Radiation-Related Heart Disease: Current Knowledge and Future Prospects

Darby IJROBP 2010;75:656

It has been recognized since the 1960s that the heart may be damaged by substantial doses of radiation [>30 Gray (Gy)], such as used to occur during mantle radiotherapy for Hodgkin lymphoma. During the last few years, however, evidence that radiation-related heart disease (RRHD) can occur following doses below 20 Gy has emerged from several independent sources. Those sources include studies of breast cancer patients who received mean cardiac doses of 3 to 17 Gy when given radiotherapy following surgery and studies of survivors of the atomic bombings of Japan who received doses of up to 4 Gy.

At doses above 30 Gy, an increased risk of RRHD can becomes apparent within a year or two of exposure, and the risk increases with higher radiotherapy dose, younger age at irradiation, and the presence of conventional risk factors. At lower doses, the typical latency period is much longer and is often more than a decade. The nature and magnitude of the risk following lower doses is not well characterized, and it is not yet clear whether there is a threshold dose below which there is no risk.

The evidence regarding RRHD comes from several different disciplines. The present review brings together information from pathology, radiobiology, cardiology, radiation oncology, and epidemiology; it summarizes current knowledge, identifies gaps in that knowledge, and outlines some potential strategies for filling them.

Clinical manifestations and management

A wide range of clinical cardiovascular problems can arise from radiation therapy (RT) for thoracic malignancy. Radiation-related pericardial and myocardial diseases are less common today than in the past, due to modifications in RT techniques, particularly for lymphomas, resulting in lower radiation doses to the heart. The predominant clinical manifestation of RRHD today is probably CAD, but the frequency with which it occurs is unknown, as it does not differ clinically from CAD from other causes, and many radiation-related cases may not be recognized as such.

Acute pericarditis is still occasionally seen within weeks after cardiac irradiation. Patients present with pleuritic chest pain, fever, tachycardia, a pericardial rub, and characteristic electrocardiographic abnormalities. Those signs and symptoms usually resolve quickly and without consequence with nonsteroidal anti-inflammatory drug therapy. A proportion of patients, however, develop chronic pericarditis up to 10 years later. The incidence of disease is related to the volume of pericardium irradiated, the dose received, and possibly the presence of effusion during the acute phase. The severity is variable, ranging from asymptomatic pericardial thickening found incidentally to cardiac tamponade requiring urgent pericardiocentesis. For those patients with recurrent effusion, some authorities recommend subtotal pericardiectomy to prevent the development of severe constrictive pericarditis, which can be difficult to treat effectively at a later stage.

Radiation-related myocardial fibrosis is often asymptomatic and is picked up only incidentally on echocardiography more than 10 years after radiation therapy. Clinically significant ventricular dysfunction is uncommon but can occur, particularly in the context of anthracycline chemotherapy and high radiation doses (>30 Gy) to large volumes of the heart. The management of radiation-related cardiomyopathy currently follows guidelines for the management of cardiac failure due to other causes.

Radiotherapy has been associated with valvular heart disease, which is common in some series. The incidence has been related to mediastinal radiation doses of >30 Gy and younger age at irradiation with an average latency of >10 years for asymptomatic and longer (median, 22 years) for symptomatic disease. Aortic disease usually consists of mixed stenosis and regurgitation and is more common than mitral and right-sided disease.

Conduction system abnormalities have also been reported, often in association with other types of RRHD. A variety of arrhythmias and conduction blocks have been observed, but these are rarely clinically serious. Autonomic nervous system dysfunction with a persistent nonvariable tachycardia has also been described

