A woman's heart

The impact of adjuvant endocrine therapy on cardiovascular health

Abstract

Adjuvant therapy in postmenopausal women with breast cancer may contribute to the expression of underlying cardiovascular disease or expose the heart to additional toxicities. Tamoxifen remains an important component of endocrine therapy for breast cancer, although major clinical trials of the aromatase inhibitors (AIs) anastrozole, letrozole, and exemestane suggest that these agents are more effective and better tolerated alternatives to tamoxifen. The AIs inhibit the conversion of androgens to estrogen in postmenopausal women; consequently, their mechanism of action differs from that of tamoxifen. Accordingly, although it has been observed that tamoxifen has some favorable effects on cardiovascular risk, such as reducing total cholesterol levels, because of its partial estrogen-agonist properties, no such effects exist for the AIs. Some studies, particularly those that compare the AIs with tamoxifen, have suggested a less favorable impact of adjuvant AI therapy on cardiovascular risk. Comorbid conditions, including cardiovascular disease, emerge as competing causes of death as women with breast cancer continue to live longer, and the potential impact of adjuvant therapies on cardiovascular risk becomes an increasingly important consideration for clinicians. Cancer 2009. © 2009 American Cancer Society.

Patients with breast cancer are at risk for death not only from cancer or its complications but also from the other major competing causes of mortality in postmenopausal women, particularly cardiovascular disease. Cardiovascular disease is a major health concern in most industrialized countries, with a prevalence in the United States of 42.7 million in 2005 and mortality of 459,000 in 2004.1 Although a lower incidence of cardiovascular disease is observed among women than among men in the group ages 20 to 39 years, this gap narrows with advancing age and eventually exceeds that of men after age 60 years (Fig. 1)1; thus, cardiovascular disease constitutes a significant health threat among older, postmenopausal women independent of breast cancer.1, 2 Endocrine interventions remain the cornerstone of adjuvant therapy for patients who have breast cancer with estrogen-sensitive, estrogen receptor (ER)-positive disease; and, as women with breast cancer live longer in the face of more effective therapies, they will be more likely to die from causes unrelated to breast cancer or its treatment, such as cardiovascular disease or bone fractures.3 Thus, there is a risk for both cardiovascular disease4 and the cardiovascular side effects of breast cancer treatments.5 Therefore, cardiovascular disease will continue to compete as a cause of death in these patients6; it has been estimated that the number of women with breast cancer who are living with such risk is 2.3 million in the United States alone.4

Figure 1. The prevalence of cardiovascular disease (as a percentage of population) in adults aged ≥20 years by age and sex. (Data from National Health and Nutrition Examination Survey 1999-2004 (see British Heart Foundation, 20052).

The selective ER modulator tamoxifen reduces both disease recurrence and breast cancer-related mortality in women with ER-positive, early disease.7 Results from the Oxford meta-analysis of tamoxifen trials demonstrated that the benefits of adjuvant hormone therapy for breast cancer in postmenopausal women are both substantive and persistent and increase over time. Among women who were randomized to receive approximately 5 years of tamoxifen treatment, the risk of breast cancer recurrence after 15 years of follow-up was reduced by 41% (absolute reduction, 11.8%), and the risk of death was reduced by 34% (absolute reduction, 9.2%).7 It has been observed that tamoxifen has lipid-lowering effects; and, in meta-analyses, a potential cardioprotective effect of the drug was observed in which the rate of death from events such as myocardial infarction (MI) was reduced with active treatment (Table 1).7-10 Despite these benefits, however, tamoxifen also has been associated with an increased risk of events such as stroke, pulmonary embolism, and deep vein thrombosis (Table 1).7-12

Table 1. Beneficial and Nonbeneficial Effects of Tamoxifen on Cardiovascular Risk and Cardiovascular Events

Event	Reference(s)	Tamoxifen Effects
EBCTCG indicates Early Breast Cancer Trialists' Collaborative Group.
Lipid effects	Herrington & Klein 200110	Beneficial: Reduces plasma total cholesterol, reduces low-density lipoprotein cholesterol; nonbeneficial: reduces high-density lipoprotein cholesterol, increases triglycerides
Coronary heart disease	EBCTCG 2005,7 McDonald 19959	Beneficial: Lower incidence of myocardial infarction and ischemic heart disease, lower rate of heart disease-related death
Vascular effects	Stamatelopoulos 200411	Beneficial: Increases flow-mediated dilatation in brachial artery, decreases combined carotid intima-media thickness
Thromboembolic events	Braithwaite 2003,8 McDonald 19959	Nonbeneficial: Increased risk of venous thromboembolic events, including pulmonary embolism and deep vein thrombosis
Cerebrovascular events	EBCTCG 2005,7 Braithwaite 2003,8 Hooning 200612	Nonbeneficial: Increased risk of stroke

