American Society of Clinical Oncology Clinical Practice Guideline: Update on Adjuvant Endocrine Therapy for Women With Hormone Receptor–Positive Breast Cancer

Abstract

Purpose

To develop evidence-based guidelines, based on a systematic review, for endocrine therapy for postmenopausal women with hormone receptor–positive breast cancer.

Methods

A literature search identified relevant randomized trials. Databases searched included MEDLINE, PREMEDLINE, the Cochrane Collaboration Library, and those for the Annual Meetings of the American Society of Clinical Oncology (ASCO) and the San Antonio Breast Cancer Symposium (SABCS). The primary outcomes of interest were disease-free survival, overall survival, and time to contralateral breast cancer. Secondary outcomes included adverse events and quality of life. An expert panel reviewed the literature, especially 12 major trials, and developed updated recommendations.

Results

An adjuvant treatment strategy incorporating an aromatase inhibitor (AI) as primary (initial endocrine therapy), sequential (using both tamoxifen and an AI in either order), or extended (AI after 5 years of tamoxifen) therapy reduces the risk of breast cancer recurrence compared with 5 years of tamoxifen alone. Data suggest that including an AI as primary monotherapy or as sequential treatment after 2 to 3 years of tamoxifen yields similar outcomes. Tamoxifen and AIs differ in their adverse effect profiles, and these differences may inform treatment preferences.

Conclusion

The Update Committee recommends that postmenopausal women with hormone receptor–positive breast cancer consider incorporating AI therapy at some point during adjuvant treatment, either as up-front therapy or as sequential treatment after tamoxifen. The optimal timing and duration of endocrine treatment remain unresolved. The Update Committee supports careful consideration of adverse effect profiles and patient preferences in deciding whether and when to incorporate AI therapy.

INTRODUCTION

The first technology assessment for the adjuvant use of aromatase inhibitors (AIs) for women with hormone receptor–positive breast cancer was published by the American Society of Clinical Oncology (ASCO) in 2002.¹ The technology assessment was updated in 2003² and 2004.³ Since then, additional reports from large-scale trials of adjuvant endocrine therapy have been published. These developments warranted an update and systematic review.

ASCO's practice guidelines reflect expert consensus based on clinical evidence and literature available at the time they are written and are intended to assist physicians in clinical decision making and to identify questions and settings for further research. Because of the rapid flow of scientific information in oncology, new evidence may have emerged since a guideline was submitted for publication. Guidelines and assessments are not continually updated and may not reflect the most recent evidence. Guidelines address only the topics specifically identified in the guideline and are not applicable to interventions, disease, or stages of disease not specifically identified. Guidelines cannot account for individual variation among patients and cannot be considered inclusive of all proper methods of care or exclusive of other treatments. It is the responsibility of the treating physician or other health care provider, relying on independent experience and knowledge of the patient, to determine the best course of treatment for the patient. Accordingly, adherence to any guideline is voluntary, with the ultimate determination regarding its application to be made by the physician in light of each patient's individual circumstances. ASCO guidelines describe the use of procedures and therapies in clinical practice and cannot be assumed to apply to the use of these interventions in the context of clinical trials. ASCO assumes no responsibility for any injury or damage to persons or property arising out of or related to any use of ASCO's guidelines, or for any errors or omissions.

UPDATE METHODOLOGY

Guideline Questions

The Update Committee (Appendix Table A1, online only) focused on the optimal adjuvant endocrine strategy with use of either tamoxifen, AIs, or both in sequence; duration of AI therapy; long-term adverse effects of AI therapy; identification of subpopulations who might derive selective benefit from either AIs or tamoxifen-based treatments; efficacy of AIs among premenopausal women; and similarities or differences among commercially available third-generation AIs.

Literature Review and Analysis

Literature search strategy.

For this update, ASCO staff evaluated a recent systematic review completed by Cancer Care Ontario (CCO) that included literature through May 2007.⁴ MEDLINE, PREMEDLINE, and the Cochrane Collaboration databases were searched from May 2007 through February 2009 (Appendix, online only), as were electronic databases for the San Antonio Breast Cancer Symposium (SABCS) and ASCO Annual Meetings from 2007 to 2009.

Inclusion and exclusion criteria.

Articles identified for inclusion in this systematic review met the following criteria: (1) the intervention was for the adjuvant therapy of breast cancer, (2) participants were randomly assigned to any of the treatments described previously, and (3) reports included at least one of the following primary outcomes of interest: overall survival, disease-free survival, or breast cancer–specific survival. Three different treatment strategies were identified on the basis of the timing of AI therapy: initial endocrine therapy (hereafter referred to as a primary adjuvant strategy), sequential therapy with treatment divergence if the patient was disease free following 1 to 4 years of initial treatment with adjuvant endocrine agents (most often tamoxifen), or extended therapy with random assignment if the patient was disease free following 5 years of treatment with adjuvant tamoxifen.

Twelve prospective, randomized clinical trials originally identified by the co-chairs were the focus of this systematic review. These same trials were identified in the CCO systematic review, as well as in a systematic review completed by the National Institute for Health Research (NIHR) in the United Kingdom.⁵

Several important limitations of the existing literature were identified. Of particular note is the timing of randomization (Fig 1). Most sequential trials and all of the extended trials randomly assigned women who were free of recurrence through multiple years of tamoxifen therapy, effectively excluding women with early recurrence. For this reason, the patient populations in the sequential and extended trials may differ importantly from one another and from patients in the primary therapy studies. Another limitation is the relatively short follow-up; disease recurrence decades after diagnosis is not uncommon. The longest available median follow-up in the present trials is slightly more than 8 years; most studies have considerably shorter follow-up. For the majority of the efficacy outcomes across all studies, the median time to event has yet to be reached. The relatively modest number of events may also limit study conclusions.

Fig 1.

Schema of included trials. ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); mos, months; HR (+), hormone receptor–positive; TAM, tamoxifen; ANA, anastrozole; BIG, Breast International Group; FU, follow-up; monotx, monotherapy; LET, letrozole; yrs, years; ABCSG, Austrian Breast and Colorectal Cancer Study Group; ER (+), estrogen receptor–positive; PR (+), progesterone receptor–positive; GOS, goserelin; ZOL, zoledronic acid; ABCSG, Austrian Breast and Colorectal Cancer Study Group; Chemo, chemotherapy; ITA, Italian Tamoxifen Anastrozole (trial); tx, therapy; TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; IES, Intergroup Exemestane Study; Unk, unknown; N-SAS, National Surgical Adjuvant Study (Group); ARNO, Arimidex-Nolvadex (trial); NSABP, National Surgical Adjuvant Breast and Bowel Project.

The number of actual events was modest because of the generally favorable prognosis among patients. Owing to the vagaries of data collection and patient participation, as well as the distribution of clinical subsets, subgroup evaluations are hindered by relatively small sample sizes. Small sample size is also a limitation in analyzing available quality-of-life data. Finally, comparisons between trials are hindered because of different definitions of study end points.⁶

Consensus Development Based on Evidence

The Update Committee was charged with reviewing evidence from the systematic review and drafting new recommendations (Table 1). Per standard ASCO practice, the guideline was submitted to the Journal of Clinical Oncology for peer review. The content of the guideline was reviewed and approved by both the ASCO Clinical Practice Guideline Committee and the Board of Directors before publication.

Table 1.

Summary of 2010 Recommendations

Clinical Question	Recommendation
1a. What adjuvant endocrine treatments should be offered to postmenopausal women with hormone receptor–positive breast cancer?	Postmenopausal women should consider taking an AI during the course of adjuvant treatment to lower recurrence risk, either as primary therapy or after 2 to 3 years of tamoxifen. Duration of AI therapy should not exceed 5 years.
1b. What is the appropriate duration of adjuvant endocrine therapy?	Therapy with an AI should not extend beyond 5 years in either the primary or extended adjuvant settings outside the clinical trials setting. In the sequential setting, patients should receive an AI after 2 or 3 years of tamoxifen for a total of 5 years of adjuvant endocrine therapy. Patients initially treated with an AI but who discontinue treatment before 5 years of therapy should consider incorporating tamoxifen for a total of 5 years of adjuvant endocrine therapy.
1c. If tamoxifen is administered first, how long should it be continued before the switch to an AI?	Patients who initially receive tamoxifen as adjuvant therapy may be offered an AI after 2 to 3 years (sequential) or after 5 years (extended) of therapy. The best time to switch from an AI to tamoxifen (or the converse) is not known. Switching at 2 to 3 years is recommended, but switching at 5 years is also supported by available data.
2. Are there specific patient populations that derive differing degrees of benefit from an AI in comparison to tamoxifen?	A specific marker or clinical subset that predicts which adjuvant treatment strategy (tamoxifen alone, AI alone, or AI and tamoxifen-based) is best has not been identified. Among men with breast cancer, tamoxifen remains the standard adjuvant endocrine treatment. The CYP2D6 genotype is not recommended to select adjuvant endocrine therapy. Caution with concurrent use of CYP2D6 inhibitors (such as bupropion, paroxetine, or fluoxetine) and tamoxifen is recommended because of drug-drug interactions.
3. What are the toxicities and risks of adjuvant endocrine therapy?	Clinicians should consider adverse effect profiles, patient preferences, and pre-existing conditions when discussing adjuvant endocrine strategies. Adverse effect profiles should be discussed with patients when presenting available treatment options. Clinicians may recommend that patients change treatments if adverse effects are intolerable or patients are persistently noncompliant with therapy.
4. Are AIs effective adjuvant therapy for women who are premenopausal at the time of diagnosis?	Women who are pre- or perimenopausal at diagnosis should be treated with 5 years of tamoxifen.
5. Can the third-generation AIs be used interchangeably?	Meaningful clinical differences between the commercially available third-generation AIs have not been demonstrated to date. The Update Committee believes that postmenopausal patients intolerant of one AI may be advised to consider tamoxifen or a different AI.

Abbreviation: AI, aromatase inhibitor.

