Venous thromboembolic events (VTEs) are among the most serious and feared complications associated with adjuvant hormone therapy for breast carcinoma. VTEs typically occur in the form of lower-extremity deep venous thrombosis (DVT) with or without pulmonary embolism (PE), but they also may involve the cerebral venous sinuses and the deep veins of the upper extremities and abdomen. Superficial venous thrombophlebitis (SVT) occupies a less ominous part of the VTE spectrum but can result in limb pain, edema, and subsequent DVT. Although lower-extremity DVT does not cause death per se, it is associated with significant morbidity and may precipitate fatal and nonfatal PE. Between 20–70% of patients with lower-extremity DVT develop the postthrombotic syndrome,1, 2 whereas 3-month mortality rates of up to 15% have been reported after PE.3 Population-based studies have shown that the annual incidence of DVT and PE increases exponentially with age, from approximately 1 in 100,000 population age < 20 years to 1 in 10,000 population ages 20–40 years, to 1 in 1000 population ages 40–75 years, and to 1 in 100 population age > 75 years.4–7
Patients with breast carcinoma may be exposed to a multitude of VTE risk factors alone or in combination during the course of their disease. The hypercoagulability of malignancy, the presence of underlying inherited or acquired (e.g., antiphospholipid antibodies) hypercoagulable states, and exposure to situational risk factors (such as prolonged immobilization, surgery, central venous access devices, and chemotherapy itself) all can promote or trigger pathologic thrombosis. The impact of these risk factors on the overall risk of VTE can be illustrated by the fact that patients with breast carcinoma have an increased risk of VTE relative to patients of the same age but without cancer and have an even greater, exponentially increased VTE risk during chemotherapy (e.g., cyclophosphamide, methotrexate, and 5-fluorouracil).8, 9 The risks with newer chemotherapy regimens (e.g., doxorubicin and cyclophosphamide) and with sequential instead of concomitant chemotherapy plus hormone therapy may be different.
From the 1940s through the 1970s, permanent ovarian ablation (either surgically or by pelvic irradiation) and the estrogen diethylstilbestrol constituted the primary methods of adjuvant endocrine therapy for breast carcinoma in premenopausal and postmenopausal women, respectively.10–12 Since the introduction of synthetic, orally administered progestins in 196713 and the approval of tamoxifen (TAM) by the U.S. Food and Drug Administration (FDA) in 1977,14 the importance of adjuvant hormone therapy for women with breast carcinoma has become well established by a number of randomized controlled trials.15–18
The five classes of hormonal agents that are available currently or that are under study for breast carcinoma therapy include progestational agents, selective estrogen receptor modulators (SERMs), aromatase inhibitors, selective estrogen receptor down-regulators (SERDs), and luteinizing hormone-releasing hormone (LHRH) analogues (Table 1). The majority of SERMs exert estrogen antagonist activity in the breast tissue and central nervous system while exhibiting full or partial estrogen agonist effects on bone, endometrium, and the cardiovascular system.19, 20 Arzoxifene, in turn, is a SERM with potent estrogen antagonist activity in the breast and endometrium.21 SERDs are pure antiestrogens that, unlike SERMs, exert estrogen antagonist activity in all tissues.20, 22 Aromatase inhibitors halt estradiol synthesis by blocking aromatase, a cytochrome P450 enzyme that catalyses the conversion of testosterone and androstenedione to estradiol and estrone, respectively.22–24 Because peripheral conversion of androgens is the main source of estradiol in postmenopausal women,22, 23 and aromatase inhibitors do not block ovarian estrogen synthesis completely,22–26 these drugs typically are used in postmenopausal (or postbilateral oophorectomy) women, but not in premenopausal women, with breast carcinoma.10, 22, 24, 25, 27 LHRH analogues have an inhibitory effect on the hypothalamic-pituitary-ovarian axis and ultimately cause suppression of ovarian function and marked decrease in plasma estradiol levels.22, 23, 28 It has been demonstrated that the reversible medical ovarian ablation produced by LHRH analogues is as effective as the irreversible surgical and radiotherapeutic methods of ovarian ablation in the treatment of premenopausal women with metastatic breast carcinoma.10, 28–30
