Homocysteine as a Predictor of Paroxysmal Atrial Fibrillation-Related Events: A Scoping Review of the Literature

High levels of homocysteine (Hcy) have been linked with adverse cardiovascular outcomes, such as arrhythmias and stroke. In the context of paroxysmal atrial fibrillation (PAF), hyperhomocysteinemia has been demonstrated to be an independent predictor of future events. The aim of this report was to address the potential value of Hcy levels in predicting future paroxysms of atrial fibrillation (AF), as well as to identify the potential mechanisms of action. We searched PubMed and the Cochrane Database on 16 January 2022. Keywords used were homocysteine or hyperhomocysteinemia paired with a total of 67 different keywords or phrases that have been implicated with the pathogenesis of AF. We included primary reports of clinical and non-clinical data in the English language, as well as systematic reviews with or without meta-analyses. We placed no time constraints on our search strategy, which yielded 3748 results. Following title review, 3293 reports were excluded and 455 reports were used for title and abstract review, after which 109 reports were finally used for full-text review. Our review indicates that Hcy levels seem to hold a predictive value in PAF. Herein, potential mechanisms of action are presented and special considerations are made for clinically relevant diagnostic procedures that could complement plasma levels in the prediction of future PAF events. Finally, gaps of evidence are identified and considerations for future clinical trial design are presented.


Introduction
Atrial fibrillation (AF) represents the most common cardiac rhythm disorder, affecting millions of individuals worldwide, while its prevalence is expected to grow alarmingly within the next 50 years [1]. In the context of the substantial burden that AF poses on global health, on health economics, and on the affected individuals' health-related quality of life, scientists have been focusing on decoding the disease's exact pathophysiology; one cannot doubt the fact that AF is a result of multiple underlying factors, while it is usually precipitated by unknown triggers [2]. Research has shown that inflammation, in all its forms, is strongly linked with AF development; circulating inflammatory factors in the context of systemic inflammatory responses lead to atrial remodeling and fibrosis, a state that serves as a preamble for AF [3]. From this perspective, the identification of potential inflammation markers may be of paramount importance for the optimal diagnosis and management of the disease.
Homocysteine (Hcy) is a non-proteinogenic amino acid that is synthesized by methionine. Hcy serves as a precursor to many different amino acids in a series of biochemical Table 1. Reports indicating a positive correlation between levels of Hcy and future paroxysms of AF.

Number of Participants Results
Kubota et al., 2019 [11] 7133 patients from Atherosclerosis Risk in Communities (ARIC) Study and Multi-Ethnic Study of Atherosclerosis (MESA) • An age-, sex-, and race-adjusted model showed dose-response relations between plasma Hcy concentrations and AF incidence in both ARIC and MESA studies • A meta-analysis of both studies showed a significant association between homocysteine and AF • Hcy may be a novel risk marker for AF. Marcucci et al., 2004 [12] 310 NVAF patients on oral anticoagulant treatment (168 patients with previous ischemic events and 142 without) and 310 controls • Hyperhomocysteinemia was independently associated with NVAF after multivariate analysis • A significant correlation was found between Hcy levels and LA diameter • As shown by multivariate analysis, elevated Hcy levels were an independent risk factor for ischemic complications during NVAF. Nasso    Based on the above findings, we ought to perform a scoping review of available clinical and non-clinical data with the aim to identify reports that could shed light on the potential mechanisms responsible for this specific effect. Scoping reviews are designed to address broader questions as compared to systematic reviews. They still rely on the established systematic review process, however, whereas a systematic review would attempt to answer a specific question (i.e., can Hcy levels be used to predict future paroxysms of AF?), scoping reviews usually tackle multiple questions, in an attempt to provide an exhaustive look at the available evidence. The goal of our effort was to identify any correlation between Hcy and PAF, identify underlying mechanisms, assess the interplay with other biomarkers, identify gaps of evidence, and ultimately provide insight that could assist in the initial clinical assessment and in the development of the optimal therapeutic strategies for each patient with PAF.
We included primary reports of clinical and non-clinical data in the English language, as well as systematic reviews with or without meta-analyses. Narrative reviews, expert opinions and other types of medical correspondence were excluded. We placed no time constraints on our search strategy. Two independent reviewers participated in the selec-tion of included reports (PC and ET). If differences were reported in included reports, a resolution was achieved via discussion and refereeing by a third reviewer (CP). Data were extracted using a custom, previously tested form for non-clinical data, while a standard PICO form was used for clinical data.

