SEARCH

SEARCH BY CITATION

Keywords:

  • bone mineral density;
  • breast cancer;
  • menopause;
  • oophorectomy;
  • tamoxifen;
  • femoral neck;
  • spine;
  • adjuvant drug therapy

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. CONFLICT OF INTEREST DISCLOSURE
  9. REFERENCES

BACKGROUND

In premenopausal women treated for breast cancer, loss of bone mineral density (BMD) follows from menopause induced by chemotherapy or loss of ovarian function biochemically or by surgical oophorectomy. The impact on BMD of surgical oophorectomy plus tamoxifen therapy has not been described.

METHODS

In 270 Filipino and Vietnamese premenopausal patients participating in a clinical trial assessing the impact of the timing in the menstrual cycle of adjuvant surgical oophorectomy on breast cancer outcomes, BMD was measured at the lumbar spine and femoral neck before this treatment, and at 6, 12, and 24 months after surgical and tamoxifen therapies.

RESULTS

In women with a pretreatment BMD assessment and at least 1 other subsequent BMD assessment, no significant change in femoral neck BMD was observed over the 2-year period (−0.006 g/cm2, −0.8%, P = .19), whereas in the lumbar spine, BMD fell by 0.045 g/cm2 (4.7%) in the first 12 months (P < .0001) and then began to stabilize.

CONCLUSIONS

Surgically induced menopause with tamoxifen treatment is associated with loss of BMD at a rate that lessens over 2 years in the lumbar spine and no significant change of BMD in the femoral neck. Cancer 2013;119:3746–3752. © 2013 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. CONFLICT OF INTEREST DISCLOSURE
  9. REFERENCES

In high-income countries, premenopausal women with hormone receptor–positive tumors account for 15% of the case burden in breast cancer; however, in low- and middle-income countries this percentage is approximately 36%.[1] Globally, one-third of the case burden in breast cancer is in this group. With increasingly effective therapies and the associated increases in long-term survival, the total “costs” of therapies are of concern to afflicted women.

Adjuvant ovarian ablation improves survival in premenopausal women with operable breast cancer unselected for hormonal receptor status.[2] Meta-analysis data also suggest a trend favoring ovarian ablation over luteinizing hormone–releasing hormone (LHRH) agonist treatment.[2] Adjuvant surgical oophorectomy plus tamoxifen improves survival with risk reduction of 0.54 in hormone receptor positive patients.[3] A variety of direct but underpowered studies and indirect risk reduction data suggest that this combined hormonal adjuvant treatment may modestly improve results over oophorectomy or tamoxifen alone.[4-8] The direct adjuvant comparison of tamoxifen with or without ovarian ablation or ovarian function suppression is being tested in the SOFT trial, in which, however, the fraction of women treated with surgical oophorectomy may be too small to allow conclusions about this specific combined hormonal treatment. In this context, in high-income countries, many premenopausal women with hormone receptor–positive tumors are treated with chemotherapy and tamoxifen, because of the conclusion that current chemotherapy regimens, particularly with taxanes, provide greater benefits than any hormonal therapies.[9] In low- and middle-income countries, however, where alkylating agent–anthracycline combinations are the standard, surgical oophorectomy plus tamoxifen and chemotherapy plus LHRH plus tamoxifen provide comparable breast cancer outcome benefits.[3, 5]

Losses of BMD are important because gains are difficult to achieve, and depending on the maximal levels of BMD women achieve and the rates of loss, losses can rapidly lead to absolute BMD levels associated with markedly greater fracture risk.[10] Loss of ovarian function is associated with loss of BMD. Because this occurs gradually, annual rates of loss of BMD vary “at menopause” but absolute loss of 5% to 10% of BMD in 5 years after menopause is common in normal women.[11] Tamoxifen alone preserves BMD in postmenopausal women[12-14] ; however, in premenopausal women treated with tamoxifen alone, Powles et al found loss of BMD (2.88% over 2 years); other data support this conclusion.[13, 15] Studies evaluating BMD changes associated with LHRH agonist plus tamoxifen and chemotherapy plus tamoxifen treatments have all demonstrated losses in BMD.[15-20]