Table 1. Hormonal Agents Available or under Study for the Prevention and Treatment of Breast Cancer
|Progestational agents||SERMs||SERD||Aromatase inhibitors||LHRH analogues|
|Megestrol acetatea||Tamoxifena||Fulvestranta||Anastrozolea||Goserelin acetatea|
|Medroxyprogesterone acetate||Raloxifeneb|| ||Letrozolea||Buserelinc|
| ||Toremifenea|| ||Exemestanea||Leuprolide|
| ||Droloxifenec|| ||Fadrozolec||Tryptorelin|
| ||Idoxifened|| ||Vorozolec|| |
| ||Arzoxifened|| || || |
| ||Lasoxifened|| || || |
When the first diagnoses of women developing VTE during TAM therapy were reported,31, 32 physicians became aware that such therapy could represent an additional risk factor for VTE in patients with breast carcinoma. Although many perceive the risk of VTE associated with most other hormonal agents as negligible, to our knowledge to date there have been no in-depth analyses of the data pertaining to the risk of VTE associated with those drugs. This systematic review was intended to update, summarize, and discuss the strengths and limitations of the available evidence on the association between VTE and agents used as adjuvant hormone therapy for breast carcinoma. Better understanding of the data supporting or refuting current perceptions should help guide more evidence-based decisions in the future.
MATERIALS AND METHODS
- Top of page
- MATERIALS AND METHODS
- RESULTS AND DISCUSSION
We performed single keyword and Boolean PubMed computerized literature searches for English language articles relating to humans published between January 1966 and December 2003 using the following keywords and phrases: deep venous thrombosis, thrombophlebitis, venous thromboembolism, pulmonary embolism, breast cancer, selective estrogen receptor modulators, aromatase inhibitors, and LHRH analogues, as well as each individual drug name. We also identified and evaluated articles by back referencing from original and relevant review articles published after 1998. Case reports, abstracts, and meeting proceedings, as well as articles from nonpeer-reviewed journals, were not included in this review. If an original study did not provide a measure of VTE risk, then we calculated the unadjusted risk whenever the study presented raw data that allowed us to reconstruct 2 × 2 tables.
All original studies that provided a measure of VTE risk were included. We did not exclude articles because of perceived poor methodology. A key objective of the current study was to present all peer-reviewed data concerning the topic as a means of highlighting the limitations of the available evidence that formulates physicians' perceptions. Studies were analyzed with regard to trial design, breast carcinoma staging, specific drugs used, definition of VTE outcomes, methods of VTE case ascertainment, and the presence of concomitant VTE risk factors. Data regarding the first-generation aromatase inhibitor aminoglutethimide were not included because the drug is used no longer.
RESULTS AND DISCUSSION
- Top of page
- MATERIALS AND METHODS
- RESULTS AND DISCUSSION
The literature pertaining to the risk of VTE associated with adjuvant endocrine therapy for patients with breast carcinoma includes 2 Phase I studies, 7 Phase II studies, 20 Phase III studies, 2 cohort studies, 1 case–control study, and 17 randomized controlled trials (Tables 2–4).33–81 To our knowledge, TAM is the only hormonal agent that has been studied both in premenopausal women and postmenopausal women, as well as in women who are healthy or who have early-stage or advanced-stage breast carcinoma.
