Implications for Use of Aromatase Inhibitors in 2003
The third-generation aromatase inhibitors (AIs) have improved efficacy and safety versus tamoxifen for treatment of advanced breast cancer. Currently, anastrozole is the only third-generation AI with adjuvant therapy data in postmenopausal women. Initial and updated results from the Arimidex, Tamoxifen, Alone or in Combination (ATAC) trial (median follow-up, 47 months) confirm it to be more effective than tamoxifen for disease-free survival with several important tolerability benefits. As a result, there has been much debate about whether or not anastrozole should be used routinely to treat postmenopausal women with early breast cancer. In its review, the American Society of Clinical Oncology Health Services Research Committee agreed that the updated ATAC analyses provided a greater level of assurance, in terms of both toxicity and efficacy, for use of anastrozole in the adjuvant setting. However, pending 5-year data from ATAC and other trials of adjuvant AI use, adjuvant anastrozole was recommended by American Society of Clinical Oncology Health Research Committee for use only under certain circumstances, with 5 years of tamoxifen remaining the standard. Anastrozole should be the preferred AI in this setting; data from the ATAC trial should not be extrapolated to other members of the class. Despite this conservative recommendation, the overall risk:benefit profile from the ATAC trial favors anastrozole, and it is expected that a more favorable efficacy and adverse effect profile will be maintained. Anastrozole should, therefore, now be considered a valid alternative option to tamoxifen for adjuvant hormonal treatment in all postmenopausal women with hormone receptor-positive early breast cancer.
The estrogen receptor antagonist tamoxifen, which first became available over 20 years ago, was the first hormonal breast cancer therapy and for many years has been considered the “gold standard” adjuvant endocrine therapy in postmenopausal women. The most recently published analysis showed that, in patients with hormone receptor-positive disease, the odds of recurrence and death were reduced by 47% and 26% respectively, after about 5 years of therapy with tamoxifen (1) . Despite its proven effectiveness, however, tamoxifen is associated with a number of serious and potentially life-threatening adverse effects, including an increased risk of endometrial cancer, uterine sarcoma, and thromboembolic disorders (1, 2, 3, 4, 5) . In particular, tamoxifen has been shown to double the risk of endometrial cancer after 1 or 2 years of treatment and to quadruple the risk after 5 years (1) . Incidence rates per 1000 women-years for these events have been estimated from the National Surgical Adjuvant Breast and Bowel Project P-1 trial. Tamoxifen citrate (Nolvadex) increased the rate of endometrial adenocarcinoma from 0.71 with placebo to 2.20 and increased the rate of uterine sarcoma from 0.0 with placebo to 0.17. For stroke, the incidence rate increased from 1.00 with placebo to 1.43 with tamoxifen citrate, whereas the incidence rate for pulmonary embolism increased from 0.25 for placebo to 0.75 with tamoxifen citrate. Some of the strokes, pulmonary emboli, and uterine malignancies were fatal (6) . These shortcomings, likely related to its partial estrogen agonist activity (7) , raise concerns regarding the use of tamoxifen in the adjuvant setting, where 5 years of treatment is usually recommended. This has prompted the search for and development of new agents with equal or improved efficacy but fewer adverse effects.
Aromatase inhibitors (AIs), developed for the treatment of women in whom ovarian function has ceased either naturally due to menopause or artificially (because of surgery or chemotherapy), have had a tremendous impact on breast cancer treatment. These agents differ from tamoxifen in their mode of action. They reduce estrogen levels by inhibiting aromatase, a cytochrome P450 enzyme that catalyzes the conversion of androgens to estrogens in peripheral tissues, such as body fat, liver, and breast and muscle cells (8 , 9) , and in the breast tumor tissue itself (10 , 11) . AIs have no estrogenic activity and therefore are unlikely to be associated with the same long-term adverse effects as tamoxifen.