Results
Our search strategy yielded 3748 results. Following title review, 3293 reports were excluded and 455 reports were used for title and abstract review, after which 109 reports were selected for full-text review. We finally included 74 reports in our scoping review ( Figure 1).

Non-Clinical Data
Our search strategy yielded 75 reports of in vitro and in animal models assessing the potential mechanisms of Hcy-induced direct myocardial toxicity. Table 2 provides a condensed overview of these reports. It should be noted that our search strategy did include neither the effects of Hcy on the vasculature nor data regarding its prothrombotic effects,

Oxidative Stress, Cardiac Fibrosis, and Remodeling
As the development of AF seems to be a result of the combination of a plethora of risk factors and comorbidities, oxidative stress has been associated with incident AF, while oxidative stress-induced atrial remodeling is considered the most common underlying mechanism; according to several reports, hyperhomocysteinemia has been linked with endothelial dysfunction, thus it represents an important factor for cardiovascular morbidity and mortality [93,94]. We were able to locate a total of 26 reports assessing oxidative stress as a potential mechanism of Hcy-induced cardiotoxicity [48][49][50]95]. Of these, 25 trials reported a positive correlation between Hcy and cardiomyocyte oxidative stress, while only one report failed to demonstrate oxidative stress as a cause of electron transport chain dysfunction, possibly due to an increased expression of the other protective mitochondrial proteins [41]. In these experiments, several different molecules were tested for their potential to inhibit the Hcy-induced oxidative stress, such as taurine and hydrogen sulfide. Taurine is an organic osmolyte involved in cell volume regulation, located in all organs throughout the body, which provides a substrate for the formation of bile salts; taurine seems to play a fundamental role in the modulation of intracellular free calcium concentration [96]. In parallel, hydrogen sulfide (H2S), the third discovered endogenous gas transmitter in mammals after NO and CO, participates in various pathophysiological processes; previous in vitro and in vivo research have revealed the protective role of H2S in the cardiovascular system that renders it useful in the protection of the myocardium against ischemia-reperfusion injury [97]. Interestingly, taurine [43,44] and hydrogen sulfide [45,46] were shown to be effective in counteracting a significant inhibiting effect against Hcy oxidative action. However, the clinical significance of that remains questionable, since none of the aforementioned agents used to hold a place in everyday clinical practice.
Furthermore, we were able to locate 16 trials reporting cardiac fibrosis as a result of Hcy-induced oxidative stress [41,[48][49][50][51][52][53][54][55][56][57][58][59][60][61][62]. Several mechanisms of action have been proposed for this effect, including increased reactive oxygen species, nitrotyrosine, matrix metalloproteinase, and decreased endothelial nitric oxide in response to antagonizing PPAR-gamma [82] and regulation of the Akt/FoxO3 pathway. More specifically, FoxO3 (forkhead box O3) is a member of the forkhead box transcription factors of the O class with a conserved helix-loop-helix DNA-binding domain; it is involved in a variety of vital cellular processes including oxidative stress, DNA repair, apoptosis, metabolism and cell cycle arrest [98,99]. Although the existing data are controversial, FoxO3 seem to play an important role in maintaining cardiovascular homeostasis.

Electrical Remodeling
Electrical remodeling, a variety of changes in the electrical substrate of the heart capable of triggering AF onset, belongs among the most common risk factors for AF development [100]. We were able to locate 11 trials assessing the effect of Hcy on the electrical conductivity of the heart and an additional 4 trials reporting on alterations of beta 2 adrenergic receptors. This was demonstrated as a significant prolongation of QRS, QTc, and PR intervals on the electrocardiogram (ECG) [71,72,81]. The above point toward a potential effect of Hcy in practically all distinct phases of the heart's systolic/diastolic cycle, demonstrating a clear arrhythmogenic potential associated with high levels of Hcy. Furthermore, Hcy has been reported to down-regulate the production of beta 2 adrenergic receptors [83,84], as well as to increase responsiveness to beta-adrenergic agonists [85]. Finally, alterations involving atrial calcium [74], sodium [77,78,80] and potassium [75,78,79,101] channels have also been demonstrated in hyperhomocysteinemia.

