Exercise and the Heart — the Harm of Too Little and Too Much

Abstract

Physical activity and exercise training are underutilized by much of Westernized society, and physical inactivity may be the greatest threat to health in the 21st century. Many studies have shown a linear relationship between one’s activity level and heart health, leading to the conclusion that “if some exercise is good, more must be better.” However, there is evolving evidence that high levels of exercise may produce similar or less overall cardiovascular (CV) benefits compared with those produced by lower doses of exercise. Very high doses of exercise may be associated with increased risk of atrial fibrillation, coronary artery disease, and malignant ventricular arrhythmias. These acute bouts of excessive exercise may lead to cardiac dilatation, cardiac dysfunction, and release of troponin and brain natriuretic peptide. The effects of too little and too much exercise on the heart are reviewed in this article, along with recommendations to optimize the dose of exercise to achieve heart health.

Introduction

Considerable evidence indicates the profound benefits of physical activity (PA), exercise training (ET), and higher levels of cardiorespiratory fitness (CRF) on lowering morbidity and mortality from heart disease (^11,38,40). Recent studies have focused on the fact that PA levels do not meet national guidelines in a substantial percentage of the population in the United States, probably in 50% to 80% of the population, and most of the world (^11,38,40), and physical inactivity may represent one of the greatest threats to health in the 21st century (⁵).

Although excessive endurance exercise (EEE), defined as ET > 60 to 90 min per session, is not nearly as prevalent as low levels of PA, occurring in 2% to 5% of the population, there is a fallible logic that “if some exercise is good, then more must be better” (¹⁸). Certainly there are a substantial number of individuals exercising for many hours and on many days while regularly competing in long road races, such as marathons, or long bicycle and swimming races, such as triathlons, and even multiday events (^18,27–30). Although most are of the belief that participation in high levels of PA is good for heart health, there is evolving evidence that a high level of EEE may be associated with risk of cardiotoxicity.

In this review, we discuss the evidence proving the tremendous CV benefit of regular exercise along with the more recent and evolving evidence that high levels of ET may provide similar or even less overall benefit compared with that provided by low doses of ET. Additionally, we review the evidence that high levels of ET may be associated with increased risk of atrial fibrillation (AF), coronary artery disease (CAD), and malignant ventricular arrhythmias (VA). Furthermore this evidence suggests that acute bouts of EEE may lead to cardiac dilatation and dysfunction, especially on the right side of the heart and ventricular septum, in turn leading to the release of cardiac biomarkers, such as cardiac troponin and brain natriuretic peptide (BNP), resulting in VA and increased risk of sudden cardiac death (SCD).

Exercise and Heart Health

Beginning in the 1950s with the seminal work of Morris et al. (^23,24) and continuing in the 1970s with early studies by Paffenbarger et al. (³¹), the positive effects of exercise on heart health have been established clearly (^11,38,40). Regular moderate exercise has been shown to be helpful for both the primary and secondary prevention of CVD in both men and women. For this and other reasons, the U.S. PA Guidelines recommend that all Americans engage in 30 min of moderate exercise (like a brisk walk) for 5 d·wk⁻¹ or more, which equates to 150 min·wk⁻¹ (^11,38,40). The guidelines also point out that similar benefits can be obtained by engaging in more strenuous exercise (like jogging) for shorter periods of time, such that 15 min of jogging done 5 d·wk⁻¹ (or 75 min·wk⁻¹) appears to provide equivalent benefit to doing 150 min·wk⁻¹ of walking.

Risk of SCD with EEE

The risk associated with EEE has been noted for centuries and evident recently with high-profile cases (^{18,27–30,32}). During the Greco-Persian war in 490 BBC, Pheidippides, a 40-year-old Greek herald, ran almost 150 miles during a 2-d period to deliver urgent military messages. On the third day, he ran the 26 miles (actually 24.8, which may be significant because most cardiac events during marathons occur in the last mile) (^29,32) from the battlefield near Marathon to Athens to deliver news of a momentous Greek victory. As per the legend, upon arriving, Pheidippides explained to the Athenians, “Victory is ours!”, and then he immediately collapsed and died. Now 2,500 years later, there are numerous examples of SCD associated with EEE, and in the best-selling book, Born to Run (Christopher McDougall, Knopf Publishers, 2009), Micah True is the mythical long distance runner, Caballo Blanco, who runs as far as 100 miles on some days. In 2012, while out on a relatively routine 12-mile training run, he died suddenly, presumably from a lethal arrhythmia. It is likely that the heart of this 58-year-old veteran endurance athlete showed manifestations of the “Pheidippides cardiomyopathy” (³²).

