Intensive recreational athletes in the prospective multinational ICD Sports Safety Registry: Results from the European cohort

Abstract

Background

In the ICD Sports Safety Registry, death, arrhythmia- or shock-related physical injury did not occur in athletes who continue competitive sports after implantable cardioverter-defibrillator (ICD) implantation. However, data from non-competitive ICD recipients is lacking. This report describes arrhythmic events and lead performance in intensive recreational athletes with ICDs enrolled in the European recreational arm of the Registry, and compares their outcome with those of the competitive athletes in the Registry.

Methods

The Registry recruited 317 competitive athletes ≥ 18 years old, receiving an ICD for primary or secondary prevention (234 US; 83 non-US). In Europe, Israel and Australia only, an additional cohort of 80 ‘auto-competitive’ recreational athletes was also included, engaged in intense physical activity on a regular basis (≥2×/week and/or ≥ 2 h/week) with the explicit aim to improve their physical performance limits. Athletes were followed for a median of 44 and 49 months, respectively. ICD shock data and clinical outcomes were adjudicated by three electrophysiologists.

Results

Compared with competitive athletes, recreational athletes were older (median 44 vs. 37 years; p = 0.0004), more frequently men (79% vs. 68%; p = 0.06), with less idiopathic ventricular fibrillation or catecholaminergic polymorphic ventricular tachycardia (1.3% vs. 15.4%), less congenital heart disease (1.3% vs. 6.9%) and more arrhythmogenic right ventricular cardiomyopathy (23.8% vs. 13.6%) (p < 0.001). They more often had a prophylactic ICD implant (51.4% vs. 26.9%; p < 0.0001) or were given a beta-blocker (95% vs. 65%; p < 0.0001). Left ventricular ejection fraction, ICD rate cut-off and time from implant were similar. Recreational athletes performed fewer hours of sports per week (median 4.5 vs. 6 h; p = 0.0004) and fewer participated in sports with burst-performances (vs. endurance) as their main sports: 4% vs. 65% (p < 0.0001). None of the athletes in either group died, required external resuscitation or was injured due to arrhythmia or shock. Freedom from definite or probable lead malfunction was similar (5-year 97% vs. 96%; 10-year 93% vs. 91%). Recreational athletes received fewer total shocks (13.8% vs. 26.5%, p = 0.01) due to fewer inappropriate shocks (2.5% vs. 12%; p = 0.01). The proportion receiving appropriate shocks was similar (12.5% vs. 15.5%, p = 0.51). Recreational athletes received fewer total (6.3% vs. 20.2%; p = 0.003), appropriate (3.8% vs. 11.4%; p = 0.06) and inappropriate (2.5% vs. 9.5%; p = 0.04) shocks during physical activity. Ventricular tachycardia/fibrillation storms during physical activity occurred in 0/80 recreational vs. 7/317 competitive athletes. Appropriate shocks during physical activity were related to underlying disease (p = 0.004) and competitive versus recreational sports (p = 0.004), but there was no relation with age, gender, type of indication, beta-blocker use or burst/endurance sports. The proportion of athletes who stopped sports due to shocks was similar (3.8% vs. 7.5%, p = 0.32).

Conclusions

Participants in recreational sports had less frequent appropriate and inappropriate shocks during physical activity than participants in competitive sports. Shocks did not cause death or injury. Recreational athletes with ICDs can engage in sports without severe adverse outcomes unless other reasons preclude continuation.

Introduction

The interim and long-term results of the Multinational ICD Sports Safety Registry showed that after a median follow-up of 44 months, there were no occurrences of death or arrhythmia- or shock-related physical injury in 440 athletes who continued organized competitive or high-risk sports after implantable cardioverter-defibrillator (ICD) implantation.^1,2 The question of safe continuation of sports comes up even more frequently in ICD recipients who want to continue recreational sports activities.³ Often, ICD patients even want to engage in intensive recreational sports. However, data from non-competitive ICD recipients participating in intensive recreational sports are lacking. In Europe, a cohort of such athletes was enrolled in a recreational arm of the ICD Registry. It was our aim to analyse arrhythmic events and lead performance in these ICD recipients compared with competitive athletes.

