Regular aerobic exercise is important for cardiovascular (CV) health (Hower et al., 2018) and decreases the risk of CVdiseases (CVD), such as myocardial infarction and heart failure (Channon, 2020). It also leads to positive chronic structural and functional adaptations to the heart (Sharma, 2003) which are more apparent if the recommended volume of exercise per week is exceeded (George et al., 2012). This has led to an increasing proportion of the general population performing aerobic exercise on a weekly basis for periods of time that greatly exceed exercise recommendations for health (Piercy et al., 2018) (termed high-volume endurance exercise). However, links between high-volume endurance exercise and adverse CV changes are increasingly being reported (Williams and Thompson, 2014; Elliott and La Gerche, 2015), such as sudden cardiac death (Tsuji et al., 1996) and electrocardiogram (ECG) abnormalities (Ector et al., 2007; Abdulla and Nielsen, 2009; Wundersitz et al., 2019). ECG abnormalities include disturbances in the cardiac rhythm (arrhythmias) and abnormalities in the ECG pattern. Arrhythmias can be further classified as those originating in the ventricles (ventricular arrhythmias) or in the atria and other regions of the heart above the ventricles (supraventricular arrhythmias). For example, a 10-fold increase in cardiac arrhythmias was found in adults pre-to-post high-volume endurance exercise performance previously. High-volume endurance exercise is also being performed by people with varied training histories and across the lifespan, which may also be of concern for older adults performing this form of exercise as age is related to ECG abnormality (Zorzi et al., 2018b; Mannina et al., 2021). An ability to understand when the heart is acutely under excess stress from the chronic performance of high-volume endurance exercise would be beneficial for safe high-volume endurance exercise performance for those in the general population, as well as reducing heart injury risks that can be severe and potentially life threatening.

Imbalances in autonomic control are increasingly being linked to chronic ECG abnormality in the literature (Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology, 1996; Sareban et al., 2020), especially in clinical populations (e.g., Smith, 1982; Binkley et al., 1991; Bigger et al., 1992; Algra et al., 1993; Kocovic et al., 1993; Hartikainen et al., 1996; de Oliveira et al., 2014). A relatively simple and non-invasive method that has shown promise for monitoring autonomic control is through heart rate variability (HRV) assessment, which is used to measure changes in intervals between successive heartbeats and is a non-invasive reflection of the health of the cardiac automatic nervous system (ANS) (Rajendra Acharya et al., 2006). A healthy ANS normally displays variability in heart rate and a stressed system is usually less variable (Kleiger et al., 1987; Nolan et al., 1998; La Rovere et al., 2003). Although, elevations in atrial fibrillation (AF) is strongly associated with increased risk of mortality (Bigger et al., 1992) and arrhythmias like AF can lead to elevations in HRV (e.g., low frequency to high frequency [LF/HF] ratio Lombardi et al., 2001).

It has been previously demonstrated that high-volume endurance exercise changes HRV in athletes (Zehender et al., 1990; Carter et al., 2003; Hynynen et al., 2006; Swanson, 2006; Seiler et al., 2007; Seidl and Asplund, 2014). For example, Seiler et al., 2007 found that highly trained athletes displayed changes in HRV, lasting 90 min after performing 60 to 120 min of exercise, and Hynynen et al. (2006) found chronic overtraining in athletes lead to significantly decreased HRV characteristics. Although, HRV analysis appears to provide a useful evaluation of autonomic control in athletes, reflecting the hearts ability to recover and respond to changing exercise-related stimuli (Acharya et al., 2007) a potential limitation of this method is that any activity during waking hours, such as physical movement (Moody, 1992; Saponznikov et al., 1992; Tobaldini et al., 2013) can introduce errors or noise within the signal. Hence, potentially important changes in ANS control might be missed. A more reliable method of assessing HRV, is to section and analyze data for periods where participants heart rate is stable, such as during deep sleep (Clifford et al., 2005; Herzig et al., 2017a,b; Kim and Shin, 2017). Deep sleep refers to sleep periods where there is regular breathing, high delta wave activity and low variation in HRV characteristics (Herzig et al., 2017b). Previous authors have reported that deep sleep HRV is more time accurate and reliable when compared to waking hours and other restful periods (Catcheside et al., 2001).

Deep sleep HRV has been assessed in several studies of athletes, but only shortly after an acute bout of exercise (Al Haddad et al., 2009; Myllymaki et al., 2012; Herzig et al., 2018). For example, Al Haddad et al., 2009 assessed the effect of supramaximal intermittent exercise to exhaustion on deep sleep HRV and found post-exercise vagal-related HRV characteristics significantly changed immediately and 12 h after exercise (i.e., LF/HF ratio expressed as a percentage significantly higher, mean RR significantly lower), but had returned to normal after 36 h. To date, however, no study has assessed the effect of high-volume endurance exercise on deep sleep HRV or investigated if changes in HRV are related to ECG abnormalities in response to high-volume endurance exercise. Therefore, the aim of this study was to assess the effect of high-volume endurance exercise in recreationally active people on HRV characteristics and if changes in HRV characteristics are related to ECG abnormalities in response to high-volume endurance exercise.

