Dermot Phelan, MD, PhD

Advances in radiotherapy over the past 50 years have dramatically improved outcomes in patients with malignancy. Five-year overall survival rates for Hodgkin lymphoma and non-Hodgkin lymphoma now stand at 80%, and breast cancer survival is 90%.¹

Increased longevity, however, has come at the cost of late side effects such as radiation-induced heart disease (RIHD). Cardiac dysfunction due to radiation involves a spectrum of disease processes in patients who have undergone mediastinal, thoracic, or breast radiotherapy and may involve any cardiac structure, including the pericardium, myocardium, valves, conduction system, and coronary arteries.

Overall, compared with nonirradiated patients, patients who have undergone chest radiotherapy have a 2% higher absolute risk of cardiac morbidity and death at 5 years and a 23% increased absolute risk after 20 years.²

This article will review the pathophysiology and epidemiology of RIHD and will offer a practical approach to its diagnosis and management.

MOST DAMAGE IS ENDOTHELIAL 

Cardiac myocytes are relatively resistant to radiation damage because of their postmitotic state. But endothelial cells remain sensitive to radiation, and the pathophysiology of most forms of RIHD appears to be associated with damage to endothelial cells. Conventional cardiac risk factors such as hyperlipidemia and smoking have been shown to compound and accelerate radiation-induced endothelial damage in animal models.³

Radiation is believed to result in transient increases in oxidative stress, resulting in formation of reactive oxygen species and a subsequent inflammatory response that includes activation of nuclear factor-kappa B. Upregulation of proinflammatory pathways results in increased expression of matrix metalloproteinases, adhesion molecules, and proinflammatory cytokines and downregulation of vasculoprotective nitric oxide.⁴ Indirect evidence for radiation-induced vascular inflammation comes from numerous studies that demonstrated increased levels of the proinflammatory cytokines interleukin 6, tumor necrosis factor alpha, and interferon gamma in Japanese atomic bomb survivors.⁵

RISK FACTORS

Risk factors for RIHD are summarized in Table 1.

The volume of heart irradiated is a major determinant of the development of RIHD.⁶ A retrospective study of 960 breast cancer patients in Stockholm between 1971 and 1976 found that those who had received the highest doses and volumes of cardiac radiation had a threefold higher risk of cardiac death. By comparison, those with lesser volumes of the heart exposed to radiation had no increase in risk of cardiac death compared with the general population.⁷

Younger age at the time of radiotherapy is associated with an increased risk of RIHD in breast cancer and lymphoma patients. A retrospective analysis of 635 patients under age 21 with Hodgkin lymphoma treated with radiotherapy showed a relative risk of fatal myocardial infarction of 41.5 compared with a general population matched for age, sex, and race.⁸

Conventional cardiac risk factors such as smoking, hypertension, diabetes, and hyperlipidemia further increase the risk of RIHD, and radiation increases the cardiotoxicity of chemotherapeutic agents such as anthracyclines.⁹

In general, high-risk patients are defined as those with at least one risk factor for RIHD who underwent anterior or left-sided chest irradiation  (Table 1).¹⁰

CORONARY ARTERY DISEASE

Ischemic heart disease is the most common cause of cardiac death in patients who have undergone radiation therapy. Atherosclerotic lesions in RIHD are morphologically identical to those in nonirradiated vessels and are characterized by intimal proliferation, accumulation of lipid-rich macrophages, and plaque formation.¹¹

A retrospective single-institution study of 415 patients with Hodgkin lymphoma who had undergone radiation therapy found the incidence of coronary artery disease 20 years later to be 10%. The mean time to development of coronary artery disease was 9 years, and all patients who developed it had at least one conventional cardiac risk factor.¹²

A meta-analysis of more than 20,000 breast cancer patients who received radiotherapy in 40 randomized controlled trials found an increase in the rate of non-breast-cancer deaths, primarily from vascular causes (annual event ratio 1.27, P <  .0001).¹³

A randomized controlled trial comparing breast cancer patients who underwent preoperative or postoperative radiotherapy vs those who had surgery alone revealed a significantly higher death rate from coronary artery disease in the postradiotherapy group.⁷

The risk of radiation-induced coronary artery disease is proportional to both the dose and the duration of radiation therapy. A retrospective study of more than 2,000 women undergoing radiotherapy for breast cancer found that the relative risk of coronary artery disease increased linearly by about 7.4% per Gy of radiation to the heart, with no apparent ceiling.¹⁴

The distribution of atherosclerotic coronary arteries correlates well with the areas exposed to the highest doses of radiation. For instance, in left-sided breast cancer, the apex and anterior wall of the heart typically receive the highest doses of radiation; consequently, the left anterior descending and distal diagonal branches are most prominently involved.¹⁵ In patients with lymphoma who undergo radiotherapy to mediastinal nodes and in breast cancer patients receiving radiotherapy to the internal mammary chain, basal structures may be exposed as well. Ostial lesions can also be seen in these patients.¹⁶

The clinical presentation of coronary artery disease in radiotherapy recipients does not differ significantly from that in the general population. Ischemia may be silent, may lead to classic anginal symptoms, or may cause sudden cardiac death. The incidence of silent myocardial infarction has been reported to be higher after mediastinal radiotherapy than it is in the general population, possibly from damage to nerve endings within the radiation field.¹⁷

Management of radiation-associated coronary artery disease

Managing patients with radiation-associated coronary artery disease is challenging, but the therapeutic options remain the same as those in nonirradiated patients and include medical therapy, percutaneous coronary intervention, and coronary artery bypass grafting, depending on the site and extent of disease.¹⁸ Although results are conflicting, there does not seem to be a significant difference in the rates of stent restenosis between patients with a history of radiation therapy and the general population.

Percutaneous coronary intervention is generally preferred to coronary artery bypass grafting in these patients for several reasons. Radiation-induced fibrosis of surrounding structures generally makes surgical procedures more difficult,¹⁹ and inclusion of the internal mammary artery or internal thoracic artery in the radiation field may result in stenosis of these vessels, rendering them unsuitable for harvesting.²⁰ Moreover, many patients with RIHD have concurrent radiation-induced lung damage, which increases the risk of perioperative pulmonary complications.²¹

If the coronary lesions are not amenable to percutaneous intervention, a careful valvular evaluation should be performed preoperatively in view of the frequency of radiation-associated valvular disease. In a study of 72 patients with RIHD undergoing coronary artery bypass grafting, 40% required valvular surgery at the time of surgery or shortly thereafter.²²

Results of studies of coronary artery bypass graft outcomes in patients with a history of thoracic radiation therapy have been conflicting, but success seems to depend on the status of the internal mammary and internal thoracic arteries.²³ Therefore, the patency of these vessels should be elucidated preoperatively by angiography and intraoperatively by visual inspection of the vessels for fibrosis.

A large single-institution study by Wu et al²⁴ revealed higher short-term and long-term mortality rates in patients with RIHD undergoing cardiac surgery than in control patients without RIHD undergoing similar procedures.

VALVULAR DISEASE

Radiation therapy may directly affect heart valves, and both stenotic (Figure 1) and regurgitant lesions have been described. Pathologic findings include leaflet retraction, fibrotic thickening, and late calcification.²⁵

The precise mechanism of radiation-induced valvular disease is unknown but is thought to be a change in the phenotype of valvular interstitial cells from a myofibroblast to an osteoblast-like cell. Radiation results in significant expression of osteogenic factors such as bone morphogenic protein 2, osteopontin, alkaline phosphatase, and runt-related transcription factor 2 by valvular interstitial cells.26

Valvular heart disease is evident in as many as 81% of patients with RIHD, with the aortic and mitral valves affected more commonly than the tricuspid and pulmonic valves.²⁷ Why there are more left-sided valve lesions than pulmonic valve lesions, despite the pulmonic valve’s anterior position in the heart, is unknown but may be due to higher pressures across the left-sided heart valves.

Although valvular disease is common in patients with RIHD, clinically significant disease is not; more than 70% of patients with radiation-induced valvular disease have no symptoms. A study of 38 cases of radiation-induced valvular disease reported a mean time to development of asymptomatic valvular lesions of 11.5 years and an average time to symptomatic valvular dysfunction of 16.5 years, indicating that 5 years seems to be the interval required for progression from asymptomatic to symptomatic valvular RIHD.²⁸

The thickness of the aortomitral curtain (the junction between the base of the anterior mitral leaflet and the aortic root) is an independent predictor of the long-term risk of death in patients with valvular RIHD.²⁹

Management of radiation-induced valvular disease

Management of patients with valvular RIHD poses a major clinical conundrum because of  the high rates of perioperative morbidity and death in patients with a history of chest radiotherapy. In one study,²³ the long-term mortality rate was 45% in postradiotherapy patients undergoing single-valve surgery and 61% in those undergoing surgery on two or more valves, compared with 13% and 17% in patients with no history of chest radiotherapy.²³

Furthermore, valve repair is an unattractive option in these patients because of high failure rates of mitral valve and tricuspid valve repair attributed to ongoing radiotherapy-induced valvular changes after repair.³⁰

As a result, valve replacement is generally preferred in this group. Patients should be advised of the higher risk of perioperative and long-term morbidity and death associated with open heart surgery than in the general population, and that the risks are even higher with repeat open heart surgery.

This risk has implications for the choice of replacement valves in younger patients. Bioprosthetic valves, which deteriorate over time, may not be advisable. Transcatheter aortic valve replacement has been successful in radiation-induced valvular disease and may become the preferred method of aortic valve replacement.³¹

PERICARDIAL DISEASE

Pericardial disease is a frequent manifestation of RIHD and covers a spectrum of manifestations from acute pericarditis, pericardial effusion, and tamponade to constrictive pericarditis. In a necropsy study, 70% of patients with RIHD were found to have pericardial involvement.³²

The mechanism is believed to be radiation-induced microvascular injury resulting in increased capillary permeability and the sometimes rapid development of a protein-rich exudate. Associated inflammation may cause acute pericarditis, which may eventually be complicated by chronic pericarditis. The parietal surface tends to be affected more severely than the epicardium.³³

Perhaps as a result of recent advances such as lower radiation doses, equal weighting of the anterior and posterior fields, and subcarinal blocking, incidence rates of pericarditis as low as 2.5% have been reported.³⁴

Pericardial RIHD may be divided into early acute pericarditis, delayed chronic pericardial effusion, and constrictive pericarditis.

Early acute pericarditis is rare and is thought to represent a reaction to tumor necrosis. It is defined as occurring during radiotherapy and occurs almost exclusively with high-dose radiotherapy for lymphoma. Due to the relatively benign course of acute pericarditis and fear of tumor recurrence, it is not an indication to withhold radiotherapy.³⁵

Delayed chronic pericardial effusion occurs months to years after radiotherapy, is typically asymptomatic, and presents as an enlarged cardiac silhouette on chest imaging.³⁵ Delayed pericardial effusion is followed with imaging. While in many cases it resolves within 2 years, it may also be long-standing. Pericardiocentesis or a pericardial window may be performed to treat symptomatic effusion or delayed effusion causing hemodynamic compromise.^35–37 Hypothyroidism should be ruled out, as it can complicate mantle irradiation and result in chronic pericardial effusion.³⁸

Constrictive pericarditis may occur as a late complication of radiotherapy and typically causes symptoms of congestive heart failure. Pericardial stripping in these patients is complicated by the possibility of coexisting RIHD of the valves, myocardium, or coronary arteries, as well as mediastinal fibrosis. A study of 163 patients who underwent pericardial stripping for chronic pericarditis found a 7-year overall survival rate of only 27%, far lower than the rate for those who had no history of radiation exposure.³⁹ Therefore, these patients are often treated for symptom control with diuretics and a low-salt diet rather than with surgery.

MYOCARDIAL DISEASE

Microvascular injury in the myocardium results in chronic ischemia, which may lead to myocardial fibrosis, typically manifesting as diastolic dysfunction. Chest radiotherapy may result in both systolic and diastolic dysfunction, and dilated and restrictive cardiomyopathy are well-recognized complications.⁴⁰

Historically, high radiation doses resulted in systolic dysfunction in more than half of patients who underwent thoracic radiotherapy.⁴¹ Now, however, fewer than 5% of patients develop reductions in left ventricular ejection fraction, and most cases of radiotherapy-induced cardiomyopathy have a restrictive pattern.⁴²

In a single-institution study, diastolic dysfunction was reported in as many as 14% of patients who underwent thoracic radiotherapy for Hodgkin lymphoma.⁴⁰ Systolic dysfunction is now seen almost exclusively in patients  treated concurrently with cardiotoxic chemotherapeutic agents such as anthracyclines in addition to radiotherapy.⁴³

In a childhood cancer survival series, the hazard ratio of congestive heart failure in patients who had undergone radiotherapy for Wilms tumor was 6.6—almost identical to the occurrence in sibling controls. By contrast, the hazard ratio increased to 18.3 in those who received doxorubicin in addition to radiotherapy.⁴⁴

Treatment of radiation-induced cardiomyopathy

Treatment of radiation-induced cardiomyopathy is similar to that for other forms of cardiomyopathy, with an emphasis on symptom management.

Heart transplant may be an option for highly selected patients with end-stage heart failure secondary to RIHD. In one report, a series of four RIHD patients received a heart transplant, and all four survived past 48 months.⁴⁵ However, data from the United Network of Organ Sharing revealed an increase in the all-cause mortality rate in patients undergoing heart transplant for RIHD compared with those undergoing transplant for cardiomyopathy due to other causes.⁴⁶ This trend may be confounded by a higher prevalence of prior cardiac surgery in the RIHD group—itself an established risk factor for poor posttransplant outcomes.

CONDUCTION SYSTEM DISEASE

Life-threatening arrhythmias have been reported that are distinct from the common, asymptomatic repolarization abnormalities that occur during radiotherapy. Atrioventricular nodal bradycardia, all degrees of heart block, and sick sinus syndrome have all been reported after chest radiotherapy. As conduction abnormalities do not typically manifest until years after radiotherapy, it is difficult to establish causation and, consequently, to define incidence.