Clinical Data
The available clinical data are highly heterogeneous. Furthermore, a positive correlation between Hcy levels and PAF has already been demonstrated in recent meta-ana-lyses. We, therefore, decided not to proceed with a meta-analysis. Available clinical evidence is presented narratively and in tabular format (Table 3).   In multivariate analysis, homocysteine and long-chain acylcarnitines were two interacting determinants of NT-pro-BNP, in addition to LVEF, BMI, creatinine, and folate.  There was no relation between Hcy level and blood pressure control • There was a significant decrease in Hcy levels in the women treated in this study; however, this effect was absent in men.

Oxidative Stress, Fibrosis, Thrombosis, and Remodeling
In a prospective cohort of 643 patients, the relationship between plasma Hcy and glutathione peroxidase (GPx)-1 levels was addressed. GPx-1 has been demonstrated to modulate cardiovascular risk related to Hcy via its antioxidant properties. After a median follow-up of 7.1 years, the authors reported an inverse relationship between Hcy and GPx-1 levels and the occurrence of cardiovascular events. In fact, patients with low GPx-1 and high Hcy levels were 3.2 times more likely to experience a cardiovascular event. The authors went on to recommend that GPx-1 levels are taken into consideration when Hcy levels are used to predict future cardiovascular events [102].
Overall, high levels of Hcy have been implicated in the induction of oxidative stress in the vasculature. While our search was not designed to locate such trials, the capacity of Hcy to induce oxidative stress with subsequent clot detachment which ultimately leads to ischemic stroke is well described in the literature [103]. In fact, patients with elevated Hcy levels have been demonstrated to have a 2.73-fold increased probability of ischemic stroke [104].
In a sample of 2697 patients from the Framingham Heart Study, left ventricular (LV) mass, wall thickness, and relative wall thickness were correlated to high Hcy levels in women but not in men. However, plasma Hcy was not related to left atrial size or LV fractional shortening in either sex [105].
In a study of 66 patients, elevated Hcy levels were linked with a disproportional LV dilatation, where the ensuing hypertrophy was not sufficient to compensate for the increased wall stress. The authors proposed that a potential mechanism is the hyperhomocysteinemiaassociated increased oxidative stress which leads to the degradation of collagen, with consecutive fiber slippage and cardiac dilatation [106].

Electrical Remodeling
In a population-based study of 7002 participants, 12-lead ECGs were performed and correlated to plasma Hcy levels. The authors demonstrated an association between the prolonged QTc interval and high Hcy levels [107]. Furthermore, in a retrospective database study, 178 patients were stratified according to QRS interval duration and plasma Hcy levels. The authors reported a significant correlation between higher levels of Hcy and longer QRS intervals. Other ECG parameters, such as PQ interval, QTc interval, and QT dispersion were not found to be statistically correlated to Hcy levels [108].

Relation to Left Ventricular Ejection Fraction and Other Biomarkers
In a prospective case-control study of 515 coronary artery disease patients and 194 controls, higher Hcy levels were correlated with reduced left ventricular ejection fraction (LVEF) (<40%). Furthermore, elevated Hcy levels were associated with increases in N-terminal pro-B-type natriuretic peptide (NT-pro-BNP) levels [109]. In another study of 1020 patients, high levels of Hcy were demonstrated to predict increased NT-pro-BNP (or BNP) levels [110]. NT-proBNP and Hcy levels were measured in 31 patients with type 2 diabetes mellitus. The authors reported a correlation between elevated NT-proBNP and Hcy levels in patients with LV diastolic dysfunction [111]. A trial of 227 patients demonstrated that the MTHFR C677T mutation which is associated with plasma Hcy levels, is also linked with plasma BNP levels, leading to the conclusion that plasma Hcy levels are positively correlated with plasma BNP levels [112].
In a community-based trial of 1497 patients, serum Hcy levels were associated with a higher likelihood of detectable high sensitivity troponin T (hs-cTnT). The effect was stronger among the elderly patients in the cohort (>65 years old), while no association was recorded in patients <65 years old. The authors concluded that the levels of serum Hcy are associated with hs-cTnT levels in the elderly, which could indicate a possible relationship between Hcy and subclinical myocardial damage [113]. In another trial of 194 consecutive patients with acute myocardial infarction, elevated serum Hcy levels were positively correlated with serum cardiac troponin-I [114].
In a prospective cohort of 1224 consecutive patients with subclinical AF (SCAF), Hcy and uric acid (UA) were significantly elevated. The authors reported that an increase of 1 standard deviation (SD) in Hcy (5.7 µmol/L) levels was associated with an increased risk of SCAF in men and women regardless of their UA levels. Similarly, a 1-SD increase in UA (91 µmol/L) was associated with an increased risk of SCAF among the patients with high levels of Hcy in men (hazard ratio, 1.81; 95% CI, 1.43-2.30) and women (hazard ratio, 2.11; 95% CI, 1. 69-2.62). This led the authors to conclude that Hcy and UA are strongly associated with SCAF [115].