In these contexts, we report here findings from longitudinal assessments of BMD at the lumbar spine (L1-L4) and femoral neck in 270 premenopausal Filipino and Vietnamese women with hormone receptor–positive breast cancer participating in a clinical trial assessing the impact of the timing in the menstrual cycle of surgical oophorectomy on breast cancer outcomes. All women in the trial also underwent treatment with tamoxifen.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. CONFLICT OF INTEREST DISCLOSURE
  9. REFERENCES

This report is of an ancillary study to a randomized phase 3 clinical trial investigating the impact of the timing of adjuvant surgical oophorectomy. Women with core biopsy hormone receptor–positive invasive breast cancer were stratified at registration according to likelihood of not being (stratum 1) or being (stratum 2) in the luteal phase of the menstrual cycle during the next 1 to 6 days. Patients in stratum 1 were randomized to (delayed) surgical oophorectomy during the estimated mid-luteal phase of the menstrual cycle or to immediate surgical oophorectomy. Patients in stratum 2 were assigned immediate surgical oophorectomy. All patients underwent surgical oophorectomy and mastectomy surgeries in that sequence under the same anesthesia on the same day according to the randomized or direct day-of-surgery assignments. All patients began on tamoxifen within 6 days of their surgeries. Patient written informed consent that included monitoring with BMD assessments was required. The study was approved at individual participating institutions in the Philippines and Vietnam, and by supervising institutional review boards for these institutions and at the lead investigators' US institutions. All of the participating institutions and institutional review boards were registered with the Office for Human Research Protections in the United States. A total of 740 patients were enrolled in the parent trial. Eligibility criteria for the parent study included age ≥ 18 years, ≤ 50 years; history of menstrual cycles for the last 3 months of ≥ 25 to ≤ 35 days and last menstrual period < 35 days ago; not taking oral contraceptives; pathologic histologic diagnosis of invasive and hormone receptor–positive (estrogen receptor–positive and or progesterone receptor–positive by immunohistochemistry) breast cancer; clinical stage II-IIIB; physical examination including gynecological examination unrevealing for any suggestion of serious illness or metastatic breast cancer; negative chest X-ray; and negative urine pregnancy test. According to the stratification and randomization schema of the parent study, of the ancillary study patients, 175 stratum 1 patients were randomized to either immediate surgical oophorectomy or to having this procedure delayed until the estimated next mid-luteal phase of their menstrual cycles; the remaining 95 stratum 2 patients underwent, by direct assignment, immediate surgical oophorectomy. This distribution of randomized or assigned treatment times was similar to the full set of patients who entered the parent trial. All patients in the parent trial were begun on tamoxifen 20 mg daily within 6 days of oophorectomy surgery, to be continued for 5 years (tamoxifen [Nolvadex] was provided by AstraZeneca Pharmaceutical, London, UK). Patients returned at 3- to 4-month intervals for tamoxifen medication and symptom and physical examination reevaluations. Because of this requirement, compliance with tamoxifen administration is thought to be high.

Statistical Analyses

A nonlinear mixed effects regression model was used to assess the change in BMD over the study period. The form was adapted from a simple exponential trend described in Stevens,[21] which allows the outcome to approach an asymptote as time increases. Specifically,

  • display math

where yijk represents the BMD measurement for kth observation on the jth subject from ith site, x represents the time from enrollment, and e the residual error. The model includes 3 fixed-effect parameters: the site-specific asymptote (αi), the change from baseline to the asymptote (β), and a rate parameter (γ) that determines how quickly the asymptote is approached over time, and a random intercept (uij) assumed to be normally distributed with mean zero and variance σu2. The random intercept adequately modeled the dependency among observations obtained from each woman. Other random coefficients (eg, rate of decline) were also tested. In addition, a site-specific asymptote parameter was included to account for absolute BMD differences at both the lumbar spine and femoral neck sites found between Vietnamese and Filipino patient populations. The difference was surmised to be at least partially due to different equipment and software analytic programs. The relationships between age, weight, and body mass index (BMI) and BMD were also explored

