The exercise electrocardiogram (ECG) is widely considered the best available test for screening asymptomatic adults without known cardiovascular (CV) disease prior to initiating a vigorous exercise programme due to its prognostic value, widespread availability and low cost. Observational studies have demonstrated an increased relative risk of CV events with positive screening exercise ECG tests in men with diabetes, advanced age, or multiple cardiac risk factors. Recent observational studies have not demonstrated similar prognostic value for exercise ECG testing in asymptomatic healthy women. Despite the predictive ability of exercise ECG testing in several groups, there have been no studies demonstrating a significant impact of screening on morbidity and mortality in completely asymptomatic patients, leading to significant discordance in consensus guidelines on screening. One prospective observational study is ongoing in Italy that may for the first time demonstrate the ability to decrease incident CV events using preparticipation screening exercise ECG testing in adult athletes with targeted exclusion from athletics. Until more conclusive data is available the authors currently recommend screening exercise ECG testing in asymptomatic men with diabetes and asymptomatic men over age 45 with two or more CV risk factors prior to initiating a vigorous exercise programme. Consideration should also be given to screening asymptomatic patients younger than 45 with particularly strong risk factor exposure or elderly patients with fewer than two risk factors.
Statistics from Altmetric.com
Coronary heart disease (CHD) is the most frequent pathological finding among adults who die during exertion,1 2 unlike younger people, for whom congenital heart disease is generally the underlying cause. Although mortality from CHD has fallen substantially over the past three decades, it remains the leading cause of death in adults in the developed world. In a study of 7733 participants in the Framingham Heart Study who were initially free of CHD, the lifetime risk of CHD (angina pectoris, coronary insufficiency, myocardial infarction, or death from coronary heart disease) for individuals at age 40 was 48.6% for men and 31.7% for women.3 Other cardiovascular (CV) diseases such as arrhythmias and valvular heart disease are also significant sources of morbidity and mortality in adults. Unfortunately, the first manifestation of underlying CV disease is often a major CV event or death.
The relative risk of CV events rises dramatically during vigorous exercise. In the medical literature vigorous exercise is usually defined as an absolute exercise work rate of at least six metabolic equivalents (METs), which is historically assumed to equal an oxygen uptake of 21 ml/kg/min and approximates the energy requirements of activities such as jogging. Evidence of acute coronary artery plaque rupture or erosion with thrombotic occlusion is common in previously asymptomatic patients who experience CV events with exercise. Possible mechanisms by which vigorous exercise precipitates CV events include exercise-induced coronary artery spasm, increased arterial shear stress from increases in heart rate and blood pressure, and increased flexing of atherosclerotic epicardial coronary arteries.4 5 Vigorous exercise also could precipitate acute coronary thrombosis by increasing catecholamine-induced platelet aggregation.6 7 In addition, ventricular arrhythmias can be precipitated during exercise through a number of mechanisms including myocardial ischaemia, electrolyte imbalance and high levels of circulating catecholamines (fig 1).8 9
Several studies have demonstrated the risk of exertion-related CV events. Thompson and colleagues studied the incidence of death during jogging in men aged 30–64 in Rhode Island between 1975 and 1980 and demonstrated one death per year for every 7620 joggers, or approximately one death per 396 000 man-hours of jogging.10 This rate was seven times the estimated death rate from CHD during more sedentary activities in Rhode Island. Siscovick and colleagues subsequently demonstrated that the relative risk of cardiac arrest during exercise compared with that at other times was 5 amongst those with the highest levels of habitual activity and 56 amongst those with the lowest levels of habitual activity.11 However, among the habitually active men, the overall risk of cardiac arrest at rest and with exertion was only 40% of that of the sedentary men. Thus, there are clear long-term benefits of habitual exercise, but there is short-term risk associated with exertion. Considerable effort has been devoted to identifying those asymptomatic, previously healthy adults most at risk of exercise-related CV events so that they can be appropriately excluded from vigorous exercise and/or medically treated to minimise morbidity and mortality.
