Effect of an Aromatase Inhibitor on BMD and Bone Turnover Markers: 2-Year Results of the Anastrozole, Tamoxifen, Alone or in Combination (ATAC) Trial (18233230)


  • Richard Eastell MD,

    Corresponding author
    1. Bone Metabolism Group, Division of Clinical Sciences (North), University of Sheffield, Sheffield, United Kingdom
    • Division of Clinical Sciences (North), University of Sheffield Clinical Sciences Centre, Northern General Hospital, Herries Road, Sheffield, South Yorkshire S5 7AU, UK
    Search for more papers by this author
    • Dr Eastell is a member of the Steering Committee at ATAC and is a consultant to AstraZeneca. Drs Eastell and Hannon receive honoraria and funding from Astrazeneca. Dr Hannon also receives honoraria from Pfizer Inc. Dr Cuzick is a statistical consultant for AstraZeneca. Dr Dowsett holds consultancies with and receives funding from Novartis and AstraZeneca. Dr Clack holds stock ownership in and is employed by AstraZeneca. All other authors state that they do not have any conflicts of interest.

  • Rosemary A Hannon,

    1. Bone Metabolism Group, Division of Clinical Sciences (North), University of Sheffield, Sheffield, United Kingdom
    Search for more papers by this author
  • Jack Cuzick,

    1. Cancer Research UK, Department of Epidemiology, Mathematics and Statistics, Wolfson Institute of Preventive Medicine, Queen Mary College, London, United Kingdom
    Search for more papers by this author
  • Mitch Dowsett,

    1. The Royal Marsden Hospital, London, United Kingdom
    Search for more papers by this author
  • Glen Clack,

    1. Global Clinical Development Department, AstraZeneca, Cheshire, United Kingdom
    Search for more papers by this author
  • Judith E Adams

    1. Clinical Radiology, Imaging Science and Biomedical Engineering, University of Manchester, Manchester, United Kingdom
    Search for more papers by this author


Aromatase inhibitors reduce estrogen levels in postmenopausal women with breast cancer. Residual estrogen is an important determinant of bone turnover. Adjuvant anastrozole was associated with significant BMD loss and increased bone remodeling, whereas tamoxifen reduced bone marker levels.

Introduction: In the Anastrozole, Tamoxifen, Alone or in Combination (ATAC) trial after a median follow-up of 68 months, a significant improvement in disease-free survival was observed with anastrozole treatment (hazard ratio [HR], 0.87; 95% CI, 0.78–0.97; p = 0.01). Anastrozole was also associated with tolerability benefits compared with tamoxifen, but with higher fracture rates. The HR of anastrozole compared with tamoxifen after 60 months of treatment was 1.49 (95% CI, 1.25–1.77).

Materials and Methods: This prospectively designed subprotocol (n = 308) of ATAC assessed changes in BMD and bone turnover markers in postmenopausal women with invasive primary breast cancer receiving anastrozole 1 mg/day, tamoxifen 20 mg/day, or combination treatment with both agents for 5 years. Patients with osteoporosis were excluded (osteopenia permitted at the investigators discretion). Lumbar spine and total hip BMD was assessed at baseline and after 1 and 2 years; bone turnover markers (serum C-telopeptide, urinary N-telopeptide [NTX], free deoxypyridinoline, serum procollagen type-1 N-propeptide, bone alkaline phosphatase [ALP]) were assessed at baseline and after 3, 6, and 12 months. Results were expressed as median percentage change.

Results: After 2 years of anastrozole treatment, BMD was lost at lumbar spine (median 4.1% loss) and total hip (median 3.9% loss) sites; increases of 2.2% and 1.2%, respectively, were observed with tamoxifen. After 1 year of anastrozole treatment, increased bone remodeling was observed (NTX, +15%; 95% CI, 3–25%; bone ALP, +20%; 95% CI, 14–25%); decreased bone remodeling was observed with tamoxifen (NTX, −52%; 95% CI, −62% to −33%; bone ALP, −16%; 95% CI, −24% to −11%).

Conclusions: Anastrozole is associated with significant BMD loss and a small increase in bone turnover, whereas tamoxifen (and the combination) is associated with increased BMD and decreased remodeling. These data may explain the increased fracture risk observed with anastrozole treatment in the ATAC trial. The impact of anastrozole on bone should be weighed against its overall superior efficacy and tolerability as observed in the main ATAC trial.