Results from the Cancer Research Network have indicated that, since 2000, the third-generation aromatase inhibitor (AI) drugs anastrozole (1 mg per day orally for 5 years), letrozole (2.5 mg per day orally for 5 years), and exemestane (25 mg per day orally until the completion of 5 years adjuvant endocrine therapy) have been replacing tamoxifen as adjuvant endocrine therapy for postmenopausal women with early breast cancer.13 Results from the Oxford meta-analysis, as noted above, demonstrated the enormous benefits of adjuvant hormone therapy in terms of reducing breast cancer recurrence and mortality, and the AIs have improved upon tamoxifen treatment still further; the AIs generally have proven to be between 15% and 25% more effective than tamoxifen in terms of reducing the relative risk of recurrence,14-16 in some cases reducing potentially lethal distant metastatic recurrences by up to 27% over tamoxifen.14 Major clinical trials, thus, have indicated a significant benefit of adjuvant AI treatment over tamoxifen in terms of disease-free survival (DFS), distant DFS, and contralateral breast cancer, but, up to this point, not in terms of overall survival.14-16 In view of the efficacy of AIs relative to tamoxifen, increased survival of women with breast cancer could allow the impact of competing causes of death to become more evident; indeed, small but significant increases in some (although not all) cardiovascular outcomes have been observed relative to tamoxifen in large clinical trials comparing the 2 treatments.14, 15, 17 This phenomenon may be related to the absence of tamoxifen-associated cardioprotective effects in women on AI therapy rather than an adverse effect of AIs per se; differences also exist in cardiovascular event definitions and the rigor with which such data are collected across the various AI trials.

With continued improvement in DFS after breast cancer diagnosis, the impact of competing causes of death and the influence of adjuvant endocrine treatments on this risk will become increasingly important. For this review, we examined the issue of heart disease in aging postmenopausal women with hormone receptor-positive breast cancer and the putative impact of adjuvant AI therapy on cardiovascular risk in this population as gleaned from the results of major adjuvant AI clinical trials.

Cardiovascular Disease Risk After Menopause

Results from the Framingham Heart Study (N = 2036) demonstrate that women have lower systolic and diastolic blood pressures as well as lower pulse pressures and mean arterial pressures than men during the fourth decade of life; however, these differences narrow by age 70 years and, in some cases (eg, in the case of systolic blood pressure and pulse pressure), reverse thereafter.18 The change in blood pressure and associated cardiovascular disease risk may be related at least in part to menopause. Obesity also has been associated with increased risk of both new hypertension and cardiovascular disease in the Framingham Study; for overweight and obese women, the age-adjusted relative risk ratios for cardiovascular disease, including angina pectoris, coronary heart disease, MI, and cerebrovascular disease, were 1.20 and 1.64, respectively.19 Weight increase is very prevalent; approximately 67% of the population in the United States is classified as overweight, and studies indicate that weight gain also occurs in patients with breast cancer who are receiving adjuvant endocrine therapy.20-22

Thus, postmenopausal women with breast cancer are at an already increased risk for cardiovascular disease,18, 23 and adjuvant endocrine therapy can increase this risk. Moreover, the populations that are studied in clinical trials are not representative of the general population. Often, major clinical studies exclude patients with pre-existing hypertension or cardiovascular disease for safety reasons; thus, the population studied may be at a comparatively reduced cardiovascular risk. The presence or absence of comorbid conditions in postmenopausal women with breast cancer also may have an impact on their cardiovascular risk. The impact of the third-generation AIs on cardiac health will be an important consideration as their use increases.

Effect of Adjuvant Aromatase Inhibitor Therapy on Serum Lipids

The influence of serum lipids, including total cholesterol, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and serum triglycerides, on atherogenesis and the cardiovascular risk profile is well established. In this regard, it is noteworthy that both tamoxifen and the AIs have demonstrated effects on serum lipids. This may explain at least in part some of the differences in cardiovascular risk that have been observed in clinical trials.

Tamoxifen

Studies have demonstrated that tamoxifen has a generally favorable impact on serum lipids. In a placebo-controlled study of postmenopausal women with breast cancer (N = 140), patients had a 12% decrease in total cholesterol and a 20% decrease in LDL-C at 2 years (P < .001).24 It is noteworthy that, in that study, HDL-C levels (“good” cholesterol) also decreased during treatment with tamoxifen; however, the decrease was only significant at 1 of 5 assessments. A longer term study of a group of patients in the Scottish adjuvant tamoxifen trial (N = 44) also indicated that tamoxifen consistently and significantly lowered cholesterol levels and that this effect ended with treatment cessation, whereas no such effect was observed in the control group (no tamoxifen).25 No significant effect on HDL-C was reported in that study, although the small number of patients may have precluded such a result. In a comparative study with another antiestrogen, torimefine, both drugs reduced total cholesterol and LDL-C levels.26 However, whereas torimefine increased HDL-C by 14%, tamoxifen decreased HDL-C levels by 5% (P = .001). Taken together, these findings provide evidence that tamoxifen decreases both total cholesterol and LDL-C, although its impact on HDL-C appears to be less favorable vis-a-vis cardiovascular health.

Aromatase inhibitors

An open-label, randomized study compared the impact of 24 weeks of treatment with anastrozole, letrozole, or exemestane on lipid parameters in healthy postmenopausal women (N = 90).27 At 12 weeks, only exemestane was associated with a significant decrease in total cholesterol from baseline (−5.5%; 95% confidence interval, −9.7 to −1.3), but there were no significant differences between anastrozole versus exemestane, anastrozole versus letrozole, or letrozole versus exemestane (overall P = .535). At 12 weeks and 24 weeks, there was a significant increase in the LDL-C:HDL-C ratio versus baseline in the exemestane group that differed significantly from both the anastrozole group and the letrozole group; this resulted from a decrease in HDL as opposed to an increase in LDL.27 Although triglyceride levels had an early, significant increase in the letrozole group compared with the exemestane and anastrozole groups at 12 weeks, this increase did not persist to the 24-week assessment. Data comparing the effects of AIs in breast cancer patients receiving adjuvant therapy are limited and mostly are derived from comparisons with tamoxifen, which, in view of its own effects on serum lipids, can confound assessments.