Table 2. The Risk of Venous Thromboembolic Events in Healthy Women Enrolled in Randomized, Double-Blind, Placebo-Controlled Cancer Prevention Trials
|Study||Study drug||Relative risk of VTE (95% CI)|
|VTE risk||DVT risk||PE risk||SVT risk|
|Fisher et al., 199833||Tamoxifen||—||1.60 (0.91–2.86)||3.01 (1.15–9.27)||—|
|Veronesi et al., 199834||Tamoxifen||2.96ab||1.93a||1.0a||3.9a|
| P value|| ||0.0053c||—||—||—|
|Powles et al., 199835||Tamoxifen||1.75ad||—||—||—|
|IBIS Investigators, 200236||Tamoxifen||2.4 (1.5–4.4)b||4.4a||1.3a||3.04a|
| P value|| || ||0.0005c||0.68c||0.004c|
|Cummings et al./MORE, 199937||Raloxifene||3.1 (1.5–6.2)d||—||—||—|
Table 3. The Risk of Venous Thromboembolic Events in Women with Early-Stage Breast Carcinoma
|Study||Study design||Risk of VTE (DVT and PE) according to treatment assignment (95% CI)a||P valueb|
|TAM vs.||Chemotherapy + TAM vs.|
|Fisher et al., 198938||DB, RCT||RR, 6.0c|| || || || || || ||Not given|
|Fisher et al., 199639||DB, RCT||RR, 3.25c|| || || || || || ||Not given|
|Fisher et al., 199940||DB, RCTd||RR, 4.0c|| || || || || || ||Not given|
|Tormey et al., 199641||Randomized|| ||2.0% vs. 0%e|| || || || || ||Not given|
|Saphner et al., 199142||Cohort|| ||RR, 5.5 (0.5–61.5)e|| || || ||RR, 3.8c|| || |
|McDonald et al., 199543||Cohort|| ||HR, 0.4 (0.18–0.90)f|| || || || || || |
|Meier and Jick, 199844||Case–control|| ||OR, 7.1 (1.5–33)d|| || || || || || |
|Rutqvist et al., 199345||Randomized|| ||RR, 1.06 (0.7–1.60)g|| || || || || || |
|Boccardo et al., 200046||Randomized|| ||1.0%h|| || || || || || |
|Holli et al., 200047||Open-label, randomized|| || ||RR, 1.33|| || || || ||> 0.5|
|Eiermann et al., 199948||Phase III, DB, RCT|| || || ||0.0% (TAM) 0.6% (letrozole)|| || || ||Not given|
|The ATAC Trialists' Group, 200249||DB, RCT|| || || || ||RR, 1.70 RR, 2.0i|| || ||0.02|
| || || || || || ||RR, 1.33j|| || || |
|Fisher et al., 198950||RCT|| || || || || ||RR, 5.3c|| ||Not given|
|Fisher et al., 199051||RCT|| || || || || || ||RR, 2.4c||Not given|
|Pritchard et al., 199652||RCT|| || || || || || ||RR, 7.7c||< 0.0001|
|Wils et al., 199953||RCT|| || || || || || ||RR, 8.67||0.045|
Table 4. The Rates of Venous Thromboembolic Events Reported from Studies in Women with Advanced Breast Carcinoma
|Study||Study design||VTE rates (%)a||P value|
|Buzdar et al., 199454||Phase I trial|| || || ||3.3b|| || || || || || || || || |
|Münster et al., 200155||Phase I trial|| || || || || ||6.3b|| || || || || || || |
|Bellmunt and Salé, 199156||Phase II trial|| || || ||2.2b|| || || || || || || || || |
|Haarstad et al., 199257||Phase II trial|| || || ||3.8c|| || || || || || || || || |
|Vogel et al., 199358||Phase II trial|| || ||1.0b|| || || || || || || || || || |
|Boccardo et al., 199759||Phase II trial|| || || || || || || || || || ||5.9b|| || |
|Jones et al., 199960||Phase II trial|| || || || || || || || || ||1.1b|| || || |
|Buzdar et al., 200361||Phase II trial|| || || || || ||1.0b|| || || || || || || |
|Paridaens et al., 200362||Phase II, randomized|| ||0.0b|| || || || || || || ||1.6b|| || ||Not given|
|Allegra et al., 198563||Phase III, RCT||1.5d||4.8d|| || || || || || || || || || ||Not given|
|Hayes et al., 199564||Phase III, RCT|| ||1.4b||1.8b|| || || || || || || || || ||NS|
|Pyrhönen et al., 199765||Phase III, DB, RCT|| ||5.5e||5.1e|| || || || || || || || || ||NS|
|Milla-Santos et al., 200166||Phase III, DB, RCT|| ||8.0e||2.8e|| || || || || || || || || ||NS|