Estrogen is a major determinant of bone turnover. In postmenopausal women, low estradiol levels are associated with increased bone turnover,(1–3) low BMD(4) and increased fracture risk.(4–6) In the first 5 years after menopause, bone loss occurs at a rate of ∼1.5% per year; slowing to 0.2–1.4% per year thereafter.(7,8)

Anastrozole is an orally active, highly selective, well-tolerated, nonsteroidal aromatase inhibitor (AI) that markedly suppresses estrogen levels in postmenopausal women.(9,10) Tamoxifen has partial estrogen-agonist activity, which results in a bone-sparing effect, but also increases the risk of endometrial cancer and thromboembolic events. Anastrozole is devoid of any such activity. Preliminary studies have shown that the third-generation nonsteroidal AI letrozole increases markers of bone resorption over a 6-month period.(11) The effect of anastrozole on markers of bone turnover and BMD are reported here for the first time.

There are data to suggest that tamoxifen may have differential effects on trabecular and cortical bone. In a preventive study in late-postmenopausal women, 2 years of tamoxifen treatment was associated with increased lumbar spine BMD (rich in trabecular bone); patients receiving placebo had a reduction in lumbar spine BMD.(12) Other groups have also reported the positive effect of tamoxifen on spine BMD(13–16); an effect associated with decreased bone resorption(12,14,17) and formation.(12,14,18,19) A protective effect against bone loss at the femoral neck has also been reported,(14,16,20) a site containing a mixture of trabecular and cortical bone. In contrast, tamoxifen does not significantly increase BMD at the radius, a site rich in cortical bone,(18,19,21) or total body BMD.(12)

The Anastrozole, Tamoxifen, Alone or in Combination (ATAC) trial is a randomized, double-blind trial evaluating the efficacy and tolerability of anastrozole 1 mg/day, tamoxifen 20 mg/day, or a combination of both agents as primary adjuvant therapy in 9366 postmenopausal women with early breast cancer who completed surgery.(22,23) The most recent results after a median follow-up of 68 months showed a significant improvement in disease-free survival with anastrozole treatment (hazard ratio [HR], 0.87; 95% CI, 0.78–0.97; p = 0.01).(24) Because of concerns about the effects of long-term estrogen suppression in postmenopausal women, the ATAC trial included a subprotocol designed to study the impact of these treatments on bone. Here, we report data from the first 2 years of the ATAC trial bone subprotocol.


The ATAC trial design has been described elsewhere.(22) Trial treatment is planned for 5 years. Patients gave their written informed consent to join the trial, and the appropriate local regulatory and ethics authority for each participating center approved the trial protocol before the enrollment of patients began.


At selected centers, postmenopausal women who were candidates to receive adjuvant endocrine treatment for invasive primary breast cancer (as part of the main ATAC trial) were recruited into the bone subprotocol. At centers using Hologic densitometers, the Hologic reference database was used to define osteoporosis and osteopenia for spine, whereas centers using Lunar densitometers used the Lunar USA reference database. The NHANES III reference database was used for hip for both types of densitometer. Patients with osteoporosis (T score < −2.5) were excluded from the primary analysis population (because treatment for osteoporosis was not allowed during the trial); those with osteopenia (T score between −1 and −2.5) were included at the investigators' discretion. A control group of postmenopausal women with early breast cancer, with good prognosis after primary surgery,(25,26) who were not to be treated with systemic adjuvant therapy, was also recruited.

Additional exclusion criteria were applied to patients participating in the bone subprotocol that were not applicable to those in the main ATAC trial. These were hormone replacement therapy or bisphosphonates within 12 months before randomization or during study treatment; bone fracture within 6 months before randomization; anticonvulsant or corticosteroid therapy; chronic renal/hepatic impairment; malabsorption syndrome; and endocrine disorders including hyperparathyroidism, untreated thyroid disease, Cushing's syndrome, and pituitary disease. Patients participating in this subprotocol and the main ATAC trial were not specifically counseled regarding osteoporosis prevention or routinely supplemented with vitamin D or calcium.


Change in BMD was the primary endpoint. Secondary endpoints included evaluation of the change in bone turnover markers.