Aromatase inhibitors versus tamoxifen

The major clinical trials that are examining the safety of AIs include the Arimidex Tamoxifen Alone or in Combination (ATAC) study, the Breast International Group (BIG) 1-98 trial, and the Intergroup Exemestane Study (IES). At a median follow-up of 68 months, the ATAC study (N = 6186) reported a higher incidence of hypercholesterolemia with anastrozole (278 of 3092 patients) than with tamoxifen (108 of 3094 patients) that was statistically significant (9% vs 3%; odds ratio [OR], 2.73; P < .0001).17 It is worth noting that lipid profiles were obtained at the investigator's discretion in that study and were not assessed systematically in the trial, and profiles were not reported at the 100-month follow-up.28 Patients in the anastrozole group also experienced significantly more hypertension (402 events [13%] vs 349 events [11%]; hazard ratio [HR], 1.18; P = .04). Unfortunately, the information gleaned from the cardiovascular data in the ATAC trial remains a bit unclear because of the trial methodology (passive collection of these adverse events).

In contrast to other adjuvant AI trials, the BIG 1-98 trial was designed to review the potential effect of AIs on lipid metabolism comprehensively and systematically. The definition of a hypercholesterolemia event in BIG 1-98 included any measurement above the upper limit of normal (ULN) that occurred at 1 or more of the 6-month visits during the study. In addition, nearly all (90%) of these cholesterol measurements were taken under nonfasting conditions and at various times of day.29

In the BIG 1-98 primary core analysis at a median follow-up of 25.8 months (N = 8010), serum cholesterol was stable with letrozole at 6 months, 12 months, and 24 months; whereas it decreased in the tamoxifen group by 12%, 13.5%, and 14.1%, respectively.14 More patients in the letrozole group experienced hypercholesterolemia, which was defined as normal cholesterol at baseline and increasing to 1.5 times greater than the ULN at least once during treatment (173 of 3203 patients [5.4%] vs 40 of 3224 patients [1.2%]).30 Most of the events in both treatment arms (>80%), however, were grade 1 (defined as >ULN −300 mg/dL or >ULN −7.75 mmol/L according to Common Toxicity Criteria guidelines) and required no intervention.14, 29, 30 When the results were assessed over 60 months, there was a 12% reduction in total cholesterol in the tamoxifen group during the first 6 months, whereas levels in the letrozole group remained stable; thereafter, total cholesterol levels remained constant in both groups.29 These results are consistent with a benefit of tamoxifen in the first 6 months in terms of lowering total cholesterol as opposed to a detrimental effect of letrozole on lipids. Results from a subset analysis that was restricted to the monotherapy arms of BIG 1-98 (n = 2448 in the letrozole arm; n = 2447 in the tamoxifen arm) demonstrated significantly more low-grade cholesterol elevation with letrozole relative to tamoxifen at 51 months (50.6% vs 24.6%; P < .001).31 Once again, the majority of these events were grade 1 (letrozole: 985 of 1238 events; 80%), and most cholesterol values were assessed under nonfasting conditions (90.8%).31

A cardiovascular safety analysis (N = 7963) that was performed at a median follow-up of 30.1 months revealed a decrease in median total cholesterol levels for patients in both the letrozole group and the tamoxifen group. The decrease for patients who were receiving tamoxifen was immediate, whereas the decrease for patients who were receiving letrozole began at approximately 30 months after randomization.32 An increase in total cholesterol levels versus baseline occurred in more patients who were receiving letrozole than patients who were receiving tamoxifen; this explains the greater proportion of low-grade hypercholesterolemia with letrozole that was reported in the primary core analysis.32

The effect of the steroidal AI exemestane on serum lipids relative to tamoxifen has been examined in the Tamoxifen Exemestane Adjuvant Multicenter (TEAM) trial and the IES. In the TEAM lipid substudy (N = 176; n = 90 in the exemestane arm; n = 86 in the tamoxifen arm), there was a trend toward reduced triglycerides with exemestane, whereas triglycerides increased with tamoxifen. In addition, although total cholesterol tended to decrease from baseline in both groups, there was a significant decrease with tamoxifen at 12 months relative to exemestane (P = .012).33 In this study, LDL levels decreased over time with tamoxifen but fluctuated with exemestane. There was no overall change in HDL with either regimen, and the total cholesterol:HDL ratio remained stable for both regimens.