The third-generation AIs, which include the nonsteroidal agents anastrozole and letrozole and the steroidal compound exemestane, are the most recent AIs to become available for use in postmenopausal women with metastatic hormone-responsive breast tumors. These drugs are administered orally and are more highly selective for the aromatase enzyme than first- and second-generation AIs.
Compelling data have led to anastrozole, letrozole, and exemestane becoming established as standard second-line therapies in postmenopausal women with hormone-dependent advanced breast cancer whose disease has recurred during tamoxifen therapy (12, 13, 14, 15) . In addition, anastrozole and letrozole have both shown superiority to tamoxifen as first-line therapy for postmenopausal women with advanced breast cancer (16 , 17) and are now increasingly being used in this setting in place of tamoxifen. As a result of their favorable efficacy and tolerability profiles in treatment of metastatic disease, these third-generation AIs now are being evaluated for treatment of early breast cancer, with several ongoing trials investigating their use as adjuvant therapy (18, 19, 20, 21, 22) . In 2001, the Arimidex, Tamoxifen, Alone or in Combination (ATAC) trial became the first of these trials to report and subsequently publish results (23) . This article provides an overview of the results of this significant study and of the implications for AIs and their future role in breast cancer treatment. At present, there are no published data for letrozole and exemestane in the adjuvant setting, and because several differences between the third-generation AIs have been noted in terms of selectivity for the aromatase enzyme, pharmacokinetics, and effects on lipid profiles, bone absorption, and steroidogenesis (24) , differences in safety profiles between AIs may become apparent after long-term dosing. For this reason, the American Society of Clinical Oncology (ASCO) Health Services Research Committee recommended that the ATAC data should not be extrapolated to other AIs (25) . This review therefore focuses on the impact of recent anastrozole data on treatment of postmenopausal women with early breast cancer.
Use of Anastrozole in the Adjuvant Setting
ATAC, the largest breast cancer trial ever undertaken, is a randomized, double-blind, multicenter study. It included postmenopausal women with invasive operable breast cancer who had completed their primary therapy (surgery, with or without chemotherapy, with or without radiotherapy) and were eligible to receive adjuvant endocrine therapy. These patients received anastrozole alone, tamoxifen alone, or a combination of both drugs. Primary end points were disease-free survival [DFS; defined as time to earliest local or distant recurrence, new primary (contralateral) breast cancer, or death, whichever occurred first] and safety/tolerability. Other end points included time to recurrence (censoring non-breast cancer deaths before recurrence) and the incidence of contralateral breast cancer. The first analysis was carried out at a median duration of therapy of 30.7 months, when the median follow-up was 33.3 months (23) .
Efficacy of Anastrozole versus Tamoxifen.
At the first analysis of the ATAC trial, anastrozole was significantly more effective than tamoxifen for all end points analyzed. In the overall population, DFS was significantly longer in the anastrozole group compared with the tamoxifen group [hazard ratio (HR), 0.83; 95% confidence interval (CI), 0.71–0.96, P = 0.013] and the combination group (HR, 0.81; 95% CI, 0.70–0.94; P = 0.0006; Ref. 23 ). In the hormone receptor-positive group (representing 84% of the total population and the target population for endocrine therapy), benefit for DFS was accentuated in patients receiving anastrozole versus tamoxifen (HR, 0.78; 95% CI, 0.65–0.93; P = 0.005) and the combination (HR, 0.76; 95% CI, 0.63–0.91; P = 0.002; Fig. 1<$REFLINK> ). Anastrozole was also superior to tamoxifen for time to recurrence in both the overall (HR, 0.79; 95% CI, 0.67–0.94; P = 0.008) and the hormone receptor-positive population (HR, 0.73; 95% CI, 0.59–0.90; P = 0.003; Ref. 23 ). In the overall population, there was a significant reduction in contralateral breast cancers as a first event in the anastrozole [0.5% (14 of 3125)] versus the tamoxifen [1.1% (33 of 3116)] group; based on the odds ratio of 0.42 (95% CI, 0.22–0.79; P = 0.007), this represented a 58% reduction in the risk of developing contralateral breast cancer for women in the anastrozole group compared with the tamoxifen group (23) . Reduction in the risk of contralateral breast cancer for the hormone receptor-positive population was consistent with the earlier results (odds ratio, 0.29; 95% CI, 0.13–0.64; P = 0.002; Ref. 23 ).
