Mendelian randomization

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1