The number of Americans participating in marathons has increased by more than 25-fold over the past 40 years (^27,29). Certainly high levels of CRF are generally associated with considerable cardioprotection and reduction in CVD. However, this is not translated into better reduction in CV or all-cause mortality among runners. Paradoxically, long-term marathon runners in some studies have been associated with increased, rather than decreased, CAD (^18,27,29).

However, perhaps the most serious consequence of EEE is SCD; although these relatively rare events often generate considerable publicity in major EEE events, SCD remains relatively uncommon (^{18,27–30,32}).

In fact, a recent study reviewed all marathons and half-marathons in the United States during a 10-year period (10.9 million runners and 59 cardiac arrests) and reported that SCD occurred in only 1 of 200,000 participants (0.54/100,000 participants) (¹³). However other studies suggest that the true occurrence may be two- to four-fold higher than this because the data from the study of Kim et al. (¹³) may be contaminated by a large number of half-marathoners and only accounted for SCD during the race itself and not soon afterwards (^18,34,35). Nevertheless, the fatalities in marathons are still relatively uncommon, although the fatality rate in triathlons may be two-fold higher than running-only races, largely because of the increased CV events and SCD during the swim portion of the race (^18,29,32); however, the data in marathons, as opposed to triathlons, are certainly more robust. The high-catecholamine state of such EEE competitions superimposed on preexisting ET-induced structural myocardial abnormalities is the most logical explanation for cases of SCD after common causes (hypertrophic cardiomyopathy, CAD, anomalous coronaries, channelopathies, and other causes) are excluded.

Effects on CV Structure and Function

There are many potential adverse effects with EEE on cardiac structure and function (Fig. 1) (²⁹). In a recent animal study, rats were trained to run strenuously for 60 consecutive minutes daily for 16 wk (³). The ET rats, compared with controls, developed left ventricular (LV) hypertrophy, right ventricular (RV) hypertrophy, diastolic ventricular dysfunction, dilation of both the left and right atria (LA and RA, respectively), as well as considerable collagen deposition within the cardiac chambers. Serious VA were inducible in 42% of the running rats compared with only 6% of controls (P = 0.05). After detraining, however, these adverse structural changes, as well as VA, largely reversed (³). Many previous animal studies also have found acute adverse CV effects of prolonged (up to 6 h) EEE, sometimes employing a rat model of cold water swimming, in which the animals were forced to swim to avoid drowning, thus reducing the clinical relevance due to excessive emotional strain in addition to the EEE (^29,33).

Figure 1: Proposed pathogenesis of cardiomyopathy in endurance athletes. BNP, B-type natriuretic peptide; CK-MB, creatine kinase MB. Reproduced with permission from O’Keefe et al. (²⁹).

However, similar adverse structural remodeling following EEE have been noted in humans (^{18,27–30,32}). Accumulating evidence suggests that the adverse effects of both short-term intense PA and cumulative EEE are most evident in the right-side cardiac chambers (RV and RA). At rest, average cardiac output in an average size human is approximately 5 L·min⁻¹, and this typically increases by five-fold to about 25 L·min⁻¹ during vigorous ET. During chronic long-term exposure to prolonged, high-intensity ET, this increased cardiac output may place more strain on the thinner wall, smaller, right-side cardiac chambers compared with the left side of the heart. Following a marathon, for example, approximately 30% of runners develop acute dilation of cardiac structures, especially the RV and RA, and dysfunction of the RV and ventricular septum. During the postrace period, the cardiac geometric dimensions are restored, but with this recurrent stretch of the chambers and reestablishment of chamber geometry, some individuals may be prone to the development of chronic structural changes, including chronic dilation of the LA, RA, and RV, as well as patchy myocardial scarring in response to the recurrent volume load and excessive cardiac strain. Although these abnormalities are typically asymptomatic and resolve over 24 to 72 h, if they accrue over many years, they may predispose to potentially serious arrhythmias, such as AF and/or VA (^{18,27–30,32}).

