Pace yourself: murine model insights into atrial fibrillation in endurance athletes

Atrial fibrillation (AF) is the most common arrhythmia seen clinically and is associated with an increased risk of heart failure and stroke. Moderate-intensity continuous training effectively reduces AF incidence and symptom burden, rendering it the currently favoured exercise therapy for AF patients. Paradoxically, master's athletes have an elevated AF risk seemingly due to very high volumes of exercise; however, the mechanisms by which excessive exercise training may increase AF risk are unclear (Wang et al., 2024). Left atrial (LA) structure, autonomic balance and inflammation are all modified by exercise training and are each independently associated with AF risk (Wang et al., 2024). Examining how the intensity and duration of exercise training (i.e. exercise dose) influence LA geometry, vagal tone and inflammatory state may improve our collective understanding of the pathogenesis of AF in athletes. A recent article in The Journal Physiology by Gorman et al. (2024) has shed important new light on the effects of exercise volume on atrial remodelling and AF vulnerability. In this study, 7-week-old CD1 mice were randomly assigned to one of four groups: 0 (sedentary), 120, 180 and 240 min/day of swim duration. Exercised mice swam up to 5 days per week, until equivalent amounts of cumulative oxygen consumption ( V ̇ O 2 ${\dot V_{{{\mathrm{O}}_{\mathrm{2}}}}}$ ) were achieved (i.e. ∼700 l O2/kg). A range of techniques were used (e.g. echocardiography, electrocardiography with pharmacological blockade, cardiac contractility assessments using invasive haemodynamics, in vivo measurement of effective refractory period and AF inducibility, heart tissue morphometry, histology and immunohistological staining, and cytochrome c oxidase measurements) to comprehensively examine LA structural and functional remodelling, along with autonomic function and cardiac inflammation. By the end of the training period, mean O2 consumption during the swim sessions increased by 20%−25% in all swim-trained groups. The authors interpreted this finding as an improvement in cardiorespiratory fitness given that mice self-regulate, or rather, naturally swim at a very high intensity (Gorman et al., 2024). In support of this, skeletal muscle mitochondrial content also increased in all swim trained groups. Indices of ventricular structure in all three groups of exercised mice were similar and demonstrated mild hypertrophy and left ventricular dilatation. Longer swim durations led to progressively greater vagal-dependent heart rate reductions, atrial-specific remodelling characterized by hypertrophy, fibrosis and macrophage accumulation, and increased AF vulnerability, without significantly affecting the left ventricle. Therefore, while aerobic conditioning led to adaptive remodelling of the ventricles and skeletal muscle, longer exercise durations led to progressively maladaptive atrial-specific remodelling and vagal-dependent AF vulnerability. These findings, and to what extent they may be translated to humans, will be discussed in this Journal Club article. Structural cardiac remodelling, such as left atrial enlargement in response to endurance exercise, is commonly seen in AF coupled with atrial fibrosis, which together may slow electrical propagation by impeding conduction. This in turn may generate a substrate for atrial arrhythmias (Wang et al., 2024). Gorman et al. (2024) assessed both macrophages and fibrosis to establish chamber unique inflammatory cell responses to exercise. Fibrosis within the atrial appendage was more pronounced compared to the LV whilst the atria showed a 2.5- to 3-fold greater increase in macrophages compared to the LV with exercise. This exercise training induced increase in fibrosis and inflammation in the LA has been observed in previous murine studies, and atrial fibrosis has been reported cross-sectionally in human athletes. However, longitudinal training studies are yet to provide causal evidence of exercise training induced fibrosis in humans. As such, whether high volumes of exercise training result in left atrial fibrosis in humans remains controversial. It has been proposed that in humans, the risk of AF is modulated by lifetime exercise hours. Athletes with over 20 years of endurance training are thought to be at increased risk (Wang et al., 2024). Gorman et al. (2024) note that endurance exercise doses ≥55 metabolic equivalents per hour (MET-h) per week may increase the relative risk of AF compared to not exercising and used this metric of MET-h as a means of translating the exercise dose from mice to humans. In humans, 1 MET-h is defined as an energy expenditure of 3.5 ml O2/kg/min and roughly equates to resting energy expenditure; 55 MET-h/week could comprise any combination of intensities (i.e. MET-h) and durations of exercise accumulated throughout the week. Traditional exercise guidelines classify exercise above 6 MET-h as ‘vigorous’ exercise. However, 6 MET-h equates to only ∼40% of maximal aerobic exercise capacity ( V ̇ O 2 max ${\dot V_{{{\mathrm{O}}_{\mathrm{2}}}{\mathrm{max}}}}$ ) for a typical male masters athlete (with a V ̇ O 2 max ${\dot V_{{{\mathrm{O}}_{\mathrm{2}}}{\mathrm{max}}}}$ of 50 ml/kg/min or 14 MET-h) (Hawkins et al., 2001). In comparison, mice have a low maximal aerobic capacity (i.e. ∼ 5 MET-h) relative to their high resting metabolic rate, and thus very high exercise intensities (i.e. 85%−90% of V ̇ O 2 max ${\dot V_{{{\mathrm{O}}_{\mathrm{2}}}{\mathrm{max}}}}$ ) and volumes (i.e. 20 h) were required to exceed 55 MET-h/week in these mice. Additionally, the mice swam for 4.5−9 weeks continuously, with no variation in intensity, or duration, which would be included in typical human endurance training programs. Whilst it is not immediately clear from the original article, the authors clarify (see peer review) that all mice received at least 2 days’ rest, even after the acclimation period. As such, daily exercise in this context reflects ∼5 days per week