Thoracic RT is now firmly established as a risk factor for CAD. Radiation-related CAD is usually not detected until at least 10 years after exposure, and the magnitude of the risk with modern RT techniques is not yet well defined. Relative risks increase with higher radiotherapy doses younger age at irradiation, and the presence of conventional risk factors for coronary artery disease. Radiation-related CAD is managed conventionally, with medical control of risk factors and percutaneous coronary intervention or surgical bypass if indicated. Surgery may be technically challenging due to the presence of mediastinal scarring, friability of the vessels, and a possible increased restenosis rate of irradiated internal mammary vessels when they are chosen for coronary artery bypass grafting. Patients most commonly presenting with RRHD are survivors of breast cancer or HL, although RRHD has also been reported after radiotherapy for testicular cancer and peptic ulcer disease

Studies of patients with HL (Hodgkin's Lymphoma)

In the general population, early changes indicative of atherosclerosis can be seen in young adults, but clinical manifestations, such as angina and MI, rarely present until patients are at least in their 50s. In contrast, young patients without conventional cardiac risk factors who received mediastinal irradiation for HL can present with CAD in their 20s. In such individuals the likely cause is clear, and they may provide an example of the hypothesis of accelerated atherosclerosis referred to above.

In older people, among whom heart disease in the absence of radiotherapy is common, RRHD is harder to identify, and much of our information about the risk comes from studies in which a large population of HL patients treated with RT have been identified and followed over many years. Information about heart disease is obtained either from patients' medical records or from disease registers or death certificates, and the risk of developing or dying from heart disease in the HL group can then be compared with that of the general population or with another suitable control group.

A recent study of 1,474 5-year survivors treated for HL before age 40 found relative risk (RR) values of 3 to 5 for cardiac morbidity of all types compared with the general population, suggesting that 66 to 80% of all cardiac disease in the HL population was due to therapy, while estimates of the RR of death from MI in HL patients compared with the general population are in the range 2 to 9

Both the radiotherapy techniques and the chemotherapy regimens used to treat HL are evolving with time. Therefore, the impact of cardiac disease in HL survivors requires periodic reevaluation. The modification of radiotherapy techniques has resulted in a reduction in the risk of death from some cardiac causes. For example, among 2,232 HL patients treated at Stanford from 1960 to 1991, the RR for non-MI cardiac death fell from 5.3 to 1.4 when subcarinal blocking was introduced. However, the RR for fatal MI was not changed significantly in this series, possibly due to continued RT exposure of the proximal coronary arteries. Patients treated during the 1990s in Canada still had an increased risk of hospitalization for cardiac disease that was more marked in those who received combination therapy with anthracyclines, suggesting that an increased risk of late cardiac effects had not yet been eliminated.

The proportional increase in the cardiac death rate is greater in survivors of childhood HL than in adults. In a report from the U.S. Childhood Cancer Survivor Study, which included 2,717 5-year survivors of childhood HL, the RR of death from all cardiac causes compared with that of the general population was 11.9 (95% confidence interval [CI], 9.1-15.3), while the RR of death from MI following mediastinal irradiation in childhood has been estimated to be as high as 41.5 (95% CI, 18.1-82.1) with RT doses of ≥42 Gy

Studies of patients with breast cancer

In studies of patients with HL, the death rates from heart disease have, in the past, often been so large that comparisons with rates in the general population gave a clear indication of the magnitude of the risk. For breast cancer patients, cardiac doses from RT have been lower and risks smaller. Therefore, the phenomenon of RRHD is less obvious in breast cancer patients than in HL patients. In breast cancer, much of our knowledge of the effect of RT on cardiovascular disease has come from long-term follow-up of women entered into trials in which all women received similar treatment in terms of surgery and drugs, and then half of the women were allocated at random also to receive adjuvant radiotherapy.

One of the first studies to examine the effect of RT on long-term survival in breast cancer was published by Cuzick et al. in 1987. This meta-analysis of randomized trials showed that survival beyond 10 years was significantly worse for those receiving RT. This study was unable to determine the disease responsible for the detrimental effect on survival, but subsequent meta-analyses by the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) have shown that mortality from heart disease was increased by 27% (2p = 0.0001) in women randomized to surgery plus RT compared with women randomized to surgery alone. Most of the increase was due to CAD. Aspects of the radiotherapy techniques used in the earlier trials that contributed to the increased cardiac mortality included field placement near the heart (particularly anterior fields used to treat the internal mammary nodes), orthovoltage radiation that delivered high doses to the anterior part of the heart, large daily fractions, and high total doses.