ECG abnormalities that were not classified as arrhythmias were noted in seven participants, three pre (2^nd degree AV block [two] and ST wave elevation) and four post (1^st degree AV block, 2^nd degree AV block, and ST-T wave changes [two]). In addition, single beat abnormalities were noted in three participants (Pre: bundle branch block, non-conducted atrial ectopic and ventricular escape beat; Post: junctional escape beat). Participants experienced significantly less total arrhythmias pre high-volume endurance exercise (median = 12, interquartile range = 4 – 62) compared to post (median = 35, interquartile range = 4 – 231; F₁, ₁₄ = 5.038, p = 0.041, η²_p = 0.265). However, the differences pre to post high-volume endurance exercise in supraventricular (pre: median = 37, interquartile range = 3 – 40; post: median = 53, interquartile range = 4 – 57; F₁, ₁₄ = 0.903, p = 0.358, η²_p = 0.061) and ventricular (pre: median = 6, interquartile range = 0 – 6; Post: median = 20, interquartile range = 0 – 20; F₁, ₁₄ = 1.060, p = 0.321, η²_p = 0.07) arrhythmias did not reach significance.

Age was the only factor to strongly correlate with total, supraventricular and ventricular arrhythmia post high-volume endurance exercise (ρ ≥ 0.54, p < 0.05). Arrhythmia pre strongly correlated with total and supraventricular arrhythmia post high-volume endurance exercise (ρ ≥ 0.66, p < 0.01), but not ventricular arrhythmia (ρ = 0.27, p < 0.05). All other factors were not significant (p > 0.05) and displayed moderate or lower correlations (ρ ≤ 0.43) with arrhythmia response post-exercise (Table 3). Deep sleep and 24-h HRV correlations are provided in Supplementary Data.

This was the first study to assess the relationship between deep sleep HRV characteristics, using a controlled measurement technique, and number of ECG abnormality in a recreational population of cyclists after performing multi-day high-volume endurance exercise. The main finding was that during deep sleep LF/HF ratio was a strong predictor of post-exercise total and supraventricular arrhythmia after controlling for age and arrhythmia number pre-exercise. We also confirmed that baseline arrhythmia number, in particular total and supraventricular arrhythmia number, and age were strongly correlated with post-exercise arrhythmia response, and that HRV characteristics were not significantly changed after performing multi-day high-volume endurance exercise.

Results from the current study suggest that post multi-day high-volume endurance exercise total arrhythmia and supraventricular number were related to baseline arrhythmia number and participant age. This result supports the widespread understanding that age is related to ECG abnormality, both in athletes (Andersen et al., 2013; Zorzi et al., 2018a,b) and the general population (Furberg et al., 1994; Lok and Lau, 1996; Mannina et al., 2021). This result is likely influenced by cardiovascular system deterioration with increasing age (Hatch et al., 2011) and accumulated demand on the heart from cumulative years of exercise training (Myrstad et al., 2014). The association between baseline total arrhythmia number with post-exercise total arrhythmia response suggests that multi-day high-volume endurance exercise is not likely to influence ECG abnormality in people with baseline arrhythmias.

Supraventricular arrhythmia number post high-volume endurance exercise was also related to baseline supraventricular number. This result is not unexpected, considering that high volume endurance exercise is accepted as a factor contributing to the development of supraventricular arrhythmias (D’Souza et al., 2019), including bradyarrhythmia, atrioventricular block (Baldesberger et al., 2008) and atrial fibrillation (Andersen et al., 2013; Myrstad et al., 2014). Further, supraventricular arrhythmias can have serious quality of life and sport participation consequences in athletes (D’Souza et al., 2019), although in the general population it has been suggested that they are common and rarely life threatening (Blomstrom-Lundqvist et al., 2003). The current findings highlight the importance of screening programs being performed prior to high-volume endurance exercise to identify unknown and potentially serious cardiac issues (Corrado et al., 2006; Mont et al., 2020).