Guideline and Conflicts of Interest

The Update Committee was assembled in accordance with ASCO's Conflict of Interest Management Procedures for Clinical Practice Guidelines (“Procedures,” summarized at www.asco.org/guidelinescoi). Members completed ASCO's disclosure form, which requires disclosure of financial and other interests relevant to the subject matter of the guideline, including relationships with commercial entities that are reasonably likely to experience direct regulatory or commercial impact as the result of promulgation of the guideline. Categories for disclosure include employment relationships, consulting arrangements, stock ownership, honoraria, research funding, and expert testimony. In accordance with the Procedures, the majority of the members of the Update Committee did not disclose any such relationships.

RESULTS

Summary of the Literature Review Results

Since the last ASCO update, 442 published articles and 42 presentations, posters, or abstracts were identified (Appendix Fig A1). Of these, 49 reported either findings from one of 12 trials^7–49 or were systematic reviews or meta-analyses of this same set of landmark trials for adjuvant endocrine therapy.^4,50–54 Thirteen articles with primary data were retrieved to complement extraction. Five reports available after completion of the initial search were also considered.^55,56

GUIDELINE RECOMMENDATIONS

Clinical question 1a.

What adjuvant endocrine treatments should be offered to postmenopausal women with hormone receptor–positive breast cancer?

Recommendation 1a.

The Update Committee recommends, on the basis of data from randomized, controlled trials, that most postmenopausal women consider taking an AI during the course of adjuvant treatment to lower recurrence risk, either as primary therapy or after 2 to 3 years of tamoxifen—strategies that yield equivalent outcomes in prospective studies. Duration of AI therapy should not exceed 5 years.

Literature update and discussion 1a.

In comparison to 5 years of tamoxifen alone, use of an AI in either primary, sequential, or extended treatment improves disease-free survival and reduces the risk of breast cancer events, including distant recurrence, locoregional recurrence, and contralateral breast cancer (Tables 2 and 3; Fig 1). In absolute terms, the reduction in risk of recurrence associated with AI-based therapy compared with tamoxifen is modest, typically amounting to < 5% through multiple years of follow-up (Table 2). Tamoxifen and AI-based therapy are equivalent in terms of overall survival when used as either a primary or extended treatment strategy (Table 4). Two of the six trials of sequential treatment strategies yielded statistically significant improvements in overall survival compared with tamoxifen alone, although the absolute difference in overall survival is modest (Table 4).

Table 2.

Disease-Free Survival

Trial	Arm	Median Follow-Up (months)	Range (months)	No. of Patients Observed		Disease-Free Survival Events				HR	95% CI	P
						AI		Comparator
				AI	Comparator	No.	%	No.	%
Primary
ATAC20	ANA v TAM ITT	100	0-126	3,125	3,116	817	26.1	887	28.5	0.90	0.82 to 0.99	.025
BIG 1-9841	LET vTAM ITT	76		2,463	2,459	509	20.7	565	23.0	0.88	0.78 to 0.99a	.03
ABCSG-1222	ANA v TAM	47.8		903	900	72	8.0	65	7.2	1.10	0.78 to 1.53	.59
Sequential
BIG 1-9841	TAM/LET v LET	71		1,548	1,546	259	16.7	248	16.0	1.05	0.84 to 1.32b	NR
	LET/TAM v LET	71		1,540	1,546	236	15.3	248	16.0	0.96	0.76 to 1.21b	NR
ABCSG-829	TAM/ANA vTAM ITT	72		1,865	1,849	227	12.2	261	14.1	0.85	0.71 to 1.01c,d	.067
ITA9	ITT	64	12-93	223	225	39	17.5	63	28	0.57	0.38 to 0.85e	.005
TEAM56	TAM/EXE v EXE ITT	61		4,868	4,898	714		712		0.97	0.88 to 1.08	.604
IES13	ITT	55.7	0-89.7	2,352	2,372	354		455		0.76	0.66 to 0.88	< .001
N-SAS BC-038		42	3.2-60	347	349	26	7.5	37	10.6	0.69	0.42 to 1.14	.14f
ARNO 9533		30.1	< 12 to ≥ 84	489	490	38	7.8	56	11.4	0.66	0.44 to 1.00	.049
Extended
MA.1725	ITTg	64	16-95	2,583	2,587	164	6.3	235	9.1	0.68	0.55 to 0.83	< .001
ABCSG-6a30		62.3		386	466	30	7.8	57	12.2	0.62	0.40 to 0.96h	.031
NSABP B-3336		30		783	779	37	4.7	52	6.7	0.68		.07

NOTE. Percent calculated as number of events divided by number of patients observed.

Abbreviations: AI, aromatase inhibitor; HR, hazard ratio; ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ANA, anastrozole; TAM, tamoxifen; ITT, intent to treat; BIG, Breast International Group; LET, letrozole; ABCSG, Austrian Breast and Colorectal Cancer Study Group; NR, not reported; ITA, Italian Tamoxifen Anastrozole (trial); TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; IES, Intergroup Exemestane Study; N-SAS, National Surgical Adjuvant Study (Group); ARNO, Arimidex-Nolvadex (trial); NSABP, National Surgical Adjuvant Breast and Bowel Project.

^aAnalysis includes only patients from monotherapy arms; crossovers not censored.

^b99% CIs used to account for multiple comparisons.

^cRelapse-free survival includes local and distant recurrence, contralateral breast cancer, and death without recurrence.

^dPatients who crossed over were censored at time of crossover.

^eEvent-free survival includes locoregional or distant recurrence, second primary (including contralateral breast cancer), and deaths without recurrence.

^fTwo different P values reported (P = .06 in abstract, P = .14 on poster).

^gSeventeen patients omitted from ITT analysis.

^hRecurrence-free survival.

Table 3.

Contralateral Breast Cancer

Trial	Arm	Median Follow-Up (months)	Range (months)	No. of Patients Observed		Contralateral Breast Cancer Events				HR	95% CI	P
						AI		Comparator
				AI	Comparator	No.	%	No.	%
Primary
ATAC20	ANA v TAM ITT	100		3,125	3,116	61	1.9	87	2.8	0.68	0.49 to 0.94	.02
BIG 1-985	LET v TAM	68		4,003	4,007	16	0.4	27	0.7	NR	NR	NR
TEAM32	EXE v TAM ITT*	33†		4,898	4,868	21	0.4	17	0.3	NR	NR	NR
Sequential
ABCSG-829	TAM/ANA v TAM ITT	72		1,865	1,849	19	1.0	29	1.6	0.64	0.36 to 1.13	NR
IES13	ITT	55.7	0-89.7	2,352	2,372	18	0.76	35	1.5	0.57	0.33 to 0.98	.04
ARNO 9533		30.1	< 12 to ≥ 84	489	490	7	1.4	5	1	NR	NR	NR
Extended
MA.1728		64		2,583	2,587	30	1.16	49	1.89	0.61	0.39 to 0.97	.033
ABCSG-6a30		62.3		386	466	6	1.6	11	2.4	0.67	0.25 to 1.80‡	.422

NOTE. Percent calculated as number of events divided by number of patients observed.

Abbreviations: AI, aromatase inhibitor; HR, hazard ratio; ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ITT, intent to treat; BIG, Breast International Group; LET, letrozole; TAM, tamoxifen; NR, not reported; TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; ABCSG, Austrian Breast and Colorectal Cancer Study Group; ANA, anastrozole; IES, Intergroup Exemestane Study; ARNO, Arimidex-Nolvadex (trial).

^*Data for contralateral risk in the TEAM trial have been reported for comparison of initial therapy of TAM v EXE but not for TAM/EXE v EXE at this time.

^†Follow-up for all TEAM patients is 33 months (data beyond that censored).

^‡Risk of contralateral breast cancer.

Table 4.

Overall Survival

Trial	Arm	Median Follow-Up (months)	Range (months)	No. of Patients Observed		Overall Survival Deaths				HR	95% CI	P
						AI		Comparator
				AI	Comparator	No.	%	No.	%
Primary
ATAC20	ANA v TAM ITT	100		3,125	3,116	629	20.1	624	20.0	1.00	0.89 to 1.12	.99
BIG 1-9841	LET v TAM ITT	76		2,463	2,459	303	12.3	343	14.0	0.81	0.69 to 0.94*	.08
ABCSG-1222	ANA v TAM	47.8		903	900	27	3.0	15	1.7	1.8	0.95 to 3.38	.70
Sequential
ABCSG-829	TAM/ANA v TAM ITT	72		1,865	1,849	138	7.4	165	9.0	0.78	0.62 to 0.98†	.032
BIG 1-9841	TAM/LET v LET	71		1,548	1,546	154	9.9	137	8.9	1.13	0.83 to 1.53‡	NR
	LET/TAM v LET	71		1,540	1,546	123	8.0	137	8.9	0.90	0.65 to 1.24‡	NR
ITA9	ITT	64	12-93	223	225	12	5.4	21	9.3	0.56	0.28 to 1.15	.1
TEAM56	TAM/EXE v EXE ITT	61		4,868	4,898	NR (90.6% survival)		NR (90.5% survival)		1.00	0.89 to 1.14	.999
IES13	ITT	55.7	0-89.7	2,352	2,372	222	9.4	261	11.0	0.85	0.71 to 1.02	.08
N-SAS BC-038		42	3.2-60	347	349	NR		NR		NR		.59
ARNO 9533		30.1	< 12 to ≥ 84	489	490	15	3.1	28	5.7	0.53	0.28 to 0.99	.045
Extended
MA.1725	ITT§	64	16-95	2,583	2,587	154	6.0	155	6.0	0.98	0.78 to 1.22	.853
ABCSG-6a30		62.3		386	466	40	10.4	55	11.8	0.89	0.59 to 1.34	.570
NSABP B-3336		30		783	779	16	2.0	13	1.7	NR		NR

NOTE. Percent calculated as number of events divided by number of patients observed.

Abbreviations: AI, aromatase inhibitor; HR, hazard ratio; ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ANA, anastrozole; TAM, tamoxifen; ITT, intent to treat; BIG, Breast International Group; LET, letrozole; ABCSG, Austrian Breast and Colorectal Cancer Study Group; NR, not reported; ITA, Italian Tamoxifen Anastrozole (trial); TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; IES, Intergroup Exemestane Study; N-SAS, National Surgical Adjuvant Study (Group); ARNO, Arimidex-Nolvadex (trial); NSABP, National Surgical Adjuvant Breast and Bowel Project.