The Japanese TEAM substudy demonstrated that tamoxifen was associated with a decrease in mean total cholesterol levels and with significantly lower total cholesterol levels compared with anastrozole or exemestane. Patients who received tamoxifen also had lower LDL levels at all time points (P ≤ .0001) and reduced lipoprotein parameters compared with patients who received anastrozole or exemestane.34 Compared with anastrozole, exemestane was associated with significantly lower mean levels of cholesterol at 3 months and at 12 months (P ≤ .01 for both).34 Mean triglyceride concentrations were significantly lower with exemestane compared with tamoxifen at all time points (P ≤ .0001) and compared with anastrozole at 3 months (P ≤ .01) and at 6 months (P ≤ .0001).34 In the IES (N = 4658), which assessed the impact of a switch to exemestane after 2 to 3 years of adjuvant tamoxifen, after 55.7 months of follow-up, hypercholesterolemia trended higher in the exemestane arm versus the tamoxifen arm (204 of 2320 patients [8.8%] vs 178 of 2338 patients [7.6%]; P = .14), but the difference did not reach statistical significance (Table 2).16

Table 2. Incidence of Hypercholesterolemia: Aromatase Inhibitors Versus Tamoxifen or Placebo

		No. of Patients (%)
Trial	AI [No.] vs Comparator [No.]	AI	Comparator
AI indicates aromatase inhibitor; ATAC, Arimidex Tamoxifen Alone or in Combination; BIG 1-98, Breast International Group trial 1-98; IES, Intergroup Exemestane Study; ITA, Italian Tamoxifen Arimidex; MA.17, a trial by the National Cancer Institute of Canada Clinical Trials Group. * These were patients who had baseline values within normal limits and then had values that were >1.5 times the upper limit of normal (see Novartis Pharmaceuticals 200630). This was a check-listed adverse event that had to be reported at every patient visit. † Lipid metabolism disorders.
ATAC	Anastrozole [3092] vs tamoxifen [3094]	278 (9)	108 (3.0)
BIG 1-98 primary core analysis	Letrozole [3203] vs tamoxifen [3224]	173 (5.4)*	40 (1.2)
IES	Exemestane [2320] vs tamoxifen [2338]	166 (8.8)	141 (7.6)
ITA	Anastrozole [233] vs tamoxifen [225]	18 (8.1)†	6 (1.4)
MA.17	Letrozole [2572] vs placebo [2577]	418 (16)	411 (16)

Aromatase inhibitors versus placebo

Given the effect of tamoxifen on serum lipids, comparisons of AIs with placebo may be more informative. In a 2-year study of postmenopausal women with breast cancer who were randomized to an exemestane arm (n = 58) or a placebo arm (n = 65), there was no significant effect of exemestane on lipid parameters, except for a 6% to 9% decrease in HDL-C and a 5% to 6% decrease in apolipoprotein A1.35, 36 It is noteworthy that the on-treatment changes in lipids with exemestane were reversed within the 3 to 12 months of follow-up (ie, after exemestane discontinuation),35, 36 suggesting that there is no carryover effect of exemestane and that exemestane is unlikely to have a detrimental effect on lipids over time.

Because lipid and cardiovascular data were not reported in the National Surgical Adjuvant Breast And Bowel Project (NSABP) B-33 trial, letrozole is also the only AI that has placebo-controlled cardiovascular data from a large adjuvant trial.37, 38 The MA.17 trial, which examined the impact of extended adjuvant letrozole after about 5 years of tamoxifen, provides for the largest comparisons of AIs with placebo. At a median follow-up of 30 months, an identical incidence of hypercholesterolemia was observed (418 of 2572 patients [16%] for letrozole vs 411 of 2577 patients [16%] for placebo; P = .79), and these findings contrast with those from the BIG 1-98 trial, in which tamoxifen was the comparator.38 There was no difference in rate of hypertension between the letrozole and placebo groups (5% for both; P = .94). In the MA.17 lipid substudy (MA.17L; n = 347), despite marginally significant changes in some parameters, there were no significant differences between the treatment groups in the percentage of patients that exceeded the defined thresholds for total cholesterol, HDL-C, or LDL-C at any time point.39 The effects of letrozole and placebo on total cholesterol, LDL-C, and HDL-C over time are shown in Figure 2; at 24 months and 36 months, there was no significant difference in the mean percentage change from baseline between groups for these parameters in patients who had samples drawn under fasting conditions.39 Similarly, in MAP.1, a prevention trial of postmenopausal women with increased mammographic breast density treated with letrozole or placebo for 12 months, letrozole was associated with a transient decrease in total cholesterol at 3 months but with no other significant changes in lipid parameters, including HDL-C, LDL-C, and triglycerides at 12 months or 24 months.40

Figure 2. The mean percentage change in total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) is illustrated (top) at 6 months and 12 months and (bottom) at 24 months and 36 months in patients who were randomized to receive either letrozole or placebo in the National Cancer Institute of Canada Clinical Trials Group MA.17 trial. Asterisks denote a significant difference between the treatments (see Wasan et al, 200539).