Controlling for age and baseline arrhythmia number, the current results showed that higher LF/HF ratio post-exercise, when recorded during deep sleep, was a strong predictor of post-exercise total arrhythmia and supraventricular arrhythmia number. This finding was not seen when LF/HF ratio data were analyzed over a 24-h period or when ventricular arrhythmias were assessed. To our knowledge, researchers have not previously assessed the effect of endurance exercise in humans on HRV measures (during deep sleep or 24 h) and used these to predict ECG abnormality response. There has been HRV and ECG abnormality research in one study of horses performing endurance exercise (Flethøj et al., 2016), as well as animal and clinical populations performing single (e.g., Guiraud et al., 2013; Broux et al., 2017; Frick et al., 2019) or multiple (training studies; e.g., Deligiannis et al., 1999; Andrade et al., 2017) sessions of short duration exercise. In these studies, mixed findings were found between LF/HF ratio and ECG abnormality. For example, similar to the current study Guiraud et al. (2013), found an association between ECG abnormality and LF/HF ratio, with a decrease in ECG abnormality strongly related to a decrease in LF/HF ratio in people with chronic heart failure performing high intensity exercise (r = 0.66; p < 0.01). However Broux et al. (2017), found that LF/HF ratio was significantly lower at rest and during exercise (lunging test) in horses with atrial fibrillation (p < 0.001). Although few exercise studies are available, there is an abundance of research in clinical populations demonstrating that either a higher (Lombardi et al., 2001; Rizzo et al., 2015) or lower (Nortamo et al., 2018; Sagnard et al., 2020) LF/HF ratio is associated with ECG abnormality. There are a number of potential factors alone or in combination that might contribute to inconsistent findings in the literature, such as the duration over which HRV was assessed (Furlan et al., 1990), as well as the population (i.e., animal or healthy/clinical human) and type of ECG abnormality (Lombardi et al., 2001) under investigation. For example, mechanisms responsible for heart rate control are not likely to, and cannot, be considered stable with longer term recordings, such as over 24-h (Furlan et al., 1990). Performing moderate physical activity (Moody, 1992; Saponznikov et al., 1992; Tobaldini et al., 2013), emotional circumstances/mental stress (Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology, 1996), and changing respiration (Malliani et al., 1991) or blood pressure (i.e., hypotension) can all induce changes in LF or HF power, hence LF/HF ratio, across longer term recordings. These should be controlled for and the results of the current study support this assumption, suggesting that measurement of HRV during deep sleep are more sensitive to detect ECG abnormalities than those calculated from 24-h recordings.

Regardless of the inconsistencies in the literature, the results of the current study suggest that there is potential to use HRV measures, such as LF/HF ratio, to acutely monitor responses to high-volume endurance exercise. If LF/HF ratio increases post high-volume endurance exercise, this could be used to indicate liability to increased ECG arrhythmia number. Further wearable devices (i.e., phones, fitness tracking watches, etc.) offer a cost-effective way to monitor exercise within 24-h of performance (i.e., whenever they go to sleep) and non-invasively as long as they produce the HRV measure LF/HF ratio during sleep. Future research should continue to assess the relationship between LF/HF ratio and ECG abnormality number after performing high-volume endurance exercise in a large cohort of people, especially those who perform exercise for health benefit. Future studies should endeavor to identify a potential threshold value where any increase in LF/HF ratio above this is of concern.

HRV measures assessed during deep sleep were unchanged as a result of high-volume endurance exercise. Recovery from acute, shorter duration exercise is normally rapid (e.g., Kaikkonen et al., 2007; Michael et al., 2016) and complete recovery might take at least 48 h after high-intensity exercise (Stanley et al., 2013). Our results extend this work and confirm that after 21-days of high-volume endurance exercise participants displayed no differences when measured 72 h after.

High-volume endurance exercise is growing in popularity across the lifespan due to perceived health benefits. However, there is a body of evidence in professional athletes suggesting that exercise might have negative CV consequences when the amount of exercise performed greatly exceeds recommendations. Results from the current study in recreationally active people found that baseline ECG arrhythmia number, in particular supraventricular arrhythmia number, was associated with post high-volume endurance exercise ECG arrhythmia number. After controlling for baseline ECG arrhythmia number, the HRV characteristic LF/HF ratio measured during deep sleep was shown to be strongly associated with ECG arrhythmia number after performing high-volume endurance exercise for 21 consecutive days. LF/HF ratio offers promise as a non-invasive measure associated with high-volume endurance exercise-induced ECG abnormality.

The original contributions presented in the study are available online (doi: 10.26181/1914974) included in Supplementary Material, further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Bendigo Health Care Group Human Research Ethics Committee and La Trobe University Human Ethics Committee (HREC/17/BHCG/9). Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

MK, BG, and CL conceptualized the project. DW and KN performed data collection. DW, KN, and SP analyzed the data. DW, MK, and BW statistically analyzed the data. SP, DW, and BG drafted the first draft. DW, MK, CL, and BG adapted the draft to the final manuscript. All authors contributed to the article and approved the submitted version.

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

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.