BMD was measured at the lumbar spine (L1–L4) and total hip by DXA, using General Electric Lunar Corp (Madison, WI, USA) and Hologic (Bedford, MA, USA) densitometers. Measurements were made at baseline and 1 and 2 years and were analyzed centrally by Bio-Imaging Technologies (Newtown, PA, USA). Bio-Imaging Technologies undertook cross-calibration and quality assurance of each densitometer. For the analysis of bone markers, nonfasting urine samples were collected as a second morning void and nonfasting serum was taken between 8:00 a.m. and 10:00 a.m. Collections were made twice at baseline (1 week apart) and at 3, 6, and 12 months. Urine samples were stored at −20°C and serum samples at −70°C until analysis. All samples from a given individual were measured in the same analytical batch in Sheffield. Replicate control samples were measured in each analytical batch, and the data were used to calculate the CV for each assay.

The bone resorption marker urinary N-telopeptide of type-I collagen (NTX) was measured using automated chemiluminescent immunoassay (Ortho-Clinical Diagnostics, Rochester, NY, USA); interassay CV was 7%. The resorption marker urinary free immunoreactive deoxypyridinoline (iFDPD) was measured by ELISA (Pyrilinks D; Quidel, Mountain View, CA, USA); interassay CV was 11%. Urinary markers were expressed as a ratio to urinary creatinine, measured by an automated dry slide method (Ortho-Clinical Diagnostics, Rochester, NY, USA). The resorption marker serum C-telopeptide of type-I collagen (CTX) and the formation marker procollagen type-I N-propeptide (PINP) were measured using electrochemiluminescence on an Elecsys 2010 autoanalyzer (Roche Diagnostics, Penzberg, Germany). The interassay CV for CTX was 7% and for PINP was 6.2%. The formation marker serum bone-specific alkaline phosphatase (bone ALP) was measured by ELISA (Alkphase B; Quidel); interassay CV was 11%. Estradiol levels were assessed from a blood sample taken at baseline as previously described(27); interassay CV was 10% (at a mean level of 26 pM).

Statistical analysis

Sixty-seven patients per arm would be required to detect a difference of 2.5% in percentage change of BMD between the anastrozole and tamoxifen groups with 90% power at the 5% (two-sided) significance level. Because it was expected that about 14% of patients would recur by the end of the 2-year follow-up and a further 10% would dropout or have incorrect baseline scans, it was recommended that 256 patients (86 per arm) be recruited.

All patients who had a valid postbaseline measurement, were recurrence-free, and were receiving trial therapy at the time of measurement were included in the analysis population. There were two prespecified comparisons of interest: (1) anastrozole compared with tamoxifen and (2) combination treatment (anastrozole plus tamoxifen) compared with tamoxifen.

Lumbar spine BMD and total hip BMD were considered separately, as were data at 1 and 2 years. Lumbar spine and total hip BMD values were log-transformed. An analysis of covariance was used to examine the change in BMD between the anastrozole alone and the tamoxifen alone groups and between the anastrozole plus tamoxifen and tamoxifen alone groups. The model was fitted with baseline (log)BMD (lumbar spine or total hip as applicable), time since last menstrual period (<1, 1–4, and >4 years), baseline (log)estradiol, and body mass index (BMI) as covariates. Where either BMI (eight patients) or baseline estradiol (eight patients) data were missing, the mean of the rest of the data were substituted for the missing values. The estimated median percentage changes from baseline and associated 95% CIs are presented. Regression analyses were undertaken to identify important cofactors, adjust for them in assessing treatment effects and explore interactions with treatment; this used a standard “step-down” procedure. All variables in the full model were entered and the least significant variable was dropped if p < 0.05. This method was continued until no further variables needed to be removed. Bone markers were log-transformed and analyzed using identical methods, with the appropriate baseline covariates.



A total of 308 postmenopausal women participating in the ATAC trial were recruited into the bone subprotocol, along with a nonrandomized control group of 46 women (Fig. 1). Of those randomized in the main study, 247 were eligible to take part in the bone subprotocol, which was in line with the power calculations. In addition, 39 women recruited into the control group were eligible (Fig. 1).

Figure Figure 1.

Flow diagram showing the number of patients in each intervention group included in the primary analysis population.