|Buzdar et al., 200267||Phase III, DB, RCT|| ||0.4f|| ||0.1f|| || || || || || || || ||Not given|
|Arpino et al., 200368||Phase III, RCT|| ||0.0f|| || ||0.6f|| || || || || || || ||Not given|
|Buzdar et al., 199669||Phase III, RCT||4.7e|| || || || || || || || ||2.6e|| || ||NS|
|Buzdar et al., 199770||Phase III, RCT||4.7e|| || || || || || || || ||2.7e|| || ||Not given|
|Gershanovich et al., 199871||Phase III, RCTg|| || || || || || || || ||0.3b|| || || || |
|Dombernowsky et al., 199872||Phase III, DB, RCT||1.6b|| || || || || || || || || || || || |
|Dombernowsky et al., 199872||Phase III, DB, RCT||3.2h|| || || || || || || ||0.3h|| || || ||Not given|
|Thürlimann et al., 199673||Phase III, RCT|| ||1.0e|| || || || || || || || || ||0.0e||0.12|
|Bonneterre et al., 200074||Phase III, DB, RCT||1.2,b 1.8h|| || || || || || ||0.6h|| || || || ||Not given|
|Nabholtz et al., 200075||Phase III, DB, RCT|| ||2.2i|| || || || || ||1.2i|| || || || ||Not given|
|Mouridsen et al., 200176||Phase III, DB, RCT|| ||2.4e|| || || || || || ||1.3e|| || || ||Not given|
|Mouridsen et al., 200377||Phase III, DB, RCT|| ||0.22f|| || || || || || ||0.22f|| || || || |
|Milla-Santos et al., 200378||Phase III, DB, RCT|| ||8.0e|| || || || || ||0.0c|| || || || ||Not given|
|Howell et al., 200279||Phase III, RCT|| || || || || || ||3.7e||1.7e|| || || || ||Not given|
|Osborne et al., 200280||Phase III, DB, RCT|| || || || || || ||3.4e||6.7e|| || || || ||Not given|
|Rose et al., 200381||Phase III, RCT|| || || || || || || ||0.3b||0.6b|| || || ||Not given|
It is interesting to note that only two studies specified their method of VTE case ascertainment. In the randomized trial by Pritchard et al., “most” cases were confirmed by objective testing,52 whereas the case–control study by Meier and Jick included only confirmed cases of VTE.44 The fact that the remaining studies did not state which methods, if any, were used to confirm the diagnosis of VTE does not imply that all diagnoses were made on clinical grounds alone. However, it has been demonstrated that < 50% of patients who are diagnosed with DVT or PE on clinical grounds alone will have the diagnosis confirmed by objective imaging tests.82–85 Thus, if some studies allowed the inclusion of cases of VTE without objective confirmation, then it likely would have led to a number of false-positive diagnoses, thereby resulting in an overestimation of the risk of VTE. In addition, none of the studies performed mandated VTE screening of all enrolled patients. Such methodological limitations may have led to significant under detection of VTE and, thus, under estimation of risk, because asymptomatic VTE may have been missed altogether. Thus, based on the methods used in most of the studies in question, accurate VTE rates are elusive.
VTE Risk During Breast Carcinoma Chemoprevention in Healthy Women
Based on four prospective, double-blind, placebo-controlled trials, the relative risk of VTE associated with TAM use in healthy women is increased by twofold to threefold (Table 2).33–36 This increased relative risk is comparable to that associated with hormone replacement therapy (HRT) or the use of oral contraceptives.86 In the study by Veronesi et al., the risk of any VTE (with a composite endpoint including all DVT, PE, and SVT cases) was increased significantly, but not the risk of DVT or PE alone or combined.34 Therefore, isolated SVT appeared to be the most common form of VTE promoted by TAM in otherwise healthy women.
