Skip to main content
NCBI home page
Europace logoLink to Europace
. 2025 May 13;27(6):euaf098. doi: 10.1093/europace/euaf098

From moderate to strenuous training: unravelling mechanistic contributors and biomarkers for atrial fibrillation in exercise

Anna Alcarraz 1,2, Aline Meza-Ramos 3,4,5, Cira Rubies 6,7, Maria Sanz-de la Garza 8,9, Carlos Eduardo Bolaños-Gomez 10, Marta Sitges 11,12,13,14, Lluis Mont 15,16,17,18, Montserrat Batlle 19,20,2, Eduard Guasch 21,22,23,24,✉,3,2
PMCID: PMC12202093  PMID: 40358233

Abstract

Aims

The impact of the transition from moderate to strenuous exercise on atrial fibrillation (AF) risk and its underlying mechanisms remain poorly understood. We aimed to analyse biatrial remodelling after moderate and strenuous exercise, compare it with pathological atrial damage, and non-invasively identify strenuous exercise insults.

Methods and results

Young male Wistar rats were trained at a moderate (MOD) or high-intensity (INT) load; sedentary rats served as controls. After 16 weeks, electrophysiological and echocardiographic studies were obtained, and atrial samples were used for fibrosis quantification. Plasmatic biomarkers (at rest and after exercise) and atrial gene expression (mRNA array) were assessed. Results were compared with a transverse aortic constriction (TAC) model.

Results

AF inducibility progressively increased with exercise load. Both trained groups presented bradycardia, an enhanced parasympathetic tone and biatrial dilatation. INT rats exhibited prolonged P-waves and greater fibrosis in the left (LA) and right atria (RA). The proarrhythmogenic remodelling substantially differed in both atria. Compared with MOD, inflammatory pathways were enriched in the RA of INT, similar to the TAC model. Only minor changes were observed after exercise in the LA. Plasma biomarkers showed unremarkable changes between groups at rest, but intensive exercise led to a transient increase in proinflammatory markers.

Conclusion

Exercise-induced-AF pathology is load-dependent: parasympathetic tone augmentation and atrial dilatation drive AF risk in moderately trained rats, whereas a further increase is associated with atrial fibrosis. Transient inflammation, identifiable through plasma biomarkers, could underpin AF susceptibility and fibrosis in the RA of INT rats, and serve as biomarkers.

Keywords: Atrial Fibrillation, Exercise load, Animal model, Inflammation, Fibrosis, Plasma biomarker, Transverse aortic constriction model

Graphical Abstract

Graphical Abstract.

Graphical Abstract

Open in a new tab

Translational Perspective.

The precise relationship between the exercise load and atrial fibrillation (AF) risk, as well as the underlying mechanisms, remain unclear. Our findings indicate a progressive increase in the risk of AF with increasing exercise load (intensity and duration), which also determines a two-stage remodelling process. The enhancement of parasympathetic tone and atrial dilation, which can be evidenced by an echocardiography and an electrocardiogram (ECG), substantiate AF risk with moderate exercise. Atrial fibrosis is a consequence of strenuous exercise and, at least partially, exercise-induced local right atrial inflammation, and it further increases the risk of AF. A distinctive transient proinflammatory biomarker fingerprint may prove useful in identifying athletes at risk.

Introduction

Although atrial fibrillation (AF) is commonly associated with cardiovascular diseases and risk factors, strenuous exercise has attracted attention as an unexpected risk factor for AF.1 To date, clinical and experimental data support a J-shaped relationship between exercise load and risk of AF, which is likely modulated by age and comorbidities.2 Under this assumption, the risk of AF results from the balance between the improved cardiovascular risk profile promoted by exercise and its potential direct proarrhythmogenic effect.2

The mechanisms by which regular high-intensity exercise promotes AF are not yet completely understood. Some of the physiological hallmarks of the athlete’s heart, including left atrial (LA) dilatation and parasympathetic tone enhancement,3 are well known to increase the risk of AF.4 Preclinical models of regular high-intensity exercise, supported by recent clinical data,5 have suggested that atrial fibrosis6 and myocardial inflammation7,8 also contribute to the atrial cardiomyopathy9 (ACM) in athletes, but their precise role remains unclear. Moreover, the differences between the structural and functional atrial changes resulting from moderate and strenuous exercise have not been previously assessed and compared. Further characterization of the ACM features underlying exercise-induced AF, as well as the mechanisms involved in the transition between the beneficial effects of moderate exercise and the proarrhythmogenic effects of strenuous training, may help in elucidating the pathophysiology of exercise-induced AF and the development of preventive and early diagnostic tools for exercise-induced AF in athletes.

Our objective in this translational project was to characterize the mechanisms underlying the differential effects of increasing loads of exercise on the atria and the risk of AF. This aim was pursued by examining the differences in both atria that occur after moderate and strenuous exercise, exploring the mechanisms of the proarrhythmogenic remodelling that is promoted by strenuous exercise, and comparing it with a model of established pathological atrial remodelling. Finally, we aimed to identify candidate plasma markers of such deleterious atrial remodelling.

Methods

Two animal models were employed to evaluate exercise-induced atrial remodelling and to contrast it with well-established atrial pathological remodelling. First, male Wistar rats (200–250 g, Charles River Laboratories, France) were randomly assigned to three groups: sedentary (SED, n = 30) that served as control, moderate exercise (MOD, n = 30; 35 cm/s for 45 min, recapitulating an active lifestyle) or high-intensity exercise (INT, n = 27; 60 cm/s for 60 min, simulating heavily trained athletes). Exercise intensity was set based on previous work6 and data on maximum oxygen consumption in rats.10 Exercise groups were conditioned to run on a treadmill for 5 days/week for 16 weeks. Second, a thoracic aortic constriction (TAC) model was used to assess whether the nature and mechanisms of exercise-induced atrial deleterious remodelling, were comparable to those promoted in pathological atrial remodelling, where male Wistar rats (300–350 g, Charles River Laboratories) were subjected to TAC (n = 15) or a Sham procedure (n = 6).

A detailed description of the methodology, including in vivo (animal models, electrocardiography, echocardiography, and electrophysiological studies), histology and in vitro (mRNA array, plasma analysis) experiments, as well as statistical analyses, is provided in the Supplementary material.

The results of the mRNA array have been deposited in the Gene Expression Omnibus (GEO) repository (GSE289211, GSE297444), and the ECGs and histological images have been uploaded to Zenodo.11

Results

After the 16-week training protocol, weight was significantly lower in both trained groups than in the SED group (see Supplementary material online, Figure S1A). Raw heart weight did not differ between groups. However, indexed cardiac mass was increased in the exercised rats, with significant differences between the two training loads, as the heart weight of the INT rats was greater than that of the MOD rats (see Supplementary material online, Figure S1B-C). Exercise did not affect left ventricle (LV) fibrosis (see Supplementary material online, Figure S1D).