Right bundle branch block is the most common conduction abnormality because of the proximity of the right bundle to the endocardium on the right side.⁴⁷

Chest radiotherapy is also associated with prolongation of the corrected QT interval (QTc). A study in patients with a history of thoracic radiotherapy found that the QTc characteristically increased with exercise, a poor prognostic indicator.⁴⁸ In a study of 134 survivors of childhood cancer, 12.5% of those who had undergone radiotherapy had a resting QTc of 0.44 msec or more.⁴⁹

Furthermore, a study of 69 breast cancer survivors found a higher incidence of conduction abnormalities at 6 months and 10 years after radiotherapy compared with baseline. The characteristic electrocardiographic changes at 6 months were T-wave changes. At 10 years, the T-wave abnormalities had resolved and were replaced by ST depression.⁵⁰

As mentioned above, establishing radiotherapy as a cause for these conduction abnormalities is challenging, given the lag between radiation therapy and electrocardiographic changes. The following criteria have been proposed for establishing a link between atrioventricular blockade and prior radiation⁵¹:

• Total radiation dose to the heart > 40 Gy
• Delay of 10 years or more since therapy
• Abnormal interval electrocardiographic changes such as bundle branch block
• Prior pericardial involvement
• Associated cardiac or mediastinal lesions.

SCREENING GUIDELINES

Consensus guidelines for identifying and monitoring RIHD have been published by the European Association of Cardiovascular Imaging and the American Society of Echocardiography (Table 2).¹⁰ The European Society of Medical Oncology has also issued guidelines for the prevention, diagnosis, and management of cardiovascular disease associated with cancer therapy.

Briefly, the guidelines call for aggressive cardiac risk-factor modification through weight loss, exercise, blood pressure control, and smoking cessation, in addition to early detection of RIHD. Cardiovascular screening for risk factors and a careful clinical examination should be performed in all patients. Baseline comprehensive transthoracic echocardiography is advocated in all patients before starting radiotherapy to detect cardiac anomalies. Beyond this, an annual history and physical examination, paying close attention to the signs and symptoms of cardiopulmonary disease, is essential. The development of new cardiopulmonary symptoms or a new physical finding such as a murmur should prompt evaluation with transthoracic echocardiography.

In patients without symptoms, screening transthoracic echocardiography at 10 years after the start of radiotherapy is recommended in light of the high probability of diagnosing cardiac disease at this juncture. In patients with no preexisting cardiac disease, surveillance transthoracic echocardiography should be at 5-year intervals thereafter.

In high-risk patients without symptoms (those who have undergone anterior or left-sided radiotherapy and have at least one risk factor for RIHD), initial screening transthoracic echocardiography is recommended 5 years after radiotherapy. These patients have a heightened risk of coronary events as described above and, consequently, are recommended to undergo noninvasive imaging 5 to 10 years after radiation exposure. If this initial examination is negative, stress testing should be repeated at 5-year intervals. Stress echocardiography and stress cardiac magnetic resonance imaging have higher specificity than stress electrocardiography and therefore are generally preferred. Stress scintigraphy should be used with caution, as it adds to the cumulative radiation exposure.

The role of magnetic resonance imaging and computed tomography depends on the results of initial transthoracic echocardiography and the clinical indication, in addition to the center’s expertise and facilities. However, there are currently no data advocating their use as screening tools, except for early detection of porcelain aorta in high-risk patients.¹⁰

MODERN RADIOTHERAPY TECHNIQUES

In recent years, there has been emphasis on exposing the patient to as little radiation as possible without compromising cure.⁵² The three major strategies employed to decrease cardiac exposure include reducing the radiation dose, reducing the radiation field and volume, and using newer planning and delivery techniques.

Reducing the radiation dose. It is well recognized that the mean dose of radiation to the heart is a significant predictor of cardiovascular disease, with one study demonstrating a linear increase in the risk of coronary artery disease with increasing mean heart radiation dose (excess relative risk per Gy 7.4%, 95% confidence interval 3.3%–14.8%).⁵³

Reducing the radiation field and volume. Modern strategies and computed tomography-based radiotherapy planning have enabled a transition from older techniques such as extended-field radiation therapy, mantle-field radiation therapy, and involved-field radiation therapy to new techniques such as involved-node and involved-site radiation therapy.⁵⁴ These have shown promise. For instance, a study in patients with early Hodgkin lymphoma found a mean heart dose of 27.5 Gy with mantle-field therapy compared with 7.7 Gy with involved-node therapy. This decrease in mean heart dose was associated with a reduction in the 25-year absolute excess cardiac risk from 9.1% to 1.4% and a reduction in cardiac mortality from 2.1% to 1%.⁵⁵

Employing newer planning and delivery systems has also demonstrated some promise in reducing rates of cardiac morbidity and mortality. Extended-field radiation therapy, mantle-field radiotherapy, and involved-field radiation therapy were traditionally based on two-dimensional planning and often resulted in large volumes of myocardium being unnecessarily exposed to large doses of radiation because of the uncertainty in targeting. Involved-site and involved-node radiotherapy are based on computed tomography, resulting in more accurate targeting and sparing of normal tissue.

In addition, newer techniques such as intensity-modulated radiotherapy and proton beam therapy have resulted in further improvements in conformality compared with three-dimensional conformal radiotherapy.^56,57 Respiratory motion management, including deep inspiration breath-holding and end-inspiration breath-holding, have decreased the radiation dose to the heart in patients undergoing mediastinal radiotherapy.^58,59

TOWARD THE GOALS OF PREVENTION AND EARLIER DETECTION

As survival from breast cancer and lymphoma has increased, we continue to see legacy or latent effects of therapy, such as RIHD. Radiation therapy can affect any cardiac structure and is a major cause of morbidity and death in cancer survivors.

Modern radiation techniques use a variety of mechanisms to decrease the radiation dose to the heart. A large body of evidence emanating from an era of higher radiation doses and a lack of knowledge of the cardiac effects of radiation highlight the perilous cardiac consequences of chest radiation. With advances in radiotherapy and the development and widespread implementation of consensus guidelines, we envision earlier detection and less frequent occurrence of RIHD, although the latter trend could be blunted by increased cardiovascular risk factors within the population. Given the lag between irradiation and the cardiac consequences, it may be a number of years before any comparisons can be drawn.

Advances in radiotherapy over the past 50 years have dramatically improved outcomes in patients with malignancy. Five-year overall survival rates for Hodgkin lymphoma and non-Hodgkin lymphoma now stand at 80%, and breast cancer survival is 90%.¹

Increased longevity, however, has come at the cost of late side effects such as radiation-induced heart disease (RIHD). Cardiac dysfunction due to radiation involves a spectrum of disease processes in patients who have undergone mediastinal, thoracic, or breast radiotherapy and may involve any cardiac structure, including the pericardium, myocardium, valves, conduction system, and coronary arteries.

Overall, compared with nonirradiated patients, patients who have undergone chest radiotherapy have a 2% higher absolute risk of cardiac morbidity and death at 5 years and a 23% increased absolute risk after 20 years.²

This article will review the pathophysiology and epidemiology of RIHD and will offer a practical approach to its diagnosis and management.

MOST DAMAGE IS ENDOTHELIAL 

Cardiac myocytes are relatively resistant to radiation damage because of their postmitotic state. But endothelial cells remain sensitive to radiation, and the pathophysiology of most forms of RIHD appears to be associated with damage to endothelial cells. Conventional cardiac risk factors such as hyperlipidemia and smoking have been shown to compound and accelerate radiation-induced endothelial damage in animal models.³

Radiation is believed to result in transient increases in oxidative stress, resulting in formation of reactive oxygen species and a subsequent inflammatory response that includes activation of nuclear factor-kappa B. Upregulation of proinflammatory pathways results in increased expression of matrix metalloproteinases, adhesion molecules, and proinflammatory cytokines and downregulation of vasculoprotective nitric oxide.⁴ Indirect evidence for radiation-induced vascular inflammation comes from numerous studies that demonstrated increased levels of the proinflammatory cytokines interleukin 6, tumor necrosis factor alpha, and interferon gamma in Japanese atomic bomb survivors.⁵

RISK FACTORS

Risk factors for RIHD are summarized in Table 1.

The volume of heart irradiated is a major determinant of the development of RIHD.⁶ A retrospective study of 960 breast cancer patients in Stockholm between 1971 and 1976 found that those who had received the highest doses and volumes of cardiac radiation had a threefold higher risk of cardiac death. By comparison, those with lesser volumes of the heart exposed to radiation had no increase in risk of cardiac death compared with the general population.⁷

Younger age at the time of radiotherapy is associated with an increased risk of RIHD in breast cancer and lymphoma patients. A retrospective analysis of 635 patients under age 21 with Hodgkin lymphoma treated with radiotherapy showed a relative risk of fatal myocardial infarction of 41.5 compared with a general population matched for age, sex, and race.⁸

Conventional cardiac risk factors such as smoking, hypertension, diabetes, and hyperlipidemia further increase the risk of RIHD, and radiation increases the cardiotoxicity of chemotherapeutic agents such as anthracyclines.⁹

In general, high-risk patients are defined as those with at least one risk factor for RIHD who underwent anterior or left-sided chest irradiation  (Table 1).¹⁰

CORONARY ARTERY DISEASE

Ischemic heart disease is the most common cause of cardiac death in patients who have undergone radiation therapy. Atherosclerotic lesions in RIHD are morphologically identical to those in nonirradiated vessels and are characterized by intimal proliferation, accumulation of lipid-rich macrophages, and plaque formation.¹¹

A retrospective single-institution study of 415 patients with Hodgkin lymphoma who had undergone radiation therapy found the incidence of coronary artery disease 20 years later to be 10%. The mean time to development of coronary artery disease was 9 years, and all patients who developed it had at least one conventional cardiac risk factor.¹²

A meta-analysis of more than 20,000 breast cancer patients who received radiotherapy in 40 randomized controlled trials found an increase in the rate of non-breast-cancer deaths, primarily from vascular causes (annual event ratio 1.27, P <  .0001).¹³

A randomized controlled trial comparing breast cancer patients who underwent preoperative or postoperative radiotherapy vs those who had surgery alone revealed a significantly higher death rate from coronary artery disease in the postradiotherapy group.⁷

The risk of radiation-induced coronary artery disease is proportional to both the dose and the duration of radiation therapy. A retrospective study of more than 2,000 women undergoing radiotherapy for breast cancer found that the relative risk of coronary artery disease increased linearly by about 7.4% per Gy of radiation to the heart, with no apparent ceiling.¹⁴

The distribution of atherosclerotic coronary arteries correlates well with the areas exposed to the highest doses of radiation. For instance, in left-sided breast cancer, the apex and anterior wall of the heart typically receive the highest doses of radiation; consequently, the left anterior descending and distal diagonal branches are most prominently involved.¹⁵ In patients with lymphoma who undergo radiotherapy to mediastinal nodes and in breast cancer patients receiving radiotherapy to the internal mammary chain, basal structures may be exposed as well. Ostial lesions can also be seen in these patients.¹⁶

The clinical presentation of coronary artery disease in radiotherapy recipients does not differ significantly from that in the general population. Ischemia may be silent, may lead to classic anginal symptoms, or may cause sudden cardiac death. The incidence of silent myocardial infarction has been reported to be higher after mediastinal radiotherapy than it is in the general population, possibly from damage to nerve endings within the radiation field.¹⁷

Management of radiation-associated coronary artery disease

Managing patients with radiation-associated coronary artery disease is challenging, but the therapeutic options remain the same as those in nonirradiated patients and include medical therapy, percutaneous coronary intervention, and coronary artery bypass grafting, depending on the site and extent of disease.¹⁸ Although results are conflicting, there does not seem to be a significant difference in the rates of stent restenosis between patients with a history of radiation therapy and the general population.

Percutaneous coronary intervention is generally preferred to coronary artery bypass grafting in these patients for several reasons. Radiation-induced fibrosis of surrounding structures generally makes surgical procedures more difficult,¹⁹ and inclusion of the internal mammary artery or internal thoracic artery in the radiation field may result in stenosis of these vessels, rendering them unsuitable for harvesting.²⁰ Moreover, many patients with RIHD have concurrent radiation-induced lung damage, which increases the risk of perioperative pulmonary complications.²¹

If the coronary lesions are not amenable to percutaneous intervention, a careful valvular evaluation should be performed preoperatively in view of the frequency of radiation-associated valvular disease. In a study of 72 patients with RIHD undergoing coronary artery bypass grafting, 40% required valvular surgery at the time of surgery or shortly thereafter.²²

Results of studies of coronary artery bypass graft outcomes in patients with a history of thoracic radiation therapy have been conflicting, but success seems to depend on the status of the internal mammary and internal thoracic arteries.²³ Therefore, the patency of these vessels should be elucidated preoperatively by angiography and intraoperatively by visual inspection of the vessels for fibrosis.

A large single-institution study by Wu et al²⁴ revealed higher short-term and long-term mortality rates in patients with RIHD undergoing cardiac surgery than in control patients without RIHD undergoing similar procedures.

VALVULAR DISEASE

Radiation therapy may directly affect heart valves, and both stenotic (Figure 1) and regurgitant lesions have been described. Pathologic findings include leaflet retraction, fibrotic thickening, and late calcification.²⁵

The precise mechanism of radiation-induced valvular disease is unknown but is thought to be a change in the phenotype of valvular interstitial cells from a myofibroblast to an osteoblast-like cell. Radiation results in significant expression of osteogenic factors such as bone morphogenic protein 2, osteopontin, alkaline phosphatase, and runt-related transcription factor 2 by valvular interstitial cells.26

Valvular heart disease is evident in as many as 81% of patients with RIHD, with the aortic and mitral valves affected more commonly than the tricuspid and pulmonic valves.²⁷ Why there are more left-sided valve lesions than pulmonic valve lesions, despite the pulmonic valve’s anterior position in the heart, is unknown but may be due to higher pressures across the left-sided heart valves.

Although valvular disease is common in patients with RIHD, clinically significant disease is not; more than 70% of patients with radiation-induced valvular disease have no symptoms. A study of 38 cases of radiation-induced valvular disease reported a mean time to development of asymptomatic valvular lesions of 11.5 years and an average time to symptomatic valvular dysfunction of 16.5 years, indicating that 5 years seems to be the interval required for progression from asymptomatic to symptomatic valvular RIHD.²⁸

The thickness of the aortomitral curtain (the junction between the base of the anterior mitral leaflet and the aortic root) is an independent predictor of the long-term risk of death in patients with valvular RIHD.²⁹

Management of radiation-induced valvular disease

Management of patients with valvular RIHD poses a major clinical conundrum because of  the high rates of perioperative morbidity and death in patients with a history of chest radiotherapy. In one study,²³ the long-term mortality rate was 45% in postradiotherapy patients undergoing single-valve surgery and 61% in those undergoing surgery on two or more valves, compared with 13% and 17% in patients with no history of chest radiotherapy.²³

Furthermore, valve repair is an unattractive option in these patients because of high failure rates of mitral valve and tricuspid valve repair attributed to ongoing radiotherapy-induced valvular changes after repair.³⁰

As a result, valve replacement is generally preferred in this group. Patients should be advised of the higher risk of perioperative and long-term morbidity and death associated with open heart surgery than in the general population, and that the risks are even higher with repeat open heart surgery.

