Marathoner’s Heart?

Efforts to evaluate the risks and benefits of exercise, especially prolonged endurance exercise, are almost as old as scientific medicine itself. Hippocrates, the father of scientific medicine, included a chapter on athletic training in his book Regimens in Health and suggested that exercise should be moderate and only part of a healthy lifestyle.¹ Hippocrates was a near contemporary of Pheideppides, an Athenian who, in 490 BC, reportedly died after running 40 km (24 miles) from Marathon to Athens to announce the Athenians’ victory. Unfortunately, this often-quoted story is probably only partly true. The runner was unlikely to have been named Pheidippides. The distance was likely much greater and probably extended from Athens to Sparta to recruit more soldiers, back to Athens to announce that the Spartans were not coming, and, finally, from Athens to Marathon and back—a total distance of approximately 500 km.¹ Furthermore, the exhausted runner probably did not die, because his death is not noted by Herodotus, the major historian of the event. There is an element of truth to the legend, however, because 50 years later, Eucles did die after running to Athens,¹ providing at least some support for the dangers of prolonged exertion.

Competitive athletics thrived in Victorian England because they were thought to build moral and ethical fitness, and the concept of an “athlete’s heart” was more a moral than a physiological concept.¹ The emergence of such sports as the Oxford–Cambridge boat race, endurance cycling, and running was accompanied by concern for the cardiac dangers of prolonged exercise. F.C. Skey, a London physician, suggested that “the hard exercise in rowing” was one of the most common causes of heart disease.² These concerns about rowers’, bicyclists’, and runners’ hearts persisted into the 20th century and were only dispelled by studies demonstrating the benefits of regular physical activity.² Nevertheless, recognition that vigorous or prolonged exercise can transiently increase the risk of sudden cardiac death in those with occult heart disease,^3–5 produce exertional rhabdomyolysis during extreme exertion⁶ (especially in susceptible individuals⁷), and lead to such conditions as dilutional hyponatremia⁸ has remained an important medical interest in the risks of prolonged exercise.

In this issue of Circulation, Neilan and colleagues⁹ provide evidence for cardiac damage during marathon running. These authors measured the serum biomarkers troponin T (cTnT) and N-terminal probrain naturetic peptide (NT-proBNP) and performed echocardiographic measurements of cardiac function in 60 recreational runners before and 20 minutes after the 2004 and 2005 Boston Marathon, a 42-km (26.2-mile) foot race. None of the runners had increased cTnT concentrations before the race, but 60% had cTnT concentrations greater than the 99th percentile of normal (≥0.01 ng/mL) after the run, and 40% had cTnT values concentrations at or above the concentration used by Neilan and colleagues to diagnose myocardial necrosis (>0.03 ng/mL). NT-proBNP concentrations roughly doubled from a median of 106 before to 182 pg/mL after the race. Left ventricular (LV) size and ejection fraction did not change, but estimated pulmonary artery (PA) systolic pressure increased, and right ventricular (RV) end-diastolic volume and the percent change in RV area, an estimate of RV systolic function, decreased. There was also a decrease in LV early diastolic filling and an increase in late diastolic filling, which was consistent with reduced LV compliance. Furthermore, changes in both cTnT and NT-proBNP correlated with changes in cardiac function, and reductions in RV systolic function correlated with the increase in cTnT. Changes in serum cTnT and NT-proBNP levels and cardiac performance were most marked in runners with the least prerace training (P=0.03 for difference in median change between groups). Specifically, runners who averaged fewer than 35 miles of running per week had the largest increases in cTnT and NT-proBNP and the biggest decreases in RV contractility. There were no changes in ischemia-modified albumin (IMA) immediately after the race. IMA is a marker used to exclude myocardial ischemia, but only if cTnT is also not increased. Moreover, it is well known that increased lactate concentrations and decreases in albumin, both of which occur during and after a marathon race, can cause false-negative IMA results. Therefore, these latter findings should be viewed with caution.

The authors duly note that this is not the first article to document increases in serological markers of myocardial damage or changes in echocardiographic indices of myocardial function with prolonged exertion. Indeed, according to a literature search for articles published from 2004 to 2006 alone, there were at least 10 articles documenting increases in cTnT with prolonged exercise, at least 5 documenting increases in serum NT-proBNP concentrations, and at least 6 suggesting decreases in left ventricular systolic or diastolic function after prolonged exercise (P.D. Thompson, unpublished data, 2006). Given such overwhelming evidence of myocardial harm from prolonged exertion and the observation that physical fitness and training do not fully protect fully against increases in cardiac biomarkers in the present⁹ or prior¹⁰ reports, should not prudent clinicians prohibit such activities, even among their healthiest and most fit patients? Perhaps F.C. Skey has been proven right after 139 years!