Atrial fibrillation inducibility progressively increases with exercise intensity

No rats presented with spontaneous AF. We therefore tested the vulnerability to AF during an electrophysiological study in all groups. A representative recording of the inducibility test in every group is shown in Figure 1A, and the outcomes of the inducibility protocol are plotted in Figure 1B. Atrial fibrillation inducibility progressively increased with the intensity of regular exercise, from <5% in the SED group to >40% in the INT group (P = 0.009, Cohran–Armitage test for trend). Notably, compared with that in the SED rats, the duration of AF was significantly prolonged in the INT rats only, indicating a more severely affected substrate (Figure 1C). We also quantified the Wenckebach cycle length (WCL), the atrial effective refractory period (ARP), and the corrected sinus recovery time (cSNRT) in all groups. The WCL was similarly prolonged in the MOD and INT groups compared with that in the SED group (Figure 1D), whereas the ARP was shorter in both trained groups than in the SED group (Figure 1E). We did not find significant differences between the groups in terms of sinus node recovery time (SNRT) or cSNRT (Figure 1F).

Figure 1.

Figure 1

Open in a new tab

Results of the electrophysiological study. (A) Recordings of burst pacing atrial fibrillation (AF) induction in SED, moderately trained (MOD), and intensively trained (INT) rats (in each group: ECG in the superior panel, intracavitary electrogram in the middle, and stimulus artefact in the lower panel). Cycle length for burst pacing was 40 ms in SED and INT, and 60 ms in MOD. (B) AF inducibility rate in the SED (n = 21), MOD (n = 20), and INT (n = 18) groups as a percentage of rats >1 s inducible AF (P for trend = 0.009). (C) AF duration (boxplot) of the irregular atrial electrograms in seconds. (D–F) Measurements (boxplot) of the electrophysiological parameters WCL (D; n = 20 SED, n = 19 MOD, n = 17 INT), ARP (E; n = 19 SED, n = 19 MOD, n = 16 INT), and cSNRT (F; n = 20 SED, n = 18 MOD, n = 17 INT). Most analyses were performed with one-way ANOVA, which was significant at P < 0.05. Post hoc comparisons were performed with the least significant difference (LSD) test. Induced-AF duration was analysed with a Kruskal–Wallis test and post hoc Dunn test. *P < 0.05, **P < 0.01, and ***P < 0.001. AF, atrial fibrillation; ARP, atrial effective refractory period; cSNRT, corrected sinus recovery time; ECG, electrocardiogram; INT, high-intensity; MOD, moderate; SED, sedentary groups; WCL, Wenckebach cycle length.

Electrocardiography reveals exercise- and load-dependent differences

Representative electrocardiogram (ECG) recordings and their corresponding measurements are shown in Figure 2. Compared with the SED group, both trained groups presented significant bradycardia (Figure 2B). Compared with the SED and MOD groups, only the INT group presented a prolonged P-wave duration (Figure 2C). The PR interval, QRS wave and QTc interval durations were unaltered due to exercise (Figure 2D–F).

Figure 2.

Figure 2

Open in a new tab

Results of the electrocardiographic analysis. (A) Representative averaged ECG recordings (thin green tracings represent 200 consecutive beats, the thick black trace is the averaged ECG recording) of a rat in the SED, MOD, and INT exercise groups. (B–F, H) Results (boxplot) and comparison of the RR interval (B), P duration (C), PR interval (D), QRS (E), and QTc duration (F). For B to F: n = 18 SED, n = 20 MOD, n = 18 INT. Analyses with a one-way ANOVA, which was significant at the P < 0.05 level for the RR interval, P duration, and LF/HF ratio. Post hoc comparisons with group levels (LSD test) are shown. (G) Representative images of the LF/HF peaks for a rat in each group and LF/HF ratio measurements (H). n = 16 SED, n = 17 MOD, n = 17 INT. *P < 0.05, **P < 0.01. ECG, electrocardiogram; HF, high frequency; INT, high-intensity; LF, low frequency; MOD, moderate; SED, sedentary groups.

Significant bradycardia along with prolonged WCL and shortened ARP suggested an enhanced parasympathetic tone in the MOD and INT groups. We next evaluated heart rate variability through the low-to-high frequency (LF/HF) ratio to estimate the drive of the autonomic tone balance on the heart (representative spectral plots in Figure 2G). Consistently, we found that the MOD and INT rats had a lower LF/HF ratio than the SED rats did (Figure 2H), indicating increased parasympathetic tone as a result of exercise, which was similar in both trained groups. After the administration of carbachol (50 µg/kg), the magnitude of changes in the PR interval and the RR cycle was greater in trained rats than in SED rats (see Supplementary material online, Table S3), further supporting enhanced vagal sensitivity as a result from chronic exercise.

Atrial dilatation and fibrosis differ according to exercise intensity

Atrial structural remodelling was assessed both macroscopically through echocardiography recordings and microscopically in histological specimens. The findings are summarized in Figure 3. In echocardiographic studies, we measured the area of both atria and the anteroposterior diameter of the LA (Figure 3A). Our results revealed a greater LA dimension at end atrial diastole (LAD) and larger LA and right atria (RA) surface areas in the MOD and INT groups compared with the SED group (Figure 3B and C), indicating that exercise caused similar atrial dilation regardless of intensity.

Figure 3.

Figure 3

Open in a new tab

Structural remodelling assessment in all groups. (A) Measurement of echocardiographic images of the left atrial diameter (LAD, longitudinal parasternal view) and the right (RA) and left atrial (LA) areas (apical 4-chamber view). RV, right ventricle; LV, left ventricle; Ao, aorta. (B, C) Atrial echocardiographic measurements (boxplot). Indexed LAD (B) was compared with one-way ANOVA, which was significant at the P < 0.001 level, and post hoc comparisons between group levels [least significant difference test (LSD)] were performed. N = 15, 20 and 15 for SED, MOD, and INT, respectively. Indexed left (n = 10, 13, 15 for SED, MOD, INT) and RA (n = 14, 16, 17 for SED, MOD, INT) areas (C) were compared with linear mixed-effects modelling, including group, atrium and their interaction as dependent variables; the group and atrial factors were significant at the P < 0.05 and P < 0.001 levels, respectively. Post hoc comparisons between group levels (LSD) are shown. (D) Representative microphotographs of atrial specimens from all groups stained with Sirius red. The scale corresponds to 100 µm. (E) Fibrosis quantification (mean ± SEM) in the LA and RA (n = 9, 8, 8 for SED, MOD, INT). Analyses were performed with linear mixed-effects modelling, including group, atria and their interaction as dependent variables; the group was significant at the P < 0.05 level. Post hoc comparisons between group levels (LSD) are shown. *P < 0.05, **P < 0.01, ****P < 0.0001. INT, high-intensity; MOD, moderate; SED, sedentary groups.

We next assessed collagen deposits in histological atrial sections (Figure 3D) and found that atrial myocardial fibrosis was significantly greater in the atria of the intensively trained rats compared with those of the moderately trained and SED rats. There were no differences between the MOD and SED groups (Figure 3E). Results were consistent in the RA and LA, and there was a moderate correlation between atria in all groups (see Supplementary material online, Figure S2). These results, which are consistent with a prolonged P wave in the ECG, indicate that only strenuous exercise, and not moderate training, promotes atrial fibrosis.