Recently, a preliminary analysis of updated EBCTCG data has related mortality from heart disease to estimated cardiac doses in over 30,000 women followed for up to 20 years. There is clear evidence that the radiation-related increase is higher in trials with larger mean cardiac RT doses and that the risk of death from heart disease increases by 3% per Gy (95% CI, 2%-5%; 2p < 0.00001). That estimate can only be taken as an approximate indication of the risk, as individual treatment plans were not available for the women in those trials. Nevertheless, the data provide strong evidence that risk of RRHD was related to cardiac dose in irradiated breast cancer patients.

Outside the context of a randomized trial, comparisons of mortality after various different treatment regimens are often misleading because the prognosis of patients given different treatments will vary. In breast cancer, however, a reliable indication of the effect of radiotherapy on heart disease can be obtained by comparing the experience of irradiated women with left-sided tumors with that of women with right-sided tumors. This can be done because cardiac radiation doses in women given radiotherapy for left-sided tumors are usually larger than the cardiac radiation doses in women with right-sided tumors, and breast cancer laterality has, in the past, played little part in determining who should be given radiotherapy.

As shown for breast cancer patients who were not treated with radiotherapy, the subsequent risk of heart disease was independent of tumor laterality, while for irradiated patients, the heart disease mortality ratio, left-sided vs. right-sided tumors, increased with increasing time since diagnosis (i.e., with increasing time since irradiation). The increase was specific to heart disease as, for mortality from all other known causes, the left-sided vs. right-sided mortality ratio was close to unity in both irradiated and unirradiated patients. That suggests that the increasing trend in the left-sided vs. right-sided mortality ratio for heart disease is caused by radiotherapy, with the bulk of the risk occurring more than a decade after exposure. The proportional increase was higher for women irradiated at ages 20 to 49 than at older ages, but the trend did not reach statistical significance.

A recent study examining the incidence of CAD following breast irradiation revealed a higher prevalence of stress test abnormalities in left-sided than in right-sided tumor patients (59% vs. 8%; p = 0.001). Among left-sided tumor patients, the disease distribution differed from that expected in women, with a preponderance of left anterior descending artery disease. The anterior portion of the heart and the left anterior descending artery territory are the parts of the heart most often within the tangential radiation fields used to treat breast cancer. Hence, this finding provides direct evidence of a causal effect of radiotherapy on the development of CAD.

Studies of atomic bomb survivors

Mortality from heart disease has been studied among atomic bomb survivors in the Hiroshima-Nagasaki Life Span Study (LSS) for over 55 years. Distinctive features of the LSS are a population of >86,000 survivors who received whole-body uniform doses; individual radiation doses; a dose range of 0 to ∼4 Gy; and complete mortality ascertainment and cause-of-death information.

Dose-related increases in heart disease mortality in the LSS occur in both genders, based on >8,400 heart disease deaths. The risk increased by 14% per Gy (95% CI, 6-23%), although it is not certain that there was any increase below about 0.5 Gy. It is unclear when the increase started, as there were substantial healthy survivor selection effects for at least a decade. However, even 50 years after irradiation, there is no suggestion that the risk has diminished.

The 14% increase in risk per Gy seen in the LSS is larger than the 3% per Gy seen in women in randomized trials of breast cancer patients. Confounding is unlikely to be responsible for the difference, as controlling for smoking, alcohol consumption, diabetes, obesity, education, and occupation had almost no effect on the estimates in the LSS. Also, the Adult Health Study biennial examinations of a subset of atomic bomb survivors provide additional support for an effect of radiation on heart disease in this population through associations of radiation exposure with alterations in blood pressure and preclinical cardiovascular disease effects, including inflammatory markers and lipid metabolism . The nominal difference in risk per Gy may be due to the fact that irradiation of the breast cancer patients is fractionated. Other possible reasons include Japanese/Western baseline differences in heart disease mortality rates, differences in age at exposure, and substantially less homogeneous cardiac radiation doses during breast cancer radiotherapy.

