Long-term risk of ischemic heart disease after adjuvant radiotherapy in breast cancer: results from a large population-based cohort

Radiotherapy (RT) is associated with ischemic heart disease (IHD) due to incidental cardiac radiation exposure, and the IHD risk increases linearly with the radiation dose to the heart and coronary arteries without an apparent threshold dose [1‐3]. Women irradiated for left-sided breast cancer (BC) receive substantially higher doses to the heart compared to women irradiated for right-sided BC, and several trials show an increased IHD risk after RT in left-sided BC [1, 4, 5]. A higher incidence of stenosis in the left anterior descending artery (LAD) has been reported after RT of left-sided compared to right-sided BC [5‐8]. A large population-based study in women with BC treated from 1976 to 2006 showed an increase in acute myocardial infarction, angina pectoris, pericarditis, and valvular heart disease after left-sided RT compared to right-sided RT, with an enhanced risk in women diagnosed with IHD prior to BC [1]. The development of new RT techniques has decreased radiation doses to the heart. Even so, doses to the anterior part of the heart, especially the LAD, may still be high [9‐11].

Many women with BC receive endocrine therapy and chemotherapy in addition to RT. Studies have shown a higher risk of cardiovascular disease in patients treated with aromatase inhibitors (AI) compared to tamoxifen, though this elevation in risk may be explained by a cardioprotective effect of tamoxifen [12, 13]. Cytotoxic agents, especially regimens including anthracyclines, are associated with an increased risk of congestive heart failure [14, 15]. Anthracyclines combined with RT have been associated with a higher risk of heart disease in left-sided BC compared to RT alone, suggesting an additive effect [16, 17]. The aim of the present study was to examine the risk of IHD in women with BC receiving RT and age-matched women without a history of BC in a large population-based cohort. The possible added risk from combining RT with endocrine therapy and chemotherapy was also studied.

The present study showed no difference in comorbidity between women with BC and women without BC at baseline, but at follow-up, a significantly lower risk of IHD was seen in women with BC compared to women without BC. When women with BC receiving left-sided RT were compared to women receiving right-sided RT, a statistically significant 18% increased risk of IHD was seen. The HR’s for IHD in women with left-sided RT compared to right-sided RT were even higher in women with more extensive lymph node metastasis, and when endocrine therapy and chemotherapy were added to RT. In concordance with the findings in the Cox regression, the cumulative incidence of IHD was higher for women receiving left-sided RT compared to right-sided RT. The difference was observed within the first years after RT, and it continued to increase with longer follow-up.

In line with other studies [22, 23], a lower risk of IHD was seen in women with BC compared to women without BC. The analysis was adjusted for educational level, CCI, and the number of previous IHD events, but the findings may, however, be due to a confounding by a different distribution of lifestyle factors not captured by our adjustment. For example, some combinations of risk factors are negatively associated with IHD but positively with BC such as differences in child-bearing patterns, in the adoption of a healthy lifestyle and in adherence to mammography screening [24, 25].

The comparison to women without BC further shows that retrospective studies on cardiac toxicity will be biased if the background population is used as the control group.

The findings of lower risk of IHD in women with BC compared to women without BC were only observed in women recommended RT and other adjuvant therapies. Women with BC who received no RT had more comorbidities according to CCI, including previous IHD events, and a selection of healthier women to active therapies has probably occurred. Retrospective comparisons between women that did receive treatments with women that did not would likely be severely biased.

Thus, the risk of IHD after RT in BC could not be assessed by comparing BC patients with age-matched women without BC, nor by comparing women with BC who received RT to women with BC who did not. Instead, we chose to analyze the risk of IHD by comparing left- and right-sided BC. The mean heart doses and doses to coronary arteries are considerably higher after left-sided RT [6, 9, 26], but as cardiac structures also receive doses after right-sided RT, this comparison probably will underestimate the total added risk of IHD after RT in BC. In a recent systematic review of trials from 2010 to 2015, mean heart doses were 5.2 Gy and 3.7 Gy in left-sided and right-sided RT, respectively [11]. We recently showed considerably lower doses in a dosimetry study of 182 BC patients selected from the same registry as the present study, although including only patients undergoing coronary angiography due to a cardiac event [27]. The median mean heart dose was 0.6 Gy (interquartile range (IQR) 0.4–1.0) after right-sided RT and 2.7 Gy (IQR 1.7–4.2) after left-sided RT. Moreover, the left/right RT dose difference was higher in the mid and distal LAD, where median mean doses were 3.6 Gy (IQR 2.4–6.2) and 26.7 Gy (IQR 7.0–41.0) after left-sided RT, and 0.5 Gy (IQR 0.1–0.8) and 0.3 Gy (IQR 0.1–0.8) after right-sided RT, respectively [27].