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. CONFLICT OF INTEREST DISCLOSURE
  9. REFERENCES

The late date of initiation of this ancillary study, site nonparticipation (because of unavailability of BMD assessment equipment), BMD measurements for individual patients done at different sites with different equipment, equipment breakdown, absence of baseline BMD assessment, or other logistical issues all unrelated to individual patient bone health led to creation of a study sample with baseline BMD assessments (taken within 1 month of randomization on study) in 332 patients. Of these, 270 patients subsequently had at least one BMD measurement taken within 30 months of randomization using the same equipment.

Table 1 presents the characteristics of the 270 studied patients; Table 2 gives the details of the number and time period of follow-up visits completed by the patients.

Table 1. Patient Characteristics at Baseline (n = 270)
VariableValue
  1. Abbreviations: BMI, body mass index; SD, standard deviation.

Age (y), mean (SD)42.7 (4.3)
Weight (kg), mean (SD)54.3 (11.3)
BMI, mean (SD)23.1 (3.7)
Pathological stage I-II, %64.6%
Pathological stage III, %35.5%
Table 2. Distribution of Patients by the Total Number of Follow-Up Measurements (Subsequent to Baseline) and Length of the Follow-Up
Total No. of Follow-Up MeasurementsLumbar Spine (n = 232)Femoral Neck (n = 244)
178 (33.6%)84 (34.4%)
290 (38.8%)83 (34.0%)
364 (27.6%)77 (31.6%)
Follow-Up Completed 
6-month228 (98.3%)239 (98.0%)
12-month191 (82.3%)194 (79.5%)
24-month85 (36.6%)89 (36.5%)

Figure 1 shows the lumbar spine BMD traces for 30 randomly selected patients. A decrease in BMD was observed following the intervention, with the rate of the decline lessening over time (Fig. 2). The nonlinear model appeared to fit the data well and all parameters were significant (Table 3). The site-specific asymptote parameters (αV and αF) in the model indicated Vietnamese patients had 0.214 g/cm2 higher lumbar spine measurements on average than Filipino patients (P value for difference < .0001). The estimated within-patient variance (σe2) was 0.0018 and the between-patient (intercept) variance (σu2) was 0.0116, which means that 85% of the total variation in the data was due to variance among patients in level of BMD. An additional random effect for the rate parameter was added to the model, but was not significant, indicating little evidence for a varying rate of decline among patients (ie, only variance for the location of the asymptote was significant). An additional parameter allowing for a site-specific rate parameter was not significant (P = .16), indicating no evidence for a difference in the rate by site. Table 4 presents the model-derived estimates for change in BMD at the 6-, 12-, and 24-month time points. As shown in Figure 2, BMD decreases most rapidly in the first 12 months after baseline/treatment initiation.

image

Figure 1. Lumbar spine bone mineral density traces are shown for 30 randomly selected patients.

Download figure to PowerPoint

image

Figure 2. Lumbar spine bone mineral density measurements and nonlinear model fits are shown. The solid circles and solid line (open circles and dashed line) represent the data and the nonlinear model fit for Vietnamese (Filipino) patients, respectively.

Download figure to PowerPoint

Table 3. Estimated Model Coefficients for Lumbar Bone Mineral Density Fit
ParameterEstimateStandard ErrorP
  1. αV and αF represent the asymptote parameters for the Vietnamese and Filipino patients, respectively (P for difference <.0001).