^*Analysis includes only patients from monotherapy arms; crossovers not censored.

^†Patients who crossed over were censored.

^‡99% CIs used to account for multiple comparisons.

^§Seventeen patients omitted from ITT analysis.

Breast cancer events such as locoregional recurrence, contralateral breast cancer, and early distant metastatic recurrence are clinically important to patients. For this reason, the Update Committee recommended consideration of AI therapy at some time during adjuvant endocrine therapy even though few trials demonstrated statistically significant differences in overall survival.

Sequential therapy.

Two trials directly compared primary monotherapy with sequential therapy as an initial 5-year adjuvant endocrine regimen. The Breast International Group 1-98 (BIG 1-98) trial compared primary tamoxifen or AI monotherapy against sequences of tamoxifen followed by an AI or an AI followed by tamoxifen.³⁹ The Tamoxifen Exemestane Adjuvant Multicenter (TEAM) trial compared a sequential treatment of tamoxifen followed by an AI against an AI alone.^56,57 Neither study demonstrated clinical or statistically significant differences between patients who received an AI alone, tamoxifen sequenced with an AI, or in the case of BIG 1-98, an AI sequenced with tamoxifen, with respect to disease-free or overall survival. In BIG 1-98, however, each AI-based therapy was superior to tamoxifen monotherapy with respect to disease-free survival (Table 2).³⁹ These data support the recommendation to incorporate AI therapy at some point during the first 5 years of adjuvant endocrine therapy, either as primary therapy or in sequence with tamoxifen.

Clinical question 1b.

What is the appropriate duration of adjuvant endocrine therapy?

Recommendation 1b.

Therapy with an AI should not extend beyond 5 years in either the primary or extended adjuvant settings, outside of clinical trials. In the sequential setting, the Update Committee recommends, on the basis of available evidence from randomized, controlled trials, that patients receive an AI after 2 or 3 years of tamoxifen for a total of 5 years of adjuvant endocrine therapy. The Update Committee recommends that patients who are initially treated with an AI but discontinue treatment before 5 years of therapy consider taking tamoxifen for a total of 5 years of adjuvant endocrine therapy.

Literature update and discussion 1b.

Importantly, no trials directly compared sequential or extended adjuvant strategies against one another or primary against extended strategies. Studies used different durations of total endocrine therapy and different durations of AI and tamoxifen treatment. It is not known whether these differences in duration of therapy are clinically significant. Optimal duration of therapy in the extended setting is unclear at this time. Safety and efficacy data from the primary trials support up to 5 years of AI therapy as a primary adjuvant strategy, a duration used in two trials of extended therapy after 5 years of tamoxifen.

The treatment regimen for patients in the sequencing trials spanned 5 years. No data support clinical benefits for durations of AIs longer than 2 or 3 years in a sequencing strategy. Data from randomized, controlled trials demonstrate that women who receive primary AI therapy should be treated for a total of 5 years; women who initially receive tamoxifen and switch to an AI should also receive at least five total years of endocrine therapy. Women who receive extended adjuvant therapy should receive 8 to 10 years of total endocrine treatment—5 years of tamoxifen followed by 3 to 5 years of an AI.

The Update Committee acknowledges that these recommendations yield an unfamiliar pattern of different durations of adjuvant endocrine treatment based on the treatment strategy used. This is a function of offering an AI at a different time point in accordance with each strategy. The recommended limit on AI treatment is 5 years total, across strategies. Two trials—MA.17R and National Surgical Adjuvant Breast and Bowel Project B-42 (NSABP B-42)—are evaluating whether longer durations of AI therapy improve outcomes, but results are not yet available.

Clinical question 1c.

If tamoxifen is administered first, how long should it be continued before the switch to an AI?

Recommendation 1c.

The Update Committee recommends that, on the basis of available evidence from randomized, controlled trials, patients who initially receive tamoxifen as adjuvant therapy may be offered an AI after 2 to 3 years (sequential) or after 5 years (extended) of therapy. The time to switch from an AI to tamoxifen (or the converse) that maximally improves outcomes is not known from available direct evidence. The Update Committee recommends switching at 2 to 3 years on the basis of data from sequential trials that used this strategy. Switching at 5 years is also a strategy supported by the extended adjuvant randomized trials.

Literature update and discussion 1c.

Most trials of sequential therapy switched patients from tamoxifen to an AI after 2 to 3 years of therapy. Some specified an exact cross-over point while others permitted switching within a broad window. Trials of extended therapy enrolled patients who already received an average of 5 years of tamoxifen (Fig 1). Of the trials of sequencing or switching, only BIG 1-98,⁵⁸ TEAM,³¹ and Austrian Breast and Colorectal Cancer Study Group ABCSG-8⁵⁹ enrolled patients before commencement of adjuvant endocrine therapy. The other sequencing and all the extended trials randomly assigned patients free of recurrence after multiple years of tamoxifen. Data are lacking to establish the best duration of tamoxifen therapy before switching to an AI.

Three trials of extended adjuvant therapy that used AI treatment for 3 to 5 years after 5 years of tamoxifen demonstrated that extended therapy can lower the risk of breast cancer recurrence (Table 2 and Appendix Table A2, online only) and contralateral breast cancer (Table 3) but does not improve overall survival (Table 4). The Update Committee recommends extended therapy with an AI for postmenopausal patients who complete 5 years of tamoxifen.

For a newly diagnosed patient or a patient who has taken tamoxifen for 2 to 5 years, it is not known whether switching from tamoxifen to an AI earlier or later is more effective for long-term disease-free survival. Lacking direct comparative data, the Update Committee recommends considering a switch to an AI after 2 or 3 years of tamoxifen therapy. This recommendation is informed by several observations. First, both the Arimidex-Nolvadex (ARNO) 95 and ABCSG-8 trials, which transitioned patients from tamoxifen to an AI after 2 or 3 years, demonstrated survival advantages, as did a meta-analysis of sequential trials^29,33,60 (Table 3). Second, each AI-based arm of BIG 1-98, including those with switching at 2 years, had improved disease-free survival relative to 5 years of tamoxifen³⁹ (Table 2). Incorporating an AI into the adjuvant treatment regimen at some point during the first 5 years of endocrine therapy yields clinical improvements.

Clinical question 2.

Are there specific patient populations that derive differing degrees of benefit from an AI in comparison to tamoxifen?

Recommendation 2.

Direct evidence from randomized trials does not identify a specific marker or clinical subset that predicted which adjuvant treatment strategy, tamoxifen or AI monotherapy or sequential therapy, would maximally improve outcomes for a given patient. Among men with breast cancer, tamoxifen remains the standard adjuvant endocrine treatment. The Update Committee recommends against using CYP2D6 genotype to select adjuvant endocrine therapy. The Update Committee encourages caution with concurrent use of CYP2D6 inhibitors (such as bupropion, paroxetine, or fluoxetine; Table 5) and tamoxifen because of the known drug-drug interactions.

Table 5.

Commonly Used CYP2D6 Inhibitors

Strong inhibitors

Bupropion

Fluoxetine

Paroxetine

Quinidine

Moderate inhibitors

Duloxetine

Terbinafine

Weak inhibitors

Amiodarone

Cimetidine

Sertraline

Flockhart DA: http://medicine.iupui.edu/clinpharm/ddis/table.asz.

Literature update and discussion 2.

Endocrine therapy is effective only among patients with tumors that express hormone receptors such as estrogen receptor (ER) and/or progesterone receptor.^61,62 Tumor size, nodal status, grade, quantitative levels of hormone receptor expression, HER2 overexpression, markers of proliferation, and the 21-gene recurrence score⁶³ are prognostic factors among patients receiving endocrine therapy.³⁷ These prognostic markers help to define relative risk of recurrence in the first 5 to 10 years after diagnosis.

The major trials of adjuvant AI therapy included patients with hormone receptor–positive tumors, generally irrespective of other markers or staging. Retrospective subset analyses from some trials considered a variety of prognostic factors among patients receiving tamoxifen or AI therapy. In these retrospective studies, tumor size, nodal status, age, quantitative ER and progesterone receptor levels, HER2 expression,^44,60 grade, Ki-67,⁴⁹ and the 21-gene recurrence score⁶⁴ seem to serve as prognostic factors for risk of breast cancer recurrence among patients receiving either tamoxifen or AI therapy (Appendix Table A3, online only).

Traditional assumptions about proportionate risk reduction achieved with adjuvant therapy suggest that differences in clinical outcome between various treatments are likely to be of greater absolute magnitude among patients with higher-risk breast cancers. Conversely, among women with lower-risk tumors, differences in outcomes between AI-based therapies and tamoxifen or between primary or sequential use of AIs are likely to be smaller if present at all. Emerging data from BIG 1-98 seem to validate these assumptions.^37,65 However, available retrospective subset analyses are constrained by missing data on some patients, technical limitations in the performance of correlative marker testing, multiple hypothesis testing, varying assays used by different trials, and lack of prospective, corroborative data.

On the basis of evidence from randomized clinical trials and consistent with the recommendation, the Update Committee recommended treatment with either upfront use of an AI or sequential therapy with tamoxifen, followed by an AI, irrespective of any specific clinical subset or prognostic marker.

Male breast cancer.

There is no evidence to evaluate the efficacy of adjuvant AI therapy in men. Tamoxifen remains the standard adjuvant endocrine therapy for male breast cancer.⁶⁶

CYP2D6 genotype.

Accumulating data suggest that variability in tamoxifen metabolism may affect serum levels of the tamoxifen metabolite endoxifen, which may in turn affect the likelihood of cancer recurrence in tamoxifen-treated patients.^66–73 Factors that contribute to this variability include concurrent use of other drugs that inhibit the CYP2D6 isoenzyme (which converts tamoxifen to endoxifen) and pharmacogenetic variation (polymorphisms) in CYP2D6 alleles. It is not known whether variations in CYP2D6 genotype account for differences in outcomes among patients treated with tamoxifen. Available data on CYP2D6 pharmacogenetics are insufficient to recommend testing as a tool to determine an adjuvant endocrine strategy.