In one prospective study of 25 runners (12 men and 13 women), Trivax et al. (³⁹) found that running a marathon caused acute dilation of the RA and RV, with a sudden fall in RV ejection fraction (EF). Similarly La Gerche et al. (¹⁴) studied 40 highly trained aerobic athletes after competing in EEE events (marathons, mean time 3 h; half Ironman triathlons, mean time 5.5 h; full Ironman, mean time 11 h; and Alpine Cycling race, mean time 8 h). On postrace echocardiograms, they noted increases in RV volume and reduced RVEF (but not LVEF) (Fig. 2) and found elevations in biomarkers (troponin and BNP), which correlated with the fall in RVEF. These abnormalities in cardiac structure returned entirely to baseline within the first few weeks and are noted more typically in races of long duration. Of this cohort, 5 of the 40 (12.5%) had myocardial scarring detected on cardiac magnetic resonance imaging (MRI), as demonstrated by late gadolinium enhancement (LGE). Therefore these studies suggest that intense EEE induces RV dysfunction, which largely spares the LV (except for possibly the ventricular septum shared by both ventricles). Even when short-term RV recovery appears to be complete, potentially long-term ET and competition in EEE races may lead to myocardial fibrosis eventually, which potentially could predispose malignant VA.

Figure 2: Differential effect of prolonged intense exercise on RV and LV volumes. Baseline volumes are shown on the left, and the changes in volume after the race are shown on the right. RV volumes increased in the postrace setting, whereas LV volumes decreased, resulting in a decrease in RVEF but not in LVEF. Reproduced with permission from La Gerche et al. (¹⁴).

Another study by Ector et al. (⁸) reported that the reduction in RVEF seemed less significant in asymptomatic athletes compared with that in EEE competitors who have symptoms of arrhythmias, in which RV dilation and reductions in RVEF were more pronounced. In another study, those endurance athletes had a 12% prevalence of RV abnormalities by MRI (⁶).

Other studies also have confirmed the long-term adverse effects on myocardial structure (^{18,27–30,32}), including one study suggesting that the CAD event rate during 2-year follow-up was significantly higher in the athletes than that in controls (P < 0.0001) (⁶).

Impact of EEE on CAD

Recent studies have suggested that long distance runners may have increased levels of atherosclerosis and CAD (^18,37). In a study 6 years ago, male marathon runners had paradoxically increased coronary artery calcification (CAC) as measured by computed tomography (CT) CAC scoring (²¹). A very recent study of men who completed at least one marathon yearly for 25 consecutive years (n = 50) compared with 23 sedentary controls demonstrated increased total plaque volume (P < 0.01), calcified plaque volume (P < 0.0001), and noncalcified plaque volume (P = 0.04) compared with those with EEE (Fig. 3) (³⁷). Despite the fact that runners have better overall CAD risk profiles, these results underscore the potential for very heavy EEE to increase the severity of CAD through mechanisms largely independent of the traditional CAD risk factors.

Figure 3: Marathoners had significantly more total coronary plaque volume, noncalcified plaque volume, and calcified plaque volume compared with those of control subjects. Reproduced with permission from Schwartz et al. (³⁷).

A very recent study by Möhlenkamp et al. (²²) assessed 108 marathon runners and 864 age-matched controls as well as 216 age- and risk factor-matched controls. Recreational marathon runners with myocardial fibrosis by MRI-LGE had higher troponin release than those without LGE. Higher CAC scores and LGE, as opposed to troponin release (which was present in 37% of the marathoners), predicted subsequent CAD events, which were overall similar between marathoners and risk factor-matched nonmarathon controls.

Impact of ET on Risk of AF

We recently evaluated various risk factors, including metabolic factors, for the risk of AF (^19,20). Certainly many epidemiologic and observational studies, although not all (²⁶), have reported a strong statistically significant association between chronic high-intensity ET and elevated risk of AF (^1,19,20). The impact of habitual PA and ET on the risk of AF appears to be nonlinear, with lower rates of AF among moderately active individuals compared with those among sedentary controls, whereas an increased risk of AF was noted among individuals performing EEE or long bouts of high-intensity ET (^{18,27–30,32}). The mechanism of AF with high ET is uncertain, but this may involve acute fluxes in cathecholamines and autonomic tone, atrial stretch, and RV cardiomyopathy.

In a study of 44,410 Swedish men, intense ET of >5 h·wk⁻¹ at age 30 years increased the risk of AF later in life, whereas moderate-intensity PA reduced the risk of AF (⁷). In a study of 5,446 older athletes (mean age, 73 years), again low- and moderate-intensity PA progressively reduced the risk of AF, whereas the rate slightly increased in those with high-intensity PA (slightly more risk than light intensity but still lower than nonexercisers) (²⁵). In nearly 53,000 long distance cross country skiers from Sweden, there was more AF with the greater amount of cross-country races and the fastest finishing times (²). However, with detraining and moderation of ET dose, there is atrial remodeling and normalization of autonomic tone and reduction of AF risk.