The combination arm was stopped after the first analysis because it showed no benefit over the tamoxifen-alone arm.
Further follow-up was carried out at a median duration of 47 months, at which time 46% of patients had been followed for more than 4 years (18 , 24) . This efficacy update showed that anastrozole was still superior to tamoxifen in terms of DFS and time to recurrence. The results of the combination arm were not significantly different from those with tamoxifen alone (18) . As might be expected, this benefit was accentuated in the known receptor-positive population (Table 1)<$REFLINK> . The absolute benefit for anastrozole versus tamoxifen in terms of DFS estimates and breast cancer events continues to increase over time; absolute difference in DFS estimates increased from 2% at year 3 to 2.4% at year 4 in the overall population and from 1.7% to 2.9%, respectively, in the hormone receptor-positive population. For breast cancer events, the absolute difference was 1.7% at year 3 and 2.3% at year 4 in the overall population and 1.8% and 2.6%, respectively, in the hormone receptor-positive population (18) .
Incidence of contralateral breast cancer continued to be lower in the anastrozole group than in the tamoxifen group and was significantly lower in the hormone receptor-positive group (Table 1)<$REFLINK> .
Comparison of Tolerability Profiles.
Anastrozole showed several important tolerability benefits over tamoxifen. At the first analysis, anastrozole was superior to tamoxifen for hot flushes (P < 0.0001), vaginal discharge (P < 0.0001), vaginal bleeding (P < 0.0001), ischemic cerebrovascular events (P = 0.0006), thromboembolic events (P = 0.0006; including deep-vein thrombosis, P = 0.02), and endometrial cancer (P = 0.02; Table 2<$REFLINK> ). Tamoxifen was superior to anastrozole for musculoskeletal disorders and fractures (P < 0.0001 for both; Table 2<$REFLINK> ; Ref. 23 ). Adverse event rates were similar between the combination and tamoxifen arms at the first analysis. Anastrozole was associated with significantly fewer withdrawals (includes completion of therapy, recurrences, deaths, or discontinuation due to any reason) from treatment than tamoxifen (21.9% versus 26.0%; P = 0.0002), including fewer withdrawals due to drug-related adverse events (5.1% versus 7.2%; Ref. 23 ).
After an additional 7 months of follow-up from the first safety analysis (median treatment duration, 37 months), an updated safety analysis was performed in line with standard United States Food and Drug Administration requirements (18 , 27) . The results showed that the favorable tolerability profile seen with anastrozole over tamoxifen is maintained in the longer term, with the relative risks for all of the predefined adverse events remaining virtually unchanged at the first and updated safety analyses (Table 2)<$REFLINK> . There was no further increase in the risks of fractures or musculoskeletal disorders with anastrozole. In addition, anastrozole continued to be associated with fewer withdrawals from treatment than tamoxifen (24.1% versus 28.3%, respectively), including fewer withdrawals due to drug-related adverse events (5.6% versus 8.1%, respectively; Ref. 27 ).
Implications of the ATAC Data for Treatment of Early Breast Cancer
Because the ATAC data show a number of significant benefits for anastrozole over tamoxifen in the adjuvant setting, both in terms of efficacy and tolerability, especially considering the relatively short follow-up of the ATAC trial, there has been much debate about whether or not anastrozole should now be used routinely to treat postmenopausal women with early breast cancer. This decision requires careful consideration, and several practice guidelines have provided recommendations to assist physicians.
ASCO Technology Assessment: Recommendations for Use of Anastrozole.