Cardiovascular Disease

Aromatase inhibitors versus tamoxifen

Unfortunately, the information gleaned from the cardiovascular data in the ATAC trial was limited to patient self-reports, because the trial was not designed to analyze cardiac risk systematically. Results from the ATAC trial (at 68 months) indicated a trend toward a higher incidence of ischemic cardiovascular events that did not reach statistical significance (most commonly mild-to-moderate angina) between women who received anastrozole relative to women who received tamoxifen (n = 127 [4.1%] vs n = 104 [3.4%]; P = .1); the incidence of other ischemic cardiovascular events (coronary artery disease, MI, or ischemia) was not different between the groups (2% for both; P = .5).17 However, recent data indicate that grade 3 and 4 cardiovascular events (ie, angina pectoris and MI) were reported more frequently in patients who received anastrozole than patients who received tamoxifen (117 patients in the anastrozole arm vs 96 patients in the tamoxifen arm).41 Furthermore, the number of deaths caused by cardiovascular events in the ATAC trial is difficult to interpret, because there was 1 report of a greater incidence with anastrozole versus tamoxifen (2% vs 1%; 49 deaths with anastrozole vs 46 deaths with tamoxifen),17 but another report indicated that the rate of deaths from ischemic cardiovascular events was 0.8% in each group. After 100 months of follow-up, the incidence of MIs was collected, and there was no significant difference; a similar incidence was observed both on treatments (anastrozole vs tamoxifen: 37 events vs 33 events; annual rate, 0.27 vs 0.27, respectively) and after the completion of treatment (26 events vs 28 events; annual rate, 0.28 vs 0.30, respectively).28 It is noteworthy that the similarity in the off-treatment incidence of MI suggests a lack of carryover effect for anastrozole. The cardiovascular data reported in the ATAC trial remain unclear; there may have been an under-reporting of these events because of the trial's methodology concerning safety evaluations.

In the IES, at 55.7 months of follow-up, the incidence of cardiovascular events (excluding vascular thrombotic events) did not differ between the groups during treatment (382 of 2320 patients [16.5%] on exemestane vs 350 of 2338 patients [15%] on tamoxifen; P = .16) or when post-treatment events were included.16 Independent cardiologists reviewed serious cardiovascular events and also defined higher level cardiovascular groupings to facilitate comparison with other studies.16 The frequency of ischemic cardiovascular disease (8% vs 6.9% for the exemestane arm vs the tamoxifen arm, respectively; P = .17) and the frequency of MI (1.3% vs 0.8%, respectively; P = .08) during treatment also trended toward significance between treatment groups; however, among patients who experienced an MI, 71.1% in the exemestane group had hypertension at baseline versus only 31.6% of the corresponding patients in the tamoxifen group. This argues strongly in favor of regular blood pressure monitoring for patients who are receiving adjuvant exemestane.16

Unlike other adjuvant AI trials, the BIG 1-98 trial was designed to methodically review the potential effect of AIs on cardiac risk, and an in-depth analysis of cardiovascular adverse events from the primary core analysis was undertaken by Mouridsen et al14, 32 The results at a median of 30.1 months of follow-up demonstrated that the overall incidence of any cardiovascular event was similar in the letrozole arm and the tamoxifen arm (191 of 3975 patients [4.8%] vs 188 of 3988 patients [4.7%], respectively; P = .87).32 The risk for any grade 3 through 5 cardiac event was higher with letrozole versus tamoxifen (HR, 1.63; P = .004); these events were rare in the overall trial population, and the risk of grade 3 through 5 ischemic heart disease was similar for both treatments (HR, 1.46; P = .12).42 In an unplanned analysis of the grade 3 through 5 events, the incidence of grade 3 through 5 cardiac events was significantly higher with letrozole (96 of 3975 patients [2.4%] vs 57 of 3988 patients [1.4%]; P = .001), whereas the events were relatively rare in both arms.32 The incidence of ischemic heart disease was higher with letrozole than with tamoxifen, but the difference did not reach significance (45 of 3975 patients [1.1%] vs 29 of 3988 patients [0.7%]; P = .06).32 In addition, the incidence of cardiac failure was significantly higher with letrozole (26 of 3975 patients [0.7%] vs 13 of 3988 patients [0.3%]; P = .04).32

Further results from a subset analysis that was restricted to the monotherapy arm of BIG 1-98 (n = 2448 patients on letrozole monotherapy; n = 2447 patients on tamoxifen monotherapy) at 51 months of follow-up revealed that, although letrozole was associated with more cardiac events in each grade compared with tamoxifen (Fig. 3), there was no significant difference in cardiac events overall (5.5% vs 5%, respectively; P = .48), ischemic heart disease (2.2% vs 1.7%, respectively; P = .21), or cardiac failure (1% vs 0.6%, respectively; P = .14) between groups.31 There was an increase in the incidence of grade 3 through 5 cardiac events with letrozole (P < .001; Fisher exact test), but the number of events was small in each arm.31 Notwithstanding the low numbers and the similar overall relative rates of cardiac events, particularly when tamoxifen was used as the comparator, it cannot be ignored that there were more than twice as many grade ≥3 events with letrozole (74 events vs 35 events) in this population (N = 4895). Despite the relatively low overall incidence, and even in the face of the observed oncologic benefit, this disparity is of potential concern and should be acknowledged; additional follow-up and study are needed. Cardiovascular events classified as “other,” although they occurred at low frequency, were more numerous with letrozole (19 of 2448 patients [0.8%] vs 62 of 2447 patients [0.2%]; P = .014).31 Because this analysis had longer follow-up (51 months), it is important to also consider the impact of patient age on cardiovascular health, because it has been demonstrated that the prevalence of comorbidities among patients with newly diagnosed breast cancer increases with age, and cardiovascular disease is the most common comorbid condition.23 It also has been demonstrated that survival decreases with the presence of comorbid conditions, illustrating the importance of comorbid conditions like pre-existing cardiovascular disease as competing causes of death in older breast cancer patients.

Figure 3. Reported incidence of any cardiac event by grade (Gr) in the letrozole and tamoxifen arms in the subset analysis restricted to the monotherapy arms of the Breast International Group 1-98 trial (see Coates et al, 200731). NS indicates nonsignificant.