This risk has implications for the choice of replacement valves in younger patients. Bioprosthetic valves, which deteriorate over time, may not be advisable. Transcatheter aortic valve replacement has been successful in radiation-induced valvular disease and may become the preferred method of aortic valve replacement.³¹

PERICARDIAL DISEASE

Pericardial disease is a frequent manifestation of RIHD and covers a spectrum of manifestations from acute pericarditis, pericardial effusion, and tamponade to constrictive pericarditis. In a necropsy study, 70% of patients with RIHD were found to have pericardial involvement.³²

The mechanism is believed to be radiation-induced microvascular injury resulting in increased capillary permeability and the sometimes rapid development of a protein-rich exudate. Associated inflammation may cause acute pericarditis, which may eventually be complicated by chronic pericarditis. The parietal surface tends to be affected more severely than the epicardium.³³

Perhaps as a result of recent advances such as lower radiation doses, equal weighting of the anterior and posterior fields, and subcarinal blocking, incidence rates of pericarditis as low as 2.5% have been reported.³⁴

Pericardial RIHD may be divided into early acute pericarditis, delayed chronic pericardial effusion, and constrictive pericarditis.

Early acute pericarditis is rare and is thought to represent a reaction to tumor necrosis. It is defined as occurring during radiotherapy and occurs almost exclusively with high-dose radiotherapy for lymphoma. Due to the relatively benign course of acute pericarditis and fear of tumor recurrence, it is not an indication to withhold radiotherapy.³⁵

Delayed chronic pericardial effusion occurs months to years after radiotherapy, is typically asymptomatic, and presents as an enlarged cardiac silhouette on chest imaging.³⁵ Delayed pericardial effusion is followed with imaging. While in many cases it resolves within 2 years, it may also be long-standing. Pericardiocentesis or a pericardial window may be performed to treat symptomatic effusion or delayed effusion causing hemodynamic compromise.^35–37 Hypothyroidism should be ruled out, as it can complicate mantle irradiation and result in chronic pericardial effusion.³⁸

Constrictive pericarditis may occur as a late complication of radiotherapy and typically causes symptoms of congestive heart failure. Pericardial stripping in these patients is complicated by the possibility of coexisting RIHD of the valves, myocardium, or coronary arteries, as well as mediastinal fibrosis. A study of 163 patients who underwent pericardial stripping for chronic pericarditis found a 7-year overall survival rate of only 27%, far lower than the rate for those who had no history of radiation exposure.³⁹ Therefore, these patients are often treated for symptom control with diuretics and a low-salt diet rather than with surgery.

MYOCARDIAL DISEASE

Microvascular injury in the myocardium results in chronic ischemia, which may lead to myocardial fibrosis, typically manifesting as diastolic dysfunction. Chest radiotherapy may result in both systolic and diastolic dysfunction, and dilated and restrictive cardiomyopathy are well-recognized complications.⁴⁰

Historically, high radiation doses resulted in systolic dysfunction in more than half of patients who underwent thoracic radiotherapy.⁴¹ Now, however, fewer than 5% of patients develop reductions in left ventricular ejection fraction, and most cases of radiotherapy-induced cardiomyopathy have a restrictive pattern.⁴²

In a single-institution study, diastolic dysfunction was reported in as many as 14% of patients who underwent thoracic radiotherapy for Hodgkin lymphoma.⁴⁰ Systolic dysfunction is now seen almost exclusively in patients  treated concurrently with cardiotoxic chemotherapeutic agents such as anthracyclines in addition to radiotherapy.⁴³

In a childhood cancer survival series, the hazard ratio of congestive heart failure in patients who had undergone radiotherapy for Wilms tumor was 6.6—almost identical to the occurrence in sibling controls. By contrast, the hazard ratio increased to 18.3 in those who received doxorubicin in addition to radiotherapy.⁴⁴

Treatment of radiation-induced cardiomyopathy

Treatment of radiation-induced cardiomyopathy is similar to that for other forms of cardiomyopathy, with an emphasis on symptom management.

Heart transplant may be an option for highly selected patients with end-stage heart failure secondary to RIHD. In one report, a series of four RIHD patients received a heart transplant, and all four survived past 48 months.⁴⁵ However, data from the United Network of Organ Sharing revealed an increase in the all-cause mortality rate in patients undergoing heart transplant for RIHD compared with those undergoing transplant for cardiomyopathy due to other causes.⁴⁶ This trend may be confounded by a higher prevalence of prior cardiac surgery in the RIHD group—itself an established risk factor for poor posttransplant outcomes.

CONDUCTION SYSTEM DISEASE

Life-threatening arrhythmias have been reported that are distinct from the common, asymptomatic repolarization abnormalities that occur during radiotherapy. Atrioventricular nodal bradycardia, all degrees of heart block, and sick sinus syndrome have all been reported after chest radiotherapy. As conduction abnormalities do not typically manifest until years after radiotherapy, it is difficult to establish causation and, consequently, to define incidence.

Right bundle branch block is the most common conduction abnormality because of the proximity of the right bundle to the endocardium on the right side.⁴⁷

Chest radiotherapy is also associated with prolongation of the corrected QT interval (QTc). A study in patients with a history of thoracic radiotherapy found that the QTc characteristically increased with exercise, a poor prognostic indicator.⁴⁸ In a study of 134 survivors of childhood cancer, 12.5% of those who had undergone radiotherapy had a resting QTc of 0.44 msec or more.⁴⁹

Furthermore, a study of 69 breast cancer survivors found a higher incidence of conduction abnormalities at 6 months and 10 years after radiotherapy compared with baseline. The characteristic electrocardiographic changes at 6 months were T-wave changes. At 10 years, the T-wave abnormalities had resolved and were replaced by ST depression.⁵⁰

As mentioned above, establishing radiotherapy as a cause for these conduction abnormalities is challenging, given the lag between radiation therapy and electrocardiographic changes. The following criteria have been proposed for establishing a link between atrioventricular blockade and prior radiation⁵¹:

• Total radiation dose to the heart > 40 Gy
• Delay of 10 years or more since therapy
• Abnormal interval electrocardiographic changes such as bundle branch block
• Prior pericardial involvement
• Associated cardiac or mediastinal lesions.

SCREENING GUIDELINES

Consensus guidelines for identifying and monitoring RIHD have been published by the European Association of Cardiovascular Imaging and the American Society of Echocardiography (Table 2).¹⁰ The European Society of Medical Oncology has also issued guidelines for the prevention, diagnosis, and management of cardiovascular disease associated with cancer therapy.

Briefly, the guidelines call for aggressive cardiac risk-factor modification through weight loss, exercise, blood pressure control, and smoking cessation, in addition to early detection of RIHD. Cardiovascular screening for risk factors and a careful clinical examination should be performed in all patients. Baseline comprehensive transthoracic echocardiography is advocated in all patients before starting radiotherapy to detect cardiac anomalies. Beyond this, an annual history and physical examination, paying close attention to the signs and symptoms of cardiopulmonary disease, is essential. The development of new cardiopulmonary symptoms or a new physical finding such as a murmur should prompt evaluation with transthoracic echocardiography.

In patients without symptoms, screening transthoracic echocardiography at 10 years after the start of radiotherapy is recommended in light of the high probability of diagnosing cardiac disease at this juncture. In patients with no preexisting cardiac disease, surveillance transthoracic echocardiography should be at 5-year intervals thereafter.

In high-risk patients without symptoms (those who have undergone anterior or left-sided radiotherapy and have at least one risk factor for RIHD), initial screening transthoracic echocardiography is recommended 5 years after radiotherapy. These patients have a heightened risk of coronary events as described above and, consequently, are recommended to undergo noninvasive imaging 5 to 10 years after radiation exposure. If this initial examination is negative, stress testing should be repeated at 5-year intervals. Stress echocardiography and stress cardiac magnetic resonance imaging have higher specificity than stress electrocardiography and therefore are generally preferred. Stress scintigraphy should be used with caution, as it adds to the cumulative radiation exposure.

The role of magnetic resonance imaging and computed tomography depends on the results of initial transthoracic echocardiography and the clinical indication, in addition to the center’s expertise and facilities. However, there are currently no data advocating their use as screening tools, except for early detection of porcelain aorta in high-risk patients.¹⁰

MODERN RADIOTHERAPY TECHNIQUES

In recent years, there has been emphasis on exposing the patient to as little radiation as possible without compromising cure.⁵² The three major strategies employed to decrease cardiac exposure include reducing the radiation dose, reducing the radiation field and volume, and using newer planning and delivery techniques.

Reducing the radiation dose. It is well recognized that the mean dose of radiation to the heart is a significant predictor of cardiovascular disease, with one study demonstrating a linear increase in the risk of coronary artery disease with increasing mean heart radiation dose (excess relative risk per Gy 7.4%, 95% confidence interval 3.3%–14.8%).⁵³

Reducing the radiation field and volume. Modern strategies and computed tomography-based radiotherapy planning have enabled a transition from older techniques such as extended-field radiation therapy, mantle-field radiation therapy, and involved-field radiation therapy to new techniques such as involved-node and involved-site radiation therapy.⁵⁴ These have shown promise. For instance, a study in patients with early Hodgkin lymphoma found a mean heart dose of 27.5 Gy with mantle-field therapy compared with 7.7 Gy with involved-node therapy. This decrease in mean heart dose was associated with a reduction in the 25-year absolute excess cardiac risk from 9.1% to 1.4% and a reduction in cardiac mortality from 2.1% to 1%.⁵⁵

Employing newer planning and delivery systems has also demonstrated some promise in reducing rates of cardiac morbidity and mortality. Extended-field radiation therapy, mantle-field radiotherapy, and involved-field radiation therapy were traditionally based on two-dimensional planning and often resulted in large volumes of myocardium being unnecessarily exposed to large doses of radiation because of the uncertainty in targeting. Involved-site and involved-node radiotherapy are based on computed tomography, resulting in more accurate targeting and sparing of normal tissue.

In addition, newer techniques such as intensity-modulated radiotherapy and proton beam therapy have resulted in further improvements in conformality compared with three-dimensional conformal radiotherapy.^56,57 Respiratory motion management, including deep inspiration breath-holding and end-inspiration breath-holding, have decreased the radiation dose to the heart in patients undergoing mediastinal radiotherapy.^58,59

TOWARD THE GOALS OF PREVENTION AND EARLIER DETECTION

As survival from breast cancer and lymphoma has increased, we continue to see legacy or latent effects of therapy, such as RIHD. Radiation therapy can affect any cardiac structure and is a major cause of morbidity and death in cancer survivors.

Modern radiation techniques use a variety of mechanisms to decrease the radiation dose to the heart. A large body of evidence emanating from an era of higher radiation doses and a lack of knowledge of the cardiac effects of radiation highlight the perilous cardiac consequences of chest radiation. With advances in radiotherapy and the development and widespread implementation of consensus guidelines, we envision earlier detection and less frequent occurrence of RIHD, although the latter trend could be blunted by increased cardiovascular risk factors within the population. Given the lag between irradiation and the cardiac consequences, it may be a number of years before any comparisons can be drawn.

Advances in radiotherapy over the past 50 years have dramatically improved outcomes in patients with malignancy. Five-year overall survival rates for Hodgkin lymphoma and non-Hodgkin lymphoma now stand at 80%, and breast cancer survival is 90%.¹

Increased longevity, however, has come at the cost of late side effects such as radiation-induced heart disease (RIHD). Cardiac dysfunction due to radiation involves a spectrum of disease processes in patients who have undergone mediastinal, thoracic, or breast radiotherapy and may involve any cardiac structure, including the pericardium, myocardium, valves, conduction system, and coronary arteries.

Overall, compared with nonirradiated patients, patients who have undergone chest radiotherapy have a 2% higher absolute risk of cardiac morbidity and death at 5 years and a 23% increased absolute risk after 20 years.²

This article will review the pathophysiology and epidemiology of RIHD and will offer a practical approach to its diagnosis and management.

MOST DAMAGE IS ENDOTHELIAL 

Cardiac myocytes are relatively resistant to radiation damage because of their postmitotic state. But endothelial cells remain sensitive to radiation, and the pathophysiology of most forms of RIHD appears to be associated with damage to endothelial cells. Conventional cardiac risk factors such as hyperlipidemia and smoking have been shown to compound and accelerate radiation-induced endothelial damage in animal models.³

Radiation is believed to result in transient increases in oxidative stress, resulting in formation of reactive oxygen species and a subsequent inflammatory response that includes activation of nuclear factor-kappa B. Upregulation of proinflammatory pathways results in increased expression of matrix metalloproteinases, adhesion molecules, and proinflammatory cytokines and downregulation of vasculoprotective nitric oxide.⁴ Indirect evidence for radiation-induced vascular inflammation comes from numerous studies that demonstrated increased levels of the proinflammatory cytokines interleukin 6, tumor necrosis factor alpha, and interferon gamma in Japanese atomic bomb survivors.⁵

RISK FACTORS

Risk factors for RIHD are summarized in Table 1.