Baseline patient demographics and bone characteristics are shown in Tables 1 and 2; the three treatment groups and the nonrandomized controls were well matched in terms of bone markers, years of treatment (excluding the control patients), and percentage osteopenic, along with most other parameters. At baseline, ∼45% of the women had osteopenia of the lumbar spine and 35% had osteopenia of the hip. Tumor characteristics were well balanced between the three groups and nonrandomized controls and were similar to those seen in the main trial.(22) However, more patients in the anastrozole group (p = 0.002 versus tamoxifen) were within 1 year of menopause; the other two groups and the nonrandomized controls were well matched in this respect.

Table Table 1.. Baseline Patient Characteristics
original image
Table Table 2.. Baseline Levels of Estradiol and Bone Biomarkers
original image

BMD results at 1 and 2 years

The unadjusted median percentage changes from baseline for lumbar spine and total hip BMD at 1 and 2 years are shown in Fig. 2. At year 1, BMD data were available for 71 patients in the anastrozole group, 69 in the tamoxifen group, 64 in the combination group, and 39 in the control group. At year 2, BMD data were available for 58 patients in the anastrozole group, 64 in the tamoxifen group, 51 in the combination group, and 32 in the control group. Patients without a 2-year DXA scan generally had lower BMD values at year 1 than patients with both scans, partially because of the need to withdraw some patients on safety grounds. The baseline characteristics of those patients with 2-year data were similar to those without 2-year data. Three patients receiving anastrozole and two receiving tamoxifen were withdrawn because of osteoporosis/osteopenia during the first 2 years of treatment.

Figure Figure 2.

Unadjusted median percentage change in (A) lumbar spine and (B) total hip BMD after 1- and 2-year treatment. Bars represent 95% CIs.

The BMD difference between anastrozole and tamoxifen was significant at both sites (p < 0.001) at years 1 and 2. No significant differences were observed between the combination (anastrozole plus tamoxifen) and tamoxifen groups in lumbar spine BMD at years 1 and 2 (p= 0.12 and p= 0.34, respectively) or total hip BMD at years 1 and 2 (p= 0.87 and p = 0.97, respectively). For the control group, only small changes were observed at both sites.

For anastrozole, there were significant (all p < 0.001) losses of BMD at the lumbar spine and hip, both at year 1 (2.6% and 1.7%, respectively) and year 2 (4.0% and 3.2%, respectively), whereas for tamoxifen, there were significant (p < 0.05) gains (1.2% and 0.8%, respectively, at year 1; 1.9% and 1.2%, respectively, at year 2).

Multivariate analyses showed an inverse correlation between baseline log(estradiol) level and BMD changes, with lower baseline estradiol predictive of greater BMD losses. A similar, significant association was observed between baseline BMD and BMD changes throughout the study. This could be partly caused by regression to the mean, although it is reassuring that patients with lower initial values did not have larger BMD changes. No other variable retained significance after these covariates were included in the model. The impact of anastrozole treatment was little affected by adjustment for these variables (mean difference from tamoxifen being 5.0% [lumbar spine] and 4.1% [total hip] versus 5.9% and 4.4%, respectively, for unadjusted values; Table 3). A significant correlation was seen between baseline BMD and baseline estradiol levels (Spearman's rank correlation coefficient ρ = 0.187 and p = 0.002 for lumbar spine, and ρ = 0.229 and p = 0.0001 for total hip).

Table Table 3.. Results From Univariate and Multiple Linear Regressions of Percentage Change in BMD From Baseline on a Range of Predictive Factors (Coefficients Give Percentage Change in BMD per Unit Change in Variable)
original image

Overall, changes in BMD at both total hip and lumbar spine sites were greatest in patients within 4 years of menopause (Table 4). Compared with patients receiving the same treatment who were >4 years postmenopause, greater BMD losses were observed in anastrozole-treated patients experiencing the menopause in the last 4 years and greater BMD increases were seen in the comparable tamoxifen group. There was no clear or significant relationship between estrogen levels and time since menopause (median values: 18.0 and 20.0 pM for women ⩽4 and >4 years after menopause).