Adrenergic Effect
In a trial of 37 patients with essential hypertension (EH) compared with 37 healthy controls, blood levels of noradrenaline and adrenaline were demonstrated to be significantly higher in the EH group. Furthermore, the left ventricular mass index (LVMI) was also significantly higher in the EH group. The authors concluded that high levels of Hcy are associated with increased adrenergic activity in EH patients [116]. In a case control study of 273 patients with EH and 103 normotensive controls, the use of angiotensin-converting enzyme (ACE) inhibitors and beta-blockers significantly decreased and hydrochlorothiazide significantly increased plasma Hcy levels. The authors speculated that this reduction in Hcy levels was due to the improvement of endothelial function along with improved renal function [117]. In another trial of 120 patients with newly diagnosed hypertension, 100 mg of metoprolol per day was demonstrated to significantly reduce plasma Hcy levels. Additionally, there was no relation between homocysteine level and blood pressure control [118].

Discussion
AF represents the most common arrhythmia worldwide and is associated with significant morbidity and mortality [8]. The prevalence of AF in adults is about 2-4% and a 2,3-fold rise is expected in the years to come as a result of extended longevity in the general population and intensifying search for undiagnosed and untreated AF. A complex interplay of triggers and risk factors has been described widely in the literature, shedding light on the pathophysiologic substrate of the disease.
Arterial hypertension, diabetes mellitus, chronic kidney disease, heart failure, valvular heart disease, coronary artery disease, inflammatory diseases, obstructive sleep apnea, chronic obstructive pulmonary disease, obesity, alcohol consumption, and smoking are only some of the risk factors implicated in the pathophysiology of AF. Thus it is evident that due to the complexity of the pathophysiologic substrate of the disease and the wide array of AF triggers, one individual biomarker is difficult to show adequate sensitivity and specificity in the prognostication of AF recurrences.
But why do we need a reliable biomarker for AF prognosis? First, AF diagnosis may be challenging due to asymptomatic and paroxysmal presentation; second, biomarkers could refine screening procedures, and third, they could help predict the risk of recurrent AF. Which is the profile of an ideal biomarker? It would be easily accessible, cost-effective, and demonstrative of consistent accuracy and reproducibility. In the present manuscript, we provide sufficient evidence connecting high levels of Hcy with AF paroxysms, through clinical research data and 2 convincing meta-analyses. Additionally, Hcy is an easily accessible biomarker with good reproducibility. In a recent paper by Chua et al. [119], a series of 40 different plasma biomarkers were validated by using logistic regression models and machine learning technologies; in this specific study, Hcy was not validated, while BNP and FGF-23 proved to hold the strongest association with AF. Interestingly, in the present scoping review we demonstrate evidence supportive of the strong connection between Hcy and BNP or NT-pro-BNP levels, which reinforces our belief that Hcy may serve in the prognostication of AF.
Inflammation and evolving fibrosis are thought to be the main pathophysiologic mechanisms responsible for the development of AF. A wide array of biomarkers have been associated with this process and have been already validated in the literature, as previously described. Which one is the best to provide the most accurate prognostic information, is still unclear. In the present manuscript, we tried to explore in depth what Hcy levels have to offer in this direction. Serious clinical data and meta-analyses, together with evidence connecting Hcy with other inflammation markers and oxidative stress, may confer adequate information. Additionally, it is well established that standard predictors of AF, such as left atrial volume, left atrial mechanical function, and left ventricular ejection fraction as evaluated by echocardiography, are strongly associated with fibrosis. Furthermore, left atrial fibrosis demonstrated by cardiac magnetic resonance imaging (CMR) has been strongly associated with AF recurrence. Since Hcy stands as a marker of cardiac inflammation, it might reflect the aforementioned left heart geometrical changes. Indeed, we present data that connect Hcy levels with LV mass hypertrophy and LV dilation.
In our scoping review, we present data that implicate a direct adrenergic effect of Hcy in special populations, such as patients with arterial hypertension. This effect may cause left ventricular hypertrophy and promote left ventricular and left atrial remodeling, all of which are associated with AF development. By inhibiting sympathetic nerve drive by a b-blocker, one can speculate that this agent may inhibit the remodeling processes. B-blockers have been widely used for AF prevention and in the present manuscript, we provide evidence of Hcy levels reduction by metoprolol. Thus, we can propose that the beneficial effect of metoprolol in AF populations might be due to a bidirectional effect of b-blockade and Hcy levels reduction.
What about electrical remodeling and Hcy levels? ECG is the simplified mirror of the electrical circuits of the heart. Whether there is any convincing evidence that Hcy is associated with ECG changes that may serve as prognostic indicators of AF recurrences is still to be proven; there are reports pointing out that even in sinus rhythm, ECG might be a reliable biomarker of future AF events, since advanced interatrial block expressed by a p-wave duration > 120 ms and biphasic p-wave pattern in inferior leads, have been correlated with left atrial dilation [120,121]. Unfortunately, we were unable to find literature connecting Hcy levels with evidence of interatrial block or left atrial dilation. On the other hand, there is enough evidence of some association of Hcy levels with QRS duration with an undetermined effect on AF prevalence or recurrence.
Machine learning applications in clinical cardiology have rapidly evolved in recent years. By using machine learning tools together with vast data sources, the management of a variety of chronic cardiac diseases including AF, is expected to change in the near future. The role of Hcy levels in this background remains unclear; large-scale trials are needed in order to establish which plasma AF biomarker is the most accurate to be incorporated in machine learning applications [122].