The Update Committee recognized the accumulating evidence on drug-drug interactions between tamoxifen and other drugs that inhibit CYP2D6 (such as buproprion, paroxetine, or fluoxetine; Table 5). Evidence linking such interactions to breast cancer outcomes remains limited and indirect. Patients clearly benefiting from known CYP2D6 inhibitors might avoid tamoxifen because of potential pharmacologic interactions. Conversely, women taking tamoxifen may prefer to avoid concurrent use of known CYP2D6 inhibitors if suitable alternatives are available.

Clinical question 3.

What are the toxicities and risks of adjuvant endocrine therapy?

Recommendation 3.

The Update Committee recommends that clinicians consider adverse effect profiles, patient preferences, and pre-existing conditions when recommending an adjuvant endocrine strategy for postmenopausal women. Clinicians should discuss adverse effect profiles when presenting available treatment options. The Update Committee suggests that clinicians consider recommending that patients change treatment if adverse effects are intolerable or if patients are persistently noncompliant with therapy.

Literature update and discussion 3.

Tamoxifen and AIs are generally well tolerated but are associated with specific toxicities including effects on bone, cardiovascular, and gynecologic health. The differing adverse effect profiles are functions of differing mechanisms of action. Tamoxifen is a selective ER modulator with mixed pro- and antiestrogenic activities, while AIs achieve near complete estrogen deprivation in postmenopausal women. The Update Committee did not find evidence that AI therapy is less toxic or better tolerated than tamoxifen. Both drug classes have distinct adverse effect profiles that are relevant to individualizing therapy for patients.

Substudies from the large, randomized trials show generally well-maintained and similar quality of life scores in women receiving any of the adjuvant endocrine therapies.^18,36,43,48,^74–77 The severity of the most common adverse effects is mild to moderate for the majority of patients; serious adverse effects are rare. The long-term adverse effect profiles for tamoxifen-treated patients are established from historical literature.⁷⁸ Late effects of AI therapy remain to be fully characterized.

Appendix Tables A4 through A8 (online only) include an abbreviated list of the adverse effects tabulated from the therapies evaluated in the prospective, randomized trials discussed herein. Tamoxifen and AIs have differing effects on cardiovascular health. Data suggest that AIs are associated with increased cardiovascular disease, possibly including ischemic cardiac disease, though in absolute terms, at present, any differences are believed to be small.^40,51 In comparison to tamoxifen, AI therapy is associated with an increased risk of both hypercholesterolemia and hypertension. Data are insufficient to exclude clinically significant differences in cardiovascular disease associated with the AIs. Tamoxifen is associated with an increased risk of venous thromboembolic events, giving rise to a 1% to 2% greater risk of deep vein thrombosis compared with women taking AIs (Appendix Table A5).⁵¹ Data on the relative incidence of stroke with either tamoxifen or an AI are inconclusive. Longer follow-up is required to better characterize the potential cardiac toxicity of AI therapy.

Tamoxifen and AIs have differing effects on musculoskeletal health and symptoms. In comparison to tamoxifen, AIs are associated with greater loss of bone mineral density¹⁷ and fractures (Appendix Table A6). The incidence of osteoporosis and bone fractures⁷⁹ differs by approximately 2% to 4% in trials of primary adjuvant endocrine therapy comparing tamoxifen with an AI^11,19,21; risk is increased with AI therapy. Randomized clinical trials suggest that bisphosphonate therapy can mitigate AI-associated loss of bone density.^21,80–82 The long-term impact of AI treatment on osteoporosis risk and fracture risk has not been characterized.

With maturation of clinical data and accumulating clinical experience, it is clear that AIs cause a musculoskeletal/arthralgia syndrome. This syndrome is characterized by bone and joint symptoms, frequently described by patients as pain, stiffness, or achiness that is symmetric and not associated with other signs of rheumatologic disorders.^46,83–86 The prevalence of this syndrome is unclear, though it seems to be widespread; most patients have mild to moderate symptoms. There are no known interventions of proven value for AI-associated musculoskeletal symptoms. Discontinuation of AI therapy usually relieves symptoms within 8 to 10 weeks.

Tamoxifen and AIs have differing effects on gynecologic health; adverse events are more common among women receiving tamoxifen. In general, tamoxifen is associated with an increased risk of uterine cancer (approximately 1% of patients), benign endometrial pathology (including bleeding, polyps, and hyperplasia), hysterectomy, and vaginal discharge (Appendix Table A7). AIs seem to be less frequently associated with hot flashes than tamoxifen is (Appendix Table A8). The Arimidex, Tamoxifen, Alone or in Combination (ATAC) and MA.17 trials both reported a lower incidence of vaginal dryness among patients treated with tamoxifen.^74,77 Alternatively, the Intergroup Exemestane Study (IES) documented similar rates of vaginal dryness among both treatment arms,¹⁸ and the TEAM trial reported a statistically higher rate of vaginal dryness in patients who received sequential tamoxifen and exemestane compared with exemestane alone.⁵⁶ Findings were mixed with respect to loss of libido: MA.17 and IES noted similar rates, while AI-treated patients had higher rates in the ATAC trial (Appendix Table A8).^31,75,76

Clinical question 4.

Are AIs effective adjuvant therapy for women who are premenopausal at the time of diagnosis?

Recommendation 4.

The Update Committee recommends that women who are pre- or perimenopausal at the time of breast cancer diagnosis be treated with 5 years of tamoxifen.

Additional considerations.

The Update Committee recommends that clinicians use caution in evaluating menopausal status of patients who were pre- or perimenopausal at diagnosis. Unequivocal determination of menopausal status may be challenging to prove. Even among women who have not experienced menses for more than 1 year, laboratory testing is inadequate because patients may recover ovarian function. This particularly applies to those patients who experience chemotherapy- or tamoxifen-induced amenorrhea.

Literature update and discussion 4.

AI therapy has been shown to be effective only in postmenopausal women and is contraindicated in patients with residual ovarian function. Patients accrued to ABCSG-12, the only trial to include premenopausal women, were all treated with gonadotropin-releasing hormone agonist therapy to achieve a postmenopausal state.²² Eligible patients had favorable prognosis and low-grade breast cancer, and none received adjuvant chemotherapy, though 5% did receive neoadjuvant chemotherapy. These patients are not necessarily representative of younger women with early-stage breast cancer. ABCSG-12 demonstrated equivalence with respect to time to recurrence, disease-free survival, and overall survival between tamoxifen and AI therapy in premenopausal women given ovarian suppression.²² Because of tamoxifen equivalence with AI therapy in that setting and the occasional failure to achieve menopausal status with ovarian suppression, the Update Committee strongly recommends tamoxifen as primary adjuvant endocrine therapy for all pre- or perimenopausal women and women with treatment-induced amenorrhea.

Some women who were pre- or perimenopausal at the time of diagnosis may become unequivocally postmenopausal in subsequent years. For these patients, the Update Committee suggests incorporating AIs as either sequential or extended adjuvant endocrine therapy. Relatively few women in the studies of sequential or extended therapy met this description. Thus, the magnitude of benefit for introducing an AI in the subsequent treatment of such women is not well characterized.⁸⁷

Both chemotherapy and tamoxifen can contribute to amenorrhea. This effect may be transient or permanent, depending on the patient's age and therapies received. Multiple reports document late clinical recovery of ovarian function among women with treatment-induced amenorrhea, which could render AI therapy ineffective.^88,89 Cessation of menses for 1 year is the clinical hallmark of menopause. However, this clinical definition applies only to women without health conditions, surgery, or medications that contribute to amenorrhea, such as chemotherapy or tamoxifen and, thus, may not apply to many breast cancer patients.

At present, the role of ovarian suppression in addition to tamoxifen for premenopausal patients is not known. The Suppression of Ovarian Function Trial (SOFT) is comparing tamoxifen, tamoxifen plus ovarian suppression, and exemestane plus ovarian suppression, and it continues to accrue patients. Findings from this trial will further define best practices for premenopausal patients as well as those patients who experience treatment-induced menopause.

Clinical question 5.

Can the third-generation AIs be used interchangeably?

Recommendation 5.

In the absence of direct comparisons, the Update Committee interprets available data as suggesting that benefits of AI therapy represent a “class effect.” Meaningful clinical differences between the commercially available third-generation AIs have not been demonstrated to date. In the clinical opinion of the Update Committee (rather than direct evidence from randomized trials), postmenopausal patients intolerant of one AI but who are still candidates for adjuvant endocrine therapy may be advised to consider tamoxifen or a different AI.

Literature update and discussion 5.

Previous results were limited to reports for principal use of a single AI in each of the clinical settings of primary, sequential, or extended adjuvant therapy. There are still no data from head-to-head comparisons of AIs. However, there are data from randomized trials for each of the commercially available third-generation AIs for all of the adjuvant treatment strategies (primary, sequential, and extended; Fig 1). The Update Committee interprets the existing data comparing the AIs with tamoxifen as qualitatively similar with respect to efficacy and tolerability. Toxicity reports (Appendix Tables A4 through A8) have not suggested obvious clinical advantages of one AI over another with respect to compliance, constitutional or menopausal symptoms, bone health, cardiovascular disease, or quality of life. Anecdotal experience suggests that patients may tolerate one AI better than another, but patterns are neither predictable nor consistent. Two trials—MA.27 and Femara versus Anastrozole Clinical Evaluation (FACE)—are directly comparing one AI against another as primary adjuvant therapy. However, data are not yet available from either trial.

PATIENT COMMUNICATION

The purpose of this section is to address aspects of patient-provider communication that play a role in decision making about the use of adjuvant endocrine therapy, the selection of agent, and the barriers to adherence to treatment regimens (taking medication as prescribed) and persistence with the medication schedule (taking medication for the full duration prescribed). Separate literature searches and Update Committee members' suggestions, rather than the systematic review, were used to prepare this section.