Table Table 4.. Unadjusted Lumbar Spine and Total Hip BMD Changes (From Baseline) in Women ≤4 years and >4 Years Since Menopause
original image

Bone turnover markers

Bone marker data were available at baseline and year 1 on all five markers for 58, 56, 54, and 25 women in the anastrozole, tamoxifen, combination, and control groups, respectively. The unadjusted median percentage changes from baseline up to 1 year for CTX, NTX, iFDPD, PINP, and bone ALP are shown in Figs. 3 and 4. Overall, patients receiving anastrozole tended to have increases in bone turnover markers, whereas those receiving tamoxifen tended to have decreases. At year 1, the resorption markers CTX (+26%) and NTX (+15%; 95% CI, 3–25%) had increased in the anastrozole group, as had the formation markers PINP (+18%) and bone ALP (+20%; 95% CI, 14–25%). The resorption marker iFDPD (0%, −15% to +8%) remained unchanged. Treatment with tamoxifen was associated with decreased bone resorption (CTX: −56%; NTX: −52%; 95% CI, −62% to −33%; iFDPD: −42%; 95% CI, −49–36%), and formation (PINP: −72%; bone ALP: −16%; 95% CI, −24% to −11%). The combination of anastrozole plus tamoxifen resulted in similar changes in bone markers as tamoxifen alone. Overall, patients in the nonrandomized control group had smaller changes in bone turnover markers at year 1 (CTX: −22%; NTX: −13%; 95% CI, −18% to +3%; iFDPD: −9%, −28% to −1%; bone ALP: +2%; 95% CI, −5–9%) than patients in the three treatment groups. However, this group did experience a 36% decrease in the bone formation marker PINP at year 1 (p = 0.001). These changes in bone turnover markers, particularly CTX, were observed even though the samples were taken in the nonfasting state.

Figure Figure 3.

Unadjusted median percentage change in the bone resorption markers (A) NTX, (B) CTX, and (C) iFDPD after 3-, 6-, and 12-month treatment. Bars represent 95% CIs.

Figure Figure 4.

Unadjusted median percentage change in the bone formation markers (A) PINP and (B) bone ALP after 3-, 6-, and 12-month treatment. Bars represent 95% CIs.


This study showed that 2 years of anastrozole treatment was associated with bone loss at the spine and hip and an increase in bone turnover. This indicates that the residual estrogen levels in the postmenopausal woman are important for the regulation of bone turnover. The effect of anastrozole was most marked in the first 4 years after menopause; this period is associated with accelerated bone loss. Estrogen deficiency bone loss can be prevented by antiresorptive agents such as bisphosphonates, and trials of such agents are underway in cancer treatment–induced bone loss.

Tamoxifen was associated with preservation of BMD at 2 years and a decrease in bone turnover. When combined with anastrozole, it prevented bone loss. The two agents are acting by differing mechanisms—tamoxifen is acting as a partial estrogen agonist on bone and anastrozole is reducing the residual level of estrogen. Unfortunately, in the larger ATAC trial, the combination arm was less effective in improving disease-free survival that anastrozole alone, so this is not a reasonable approach to preventing bone loss.

The ATAC trial did not include a pure placebo group, so we enrolled a group of women with breast cancer after surgery who had a good prognosis and did not require endocrine therapy.(25,26) This allowed us to have a control group and to observe that some markers are affected by the surgery—PINP decreased in this group and that was probably a result of wound healing after surgery.

The third-generation AIs, anastrozole,(22–24) letrozole,(28,29) and exemestane,(30) have all been associated with an increased risk of osteoporosis and/or fractures in the large adjuvant breast cancer trials. These trials are somewhat difficult to interpret because vertebral fracture events were not captured by systematic radiographs of the spine, loose definitions of “osteoporosis” were used, and the control (or prior treatment) was itself active on bone, namely tamoxifen. Tamoxifen preserves BMD and may(31) or may not(32) reduce the risk of fractures.

Studies on mechanism have indicated that letrozole and exemestane are also associated with higher levels of bone resorption markers(11,33,34) and faster rates of bone loss(34) than controls (although in the exemestane study, the control group had unexplained accelerated bone loss).