Effect of exercise intensity on atrial molecular remodelling

To determine the causes of the deleterious atrial remodelling in the INT group, which was characterized by increased fibrosis, we performed microarray expression profiling in each atrial chamber of the three groups.

Atrial laterality is a major determinant of gene expression

We initially performed unsupervised clustering analysis [principal component analysis (PCA)], which demonstrated that atrial laterality drove most of the variability in atrial gene expression (see Supplementary material online, Figure S3A). A total of 5726 genes were differentially expressed between the right and left atria [1451 with an absolute fold change (|FC|)>1.5]. We identified several genes as markers of laterality [volcano plot and the top 10 differentially expressed genes (DEG) are shown in Supplementary material online, Figures S3B and C]. Bone morphogenetic protein 10 (Bmp10) expression was almost exclusive to the RA (∼9000-fold higher in the RA than in the LA). The expression of Kv channel-interacting protein 2 (Kchip2 and Kcnip2) was ∼30-fold greater in the RA. Paired-like homeodomain 2 (Pitx2), chemokine ligand 21 (Ccl21), and ADAM metallopeptidase with thrombospondin type 1 motif 8 (Adamts8) were distinctive markers of the LA. The most frequently overrepresented ontologies were biological processes (see Supplementary material online, Figure S3D), with angiogenesis being the most significant. The top overrepresented pathways identified in WikiPathways are shown in Supplementary material online, Figure S3E and include cellular pathways such as senescence and autophagy, calcium regulation in cardiac cells and focal adhesions, but also those related to inflammation and oxidative stress. Importantly, exercise load did not have a major effect on gene expression laterality. Only 66 genes shared a significant interaction between atrial side and group in both INT and SED, and INT and MOD comparisons, with 52 of these genes exhibiting an opposite direction of the effect between atrial side and INT vs MOD-SED (see [Supplementary material online, Table S4 and Supplementary material online, Figure S3F]). Hence, to determine the transcriptomic differences between exercise groups, we analysed the effects of different exercise loads separately in each atrium.

Atrial gene expression changes in response to regular exercise

We subsequently aimed to identify the mechanisms underlying profibrotic exercise in each atrium. We compared the remodelling induced by moderate (i.e. non-fibrotic atrial substrate) and intense (increasing atrial fibrosis) exercise. In the LA, 419 genes were differentially expressed (|FC|>1.5) between the two exercise intensities (Figure 4A, complete list in Supplementary material online, Table S5). Two pathways involved in complement activation were significantly enriched (Figure 4B). In the RA, 384 genes were deregulated (Figure 4C, complete list in Supplementary material online, Table S6). Pathway enrichment analyses revealed a predominance of inflammation-related pathways among the DEGs in the MOD vs. INT comparison, including cytokines and inflammation, the transforming growth factor beta (TGF-β) signalling pathway, the IL-1 signalling pathway, and senescence and autophagy, among others (Figure 4D). The most significant pathway, cytokines and inflammatory response, is represented in Figure 4E (DEGs are highlighted) and suggests a central role of macrophage-related inflammation.

Figure 4.

Figure 4

Open in a new tab

Transcriptomic analysis. (A, C) Volcano plot of the moderate (MOD) vs. intensive (INT) trained group comparison in the left atrium (LA; MOD n = 8, INT n = 8) (A) and right atrium (RA; MOD n = 7, INT n = 8) (C). In each atrium, dark blue points (right side of Volcano plots) represent genes significantly up-regulated (fold-change > 1.5) in the INT group compared with MOD, light blue points (left side of Volcano plots) represent genes significantly up-regulated (fold-change > 1.5) in the MOD group compared with INT, and grey points are genes not differentially expressed between MOD or INT or with a fold-change <1.5. (B, D) Overrepresentation analysis of the differentially expressed genes (DEGs) in WikiPathways in the LA (B) and RA (D) of the MOD vs. INT comparison. (E) Genes deregulated in the cytokines and inflammatory response according to the WikiPathway analysis.

Overall, these data indicated differential remodelling induced by strenuous exercise in the LAs and RAs, with a greater impact on the latter and mainly involving inflammatory-related pathways.

Local inflammation assessment

Because of the suspected potential involvement of local inflammation in the pathology of exercise-induced fibrosis, we next quantified macrophage infiltration in the atrial myocardium of all groups (representative microphotographs in Figure 5A). Compared with those in the SED and MOD rats, the number of CD68-positive cells, primarily macrophages, was significantly greater only in the RA of the INT rats (Figure 5B). CD68+ infiltration was also correlated with the myocardial fibrotic burden in the RA [slope 0.35 (95% CI 0.17, 0.52)] but not in the LA [slope 0.100 (95% CI −0.009, 0.21)] (Figure 5C).

Figure 5.

Figure 5

Open in a new tab

Assessment of local inflammation. (A) Representative images of immunochemical staining of CD68-positive cells (brown) in the left (LA) and right (RA) atria. The scale corresponds to 50 µm. (B) Number of CD68-positive cells per square millimetres in the LA and RA (mean ± SEM). N = 9, 7, 8 for SED, MOD, INT, respectively. Two-way ANOVA was performed, with group, atrium and their interaction as dependent variables. The group × atrium interaction was significant at the P < 0.01 level; within each atrium, significant post hoc pairwise comparisons between groups [least significant difference test (LSD)] are shown. (C) Correlation between the degree of atrial fibrosis and the presence of CD68-positive inflammatory cells. Analysis with a linear mixed model including CD68 infiltration, atria, and their interaction as predictors in a model with a repeated measurements covariate structure. Atria × CD68 infiltration was significant at the P < 0.05 level; the slope at every atrium was subsequently tested and is shown. The CD68 infiltration dependent variable was log-transformed to fulfil the normality of the residual requirement. *P < 0.05, **P < 0.01, ****P < 0.0001. INT, high-intensity; MOD, moderate; SED, sedentary groups.

Comparison to pathological atrial remodelling

Our results demonstrated that intensive training leads to atrial myocardial fibrosis, similar to the pathological remodelling observed in hypertension, heart failure, and other cardiovascular diseases.12 To investigate the potential shared pathways regulating the deleterious atrial remodelling in trained rats and rats with cardiac structural diseases, we performed an mRNA array in the atria of a pathological rat model (TAC) presenting with balanced hypertrophy or overt heart failure (see Supplementary material online, Figure S4A–C).

Transverse aortic constriction induced significant changes (|FC|>1.5) in 1288 transcripts in the LA and 1700 transcripts in the RA. None of the enriched pathways in the LA of TAC (see Supplementary material online, Table S7, the top 15 pathways are shown in Supplementary material online, Figure S4D) coincided with those pathways significantly involved in strenuous exercise-induced LA remodelling (Figure 4B). Pathway overrepresentation analysis in the RA revealed that focal adhesion and adipogenesis presented the lowest P-values (all significant pathways are shown in Supplementary material online, Table S8, the top 15 are shown in Supplementary material online, Figure S4E). All but five of the pathways overrepresented in the LA were also flagged in the RA, demonstrating comparable TAC-induced remodelling in both atria. Interestingly, six pathways were also involved in the remodelling induced by strenuous exercise in RA, including the TGF-β signalling pathway, spinal cord injury, and senescence and autophagy (highlighted in yellow in Supplementary material online, Table S8).