Participation of the European sites was endorsed by the Scientific Committee of the European Heart Rhythm Association and the Sports Cardiology Section of the European Association of Preventive Cardiology, both associations of the European Society of Cardiology.

Methods

Methods for this study have been described in detail previously¹ and are summarized here.

Participants and recruitment

The Multinational Registry included athletes with ICDs implanted for primary or secondary prevention, age 10–60 years. They had to perform regular practice of sports with static and dynamic components greater than ‘IA’, as defined by competitive sports recommendations (i.e. more vigorous than golf or bowling).⁴ In Europe, Israel and Australia only, an additional cohort of 80 ‘auto-competitive’ recreational athletes of ≥ 18 years old was also included. They had to be engaged in intense physical activity on a regular basis (≥2×/week and/or ≥ 2 h/week) with the explicit aim to improve their physical performance limits but without competitive participation. These recreational athletes were compared with the 317 competitive athletes 18 years or older of the main study, 234 from the USA and 83 from Europe and Australia. Athletes participating in potentially high-risk sports, defined as those in which a brief loss of control could result in injury (e.g. skiing or surfing) were also included in the main registry,^1,2 but were not included in the current analysis focusing on competitive versus recreational athletes. Participants were enrolled, starting December 2006, directly by participating sites (41 North American, 18 European or Australian) or by ‘self-enrolment’ by contacting the central site (Yale) directly (n = 161). Information was disseminated via mailings to physicians and by direct communication with patients via patient group Internet sites and mailing lists. This study was approved by the Yale Human Investigation Committee (central site) and by the institutional review boards of all participating sites. All participants provided informed signed consent.

Study design

Data were collected and entered in a secure web-based database by personnel at Yale or University Hospital Leuven, Belgium. Athletes were contacted by phone to obtain information on sports participation, and medical records were obtained from their treating cardiologists/electrophysiologists for clinical information at initiation of the study and during follow-up in the case of an event. Participants were asked to call the central site if they received an ICD shock, and were queried regarding preceding activity and any sequelae. Pre-shock activity was categorized as either: 1) during or after competition or practice (within 2 h), 2) other physical activity (recreational activity or other exertion, e.g. running for a bus), or 3) at rest. In the US, participants were required to contact Yale every six months regarding shocks received, and any change in sports participation, health or ICD status, or injuries received. Overall compliance with scheduled every-six-months interviews was 77%. If interviews were missed, information regarding the missed six-month period was obtained at the next call. If not, the treating physicians were contacted to obtain information about shocks and the circumstances. Moreover, in Europe/Australia, all participants were directly contacted every six months, mostly through their own physicians (because of different languages). In only 3.2% of planned contacts, the athlete could not be reached; for most, information was obtained at the next contact. Medical records were obtained from sites and/or participants' treating physicians and/or facilities, at initiation of the study and following a patient-reported event, and were reviewed by study personnel for any shock or changes in health or ICD status (e.g. lead replacement, death). As per study design, event-ascertainment for shocks was by patient report, either spontaneous or in the six-month follow-up calls. Neither office interrogations nor remote monitor reports were systematically collected. Vital status was determined for participants lost to follow-up (n = 20) through contact with their treating physician and could be ascertained for all. The database was closed on 31 January 2015.

Follow-up and endpoints

The median follow-up duration was slightly longer for the recreational athletes than for the competitive athletes (49 vs. 43.9 months; p < 0.001) (Table 1). Event detail data and stored ICD electrograms were each reviewed for rhythm diagnosis and shock outcome by two of three electrophysiologists (RL, BO, HH), as were clinical data on system malfunction (RL, HH)

The primary predetermined endpoint was a serious adverse event during or up to 2 h after sports. This included: 1) tachyarrhythmic death or externally resuscitated tachyarrhythmia due to shock failure, incessant ventricular arrhythmia, or post-shock pulseless electrical activity, and 2) severe injury requiring hospitalization, due to shock or syncopal arrhythmia. All ICD shock data and clinical outcomes were adjudicated by two of three electrophysiologists (HH, RL, BO).