αV1.01860.01841<.0001
αF0.80460.01704<.0001
β0.078800.01389<.0001
γ−0.070210.02350.0031
σu20.011580.001135<.0001
σe20.0018020.000120<.0001
Table 4. Model Estimated Absolute Change From Baseline (Standard Errors) and Percent Change (95% CI) at 6, 12, and 24 Months After Randomization for Lumbar (All P Values <.0001) and Femoral Neck (all P values > 0.19) BMD
Follow-Up MeasurementLumbar BMD ChangeFemoral BMD Change
Absolute (SE)Percent (95% CI)Absolute (SE)Percent (95% CI)
  1. Abbreviations: BMD, bone mineral density; CI, confidence interval, SE, standard error.

6-month−0.027 (0.004)−2.8% (−3.5, −2.1)−0.002 (0.001)−0.2% (−0.5, 0.1)
12-month−0.045 (0.004)−4.7% (−5.5, −3.8)−0.003 (0.002)−0.4% (−1.0, 0.2)
24-month−0.064 (0.005)−6.7% (−7.7, −5.6)−0.006 (0.005)−0.8% (−1.9, 0.4)

For femoral neck BMD, there was no clear change in BMD following the intervention (Fig. 3). Estimates from the nonlinear model were somewhat unstable due to the flat nature of the relationship between BMD and time, but consistently demonstrated a lack of significance for the change and rate parameters of the model. The difference in the site-specific asymptote parameters was significant (P < .0001) and indicated 0.208 g/cm2 higher femoral neck measurements for Vietnamese patients, on average. The estimated within-patient variance was 0.0015 compared with 0.0079 for the between-patient (asymptote) variance, again roughly 85% of the total. A fixed effect for site for the rate parameter was not significant (P = .22), indicating no evidence for a difference in the rate by site. The estimated changes from baseline presented in Table 4 are small, and the 95% confidence intervals are narrow, indicating no evidence for a clinically meaningful change over the study time period. A simple linear mixed model with site and time as the fixed effects gave very similar results.

image

Figure 3. Femoral neck bone mineral density measurements and nonlinear model fits are shown. The solid circles and solid line (open circles and dashed line) represent the data and the nonlinear model fit for Vietnamese (Filipino) patients, respectively.

Download figure to PowerPoint

Changes in BMD by patient age, weight, and BMI were explored. No impact was observed on the rate of decline by age (< 43 years versus ≥ 43 years) for either anatomic site, but patients aged 43 years or older exhibited 0.030 g/cm2 lower baseline femoral neck BMD on average than those under 43 years. No difference by age group was observed in baseline lumbar BMD. No changes in the rates of decline in BMD were observed by weight or BMI. However, clear positive relationships between weight and BMI and baseline BMD were observed. The BMI correlations with BMD were similar in Vietnamese and Filipino patients both for lumbar (r = 0.32, P = .0029 and r = 0.38, P < .0001, respectively) and femoral neck (r = 0.33, P = .0011 and r = 0.32, P < .0001, respectively).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. CONFLICT OF INTEREST DISCLOSURE
  9. REFERENCES