In 2002, the ASCO Health Services Research Committee reviewed the first analysis of the ATAC trial and how the results should be applied to clinical practice (28) . The results were published in a technology assessment, in which the panel observed that follow-up data beyond 5 years in patients who received anastrozole is not yet available. However, anastrozole showed significantly improved efficacy (in terms of DFS) compared with tamoxifen at a median follow-up of 33 months. At this study point, anastrozole was also associated with short-term adverse effects comparable with or fewer than those seen with tamoxifen, except for musculoskeletal disorders and fractures. Nonetheless, it was recommended that until updated data from the ATAC trial and other trials of third-generation AIs in the adjuvant setting were published, the standard adjuvant hormonal therapy should remain 5 years of tamoxifen therapy (28) . There were certain circumstances (if patients experienced intolerable adverse effects or a complication attributable to tamoxifen, or if they had a history of cerebrovascular disease or thromboembolic events) under which the panel did consider that it would be reasonable for patients to be treated with an AI as adjuvant therapy. In this case, anastrozole should be the preferred agent because it is currently the only AI with data in the adjuvant setting.
This assessment was reviewed and updated in 2003 by the ASCO Working Group, based on the published results from the updated ATAC analyses (18) . In this update report, a prime concern was the apparent reduction in the recurrence advantage of anastrozole over tamoxifen seen in the updated efficacy analysis compared with the first analysis. This concern was based on a nonsignificant increase in the HR, although this was within normal statistical variation. The slight increase occurred due to fluctuations in the unstable tail of the survival curve (Fig. 2)<$REFLINK> . At the time of the first analysis, there was an unrealistic separation of the curves in favor of anastrozole after 42 months due to small numbers of events that was not reported because it was thought to be an artifact. Events occurring after 42 months were included in the summary HRs and Ps, in keeping with good practice. With more data, the curves stabilized, leading to a slightly less significant P. However, the curves are virtually unchanged for the first 3 years, indicating that, in this period, the benefit in terms of recurrence for anastrozole is not being lost, and the larger separation at year 4 than at year 3, with a continuing improvement in absolute difference, suggests an increasing benefit of anastrozole with longer follow-up (18) .
Despite this concern, the ASCO Working Group acknowledged that the additional follow-up provides a greater level of assurance, in terms of both toxicity and efficacy, for physicians and their patients considering the use of anastrozole in the adjuvant setting. However, they maintained the recommendations that they initially released in 2002 (25) .
Although it cannot be denied that more time is needed to assess distant recurrences and breast cancer mortality in the ATAC trial, a substantial number of patients (46%) have now completed 4 or more years of treatment, and the adverse effect profile has remained virtually unchanged since the first analysis (18 , 27) . In particular, there is no indication of any acceleration of fracture rates with anastrozole (29) . The ATAC Trialists’ Group will continue to be vigilant for any late effects but feels confident that the updated report will accurately reflect the adverse effect profile for a full 5 years of active treatment.
The benefits of tamoxifen appear to be prolonged. In patients receiving 5 years of tamoxifen treatment, benefits are still apparent up to 9 years from treatment initiation (1) . However, this effect is probably because tamoxifen is curative in a proportion of patients. The extended duration of benefit is therefore also likely to be seen with anastrozole, perhaps to an even greater extent. It should also be noted that the 22% advantage of anastrozole versus tamoxifen (HR, 0.78) seen in patients with receptor-positive tumors in years 0–4 of the ATAC trial is already large enough to negate the extended duration of response to tamoxifen compared with placebo in years 5–9. These data indicate that anastrozole is unlikely to be less efficacious than tamoxifen at 10 years of follow-up. Further follow-up will confirm whether this will be the case and will also demonstrate whether anastrozole retains a more favorable adverse effect profile as anticipated.
Finally, it should be pointed out that the seminal paper on CMF chemotherapy was published with a 27-month median follow-up and no survival advantage (30) . Despite its toxicities, CMF was adopted rapidly by many oncologists (31) and certainly before survival data were available.
National Comprehensive Cancer Network Guidelines on the Use of Adjuvant Hormonal Therapy.