Secondary endpoints were adjudicated based on review of electrograms and clinical data and included: 1) number of appropriate (for ventricular tachyarrhythmia) and inappropriate shock episodes; 2) multiple shocks within one appropriate shock episode (i.e. failure of first maximum-energy shock or recurrent arrhythmia) or electrical storm (>1 event in 24 h); 3) moderate injury (requiring emergency room visit) associated with a shock; 4) ICD lead/system damage including definite lead malfunction, defined as change in pacing function and/or documented noise on electrogram and/or visible lead abnormality, or possible malfunction, that is, change in pacing function only requiring lead revision. Site-reported lead malfunctions for which no clinical data could be obtained were also classified as probable (n = 4). All endpoints were reviewed periodically by the Data Safety Monitoring Board.

Statistical analysis

Descriptive statistics were used to summarize participants' characteristics, sports participation and shock episodes. The binomial 95% confidence interval for adverse event was estimated using exact method. Fisher's exact test was performed to compare the frequency of multiple shocks across activity category and to compare incidence of sports-related shocks for demographic and clinical variables. The McNemar test was performed to compare in a pairwise fashion the fraction of people shocked during competition versus other physical activity versus at rest.

Survival analysis was performed for time to lead malfunction using the Kaplan–Meier method, based on date of lead implantation. All analyses were performed using SAS 9.2 (Cary, NC, USA).

As a post-hoc exploratory analysis, the competitive athlete group was subdivided into competitive athletes from the US or EU and analyses repeated.

Results

Recreational versus competitive athlete cohorts

Demographic and clinical characteristics are shown in Table 1, for US and EU competitive athletes and for EU recreational athletes.

Compared with competitive athletes, recreational athletes were older (median 44 vs. 37 years; p = 0.0004) and tended more frequently to be men (79% vs. 68%; p = 0.06). Long-QT syndrome (LQTS) and hypertrophic cardiomyopathy (HCM) were the most common diagnoses in competitive athletes, while arrhythmogenic right ventricular cardiomyopathy (ARVC) was the most prevalent condition in the recreational cohort (p < 0.0001). LQTS and HCM were also much less prevalent in the EU than in the US competitive cohort. The recreational athletes also had less idiopathic ventricular fibrillation or catecholaminergic polymorphic ventricular tachycardia (CPVT) (1.3% vs. 15.4%) and less congenital heart disease (1.3% vs. 6.9%). They more often had a prophylactic ICD implant (51.4% vs. 26.9%; p < 0.0001) and more were given a beta-blocker (95% vs. 65%; p < 0.0001). Left ventricular ejection fraction, ICD rate cut-off and time from implant were similar. Recreational athletes performed fewer hours of sports per week (median 4.5 vs. 6 h, p = 0.0004).

The most common sports were running in the competitive group (39.2%) and cycling in the recreational group (57.5%) (Table 2). Nearly all recreational athletes performed endurance sports while this was only the case for 74% of the competitive athletes. Only 4% of recreational athletes participated in sports with burst-performances as their main sports versus 65% of the competitive athletes (p < 0.0001).

Primary endpoints

There were no tachyarrhythmic deaths or externally resuscitated tachyarrhythmia episodes during or after sports participation in either group, nor any severe injury due to arrhythmia-related syncope or shocks during sports.

Lead survival

There were 21 definite and 12 probable lead malfunctions. Freedom from definite lead malfunction (from time of implant) was similar in competitive and recreational athletes (5-year 97% competitive vs. 99% recreational; 10-year 94% vs. 97%) (Figure 1), as was freedom from definite or probable lead malfunction (5-year 96% competitive vs. 97% recreational; 10-year 91% vs. 93%). There were no generator malfunctions.