Identification of plasma biomarkers

To describe the plasma protein fingerprint and identify potential biomarkers of deleterious exercise-induced atrial remodelling, we conducted high-throughput proteomic analyses of plasma samples. In an Olink® panel quantifying cytokines, chemokines, and growth factors, we aimed to identify biomarkers that were differentially expressed in the plasma of the INT rats compared with those of both the SED and MOD rats. The results at rest are presented in Supplementary material online, Figure S5 and Supplementary material online, Table S9 and revealed that only six proteins were differentially expressed between the groups. Notably, in most cases, the results were similar between the exercised groups (MOD and INT) and the SED group, ruling out a potential biomarker to identify excessive exercise. Only Fli1, whose deficit promotes cardiac fibrosis,13 was increased in the MOD group compared with the SED group and evolved as a potential contributor to atrial fibrosis, but the results did not significantly differ from those of the INT group. Notably, supervised clustering algorithms demonstrated that both exercised groups presented similar profiles (see Supplementary material online, Figure S5B), and a loading plot did not flag Fli1 as a potential biomarker (see Supplementary material online, Figure S5C).

Given the absence of candidate biomarkers at rest and because of the suggested transient exercise-induced damage, we subsequently analysed plasma at rest and after an exercise bout of moderate or strenuous intensity. A diverging change in plasma biomarker concentrations between both groups was observed for 28 plasma proteins (denoted by a significant interaction in Figure 6A; main results in Figure 6B and all of them in Supplementary material online, Table S10). Differences in inflammatory-related biomarkers were particularly relevant: chronic intensive training was associated with an anti-inflammatory effect at rest but with a greater increase in proinflammatory marker levels after exercise. For example, IL-1α levels were comparable at rest between the MOD and INT rats but were greater in the INT rats after exercise. IL-6 levels were lower in the INT group than in the MOD group at rest but dramatically increased after exercise only in the INT group; the magnitude of the difference was 2.01 (95% CI 1.57–2.45) relative units. Similar patterns were found for TGF-α and platelet derived growth factor-beta (PDGFB). In general, the levels of immune (e.g. CCL2, CCL3, CXCL1, and CXCL9) and fibrotic regulators [e.g. vascular endothelial growth factor-D (VEGFD) and hepatocyte growth factor (HGF)] were lower at rest in the INT group than in the MOD group but increased and reached similar values in both groups after exercise (Figure 6B), suggesting a greater exercise-related increase as a consequence of INT exercise.

Figure 6.

Figure 6

Open in a new tab

Plasmatic biomarker profile in trained groups after exercise. (A) Plasma proteins showing a significant group × timepoint interaction [e.g. different behaviours after MOD and intensive (INT) exercise] are coloured vs. the –log(P-value); dot size is inverse to P-value. (B) Cytokine, chemokine, and growth factor boxplots of altered proteins analysed in plasma. The results were analysed with a linear mixed model with a repeated measures covariance structure including group (MOD and INT) and timepoint (at rest, post-exercise), and their interaction as independent variables; all interactions were significant (see panel A). Post hoc comparisons are shown. N = 7 for MOD, n = 4 for INT at both timepoints. (C) Plasmatic biomarkers significantly increased after exercise and in the pathological pressure overload model (boxplots). For the pathological model (left panels), the results were analysed with a t-test. *P < 0.05, **P < 0.01, ****P < 0.0001 basal vs. post-exercise within each group; #P < 0.05, ##P < 0.01, ##P < 0.001 MOD vs. INT within time-point. INT, high-intensity; MOD, moderate; SED, sedentary groups.

The atrial expression of four central genes (Cxcl2, Il1b, Il6, and Tnf) was quantified in these rats and results correlated with protein plasma quantification (see Supplementary material online, Figure S6). In general, there was a positive trend between the increase in plasma biomarkers and proinflammatory genes in the RA, but a negative relationship in the LA, yet most did not reach statistical significance.

Finally, we compared the plasma protein profile of the strenuous exercise model with that of the TAC-induced model. Three proteins (TNFRSF11B, TNNI3, and VSIG2) were significantly increased in both the acute strenuous vs. moderate exercise, and sham vs. TAC comparisons (Figure 6C).

Discussion

In the present study, we aimed to evaluate atrial remodelling at increasing doses of exercise to better understand the pathophysiology of the J-shaped relationship between exercise load and AF risk. Our results show that (i) AF vulnerability progressively increases from sedentary to moderately exercised to heavily exercised rats; (ii) both exercise loads similarly increase parasympathetic tone and induce atrial dilatation; (iii) strenuous exercise promotes deleterious atrial remodelling characterized by atrial myocardial fibrosis; (iv) atrial remodelling occurs in an atrial side-dependent manner, with changes in the RA (but not the LA) resembling those of TAC, including local inflammation; and (5) biomarker quantification at rest yields minimal discriminatory capacity between exercise loads, but acute strenuous exercise is associated with a transient, distinctive fingerprint of plasmatic biomarkers that could be relevant for identifying excessive exercise.

Moderate- and high-intensity exercise increases the risk of atrial fibrillation through different mechanisms

Aging and the presence of cardiovascular risk factors have been reported to influence the relationship between physical activity load and risk of AF.2 In middle-aged individuals, who commonly present one or more cardiovascular risk factors, the relationship between exercise load and risk of AF follows a U-shaped curve.14 Conversely, such a relationship is not evident at the extremes: exercise seems to convey minimal, yet not zero,15 risk of AF in aged individuals16 and those with risk factors (i.e. a reversed-J-shaped curve),17 whereas in young men with almost no comorbidities, even moderate loads of exercise are associated with an increased risk of AF (a J-shaped curve).18 Our results in a young, healthy animal model are in agreement with the later and previous preclinical models,19 and provide their pathophysiological basis. Atrial fibrillation inducibility progressively increased with increasing exercise load from sedentary to moderate to strenuous, supporting results from Gorman et al.19 in swimmer mice. Further experiments presented here suggest that components of the ACM differ in both groups driven by a two-hit mechanism. First, moderate intensity exercise leads to increased parasympathetic tone, which primarily contributes to the increased AF inducibility. In our animal model, a shortening of the ARP, a prolongation of the WCL, a decrease in LF/HF and augmented carbachol sensitivity all pointed to augmented parasympathetic tone in both trained groups, consistent with previous reports comprehensively assessing the parasympathetic tone in the same strenuously trained model.6 Enhanced parasympathetic tone in regular exercisers enables an acute, rapid increase in cardiac output during exercise bouts. At rest, this process leads to bradycardia and a spaciously heterogeneous shortening of the ARP,20 thereby creating a dynamic arrhythmogenic substrate for AF.21 Second, high-intensity exercise still presents with parasympathetic enhancement and introduces atrial fibrosis as an additional proarrhythmogenic hit, creating a fixed substrate for re-entry. It is therefore possible that, depending on its load, exercise-induced AF may encompass two different pathophysiological forms of AF: a parasympathetic tone driven at low to moderate loads and an atrial fibrosis dominant at high or very high doses. In humans, parasympathetic tone enhancement and atrial dilation are well known components of the athlete’s heart, and recent data confirmed the increased fibrotic burden in heavily trained individuals in the absence of AF;5 however, the relative weight of each of these contributors and potential forms of exercise-induced AF in clinical practice still need to be elucidated. It is important to note, however, that only some athletes will develop AF, suggesting variability in individual predisposition. Latent cardiomyopathies or polygenic risk, as recently suggested for the apparently physiological reduction in ejection fraction in athletes,22 may contribute to this variability.