Recent Running Studies and U-Shaped Curve

Three very recent running studies have demonstrated the U-shaped relationship between running doses and CV diseases and all-cause mortality (^15,36,41).

The Copenhagen City Heart Study followed 1,878 runners and 10,158 nonrunners for up to 35 years (^28,36). The runners had an impressive 44% lower risk of mortality during follow-up, with an average 6-year extension in life expectancy. However, a U-shaped curve was apparent for mortality with respect to running dose, with the peak benefit noted with slow-to-moderate running speeds, frequency of about 3 times per week, and 1 to 2 h·wk⁻¹ of running. Very high doses of running, however, were associated with trends of worse survival compared with either nonrunners or groups of low- and moderate-dose runners.

Data from 55,000 people from the Aerobic Center Longitudinal Study, including 13,000 runners and 42,000 nonrunners, followed for an average of 15 years, have been reported recently (¹⁵). During follow-up, runners had impressive reductions in all-cause and CV mortality of 30% and 45%, respectively, with an average increase in life expectancy of 3 years. Persistent runners had the greatest event reduction, whereas those who began running and stopped or vice versa received about half of the full beneficial effect. However, when dividing runners into quintiles of doses (miles·wk⁻¹, running days per week, min·wk⁻¹, and running speed), with the exception of speed (faster running always had a trend for better survival), quintile 1 (<6 miles·wk⁻¹, 1 to 2 times per week, <51 min·wk⁻¹) had similar mortality reductions as those in quintiles 2 to 4 and a trend to slightly greater benefit than those in quintile 5 (Fig. 4). The individuals performing higher running doses generally have higher levels of CRF; and in almost all studies in the literature, higher levels of CRF are associated with lower CV and all-cause mortality rates. However, among runners (who already have high levels of CRF even when running at lower volumes), low doses of running provide maximal protection against all-cause and CV mortality (¹⁵).

Figure 4: Central illustration: hazard ratios (HR) of all-cause and CV mortality by running characteristic (weekly running time, distance, frequency, total amount, and speed). Participants were classified into six groups: nonrunners (reference group) and five quintiles of each running characteristic. All HR were adjusted for baseline age (yr), sex, examination year, smoking status (never, former, or current), alcohol consumption (heavy drinker or not), other physical activities except running (0, 1 to 499, or 500MET·min·wk⁻¹), and parental history of CV disease (yes or no). All P values for HR across running characteristics were <0.05 for all-cause and CV mortality except for running frequency of six times a week (P = 0.11) and speed of <6.0 miles/hyperlipidemia (P = 0.10) for CV mortality. Reproduced with permission from Lee et al. (¹⁵).

Additionally, in a study of 24,000 patients with CAD with a history of myocardial infarction, those doing more ET had progressive reductions in CV mortality, up to a point (Figs. 5 and 6) (⁴¹). At running doses of >30 miles·wk⁻¹ (or walking >46 miles·wk⁻¹), there appears to be partial loss of the ET benefit on CV mortality (⁴¹).

Figure 5: Categorical model. Cox proportional survival analyses of the risk of CVD-related mortality vs MET-hours per day run or walked. Relative risk is calculated for 1.07 to 1.8, 1.8 to 3.6, 3.6 to 5.4, 5.4 to 7.2, and 7.2 MET·h·d⁻¹ or more relative to the inadequate exercisers (<1.07 MET·h·d⁻¹). “All CVD-related” mortality includes both “CVD as an underlying cause” and “CVD as a contributing cause for some other underlying cause.” Significance levels are coded as follows: ^a P ≤ 0.05; ^b P ≤ 0.01; and ^c P ≤ .001. The significance levels for 7.2 MET·h·d⁻¹ or more vs less than 1.07 MET·h·d⁻¹ were all nonsignificant, that is, P = 0.99 for all-cause mortality, P = 0.68 for all CVD-related mortality, and P = 0.46 for CVD as the underlying cause of death. MET·h·d⁻¹, metabolic equivalent of task-hours per day. Reproduced with permission from Williams et al. (⁴¹).