Shocks

Recreational athletes received fewer total shocks at any time (13.8% vs. 26.5%, p = 0.01) due to fewer inappropriate shocks (Table 3). The proportion receiving appropriate shocks at any time was similar (12.5% vs. 15.5%, p = 0.21) but the average number of appropriate shocks per athlete was significantly higher in competitive athletes (1.57 vs. 1.36, p = 0.01). Fewer recreational athletes received total (6.3% vs. 20.2%; p = 0.003) and appropriate (3.8% vs. 11.4%; p = 0.04) shocks during physical activity (Table 3 and Figure 2). Ventricular tachycardia/fibrillation storms during physical activity occurred in none of the 80 recreational athletes versus in seven of the 317 competitive athletes (Table 4).

A minority of athletes stopped sports during the course of the study due to shocks. This proportion was not significantly different between competitive and recreational athletes (3.8% vs. 7.5%, p = 0.32). Among those receiving shocks, 23/82 (28%) vs. 3/11 (27%) stopped sports, at least temporarily (NS).

In a bivariate analysis, apart from recreational-vs.-competitive, appropriate shocks during physical activity in both groups combined were related to underlying disease (mainly ARVC, of whom 23% experienced a shock during the four-year follow-up; p = 0.007), secondary prevention indication (17% vs. 7%; p = 0.003) and competitive sports (12% vs. 4%; p = 0.04), but there was no relation with age, gender, beta-blocker use or burst versus endurance sports. The three recreational athletes who experienced ventricular tachycardia during sports participated in mountain hiking (one) and cycling (two). Two of them had a prophylactic ICD in the context of ARVC, and one a secondary implantation after idiopathic ventricular fibrillation. All were Caucasian men, with an ICD implanted for 28–91 months before entering the Registry. Their ages were 23, 44 and 53 years. They performed 5, 3 and 6 h of sports per week respectively. Two of the three were taking beta-blockers. The sinus rate at the time of intervention was 142, 103 and 113 beats/min, or 58–72% of maximum predicted for age. The numbers are small, precluding any definitive conclusions, except for the fact that it confirms that exercise is a particular trigger for arrhythmias in ARVC patients, and that they were doing moderate to vigorous activity at the time of shocks.

Exploratory analysis

In an exploratory analysis, we noticed different shock patterns in the competitive athletes on either side of the Atlantic: US competitive athletes experienced more total shocks and shocks during physical activity than their EU counterparts (Table 3). In a multivariable analysis of predictors of appropriate shock during physical activity/sports, cardiac diagnosis remained significantly related, as was the level of sports activity (competitive vs. recreational), with odds ratios of 2.4 and 6.9 respectively in EU and US competitive athletes compared with recreational athletes (Table 5).

Given the small numbers of events even in this large registry, the explanation for the difference among the US and EU competitive athletes remains elusive. The Supplementary Material online lists all athletes with appropriate shocks during physical activity. The type of sports at the time of shock was mainly running or cycling in all cohorts, without indication for any particular sports as trigger in the US cohort when looking at the individual cases. Differences between US and EU competitive athletes included level of competition, with more US athletes competing at the national or international level (Table 1: 15.1% vs. 10.2%; p = 0.046) and the number of hours of competitive sports they performed tended to be higher. The rate cut-off in the US athletes was lower (200 vs. 207 beats/min; p = 0.03). Fewer US competitive athletes received a primary implant (24.8% vs. 32.6%; p = 0.002) and fewer were taking beta-blockers (65.4% vs. 76.1%; p = 0.03). Although secondary implantation but not beta-blockers was associated with more appropriate shocks during sports in univariate analysis (p = 0.003 respectively p = 0.12), these factors were not associated with shocks during sports in a multivariable analysis (Table 5). US competitive athletes were less frequently men, and had more LQTS and HCM, but none of these were related to more shocks in the main registry.^1,2 Other characteristics associated with shock did not differ between US and EU athletes: ARVC, the cardiac disease that was most significantly related to appropriate sports-related shocks (Table 5 and main Registry) was equally prevalent in both competitive cohorts (Table 1).