The volume of heart irradiated is a major determinant of the development of RIHD.⁶ A retrospective study of 960 breast cancer patients in Stockholm between 1971 and 1976 found that those who had received the highest doses and volumes of cardiac radiation had a threefold higher risk of cardiac death. By comparison, those with lesser volumes of the heart exposed to radiation had no increase in risk of cardiac death compared with the general population.⁷

Younger age at the time of radiotherapy is associated with an increased risk of RIHD in breast cancer and lymphoma patients. A retrospective analysis of 635 patients under age 21 with Hodgkin lymphoma treated with radiotherapy showed a relative risk of fatal myocardial infarction of 41.5 compared with a general population matched for age, sex, and race.⁸

Conventional cardiac risk factors such as smoking, hypertension, diabetes, and hyperlipidemia further increase the risk of RIHD, and radiation increases the cardiotoxicity of chemotherapeutic agents such as anthracyclines.⁹

In general, high-risk patients are defined as those with at least one risk factor for RIHD who underwent anterior or left-sided chest irradiation  (Table 1).¹⁰

CORONARY ARTERY DISEASE

Ischemic heart disease is the most common cause of cardiac death in patients who have undergone radiation therapy. Atherosclerotic lesions in RIHD are morphologically identical to those in nonirradiated vessels and are characterized by intimal proliferation, accumulation of lipid-rich macrophages, and plaque formation.¹¹

A retrospective single-institution study of 415 patients with Hodgkin lymphoma who had undergone radiation therapy found the incidence of coronary artery disease 20 years later to be 10%. The mean time to development of coronary artery disease was 9 years, and all patients who developed it had at least one conventional cardiac risk factor.¹²

A meta-analysis of more than 20,000 breast cancer patients who received radiotherapy in 40 randomized controlled trials found an increase in the rate of non-breast-cancer deaths, primarily from vascular causes (annual event ratio 1.27, P <  .0001).¹³

A randomized controlled trial comparing breast cancer patients who underwent preoperative or postoperative radiotherapy vs those who had surgery alone revealed a significantly higher death rate from coronary artery disease in the postradiotherapy group.⁷

The risk of radiation-induced coronary artery disease is proportional to both the dose and the duration of radiation therapy. A retrospective study of more than 2,000 women undergoing radiotherapy for breast cancer found that the relative risk of coronary artery disease increased linearly by about 7.4% per Gy of radiation to the heart, with no apparent ceiling.¹⁴

The distribution of atherosclerotic coronary arteries correlates well with the areas exposed to the highest doses of radiation. For instance, in left-sided breast cancer, the apex and anterior wall of the heart typically receive the highest doses of radiation; consequently, the left anterior descending and distal diagonal branches are most prominently involved.¹⁵ In patients with lymphoma who undergo radiotherapy to mediastinal nodes and in breast cancer patients receiving radiotherapy to the internal mammary chain, basal structures may be exposed as well. Ostial lesions can also be seen in these patients.¹⁶

The clinical presentation of coronary artery disease in radiotherapy recipients does not differ significantly from that in the general population. Ischemia may be silent, may lead to classic anginal symptoms, or may cause sudden cardiac death. The incidence of silent myocardial infarction has been reported to be higher after mediastinal radiotherapy than it is in the general population, possibly from damage to nerve endings within the radiation field.¹⁷

Management of radiation-associated coronary artery disease

Managing patients with radiation-associated coronary artery disease is challenging, but the therapeutic options remain the same as those in nonirradiated patients and include medical therapy, percutaneous coronary intervention, and coronary artery bypass grafting, depending on the site and extent of disease.¹⁸ Although results are conflicting, there does not seem to be a significant difference in the rates of stent restenosis between patients with a history of radiation therapy and the general population.

Percutaneous coronary intervention is generally preferred to coronary artery bypass grafting in these patients for several reasons. Radiation-induced fibrosis of surrounding structures generally makes surgical procedures more difficult,¹⁹ and inclusion of the internal mammary artery or internal thoracic artery in the radiation field may result in stenosis of these vessels, rendering them unsuitable for harvesting.²⁰ Moreover, many patients with RIHD have concurrent radiation-induced lung damage, which increases the risk of perioperative pulmonary complications.²¹

If the coronary lesions are not amenable to percutaneous intervention, a careful valvular evaluation should be performed preoperatively in view of the frequency of radiation-associated valvular disease. In a study of 72 patients with RIHD undergoing coronary artery bypass grafting, 40% required valvular surgery at the time of surgery or shortly thereafter.²²

Results of studies of coronary artery bypass graft outcomes in patients with a history of thoracic radiation therapy have been conflicting, but success seems to depend on the status of the internal mammary and internal thoracic arteries.²³ Therefore, the patency of these vessels should be elucidated preoperatively by angiography and intraoperatively by visual inspection of the vessels for fibrosis.

A large single-institution study by Wu et al²⁴ revealed higher short-term and long-term mortality rates in patients with RIHD undergoing cardiac surgery than in control patients without RIHD undergoing similar procedures.

VALVULAR DISEASE

Radiation therapy may directly affect heart valves, and both stenotic (Figure 1) and regurgitant lesions have been described. Pathologic findings include leaflet retraction, fibrotic thickening, and late calcification.²⁵

The precise mechanism of radiation-induced valvular disease is unknown but is thought to be a change in the phenotype of valvular interstitial cells from a myofibroblast to an osteoblast-like cell. Radiation results in significant expression of osteogenic factors such as bone morphogenic protein 2, osteopontin, alkaline phosphatase, and runt-related transcription factor 2 by valvular interstitial cells.26

Valvular heart disease is evident in as many as 81% of patients with RIHD, with the aortic and mitral valves affected more commonly than the tricuspid and pulmonic valves.²⁷ Why there are more left-sided valve lesions than pulmonic valve lesions, despite the pulmonic valve’s anterior position in the heart, is unknown but may be due to higher pressures across the left-sided heart valves.

Although valvular disease is common in patients with RIHD, clinically significant disease is not; more than 70% of patients with radiation-induced valvular disease have no symptoms. A study of 38 cases of radiation-induced valvular disease reported a mean time to development of asymptomatic valvular lesions of 11.5 years and an average time to symptomatic valvular dysfunction of 16.5 years, indicating that 5 years seems to be the interval required for progression from asymptomatic to symptomatic valvular RIHD.²⁸

The thickness of the aortomitral curtain (the junction between the base of the anterior mitral leaflet and the aortic root) is an independent predictor of the long-term risk of death in patients with valvular RIHD.²⁹

Management of radiation-induced valvular disease

Management of patients with valvular RIHD poses a major clinical conundrum because of  the high rates of perioperative morbidity and death in patients with a history of chest radiotherapy. In one study,²³ the long-term mortality rate was 45% in postradiotherapy patients undergoing single-valve surgery and 61% in those undergoing surgery on two or more valves, compared with 13% and 17% in patients with no history of chest radiotherapy.²³

Furthermore, valve repair is an unattractive option in these patients because of high failure rates of mitral valve and tricuspid valve repair attributed to ongoing radiotherapy-induced valvular changes after repair.³⁰

As a result, valve replacement is generally preferred in this group. Patients should be advised of the higher risk of perioperative and long-term morbidity and death associated with open heart surgery than in the general population, and that the risks are even higher with repeat open heart surgery.

This risk has implications for the choice of replacement valves in younger patients. Bioprosthetic valves, which deteriorate over time, may not be advisable. Transcatheter aortic valve replacement has been successful in radiation-induced valvular disease and may become the preferred method of aortic valve replacement.³¹

PERICARDIAL DISEASE

Pericardial disease is a frequent manifestation of RIHD and covers a spectrum of manifestations from acute pericarditis, pericardial effusion, and tamponade to constrictive pericarditis. In a necropsy study, 70% of patients with RIHD were found to have pericardial involvement.³²

The mechanism is believed to be radiation-induced microvascular injury resulting in increased capillary permeability and the sometimes rapid development of a protein-rich exudate. Associated inflammation may cause acute pericarditis, which may eventually be complicated by chronic pericarditis. The parietal surface tends to be affected more severely than the epicardium.³³

Perhaps as a result of recent advances such as lower radiation doses, equal weighting of the anterior and posterior fields, and subcarinal blocking, incidence rates of pericarditis as low as 2.5% have been reported.³⁴

Pericardial RIHD may be divided into early acute pericarditis, delayed chronic pericardial effusion, and constrictive pericarditis.

Early acute pericarditis is rare and is thought to represent a reaction to tumor necrosis. It is defined as occurring during radiotherapy and occurs almost exclusively with high-dose radiotherapy for lymphoma. Due to the relatively benign course of acute pericarditis and fear of tumor recurrence, it is not an indication to withhold radiotherapy.³⁵

Delayed chronic pericardial effusion occurs months to years after radiotherapy, is typically asymptomatic, and presents as an enlarged cardiac silhouette on chest imaging.³⁵ Delayed pericardial effusion is followed with imaging. While in many cases it resolves within 2 years, it may also be long-standing. Pericardiocentesis or a pericardial window may be performed to treat symptomatic effusion or delayed effusion causing hemodynamic compromise.^35–37 Hypothyroidism should be ruled out, as it can complicate mantle irradiation and result in chronic pericardial effusion.³⁸

Constrictive pericarditis may occur as a late complication of radiotherapy and typically causes symptoms of congestive heart failure. Pericardial stripping in these patients is complicated by the possibility of coexisting RIHD of the valves, myocardium, or coronary arteries, as well as mediastinal fibrosis. A study of 163 patients who underwent pericardial stripping for chronic pericarditis found a 7-year overall survival rate of only 27%, far lower than the rate for those who had no history of radiation exposure.³⁹ Therefore, these patients are often treated for symptom control with diuretics and a low-salt diet rather than with surgery.

MYOCARDIAL DISEASE

Microvascular injury in the myocardium results in chronic ischemia, which may lead to myocardial fibrosis, typically manifesting as diastolic dysfunction. Chest radiotherapy may result in both systolic and diastolic dysfunction, and dilated and restrictive cardiomyopathy are well-recognized complications.⁴⁰

Historically, high radiation doses resulted in systolic dysfunction in more than half of patients who underwent thoracic radiotherapy.⁴¹ Now, however, fewer than 5% of patients develop reductions in left ventricular ejection fraction, and most cases of radiotherapy-induced cardiomyopathy have a restrictive pattern.⁴²

In a single-institution study, diastolic dysfunction was reported in as many as 14% of patients who underwent thoracic radiotherapy for Hodgkin lymphoma.⁴⁰ Systolic dysfunction is now seen almost exclusively in patients  treated concurrently with cardiotoxic chemotherapeutic agents such as anthracyclines in addition to radiotherapy.⁴³

In a childhood cancer survival series, the hazard ratio of congestive heart failure in patients who had undergone radiotherapy for Wilms tumor was 6.6—almost identical to the occurrence in sibling controls. By contrast, the hazard ratio increased to 18.3 in those who received doxorubicin in addition to radiotherapy.⁴⁴

Treatment of radiation-induced cardiomyopathy

Treatment of radiation-induced cardiomyopathy is similar to that for other forms of cardiomyopathy, with an emphasis on symptom management.

Heart transplant may be an option for highly selected patients with end-stage heart failure secondary to RIHD. In one report, a series of four RIHD patients received a heart transplant, and all four survived past 48 months.⁴⁵ However, data from the United Network of Organ Sharing revealed an increase in the all-cause mortality rate in patients undergoing heart transplant for RIHD compared with those undergoing transplant for cardiomyopathy due to other causes.⁴⁶ This trend may be confounded by a higher prevalence of prior cardiac surgery in the RIHD group—itself an established risk factor for poor posttransplant outcomes.

CONDUCTION SYSTEM DISEASE

Life-threatening arrhythmias have been reported that are distinct from the common, asymptomatic repolarization abnormalities that occur during radiotherapy. Atrioventricular nodal bradycardia, all degrees of heart block, and sick sinus syndrome have all been reported after chest radiotherapy. As conduction abnormalities do not typically manifest until years after radiotherapy, it is difficult to establish causation and, consequently, to define incidence.

Right bundle branch block is the most common conduction abnormality because of the proximity of the right bundle to the endocardium on the right side.⁴⁷

Chest radiotherapy is also associated with prolongation of the corrected QT interval (QTc). A study in patients with a history of thoracic radiotherapy found that the QTc characteristically increased with exercise, a poor prognostic indicator.⁴⁸ In a study of 134 survivors of childhood cancer, 12.5% of those who had undergone radiotherapy had a resting QTc of 0.44 msec or more.⁴⁹

Furthermore, a study of 69 breast cancer survivors found a higher incidence of conduction abnormalities at 6 months and 10 years after radiotherapy compared with baseline. The characteristic electrocardiographic changes at 6 months were T-wave changes. At 10 years, the T-wave abnormalities had resolved and were replaced by ST depression.⁵⁰

As mentioned above, establishing radiotherapy as a cause for these conduction abnormalities is challenging, given the lag between radiation therapy and electrocardiographic changes. The following criteria have been proposed for establishing a link between atrioventricular blockade and prior radiation⁵¹:

• Total radiation dose to the heart > 40 Gy
• Delay of 10 years or more since therapy
• Abnormal interval electrocardiographic changes such as bundle branch block
• Prior pericardial involvement
• Associated cardiac or mediastinal lesions.

SCREENING GUIDELINES

Consensus guidelines for identifying and monitoring RIHD have been published by the European Association of Cardiovascular Imaging and the American Society of Echocardiography (Table 2).¹⁰ The European Society of Medical Oncology has also issued guidelines for the prevention, diagnosis, and management of cardiovascular disease associated with cancer therapy.

Briefly, the guidelines call for aggressive cardiac risk-factor modification through weight loss, exercise, blood pressure control, and smoking cessation, in addition to early detection of RIHD. Cardiovascular screening for risk factors and a careful clinical examination should be performed in all patients. Baseline comprehensive transthoracic echocardiography is advocated in all patients before starting radiotherapy to detect cardiac anomalies. Beyond this, an annual history and physical examination, paying close attention to the signs and symptoms of cardiopulmonary disease, is essential. The development of new cardiopulmonary symptoms or a new physical finding such as a murmur should prompt evaluation with transthoracic echocardiography.

In patients without symptoms, screening transthoracic echocardiography at 10 years after the start of radiotherapy is recommended in light of the high probability of diagnosing cardiac disease at this juncture. In patients with no preexisting cardiac disease, surveillance transthoracic echocardiography should be at 5-year intervals thereafter.

In high-risk patients without symptoms (those who have undergone anterior or left-sided radiotherapy and have at least one risk factor for RIHD), initial screening transthoracic echocardiography is recommended 5 years after radiotherapy. These patients have a heightened risk of coronary events as described above and, consequently, are recommended to undergo noninvasive imaging 5 to 10 years after radiation exposure. If this initial examination is negative, stress testing should be repeated at 5-year intervals. Stress echocardiography and stress cardiac magnetic resonance imaging have higher specificity than stress electrocardiography and therefore are generally preferred. Stress scintigraphy should be used with caution, as it adds to the cumulative radiation exposure.