Preparticipation screening in asymptomatic adults without known cardiovascular disease
The performance of a screening test is dictated by the prevalence of the disease in a given population, which determines the pretest probability and the test characteristics (sensitivity and specificity). The utility of a screening test is a function of both the test performance and the ability to successfully intervene upon a disease state to prevent morbidity or mortality in a cost-effective manner. We will first focus on screening test performance and then discuss the possibility of fruitful intervention later in the review. While multiple tests have been evaluated for their ability to screen for CHD, the vast majority of studies have focused on exercise ECG testing, and our review will focus largely on this testing modality.
Exercise ECG testing in asymptomatic healthy adults
The Multiple Risk Factor Intervention Trial (MRFIT), published in 1985, demonstrated that the presence of ST depression in 6438 asymptomatic healthy men during baseline exercise treadmill testing predicted an increased risk of coronary events and cardiac death. The relative risk (RR) of coronary death after 7 years of follow-up for those with an abnormal exercise treadmill test (ETT) after adjustment for standard risk factors was 3.5.12 A similar relationship has been noted in multiple other studies published subsequently, generally demonstrating a RR of between 2 and 5 for those asymptomatic healthy patients with a positive exercise electrocardiogram (ECG) test.13 14 15 16 17 18 19 20 21
Sensitivity and specificity of exercise ECG testing
As the number of studies evaluating exercise ECG testing proliferated, investigators sought to define the sensitivity and specificity of the test. Gianrossi and colleagues applied meta-analysis to 147 published reports up to 1989 including 24 074 patients and comparing exercise-induced ST depression with coronary angiography. They demonstrated wide variability in the sensitivity and specificity of exercise testing to detect CHD with a mean sensitivity of 68% (range 23–100%) and a mean specificity of 77% (range 17–100%).22 However, in most of these studies patients with positive exercise tests were more likely to undergo coronary angiography and therefore be diagnosed with CHD, causing significant work-up bias. To address this issue, Froelicher and colleagues conducted exercise testing and coronary angiography on 814 patients with angina pectoris and they demonstrated more realistic test characteristics with a sensitivity of 45% and a specificity of 85%.23 A similar analysis in asymptomatic patients, in whom the prevalence of CHD is expected to be lower, would likely yield similar or worse test characteristics.
Exercise ECG testing in different subpopulations
Given the limited sensitivity and specificity of exercise ECG testing, considerable effort has been devoted to defining categories of patients in whom screening would be of most benefit, due to either improved test characteristics in the subpopulation or higher disease prevalence leading to a higher pretest probability. The American College of Cardiology/American Heart Association (ACC/AHA), the American College of Sports Medicine (ASCM) and the United States Preventative Services Task Force (USPSTF) have all issued recent guidelines defining their recommendations for exercise ECG testing in asymptomatic healthy patients planning to undertake a vigorous exercise programme.24 25 26 While the USPSTF concluded that no clear recommendations could be made, the ACC/AHA and ACSM defined several subcategories of asymptomatic healthy patients that would most benefit from exercise ECG testing (table 1).
We will review the evidence behind the use of exercise ECG testing for screening in the subgroups defined by the ACC/AHA and ACSM guidelines, including men >45 years of age, women >55 years, patients with diabetes mellitus, and patients with more than two cardiac risk factors.
Exercise ECG testing in men >45 years and women >55 years of age
It has been well demonstrated that the prevalence of CV disease increases with age in both men and women. Beyond a certain prevalence threshold, the pretest probability of a disease and the screening test performance (sensitivity and specificity) will improve sufficiently to merit screening. Using such calculations, consensus groups have arrived at their age cut-offs for appropriate screening, recognising that risk increases in a continuous fashion. However, these calculations assume a standard set of test characteristics which until recently were derived from studies primarily of men.