Patients need to be informed about risk factors for tumor recurrence, the role of residual subclinical (ie, microscopic) disease in causing recurrence, and the potential benefit of adjuvant endocrine therapy. Clinicians can base risk estimates in women with hormone receptor–positive disease on well-established prognostic markers such as stage, HER2 status, and grade. Emerging molecular diagnostic assays such as the 21-gene recurrence score also seem to serve as prognostic markers in ER-positive, node-negative breast cancer.^90,91 Decision tools such as Adjuvant! Online⁹² quantify and communicate, in broad terms, both the risk of cancer recurrence and the benefit of various adjuvant endocrine treatment strategies on the basis of a patient's tumor characteristics, comorbidity, age, and receipt of chemotherapy.⁹³

Rates of nonpersistence (early discontinuation of medications) in women who start taking tamoxifen are as high as 30% at 3 years after filling a first prescription. For example, a study of 2,816 women identified a 22% rate of nonpersistence 12 months after starting tamoxifen.⁹⁴ In another study of 392 patients, 32% of women discontinued adjuvant tamoxifen at 2 years and 39% at 3 years.⁹⁵ Similarly, in a sample of 961 women 65 years of age or older, 24% discontinued tamoxifen after 2 years of treatment, 33% discontinued tamoxifen after 3 years of treatment, and 50% discontinued tamoxifen before 5 years; discontinuation was not due to recurrent disease.⁹⁶

Data from clinical trials⁹⁷ and claims-based research⁹⁸ indicate that persistence is no better with AIs. Physicians may underestimate rates of nonadherence and nonpersistence, and patients may be reluctant to disclose problems with adherence and persistence. Patient beliefs about the benefits and risks of medications are associated with adherence and persistence; thus, discussing and addressing such beliefs is warranted.⁹⁹

Adverse effects are particularly important in precipitating early discontinuation of therapy. Patients who experience treatment-related adverse effects are more likely to discontinue adjuvant endocrine therapy, particularly during the first years.^100,101 Adverse effects were the most common reason for discontinuation, particularly during the first year according to one study. Among those who stopped tamoxifen in the first year, 70% stopped because of adverse effects.⁹⁶ Patients who experience adverse effects for which they were not prepared seem to be at particularly high risk for nonpersistence.¹⁰¹ Similarly, musculoskeletal adverse effects of the AIs have been shown to prompt discontinuation in more than 10% of patients who start an AI.⁸⁵ Information support for patients about anticipated adverse effects and management of those adverse effects may increase persistence.⁹¹

Financial constraints are another reported cause of nonpersistence with therapy. One study of patients taking tamoxifen noted that 60% of patients who discontinued therapy early reported this issue as a key factor.⁹⁶ It is likely that the out-of-pocket costs of AIs (Table 6) pose an even greater barrier to patients. Although physicians are generally reluctant to discuss costs of cancer therapies,¹⁰² inquiring about patients' cost concerns may help direct patients to assistance programs and create opportunities to stress the benefits of persistence with adjuvant endocrine therapy. ASCO supports the development of resources to facilitate patient-provider communication about costs of cancer care.¹⁰³

Table 6.

Cost of Therapies

Drug	Cost ($)
	Per Unit	30-Day Supply
Anastrozole (Arimidex), 1 mg	12.66	379.80
Exemestane (Aromasin), 25 mg	11.38	341.40
Letrozole (Femara), 2.5 mg	13.91	417.30
Tamoxifen (Nolvadex), 20 mg	3.67	110.10
Tamoxifen, 20 mg	0.73	21.90

NOTE. US dollars; www.drugstore.com.

In summary, optimal adjuvant endocrine therapy includes careful consideration of tumor risk, treatment benefits and adverse effects, and patient adherence. Clinicians should discuss realistic, quantitative risks of cancer recurrence and death and benefits from cancer therapy as part of the adjuvant therapy decision-making process. Clinicians should alert patients to common adverse effects of therapy and serially inquire about treatment-related toxicities, patient adherence, and factors that may affect adherence and persistence.

HEALTH DISPARITIES

ASCO clinical practice guidelines represent expert recommendations derived from high-quality evidence on the best practices in disease management. However, racial, ethnic, and socioeconomic disparities in quality of health care exist and persist in the United States. Members of racial and ethnic minority groups and patients with fewer financial resources tend to have a higher burden of comorbid illness, are more likely to be uninsured or underinsured, face more challenges in accessing care, and are at greater risk of receiving poor-quality care than other Americans.^104–108

Representation of minorities in clinical trials of adjuvant endocrine therapy is low. To date, no evidence indicates differences in therapeutic benefit between black and white women receiving adjuvant tamoxifen in the clinical trials setting. A few studies suggest that women who belong to minority groups are less likely to receive guideline-concordant endocrine therapy. One study of prescribing patterns¹⁰⁹ indicated that rates of guideline-concordant adjuvant endocrine therapy prescribing were lower in Hispanic (71%) and black (75%) patients compared with non-Hispanic white (85%) patients.

Only small samples of minority patients were included in persistence studies, largely preventing an examination of correlates and mechanisms of optimal therapy in minorities. Hispanics have largely been omitted. In one study,⁹⁵ nonwhite patients (who were pooled because they represented only 17% of the sample) were more likely to stop adjuvant endocrine therapy before completing planned treatment. Other studies addressing racial differences in adherence to or persistence with endocrine therapy generated mixed results.^96,101 Race was not associated with adherence or persistence in one recently published claims-based study of low-income patients with Medicaid¹¹⁰; 80% of patients who started adjuvant endocrine therapy were persistent with medication at 1 year of follow-up. Awareness of disparities in quality of care should be considered in the context of this clinical practice guideline.

SUMMARY AND FUTURE DIRECTIONS

Incorporation of an AI improves disease-free survival in postmenopausal women with hormone receptor–positive breast cancer compared with tamoxifen alone. Thus, the Update Committee recommends AI therapy at some point during adjuvant treatment, either as upfront therapy or as sequential or extended treatment after tamoxifen. The optimal timing and duration of AI treatment remain unresolved; it is unclear whether sequential treatment strategies yield advantages over monotherapy with AIs. The Update Committee recognizes distinct adverse effect profiles of tamoxifen and AIs and believes that consideration of adverse effect profiles and patient preferences are relevant to deciding whether and when to incorporate AI therapy.

Important and unanswered questions about adjuvant endocrine therapy in postmenopausal women persist. The Update Committee identified the following issues as clinically significant for ongoing research studies and treatment recommendations:

• Long-term follow-up to characterize the lasting clinical effects of AIs and impact on survival, survivorship, and quality of life

• Comparisons of currently available AIs against one another

• Determination of optimal schedules for endocrine therapy, including duration of treatment, interrupted treatment schedules, and sequencing

• Late adverse effects of AI therapy

• Strategies to improve adherence and minimize disparities in access to therapy

• Interventions to minimize treatment-related adverse effects among women receiving adjuvant endocrine therapy

• Comparative effectiveness analyses of adjuvant endocrine strategies based on efficacy, toxicity, and cost

• Development of biomarker(s) for selection of endocrine strategies and for refining risk estimates in postmenopausal, ER-positive breast cancer

• Identification of predictors of late (after 5 or 10 years) recurrence to tailor durations of therapy

• Definitive analyses of the role of drug metabolism and pharmacogenetics as predictors of benefit and/or treatment options in adjuvant endocrine therapy

• Incorporation of novel antiestrogens or other treatments to enhance endocrine therapy and reduce recurrence risk

• Clarification of the role of AI therapy and ovarian suppression among women who are premenopausal at the time of breast cancer diagnosis

The Update Committee anticipates that the results from prospective randomized trials, ongoing correlative science studies, long-term follow-up from major adjuvant studies, and smaller, focused investigations of related scientific and clinical questions regarding endocrine treatment will continue to inform and revise the recommendations for adjuvant endocrine therapy in the years ahead.

Appendix

Search Strategy for MEDLINE (PUBMED)

(Breast neoplasms[mesh] NOT (“Breast Neoplasms, Male”[Mesh] OR “Phyllodes Tumor”[Mesh] OR “Carcinoma, Ductal, Breast”[Mesh]) OR (“Breast”[Mesh] AND (“Carcinoma”[Mesh] OR “Adenocarcinoma”[Mesh] OR “Neoplasms”[Mesh]))) OR ((breast[tiab] OR mammary[tiab]) AND (cancer[tiab] OR cancers[tiab] OR tumor[tiab] OR tumors[tiab] OR tumour[tiab] OR tumours[tiab] OR malignan*[tiab] OR carcinoma[tiab] OR carcinomas[tiab] OR adenocarcinoma[tiab] OR adenocarcinomas[tiab])) AND

((“Aromatase Inhibitors”[Mesh] OR aromatase inhibitors[nm] OR “aromatase inhibitor”[TIAB] OR “aromatase inhibitors” [TIAB] OR “aromatase inhibition”[TIAB] OR “anastrozole”[Substance Name] OR anastrozole [tiab] OR arimedex[tiab] OR “letrozole”[Substance Name] OR “letrozole”[tiab] OR femara[tiab] OR aromasil[tiab]OR exemestane [Substance Name] OR “exemestane”[tiab] OR aromasin[tiab] OR aromasine[TIAB] OR Tamoxifen[Mesh] OR Tamoxifen[substance name] OR “Tamoxifen”[tiab] OR novaldex [tiab]) NOT (aminoglutethimide[mesh] OR aminoglutethimide[tiab])) AND

(((randomized controlled trial[pt] OR controlled clinical trial[pt] OR randomized controlled trials[mh] OR clinical trial[pt] OR “clinical trial”[tiab] OR “clinical trials”[tiab] OR clinical trials as topic[mh] OR controlled clinical trials as topic[mh] OR randomized controlled trials as topic[mh] OR clinical trials, phase II as topic[mh] OR clinical trials, phase III as topic[mh] OR clinical trials, phase IV as topic[mh] OR clinical trial, phase II[pt] OR clinical trial, phase III[pt] OR clinical trial, phase IV[pt] OR random allocation[mh] OR “random allocation”[tiab] OR “randomly allocated”[tiab] OR double-blind method[mh] OR single-blind method[mh]) OR ((random[tiab] OR randomly[tiab] OR randomized[tiab] OR randomised[tiab] OR randomization[tiab] OR randomisation[tiab]) AND (clinical[tiab] OR control[tiab] OR controlled[tiab] or control groups[mh])) OR ((single[tiab] OR single-[tiab] OR double[tiab] OR double-[tiab] OR triple[tiab] OR triple-[tiab] OR multi[tiab] OR multi-[tiab]OR evaluator[tiab] OR assessor[tiab] OR interviewer[tiab]) AND (mask[tiab] OR masked[tiab] OR masking[tiab] OR blind[tiab] OR blinded[tiab] OR blinding[tiab])) OR ((placebos[mh] OR placebo[tiab] OR placebos[tiab] OR random[tiab] OR randomly[tiab] OR randomized[tiab] OR randomised[tiab] OR randomization[tiab] OR randomization[tiab]) AND (research design[mh] OR “comparative study”[tiab] OR comparative study[pt] OR evaluation studies as topic[mh:noexp] OR evaluation studies[pt] OR “evaluation study”[tiab] OR “evaluation studies”[tiab] OR validation studies as topic[mh] OR follow-up studies[mh] OR “follow-up study”[tiab] OR “follow up study”[tiab] OR “follow-up studies”[tiab] OR “follow up studies”[tiab] OR prospective studies[mh] OR prospective[tiab] OR epidemiologic research design[mh] OR epidemiologic methods[mh] OR epidemiologic study characteristics as topic[mh] OR epidemiologic studies[mh] OR intervention studies[mh] OR cross-over studies[mh]))) NOT (clinical trial, phase I[pt] OR clinical trials, phase I as topic[mh]) NOT (animals[mh] NOT humans[mh])) AND English[la]

Fig A1.