The role of magnetic resonance imaging and computed tomography depends on the results of initial transthoracic echocardiography and the clinical indication, in addition to the center’s expertise and facilities. However, there are currently no data advocating their use as screening tools, except for early detection of porcelain aorta in high-risk patients.¹⁰

MODERN RADIOTHERAPY TECHNIQUES

In recent years, there has been emphasis on exposing the patient to as little radiation as possible without compromising cure.⁵² The three major strategies employed to decrease cardiac exposure include reducing the radiation dose, reducing the radiation field and volume, and using newer planning and delivery techniques.

Reducing the radiation dose. It is well recognized that the mean dose of radiation to the heart is a significant predictor of cardiovascular disease, with one study demonstrating a linear increase in the risk of coronary artery disease with increasing mean heart radiation dose (excess relative risk per Gy 7.4%, 95% confidence interval 3.3%–14.8%).⁵³

Reducing the radiation field and volume. Modern strategies and computed tomography-based radiotherapy planning have enabled a transition from older techniques such as extended-field radiation therapy, mantle-field radiation therapy, and involved-field radiation therapy to new techniques such as involved-node and involved-site radiation therapy.⁵⁴ These have shown promise. For instance, a study in patients with early Hodgkin lymphoma found a mean heart dose of 27.5 Gy with mantle-field therapy compared with 7.7 Gy with involved-node therapy. This decrease in mean heart dose was associated with a reduction in the 25-year absolute excess cardiac risk from 9.1% to 1.4% and a reduction in cardiac mortality from 2.1% to 1%.⁵⁵

Employing newer planning and delivery systems has also demonstrated some promise in reducing rates of cardiac morbidity and mortality. Extended-field radiation therapy, mantle-field radiotherapy, and involved-field radiation therapy were traditionally based on two-dimensional planning and often resulted in large volumes of myocardium being unnecessarily exposed to large doses of radiation because of the uncertainty in targeting. Involved-site and involved-node radiotherapy are based on computed tomography, resulting in more accurate targeting and sparing of normal tissue.

In addition, newer techniques such as intensity-modulated radiotherapy and proton beam therapy have resulted in further improvements in conformality compared with three-dimensional conformal radiotherapy.^56,57 Respiratory motion management, including deep inspiration breath-holding and end-inspiration breath-holding, have decreased the radiation dose to the heart in patients undergoing mediastinal radiotherapy.^58,59

TOWARD THE GOALS OF PREVENTION AND EARLIER DETECTION

As survival from breast cancer and lymphoma has increased, we continue to see legacy or latent effects of therapy, such as RIHD. Radiation therapy can affect any cardiac structure and is a major cause of morbidity and death in cancer survivors.

Modern radiation techniques use a variety of mechanisms to decrease the radiation dose to the heart. A large body of evidence emanating from an era of higher radiation doses and a lack of knowledge of the cardiac effects of radiation highlight the perilous cardiac consequences of chest radiation. With advances in radiotherapy and the development and widespread implementation of consensus guidelines, we envision earlier detection and less frequent occurrence of RIHD, although the latter trend could be blunted by increased cardiovascular risk factors within the population. Given the lag between irradiation and the cardiac consequences, it may be a number of years before any comparisons can be drawn.

Aortic stenosis is the most common valvular heart condition in the developed world, affecting 3% of people between ages 75 and 85¹ and 4% of people over age 85.² Aortic valve replacement remains the only treatment proven to reduce the rates of mortality and morbidity in this condition.³ Under current guidelines,^4,5 the onset of symptoms of exertional angina, syncope, or dyspnea in a patient who has severe aortic stenosis is a class I indication for surgery—ie, surgery should be performed.

However, high-gradient, severe aortic stenosis that is asymptomatic often poses a dilemma. The annual rate of sudden death in patients with this condition is estimated at 1% to 3%,^6–9 but the surgical mortality rate in aortic valve replacement has been as high as 6% in Medicare patients (varying by center and comorbidities).¹⁰ Therefore, the traditional teaching was to not surgically replace the valve in asymptomatic patients, based on an adverse risk-benefit ratio. But with improvements in surgical techniques and prostheses, these rates have been reduced to 2.41% at high-volume centers¹¹ (and to less than 1% at some hospitals),¹² arguing in favor of earlier intervention.

Complicating the issue, transcatheter aortic valve replacement has become widely available, but further investigation into its use in this patient cohort is warranted.

Furthermore, many patients with severe but apparently asymptomatic aortic stenosis and normal left ventricular ejection fraction may actually have impaired exercise capacity, or they may have structural left ventricular changes such as severe hypertrophy or reduction in global strain, which may worsen the long-term survival rate.^13,14

A prospective trial in patients with severe aortic stenosis found that mortality rates were significantly lower in those who underwent surgery early than in those who received conventional treatment, ie, watchful waiting (no specific medical treatment for aortic stenosis is available).¹⁵

Patients with asymptomatic severe aortic stenosis are a diverse group; some have a far worse prognosis than others, with or without surgery.

This paper reviews the guidelines for valve replacement in this patient group and the factors useful in establishing who should be considered for early intervention even if they have no classic symptoms (Figure 1).

SIGNS AND SYMPTOMS OF STENOSIS

Aortic stenosis is often first suspected when a patient presents with angina, dyspnea, and syncope, or when an ejection systolic murmur is heard incidentally on physical examination—typically a high-pitched, crescendo-decrescendo, midsystolic ejection murmur that is best heard at the right upper sternal border and that radiates to the carotid arteries.

Several physical findings may help in assessing the severity of aortic stenosis. In mild stenosis, the murmur peaks in early systole, but as the disease progresses the peak moves later into systole. The corollary of this phenomenon is a weak and delayed carotid upstroke known as “pulsus parvus et tardus.” This can be assessed by palpating the carotid artery while auscultating the heart.

Aortic stenosis is often first suspected when a patient has angina, dyspnea, and syncope or an ejection systolic murmur

The second heart sound becomes progressively softer as the stenosis advances until it is no longer audible. If a fourth heart sound is present, it may be due to concentric left ventricular hypertrophy with reduced left ventricular compliance, and a third heart sound indicates severe left ventricular dysfunction. Both of these findings suggest severe aortic stenosis.

ECHOCARDIOGRAPHIC MEASURES OF SEVERITY

Echocardiography is the best established and most important initial investigation in the assessment of a patient with suspected aortic stenosis. It usually provides accurate information on the severity and the mechanism of stenosis. The following findings indicate severe aortic stenosis:

• Mean pressure gradient > 40 mm Hg
• Peak aortic jet velocity > 4.0 m/s
• Aortic valve area < 1 cm².

RECOMMENDATIONS FOR SURGERY BASED ON SEVERITY AND SYMPTOMS

The American College of Cardiology and American Heart Association (ACC/AHA)⁴ have issued the following recommendations for aortic valve replacement, based on the severity of stenosis and on whether the patient has symptoms (Figure 2):

Severe stenosis, with symptoms: class I recommendation (surgery should be done). Without surgery, these patients have a very poor prognosis, with an overall mortality rate of 75% at 3 years.³

Severe stenosis, no symptoms, in patients undergoing cardiac surgery for another indication (eg, coronary artery bypass grafting, ascending aortic surgery, or surgery on other valves): class I recommendation for concomitant aortic valve replacement.

Moderate stenosis, no symptoms, in patients undergoing cardiac surgery for another indication: class IIa recommendation (ie, aortic valve replacement “is reasonable”).

Very severe stenosis (aortic peak velocity > 5.0 m/s or mean pressure gradient ≥ 60 mm Hg), no symptoms, and low risk of death during surgery: class IIa recommendation.

Severe stenosis, no symptoms, and an increase in transaortic velocity of 0.3 m/s or more per year on serial testing or in patients considered to be at high risk for rapid disease progression, such as elderly patients with severe calcification: class IIb recommendation (surgery “can be considered”). The threshold to replace the valve is lower for patients who cannot make serial follow-up appointments because they live far away or lack transportation, or because they have problems with compliance.

Surgery for those with left ventricular dysfunction

Echocardiography also provides information on left ventricular function, and patients with left ventricular dysfunction have significantly worse outcomes. Studies have shown substantial differences in survival in patients who had an ejection fraction of less than 50% before valve replacement compared with those with a normal ejection fraction.³

Thus, the ACC/AHA guidelines recommend immediate referral for aortic valve replacement in asymptomatic patients whose left ventricular ejection fraction is less than 50% (class I recommendation, level of evidence B) in the hope of preventing irreversible ventricular dysfunction.⁴

TREADMILL EXERCISE TESTING UNMASKS SYMPTOMS

Treadmill testing is absolutely contraindicated in patients with severe symptomatic aortic stenosis

In the past, severe aortic stenosis was considered a contraindication to stress testing because of concerns of precipitating severe, life-threatening complications. However, studies over the past 10 years have shown that a supervised modified Bruce protocol is safe in patients with severe asymptomatic aortic stenosis.^16,17

However, treadmill exercise testing clearly is absolutely contraindicated in patients with severe symptomatic aortic stenosis because of the risk of syncope or of precipitating a malignant arrhythmia. Nevertheless, it may play an essential role in the workup of a physically active patient with no symptoms.

Symptoms can develop insidiously in patients with chronic valve disease and may often go unrecognized by patients and their physicians. Many patients who state they have no symptoms may actually be subconsciously limiting their exercise to avoid symptoms.

Amato et al¹³ examined the exercise capacity of 66 patients reported to have severe asymptomatic aortic stenosis. Treadmill exercise testing was considered positive in this study if the patient developed symptoms or complex ventricular arrhythmias, had blood pressure that failed to rise by 20 mm Hg, or developed horizontal or down-sloping ST depression (≥ 1 mm in men, ≥ 2 mm in women). Twenty (30.3%) of the 66 patients developed symptoms during exercise testing, and they had a significantly worse prognosis: the 2-year event-free survival rate was only 19% in those with a positive test compared with 85% in those with a negative test.¹³ This study highlights the problem of patients subconsciously reducing their level of activity, thereby masking their true symptoms.

A meta-analysis by Rafique et al¹⁸ found that asymptomatic patients with abnormal results on exercise testing had a risk of cardiac events during follow-up that was eight times higher than normal, and a risk of sudden death 5.5 times higher.

With trials demonstrating that exercise testing is safe and prognostically useful in patients with aortic stenosis, the ACC/AHA guidelines emphasize its role, giving a class I recommendation for aortic valve replacement in patients who develop symptoms on exercise testing, and a class IIa recommendation in asymp­tomatic patients with decreased exercise tolerance or an exercise-related fall in blood pressure (Figure 2).⁴

STRESS ECHOCARDIOGRAPHY

Stress echocardiography has been used since the 1980s to assess the hemodynamic consequences of valvular heart disease, and many studies highlight its prognostic usefulness in patients with asymptomatic aortic stenosis.

In a 2005 study by Lancellotti et al,¹⁹ 69 patients with severe asymptomatic aortic stenosis underwent a symptom-limited bicycle exercise stress test using quantitative Doppler echocardiography both at rest and at peak exercise, and a number of independent predictors of poor outcome (ie, symptoms, aortic valve replacement, death) were identified. These predictors included an abnormal test result, defined as any of the following: angina, dyspnea, ST-segment depression of 2 mm Hg or more, a fall or a small (< 20 mm Hg) rise in systolic blood pressure during the test, an aortic valve area of 0.75 cm² or less, or a mean increase in valve gradient of 18 mm Hg or more.

Subsequently, a multicenter prospective trial assessed the value of exercise stress echocardiography in 186 patients with asymptomatic moderate or severe aortic stenosis.²⁰ A mean increase in the aortic valve gradient of 20 mm Hg or more after exercise was associated with a rate of cardiovascular events (death, aortic valve replacement) 3.8 times higher, independent of other risk factors and whether moderate or severe stenosis was present (Table 1).²⁰

Exercise-induced changes in systolic pulmonary artery pressure, which can be assessed using stress echocardiography, also have prognostic utility. Elevated systolic pulmonary artery pressure (> 50 mm Hg) seems to portend a poorer prognosis^21,22 and a higher mortality rate after valve replacement,²³ making it an independent predictor of hospital mortality and postoperative major adverse cardiovascular and cerebrovascular events (Table 1).

Exercise echocardiography also can be used to assess the patient’s contractile reserve. Left ventricular contractile reserve can be defined as an exercise-induced increase in left ventricular ejection fraction. In a study by Maréchaux et al²⁴ in 50 patients with asymptomatic aortic stenosis and a normal resting left ventricular ejection fraction (> 50%), 40% of patients did not have left ventricular contractile reserve. In fact, their left ventricular ejection fraction decreased with exercise (from 64 ± 10% to 53 ± 12%). The subgroup of patients without contractile reserve developed symptoms more frequently during exercise and had lower event-free survival (Table 1).

Stress echocardiography has recently been introduced into the European Society of Cardiology guidelines, which give a class IIb indication for aortic valve replacement in asymp­tomatic patients who have severe aortic stenosis, a normal ejection fraction, and a greater than 20-mm Hg increase in mean gradient on exercise.⁵ But it has yet to be introduced into the ACC/AHA guidelines as a consideration for surgery.

LEFT VENTRICULAR FUNCTION: BEYOND EJECTION FRACTION

Left ventricular dysfunction is a bad sign for patients with aortic stenosis. Struggling to empty its contents through the narrowed aortic valve, the left ventricle is subjected to increased wall stress and eventually develops hypertrophy. The hypertrophied heart muscle requires more oxygen but receives less perfusion. Eventually, myocardial fibrosis develops, leading to systolic dysfunction and a reduction in the ejection fraction. As described above, patients with asymptomatic aortic stenosis and a left ventricular ejection fraction less than 50% have a poor prognosis,¹⁴ and therefore the ACC/AHA guidelines give this condition a class I recommendation for surgery.⁴

The ejection fraction has limitations as a marker of left ventricular function

However, the ejection fraction has limitations as a marker of left ventricular function. It reflects changes in left ventricular cavity volume but not in the complex structure of the left ventricle. Several studies show that up to one-third of patients with severe aortic stenosis have considerable impairment of intrinsic myocardial systolic function despite a preserved ejection fraction.^8,25,26

Thus, other variables such as left atrial size, left ventricular hypertrophy, myocardial deformation (assessed using strain imaging), and B-type natriuretic peptide (BNP) level may also be considered in assessing the effect of severe aortic stenosis on left ventricular function in the context of a normal ejection fraction (Table 2).