Recently, however, Mora and colleagues studied the female participants of the Lipid Research Clinics Prevalence Study and demonstrated that ST-segment response did not predict future risk for CHD events.28 Low exercise capacity and low heart rate recovery after exercise were independent predictors of death from CHD (RR 3.52) and of all-cause mortality (RR 2.11) in asymptomatic women. Gulati and colleagues studied asymptomatic female volunteers in Chicago and found that only exercise capacity predicts with statistical significance all-cause mortality.29 Thus, it appears that the standard exercise ECG criteria of ST changes as a marker of myocardial ischaemia cannot be employed in women to reliably prognosticate CV events. Though variables such as exercise capacity and heart rate recovery provide some valuable prognostic information, these parameters do little to guide therapeutic intervention.
Exercise testing in asymptomatic healthy adults with cardiac risk factors
The predictive value of the exercise test improves for patients with cardiac risk factors.14 17 30 Gibbons et al followed 25 927 healthy men, 20 to 82 years of age at baseline (mean 43 years), between 1970 and 1989 at the Cooper Clinic in Dallas Texas for a mean follow-up of 8.4 years. The sensitivity of ST changes during an exercise test to predict coronary death was 61% in their study. The age-adjusted RR of an abnormal exercise test for CHD death was 21 in those with no risk factors, 27 in those with one risk factor, 54 in those with two risk factors, and 80 in those with three or more factors (fig 2). Similarly, Laukkanen et al followed 2682 men in Finland for a mean period of 10 years and demonstrated that ST depression during exercise and in recovery was associated with an increased relative risk of CHD mortality of 5.9 in smokers, 3.8 in hypercholesterolaemic men and 4.7 in hypertensive men, with adjustment for other risk factors. Thus, the prognostic value of exercise ECG testing in asymptomatic healthy men with two or more cardiac risk factors has been well demonstrated.
Exercise testing in asymptomatic adults with diabetes
In 2005, 7% of the US population was affected by diabetes, and it is known that individuals with diabetes are 2–4 times more likely to develop premature CV disease and death than unaffected individuals.31 32 In addition, given the neuropathic changes that can accompany diabetes, there is theoretical concern that these patients may have more silent myocardial ischaemia than non-diabetic individuals. Several studies have evaluated the predictive potential of exercise testing diabetic patients with symptoms or known CHD and have demonstrated that the prognostic performance of the test is similar to or better than exercise testing in non-diabetic patients.33 34 Rutter et al performed exercise ECG tests on 86 patients with type 2 diabetes and no history of CHD, and demonstrated that patients with ST segment depression had a RR of 21 for CHD events.35 In the prediction of CHD events, silent myocardial ischaemia showed higher sensitivity and positive predictive value than the Framingham calculated CHD risk. Lyerly et al performed exercise ECG testing in 2854 asymptomatic adult men without previous CHD and with diabetes mellitus. They demonstrated that across normal, equivocal, and abnormal ECG groups and after adjustment for standard cardiac risk factors the hazard ratios for CHD mortality were 1.00, 1.68, and 2.21 respectively (p(trend)<0.001).36 Thus, the prognostic value of exercise ECG testing in asymptomatic diabetic men has been well demonstrated.
Exercise test parameters other than ST segment changes
Several recent studies have evaluated the prognostic value of exercise-associated risk markers including exercise capacity, chronotropic incompetence, heart rate recovery, exercise-induced premature ventricular contractions, heart rate variability, and T wave alternans.37 38 39 40 41 42 43 These variables may go beyond the detection of myocardial ischaemia and indicate other CV abnormalities, such as aberrant autonomic regulation, that predispose people to CV events. In general, abnormal findings for any of these variables are associated with moderate increases in risk for CV events (RR 1.7 to 3.5).