Quorum diagram. Inclusion and exclusion of publications identified for this systematic review.

Table A1.

Update Committee Members

Member	Affiliation
Harold J. Burstein, MD, PhD, Co-Chair	Dana-Farber Cancer Institute
Jennifer J. Griggs, MD, MPH, Co-Chair	University of Michigan Comprehensive Cancer Center
Holly Anderson, RN, BSN, Patient Representative	Breast Cancer Coalition of Rochester
Diana Rowden, MS, Patient Representative	Susan G. Komen for the Cure
Thomas A. Buchholz, MD	M. D. Anderson Cancer Center
Nancy E. Davidson, MD	University of Pittsburgh Cancer Institute
Karen A. Gelmon, MD	British Columbia Cancer Agency
Sharon Hermes Giordano, MD	M. D. Anderson Cancer Center
Clifford Hudis, MD	Memorial Sloan-Kettering Cancer Center
Jennifer Malin, MD, PhD	Greater Los Angeles Veterans Affairs Healthcare System
Eleftherios P. Mamounas, MD	Aultman Health Foundation
Alexander J. Solky, MD	Interlakes Oncology and Hematology
MaryFran R. Sowers, PhD	University of Michigan
Vered Stearns, MD	Johns Hopkins School of Medicine
Eric P. Winer, MD	Dana-Farber Cancer Institute

Table A2.

Time to Recurrence

Trial	Median Follow-Up (months)	Range (months)	No. of Patients Observed		Time to Recurrence Events				HR	95% CI	P
					AI		Comparator
			AI	Comparator	No.	%	No.	%
Primary
ATAC20 ITT	100		3,125	3,116	538		645		0.81	0.73 to 0.91	< .001
Sequential
TEAM56 EXE/TAM v EXE ITT	61		4,868	4,898		10.9		10.2	0.94	0.83 to 1.06	.293
IES13	55.7	0-89.7	2,352	2,372	265		325		0.7	0.58 to 0.83*	NR
N-SAS BC-038	42	3.2-60	347	349	15	4.3	27	7.7	0.54	0.29 to 1.02†‡	.06‡
ARNO 9533	30.1	< 12 to ≥ 84	489	490	36	7.4	47	9.6	NR		NR
Extended
MA.1725	64		2,583	2,587	136	5.3	190	7.3	NR		NR
NSABP B-3336	30		783	779	17	2.2	37	4.7	0.44	NR§	.004

NOTE. Percent calculated as number of events divided by number of patients observed.

Abbreviations: AI, aromatase inhibitor; HR, hazard ratio; ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ITT, intent to treat; TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; TAM, tamoxifen; IES, Intergroup Exemestane Study; N-SAS, National Surgical Adjuvant Study (Group); ARNO, Arimidex-Nolvadex (trial); NR, not reported; NSABP, National Surgical Adjuvant Breast and Bowel Project.

^*Disease recurrence.

^†Relapse-free survival.

^‡Two different sets of values reported on abstract and poster; table reflects poster (abstract: aromatase inhibitor, N = 15 [4.3%]; comparator: N = 28 [NR]; hazard ratio, 0.52; 95% CI, 0.28 to 0.98; P = .02).

^§Relapse-free survival follow-up for all TEAM patients is 33 months (data beyond that point censored).

Table A3.

Disease-Free Survival Subset Analyses

Trial	Median Follow-Up (months)	Range (months)	No. of Patients Observed		Disease-Free Survival Events				HR	95% CI	P
					AI		Comparator
			AI	Comparator	No.	%	No.	%
Primary
ATAC20,60(Goss PE, et al: J Natl Cancer Inst 97:1262-1271, 2005) ANA v TAM ITT	100	0-126	3,125	3,116	817	26.1	887	28.5	0.90	0.82 to 0.99	.025
Hormone receptor–positive	100	0-126	2,618	2,598	618	23.6	702	27.0	0.85	0.76 to 0.94	.003
ANA v TAM ITT	68		Total: 9,366		402	12.9	498	16.0	0.79	0.70 to 0.90*	< .001
Node-positive	68		NR		NR		NR		0.84	0.70 to 1.00	NR
Node-negative	68		NR		NR		NR		0.68	0.55 to 0.84	NR
Node-unknown	68		NR		NR		NR		0.48	0.23 to 1.00	NR
ER-positive/PR-positive	68		Total: 5,719		191	10	222	12	0.84	0.69 to 1.02*	.07
ER-positive/PR-negative	68		Total: 1,372		50	11	102	24	0.43	0.31 to 0.61*	< .001
ER-negative/PR-positive	68		Total: 220		17	27	25	33	0.79	0.43 to 1.47*	.5
ER-negative/PR-negative	68		Total: 703		66	28	79	32	0.90	0.65 to 1.25*	.5
ER-positive/PR-unknown	68		Total: 518		22	13	20	11	1.29	0.71 to 2.37*	.4
ER- and PR-unknown	68		Total: 743		46	19	47	19	0.96	0.64 to 1.44*	.8
BIG 1-9841,44,49 LET v TAM	76		2,463	2,459	509	20.7	565	23.0	0.88	0.78 to 0.99	.03
Node-negative or node-unknown	71		1,376	1,404	200	14.5	227	16.2	0.89	0.74 to 1.08	NR†
Node-positive	71		1,050	1,017	303	28.9	332	32.6	0.83	0.71 to 0.98	NR
ERBB2-negative	51		1,648	1,646	178	1.08	240	14.6	0.72	0.59 to 0.87	NR
ERBB2-positive	51		134	105	27	20.1	32	30.5	0.62	0.37 to 1.03	NR
Low Ki-67	51		730	703	56	7.7	66	9.4	0.81	0.57 to 1.15	.09 (for interaction between low and high Ki-67)
High Ki-67	51		631	621	66	10.5	115	18.5	0.53	0.39 to 0.72
ER-positive/PR-positive	25.8		2,542	2,513	179	7	208	8.3	0.84	0.49 to 1.03	.09†
ER-positive/PR-negative	25.8		808	823	89	11	107	13	0.83	0.62 to 1.10	.18†
ER-positive/PR-unknown	25.8		579	575	70	12.1	92	16	0.72	0.53 to 0.98	.04†
Sequential
BIG 1-9841 TAM/LET v LET	71		1,548	1,546	259	16.7	248	16.0	1.05	0.84 to 1.32‡	NR
Node-negative or node-unknown	71		894	886	100	11.2	102	11.5	0.96	0.67 to 1.38	NR
Node-positive	71		635	645	154	24.2	145	22.5	1.13	0.84 to 1.52	NR
LET/TAM v LET	71		1,540	1,546	236	15.3	248	16.0	0.96	0.76 to 1.21‡	NR
Node-negative or node-unknown	71		920	886	101		102		0.97	0.67 to 1.39	NR
Node-positive	71		611	645	135		145		0.98	0.72 to 1.59	NR
IES13 ITT	55.7	0-89.7	2,352	2,372	354		455		0.76	0.66 to 0.88	< .001
Node-negative	55.7	0-89.7	1,217	1,230	NR		NR		0.74	0.58 to 0.94	NR
Node-positive	55.7	0-89.7	1,050	1,039	NR		NR		0.72	0.61 to 0.86	NR
ER-positive/ER-unknown	55.7	0-89.7	2,296	2,306	339	14.8	439	19.0	0.75	0.65 to 0.87	< .001
ER and PR high	NR		NR	NR	NR		NR		0.686		NR
Extended
MA.1725 ITT§	64	16-95	2,583	2,587	164	6.3	235	9.1	0.68	0.55 to 0.83	< .001
Node-positive	64		1,174	1,194		91.7 DFS at 4 years		87.8 DFS at 4 years	0.74	0.58 to 0.94	.01
Node-negative	64		1,295	1,281		97.0 DFS at 4 years		94.6 DFS at 4 years	0.51	0.35 to 0.75	< .001
ITT§	30		2,583	2,587	92	3.6	155	6.0	0.58	0.45 to 0.76	< .001
ER-positive/PR-positive	30		1,907	1,902	60	3.0	117	6.0	0.49	0.36 to 0.67	NR
ER-positive/PR-negative	30		309	327	19	6.0	17	5.0	1.21	0.63 to 2.34	NR
ER-negative/PR-positive	30		100	100	4	4.0	5	5.0	0.56	0.15 to 2.12	NR
NSABP B-3336	30		783	779	37	4.7	52	6.7	0.68		.07
Node-positive	30		Total: 755		NR		NR		0.5	0.30 to 0.86	.01
Node-negative	30		Total: 807		NR		NR		1.13	NR	.74
ER-positive and PR-positive	30		Total: 1,266		NR		NR		0.7	NR	.15
ER-positive or PR-positive	30		Total: 248		NR		NR		0.52	NR	.17

NOTE. Percent calculated as number of events divided by number of patients observed.