In summary, adjuvant anastrozole treatment results in increased bone turnover and increased bone loss, whereas tamoxifen (alone or in combination with anastrozole) results in decreased bone turnover and decreased bone loss. A 5-year assessment of BMD is planned, and these data are required to define long-term changes in BMD. Other factors (including age, family history, smoking, and concomitant use of certain drugs, such as corticosteroids) should also be taken into account when assessing the risk of fracture. Until further data are available, patients considered to be at risk should be identified, managed, and monitored according to local practice, making use of the American Society for Clinical Oncology guidelines.(35)


The authors thank the patients who are participating in the ATAC trial and all listed below. Members of the Writing Group for this paper are asterisked. This study was conducted under the auspices of the ATAC trial steering committee, who were responsible for the study design, interpretation of the data, and preparation of the manuscript. The sponsor, AstraZeneca, provided support for the conduct of the study, data collection, and project management. The data analysis was performed under the direction of the independent statistician (J Cuzick). R Eastell wrote the manuscript and reviewed changes suggested by the ATAC trial steering committee.


Prof M Baum (Chairman and Principal Investigator for the main ATAC trial), University College London, London, UK; Prof AR Bianco, Universita Degli Studi Di Napoli Federico II, Napoli, Italy; Dr A Buzdar, The University of Texas, M.D. Anderson Cancer Centre, Houston, TX, USA; Dr M Coibion, Institut Bordet, Bruxelles, Belgium; Prof Robert Coleman, Cancer Research Centre, Weston Park Hospital, Sheffield, UK; Dr M Constenla, Hospital Montecelo, Pontevedra, Spain; *Prof J Cuzick (independent statistician), Cancer Research UK, London, UK; Prof Dr W Distler, Universitdtsklinikum Carl Gustav Carus Dresden, Dresden, Germany; Prof M Dowsett, The Royal Marsden Hospital, London, UK; Prof J Forbes, Newcastle Mater Misericordiae Hospital, New South Wales, Australia; Prof WD George, Beatson Oncology Centre, Western Infirmary, Glasgow, UK; J Gray, Belfast City Hospital, Belfast, UK; Dr JP Guastalla, Centre Leon Berard, Lyon, France; J Houghton, Dr N Williams, Clinical Trials Group of the Department of Surgery, UCL, London, UK, *Prof A Howell, Christie Hospital, Manchester, UK; Prof Dr JGM Klijn, Dr Daniel den Hoed Kliniek, University Hospital Rotterdam, Rotterdam, The Netherlands; *Dr GY Locker, Evanston Hospital, Kellogg Cancer Care Center, Evanston IL, USA; Dr John Mackey, Cross Cancer Institute, Edmonton, Canada; Prof RE Mansel, University of Wales College of Medicine, Cardiff, UK; Dr JM Nabholtz, University of California at Los Angeles, Los Angeles, CA, USA; Dr T Nagykalnai, Uzsoki U Hospital, Budapest, Hungary; Dr A Nicolucci, GIVIO Co-ordinating Centre, Consorzio Mario Negri Sud, Centro Di Ricerchi Farmacologichi, E Biomedichi, Chieta, Italy; Dr U Nylen, Radiumhemmet, Karolinska Sjukhuset, Stockholm, Sweden; Dr T Sahmoud, R Hellmund, AstraZeneca Pharmaceuticals, Wilmington, DE, USA; R Sainsbury, Royal Free and University College Medical School, London, UK; Dr N Griffiths, Dr G Hoctin-Boes, AstraZeneca Pharmaceuticals, Macclesfield, UK; Dr JS Tobias, The Meyerstein Institute of Clinical Oncology, London, UK.