Left ventricular hypertrophy

The development of left ventricular hypertrophy is one of the earliest compensatory responses of the ventricle to the increase in afterload. This leads to impaired myocardial relaxation and reduced myocardial compliance, with resultant diastolic dysfunction with increased filling pressures.

Cioffi et al,²⁷ in a study in 209 patients with severe but asymptomatic aortic stenosis, found that inappropriately high left ventricular mass (> 110% of that expected for body size, sex, and wall stress) portended a 4.5-times higher risk of death, independent of other risk factors.

Severe left ventricular hypertrophy may have a long-term effect on prognosis irrespective of valve replacement. An observational study¹⁴ of 3,049 patients who underwent aortic valve replacement for severe aortic stenosis showed that the 10-year survival rate was 45% in those whose left ventricular mass was greater than 185 g/m², compared with 65% in patients whose left ventricular mass was less than 100 g/m².

Thus, as surgical mortality and morbidity rates decrease, the impact of these structural changes in left ventricular wall thickness may affect the decision to intervene earlier in order to improve longer-term outcomes in select asymptomatic patients with high-risk features.

Left atrial size

Diastolic dysfunction is caused by increased afterload and results in elevated left ventricular end-diastolic pressure and elevated left atrial pressure. The left atrium responds by dilating, which increases the risk of atrial fibrillation.

Lancellotti et al⁸ investigated the negative prognostic implications of a large indexed left atrial area in asymptomatic patients with severe aortic stenosis. They found that patients with an indexed left atrial area greater than 12.2 cm²/m² had a 77% 2-year probability of aortic valve replacement or death.

Beach et al²⁸ examined cardiac remodeling after surgery and found that the left atrial diameter did not decrease after aortic valve replacement, even after left ventricular hypertrophy reversed. This observation has major prognostic implications. Patients with a severely enlarged left atrium (> 5.0 cm in diameter) had considerably lower survival rates than patients with a diameter less than 3.55 cm at 5 years (61% vs 85%) and at 10 years (28% vs 62%) after aortic valve replacement.

Therefore, left atrial size appears to have an important long-term impact on prognosis in patients with aortic stenosis even after aortic valve replacement and adds valuable information when assessing the effect of aortic stenosis on myocardial function.

B-type natriuretic peptide

Natriuretic peptides are cardiac hormones released in response to myocyte stretch. In aortic stenosis, increased afterload induces significant expression of BNP, N-terminal proBNP,²⁹ and atrial natriuretic peptide,³⁰ with numerous studies showing a good correlation between plasma natriuretic peptide levels and severity of aortic stenosis.^31–34

Natriuretic peptides, though not specific, are an easy and low-cost way to assess left ventricular function

Bergler-Klein et al³³ showed that patients with asymptomatic aortic stenosis who developed symptoms during follow-up had higher levels of these biomarkers than patients who remained asymptomatic. Of note, patients with BNP levels lower than 130 pg/mL had significantly better symptom-free survival than those with higher levels, 66% vs 34% at 12 months.

However, these biomarkers are not specific to aortic stenosis and can be elevated in any condition that increases left ventricular stress. Nevertheless, they offer an easy and low-cost way to assess left ventricular function and may give an indication of the total burden of disease on the left ventricle.

Global left ventricular longitudinal strain

In view of the limitations of the left ventricular ejection fraction in identifying changes in the structure of the heart and in early detection of myocardial dysfunction, assessment of myocardial deformation using strain imaging is proving an attractive alternative.

Strain is the normalized, dimensionless measure of deformation of a solid object (such as a segment of myocardium) in response to an applied force or stress.³⁵ A novel echocardiographic technique allows assessment of segmental myocardial deformation and thereby overcomes the limitation of tethering, which limits other echocardiographic techniques in the assessment of systolic function. Strain can be circumferential, longitudinal, or radial and is generally assessed using either tissue Doppler velocities or 2D echocardiographic speckle-tracking techniques. Longitudinal strain has proven to be a more sensitive method than left ventricular ejection fraction in detecting subclinical myocardial dysfunction and is a superior prognosticator in a variety of clinical conditions.^36,37

Abnormal strain develops very early in the disease process and can even be seen in patients with mild aortic stenosis.

A study by Kearney et al³⁸ in 146 patients with various degrees of aortic stenosis (26% mild, 21% moderate, and 53% severe) and preserved left ventricular ejection fraction demonstrated that global longitudinal strain worsened with increasing severity of aortic stenosis. Furthermore, global longitudinal strain was a strong independent predictor of all-cause mortality (hazard ratio 1.38, P < .001).

Similarly, in a study by Lancellotti et al⁸ in 163 patients with at least moderate to severe asymptomatic aortic stenosis, impaired longitudinal myocardial strain was an independent predictor of survival. Patients with longitudinal strain greater than 15.9% had significantly better outcomes than patients with strain of 15.9% or less (4-year survival 63% vs 22%, P < .001).

Hence, left ventricular global longitudinal strain offers an alternative—perhaps a superior alternative—to left ventricular ejection fraction in detecting and quantifying left ventricular dysfunction in asymptomatic aortic stenosis. It is an exciting new marker for the future in aortic stenosis, with a threshold of strain below 15.9% as a possible cutoff for those at higher risk of poorer outcomes.

WHERE ARE WE NOW? WHERE ARE WE GOING?

Aortic valve replacement in patients with severe but asymptomatic aortic stenosis remains a topic of debate, but support is growing for earlier intervention.

Now that concerns over the safety of exercise stress testing in patients with severe asymptomatic aortic stenosis have subsided following multiple studies,^16,17 exercise testing should be performed in patients with asymp­tomatic severe aortic stenosis suspected of having reduced exercise capacity, with stress echocardiography providing added prognostic information through its assessment of exercise-induced changes in mean pressure gradient¹⁹ and systolic pulmonary artery pressure.^21–23

Further study of the newer evaluation techniques is needed to evaluate long-term

Assessing left ventricular function provides important information about prognosis, with left ventricular ejection fraction, left ventricular diameter, left atrial size, BNP, and global longitudinal strain all helping identify asymptomatic patients at higher risk of death. Surgical intervention in asymptomatic patients with severe aortic stenosis may be considered when there is evidence of higher longer-term mortality risk based on reduced functional capacity, excess left ventricular hypertrophy, and abnormal left ventricular function as detected by ancillary methods such as global longitudinal strain and BNP elevation despite a normal left ventricular ejection fraction.

Figure 3 shows a possible algorithm to define which patients would benefit from earlier intervention. However, left ventricular hypertrophy, left atrial diameter, BNP, left ventricular longitudinal strain, and changes in systolic pulmonary artery pressure are not included in the current ACC/AHA guidelines for the management of asymptomatic patients with severe aortic stenosis. Further study is needed to determine whether earlier intervention in those with adverse risk profiles based on the newer evaluation techniques described above leads to better long-term outcomes.

Intervention should especially be considered in those in whom the measured surgical risk is low and in surgical centers at which the mortality rate is low.

Aortic stenosis is the most common valvular heart condition in the developed world, affecting 3% of people between ages 75 and 85¹ and 4% of people over age 85.² Aortic valve replacement remains the only treatment proven to reduce the rates of mortality and morbidity in this condition.³ Under current guidelines,^4,5 the onset of symptoms of exertional angina, syncope, or dyspnea in a patient who has severe aortic stenosis is a class I indication for surgery—ie, surgery should be performed.

However, high-gradient, severe aortic stenosis that is asymptomatic often poses a dilemma. The annual rate of sudden death in patients with this condition is estimated at 1% to 3%,^6–9 but the surgical mortality rate in aortic valve replacement has been as high as 6% in Medicare patients (varying by center and comorbidities).¹⁰ Therefore, the traditional teaching was to not surgically replace the valve in asymptomatic patients, based on an adverse risk-benefit ratio. But with improvements in surgical techniques and prostheses, these rates have been reduced to 2.41% at high-volume centers¹¹ (and to less than 1% at some hospitals),¹² arguing in favor of earlier intervention.

Complicating the issue, transcatheter aortic valve replacement has become widely available, but further investigation into its use in this patient cohort is warranted.

Furthermore, many patients with severe but apparently asymptomatic aortic stenosis and normal left ventricular ejection fraction may actually have impaired exercise capacity, or they may have structural left ventricular changes such as severe hypertrophy or reduction in global strain, which may worsen the long-term survival rate.^13,14

A prospective trial in patients with severe aortic stenosis found that mortality rates were significantly lower in those who underwent surgery early than in those who received conventional treatment, ie, watchful waiting (no specific medical treatment for aortic stenosis is available).¹⁵

Patients with asymptomatic severe aortic stenosis are a diverse group; some have a far worse prognosis than others, with or without surgery.

This paper reviews the guidelines for valve replacement in this patient group and the factors useful in establishing who should be considered for early intervention even if they have no classic symptoms (Figure 1).

SIGNS AND SYMPTOMS OF STENOSIS

Aortic stenosis is often first suspected when a patient presents with angina, dyspnea, and syncope, or when an ejection systolic murmur is heard incidentally on physical examination—typically a high-pitched, crescendo-decrescendo, midsystolic ejection murmur that is best heard at the right upper sternal border and that radiates to the carotid arteries.

Several physical findings may help in assessing the severity of aortic stenosis. In mild stenosis, the murmur peaks in early systole, but as the disease progresses the peak moves later into systole. The corollary of this phenomenon is a weak and delayed carotid upstroke known as “pulsus parvus et tardus.” This can be assessed by palpating the carotid artery while auscultating the heart.

Aortic stenosis is often first suspected when a patient has angina, dyspnea, and syncope or an ejection systolic murmur

The second heart sound becomes progressively softer as the stenosis advances until it is no longer audible. If a fourth heart sound is present, it may be due to concentric left ventricular hypertrophy with reduced left ventricular compliance, and a third heart sound indicates severe left ventricular dysfunction. Both of these findings suggest severe aortic stenosis.

ECHOCARDIOGRAPHIC MEASURES OF SEVERITY

Echocardiography is the best established and most important initial investigation in the assessment of a patient with suspected aortic stenosis. It usually provides accurate information on the severity and the mechanism of stenosis. The following findings indicate severe aortic stenosis:

• Mean pressure gradient > 40 mm Hg
• Peak aortic jet velocity > 4.0 m/s
• Aortic valve area < 1 cm².

RECOMMENDATIONS FOR SURGERY BASED ON SEVERITY AND SYMPTOMS

The American College of Cardiology and American Heart Association (ACC/AHA)⁴ have issued the following recommendations for aortic valve replacement, based on the severity of stenosis and on whether the patient has symptoms (Figure 2):

Severe stenosis, with symptoms: class I recommendation (surgery should be done). Without surgery, these patients have a very poor prognosis, with an overall mortality rate of 75% at 3 years.³

Severe stenosis, no symptoms, in patients undergoing cardiac surgery for another indication (eg, coronary artery bypass grafting, ascending aortic surgery, or surgery on other valves): class I recommendation for concomitant aortic valve replacement.

Moderate stenosis, no symptoms, in patients undergoing cardiac surgery for another indication: class IIa recommendation (ie, aortic valve replacement “is reasonable”).

Very severe stenosis (aortic peak velocity > 5.0 m/s or mean pressure gradient ≥ 60 mm Hg), no symptoms, and low risk of death during surgery: class IIa recommendation.

Severe stenosis, no symptoms, and an increase in transaortic velocity of 0.3 m/s or more per year on serial testing or in patients considered to be at high risk for rapid disease progression, such as elderly patients with severe calcification: class IIb recommendation (surgery “can be considered”). The threshold to replace the valve is lower for patients who cannot make serial follow-up appointments because they live far away or lack transportation, or because they have problems with compliance.

Surgery for those with left ventricular dysfunction

Echocardiography also provides information on left ventricular function, and patients with left ventricular dysfunction have significantly worse outcomes. Studies have shown substantial differences in survival in patients who had an ejection fraction of less than 50% before valve replacement compared with those with a normal ejection fraction.³

Thus, the ACC/AHA guidelines recommend immediate referral for aortic valve replacement in asymptomatic patients whose left ventricular ejection fraction is less than 50% (class I recommendation, level of evidence B) in the hope of preventing irreversible ventricular dysfunction.⁴

TREADMILL EXERCISE TESTING UNMASKS SYMPTOMS

Treadmill testing is absolutely contraindicated in patients with severe symptomatic aortic stenosis

In the past, severe aortic stenosis was considered a contraindication to stress testing because of concerns of precipitating severe, life-threatening complications. However, studies over the past 10 years have shown that a supervised modified Bruce protocol is safe in patients with severe asymptomatic aortic stenosis.^16,17

However, treadmill exercise testing clearly is absolutely contraindicated in patients with severe symptomatic aortic stenosis because of the risk of syncope or of precipitating a malignant arrhythmia. Nevertheless, it may play an essential role in the workup of a physically active patient with no symptoms.

Symptoms can develop insidiously in patients with chronic valve disease and may often go unrecognized by patients and their physicians. Many patients who state they have no symptoms may actually be subconsciously limiting their exercise to avoid symptoms.

Amato et al¹³ examined the exercise capacity of 66 patients reported to have severe asymptomatic aortic stenosis. Treadmill exercise testing was considered positive in this study if the patient developed symptoms or complex ventricular arrhythmias, had blood pressure that failed to rise by 20 mm Hg, or developed horizontal or down-sloping ST depression (≥ 1 mm in men, ≥ 2 mm in women). Twenty (30.3%) of the 66 patients developed symptoms during exercise testing, and they had a significantly worse prognosis: the 2-year event-free survival rate was only 19% in those with a positive test compared with 85% in those with a negative test.¹³ This study highlights the problem of patients subconsciously reducing their level of activity, thereby masking their true symptoms.

A meta-analysis by Rafique et al¹⁸ found that asymptomatic patients with abnormal results on exercise testing had a risk of cardiac events during follow-up that was eight times higher than normal, and a risk of sudden death 5.5 times higher.

With trials demonstrating that exercise testing is safe and prognostically useful in patients with aortic stenosis, the ACC/AHA guidelines emphasize its role, giving a class I recommendation for aortic valve replacement in patients who develop symptoms on exercise testing, and a class IIa recommendation in asymp­tomatic patients with decreased exercise tolerance or an exercise-related fall in blood pressure (Figure 2).⁴

STRESS ECHOCARDIOGRAPHY

Stress echocardiography has been used since the 1980s to assess the hemodynamic consequences of valvular heart disease, and many studies highlight its prognostic usefulness in patients with asymptomatic aortic stenosis.