Abbreviations: AI, aromatase inhibitor; HR, hazard ratio; ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ANA, anastrozole; TAM, tamoxifen; ITT, intent to treat; NR, not reported; ER. estrogen receptor; PR, progesterone receptor; BIG, Breast International Group; LET, letrozole; IES, Intergroup Exemestane Study; DFS, disease-free survival; NSABP, National Surgical Adjuvant Breast and Bowel Project.

^*Time to recurrence.

^†Relapse-free survival, local and distant recurrence, contralateral breast cancer, and death without recurrence.

^‡99% CIs used to account for multiple comparisons.

^§Seventeen patients omitted from ITT analysis.

Table A4.

Adverse Cardiovascular (cardiovascular deaths, ischemic cardiac disease, hypertension, high cholesterol) Effects

Trial	Median Follow-Up (months)	Arm	No. of Patients	Cardiovascular Events
				All Cardiac Events		Grades 3 to 5 Cardiac Events		Cardiovascular Deaths		All Ischemic Cardiac Disease		Grades 3 to 5 Ischemic Cardiac Disease		Any Grade Hypertension		High Cholesterol
				No.	%	No.	%	No.	%	No.	%	No.	%	No.	%	No.	%
ATAC (Howell A, et al: Lancet 365:60-62, 2005; Buzdar et al: Lancet Oncol 7:633-643, 2006)	68	ANA	3,092					49	2	127	4.10			402	13	278	9
		TAM	3,094					46	1	104	3.40			349	11	108	3
		OR						NR		1.23				1.18		2.73
		95% CI						NR		0.95 to 1.60				0.69 to 1.44		2.02 to 3.69
		P						NR		.1				.04		< .001
BIG 1-9814,35,40	30.1	LET	3,975	140	5.7	96	2.4			54	2.2	45	1.1	150	3.8		43.60*
		TAM	3,988	127	5.2	57	1.4			45	1.8	29	0.7	135	3.4		19.20
		RR		NR		1.68				NR		1.55		1.11		NR
		P		.45		.001				.42		.06		.37
ABCSG-829	75	TAM/ANA	1,469	127	6.5
		TAM	1,453	130	6.0
		P		.479
ITA9 (Boccardo F, et al: J Clin Oncol 23:5138-5147, 2005)	64	ANA	223	17	7.6†											19	9.3*
		TAM	225	14	6.2											6	4.0
		P		.6												.04
TEAM56	61	TAM/EXE	4,814							60	1.2‡			215	4.5	134	2.8
		EXE	4,852							77	1.6			293	6.0	227	4.7
		P								.183				.001		< .001
IES13	55.7	EXE	2,320	382	16.5§	23	1	14	< 1	185	8.0	185	8	830	35.8	166	7.2
		TAM	2,338	350	15.0	18	< 1	13	< 1	162	6.9	162	6.9	772	33.0	141	6.0
		P		.16		NR		NR		.17		.17		.05		.12
ARNO 9533	30.1	ANA	445							9	2.0
		TAM	452							4	0.9
N-SAS BC-038	42	TAM/ANA	347		0.6¶
		TAM	349		0.9
MA.1724	30	LET	2,572		5.80												16.00
		Placebo	2,577		5.60												16.00
		P		.76												.79

Abbreviations: ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ANA, anastrozole; TAM, tamoxifen; OR, odds ratio; NR, not reported; BIG, Breast International Group; LET, letrozole; RR, risk ratio; ABCSG, Austrian Breast and Colorectal Cancer Study Group; ITA, Italian Tamoxifen Anastrozole (trial); TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; IES, Intergroup Exemestane Study; ARNO, Arimidex-Nolvadex (trial); N-SAS, National Surgical Adjuvant Study (Group).

^*Lipid metabolism disorder.

^†Cardiovascular disease.

^‡Myocardial ischemia/infarction.

^§Excludes venous thromboembolic events.

^¶Heart disease.

Table A5.

Adverse Cardiovascular (ischemic cerebrovascular, stroke, or TIA thrombosis) Effects

Trial	Median Follow-Up (months)	Arm	No. of Patients	Cardiovascular Events
				All Ischemic Cerebrovascular Events		Grades 3 to 5 Stroke or TIA		All Venous Thrombosis Events		Grades 3 to 5 Venous Thrombosis Events
				No.	%	No.	%	No.	%	No.	%
ATAC20 (Howell A, et al: Lancet 365:60-62, 2005; Arimidex, Tamoxifen, Alone or in Combination Trialists' Group, et al: Lancet Oncol 7:633-643, 2006)	68	ANA	3,092	62	2	20	0.16*	87	2.80
		TAM	3,094	88	2.8	34	0.28	140	4.50
		OR		0.70		NR		0.61
		95% CI		0.50 to 0.97		NR		0.46 to 0.80
		P		.03		NR		< .001
BIG 1-9840	30.1	LET	3,975			47	1.2	68	1.7	35	0.9
		TAM	3,988			49	1.2	154	3.9	92	2.3
		RR				1		0.44		0.38
		P				0.99		< .001		< .001
ABCSG-1222	47.8	ANA + GOS	453					0
		TAM + GOS	451					0
ABCSG-829	75	TAM/ANA	1,469					44	2.3†
		TAM	1,453					47	2.2
		P						.833
TEAM56	61	TAM/EXE	4,814	35	0.7			97	2.0
		EXE	4,852	51	1.1			45	0.9
		P		.112				< .001
IES13	55.7	EXE	2,320					28	1.2	7	< 1
		TAM	2,338					54	2.3	19	< 1
		P						.004		NR
ARNO 9533	30.1	ANA	445	3	0.7
		TAM	452	1	0.2
N-SAS BC-038	42	TAM/ANA	347								0.3
		TAM	349								0.0

Abbreviations: TIA, transient ischemic attack; ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ANA, anastrozole; TAM, tamoxifen; OR, odds ratio; NR, not reported; BIG, Breast International Group; LET, letrozole; RR, risk ratio; ABCSG, Austrian Breast and Colorectal Cancer Study Group; GOS, goserelin; TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; IES, Intergroup Exemestane Study; ARNO, Arimidex-Nolvadex (trial); N-SAS, National Surgical Adjuvant Study (Group).

^*Cerebrovascular accidents.

^†Embolism, thromboses.

Table A6.

Adverse Musculoskeletal Effects

Trial	Median Follow-Up (months)	Arm	No. of Patients	Musculoskeletal Events
				Bone Pain		Musculoskeletal Pain		Arthralgia		Osteoporosis		Fracture Rate		Decrease in Hip BMD	Decrease in Lumbar Spine BMD
				No.	%	No.	%	No.	%	No.	%	No.	%
ATAC17,20 (Howell A, et al: Lancet 365:60-62, 2005; Arimidex, Tamoxifen, Alone or in Combination Trialists' Group, et al: Lancet Oncol 7:633-643, 2006)	68	ANA	3,092			132	4a	1,100	35.6	325	11b	375	12.1	−7.24c	− 6.08c
		TAM	3,094			235	8	911	29.4	226	7	234	7.6	0.74	−2.77
		OR				0.54		1.32		1.49		1.55		NR	NR
		95% CI				0.41 to 0.72		1.19 to 1.47		1.18 to 1.88		1.25 to 1.77		NR	NR
		P				< .001		< .001		< .001		< .001		NR	NR
BIG 1-9814,35	30.1	LET	3,975									196	8
		TAM	3,988									132	5.4
		P										< .001
ABCSG-1222	47.8	ANA + GOS	451	128	28.3			112	4.7			1	0.2		−7.8%d
		TAM + GOS	453	94	20.8			52	11.5			1	0.2		−4.5%
ABCSG-829	75	TAM/ANA	1,469	223	11.4e							69	3.5
		TAM	1,453	179	8.2							35	1.6
		P		< .001								< .001
ITA9	64	ANA	223			22	9.9f
		TAM	225			15	6.7
		P				.2
TEAM56	61	TAM/EXE	4,814			302	6.3g	961	20.0	259	5.4	170	3.5
		EXE	4,852			188	3.9	1,140	23.5	478	9.9	249	5.1
		P				< .001		< .001		< .001		< .001
IES12,13	55.7	EXE	2,320			488	21.0	432	18.6	169	7.3	100	4.3	−2.799	−3.726
		TAM	2,338			376	16.1	275	11.8	128	5.5	73	3.1	−1.011	−0.412
		P				< .001		< .001		.01		.03		NR	NR
N-SAS BC-038	42	TAM/ANA	347						50.4
		TAM	349						31.8
ARNO 9533	30.1	ANA	445					52	11.7h	13	2.9	10	2.2
		TAM	452					22	4.9	4	0.9	10	2.2
ABCSG-6a30	62.3	ANA	387	86	18.3i
		Placebo	469	95	24.5
		OR		1.55
		95% CI		1.11 to 2.17
		P		.009
NSABP B-3336	30	EXE	799		0.50j				1.00j			28
		Placebo	799		0.70				0.50			20
		P		NR				NR				.33
MA.1724 (Goss PE, et al: J Natl Cancer Inst 97:1262-1271, 2005)	30	LET	2,572		5		15k		25		8.10		5.30	−3.60%l	−5.35%l
		Placebo	2,577		6		12		21		6		4.60	−0.71%	−0.70%
		P		.67		.004		< .001		.003		.25		.044	.008

Abbreviations: BMD, bone mineral density; ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ANA, anastrozole; TAM, tamoxifen; OR, odds ratio; NR, not reported; BIG, Breast International Group; LET, letrozole; ABCSG, Austrian Breast and Colorectal Cancer Study Group; GOS, goserelin; ITA, Italian Tamoxifen Anastrozole (trial); TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; IES, Intergroup Exemestane Study; N-SAS, National Surgical Adjuvant Study (Group); ARNO, Arimidex-Nolvadex (trial); NSABP, National Surgical Adjuvant Breast and Bowel Project.

^aMuscle cramps.

^bOsteoporosis or osteopenia.

^cChange in BMD from baseline to year 5.

^dChanges at 60 months.

^eBone or joint disorder.

^fIncludes both musculoskeletal disorders and bone fractures.

^gMuscle cramps/disorders.

^hBoth arthralgia and bone pain.