The results presented here are from the largest study reported to date of BMD changes over time after adjuvant therapy for premenopausal breast cancer. We found compelling evidence in longitudinal data in 270 patients that surgical oophorectomy followed by immediate tamoxifen treatment is associated with loss of BMD at the lumbar spine, the rate of which declines over 2 years and no significant change in BMD in the femoral neck. Sverrisdottir et al reported on 14 patients treated with LHRH plus tamoxifen 40 mg and 13 patients treated with LHRH alone. At 2 years, in the patients treated with LHRH plus tamoxifen, total body BMD had decreased by 1.4% (0.015 g/cm2) and 5.0% in the patients treated with LHRH alone. In this study, total body BMD was correlated with lumbar spine BMD (r = 0.8).[15] Gnant et al studied 82 patients, who were treated with LHRH plus tamoxifen, 26 of whom also had 3-year BMD assessments; baseline lumbosacral spine BMD was 1.028 g/cm2 at baseline and at 36 months had decreased by 0.072 g/cm2, of 9.0% observed or estimated 7.2% (from a model considering missing data); trochanter BMD was 0.715 g/cm2 at baseline with losses of 5.1% observed and estimated of 2.9%; In both sites, the BMD loss stabilized after 24 months.[16] Similar BMD losses have been consistently reported in premenopausal populations treated with chemotherapy and tamoxifen. Shapiro et al studied 35 women, 9 of whom received tamoxifen; lumbar spine BMD, which was 0.999 g/cm2 before treatment, fell 7.7% in 1 year; femoral neck BMD, which was 0.782 g/cm2 before treatment, fell 4.6%.[17] Powles et al, evaluating 118 patients, most also treated with tamoxifen and all with ovarian failure, found lumbar spine BMD losses of 4.0% at 1 year.[18] Saarto et al found BMD losses of 6.8% at the lumbar spine and 1.9% in the femoral neck at 1 year in 22 patients.[19] Finally, Delmas et al reported a lumbar spine BMD loss of 2.7% at 1 year.[20]

In summary, in studies involving mostly small numbers of patients, treatments with LHRH and chemotherapy with or without tamoxifen have demonstrated loss of BMD in all treatment groups; the losses appear to be of lower magnitude at the femoral region than in the lumbar spine, there are trends for the initial losses to stabilize after 1 to 3 years, and tamoxifen lessens the BMD losses. The results of the current large study in 270 women, with considerably less missing data than has been reported in the noted studies (16 for example), are different than all of these reports in demonstrating no significant change in BMD at the femoral neck. Multiple complex hormonal-bone interactions may explain these study result differences. Because Powles et al found increases of 1.17% per year in the lumbar spine and 1.71% per year in the trochanter in postmenopausal women treated with tamoxifen, it is reasonable to expect that in our populations further stabilization and possibly BMD increases at both lumbar spine and femoral neck sites may occur with continuing tamoxifen treatment.[13]

Interpreting these data is partially compromised by the absence of comparison groups of patients. There are 2 general possible comparison groups of interest: 1) Women with breast cancer (DCIS, for example) who have not received systemic therapy or oophorectomy, or healthy women without breast cancer; and 2) Women with similar breast cancers treated with more usual high-income country combined chemotherapy/hormonal therapy. In the clinical settings in which this study took place, the numbers of such possible comparison patients were small. A second broad global population applicability issue about the data reported here concerns possible ethnic/genetic differences in bone and drug metabolism in the studied women, which might influence these results. We are unaware of any data that would suggest that Filipino or Vietnamese women might differ in basic hormonal/bone biology in significant ways from women in Europe or the United States. The absolute BMD levels in our patients, despite short height and Asian ethnicity, which are characteristics that have been associated with lower BMD, are very similar to those populations (primarily white patients) in other studied high-income countries. Although the apparent absolute differences in BMD at both measured sites between the ethnically different populations with higher levels seen in the Vietnamese patients can be explained by differences in equipment used, analytic software, and environmental BMD determinants, it is possible that genetic population differences exist. However, we found no significant differences between the 2 ethnic groups other than this absolute level parameter. Interestingly, the absolute BMD levels at the lumbar spine and femoral neck in our Vietnamese population were higher than in a study of European women.[16]

A third possible weakness of the current report is that we have no specific data on possible confounding factors such as family history of osteoporosis, alcohol abuse, smoking, or calcium intake, which might lower BMD, or vitamin D or activity levels, which in these populations might increase BMD (because of large sun exposure and high levels of physical activity). Fourth, we have missing data on BMD measurements, but as discussed above these are less extensive than in many reported studies.