In a meta-analysis of randomized trials of adjuvant AI therapy versus tamoxifen in early breast cancer, all AIs were associated with a significantly higher risk of grade 3 and 4 cardiovascular adverse events. The absolute differences, however, were small (<1%), and 160 to 180 patients were needed to observe 1 event.20 These results do not suggest detrimental effects of the AIs but, rather, provide further support for the cardioprotective effects of tamoxifen.

Aromatase inhibitors versus placebo

Comparisons with placebo may be the best way to truly understand the impact of adjuvant AI therapy on cardiac health. When comparing letrozole with placebo in the MA.17 trial (median follow-up, 2.5 years; N = 5187), similar incidences of cardiovascular disease were observed (letrozole vs placebo: 149 of 2572 patients [5.8%] vs 144 of 2577 patients [5.6%]; P = .76).38 MIs were rare and occurred in 0.3% (9 of 2572) and 0.4% (11 of 2577) of patients in the respective groups. A recent study examining causes of patient death in MA.17 indicated that nonbreast cancer deaths accounted for 60% of the known deaths in the study, with higher rates observed in the group aged ≥70 years (72% nonbreast cancer deaths) and lower rates in the group aged <70 years (48% nonbreast cancer deaths). Women with pre-existing cardiovascular disease also were significantly more likely to die from causes other than breast cancer (P = .02).3 Thus, as women survive for longer periods free from breast cancer, they are more likely to die from causes other than breast cancer.

Thromboembolic and Cerebrovascular Events

Aromatase inhibitors versus tamoxifen

At the 68-month follow-up in the ATAC trial, there was a significantly reduced incidence of both vascular thrombotic events overall (87 of 3092 patients [2.8%] vs 140 of 3094 patients [4.5%]; P = .0004) and deep vein thrombosis events (48 of 3092 patients [1.6%] vs 74 of 3094 patients [2.4%]; P = .02) with anastrozole (n = 3092) versus tamoxifen (n = 3094), respectively.15 In the ATAC trial, thromboembolic events were recorded only during active treatment 14 days thereafter; thus, occurrences of thromboembolic events at 100 months were similar to those at 68 months.28 A detailed analysis examining the long-term effects of anastrozole versus tamoxifen on the risk of vascular thrombotic events occurring in the ATAC trial at 68 months (n = 51 pulmonary embolism, n = 147 deep vein thrombosis, n = 178 superficial thrombophlebitis) indicated a 39% decrease in vascular thrombotic events for patients on anastrozole relative to tamoxifen (OR, 0.61; P < .0001), and the rates of vascular thrombotic events were similar in the tamoxifen/combination groups (OR, 1.10, P = .38).43 Results also demonstrated that the risk of a vascular thrombotic event was greatest during the first 2 years but persisted throughout the 5-year period for patients on tamoxifen. Thus, the risk for vascular thrombotic events appears very early, in the first 2 years, in tamoxifen-treated patients.43 A significantly lower incidence of cerebrovascular events with anastrozole versus tamoxifen (2% vs 2.8%; P = .03) was reported.17 After 100 months of follow-up, the incidence of cerebrovascular accidents continued to be lower with anastrozole during treatment (20 events, annual rate 0.16 [0.64%] vs 34 events, annual rate 0.28 [1.1%]; OR, 0.59; P = .056) but not off treatment (22 events, annual rate 0.24 vs 20 events, annual rate 0.21; OR, 1.10; P = .75), which suggests a lack of tamoxifen carryover effect on cerebrovascular events.28 Deaths resulting from cerebrovascular events were not different between the anastrozole and tamoxifen groups (0.8% vs 0.9%, respectively).

In BIG 1-98, there was a significantly lower incidence of thromboembolic events and a similar incidence of cerebrovascular accidents/transient ischemic attacks (39 of 3975 patients [1%] vs 41 of 3988 patients [1%]; P = .91) with letrozole relative to tamoxifen at the 25.8-month follow-up (61 of 3975 patients [1.5%] vs 140 of 3988 patients [3.5%]; P < .001).14 In the safety analysis of the primary core analysis at a median follow-up of 30.1 months, there was a significantly lower incidence of thromboembolic events in the letrozole group versus tamoxifen (68 of 3975 patients [1.7%] vs 154 of 3988 patients [3.9%]; P < .001), and the incidence of cerebrovascular accidents/transient ischemic attacks was 1.2% in both the letrozole arm (47 of 3975 patients) and the tamoxifen arm (49 of 3988 patients; P = .92).32 Similarly, in the subset analysis of the monotherapy arms at 51 months, the reduction in thromboembolic events with letrozole remained significant (50 of 2448 patients [2%] vs 94 of 2447 patients [3.8%]; P < .001).31 Identical incidences of cerebrovascular accidents/transient ischemic attacks were observed (1.4% and 1.4%; P = .90).14, 31

The benefit of AI therapy over tamoxifen in terms of reducing vascular thrombotic events also was evident in women with previous exposure to tamoxifen (2-3 years of prior therapy). In the IES at a median follow-up of 55.7 months, vascular thrombotic events occurred in 1.2% (28 of 2320) of patients who switched to exemestane, compared with 2.3% (54 of 2338) of patients who continued on tamoxifen (P = .004).16 Similar results were observed when considering events that occurred both during treatment and after treatment (1.9% vs 3.1%, respectively; P = .01). The incidence of cerebrovascular accidents during treatment was 1.4% and was identical to that observed in patients who continued on tamoxifen (P = .98), and similar results were reported when events both during treatment and after treatment were considered (2.5% vs 2.4%, respectively; P = .89).16