In the International Breast Cancer Intervention Study (IBIS-I) trial, 42% of the VTE in TAM users occurred within 3 months of major surgery or prolonged immobilization and, thus, may be deemed “situational”.36 Therefore, it is unclear whether the risk of idiopathic (i.e., spontaneous) VTE still would be increased significantly after the removal of patients with situational DVT from the analysis.36 Although both situational and idiopathic DVT can result in PE, situational DVT is less likely to result in recurrent VTE provided the situational risk has passed.
One important limitation of the three European studies of TAM therapy for breast carcinoma prevention was that many of the enrolled women (up to 40%) also were receiving HRT.34–36 Because HRT itself is associated with a twofold to threefold increased relative risk of VTE,86 concurrent use of HRT and TAM is a potential confounder that may have resulted in an overestimation of the TAM-related VTE risk in healthy women. The IBIS-I investigators commented that there appeared to be no synergy between concomitant TAM and HRT use with regard to the observed rates of VTE.36
A recent meta-analysis of the randomized breast carcinoma prevention trials that compared TAM with placebo concluded that the risk of VTE, excluding isolated SVT, was increased nearly 2-fold in TAM users (relative risk, 1.9; 95% confidence interval, 1.4–2.6; P < 0.0001).87 The Multiple Outcomes of Raloxifene Evaluation trial revealed a 3-fold increased risk of VTE in raloxifene users and a 76% reduction in the incidence of newly diagnosed breast carcinoma compared with nonusers.37 This increase in VTE risk is comparable to, but not worse than, the risk associated with TAM. The Study of Tamoxifen and Raloxifene (STAR) trial is an ongoing, large (> 20,000 patients projected), randomized, direct comparison study of TAM and raloxifene for the prevention of breast carcinoma in high-risk, postmenopausal women.20, 21 The STAR trial hopefully will shed more light on the comparative risk of VTE in healthy individuals who are exposed to TAM or raloxifene.
It is unclear whether TAM chemoprevention confers an exponentially increased risk of VTE in carriers of inherited hypercoagulable states compared with noncarriers. To our knowledge to date, the only published study that attempted to address this issue was a nested case–control study that was based on the IBIS-I cohort of 96 patients with VTE.88 However, only 35 patients (36%) had blood samples available for specific testing, and none tested positive for the factor V Leiden or prothrombin G20210A mutations.88 Therefore, the VTE risk associated with TAM therapy in women with inherited hypercoagulable states remains unknown.
VTE Risk during Hormone Therapy for Early-Stage Breast Carcinoma
Studies of TAM in women with early-stage breast carcinoma evaluated patients who had undergone surgery, with or without chemotherapy and/or radiation therapy, and randomized them to treatment with either TAM or placebo (Table 3). Three studies randomized patients to either TAM therapy or observation alone and therefore were not placebo-controlled.41, 45, 46 The study by Saphner et al. was comprised of a cohort of women from seven consecutive Eastern Cooperative Oncology Group trials, one of which included patients with Stage III breast carcinoma.42
Compared with women treated with placebo or observation only, those who were treated with TAM had an estimated 1.5-fold to 7.1-fold increased risk of VTE (Table 3). In the study by McDonald et al., 40% of women with VTE among TAM users occurred after major inpatient surgery and therefore should be viewed as situational in nature.43
The study by Saphner et al. reported that postmenopausal women who received TAM were three times more likely to develop VTE compared with premenopausal women receiving the same therapy and that the risk of VTE associated with TAM use increased with age.42 In 1996, Fisher et al. also reported that the greatest increase in VTE was observed among women age > 50 years, and particularly in women age > 60 years.39
Eiermann et al. used TAM or letrozole preoperatively in women with nonmetastatic breast carcinoma.48 In that study, no VTEs were reported among 170 TAM users, and 1 episode of PE was reported among 157 letrozole users (Table 3).48
The Arimidex, Tamoxifen Alone, or in Combination trial is to our knowledge the largest study published to date in which an aromatase inhibitor has been compared directly with TAM in patients with early-stage breast carcinoma.49 The study enrolled > 9000 women and reported a small but significantly greater risk of VTE among TAM users compared with anastrozole users; the risk was slightly greater in women who were receiving both drugs in combination compared with women who were receiving TAM alone (Table 3).49 The incidence of DVT and PE during the study period (median follow-up, 33.3 months) among the 3092 women (mean age, 64 years) who were prescribed anastrozole was 1.0%.49 This absolute risk of VTE (approximately 3.6 per 1000 individuals per year) still is greater compared with the estimated risk in untreated, healthy women.4–7
More recently, Goss et al. published a randomized, placebo-controlled trial of letrozole in women with early-stage breast carcinoma who already had received 5 years of TAM.89 Although the rates of cardiovascular events between the letrozole group and the placebo group was similar in their study (4.1% vs. 3.6%; P = 0.40),89 specific rates of VTE were not reported. Thus, no comment can be made regarding the letrozole-associated VTE risk in this group of patients.