In a 2005 study by Lancellotti et al,¹⁹ 69 patients with severe asymptomatic aortic stenosis underwent a symptom-limited bicycle exercise stress test using quantitative Doppler echocardiography both at rest and at peak exercise, and a number of independent predictors of poor outcome (ie, symptoms, aortic valve replacement, death) were identified. These predictors included an abnormal test result, defined as any of the following: angina, dyspnea, ST-segment depression of 2 mm Hg or more, a fall or a small (< 20 mm Hg) rise in systolic blood pressure during the test, an aortic valve area of 0.75 cm² or less, or a mean increase in valve gradient of 18 mm Hg or more.

Subsequently, a multicenter prospective trial assessed the value of exercise stress echocardiography in 186 patients with asymptomatic moderate or severe aortic stenosis.²⁰ A mean increase in the aortic valve gradient of 20 mm Hg or more after exercise was associated with a rate of cardiovascular events (death, aortic valve replacement) 3.8 times higher, independent of other risk factors and whether moderate or severe stenosis was present (Table 1).²⁰

Exercise-induced changes in systolic pulmonary artery pressure, which can be assessed using stress echocardiography, also have prognostic utility. Elevated systolic pulmonary artery pressure (> 50 mm Hg) seems to portend a poorer prognosis^21,22 and a higher mortality rate after valve replacement,²³ making it an independent predictor of hospital mortality and postoperative major adverse cardiovascular and cerebrovascular events (Table 1).

Exercise echocardiography also can be used to assess the patient’s contractile reserve. Left ventricular contractile reserve can be defined as an exercise-induced increase in left ventricular ejection fraction. In a study by Maréchaux et al²⁴ in 50 patients with asymptomatic aortic stenosis and a normal resting left ventricular ejection fraction (> 50%), 40% of patients did not have left ventricular contractile reserve. In fact, their left ventricular ejection fraction decreased with exercise (from 64 ± 10% to 53 ± 12%). The subgroup of patients without contractile reserve developed symptoms more frequently during exercise and had lower event-free survival (Table 1).

Stress echocardiography has recently been introduced into the European Society of Cardiology guidelines, which give a class IIb indication for aortic valve replacement in asymp­tomatic patients who have severe aortic stenosis, a normal ejection fraction, and a greater than 20-mm Hg increase in mean gradient on exercise.⁵ But it has yet to be introduced into the ACC/AHA guidelines as a consideration for surgery.

LEFT VENTRICULAR FUNCTION: BEYOND EJECTION FRACTION

Left ventricular dysfunction is a bad sign for patients with aortic stenosis. Struggling to empty its contents through the narrowed aortic valve, the left ventricle is subjected to increased wall stress and eventually develops hypertrophy. The hypertrophied heart muscle requires more oxygen but receives less perfusion. Eventually, myocardial fibrosis develops, leading to systolic dysfunction and a reduction in the ejection fraction. As described above, patients with asymptomatic aortic stenosis and a left ventricular ejection fraction less than 50% have a poor prognosis,¹⁴ and therefore the ACC/AHA guidelines give this condition a class I recommendation for surgery.⁴

The ejection fraction has limitations as a marker of left ventricular function

However, the ejection fraction has limitations as a marker of left ventricular function. It reflects changes in left ventricular cavity volume but not in the complex structure of the left ventricle. Several studies show that up to one-third of patients with severe aortic stenosis have considerable impairment of intrinsic myocardial systolic function despite a preserved ejection fraction.^8,25,26

Thus, other variables such as left atrial size, left ventricular hypertrophy, myocardial deformation (assessed using strain imaging), and B-type natriuretic peptide (BNP) level may also be considered in assessing the effect of severe aortic stenosis on left ventricular function in the context of a normal ejection fraction (Table 2).

Left ventricular hypertrophy

The development of left ventricular hypertrophy is one of the earliest compensatory responses of the ventricle to the increase in afterload. This leads to impaired myocardial relaxation and reduced myocardial compliance, with resultant diastolic dysfunction with increased filling pressures.

Cioffi et al,²⁷ in a study in 209 patients with severe but asymptomatic aortic stenosis, found that inappropriately high left ventricular mass (> 110% of that expected for body size, sex, and wall stress) portended a 4.5-times higher risk of death, independent of other risk factors.

Severe left ventricular hypertrophy may have a long-term effect on prognosis irrespective of valve replacement. An observational study¹⁴ of 3,049 patients who underwent aortic valve replacement for severe aortic stenosis showed that the 10-year survival rate was 45% in those whose left ventricular mass was greater than 185 g/m², compared with 65% in patients whose left ventricular mass was less than 100 g/m².

Thus, as surgical mortality and morbidity rates decrease, the impact of these structural changes in left ventricular wall thickness may affect the decision to intervene earlier in order to improve longer-term outcomes in select asymptomatic patients with high-risk features.

Left atrial size

Diastolic dysfunction is caused by increased afterload and results in elevated left ventricular end-diastolic pressure and elevated left atrial pressure. The left atrium responds by dilating, which increases the risk of atrial fibrillation.

Lancellotti et al⁸ investigated the negative prognostic implications of a large indexed left atrial area in asymptomatic patients with severe aortic stenosis. They found that patients with an indexed left atrial area greater than 12.2 cm²/m² had a 77% 2-year probability of aortic valve replacement or death.

Beach et al²⁸ examined cardiac remodeling after surgery and found that the left atrial diameter did not decrease after aortic valve replacement, even after left ventricular hypertrophy reversed. This observation has major prognostic implications. Patients with a severely enlarged left atrium (> 5.0 cm in diameter) had considerably lower survival rates than patients with a diameter less than 3.55 cm at 5 years (61% vs 85%) and at 10 years (28% vs 62%) after aortic valve replacement.

Therefore, left atrial size appears to have an important long-term impact on prognosis in patients with aortic stenosis even after aortic valve replacement and adds valuable information when assessing the effect of aortic stenosis on myocardial function.

B-type natriuretic peptide

Natriuretic peptides are cardiac hormones released in response to myocyte stretch. In aortic stenosis, increased afterload induces significant expression of BNP, N-terminal proBNP,²⁹ and atrial natriuretic peptide,³⁰ with numerous studies showing a good correlation between plasma natriuretic peptide levels and severity of aortic stenosis.^31–34

Natriuretic peptides, though not specific, are an easy and low-cost way to assess left ventricular function

Bergler-Klein et al³³ showed that patients with asymptomatic aortic stenosis who developed symptoms during follow-up had higher levels of these biomarkers than patients who remained asymptomatic. Of note, patients with BNP levels lower than 130 pg/mL had significantly better symptom-free survival than those with higher levels, 66% vs 34% at 12 months.

However, these biomarkers are not specific to aortic stenosis and can be elevated in any condition that increases left ventricular stress. Nevertheless, they offer an easy and low-cost way to assess left ventricular function and may give an indication of the total burden of disease on the left ventricle.

Global left ventricular longitudinal strain

In view of the limitations of the left ventricular ejection fraction in identifying changes in the structure of the heart and in early detection of myocardial dysfunction, assessment of myocardial deformation using strain imaging is proving an attractive alternative.

Strain is the normalized, dimensionless measure of deformation of a solid object (such as a segment of myocardium) in response to an applied force or stress.³⁵ A novel echocardiographic technique allows assessment of segmental myocardial deformation and thereby overcomes the limitation of tethering, which limits other echocardiographic techniques in the assessment of systolic function. Strain can be circumferential, longitudinal, or radial and is generally assessed using either tissue Doppler velocities or 2D echocardiographic speckle-tracking techniques. Longitudinal strain has proven to be a more sensitive method than left ventricular ejection fraction in detecting subclinical myocardial dysfunction and is a superior prognosticator in a variety of clinical conditions.^36,37

Abnormal strain develops very early in the disease process and can even be seen in patients with mild aortic stenosis.

A study by Kearney et al³⁸ in 146 patients with various degrees of aortic stenosis (26% mild, 21% moderate, and 53% severe) and preserved left ventricular ejection fraction demonstrated that global longitudinal strain worsened with increasing severity of aortic stenosis. Furthermore, global longitudinal strain was a strong independent predictor of all-cause mortality (hazard ratio 1.38, P < .001).

Similarly, in a study by Lancellotti et al⁸ in 163 patients with at least moderate to severe asymptomatic aortic stenosis, impaired longitudinal myocardial strain was an independent predictor of survival. Patients with longitudinal strain greater than 15.9% had significantly better outcomes than patients with strain of 15.9% or less (4-year survival 63% vs 22%, P < .001).

Hence, left ventricular global longitudinal strain offers an alternative—perhaps a superior alternative—to left ventricular ejection fraction in detecting and quantifying left ventricular dysfunction in asymptomatic aortic stenosis. It is an exciting new marker for the future in aortic stenosis, with a threshold of strain below 15.9% as a possible cutoff for those at higher risk of poorer outcomes.

WHERE ARE WE NOW? WHERE ARE WE GOING?

Aortic valve replacement in patients with severe but asymptomatic aortic stenosis remains a topic of debate, but support is growing for earlier intervention.

Now that concerns over the safety of exercise stress testing in patients with severe asymptomatic aortic stenosis have subsided following multiple studies,^16,17 exercise testing should be performed in patients with asymp­tomatic severe aortic stenosis suspected of having reduced exercise capacity, with stress echocardiography providing added prognostic information through its assessment of exercise-induced changes in mean pressure gradient¹⁹ and systolic pulmonary artery pressure.^21–23

Further study of the newer evaluation techniques is needed to evaluate long-term

Assessing left ventricular function provides important information about prognosis, with left ventricular ejection fraction, left ventricular diameter, left atrial size, BNP, and global longitudinal strain all helping identify asymptomatic patients at higher risk of death. Surgical intervention in asymptomatic patients with severe aortic stenosis may be considered when there is evidence of higher longer-term mortality risk based on reduced functional capacity, excess left ventricular hypertrophy, and abnormal left ventricular function as detected by ancillary methods such as global longitudinal strain and BNP elevation despite a normal left ventricular ejection fraction.

Figure 3 shows a possible algorithm to define which patients would benefit from earlier intervention. However, left ventricular hypertrophy, left atrial diameter, BNP, left ventricular longitudinal strain, and changes in systolic pulmonary artery pressure are not included in the current ACC/AHA guidelines for the management of asymptomatic patients with severe aortic stenosis. Further study is needed to determine whether earlier intervention in those with adverse risk profiles based on the newer evaluation techniques described above leads to better long-term outcomes.

Intervention should especially be considered in those in whom the measured surgical risk is low and in surgical centers at which the mortality rate is low.

Aortic stenosis is the most common valvular heart condition in the developed world, affecting 3% of people between ages 75 and 85¹ and 4% of people over age 85.² Aortic valve replacement remains the only treatment proven to reduce the rates of mortality and morbidity in this condition.³ Under current guidelines,^4,5 the onset of symptoms of exertional angina, syncope, or dyspnea in a patient who has severe aortic stenosis is a class I indication for surgery—ie, surgery should be performed.

However, high-gradient, severe aortic stenosis that is asymptomatic often poses a dilemma. The annual rate of sudden death in patients with this condition is estimated at 1% to 3%,^6–9 but the surgical mortality rate in aortic valve replacement has been as high as 6% in Medicare patients (varying by center and comorbidities).¹⁰ Therefore, the traditional teaching was to not surgically replace the valve in asymptomatic patients, based on an adverse risk-benefit ratio. But with improvements in surgical techniques and prostheses, these rates have been reduced to 2.41% at high-volume centers¹¹ (and to less than 1% at some hospitals),¹² arguing in favor of earlier intervention.

Complicating the issue, transcatheter aortic valve replacement has become widely available, but further investigation into its use in this patient cohort is warranted.

Furthermore, many patients with severe but apparently asymptomatic aortic stenosis and normal left ventricular ejection fraction may actually have impaired exercise capacity, or they may have structural left ventricular changes such as severe hypertrophy or reduction in global strain, which may worsen the long-term survival rate.^13,14

A prospective trial in patients with severe aortic stenosis found that mortality rates were significantly lower in those who underwent surgery early than in those who received conventional treatment, ie, watchful waiting (no specific medical treatment for aortic stenosis is available).¹⁵

Patients with asymptomatic severe aortic stenosis are a diverse group; some have a far worse prognosis than others, with or without surgery.

This paper reviews the guidelines for valve replacement in this patient group and the factors useful in establishing who should be considered for early intervention even if they have no classic symptoms (Figure 1).

SIGNS AND SYMPTOMS OF STENOSIS

Aortic stenosis is often first suspected when a patient presents with angina, dyspnea, and syncope, or when an ejection systolic murmur is heard incidentally on physical examination—typically a high-pitched, crescendo-decrescendo, midsystolic ejection murmur that is best heard at the right upper sternal border and that radiates to the carotid arteries.

Several physical findings may help in assessing the severity of aortic stenosis. In mild stenosis, the murmur peaks in early systole, but as the disease progresses the peak moves later into systole. The corollary of this phenomenon is a weak and delayed carotid upstroke known as “pulsus parvus et tardus.” This can be assessed by palpating the carotid artery while auscultating the heart.

Aortic stenosis is often first suspected when a patient has angina, dyspnea, and syncope or an ejection systolic murmur

The second heart sound becomes progressively softer as the stenosis advances until it is no longer audible. If a fourth heart sound is present, it may be due to concentric left ventricular hypertrophy with reduced left ventricular compliance, and a third heart sound indicates severe left ventricular dysfunction. Both of these findings suggest severe aortic stenosis.

ECHOCARDIOGRAPHIC MEASURES OF SEVERITY

Echocardiography is the best established and most important initial investigation in the assessment of a patient with suspected aortic stenosis. It usually provides accurate information on the severity and the mechanism of stenosis. The following findings indicate severe aortic stenosis:

• Mean pressure gradient > 40 mm Hg
• Peak aortic jet velocity > 4.0 m/s
• Aortic valve area < 1 cm².

RECOMMENDATIONS FOR SURGERY BASED ON SEVERITY AND SYMPTOMS

The American College of Cardiology and American Heart Association (ACC/AHA)⁴ have issued the following recommendations for aortic valve replacement, based on the severity of stenosis and on whether the patient has symptoms (Figure 2):

Severe stenosis, with symptoms: class I recommendation (surgery should be done). Without surgery, these patients have a very poor prognosis, with an overall mortality rate of 75% at 3 years.³

Severe stenosis, no symptoms, in patients undergoing cardiac surgery for another indication (eg, coronary artery bypass grafting, ascending aortic surgery, or surgery on other valves): class I recommendation for concomitant aortic valve replacement.