The National Comprehensive Cancer Network, which also produces practice guidelines on the use of adjuvant hormonal therapy, has already acknowledged that anastrozole is a suitable alternative to tamoxifen in the advanced setting (32) . Based on the evidence from the ATAC trial, the National Comprehensive Cancer Network has advised that, at the current time, in postmenopausal women with hormone receptor-positive tumors, anastrozole may be considered an option to adjuvant tamoxifen. However, discussion of the available data from the ATAC trial between the physician and patient is advised first (32) .
International Breast Cancer Study Group: Updated Consensus Guidelines.
At the eighth meeting of the International Breast Cancer Study Group held in St. Gallen (Switzerland) in 2003, the implications of evidence for patient treatment selection were discussed. In contrast to the National Comprehensive Cancer Network guidelines, the St. Gallen guidelines are conservative. As in the initial ASCO Technology Assessment Report (28) , the International Breast Cancer Study Group noted that additional follow-up is required to determine the ultimate comparison between anastrozole and tamoxifen in the adjuvant setting (33) ; however, an AI may be substituted in postmenopausal women in whom tamoxifen is contraindicated or not tolerated, and currently this should be anastrozole, the only AI with data in the adjuvant setting. The panel indicated that the evidence that tamoxifen increased deep vein thrombosis and pulmonary embolism, especially among patients immobilized by surgery for bone fractures, should be borne in mind when considering the benefits of anastrozole in this setting (33) .
The third-generation AIs have provided improved efficacy and safety versus tamoxifen for treatment of advanced breast cancer and thus show potential as adjuvant therapies. Anastrozole is the only third-generation AI for which data on adjuvant treatment of early breast cancer in postmenopausal women are available. Initial and updated results from the ATAC trial confirm it to be more effective than tamoxifen for postmenopausal women with hormone-responsive tumors with several important tolerability benefits. Therefore, for the first time, anastrozole has provided a valid alternative endocrine option to tamoxifen for adjuvant therapy of postmenopausal women with early breast cancer and should be the preferred AI in this setting at this time. However, longer-term follow-up is required to conclusively assess the final risk benefit profile of adjuvant anastrozole therapy and its role in comparison with tamoxifen. Although there is no suggestion that this agent combined with tamoxifen offers any advantage, the sequential use of this and other AIs with tamoxifen is under investigation.
Dr. James Ingle: One of the issues with the AIs relates to duration of therapy. Even the United States Food and Drug Administration doesn’t know how long to give it; the package insert states that the optional duration is not known. The only data in the literature are the Trevor Powles study (Jones, A. L., Powles, T. J., Law, M., Tidy, A., Easton, D., Coombes, R. C., Smith, I. E., McKinna, J. A., Nash, A., Ford, H. T., et al. Adjuvant aminoglutethimide for postmenopausal patients with primary breast cancer: analysis at 8 years. J. Clin. Oncol., 10: 1547–1552, 1992) with aminoglutethimide, which shows the paucity of data. That was a study where patient were randomized to 2 years of aminoglutethimide versus observation. If you look at the HRs, it was significantly in favor of aminoglutethimide for the 2 years that aminoglutethimide was given; from years 2–4, it was still in favor of aminoglutethimide, but it was not statistically significant; and beyond that point, there was an increasing rate of events that actually was in favor of observation alone. What is your thought on that? Is this something that once you start, you give it forever?
Dr. Buzdar: We have some indirect evidence from the reversible ovarian suppression data, which is analogous to this. Reversible ovarian suppression with pharmacological intervention for 2–3 years has exerted almost minimal difference in findings compared to permanent ovarian suppression, and in a sizeable number of those patients ovarian function does come back to normal if they have not gone through the natural menopause. There was no increase in relapses in patients in whom ovarian function did return. I think aminoglutethimide is a poor drug.
Dr. Myles Brown: I was interested in the prevention data indicating that the contralateral breast cancer prevention was 50% better than tamoxifen, which we already know prevents 50% of the estrogen receptor-positive breast cancers. What are your current thoughts regarding prevention trials using AIs?