The impact of a two-hit substrate for AF on long-term prognosis remains to be elucidated. It is unclear whether purely parasympathetic tone-driven AF in moderately trained rats results in the same long-term prognosis as those dependent on strenuous exercise-induced AF and atrial fibrosis. In this context, recent data have demonstrated that AF at a young age, which is common in athletes, conveys a similar or poorer prognosis than that at older ages, even in the absence of relevant comorbidities.23 Moreover, data on the effects of deconditioning6 demonstrated that vagal tone is reduced after 4 weeks, yet atrial fibrosis persists for at least 8 weeks following cessation of regular exercise. This finding indicates that the arrhythmogenic substrate is partially reversible, particularly at early stages or after moderate exercise, although it cannot be entirely regressed after a certain time or intensity threshold has been surpassed.

The mechanisms behind atrial cardiomyopathy in athletes differ from those promoted by cardiac diseases

The finding of atrial fibrosis in strenuous exercise animal models6 and athletes5 has prompted the question of whether myocardial fibrosis nature and mechanisms are analogous to those observed in other causes of AF. Our results showed that, in general, TAC induced deeper and more severe remodelling than strenuous exercise did. The number of DEGs was four-fold greater in the TAC group than in the control group, and the magnitude of their deregulation was greater than that in the intense exercise group. Recent findings may help explain this discrepancy.24 The continuous, intense injury caused by continuous haemodynamic, proinflammatory overload in TAC results in more severe and persistent remodelling than does the discontinuous effect of exercise. The transient nature of strenuous exercise insults may inflict microinjuries that may partially resolve after every training bout. Notably, analyses in our trained rats were performed at least 24 h after the last training session, quite likely at a reparative stage, thereby explaining the relatively low number of proinflammatory and pro-fibrotic affected genes. Indeed, previous reports had already identified an LA anti-fibrotic and anti-inflammatory gene expression fingerprint in the chronic setting, contrasting with an up-regulation of tumour necrosis factor (TNF) and TNF-regulated genes early after exercise.24 Our study of plasma biomarkers distinguished between acute and chronic profiles: at rest, minimal differences were observed between groups. However, a pronounced increase in inflammatory biomarkers was detected following each exercise session in the intensive group.

Overall, these findings are consistent with those of previous human studies demonstrating transient, load-dependent atrial dysfunction25 after exercise. Interestingly, a repetitive injury-and-repair process is also known to occur in the right ventricle (RV) depending on the exercise duration and intensity,26,27 eventually resulting in the development of a load-dependent pro-arrhythmogenic substrate.28

Excessive haemodynamic overload and subsequent local inflammation in the right atria might contribute to a different left–right remodelling after strenuous exercise

The RA and LA exhibit remarkable structural, functional, haemodynamic, and molecular differences.29,30 Consistent with these findings, we confirmed that gene expression significantly differed between the two atria, underscoring the different environment and warranting separate analysis of each atrium. Further delving into baseline differences, certain conditions superimpose a more severe insult on either the RA or the LA. For example, pulmonary hypertension selectively increases right atrial pressure, and hypertension primarily impinges the LA pressure.

Despite these differences, in this and previous work6 we reported that strenuous exercise promoted a similar degree of atrial fibrosis in both atria. To our knowledge, this is the first study assessing the effects of exercise in gene expression of the RA and LA separately. This enabled us to determine evident laterality differences in the molecular remodelling underpinning exercise-induced atrial fibrosis. On the one hand, remodelling in the LA largely differs between INT and TAC. On the other hand, both models share common molecular changes in the RA. The rationale behind the observed left-right differences despite a similar degree of fibrosis remains unknown but the expected pathophysiology of each model may provide some insights.

During strenuous exercise the haemodynamic overload in right-sided cavities exceeds that of the left cardiac chambers in athletes31 and in mice.27 Such increased stress may be on the basis of an earlier and more severe dysfunction post-exercise in the RA than in the LA.25 Regular, intensive training has been shown to result in impaired systolic and diastolic function, as well as myocardial fibrosis in the RV, though not in the LV. This may also contribute to selective retrograde chronic overload in the RA.28 In keeping with a more severe insult to the RA, we found remarkable activation of pro-inflammatory pathways in the RA, which was not evident in the LA, and may mediate atrial fibrosis and proarrhythmogenicity.32 Local myocardial inflammation could result from transient exercise-induced systemic inflammation and, at least partially, stretch-induced local myocardial inflammation.8,33 It is therefore likely that the more severe overload on the RA may lead to persistent proinflammatory changes, whereas the lower intensity of stretch-activated inflammation in the LA may motivate its faster regression in-between exercise bouts. Consistent with this hypothesis, in a model of pulmonary hypertension-induced AF that exhibited a similar degree of fibrosis and inflammation in both atria, changes were more persistent in the RA than in the LA after anti-inflammatory therapy.34 Overall, the question of whether the mechanisms leading to fibrosis in the LA are similar to those in the RA and the role of non-inflammatory mechanisms remains uncertain.

Clinical implications

If confirmed, our translational study may have several direct clinical implications. A controversial issue in the field is whether an exercise safety threshold exists for AF. Some methods have been proposed, including lifetime cumulative >2000 h of physical activity35 or jogging >5 days/week,36 but it remains unknown whether lower amounts may also increase AF risk, particularly in individuals at low baseline risk. Our findings indicate that even an active lifestyle, which is modelled in our study through moderate physical activity, may heighten the risk of AF in healthy individuals, supporting the findings of several large studies showing a progressive increase of AF risk with cardiorespiratory fitness.18 Notably, neither the latter nor our results seem to align with the U-shaped relationship established between exercise load and AF risk. This discrepancy may be attributed to the fact that the aforementioned studies were conducted in young, healthy individuals, or rats. In the presence of cardiovascular risk factors, the beneficial effects of regular physical activity could offset deleterious pro-arrhythmogenic remodelling, serving as the basis for a decrease in the risk of AF with exercise of moderate intensity.16,17 However, the balance is complex and contingent upon a multitude of conditioning factors, which should be addressed in dedicated studies.