Occupational studies

Radiologists and radiologic technologists undoubtedly received substantial fractionated radiation doses in the early years. Although individual dose estimates are not available, several studies have subdivided them by year of registration to groups likely to have received different doses. In the United Kingdom, cancer mortality was higher in radiologists than in other medical practitioners for those registered during 1897 to 1920, suggesting a radiation-related excess (RR = 1.75; 95% CI, 1.34-2.26), and mortality declined significantly with the increasing year of registration, as would be expected from the declining doses over the 20th Century. In contrast, mortality from circulatory disease for those registered during 1897 to 1921 was lower in radiologists than in other medical practitioners (RR = 0.79; 95% CI, 0.66-0.94), and there was no suggestion of any trend with year of registration. Those data appear incompatible with a substantial radiation-related excess of circulatory disease in the early group. However, a study of 90,000 US radiologic technologists found increased risks of circulatory disease in those starting work prior to 1940 compared with those starting after 1960 (RR = 1.42; 95% CI, 1.04-1.94). This increase was mostly due to cerebrovascular disease (RR = 2.40; 95% CI, 1.09-5.31), rather than ischemic heart disease (RR = 1.22; 95% CI, 0.81-1.82)

Future Prospects

Gaps in knowledge

There are major gaps in current knowledge of the structures at risk and mechanisms of damage in RRHD. It is uncertain whether there is a threshold dose of radiation to the heart below which no risk is incurred. Even at doses sufficient to increase risk, estimates of its magnitude for a given cardiac dose are still subject to considerable uncertainty, and knowledge about the shape of the dose-response relationship and factors that modify it is rudimentary, both for heart disease as a whole and for specific heart diseases. There are substantial dose inhomogeneities within the heart during radiation therapy, but it is not known which part of the heart is the most radiosensitive, nor which structure or structures at risk should be chosen as a reference point for tolerance doses in clinical practice. Additional important gaps in current knowledge include the relationship between short-term effects and long-term risk and the extent to which cardiac risk may be modified by other factors such as irradiation of other organs (e.g., kidneys), preexisting cardiovascular disease or diabetes, lifestyle factors including smoking, and interactions with cardiotoxic drugs used in combination with radiotherapy during cancer treatment.

Radiological protection

There are approximately 400 million diagnostic medical examinations performed annually in the United States and some 3.6 billion worldwide. In addition, there are almost 4 million workers monitored occupationally for potential radiation exposure in the United States and over 15 million worldwide. Present radiological protection guidelines and regulations are concerned solely with radiation-related cancers and hereditary effects and do not consider RRHD. Virtually all the exposure occurs at doses of below 0.5 Gy, and, at the present time, it is unclear whether such low doses pose any cardiac risk. Therefore, it is as yet unclear whether there will be any future need to take explicit account of RRHD in radiation protection.

Conclusions

Further knowledge about the nature and magnitude of radiation-related heart disease would have immediate application in radiation oncology. It would also provide a basis for radiation protection policies for use in diagnostic and occupational exposure.

To our knowledge, this review is first time that individuals from such a wide range of disciplines, including pathology, radiobiology, cardiology, radiation oncology, epidemiology, statistics, and diagnostic radiology have collaborated to share knowledge about RRHD. All those different disciplines have roles to play in furthering knowledge and in reducing the future burden of RRHD. Knowledge accumulated in one discipline may often be published in specialized journals that are not regularly read by those working in other disciplines, and technical terminology often impedes ready interpretation by those in other disciplines. We therefore conclude that future multidisciplinary reviews examining the topics raised here in greater depth and summarizing new findings are needed.