Discussion

The main findings of our study are that in ICD patients at least 18 years old, a) overall, competitive athletes received more total shocks at any time, due to a higher number of inappropriate shocks, without difference in overall numbers of appropriate shocks; b) both appropriate and inappropriate shocks during physical activity were more common in the competitive athletes; and c) shocks during recreational or competitive activity were not associated with death or bodily harm, and d) there was a higher rate of shocks during competition/physical activity in competitive athletes in the US compared with the EU.

Multiple prior studies have shown that physical activity acts as a trigger for arrhythmias in patients with underlying cardiovascular conditions.^5–7 In fact, this long-standing observation forms the basis for preparticipation screening. A higher rate of both appropriate and inappropriate shocks was observed during exercise in competitive than in recreational athletes (Figure 2), which indicates that the intensity of physical activity leads to a graded risk for more shocks. Also the observation that the average number of appropriate shocks per athlete was significantly higher in competitive athletes (81/317 = 0.26 vs. 13/80 = 0.16) is another indication that sports triggers arrhythmias. However, it is important to note that the proportion of athletes with appropriate shocks did not differ between competitive and recreational athletes: while the higher-intensity athletes received more shocks during physical activity, the lower-intensity received more at rest. The observation is also consistent with the ‘paradox of exercise’: while exercise can be an acute trigger of sudden death, even in the very fit, overall, exercise is protective against sudden death.⁷

In line with the findings in the main Registry,^1,2 recreational athletes, like competitive, did not incur any major consequences of shocks during sports. This is an important observation, since fear of untoward consequences was historically one of the main reasons to restrict ICD recipients from more than light competitive or moderate recreational sports activity.^3,8,9 The findings of the competitive cohort in the main Registry have already led to a more permissive attitude in the American Heart Association/American College of Cardiology recommendations for sports participation with an ICD, such that competitive sports now ‘may be considered’ for these athletes.¹⁰ This may now be extended to the much larger group of ICD patients who want to engage in recreational sports, especially when considering that the recreational athletes in this study were selected as intensive, and mainly endurance, sports participants.

One has to keep in mind that although ours is the largest observational study of its kind, with a follow-up of almost four years, untoward effects of sports continuation cannot be excluded. Therefore, a decision for sports participation should be a shared one between athlete/patient and physician, after description of the available information and gaps in that information.^10,11 Such a decision needs consideration of the underlying condition itself, since some are associated with more frequent activity-induced shocks (such as ARVC, CPVT or idiopathic ventricular fibrillation)¹² and since physical activity may lead to progression of the underlying substrate itself (especially in the case of familial or exercise-induced gene-elusive ARVC, and in athletes with post-myocarditis fibrotic areas in their left ventricle).^13–15 ICD implantation therefore does not necessarily provide simple clearance for competitive or intensive recreational sports activity.

Apart from the absence of physical harm, consideration of the possible psychological impact of appropriate or inappropriate ICD shocks during sports participation is needed. Of the athletes who experienced shocks in our study 30–40% at least temporarily stopped participation. ICD therapy is a lifelong therapy, where quality of life is not only dependent on the ability to perform (competitive) sports, but also on lifetime trust in the device. Therefore, physicians should be aware that findings about the effectiveness and safety of ICD therapy during sports do not implicitly create pressure on the athlete to continue sports, but leave room for the athlete's own decision. Informed decision making should take place to reevaluate the options of the athlete after any shocks, in addition to evaluation of need for any further evaluation or change in therapy.^16,17 A high proportion of the athletes in the multinational Registry, both competitive and recreational, were taking beta-blockers. Beta-blocker use was not associated with decreased likelihood of shock, but use of beta-blockers was not randomized, and it is possible that arrhythmia was thought to be more likely in these patients. It is unclear how far the results of our study can be extrapolated to athletes without beta-blockers, or how athletes should be counselled if they wish to cease beta-blockers in order to continue high-level practice. Stopping beta-blockers may be associated with more appropriate and more inappropriate shocks due to rapidly conducted atrial rhythms, with potential psychological impact.^18,19