Our results prompt, and may be the basis for, early identification of pro-arrhythmogenic atrial remodelling in athletes. An enhanced, potentially reversible parasympathetic tone6 could be estimated through heart rate variability parameters or, more simply, heart rate. The early identification of late irreversible stages, when fibrosis is present, might be particularly relevant. Indeed, in magnetic resonance imaging (MRI) scans, the atrial fibrosis burden is greater in endurance athletes than in sedentary individuals,5 but the feasibility and scalability of an MRI-driven approach is currently unlikely. Plasma biomarkers have become attractive tools for identifying individuals at risk of exercise-induced AF. Our biomarker analysis revealed no insights into deleterious atrial remodelling at rest. Interestingly, however, pro-inflammatory and pro-fibrotic biomarkers significantly increased after a single bout of strenuous, but not moderate, training. Exercise-triggered increases suggest that the identification of the intensity of transient insults during strenuous bouts of exercise may better flag athletes at risk of AF than measurements at rest. Similarly, a previous study identified that plasma inflammatory biomarkers (IL-12p70, TNF) at post-exercise correlated with exercise-induced myocardial dysfunction.37 Nonetheless, the interpretation of plasma biomarkers in the daily clinical practice remains complex: the source of these biomarkers is likely extracardiac and correlation analyses cannot infer causality. Three specific biomarkers, TNFRSF11B, which has been linked to AF pathology,38 the immunomodulator VSIG2, and TNNI3, are shared between exercise and pathological pressure overload and could be of particular interest.

Finally, potential future therapeutic implications of our findings should also be considered. Antiarrhythmic drugs, ablation procedures or novel therapeutic approaches targeting parasympathetic tone enhancement might prove particularly effective in the management of patients with AF who regularly exercise, irrespective of its intensity; a more extensive substrate modification might be necessary in highly trained individuals.

Limitations

Extrapolation of animal model results to humans is challenging, particularly in terms of different exercise loads. Although estimates suggest that the exercise loads are at ∼60 and 85% of the maximum O2 capacity for MOD and INT, it remains uncertain whether these approximations and loads are reasonable in freely moving rats. The use of young male rats precludes the extrapolation of our results to female or elderly rats. Indeed, we focused our work on males because evidence supporting the existence of exercise-induced AF in women is inconclusive.39 Wild-type rodent models seldom develop spontaneous AF, and induction testing is warranted. Therefore, our rat model should be considered an early model of pre-AF, in contrast to athletes, who are clinically evaluated only once AF has developed. Large animal models have also shown to develop exercise-induced AF (e.g. horses40 and goats41) and may offer certain physiological similarities to human, but they involve greater ethical concerns and logistical needs. Of note, previous findings in our rat model have been later confirmed by clinical data,5 highlighting its clinical relevance. Finally, our results are based on endurance exercise, and other types of training may lead to different outcomes.42 On the other hand, our data offer no definitive evidence to substantiate a causal relationship between parasympathetic tone enhancement and the risk of AF, nor between right atrial inflammation, fibrosis, and AF; these should be the focus of future, dedicated research.

Conclusions

Endurance exercise load drives a progressive increase in the risk of AF in a young and healthy animal model. Vagal hypertonia and atrial dilatation emerged as early atrial arrhythmogenic substrate biomarkers, whereas atrial fibrosis in the intensive exercise group was crucial for further increasing AF risk at later stages. Local atrial inflammation may play a role in strenuous exercise-induced fibrosis, particularly in the RA. The increase in inflammatory biomarkers in plasma after a bout of exercise could serve as an indicator of advanced stages of pro-arrhythmogenic atrial remodelling.

Supplementary Material

euaf098_Supplementary_Data
euaf098_supplementary_data.zip (3.4MB, zip)

Acknowledgements

We want to thank Nadia Castillo for her exceptional technical support throughout the project and the Functional Genomics platform of the FRCB-IDIBAPS for transcriptomics sample processing.

Contributor Information

Anna Alcarraz, Fundació de Recerca Clínic Barcelona-Institut d’Investigacions Biomèdiques August Pi i Sunyer (FRCB-IDIBAPS), Rosselló 149, 08036 Barcelona, Catalonia, Spain; Medicine Department, Universitat de Barcelona (UB), Casanova 143, 08036 Barcelona, Catalonia, Spain.

Aline Meza-Ramos, Fundació de Recerca Clínic Barcelona-Institut d’Investigacions Biomèdiques August Pi i Sunyer (FRCB-IDIBAPS), Rosselló 149, 08036 Barcelona, Catalonia, Spain; Medicine Department, Universitat de Barcelona (UB), Casanova 143, 08036 Barcelona, Catalonia, Spain; Consejo Nacional de Humanidades, Ciencias y Tecnologías CONAHCYT, Ciudad de Mexico, Mexico.

Cira Rubies, Fundació de Recerca Clínic Barcelona-Institut d’Investigacions Biomèdiques August Pi i Sunyer (FRCB-IDIBAPS), Rosselló 149, 08036 Barcelona, Catalonia, Spain; Medicine Department, Universitat de Barcelona (UB), Casanova 143, 08036 Barcelona, Catalonia, Spain.

Maria Sanz-de la Garza, Fundació de Recerca Clínic Barcelona-Institut d’Investigacions Biomèdiques August Pi i Sunyer (FRCB-IDIBAPS), Rosselló 149, 08036 Barcelona, Catalonia, Spain; Cardiovascular Institute, Clínic Barcelona, Villarroel 170, 08036 Barcelona, Catalonia, Spain.

Carlos Eduardo Bolaños-Gomez, Cardiovascular Institute, Clínic Barcelona, Villarroel 170, 08036 Barcelona, Catalonia, Spain.

Marta Sitges, Fundació de Recerca Clínic Barcelona-Institut d’Investigacions Biomèdiques August Pi i Sunyer (FRCB-IDIBAPS), Rosselló 149, 08036 Barcelona, Catalonia, Spain; Medicine Department, Universitat de Barcelona (UB), Casanova 143, 08036 Barcelona, Catalonia, Spain; Cardiovascular Institute, Clínic Barcelona, Villarroel 170, 08036 Barcelona, Catalonia, Spain; Centro de Investigación Biomédica en Red Enfermedades Cardiovasculares (CIBERCV), Av. Monforte de Lemos, 3-5, 28029 Madrid, Spain.

Lluis Mont, Fundació de Recerca Clínic Barcelona-Institut d’Investigacions Biomèdiques August Pi i Sunyer (FRCB-IDIBAPS), Rosselló 149, 08036 Barcelona, Catalonia, Spain; Medicine Department, Universitat de Barcelona (UB), Casanova 143, 08036 Barcelona, Catalonia, Spain; Cardiovascular Institute, Clínic Barcelona, Villarroel 170, 08036 Barcelona, Catalonia, Spain; Centro de Investigación Biomédica en Red Enfermedades Cardiovasculares (CIBERCV), Av. Monforte de Lemos, 3-5, 28029 Madrid, Spain.