ⁱGrades 1 to 4 only.

^jGrades 3 and 4 only.

^kMuscle pain.

^lBMD from 24 months: N (LET) = 122; N (placebo) = 104 cases only.

Table A7.

Adverse Gynecologic Effects

Trial	Median Follow-Up (months)	Arm	No. of Patients	Gynecologic Events
				Gynecologic Disorders		Vaginal Bleeding		Vaginal Discharge		Endometrial Biopsy		Endometrial Cancer		Uterine Polyps
				No.	%	No.	%	No.	%	No.	%	No.	%	No.	%
ATAC (Howell A, et al: Lancet 365:60-62, 2005)	68	ANA	3,092	95	3a	167	5.4	109	3.5			5	0.2
		TAM	3,094	324	10	317	10.2	408	13.2			17	0.8
		OR		0.27		0.50		0.24				0.29
		95% CI		0.20 to 0.37		0.41 to 0.61		0.19 to 0.30				0.11 to 0.80
		P		< .001		< .001		< .001				.02
BIG 1-985,35	25.8	LET	3,975				3.3				2.3	6	0.1
		TAM	3,988				6.6				9.1	15	0.3
		P				< .001				< .001		NR
ABCSG-1222	47.8	ANA + GOS	453											1	0.2
		TAM + GOS	451											5	1.1
ABCSG-829	75	TAM/ANA		442	22.6
		TAM		681	31.2
		P		< .001
ITA9	64	ANA	223	16	7.2b
		TAM	225	9	4.0
		P		.1
TEAM56	61	TAM/EXE	4,814	94	2.0c	243	5.0	402	8.4	187	3.9 d	17	0.4
		EXE	4,852	64	1.3	114	2.3	123	2.5	20	0.4	7	0.1
		P		.017		< .001		< .001		< .001		.040
IES5,13	55.7	EXE	2,320			91	4.6	65	2.8	1	< 0.1e	5	0.2	24	1.2f
		TAM	2,338			131	6.5	91	3.9	20	0.9	11	0.5	65	(3.2)
		P				.008		.04		< .001		NR		< .001
N-SAS BC-038	42	TAM/ANA	347				9.2g		16.1			0
		TAM	349				8.0		24.4				0.3
ARNO 9533	30.1	ANA	445			7	1.6			8	1.8e	0		4	0.9
		TAM	452			9	2.0			39	8.6	2	0.4	15	3.3
ABCSG-6a30	62.3	ANA	387			3	0.8h	23	59
		Placbeo	469			1	0.2	13	2.8
		OR				3.84		2.34
		95% CI				1.40 to 37.06		1.17 to 4.68
		P				.245		.017
MA.175 (Goss PE, et al: J Natl Cancer Inst 97:1262-1271, 2005)	30	LET	2,572			145	6.0					4
		Placebo	2,577			196	8.0					11
		P				.005						.12

Abbreviations: ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ANA, anastrozole; TAM, tamoxifen; OR, odds ratio; BIG, Breast International Group; LET, letrozole; NR, not reported; ABCSG, Austrian Breast and Colorectal Cancer Study Group; GOS, goserelin; ITA, Italian Tamoxifen Anastrozole (trial); TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; IES, Intergroup Exemestane Study; N-SAS, National Surgical Adjuvant Study (Group); ARNO, Arimidex-Nolvadex (trial).

^aGynecologic events.

^bGynecologic symptoms.

^cOther vulvovaginal disorders.

^dEndometrial abnormalities.

^eEndometrial hyperplasia.

^fUterine fibroids or polyps.

^gGenital bleeding.

^hGrades 1 and 2 only.

Table A8.

Adverse Climacteric Effects

Trial	Median Follow-Up (months)	Arm	No. of Patients	Climacteric Symptoms
				Hot Flashes		Fatigue/Asthenia		Mood Disturbance		Decreased Libido		Nausea and Vomiting		Headache		Cognitive Disorder		GI Symptoms
				No.	%	No.	%	No.	%	No.	%	No.	%	No.	%	No.	%	No.	%
ATAC (Howell A, et al: Lancet 365:60-62, 2005; Arimidex, Tamoxifen, Alone or in Combination Trialists' Group, et al: Lancet Oncol 7:633-643, 2006)	68	ANA	3,092	1,104	35.7	575	18.6	597	19.3	39	1	393	12.7					265	9a
		TAM	3,094	1,264	40.9	544	17.6	554	17.9	12	< 1	384	12.4					216	7
		OR		0.80		1.07		1.10		3.28		1.03						1.25
		95% CI		0.73 to 0.89		0.94 to 1.22		0.97 to 1.25		1.4 to 7.7		0.88 to 1.19						0.98 to 1.60
		P		< .001		.3		.2		< .001		.7						.02
BIG 1-9835	25.8	LET	3,975		33.5
		TAM	3,988		38
		P		< .001
ABCSG- 1222	47.8	ANA + GOS	453	25	5.5	93	20.5	97	21.4b			32	7.1	63	13.9	3	0.7
		TAM + GOS	451	28	6.2	70	15.5	70	15.5			23	5.1	59	13.1		0
ABCSG-829	75	ANA/TAM	1,469													80	4.1c	67	3.4
		TAM	1,453													51	2.3	82	3.8
		P														.001		.616
ITA9	64	ANA	223			4	1.8											26	11.7d
		TAM	225			0												15	6.6
		P				.045												.07
TEAM56	61	TAM/EXE	4,814	2,127	44.2
		EXE	4,852	1,812	37.3
		P		< .001
IES13,18	55.7	EXE	2,320	957	41.3	526	22.7	228	9.8e		50.6	225	9.7	416	17.9			97	4.2a
		TAM	2,338	903	38.6	522	22.3	205	8.8		56.9	248	10.6	363	15.5			51	2.2
		P		.07		.78		.21				.30		.03				< .001
N-SAS BC-038	42	TAM/ANA	347		36.3		26.5		20.5f						18.2
		TAM	349		44.7		25.5		18.1						16.0
ARNO 9533	30.1	ANA	445	29	6.5
		TAM	452	29	6.4
ABCSG-6a30	62.3	ANA	387	151	39.0g	41	10.6h					32	8.3					16	4.1a
		Placebo	469	105	22.4	20	4.3					11	2.3					12	2.6
		OR		2.44		2.82						3.97						1.73
		95% CI		1.80 to 3.31		1.62 to 4.90						1.97 to 7.9						0.81 to 3.70
		P		< .001		< .001						< .001						.159
NSABP B-3336	30	EXE	799				0.90i
		Placebo	799				0.50
		P				NR
MA.174 (Goss PE, et al: J Natl Cancer Inst 97:1262-1271, 2005)	30	LET	2,572	1,486	58	999	39			272	16j
		Placebo	2,577	1,383	54	998	39			249	15
		P		.003		.95				NR

Abbreviations: ATAC, Arimidex, Tamoxifen, Alone or in Combination (trial); ANA, anastrozole; TAM, tamoxifen; OR, odds ratio; BIG, Breast International Group; LET, letrozole; ABCSG, Austrian Breast and Colorectal Cancer Study Group; GOS, goserelin; ITA, Italian Tamoxifen Anastrozole (trial); TEAM, Tamoxifen Exemestane Adjuvant Multinational (trial); EXE, exemestane; IES, Intergroup Exemestane Study; N-SAS, National Surgical Adjuvant Study (Group); ARNO, Arimidex-Nolvadex (trial); NSABP, National Surgical Adjuvant Breast and Bowel Project; NR, not reported.

^aDiarrhea.

^bDepression or sleep disturbance.

^cNeurologic or psychiatric disorder.

^dGI complaints.

^eDepression.

^fMood fluctuations.

^gGrade 1 only.

^hGrades 1 to 4.

ⁱGrade 3 or 4 toxicity.

^jSubset analysis: N (LET) = 1,813; N (placebo) = 1,799.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.

Employment or Leadership Position: None Consultant or Advisory Role: Jennifer Malin, Pfizer (C); Eleftherios P. Mamounas, Novartis (C), Pfizer (C) Stock Ownership: None Honoraria: Karen E. Gelmon, Novartis, AstraZeneca, Pfizer; Eleftherios P. Mamounas, Pfizer, Novartis; Vered Stearns, AstraZeneca Research Funding: Vered Stearns, Novartis, Pfizer Expert Testimony: None Other Remuneration: None

AUTHOR CONTRIBUTIONS

Conception and design: Harold J. Burstein, Jennifer J. Griggs

Administrative support: Ann Alexis Prestrud, Jerome Seidenfeld, Mark R. Somerfield

Collection and assembly of data: Harold J. Burstein, Ann Alexis Prestrud, Jerome Seidenfeld, Jennifer J. Griggs

Data analysis and interpretation: Harold J. Burstein, Ann Alexis Prestrud, Jerome Seidenfeld, Holly Anderson, Thomas A. Buchholz, Nancy E. Davidson, Karen E. Gelmon, Sharon H. Giordano, Clifford A. Hudis, Jennifer Malin, Eleftherios P. Mamounas, Diana Rowden, Alexander J. Solky, MaryFran R. Sowers, Vered Stearns, Eric P. Winer, Mark R. Somerfield, Jennifer J. Griggs

Manuscript writing: Harold J. Burstein, Ann Alexis Prestrud, Jerome Seidenfeld, Holly Anderson, Thomas A. Buchholz, Nancy E. Davidson, Karen E. Gelmon, Sharon H. Giordano, Clifford A. Hudis, Jennifer Malin, Eleftherios P. Mamounas, Diana Rowden, Alexander J. Solky, MaryFran R. Sowers, Vered Stearns, Eric P. Winer, Mark R. Somerfield, Jennifer J. Griggs

Final approval of manuscript: Harold J. Burstein, Ann Alexis Prestrud, Jerome Seidenfeld, Holly Anderson, Thomas A. Buchholz, Nancy E. Davidson, Karen E. Gelmon, Sharon H. Giordano, Clifford A. Hudis, Jennifer Malin, Eleftherios P. Mamounas, Diana Rowden, Alexander J. Solky, MaryFran R. Sowers, Vered Stearns, Eric P. Winer, Mark R. Somerfield, Jennifer J. Griggs