with at least 2 days of recovery, which may have mitigated further maladaptation. Nonetheless, this volume (i.e. 20 h) and relative intensity of exercise (i.e. 90% of max) for a human with a V ̇ O 2 max ${\dot V_{{{\mathrm{O}}_{\mathrm{2}}}{\mathrm{max}}}}$ of 50 ml/kg/min is estimated to be ∼260 MET-h/week, which is far beyond (∼6-fold higher) that of highly trained human endurance athletes (Torvik et al., 2021). Clearly, these simplistic extrapolations of murine training doses to humans have limitations, and therefore translation of the findings is extremely difficult. Nonetheless, excessive exercise loads, coupled with insufficient recovery, against a backdrop of heightened inflammation, is likely to also provoke maladaptive remodelling in humans. Despite some of the inherent challenges that accompany the translatability of using the techniques employed by Gorman et al. (2024), these data are crucial in better understanding mechanisms of AF and atrial remodelling. Techniques such as atrial tissue morphometry, histology and immunohistological staining are likely not achievable in human subjects. Therefore, Gorman et al. contribute important mechanistic data to this area which are crucial precursors to inform subsequent human research in relation to exercise-induced adverse atrial remodelling. As noted by the authors, the underlying mechanisms behind some of the adverse atrial remodelling (i.e. fibrosis, increased inflammatory cells and hypertrophy) observed in the atria in swim-exercised mice are complex and may be haemodynamically driven. Atrial remodelling is likely driven by a pressure and/or volume stimulus and associated increases in wall stress. Additionally, cardiac time intervals influence both the proportional and absolute duration that the atrium is exposed to filling ‘stretch’. During left ventricular systole, the mitral valve is closed and the LA undergoes stress during its reservoir (filling) phase. This phase represents approximately one-third of the cardiac cycle at rest, but during exercise increases to over half of the cardiac cycle. Therefore, the proportion of time that the LA undergoes stress increases with exercise intensity. Additionally, the magnitude of the stress may increase as LA chamber volume and pressure rise during exercise. Consequently, the thin-walled LA may be susceptible to remodelling due to a combination of haemodynamic challenges observed during intensive exercise including greater circulatory flow (more volume), greater pressures and greater proportion of time that the mitral valve is closed (La Gerche & Claessen, 2016). Thus, the higher volumes of high-intensity exercise likely contributed to greater haemodynamic stress and progressive atrial remodelling. Less clear is whether pressure or volume changes are comparatively similar between humans and mice. Gorman et al. (2024) suggest that cardiovascular responses to exercise in mice may mimic responses in humans. However, there are overt differences between mice and humans which likely impact atrial remodelling and the adaptation to high intensity exercise (e.g. haemodynamic load associated with quadrapedalism vs. bipedalism). Humans have a much greater capacity to increase cardiac output (and V ̇ O 2 ${\dot V_{{{\mathrm{O}}_{\mathrm{2}}}}}$ ) during exercise compared to mice. The average heart rate of the mouse at rest is roughly 600 beats per minute (bpm) and 750 bpm at peak exercise (10%–20% increase), whereas the human's resting heart rate is around 60 bpm and roughly 170 bpm at peak exercise, reflecting upwards of a 300% increase (Milani-Nejad & Janssen, 2014). It could be hypothesized that the constrained ability of mice to increase heart rate leads to more conservative and manageable increase in atrial pressures during systole, as the mitral valve closure time would remain closer to its resting state. Therefore, it is plausible that the mechanisms underpinning atrial remodelling and (mal)adaptation to exercise may differ from humans. The difference in the heart rate response between the two species suggests that while mouse models offer valuable insights, these limitations should be considered when extrapolating findings to human physiology. Further research exploring species specific cardiovascular responses to exercise is warranted, to fully contextualise comparative studies. Lastly, we would like to see studies in this area inclusive of females, to better understand potential sex differences in atrial remodelling and AF risk. Given the growing number of females participating in endurance sport, understanding sex-related differences or similarities in LA loading, remodelling and AF risk is crucial. Integrating female mice into this research design would provide a valuable opportunity to explore potential cardioprotective mechanisms observed in the female heart, and factors that may mediate sex-related differences. In conclusion, using a series of integrative techniques, Gorman et al. (2024) have elegantly provided valuable insights into the impact of incremental exercise volume on atrial remodelling and AF vulnerability in male mice. The authors should be commended for a thorough and comprehensive analysis on an important aspect of sports cardiology. The study prompts consideration of the translatability of these findings to humans, emphasising the need for an integrative, prospective and longitudinal approach to understand the potential pathogenesis of AF in human athletes. Importantly, inclusivity of female athletes in future research endeavours will enrich our understanding of the nuanced interactions between vagal tone, inflammatory state, and atrial structure and function in the context of exercise-induced remodelling and AF risk. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. The authors declare they have no competing interests. All authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship are listed. None. The authors acknowledge the valuable insights and feedback provided by Stephen Wright and John Sasso.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.