These results should provoke concern, but it may too early to conclude that clinically important myocardial damage occurs with prolonged endurance exercise. First, to our knowledge, there is no evidence that former endurance athletes suffer more heart failure or cardiac dysfunction from accumulated myocardial damage, although we are unaware of long-term studies evaluating the clinical status of recreational marathon runners similar to those who comprise the present report. Indeed, marathon participation by large numbers of modestly trained runners is a relatively recent phenomenon. We do not know the long-term effects of such participation, although nearly all epidemiological studies support a cardiovascular benefit from habitual physical activity.^11,12

Second, although these increased serum markers must be assumed to be of cardiac origin, reports of cTnT increases with prolonged exercise are reminiscent of earlier reports by some of these same authors¹³ using increased creatine kinase-MB (CK-MB) concentrations to document myocardial damage after similar exercise. These concerns were alleviated when increased concentrations of CK-MB concentrations in skeletal muscle were demonstrated in biopsies from distance runners¹⁴ and the percentage of CK-MB in the gastrocenemius muscle was shown to increase during marathon training.¹⁵ CK-MB is the developmental form of CK and is expressed in embryonal muscle.¹⁶ Consequently, because exercise training can injure skeletal muscle, the repair process may recruit primitive regenerating muscle fibers containing increased muscle CK-MB, which is then released with the muscle injury produced by the exercise event. cTnT is also the predominant early developmental TnT isoform¹⁷ and is present in regenerating skeletal muscle. Among patients with polymyositis or Duchenne dystrophy, cTnT was present in skeletal muscle samples in 8 of 13 and 6 of 6 patients, respectively.¹¹ However, the currently available clinical cTnT assay used by Neilan et al⁹ does not detect the regenerating isoform of cTnT or cTnT in muscle from marathon runners (Apple, unpublished data, 1999). Consequently, it is most likely that the cTnT is of cardiac origin. The long-term significance of this possible cardiac injury is unknown. Animal models support the concept that cTnT is released from cardiac muscle with strenuous exercise. Rats forced to swim for 5 hours had increases in serum cTnT concentrations as well as histologic evidence of myocyte damage 24 and 48 hours after exertion.¹⁸ Interestingly, exercise training for only 8 days before the swim reduced the histologic evidence of cardiac injury.

Third, echocardiography and, indeed, all noninvasive measures of myocardial function are only estimates of cardiac performance. Echocardiogram results have been validated by other methods, but, to our knowledge, they have not been validated in the population in this study. For example, PA pressure measurements with echocardiography often assume a right atrial (RA) pressure of 10 mm Hg, although this is not specified in the present report.⁹ Reductions in intravascular volume could decrease RA pressure and overestimate PA pressure. Runners in the present study lost an average of only 3 lb, but changes in body weight underestimate fluid shifts into the muscle with exercise.¹⁹ Body builders use this phenomenon to produce the pumped muscle appearance before competition. Indices of LV diastolic function are also affected by changes in intravascular volume²⁰ and are not totally independent of heart rate. Heart rate was 40 beats per minute faster after the marathon and could have decreased early, and increased late, diastolic filling. Changes in cardiac performance did correlate directly with changes in serum markers of injury and inversely with training mileage, adding support to the concept of real myocardial damage; however, statistical associations should not be used to indicate causation.

The study by Neilan et al⁹ also adds to the debate over whether cardiac troponin can be released in patients with reversible myocardial injury. In the editorial accompanying the redefinition of acute myocardial infarction, Jaffe et al²¹ suggested that although increases in myocardial biomarkers probably reflected irreversible myocardial injury, there might be a continuum of reversible to irreversible release, and it might be impossible to determine the degree of injury from biomarkers. Guidelines authored by the Cardiac Society of Australia and New Zealand²² were also noncommittal, noting the controversy over whether cytosolic troponin could be released with reversible myocardial ischemia. More recently, Lakkireddy et al²³ reported transient cardiac troponin and BNP increases as well as echocardiographic changes in a patient with myocardial injury caused by Gram-positive aerobic diphtheria. This report indicated that troponin can be released in what appears to be reversible cardiac injury. In contrast to the controversy on troponin release in reversible injury, it has not been as rigorously examined whether B-type natriuretic peptide and NT-proBNP can be released in reversible injury or synthesized after the stress of long-distance running. In patients undergoing radionuclide exercise stress testing, those with myocardial ischemia had 4-fold higher NT-proBNP concentrations than those without ischemia.²⁴ The novelty of this article is the demonstration of an inverse relationship between cardiac biomarker release with V̇o₂max, miles of training in the 8 weeks before the race, finishing time in the race, and muscle soreness.²⁵ A study on the mechanism of cardiac injury, as suggested by echocardiography or imaging, is warranted.

These comments are not intended to detract from the present report,⁹ but we urge caution in the interpretation of this report, both by physicians and the lay press, and we urge examination of the issues raised. It is critically important that future investigators use other measures to examine myocardial function after exercise and to validate the many echocardiographic reports of myocardial dysfunction. Changes in cTnT should be confirmed by other markers of myocardial injury such as cTnI. If concerns about myocardial damage cannot be dismissed, it is important to determine the long-term sequelae of prolonged exercise among minimally trained marathoners. It is also important to determine with certainty whether skeletal muscle is a possible source of some of the biomarkers reported. This seems unlikely, but it should not be dismissed without additional consideration, given the historical experience with CK-MB in similar athletes. On the other hand, given the present results, it is possible in retrospect that even the earlier reports of CK-MB increases with endurance exercise were indicative of myocardial injury. Clinically, physicians should be aware that myocardial markers can increase after prolonged exercise and that such increases require confirmation and other symptoms before assuming that athletes have suffered classical ischemic myocardial injury. Finally, participants in such events should adequately prepare for such events because all markers of possible myocardial dysfunction were more pronounced in the least-trained runners. Even the Victorians would agree with the latter suggestion.

The present report is disturbing, considering the present widespread interest and participation in recreational endurance events. There is, however, compelling evidence for the cardiovascular benefits of regular physical activity, and we should be circumspect in once again evaluating the possibility that there are indeed deleterious consequences of the “marathoner’s heart.”

The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.

Footnotes