Prof Apffelstaedt, Tyerberg Hospital, Cape Town, South Africa; Dr WW Bate, Mercy Cancer Centre, Mercy Medical Centre, Mason City, IA, USA; Prof Dr Med M Beckmann, Universitat Erlangen-Nurnberg, Erlangen, Germany; Dr MJ Burnell, Atlantic Health Science Group, Saint John, New Brunswick, Canada; Prof A Buzdar, The University of Texas, M.D. Anderson Cancer Center Breast Oncology Clinic Station, Houston, TX, USA; Dr PD Byeff, University of Connecticut, John Dempsey Hospital, Farmington, CT, USA; Dr S Cawthorn, Frenchay Healthcare NHS Trust, Frenchay Hospital, Bristol, UK; Prof R Coleman, Weston Park Hospital, Sheffield, UK; Dr R Coquard, Clinique St Jean, Lyon, France; Dr AC de Boer, Ijsselland Ziekenhuis, Ijssel, Holland; Dr AKR al Debbagh, Trafford General Hospital, Trafford, UK; Dr MA Deutsch, Raleigh Internal Medicine and Wake Haematology/Oncology Clinic, Raleigh, NC, USA; Dr C Dijkhuis, Oosterschelde Ziekenhuis, Goes, Holland; Dr R Fernstad, St Goran's Hospital, Stockholm, Sweden; Dr JP Guestalla, Centre Leon Berard, Lyon, France; Dr D Halkema, Albert Schweitzer Ziekenhuis, Dordrecht, Holland; B Harrison, Northern General Hospital, Sheffield, UK; Dr S Holmberg, SU/Molndal Hospital, Molndal, Sweden; Dr IA Jaiyesimi, Wayne State University School of Medicine, Royal Oak, MI, USA; Dr EP Lester, Oncology Care Associates, PLLC, St Joseph, MI, USA; Dr G Locker, Kellogg Cancer Care Center, Evanston, IL, USA; Dr PA Lyss, Missouri Baptist Cancer Center, St Louis, MO, USA; Dr JR Mackey, Cross Cancer Institute, Edmonton, Alberta, Canada; Prof R Mansel, University Hospital of Wales NHS Trust, Cardiff Breast Unit, Cardiff, UK; Dr U Nylén, Karolinska Sjukhuset, Stockholm, Sweden; Dr P Paterson, Royal Cornwall Hospitals NHS Trust, Treliske Hospital, Truro, UK; Dr KB Pendergrass, Kansas City Oncology and Hematology Group, Lenexa, KS, USA; Dr JG Posada, Division of Haematology/Oncology Scott and White Memorial Hospital, Temple, TX, USA; Dr O Rixe, Hopital Clinique Claude Bernard, Metz, France; Dr J Robert, CHAUQ-Hospital du St-Sacrement, Quebec City, Quebec, Canada; Prof JFR Robertson, Nottingham City Hospital NHS Trust, Nottingham, UK; Dr S Rotstein, Onkologiska Kliniken, Danderyd, Sweden; Dr A Sami, Saskatoon Cancer Centre, University of Saskatchewan Campus, Saskatoon, Saskatchewan, Canada; Dr JI Spector, Berkshire Hematology Oncology, PC, Pittsfield, MA, USA; Dr L Strobbe, Nijmeegs Interconfessioneel, Ziekenhuis Canisius Wilhelmina, Nijmgen, Holland; Dr N Tirumali, Kaiser Permanente/Oncology Research, Portland, OR, USA; Dr A Wardley, Christie Hospital NHS Trust, Manchester, UK.


International project team

E Foster, SCTN Central Office, Information and Statistics Division, Edinburgh, UK; N Griffiths, A Doe, F Sapunar AstraZeneca Pharmaceuticals, Macclesfield, UK; J Houghton, N Williams, Clinical Trials Group of the Department of Surgery, UCL, London, UK; A Nicolucci, Mario Negri Institute, Chieta, Italy; S Pollard, Northern Yorkshire Clinical Trials Research Unit, Leeds, UK.

Independent data monitoring committee

Dr M Buyse, International Institute for Drug Development (ID squared), Brussels, Belgium; Dr R Margolese, McGill University, The Sir Mortimer B Davis Jewish General Hospital, Montreal, Quebec, Canada; Prof J J Body, Institute J Bordet, Bruxelles, Belgium.

Collaborative/operational groups

JF Forbes (group coordinator), JK Wakeham (study coordinator): Australian New Zealand Breast Cancer Trials Group Operations Office, Waratah, Australia; S de Placido (study coordinator), C Carlomagno (study coordinator): Universita Degli Studi Di Napoli Federico II, Naples, Italy; A Nicolucci (group coordinator), M Belfiglio (study coordinator), M Valentini (study coordinator): GIVIO Group, Consorzio Mario Negri Sud, Abruzzi, Italy; E Foster (trial coordinator): Scottish Cancer Therapy Network (SCTN), Information & Statistics Division, Edinburgh, UK; C Lacey (trial monitor): North West Breast Group, Burnley, Lancashire, UK; S Pollard (trial coordinator): Northern & Yorkshire Clinical Trials Research Unit (NCTRU), University of Leeds, Leeds, UK. J Houghton (senior lecturer in clinical trials), N Williams (trial coordinator): Clinical Trials Group of the Department of Surgery, UCL, London, UK. Changes (From Baseline) in Women ≤4 years and >4 Years Since Menopause