In addition, five studies of patients with early-stage breast carcinoma reported data concerning the risk of VTE in relation to concurrent hormone therapy and chemotherapy.42, 50–53 The relative risk of VTE in women receiving chemotherapy plus TAM was threefold to eightfold greater compared with the risk for women treated with TAM alone,42, 51–53 threefold to fivefold greater compared with the risk for women treated with chemotherapy alone (without adjuvant TAM),42, 50 and 20fold greater compared with the risk for women who received placebo or who only were observed.42 Thus, chemotherapy and hormone therapy should be viewed as independent risk factors for VTE. The studies by Rutqvist et al.45 and Boccardo et al.46 do not allow for any conclusions regarding the TAM-associated risk of VTE, because the former authors reported the risk of all VTE and stroke as a composite endpoint without reporting the risk of VTE alone,45 whereas the latter authors reported the rate of VTE in users of TAM and goserelin acetate combined.46
VTE Risk during Hormone Therapy for Patients with Advanced Stage Breast Carcinoma
All 28 studies shown in Table 4 reported the incidence rates of symptomatic VTE occurring during the study periods, and 19 studies compared 2 different hormonal agents directly. However, after excluding Phase I and II trials, the methodologic limitations of most Phase III trials significantly impaired the ability to determine and compare the true risk of DVT and PE associated with the therapies under study. Nine studies reported the rates of “all thromboses” (including arterial thromboses) but not the rates of VTE alone65, 66, 69, 70, 73, 76, 78–80; whereas two other studies analyzed VTE outcomes, including DVT, PE, and isolated SVT, without specifically reporting the rates of isolated SVT.63, 75 Moreover, nonblinded studies that compared TAM or megestrol acetate with newer SERMs or aromatase inhibitors may have been influenced by diagnostic suspicion and referral bias.63, 64, 68–71, 73, 78 This is because clinicians may have been more alert and responsive to signs and symptoms of VTE in patients who were using the older drugs, which for decades have been perceived as being associated with a significant risk of VTE.
The fact that none of the studies were designed specifically to address the risk of VTE also may have led to bias. Cancer is a strong, independent risk factor for VTE, and DVT often is asymptomatic as well as potentially under diagnosed, particularly in patients with advanced disease who may have pulmonary or limb symptoms attributed to their underlying malignancy rather than VTE.90 Thus, in the absence of baseline (prehormone therapy initiation) imaging assessments, it is possible that some asymptomatic cases of DVT may have been present already at the time HRT was initiated but only became clinically declared or detected during the study follow-up. This would result in overestimation of the risk of VTE. Conversely, a lack of routine imaging assessments may have resulted in failure to detect asymptomatic DVT, hence leading to an under estimation of the VTE risk.
The following series of conclusions and statements can be derived from the available evidence regarding the VTE risk associated with adjuvant hormone therapy for breast carcinoma.