Screening to prevent CV events during exercise
There have been relatively few studies that have specifically evaluated the ability of exercise ECG testing to prognosticate exertion-related CV events. Siscovick and colleagues studied 3617 asymptomatic, hypercholesterolaemic men (age range 35 to 59 years) who were enrolled in the Lipid Research Clinics Coronary Primary Prevention Trial.13 The sensitivity and specificity of an exercise test performed at enrolment into the study to predict a subsequent exertion-related CV event were 18% and 92%, respectively. This suggests that the ability of the exercise ECG test to predict CV events occurring specifically during exertion is very limited.
More recently Sofi and colleagues performed a cross-sectional study over 5 years of 30 065 people (23 570 men) in Italy who underwent preparticipation screening to obtain clinical eligibility for competitive sports.44 Exercise ECG testing was abnormal in 1459 (4.9%) participants, 1227 of whom had normal resting ECGs. At the end of screening and additional testing, 196 (0.6%) participants were considered ineligible for competitive sports. Cardiovascular causes for disqualification included valvular disease (24%), hypertension (19%), arrhythmia (18%), CHD (9%), conduction disorders (7%), and cardiomyopathies (5%). Among the 159 participants disqualified because of cardiac abnormalities, personal history or physical examination had suggested problems in only six (3.7%) and a large proportion (126, 79.2%) had a normal pattern on resting ECG. Conversely, almost all the disqualified participants showed some cardiac abnormalities during the exercise ECG. The authors concluded that, amongst adults seeking to take part in competitive sports, exercise ECG testing can identify those with cardiac abnormalities. However, without gold standard testing such as universal angiography and echocardiography it is difficult to assess the true effectiveness of screening exercise ECG testing. This study may prove very significant in the future if the investigators can demonstrate that disqualification of selected athletes reduces the incidence of athletic-related CV events.
Intervention in patients with a positive screening test
As previously stated, the utility of a screening test is a function of both the test performance and the ability to successfully intervene in a patient with a positive test to prevent morbidity or mortality. The Atenolol Silent Ischemia Study (ASIST) was a multicentre, randomised, double-blind, placebo-controlled study that evaluated 306 outpatients with mild angina, abnormal exercise tests, and ischaemia on ambulatory monitoring who were randomised to receive either atenolol (100 mg/d) or placebo.45 The primary outcome measure was event-free survival at 1 year by Kaplan-Meier analysis, and events were death and other major CV events. Event-free survival improved in atenolol-treated patients who had an increased time to onset of first adverse event (120 versus 79 days) and fewer total first events compared with placebo. There was a non-significant trend for fewer serious events (death, resuscitation from ventricular tachycardia/fibrillation, non-fatal myocardial infarction, or hospitalisation for unstable angina) in atenolol-treated patients. Thus, the detection of periods of asymptomatic ischaemia in patients with minimal CV symptoms by exercise ECG testing may prevent serious morbidity or mortality through the initiation of beta blocker medical therapy. However, it is important to realise that this study did not enrol completely asymptomatic patients in whom the prevalence of CHD is likely to be significantly lower, thereby worsening the performance of the screening test and the significance of any intervention based on test results.
The Asymptomatic Cardiac Ischemia Pilot (ACIP) study enrolled 558 patients with coronary anatomy amenable to revascularisation, at least one episode of asymptomatic ischaemia on the 48 h ambulatory ECG and ischaemia on treadmill exercise testing. Patients were randomised to one of three treatment strategies: 1) medication to suppress angina (angina-guided strategy, n = 183); 2) medication to suppress both angina and ambulatory ECG ischaemia (ischaemia-guided strategy, n = 183); or 3) revascularisation strategy (angioplasty or bypass surgery, n = 192).46 Medication was titrated atenolol–nifedipine or diltiazem–isosorbide dinitrate. At 1 year, the mortality rate was 4.4% in the angina-guided group (8 of 183), 1.6% in the ischaemia-guided group (3 of 183) and 0% in the revascularisation group. Frequency of myocardial infarction, unstable angina, stroke and congestive heart failure was not significantly different among the three strategies. The revascularisation group had significantly fewer hospital admissions and non-protocol revascularisations at 1 year. This study suggested that revascularisation in minimally symptomatic patients with periods of asymptomatic ischaemia detected on exercise ECG may prevent serious morbidity and mortality. However, these findings have not been confirmed in recent years when much of coronary revascularisation is performed percutaneously with intracoronary stenting, and as with the ASIST trial these results have not been replicated in a truly asymptomatic patient population.