Figure 6: Continuous model. Cox proportional survival analyses of the risk of CVD-related mortality vs MET-hours per day run or walked. In the model “α MET·h·d⁻¹ _trimmed (MET·h·d⁻¹ if MET·h·d⁻¹ ≤ 7.2, 7.2 otherwise) + β indicator function (1 = MET·h·d⁻¹ ≥ 7.2, 0 otherwise) + covariates,” the hypothesis β = 0 tests whether the HR is increased significantly above 7.2 MET·h·d⁻¹ relative to the HR at 7.2. Shown is the 15.4% average decrease in the risk for CVD-related mortality per MET-hours per day between 0 and 7.2 (95% confidence interval (CI), 8.9%–21.5%; P < 0.001) and a 2.62-fold risk increase above 7.2 MET·h·d⁻¹ relative to the risk at 7.2 MET·h·d⁻¹ (95%CI, 1.29- to 5.06-fold; P = 0.009). Reproduced with permission from Williams et al. (⁴¹).

These studies all point to the fact that more does not appear to be better and low-to-moderate doses of running seem to be ideal for conferring long-term CV health and enhanced longevity (^{15,18,27–30,32,36,41}).

Nevertheless, over a decade ago, a study on more than 73,000 participants, mostly nonelite skiers, who had competed in a major ski race in Sweden demonstrated that participants had lower mortality from all causes, CV diseases, and cancers compared with age-adjusted mortality rates (¹⁰). Another recent study indicates that participants in the Olympics and professional sports also were associated with long-term protection against mortality from all causes, CV diseases, and cancers, suggesting that previous sports stardom is associated with overall excellent long-term prognosis (¹²).

Optimal ET Dosing

As Hippocrates said centuries ago, “Everything in excess is opposed to nature” (⁹). From a population perspective, lack of PA is much more prevalent than excessive ET. The PA Federal Guidelines call for 150 min·wk⁻¹ of moderate PA or 75 min·wk⁻¹ of vigorous PA. The Institute of Medicine suggests 60 min·d⁻¹ of some PA. Recent evidence has suggested that very few are meeting these minimal requirements (^11,38,40), and we have made a call to action that all clinicians should be promoting PA throughout the health care system (⁴⁰).

Evidence is mounting, however, that indicates a substantial number of individuals (perhaps 2% or 5% of the population) may be overdoing ET, at least from a health standpoint. If the current mantra “exercise is medicine” is embraced, PA/ET may be analogous to a drug, with indications and contraindications as well as issues related to underdosing and overdosing (²⁷). The recent running articles certainly raise the idea that running doses of less than the PA guidelines may produce maximal protection from all-cause and CV mortality (^15,36). On the other hand, very high doses of ET could have adverse effects by increasing the risk of arrhythmias (especially VA and AF) but also with acute bouts of high-volume, high-intensity ET increasing release of troponin and BNP and causing acute cardiac dilation and dysfunction of cardiac chambers. Returning to high bouts of ET before cardiac recovery could prove to be particularly toxic to the heart. Chronic EEE can produce myocardial scarring and potentially produce cardiomyopathies and substrate for sustained and life-threatening VA.

The Other Side of the Story — Providing Fair Balance

Others feel that the evidence that years of high-intensity ET accelerates atherosclerosis or causes cardiac fibrosis is relatively weak, and given the known benefits of competitive ET, discussed previously, on both cardiac and vascular structure and function, this may not be clinically important (^4,16,17). Although there is no argument that EEE can be potentially harmful, it may be inappropriate to overly frighten individuals who want to participate in competitive ET, even EEE, such as marathons, triathlons, or even ultraendurance races (¹⁶). In fact, very high levels of ET may prevent the decreased compliance and dispensability observed with healthy, sedentary aging (^4,17).

Summary and Conclusions

On the basis of multiple studies, it would seem most advisable, at least from a population-wide public health standpoint, to limit vigorous ET to no more than 60 min·d⁻¹, and indeed it is likely that near-maximal health benefits occur at even much lower doses (^15,27,36). On the other hand, people participate in vigorous ET for non-health reasons, including competition/challenge, fun/exhilaration, ego, psychological benefits, friendships, etc. These individuals should recognize that there may be some increased risk, although relatively small, of such vigorous EEE. Ideally, however, a weekly cumulative dose of vigorous ET of not more than 5 h may be ideal, including taking 1 to 2 d·wk⁻¹ off from vigorous and high-intensity ET (²⁷).

Finally as Hippocrates noted centuries ago, “If we could give every individual the right amount of nourishment and exercise, not too little and not too much, we would have found the safest way to health” (⁹). These words from 2,500 years ago still seem prudent and wise today.

The authors declare no conflicts of interest and do not have any financial disclosures.