Montserrat Batlle, Fundació de Recerca Clínic Barcelona-Institut d’Investigacions Biomèdiques August Pi i Sunyer (FRCB-IDIBAPS), Rosselló 149, 08036 Barcelona, Catalonia, Spain; Centro de Investigación Biomédica en Red Enfermedades Cardiovasculares (CIBERCV), Av. Monforte de Lemos, 3-5, 28029 Madrid, Spain.

Eduard Guasch, Fundació de Recerca Clínic Barcelona-Institut d’Investigacions Biomèdiques August Pi i Sunyer (FRCB-IDIBAPS), Rosselló 149, 08036 Barcelona, Catalonia, Spain; Medicine Department, Universitat de Barcelona (UB), Casanova 143, 08036 Barcelona, Catalonia, Spain; Cardiovascular Institute, Clínic Barcelona, Villarroel 170, 08036 Barcelona, Catalonia, Spain; Centro de Investigación Biomédica en Red Enfermedades Cardiovasculares (CIBERCV), Av. Monforte de Lemos, 3-5, 28029 Madrid, Spain.

Supplementary material

Supplementary material is available at Europace online.

Funding

This work was partially supported by grants from the Instituto de Salud Carlos III (PI19/00443 and PI22/00953), Consejo Superior de Deportes and the Ministerio de Cultura y Deporte (EXP_75119), and CIBERCV (CB/16/11/00354). A.A. has received a personal scholarship from Instituto de Salud Carlos III (FI20/00080).