Moderate stenosis, no symptoms, in patients undergoing cardiac surgery for another indication: class IIa recommendation (ie, aortic valve replacement “is reasonable”).

Very severe stenosis (aortic peak velocity > 5.0 m/s or mean pressure gradient ≥ 60 mm Hg), no symptoms, and low risk of death during surgery: class IIa recommendation.

Severe stenosis, no symptoms, and an increase in transaortic velocity of 0.3 m/s or more per year on serial testing or in patients considered to be at high risk for rapid disease progression, such as elderly patients with severe calcification: class IIb recommendation (surgery “can be considered”). The threshold to replace the valve is lower for patients who cannot make serial follow-up appointments because they live far away or lack transportation, or because they have problems with compliance.

Surgery for those with left ventricular dysfunction

Echocardiography also provides information on left ventricular function, and patients with left ventricular dysfunction have significantly worse outcomes. Studies have shown substantial differences in survival in patients who had an ejection fraction of less than 50% before valve replacement compared with those with a normal ejection fraction.³

Thus, the ACC/AHA guidelines recommend immediate referral for aortic valve replacement in asymptomatic patients whose left ventricular ejection fraction is less than 50% (class I recommendation, level of evidence B) in the hope of preventing irreversible ventricular dysfunction.⁴

TREADMILL EXERCISE TESTING UNMASKS SYMPTOMS

Treadmill testing is absolutely contraindicated in patients with severe symptomatic aortic stenosis

In the past, severe aortic stenosis was considered a contraindication to stress testing because of concerns of precipitating severe, life-threatening complications. However, studies over the past 10 years have shown that a supervised modified Bruce protocol is safe in patients with severe asymptomatic aortic stenosis.^16,17

However, treadmill exercise testing clearly is absolutely contraindicated in patients with severe symptomatic aortic stenosis because of the risk of syncope or of precipitating a malignant arrhythmia. Nevertheless, it may play an essential role in the workup of a physically active patient with no symptoms.

Symptoms can develop insidiously in patients with chronic valve disease and may often go unrecognized by patients and their physicians. Many patients who state they have no symptoms may actually be subconsciously limiting their exercise to avoid symptoms.

Amato et al¹³ examined the exercise capacity of 66 patients reported to have severe asymptomatic aortic stenosis. Treadmill exercise testing was considered positive in this study if the patient developed symptoms or complex ventricular arrhythmias, had blood pressure that failed to rise by 20 mm Hg, or developed horizontal or down-sloping ST depression (≥ 1 mm in men, ≥ 2 mm in women). Twenty (30.3%) of the 66 patients developed symptoms during exercise testing, and they had a significantly worse prognosis: the 2-year event-free survival rate was only 19% in those with a positive test compared with 85% in those with a negative test.¹³ This study highlights the problem of patients subconsciously reducing their level of activity, thereby masking their true symptoms.

A meta-analysis by Rafique et al¹⁸ found that asymptomatic patients with abnormal results on exercise testing had a risk of cardiac events during follow-up that was eight times higher than normal, and a risk of sudden death 5.5 times higher.

With trials demonstrating that exercise testing is safe and prognostically useful in patients with aortic stenosis, the ACC/AHA guidelines emphasize its role, giving a class I recommendation for aortic valve replacement in patients who develop symptoms on exercise testing, and a class IIa recommendation in asymp­tomatic patients with decreased exercise tolerance or an exercise-related fall in blood pressure (Figure 2).⁴

STRESS ECHOCARDIOGRAPHY

Stress echocardiography has been used since the 1980s to assess the hemodynamic consequences of valvular heart disease, and many studies highlight its prognostic usefulness in patients with asymptomatic aortic stenosis.

In a 2005 study by Lancellotti et al,¹⁹ 69 patients with severe asymptomatic aortic stenosis underwent a symptom-limited bicycle exercise stress test using quantitative Doppler echocardiography both at rest and at peak exercise, and a number of independent predictors of poor outcome (ie, symptoms, aortic valve replacement, death) were identified. These predictors included an abnormal test result, defined as any of the following: angina, dyspnea, ST-segment depression of 2 mm Hg or more, a fall or a small (< 20 mm Hg) rise in systolic blood pressure during the test, an aortic valve area of 0.75 cm² or less, or a mean increase in valve gradient of 18 mm Hg or more.

Subsequently, a multicenter prospective trial assessed the value of exercise stress echocardiography in 186 patients with asymptomatic moderate or severe aortic stenosis.²⁰ A mean increase in the aortic valve gradient of 20 mm Hg or more after exercise was associated with a rate of cardiovascular events (death, aortic valve replacement) 3.8 times higher, independent of other risk factors and whether moderate or severe stenosis was present (Table 1).²⁰

Exercise-induced changes in systolic pulmonary artery pressure, which can be assessed using stress echocardiography, also have prognostic utility. Elevated systolic pulmonary artery pressure (> 50 mm Hg) seems to portend a poorer prognosis^21,22 and a higher mortality rate after valve replacement,²³ making it an independent predictor of hospital mortality and postoperative major adverse cardiovascular and cerebrovascular events (Table 1).

Exercise echocardiography also can be used to assess the patient’s contractile reserve. Left ventricular contractile reserve can be defined as an exercise-induced increase in left ventricular ejection fraction. In a study by Maréchaux et al²⁴ in 50 patients with asymptomatic aortic stenosis and a normal resting left ventricular ejection fraction (> 50%), 40% of patients did not have left ventricular contractile reserve. In fact, their left ventricular ejection fraction decreased with exercise (from 64 ± 10% to 53 ± 12%). The subgroup of patients without contractile reserve developed symptoms more frequently during exercise and had lower event-free survival (Table 1).

Stress echocardiography has recently been introduced into the European Society of Cardiology guidelines, which give a class IIb indication for aortic valve replacement in asymp­tomatic patients who have severe aortic stenosis, a normal ejection fraction, and a greater than 20-mm Hg increase in mean gradient on exercise.⁵ But it has yet to be introduced into the ACC/AHA guidelines as a consideration for surgery.

LEFT VENTRICULAR FUNCTION: BEYOND EJECTION FRACTION

Left ventricular dysfunction is a bad sign for patients with aortic stenosis. Struggling to empty its contents through the narrowed aortic valve, the left ventricle is subjected to increased wall stress and eventually develops hypertrophy. The hypertrophied heart muscle requires more oxygen but receives less perfusion. Eventually, myocardial fibrosis develops, leading to systolic dysfunction and a reduction in the ejection fraction. As described above, patients with asymptomatic aortic stenosis and a left ventricular ejection fraction less than 50% have a poor prognosis,¹⁴ and therefore the ACC/AHA guidelines give this condition a class I recommendation for surgery.⁴

The ejection fraction has limitations as a marker of left ventricular function

However, the ejection fraction has limitations as a marker of left ventricular function. It reflects changes in left ventricular cavity volume but not in the complex structure of the left ventricle. Several studies show that up to one-third of patients with severe aortic stenosis have considerable impairment of intrinsic myocardial systolic function despite a preserved ejection fraction.^8,25,26

Thus, other variables such as left atrial size, left ventricular hypertrophy, myocardial deformation (assessed using strain imaging), and B-type natriuretic peptide (BNP) level may also be considered in assessing the effect of severe aortic stenosis on left ventricular function in the context of a normal ejection fraction (Table 2).

Left ventricular hypertrophy

The development of left ventricular hypertrophy is one of the earliest compensatory responses of the ventricle to the increase in afterload. This leads to impaired myocardial relaxation and reduced myocardial compliance, with resultant diastolic dysfunction with increased filling pressures.

Cioffi et al,²⁷ in a study in 209 patients with severe but asymptomatic aortic stenosis, found that inappropriately high left ventricular mass (> 110% of that expected for body size, sex, and wall stress) portended a 4.5-times higher risk of death, independent of other risk factors.

Severe left ventricular hypertrophy may have a long-term effect on prognosis irrespective of valve replacement. An observational study¹⁴ of 3,049 patients who underwent aortic valve replacement for severe aortic stenosis showed that the 10-year survival rate was 45% in those whose left ventricular mass was greater than 185 g/m², compared with 65% in patients whose left ventricular mass was less than 100 g/m².

Thus, as surgical mortality and morbidity rates decrease, the impact of these structural changes in left ventricular wall thickness may affect the decision to intervene earlier in order to improve longer-term outcomes in select asymptomatic patients with high-risk features.

Left atrial size

Diastolic dysfunction is caused by increased afterload and results in elevated left ventricular end-diastolic pressure and elevated left atrial pressure. The left atrium responds by dilating, which increases the risk of atrial fibrillation.

Lancellotti et al⁸ investigated the negative prognostic implications of a large indexed left atrial area in asymptomatic patients with severe aortic stenosis. They found that patients with an indexed left atrial area greater than 12.2 cm²/m² had a 77% 2-year probability of aortic valve replacement or death.

Beach et al²⁸ examined cardiac remodeling after surgery and found that the left atrial diameter did not decrease after aortic valve replacement, even after left ventricular hypertrophy reversed. This observation has major prognostic implications. Patients with a severely enlarged left atrium (> 5.0 cm in diameter) had considerably lower survival rates than patients with a diameter less than 3.55 cm at 5 years (61% vs 85%) and at 10 years (28% vs 62%) after aortic valve replacement.

Therefore, left atrial size appears to have an important long-term impact on prognosis in patients with aortic stenosis even after aortic valve replacement and adds valuable information when assessing the effect of aortic stenosis on myocardial function.

B-type natriuretic peptide

Natriuretic peptides are cardiac hormones released in response to myocyte stretch. In aortic stenosis, increased afterload induces significant expression of BNP, N-terminal proBNP,²⁹ and atrial natriuretic peptide,³⁰ with numerous studies showing a good correlation between plasma natriuretic peptide levels and severity of aortic stenosis.^31–34

Natriuretic peptides, though not specific, are an easy and low-cost way to assess left ventricular function

Bergler-Klein et al³³ showed that patients with asymptomatic aortic stenosis who developed symptoms during follow-up had higher levels of these biomarkers than patients who remained asymptomatic. Of note, patients with BNP levels lower than 130 pg/mL had significantly better symptom-free survival than those with higher levels, 66% vs 34% at 12 months.

However, these biomarkers are not specific to aortic stenosis and can be elevated in any condition that increases left ventricular stress. Nevertheless, they offer an easy and low-cost way to assess left ventricular function and may give an indication of the total burden of disease on the left ventricle.

Global left ventricular longitudinal strain

In view of the limitations of the left ventricular ejection fraction in identifying changes in the structure of the heart and in early detection of myocardial dysfunction, assessment of myocardial deformation using strain imaging is proving an attractive alternative.

Strain is the normalized, dimensionless measure of deformation of a solid object (such as a segment of myocardium) in response to an applied force or stress.³⁵ A novel echocardiographic technique allows assessment of segmental myocardial deformation and thereby overcomes the limitation of tethering, which limits other echocardiographic techniques in the assessment of systolic function. Strain can be circumferential, longitudinal, or radial and is generally assessed using either tissue Doppler velocities or 2D echocardiographic speckle-tracking techniques. Longitudinal strain has proven to be a more sensitive method than left ventricular ejection fraction in detecting subclinical myocardial dysfunction and is a superior prognosticator in a variety of clinical conditions.^36,37

Abnormal strain develops very early in the disease process and can even be seen in patients with mild aortic stenosis.

A study by Kearney et al³⁸ in 146 patients with various degrees of aortic stenosis (26% mild, 21% moderate, and 53% severe) and preserved left ventricular ejection fraction demonstrated that global longitudinal strain worsened with increasing severity of aortic stenosis. Furthermore, global longitudinal strain was a strong independent predictor of all-cause mortality (hazard ratio 1.38, P < .001).

Similarly, in a study by Lancellotti et al⁸ in 163 patients with at least moderate to severe asymptomatic aortic stenosis, impaired longitudinal myocardial strain was an independent predictor of survival. Patients with longitudinal strain greater than 15.9% had significantly better outcomes than patients with strain of 15.9% or less (4-year survival 63% vs 22%, P < .001).

Hence, left ventricular global longitudinal strain offers an alternative—perhaps a superior alternative—to left ventricular ejection fraction in detecting and quantifying left ventricular dysfunction in asymptomatic aortic stenosis. It is an exciting new marker for the future in aortic stenosis, with a threshold of strain below 15.9% as a possible cutoff for those at higher risk of poorer outcomes.

WHERE ARE WE NOW? WHERE ARE WE GOING?

Aortic valve replacement in patients with severe but asymptomatic aortic stenosis remains a topic of debate, but support is growing for earlier intervention.

Now that concerns over the safety of exercise stress testing in patients with severe asymptomatic aortic stenosis have subsided following multiple studies,^16,17 exercise testing should be performed in patients with asymp­tomatic severe aortic stenosis suspected of having reduced exercise capacity, with stress echocardiography providing added prognostic information through its assessment of exercise-induced changes in mean pressure gradient¹⁹ and systolic pulmonary artery pressure.^21–23

Further study of the newer evaluation techniques is needed to evaluate long-term

Assessing left ventricular function provides important information about prognosis, with left ventricular ejection fraction, left ventricular diameter, left atrial size, BNP, and global longitudinal strain all helping identify asymptomatic patients at higher risk of death. Surgical intervention in asymptomatic patients with severe aortic stenosis may be considered when there is evidence of higher longer-term mortality risk based on reduced functional capacity, excess left ventricular hypertrophy, and abnormal left ventricular function as detected by ancillary methods such as global longitudinal strain and BNP elevation despite a normal left ventricular ejection fraction.

Figure 3 shows a possible algorithm to define which patients would benefit from earlier intervention. However, left ventricular hypertrophy, left atrial diameter, BNP, left ventricular longitudinal strain, and changes in systolic pulmonary artery pressure are not included in the current ACC/AHA guidelines for the management of asymptomatic patients with severe aortic stenosis. Further study is needed to determine whether earlier intervention in those with adverse risk profiles based on the newer evaluation techniques described above leads to better long-term outcomes.

Intervention should especially be considered in those in whom the measured surgical risk is low and in surgical centers at which the mortality rate is low.