Implications

Although fracture risks increase with decreased BMD, the magnitude of the changes with the observed changes in the studied populations is extremely difficult to assess. There is a marked (as great as 10-fold variation) in age-adjusted fracture risk among countries.[11] The World Health Organization technical report recommends interpretation of BMD results using country-specific reference populations.[11] As an example, by standards for a white population in which absolute femoral neck BMD of 0.62 to 0.85 g/cm2 are classified as osteopenic, two-thirds of the 270 patients studied here would be considered osteopenic at baseline.

The results of this study have major implications for the adjuvant treatment of premenopausal women with hormone receptor–positive breast cancer, as treatment is practiced in low- and middle-income countries. Because there are data suggesting that surgical oophorectomy plus tamoxifen is superior to LHRH plus tamoxifen (reviewed above), and of equivalent benefit compared with chemotherapy with cyclophosphamide, doxorubicin, and 5-fluorouracil plus LHRH plus tamoxifen, the similar lumbar spine BMD losses and absence of BMD loss in the femoral neck with the former treatment alters the overall risk–benefit analysis for these treatments in favor of the studied hormonal therapy.[3, 5] Should the primary hypothesis being tested in the parent trial be supported, that luteal phase oophorectomy is more effective, then additional evidence would further favor this combined hormonal approach.[21] The possibility that Her-2/neu–positive tumor-bearing patients also benefit significantly from surgical oophorectomy plus tamoxifen also adds another important consideration in favor of this approach.[22] Finally, the repeated demonstrations of significant losses of BMD at both lumbar and femoral neck sites with chemotherapy and LHRH treatments have led to a multiplicity of studies evaluating bisphosphonate adjuvant therapies. These have generally demonstrated benefits in stabilizing or achieving gains in BMD with this new treatment, but these therapies come with their own side effects (eg, jaw osteonecrosis) and particularly high financial costs. In low- and middle-income countries where the majorities of premenopausal women with breast cancer reside, and where oral hygiene is sadly poor, the results of this study are particularly important because they suggest that such additional bisphosphonate treatment may not be necessary to maintain BMD and decrease long-term fracture risk.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. CONFLICT OF INTEREST DISCLOSURE
  9. REFERENCES