A higher risk of IHD in women with BC receiving left-sided RT compared to right-sided RT has also been shown in previous studies [1, 4, 5, 8]. In this study, we show that this increase in risk persisted in women receiving RT for BC during 1992 to 2012. Most women were likely to receive three-dimensional conformal radiotherapy (3DCRT) given with tangential fields, but were treated before breathing adaption techniques, e.g., deep inspiratory breath-hold (DIBH), were implemented. The pattern of cumulative incidence further shows that the increase in the incidence of IHD in women with left-sided RT starts in the early years after RT and continues throughout follow-up.

An even higher risk of IHD was seen in women with more advanced pathological lymph node status. It is plausible that this group received more extensive RT including axillary and supraclavicular lymph nodes, and for some patients, the internal mammary chain (IMC), which is associated with higher cardiac radiation doses [6]. An increase in the risk of IHD was also seen when endocrine therapy and chemotherapy were added to RT in women with left-sided BC compared to right-sided BC. Anthracycline-based chemotherapy significantly increased the risk of cardiac death according to the 2005 EBCTCG meta-analysis, while tamoxifen use was associated with a lower cardiac death risk although the difference was not statistically significant (p = 0.06) [28]. A meta-analysis of published data from trials comparing aromatase inhibitors and tamoxifen reported a significant 30% increase in cardiovascular morbidity with aromatase inhibitors [29]. Whether aromatase inhibitors increase cardiovascular disease compared to no endocrine treatment is still unclear. Two recent reports [16, 17] indicate that adjuvant anthracycline treatment may increase the risk of IHD after RT in breast cancer, especially when combined with RT to the IMC [17]. Since women who received adjuvant endocrine therapy and chemotherapy also had more advanced pathological lymph node status, the increased risk of IHD in women receiving both RT and systemic therapy has to be interpreted with caution.

Strengths of this study include the population-based setting, the large size of the cohort, and the linkage with other registries enabling adjustments for comorbidity and socioeconomic factors.

Several limitations need mentioning. Causes of death were not analyzed since in women with BC, dying of BC is in this study design a competing risk to developing IHD, and the analysis comparing women with BC to women without BC diagnosis may bias the risk of IHD in women with BC. Another limitation is that information concerning individual radiation doses and targets were not available. However, most women received a dose of 2 to 50 Gy over 5 weeks, according to Regional BC treatment guidelines throughout most of the study period. A few patients may have received hypofractionated RT schedules of 2.66 to 42.56 Gy or of 2.67 to 40.05 Gy, as the guidelines allowed these schedules for selected cases from 2010. Most women with lymph node metastases also received RT to the axilla, the supraclavicular fossa, and in some cases the IMC, according to the Regional BC treatment guidelines. A further limitation is that the median follow-up of 8.1 years for women with BC may be too short to draw conclusions regarding the risk of IHD for those women receiving RT during the latter part of the inclusion period. Previous studies have shown that the risk of radiation-induced IHD begins within a few years from RT and continues for a long period of time, with the highest risk seen between 10 and 14 years from the RT [2], and it is thus likely that the elevation in risk of IHD in women receiving left-sided RT in the present study will be even more pronounced at longer follow-up. One cannot exclude the possibility that women with a history of left-sided BC are observed with more caution due to awareness of side effects of the RT, which may lead to more frequent cardiovascular examinations and thus an increased opportunity to detect IHD. On the other hand, there were proportionally slightly fewer women with left-sided BC selected for RT compared to right-sided BC in this study, suggesting an initial selection of healthier women for left-sided RT.

We found an increase in the risk of IHD in women with left-sided BC compared to right-sided BC, given adjuvant RT during 1992 to 2012. Most of these women likely received 3DCRT given with tangential fields, but without DIBH techniques. The risk of radiation-induced IHD is expected to decrease when DIBH or other techniques that lower the cardiac and coronary doses are used. The results of this study emphasize the importance of fully implementing and further developing heart-sparing techniques, without compromising the target coverage and the beneficial effects of RT. Even though the rate of IHD events after RT is relatively low, long-term side effects of adjuvant treatment have to be taken into consideration, to ensure health and quality of life for BC survivors.