Aromatase inhibitors versus placebo

Limited data are available comparing thromboembolic events and cerebrovascular events between AIs and placebo. Among postmenopausal women with previous exposures to tamoxifen (4.5-6 years) in the MA.17 trial, vascular thrombotic events occurred at low frequency (11 of 2572 patients [0.4%] on letrozole vs 6 of 2577 patients [0.2%] on placebo), and the incidence of cardiovascular disease was not statistically significant (P = .76) between letrozole and placebo.38 The incidence of cerebrovascular events (stroke and/or transient ischemic attacks) was low, with 17 events in 2572 patients on letrozole (0.7%) compared with 15 events in 2577 patients on placebo (0.6%).38 Further data on vascular thrombotic events with AIs versus placebo are likely to be gained from the NSABP B-42 trial, which is investigating the optimal duration of adjuvant letrozole versus placebo. The B-42 trial will compare 5 years of treatment with letrozole or placebo in postmenopausal women who have completed 5 years of previous endocrine therapy; it includes the incidence of arterial thromboembolic events as a secondary endpoint.44

Summary of Aromatase Inhibitor Effects on Cardiovascular Health

All AIs, steroidal and nonsteroidal alike, have similar lipid and cardiovascular profiles compared with tamoxifen; in addition, letrozole has been studied versus placebo, and no significant differences have been observed. The increased lipid levels reported with AIs versus tamoxifen may reflect the lipid-lowering effect of tamoxifen.38 There is no clear evidence of a long-term detrimental effect of AI therapy on lipids, despite the finding that significantly more patients experienced hypercholesterolemia in the AI group compared with patients who were exposed to tamoxifen in the ATAC and BIG 1-98 trials14, 15; indeed, hypercholesterolemia and cardiovascular events were observed at similar frequencies when letrozole was compared with placebo in the MA.17 trial (Table 2).38 Results from the MA.17 lipid substudy also indicated no significant adverse effect of letrozole on lipid parameters compared with placebo (Fig. 2).39

Similarly, with respect to overall cardiovascular safety, the weight of evidence from clinical trials suggests no major differences among the AIs.6, 45 Although it has been suggested that a steroidal AI may have beneficial effects on lipid metabolism,46 the effect of steroidal (exemestane) and nonsteroidal (letrozole and anastrozole) AIs on lipid parameters appears to be similar27; and, in the absence of comparative studies, it should not be assumed that the AIs differ in this regard. It is more likely that these slight increases are largely because of the lipid-lowering effects of tamoxifen rather than differing effects of the AIs on lipids based on their selectivity for the aromatase enzyme or their class (nonsteroidal vs steroidal).14, 16, 17 Although AIs may lack the lipid-lowering and cardioprotective effect of tamoxifen, the risk of cardiovascular events such as fatal MI and other forms of cardiovascular disease is well within the range observed in an age-matched, nonbreast cancer population in Finland.47 This also was observed when comparing MI rates for letrozole and tamoxifen in the BIG 1-98 trial with MI rates in the United Kingdom General Practice Research Database and the Swedish MI register.6 The increasing use of AIs as alternatives to tamoxifen also appears to provide benefit in lowering the incidence of potentially serious events, such as deep vein thrombosis, pulmonary embolism, and stroke, all of which are increased among patients who are treated with tamoxifen.14-16

Directly comparative trials under way include the Femara Anastrozole Clinical Evaluation (FACE) trial (anastrozole vs letrozole) and the MA.27 trial (anastrozole vs exemestane); these should provide more reliable data with respect to any potential differences among the AIs on cardiovascular risk factors and disease.48, 49 In the FACE trial, which recently completed enrollment, the incidence of cardiovascular events and changes in serum lipids are being evaluated as secondary safety endpoints.48