Future Research
Our systematic search of the literature shed light on a wide array of existing data concerning the potential role of Hcy in the modification of risk factors for AF development.
More specifically, our search demonstrated a direct effect of high Hcy levels on the adrenergic system in various laboratory models, while we highlighted the rather limited clinical evidence which presented a potential Hcy-lowering effect of beta-blockers. This potential for a dual benefit warrants further investigation in the clinical setting.
Furthermore, we were able to identify several different biomarkers that could potentially be part of a predictive algorithm. Glutathione peroxidase is an endogenous antioxidant that may be inversely correlated to Hcy levels, while NT-pro-BNP and troponin levels also seem to hold a significant role. As NT-pro-BNP levels are mostly used in the diagnosis and monitoring of heart failure, this prohormone has been used in other disease states, such as myocardial ischemia [123]. Finally, UA has also been implicated as a predictor of future PAF events. While there does not seem to be one single biomarker that has been proven to consistently and accurately predict PAF, it seems very likely that a combination of these biomarkers may provide a valuable algorithm in that direction. Future research should be focused on combining the predictive value of these biomarkers to generate a reliable stratification algorithm.

Conclusions
In the context of PAF, there is sufficient evidence to support the use of Hcy levels as an independent predictor of future events. However, the pathogenesis of cardiac arrhythmias involves complex biological mechanisms that cannot be explained by or attributed to a single biomarker. Clinicians should take several other factors into consideration when assessing the likelihood of AF. In the near future, artificial intelligence with machine learning modalities could allow better prognostication and improved therapeutic management of the disease.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/diagnostics12092192/s1, Table S1: Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) Checklist. Reference [124] is cited in Supplementary Materials. Author Contributions: P.C. and C.E.P. conceptualized the central idea of the scoping review. E.T. and A.B. performed the literature search and, along with P.R. and K.T., they interpreted the data and drafted the manuscript. E.T., A.B. and V.P. substantially revised the manuscript. T.K., C.E.P. and V.V. supervised the project and approved the final version of the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.