In addition, in 2007 Boden and colleagues published the results of the COURAGE trial, a randomised trial involving 2287 patients who had objective evidence of myocardial ischaemia and significant CHD, which suggested that, with recent improvements in non-interventional therapies, optimal medical treatment is as effective as percutaneous intervention in preventing death, myocardial infarction, or other major CV events.47 We do not know whether these results can be extrapolated to patients with asymptomatic ischaemia. But if this is the case then optimal medical therapy using widely available and inexpensive medications could improve outcomes in these patients, thereby increasing the utility of effective screening tools such as exercise ECG testing in asymptomatic patients.
Cost-effectiveness of screening exercise testing
Several studies have attempted to estimate the cost-effectiveness of screening to identify CHD. Sox and colleagues used a decision analysis model that was structured so that the benefit of screening was achieved through detection of patients with severe disease who would benefit from revascularisation.48 Only direct costs were considered and they were based on reimbursement rates in the late 1980s. Screening 60-year-old men had a cost per life-year saved of $24 600; for 60-year-old women, the cost was $47 606. The presence or absence of risk factors for coronary heart disease affected the cost-effectiveness ratios. The cost per life-year saved was $44 332 for 60-year-old men with no risk factors and $20 504 for those with one or more risk factors. For persons 40 years of age, the cost-effectiveness ratios were $80 349 per life-year saved for men and $216 496 per life-year saved for women. Their model predicted that screening would increase the life expectancy of 60-year-old men at average risk by at most 12 days. The investigators concluded that routine screening was not warranted in general but that it may be beneficial for persons at increased risk for coronary heart disease.
Pilote and colleagues studied over 4000 persons referred to the Cleveland Clinic for screening exercise tolerance testing.49 Data on cost were obtained from 1994 Medicare reimbursement rates: $110 for exercise testing, $1780 for angiography, and $27 270 for coronary artery bypass surgery. Screening identified 19 patients with severe coronary artery obstruction (0.44% of the cohort); of these, 14 had subsequent coronary artery bypass graft surgery. The investigators estimated a cost of $39 623 to identify one case of severe coronary artery disease by screening exercise tolerance testing. The estimated cost per year of life saved was $55 274.
These studies suggest that screening with exercise testing and performing bypass surgery on persons with severe obstructions are not cost-effective in low-risk groups but may be cost-effective in older patients with cardiac risk factors. There have been no cost-effectiveness analyses over the last decade, during which medical therapy has markedly improved and the majority of revascularisation has been performed using percutaneous stenting.
Other screening modalities
Multiple studies have evaluated the use of other screening modalities in asymptomatic patients without known CHD, including resting ECG, 24–48 h ambulatory ECG monitoring, nuclear imaging and stress echocardiography, and computed tomography (CT).
In different epidemiological studies, the presence of variably defined ECG abnormalities has increased the adjusted relative risk for CV mortality and morbidity by 1.5 to 2.5-fold.50 51 However, the utility of these findings for screening is limited. Approximately 1/3–1/2 of individuals with a normal coronary angiogram have ECG abnormalities, approximately 30% of individuals with angiographically proven CHD have a normal resting ECG, and most coronary events occur in individuals without resting ECG abnormalities.52 53 Similarly, ischaemia detected by ambulatory ECG confers an increased RR of CV events but poorly correlates with other measures of CHD.54
Exercise ECG testing can only be performed in patients who are able to exercise and who do not have baseline ECG abnormalities that can interfere with interpretation, including pre-excitation (Wolff–Parkinson–White) syndrome, left bundle branch block, ST depression >1 mm at rest, or a paced ventricular rhythm. Stress radionuclide myocardial perfusion imaging and stress echocardiography are alternative approaches to exercise ECG testing in such patients. Both have been demonstrated to have better test characteristics than exercise ECG testing55; however, these techniques are expensive and not readily available for routine screening purposes. In addition, though there is minimal radiation exposure with perfusion imaging, potential morbidity from increased neoplastic risk is a concern, particularly for any potential screening tool.