Based on the available literature, a truly accurate rate of VTE associated with hormone therapies for the prevention and/or treatment of breast carcinoma is impossible to determine because most studies failed to state their methods of VTE case ascertainment explicitly and because none employed routine VTE screening of all enrolled patients. These important methodologic limitations may have resulted in either overestimating or underestimating the risk of VTE, and they may limit all risk estimates regarding the risk of symptomatic VTE.
In healthy women, the relative risk of symptomatic VTE associated with TAM or raloxifene use is increased 2-fold to 3-fold compared with nonusers. Such relative risk translates into an estimated absolute risk of 2 or 3 cases per 1000 population per year,33, 91 which is similar to the absolute risk of VTE observed in women using HRT.92 Because the risk of VTE increases with increasing age, such risk during breast carcinoma chemoprevention is expected to be higher in older compared with younger women.
To our knowledge, the risk of VTE during breast carcinoma chemoprevention in women who are known carriers of inherited hypercoagulable states or who have a past history of idiopathic VTE is unknown. Women with the factor V Leiden or the prothrombin G20210A mutation have exponentially increased VTE risk during HRT and oral contraceptive use,93–98 and the risk of recurrent VTE during HRT is increased significantly in women who have a past history of VTE.99 Thus, extrapolation from the HRT and oral contraceptives literature suggests that it is prudent to view TAM and raloxifene as contraindicated in women who have a past history of VTE and at least relatively contraindicated in asymptomatic carriers of inherited hypercoagulable states. The development of any form of thrombosis during breast carcinoma chemoprevention should prompt drug discontinuation because of a likely accentuated risk of nonfatal and fatal VTE recurrence.
TAM use increases the risk of symptomatic VTE in women with early-stage breast carcinoma by a factor of twofold to sevenfold, but the risk is much more pronounced when chemotherapy and TAM are used in combination. The absolute risk of VTE is difficult to estimate, because it will be dependent on the coexistence of multiple risk factors. In women with breast carcinoma who have a past history of VTE, a strong family history of VTE, or known hematologic risk factors for VTE, the benefits of adjuvant hormone therapy must be weighed against the risks.
In healthy women and in those with early-stage breast carcinoma, major surgery appears to be the single most important determinant (trigger) of VTE during TAM therapy.
The risk of VTE associated with TAM therapy in women with early-stage breast carcinoma appears to increase with age.39, 42 TAM is not necessarily more toxic in older women; rather, older individuals simply may have a greater incremental increase in absolute VTE risk because of a higher baseline risk.
Therapy with anastrozole is an independent risk factor of VTE because anastrozole use increases the incidence of VTE compared with the incidence among healthy women who do not receive any form of hormone therapy.
There is mounting evidence suggesting that the use of anastrozole results in a lower risk of VTE than the use of TAM in women with early-stage disease. Nevertheless, such lower risk does not imply “negligible” or “nonexistent” risk.
To our knowledge, there currently is insufficient evidence in women with advanced breast carcinoma to support any assumptions that therapy with newer hormonal agents indeed is associated with a lower risk of VTE compared with therapy with TAM or megestrol acetate. The evidence is flawed with regard to determining the true risk of DVT and/or PE associated with various hormonal agents, because many studies do not report arterial and VTE complications (or isolated SVT events) separately. Moreover, in nonblinded studies, the lower rates of VTE noted with newer agents should be interpreted with caution, because they may have been impacted by diagnostic suspicion and referral bias.
Conclusive evidence supporting or refuting a differential risk of VTE between different agents can be obtained best by double-blind, randomized trials that minimize the possibility of bias; explicitly report the incidence of VTE; and utilize predetermined diagnostic algorithms for the screening, detection, and confirmation of VTE. Until such trials become available, the decision regarding whether to treat with a given agent must be individualized and always should be based on actual, calculated benefits and risks. The routine prescription of any antithrombotic therapy theoretically to off-set the risk of VTE associated with a hormonal agent cannot be advocated in the absence of controlled trials that demonstrate risk neutralization with acceptable bleeding risk.