Management of Cardiovascular Risks in Breast Cancer Patients

Regardless of adjuvant therapy choices for postmenopausal women with breast cancer, several recommendations can be made to improve the cardiovascular risk profile. The use of statins, beta-blockers, and/or angiotensin-converting enzyme inhibitors and sulfonylureas or metformin to manage hyperlipidemia, hypertension, and diabetes, respectively, is appropriate.4 An emphasis on lifestyle modification (eg, smoking cessation, exercise training, dietary management) also may be helpful. Body weight and physical activity level are modifiable risk factors for cardiovascular disease and may be overlooked when considering the long-term impact of adjuvant therapy.4 Results from the Women's Healthy Eating and Living Study Group have indicated that >70% of women with breast cancer had weight gains of 2.5 to 6.2 kg during adjuvant therapy (mean change, +2.7 kg [6 pounds]).50 It is noteworthy that factors that were associated positively with weight gain in that study included adjuvant therapy and postmenopausal status, whereas negative factors included an exercise index score. Another study demonstrated that early breast cancer patients decreased their physical activity significantly in the year after diagnosis by approximately 2 hours per week, an 11% decrease (P < .05).51 In addition to increasing risks for cardiovascular disease, decreases in activity level and the associated increases in weight among women with newly diagnosed breast cancer also may have unfavorable consequences for the disease process. In 1 study, 51 postmenopausal women who were overweight or obese had significantly higher serum sex hormone levels, including estradiol, bioavailable estradiol, estrone, and estrone sulfate; whereas an inverse relation was observed for sex hormone-binding globulin (all P < .05).52 Current body mass index (calculated as weight in kilograms divided by height in meters squared) at ages 50 years and 35 years and waist-hip ratios were associated with increased breast cancer risk, particularly in women who were not using menopausal hormone therapy.53 A comprehensive review of published observational studies further suggests that women with breast cancer who are overweight or who gained weight after the diagnosis of malignancy are at greater risk for recurrence and death, and this may relate in part to unfavorable hormone profiles resulting from adiposity.54 Thus, interventions like weight reduction, increasing physical activity, and pharmacologic management of dyslipidemia, hypertension, and/or hyperglycemia can provide meaningful reductions in the risk of cardiovascular disease and can improve overall health for patients who are on adjuvant therapy for breast cancer. The general population of women receiving adjuvant endocrine therapy is more likely to have pre-existing coronary heart disease risk factors than women who are enrolled in clinical studies. Such women should be monitored and treated according to the appropriate guidelines.55, 56 On the basis of the National Cholesterol Education Program Adult Treatment Panel III guidelines, the diagnostic criteria for hypercholesterolemia are LDL-C ≥160 mg/dL, HDL-C <40 mg/dL, and total cholesterol ≥240 mg/dL.55, 56

In conclusion, the cardiovascular risk factors present in breast cancer patients reflect those in the general population. Regular assessment and management of cardiovascular risk is an important component of therapy for women with breast cancer. Periodic assessment of serum lipids and management of hypertension and weight are essential to minimize cardiovascular risk, especially in women aged >65 years, who constitute >50% of the breast cancer population.56 Studies with tamoxifen have suggested a beneficial effect of the drug on serum lipids, particularly total cholesterol and LDL (but not triglycerides) for women who receive adjuvant therapy for breast cancer. These benefits may contribute to the apparent cardioprotective effect of tamoxifen and the reductions in some cardiovascular disease endpoints. The potential benefits of tamoxifen therapy on cardiovascular health, however, must be balanced with the risk for recurrence of breast cancer. Major studies, including ATAC, BIG 1-98, and the IES, have demonstrated a significant DFS advantage for AIs over tamoxifen.14-16 The results of these studies clearly demonstrate the benefits of adjuvant hormone treatment with the AIs relative to tamoxifen in terms of reducing the risk of recurrence, especially with regard to reductions in eventually lethal distant metastatic events.14 Regardless of the potential for cardiovascular events, the substantive benefits of AI therapy with respect to breast cancer far outweigh the risks, and this should be taken into account when considering treatment options for patients. With this improved efficacy in reducing breast cancer recurrences and longer survival after breast cancer diagnosis, cardiovascular disease and other comorbidities undoubtedly will continue to emerge as competing causes of death in women who receive adjuvant AI therapy. Although not specifically designed or powered to assess cardiovascular endpoints, AI clinical trials have provided some evidence for an increase in selected cardiovascular events, particularly when tamoxifen was the comparator. This raises potential concerns that suggest the need for additional follow-up and study. The evidence suggests, however, that these findings are due to the lack of tamoxifen's putative cardioprotective effects rather than to a detrimental effect of the AIs on cardiovascular health. This is supported by the lack of significant differences in lipid and cardiovascular endpoints observed with letrozole versus placebo in the MA.17 trial (Fig. 2).38, 39 A further benefit of AI therapy is the decreased incidence of cerebrovascular events and thromboembolic disease versus tamoxifen. Limited data from the ATAC trial also suggest a lack of carryover effects from AI therapy on cardiovascular health.

Cardiovascular events that occur in AI clinical trials should continue to be monitored. Additional follow-up should elucidate further the potential impact of AIs on cardiovascular disease. Clearly, the rigorous safety assessments performed in the BIG 1-98 trial have set new standards in the reporting of cardiovascular data in adjuvant AI trials, and similar safety reporting should be encouraged in other adjuvant AI trials.14, 32

There is no evidence of an overtly increased cardiovascular risk with AI therapy in relation to placebo, despite some increases in comparisons with tamoxifen. The 2005 American Society of Clinical Oncology Technology Assessment remains accurate in its assertion that current information is insufficient to fully assess the impact of AIs on cardiovascular disease.57 Nevertheless, postmenopausal women who are receiving AIs remain at risk for such complications, because they continue to live for longer periods with their disease. Like all patients who are at risk for cardiovascular effects, these patients should be assessed regularly and should be managed to minimize cardiovascular risks during therapy.

Conflict of Interest Disclosures

An early version of this article was drafted in part with writing assistance sponsored by Novartis Pharmaceuticals. The article subsequently underwent rigorous revisions and multiple reviews by the authors before submission, and the authors assume full responsibility for the content of the final article.

M. S. Ewer is a member of the speakers' bureaus of Sanofi-Aventis, F. Hoffmann La Roche, Novartis, AstraZeneca, and Glaxo-Smith-Kline and is a member of the advisory boards of Genentech and F. Hoffmann-La Roche.

S. Gluck has received honoraria and is a member of the research and advisory boards of AstraZeneca, Novartis, and Pfizer.