There has been considerable interest in the use of electron beam or multidetector row CT to detect and quantify coronary artery calcification (CAC) to screen for coronary artery disease. In both men and women, CAC detected by electron beam computed tomography (EBCT) is highly sensitive for the presence of ⩾50 percent angiographic stenoses but only moderately specific.56 57 The 2007 ACCF/AHA expert consensus document on coronary artery calcium scoring concluded it cannot be recommended as a screening tool for the diagnosis of obstructive CHD because of its low specificity, with a resulting high false positive rate when the test is applied to low-risk populations. But the panel did feel that selective use of coronary calcium scores in patients at intermediate risk may be appropriate.58
Based on its test characteristics, widespread availability, and cost, exercise ECG testing appears to be the best available screening test for several subpopulations of asymptomatic adults without known CV disease prior to initiating a vigorous exercise programme. There is extensive observational data demonstrating the increased relative risk of CV events with positive screening exercise ECG tests in men with diabetes, advanced age or multiple cardiac risk factors. However, recent observational literature does not support the use of exercise ECG testing in asymptomatic healthy women. In this subpopulation, parameters such as exercise capacity and heart rate recovery have been found to be predictive, but at this time these parameters cannot be used to guide therapeutic intervention and thus are of very limited utility.
In subgroups for which exercise ECG testing has been demonstrated to be predictive, the question remains whether positive test results will productively change the management of these patients. Positive results could lead to three possible interventions that may be of benefit: 1) change in medical management, 2) coronary revascularisation or 3) exclusion from vigorous exertion. Several studies such as ASIST and ACIP have demonstrated that detection of periods of asymptomatic ischaemia by exercise ECG or ambulatory ECG monitoring in patients who have minimal CV symptoms can guide medical intervention or surgical revascularisation that significantly reduces morbidity and mortality.45 46 Yet it is difficult to generalise these findings to a completely asymptomatic population, as the CHD prevalence is likely to be higher in those with even minor symptoms. One major prospective observational study, currently being conducted by Sofi and colleagues in Italy, may for the first time demonstrate the ability to decrease incident CV events using preparticipation screening exercise ECG testing in adult athletes with targeted exclusion from athletics. But, until this study is complete, the medical community is left to make screening recommendations based on incomplete evidence. This has resulted in the significant discordance between the screening guidelines and the lack of uniformity in clinical practice. Based on the currently available evidence, the authors recommend screening exercise ECG testing in asymptomatic men with diabetes and asymptomatic men more than 45 years old with two or more CV risk factors prior to initiating a vigorous exercise programme. Consideration should also be given to screening asymptomatic patients younger than 45 with particularly strong risk factor exposure (i.e. familial hypercholesterolaemia) or elderly patients with fewer than two risk factors. We anticipate that the study by Sofi et al will offer more clarity on the utility of exercise ECG testing and we hope that further investigation, including randomised clinical trials, is undertaken to guide medical practice in the future.
What is already known on this topic
Multiple consensus guidelines have been issued over the last decade on the topic of screening asymptomatic adults without known cardiovascular disease prior to initiating a vigorous exercise programme and there has been significant discordance in the recommendations.
What this study adds
The medical literature currently best supports screening exercise ECG testing in asymptomatic men with diabetes and asymptomatic men over age 45 with two or more cardiovascular risk factors prior to initiating a vigorous exercise programme.
Competing interests None.
Provenance and peer review Commissioned; not externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.