Dr. Buzdar: I think it is ready for prime time to be tested in that subset of patients. The Europeans have started an IBIS-II trial, in which they elected to go with a placebo instead of the tamoxifen. I think it needs to be tested against our standard, which is tamoxifen.
Dr. Brown: Do we have any data about the phenotype of the tumors that have occurred on the ATAC trial?
Dr. Buzdar: The majority of events with anastrozole are exactly as seen with tamoxifen. Most of the benefit and the events have been seen in the hormone receptor-positive patient population. For the estrogen receptor-negative group, those data are still in the collection phase.
Dr. Steven Come: The fracture data have never made complete sense to me. It seems as though there is a large separation really quickly, given how slowly you expect bone to change, and then you are showing that they remain stable over time. I would have thought there would have been a progressive effect and the fracture rate would have increased.
Dr. Richard Santen: Recent studies have shown that any of the agents that are antiresorptive block fractures within 6 months to a year, well before there is a change in bone density. In the bone endocrinology field, the issues of bone quality, microfractures, and areas of trabecular thinning are thought to be quite important. With a reduction in bone quality, you can get fractures very quickly, and with an improvement in bone quality, you can prevent fractures very quickly. Bone quantity and quality represent different parameters. With an AI where the estrogen levels go down quickly, one might expect to see a difference within a year in comparison with tamoxifen, because of this bone quality issue. In regard to bone quantity, the bone densities are different, a little bit after a year, but not substantially different. It takes a longer time for bone quantity to change.
Dr. Ingle: As I remember from the ASCO technology assessment panel paper (25) , one of the comments was that the HR for a first event (anastrozole:tamoxifen) was higher in the report with 47-month median follow-up than the earlier report with the 33-month median follow-up, indicating, if anything, less benefit from anastrozole with the longer follow-up.
Dr. Buzdar: The HR is calculated on the total event rates of the curve, including the tail end.
Dr. Brown: The error bars were just bigger in your first analysis.
Dr. Buzdar: But these fall within the same CI, where there is very little difference.
Dr. Ingle: At least you would agree that there is no statistical basis for saying that the differences are increasing on the basis of HRs.
Dr. Buzdar: Absolute differences are increasing with time. HRs are similar to initial publication, statistically speaking.
Dr. Arteaga: There is a big difference in patients with prior chemotherapy versus no prior chemotherapy. Are the numbers enough to make the statement that prior chemotherapy negates the possible benefit?
Dr. Buzdar: No, there are about 700 patients in that subgroup, and the patients who got anthracycline are quite a bit different in the two arms. Actually, we are looking at this now, because using the information from the Oxford Overview you can weight the data and ask: what is the reduction in recurrence with CMF or anthracycline-based or taxane-based chemotherapies? We may be able to adjust, but it will be a retrospective look at the data.
Dr. Schiff: What can you tell us about quality-of-life function with anastrozole?
Dr. Buzdar: Quality of life has been prospectively evaluated in a blinded fashion. There is a paper on these results that has just been submitted for publication.1 In the first few months, the overall quality of life deteriorates in all three arms, and subsequently there is an improvement with no significant differences, except for issues of vaginal dryness and related symptoms from estrogen deprivation, that are more in favor of tamoxifen.
Presented at the Third International Conference on Recent Advances and Future Directions in Endocrine Manipulation of Breast Cancer, July 21–22, 2003, Cambridge, MA.
Requests for reprints: Aman U. Buzdar, Department of Breast Medical Oncology, Box 424, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas 77030. Phone: (713) 792-2817; Fax: (713) 794-4385; E-mail:
↵1 Assessing the quality of life of postmenopausal women randomized into the ATAC (Arimidex*, Tamoxifen, Alone or in Combination) adjuvant breast cancer trial. The ATAC QoL Writing Committee on behalf of the ATAC Trialists’ Group. Lesley Fallowfield, David Cella, Jack Cuzick, Stephen Francis, Gershon Locker and Anthony Howell. Submitted.