References

  • 1. Guasch  E, Mont  L. Diagnosis, pathophysiology, and management of exercise-induced arrhythmias. Nat Rev Cardiol  2017;14:88–101. [DOI] [PubMed] [Google Scholar]
  • 2. Guasch  E, Nattel  S. Ageing, comorbidities, and the complex determinants of atrial fibrillation in athletes. Eur Heart J  2021;42:3526–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Lasocka-Koriat  Z, Lewicka-Potocka  Z, Kaleta-Duss  A, Siekierzycka  A, Kalinowski  L, Lewicka  E  et al.  Differences in cardiac adaptation to exercise in male and female athletes assessed by noninvasive techniques: a state-of-the-art review. Am J Physiol Heart Circ Physiol  2024;326:H1065–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Elliott  AD, Middeldorp  ME, Van Gelder  IC, Albert  CM, Sanders  P. Epidemiology and modifiable risk factors for atrial fibrillation. Nat Rev Cardiol  2023;20:404–17. [DOI] [PubMed] [Google Scholar]
  • 5. Peritz  DC, Catino  AB, Csecs  I, Kaur  G, Kheirkhahan  M, Loveless  B  et al.  High-intensity endurance training is associated with left atrial fibrosis. Am Heart J  2020;226:206–13. [DOI] [PubMed] [Google Scholar]
  • 6. Guasch  E, Benito  B, Qi  X, Cifelli  C, Naud  P, Shi  Y  et al.  Atrial fibrillation promotion by endurance exercise: demonstration and mechanistic exploration in an animal model. J Am Coll Cardiol  2013;62:68–77. [DOI] [PubMed] [Google Scholar]
  • 7. Oláh  A, Németh  BT, Mátyás  C, Horváth  EM, Hidi  L, Birtalan  E  et al.  Cardiac effects of acute exhaustive exercise in a rat model. Int J Cardiol  2015;182:258–66. [DOI] [PubMed] [Google Scholar]
  • 8. Aschar-Sobbi  R, Izaddoustdar  F, Korogyi  AS, Wang  Q, Farman  GP, Yang  F  et al.  Increased atrial arrhythmia susceptibility induced by intense endurance exercise in mice requires TNFα. Nat Commun  2015;6:6018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Goette  A, Corradi  D, Dobrev  D, Aguinaga  L, Cabrera  J-A, Chugh  SS  et al.  Atrial cardiomyopathy revisited—evolution of a concept: a clinical consensus statement of the European Heart Rhythm Association (EHRA) of the ESC, the Heart Rhythm Society (HRS), the Asian Pacific Heart Rhythm Society (APHRS), and the Latin American Heart Rhythm Society (LAHRS). Europace  2024;26:euae204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Wisløff  U, Helgerud  J, Kemi  OJ, Ellingsen  O. Intensity-controlled treadmill running in rats: VO(2 max) and cardiac hypertrophy. Am J Physiol Heart Circ Physiol  2001;280:H1301–10. [DOI] [PubMed] [Google Scholar]
  • 11. Rubies  C, Alcarraz Garca  A, Meza Ramos  A, Batlle  M, Guasch  E. From moderate to strenuous training: unraveling mechanistic contributors and biomarkers for atrial fibrillation in exercise 1.0. Zenodo  2025; doi: 10.5281/zenodo.15510943 [DOI] [PMC free article] [PubMed]
  • 12. Nattel  S, Heijman  J, Zhou  L, Dobrev  D. Molecular basis of atrial fibrillation pathophysiology and therapy: a translational perspective. Circ Res  2020;127:51–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Mikhailova  EV, Romanova  IV, Bagrov  AY, Agalakova  NI. Fli1 and tissue fibrosis in various diseases. Int J Mol Sci  2023;24:1881. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Khan  H, Kella  D, Rauramaa  R, Savonen  K, Lloyd  MS, Laukkanen  JA. Cardiorespiratory fitness and atrial fibrillation: a population-based follow-up study. Heart Rhythm  2015;12:1424–30. [DOI] [PubMed] [Google Scholar]
  • 15. Meza-Ramos  A, Alcarraz  A, Lazo-Rodriguez  M, Sangüesa  G, Banon-Maneus  E, Rovira  J  et al.  High-intensity exercise promotes deleterious cardiovascular remodeling in a high-cardiovascular-risk model: a role for oxidative stress. Antioxidants (Basel)  2023;12:1462. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Mozaffarian  D, Furberg  CD, Psaty  BM, Siscovick  D. Physical activity and incidence of atrial fibrillation in older adults: the cardiovascular health study. Circulation  2008;118:800–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Qureshi  WT, Alirhayim  Z, Blaha  MJ, Juraschek  SP, Keteyian  SJ, Brawner  CA  et al.  Cardiorespiratory fitness and risk of incident atrial fibrillation: results from the Henry Ford Exercise Testing (FIT) project. Circulation  2015;131:1827–34. [DOI] [PubMed] [Google Scholar]
  • 18. Andersen  K, Rasmussen  F, Held  C, Neovius  M, Tynelius  P, Sundström  J. Exercise capacity and muscle strength and risk of vascular disease and arrhythmia in 1.1 million young Swedish men: cohort study. BMJ  2015;351:h4543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Gorman  RA, Yakobov  S, Polidovitch  N, Debi  R, Sanfrancesco  VC, Hood  DA  et al.  The effects of daily dose of intense exercise on cardiac responses and atrial fibrillation key points. J Physiol  2024;602:569–96. [DOI] [PubMed] [Google Scholar]
  • 20. Liu  L, Nattel  S. Differing sympathetic and vagal effects on atrial fibrillation in dogs: role of refractoriness heterogeneity. Am J Physiol  1997;273:H805–16. [DOI] [PubMed] [Google Scholar]
  • 21. Linz  D, Elliott  AD, Hohl  M, Malik  V, Schotten  U, Dobrev  D  et al.  Role of autonomic nervous system in atrial fibrillation. Int J Cardiol  2019;287:181–8. [DOI] [PubMed] [Google Scholar]
  • 22. Claessen  G, De Bosscher  R, Janssens  K, Young  P, Dausin  C, Claeys  M  et al.  Reduced ejection fraction in elite endurance athletes: clinical and genetic overlap with dilated cardiomyopathy. Circulation  2024;149:1405–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Paludan-Müller  C, Vad  OB, Stampe  NK, Diederichsen  SZ, Andreasen  L, Monfort  LM  et al.  Atrial fibrillation: age at diagnosis, incident cardiovascular events, and mortality. Eur Heart J  2024;45:2119–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Oh  Y, Yang  S, Liu  X, Jana  S, Izaddoustdar  F, Gao  X  et al.  Transcriptomic bioinformatic analyses of atria uncover involvement of pathways related to strain and post-translational modification of collagen in increased atrial fibrillation vulnerability in intensely exercised mice. Front Physiol  2020;11:605671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Sanz-de la Garza  M, Grazioli  G, Bijnens  BH, Sarvari  SI, Guasch  E, Pajuelo  C  et al.  Acute, exercise dose-dependent impairment in atrial performance during an endurance race: 2D ultrasound speckle-tracking strain analysis. JACC Cardiovasc Imaging  2016;9:1380–8. [DOI] [PubMed] [Google Scholar]
  • 26. La Gerche  A, Burns  AT, Mooney  DJ, Inder  WJ, Taylor  AJ, Bogaert  J  et al.  Exercise-induced right ventricular dysfunction and structural remodelling in endurance athletes. Eur Heart J  2012;33:998–1006. [DOI] [PubMed] [Google Scholar]
  • 27. Lakin  R, Debi  R, Yang  S, Polidovitch  N, Goodman  JM, Backx  PH. Differential negative effects of acute exhaustive swim exercise on the right ventricle are associated with disproportionate hemodynamic loading. Am J Physiol Heart Circ Physiol  2021;320:H1261–75. [DOI] [PubMed] [Google Scholar]
  • 28. Sanz-de la Garza  M, Rubies  C, Batlle  M, Bijnens  BH, Mont  L, Sitges  M  et al.  Severity of structural and functional right ventricular remodeling depends on training load in an experimental model of endurance exercise. Am J Physiol Heart Circ Physiol  2017;313:H459–68. [DOI] [PubMed] [Google Scholar]
  • 29. Kahr  PC, Piccini  I, Fabritz  L, Greber  B, Schöler  H, Scheld  HH  et al.  Systematic analysis of gene expression differences between left and right atria in different mouse strains and in human atrial tissue. Windt LJ de, ed. PLoS One  2011;6:e26389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Hsu  J, Hanna  P, Van Wagoner  DR, Barnard  J, Serre  D, Chung  MK  et al.  Whole genome expression differences in human left and right atria ascertained by RNA sequencing. Circ Cardiovasc Genet  2012;5:327–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. La Gerche  A, Heidbüchel  H, Burns  AT, Mooney  DJ, Taylor  AJ, Pfluger  HB  et al.  Disproportionate exercise load and remodeling of the athlete’s right ventricle. Med Sci Sports Exerc  2011;43:974–81. [DOI] [PubMed] [Google Scholar]
  • 32. Dobrev  D, Heijman  J, Hiram  R, Li  N, Nattel  S. Inflammatory signalling in atrial cardiomyocytes: a novel unifying principle in atrial fibrillation pathophysiology. Nat Rev Cardiol  2023;20:145–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Lakin  R, Polidovitch  N, Yang  S, Parikh  M, Liu  X, Debi  R  et al.  Cardiomyocyte and endothelial cells play distinct roles in the tumour necrosis factor (TNF)-dependent atrial responses and increased atrial fibrillation vulnerability induced by endurance exercise training in mice. Cardiovasc Res  2023;119:2607–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Hiram  R, Xiong  F, Naud  P, Xiao  J, Sirois  M, Tanguay  J-F  et al.  The inflammation-resolution promoting molecule resolvin-D1 prevents atrial proarrhythmic remodelling in experimental right heart disease. Cardiovasc Res  2021;117:1776–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Calvo  N, Ramos  P, Montserrat  S, Guasch  E, Coll-Vinent  B, Domenech  M  et al.  Emerging risk factors and the dose-response relationship between physical activity and lone atrial fibrillation: a prospective case-control study. Europace  2016;18:57–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Aizer  A, Gaziano  JM, Cook  NR, Manson  JE, Buring  JE, Albert  CM. Relation of vigorous exercise to risk of atrial fibrillation. Am J Cardiol  2009;103:1572–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. La Gerche  A, Inder  WJ, Roberts  TJ, Brosnan  MJ, Heidbuchel  H, Prior  DL. Relationship between inflammatory cytokines and indices of cardiac dysfunction following intense endurance exercise. Tauler P, ed. PLoS One  2015;10:e0130031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Cao  H, Wang  J, Xi  L, Røe  OD, Chen  Y, Wang  D. Dysregulated atrial gene expression of osteoprotegerin/receptor activator of nuclear factor-κB (RANK)/RANK ligand axis in the development and progression of atrial fibrillation. Circ J  2011;75:2781–8. [DOI] [PubMed] [Google Scholar]
  • 39. Guasch  E, Mont  L. Exercise, sex and atrial fibrillation: arrhythmogenesis beyond Y-chromosome?  Heart  2015;101:1607–9. [DOI] [PubMed] [Google Scholar]
  • 40. Carstensen  H, Nissen  SD, Saljic  A, Hesselkilde  EM, van Hunnik  A, Hohl  M  et al.  Long-term training increases atrial fibrillation sustainability in standardbred racehorses. J Cardiovasc Transl Res  2023;16:1205–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Regouski  M, Galenko  O, Doleac  J, Olsen  AL, Jacobs  V, Liechty  D  et al.  Spontaneous atrial fibrillation in transgenic goats with TGF (transforming growth factor)-β1 induced atrial myopathy with endurance exercise. Circ Arrhythm Electrophysiol  2019;12:e007499. [DOI] [PubMed] [Google Scholar]
  • 42. Martinez  V, de la Garza  MS, Grazioli  G, Roca  E, Brotons  D, Sitges  M. Cardiac adaptation to endurance exercise training: differential impact of swimming and running. Eur J Sport Sci  2021;21:844–53. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Citations

  1. Rubies  C, Alcarraz Garca  A, Meza Ramos  A, Batlle  M, Guasch  E. From moderate to strenuous training: unraveling mechanistic contributors and biomarkers for atrial fibrillation in exercise 1.0. Zenodo  2025; doi: 10.5281/zenodo.15510943 [DOI] [PMC free article] [PubMed]

Supplementary Materials

euaf098_Supplementary_Data
euaf098_supplementary_data.zip (3.4MB, zip)

Articles from Europace are provided here courtesy of Oxford University Press

RESOURCES