This work was supported by the National Cancer Institute at the National Institutes of Health (R01 CA 097375); the Breast Cancer Research Foundation; and the International Breast Cancer Research Foundation.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. CONFLICT OF INTEREST DISCLOSURE
  9. REFERENCES
  • 1
    Economist Intelligence Unit report, sponsored by Livestrong. Breakaway: The global burden of cancer—challenges and opportunities. http://www.livestrong.org/pdfs/GlobalEconomicImpact. August 2009.
  • 2
    Early Breast Cancer Trialists' Collaborative Group. Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: An overview of the randomised trials. Lancet. 2005;365:16871717.
  • 3
    Love RR, Van Dinh N, Quy TT, et al. Survival after adjuvant oophorectomy and tamoxifen in operable breast cancer in premenopausal women. J Clin Oncol. 2008;26:253257.
  • 4
    Robert H, Wang M, Cella D, et al. Phase III comparison of tamoxifen versus tamoxifen with ovarian ablation in premenopausal women with axillary node negative receptor-positive breast cancer ≤ 3 cm [Abstract]. Proc Am Soc Clin Oncol. 2003;22:5 (abstract 16).
  • 5
    Davidson NE, O'Neill AM, Vukov AM, et al. Chemoendocrine therapy for premenopausal women with axillary lymph node-positive, steroid hormone receptor-positive breast cancer: Results from INT 0101 (E5188). J Clin Oncol. 2005;23:59735982.
  • 6
    Klijn JG, Beex LV, Mauriac L, et al. Combined treatment with buserelin and tamoxifen in premenopausal metastatic breast cancer: A randomised study. J Natl Cancer Inst. 2000;92:903911.
  • 7
    Klijn JGM, Blamey RW, Boccardo F, et al. Combined tamoxifen and luteinizing hormone-releasing hormone (LHRH) agonist alone in premenopausal advanced breast cancer: A meta-analysis of four randomized trials. J Clin Oncol. 2001;19:343353.
  • 8
    Cuzick J, Ambroisine L, Davidson N, et al. Use of luteinising-hormone-releasing hormone agonists as adjuvant treatment in premenopausal patients with hormone-receptor-positive breast cancer: A meta-analysis of individual patient data from randomised adjuvant trials. Lancet. 2007;369:17111723.
  • 9
    Parton M, Smith IE. Controversies in the management of patients with breast cancer: adjuvant endocrine therapy in premenopausal women. J Clin Oncol. 2008;26:745762.
  • 10
    Siris ES, Chen YT, Abbott TA, et al. Bone mineral density thresholds for pharmacologic intervention to prevent fractures. Arch Int Med. 2004;164:11081112.
  • 11
    WHO Scientific Group. Prevention and Management of Osteoporosis. WHO Technical Report Series 921. Geneva, Switzerland: World Health Organization; 2003.
  • 12
    Love RR, Mazess RB, Barden HS, et al. Effects of tamoxifen on bone mineral density in postmenopausal women with breast cancer. N Engl J Med. 1992;326:852856.
  • 13
    Powles TJ, Hickish T, Kanis JA, et al. Effect of tamoxifen on bone mineral density measured by dual-energy x-ray absorptiometry in healthy premenopausal and postmenopausal women. J Clin Oncol. 1996;14:7884.
  • 14
    Eastell R, Admas J, Glack G, et al. Long term effects of anastrozole on bone mineral density: 7 year results from the ATC trial. Ann Oncol. 2011;22:857862.
  • 15
    Sverrisdottir A, Fortnander H, Jabcobsson H, von Schoultz E, Rutqvist LE. Bone mineral density among premenopausal women with early breast cancer in a randomized trial of adjuvant endocrine therapy. J Clin Oncol. 2004;22:36943699.
  • 16
    Gnant MF, Mlineritsch B, Luschin-Ebengreuth G, et al. Zoledronic acid prevents cancer treatment-induced bone loss in premenopausal women receiving adjuvant endocrine therapy for hormone-responsive breast cancer: a report from the Austrian Breast and Colorectal Cancer Study Group. J Clin Oncol. 2007;25:820828.
  • 17
    Shapiro CL, Manola J, Leboff M. Ovarian failure after adjuvant chemotherapy is associated with rapid bone loss in premenopausal women with early stage breast cancer. J Clin Oncol. 2001;19:33063311.
  • 18
    Powles TJ, McCloskey E, Paterson AH, et al. Oral clodronate and reduction in loss of bone mineral density in women with operable primary breast cancer. J Natl Cancer Inst. 1998;90:704708.
  • 19
    Saarto T, Blomqvist C, Valimaki M, et al. Chemical castration induced by adjuvant cyclophosphamide, methotrexate, and fluorouracil chemotherapy causes rapid bone loss that is reduced by clodronate: A randomized study in premenopausal breast cancer patients. J Clin Oncol. 1997;15:13411347.
  • 20
    Delmas PD, Balena R, Confravreux E, Hardouin C, Hardy P, Bremond A. Bisphosphonate risedronate prevents bone loss in women with artificial menopause due to chemotherapy of breast cancer: A double blind, placebo-controlled study. J Clin Oncol. 1997;15:955962.
  • 21
    Stevens W. Asymptotic regression. Biometrics. 1951;7:247267.
  • 22
    Love RR, Duc NB, Dinh NV, et al. Mastectomy and oophorectomy by menstrual cycle phase in operable breast cancer. J Natl Cancer Inst. 2002;94:662669.
  • 23
    Love RR, Duc NB, Havighurst TC, et al. HER-2/neu overexpression and response to oophorectomy plus tamoxifen adjuvant therapy in estrogen receptor-positive premenopausal women with operable breast cancer. J Clin Oncol. 2003;21:453457.