The lead failure rate in the competitive athletes in this analysis was lower than in the main Registry (6% vs. 10% definite lead malfunction at 10 years)¹ and lower than in other cohorts.²⁰ This analysis included only athletes over age 18 (to be comparable to the recreational cohort). The lead malfunction rate in the main Registry was twice as high in the younger age group, as we have reported before.²¹

In an exploratory analysis, we noted a higher rate of shocks during competition or physical activity in US competitive athletes compared with EU competitive athletes. In fact, when comparing only European athletes, the odds ratio (OR) of shocks during sports/physical activity was not significantly higher in competitive versus recreational athletes (OR 2.2; p = 0.32), although the difference was clearly significant for the US competitive versus recreational groups (OR 7.2; p = 0.004). We have explored characteristics of the US and EU competitive athletes which might explain the difference: level of competition, exposure (hours of sports per week), idiopathic ventricular fibrillation, lower rate cut-off, more secondary implants and lower beta-blocker could all contribute to explain this difference. Unfortunately, by design the Registry did not include a recreational cohort in the US, which could have shed more light on reasons for different shock rates in EU versus US athletes in general. Given the small number of events, and the highly significant difference between competitive and recreational athletes overall, it is possible that the US–EU difference is due to chance.

Limitations

Although our Registry is the largest to date exploring the relation between sports and ICD therapies, it is clearly limited by its size and by the fact that no recreational athletes were included in the US to come to definitive conclusions. Larger registry data will be needed to explore this further in the future. Since our study design was based on the reporting of shocks by the athletes themselves (with consecutive inquiry of the source documents by the investigators) some athletes might deliberately not have reported shocks when contacted by the study personnel, which may have led to an underestimation of shocks. Further, it is possible that shocks may have occurred during sleep and gone unnoticed and thus unreported. This would not impact our findings regarding shocks during exercise, but may have influenced our findings on total shocks. Study design did not entail systematic office interrogations nor remote monitoring. This may have led to under-detection of shocks. However, such bias would work against our observation that the US competitive cohort had a higher rate of shocks, since in the EU the six-month follow-up data were mostly relayed to the central study site through the local physicians themselves after read-out of the devices (due to the different languages in Europe). Moreover, there was no registration of all antitachycardia pacing (ATP) sequences. Because ATP is often asymptomatic, we could not evaluate whether ATP-terminated ventricular tachycardia was common during sports. However, if so, it did not result in arrhythmia-related injury. On the other hand, if ATP of ventricular tachycardia induced secondary ventricular fibrillation (and hence shocks), this was accounted for in the results of the study if the athlete reported the shock.

Conclusions

Participants in recreational sports had less frequent appropriate and inappropriate shocks during physical activity than participants in competitive sports. Differences between athletes in the US and the EU may underlie some of this difference. Nonetheless, shocks did not cause death, bodily harm or accelerated lead malfunction, and a majority of athletes continued their sports activities after activity-related shocks. Therefore, recreational athletes with ICDs can safely engage in sports unless other reasons preclude continuation, such as concerns about progression of the underlying disease induced by physical activity, or psychological coping.

Acknowledgements

We want to acknowledge the study coordinators on both sides of the Atlantic (Karin Broos, Cheryl Barth) as well as Agnes Muskens, research nurse at the most active European site (Rotterdam), who regularly contacted the patients and referring physicians for information, and who ensured a high-quality database.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: this study was supported by investigator‐initiated grants from Boston Scientific, Medtronic, and St Jude Medical. These entities had no role in design and conduct of the study; collection, management, analysis, nor in the interpretation of the data and in the preparation, review, or approval of the manuscript.

References