Dr Cauley has received funding from Eli Lilly and Company, Merck & Co., Inc., Novartis, and Pfizer Inc. She served on the speaker's bureau for Eli Lilly and Company and received honorarium from Eli Lilly and Company, Merck & Co., Inc., and Novartis. Dr Cox, Geiger, Kulkarn, and Sashegyi are employees of and hold stock in Eli Lilly and Company. All other authors have no conflict of interest
Risk-Benefit Profile for Raloxifene: 4-Year Data From the Multiple Outcomes of Raloxifene Evaluation (MORE) Randomized Trial
Article first published online: 12 APR 2004
Copyright © 2004 ASBMR
Journal of Bone and Mineral Research
Volume 19, Issue 8, pages 1270–1275, August 2004
How to Cite
Barrett-Connor, E., Cauley, J. A., Kulkarni, P. M., Sashegyi, A., Cox, D. A. and Geiger, M. J. (2004), Risk-Benefit Profile for Raloxifene: 4-Year Data From the Multiple Outcomes of Raloxifene Evaluation (MORE) Randomized Trial. J Bone Miner Res, 19: 1270–1275. doi: 10.1359/JBMR.040406
- Issue published online: 2 DEC 2009
- Article first published online: 12 APR 2004
- Manuscript Accepted: 12 APR 2004
- Manuscript Revised: 16 MAR 2004
- Manuscript Received: 13 NOV 2003
- selective estrogen receptor modulator;
- hormones and receptors;
- estrogens and selective estrogen receptor modulators;
Posthoc analysis of the MORE osteoporosis treatment trial assessed risk-benefit profile of raloxifene in 7705 postmenopausal women. A major disease outcomes global index resulted in annual rates of 1.39% and 1.83% in the raloxifene and placebo groups, respectively (HR, 0.75; 95% CI, 0.62-0.92), compatible with a favorable risk-benefit profile for raloxifene for treating postmenopausal osteoporosis.
Introduction: The Women's Health Initiative (WHI) trial reported overall risks that exceeded benefits from use of estrogen-progestin in healthy postmenopausal women. The objective of this posthoc analysis of the Multiple Outcomes of Raloxifene Evaluation (MORE) trial was to assess the safety profile of raloxifene, a selective estrogen receptor modulator indicated for the prevention and treatment of osteoporosis, using the global index method from the WHI trial.
Materials and Methods: A total of 7705 postmenopausal women (mean age, 67 years) were enrolled in the MORE osteoporosis treatment trial and randomly assigned to receive placebo or one of two doses of raloxifene (60 or 120 mg/day) for 4 years. A global index of clinical outcomes, defined as described for the WHI trial (the earliest occurrence of coronary heart disease, stroke, pulmonary embolism, invasive breast cancer, endometrial cancer, colorectal cancer, hip fracture, or death because of other causes) was applied to the MORE trial data. Physicians blinded to treatment assignment adjudicated events. Intention-to-treat survival analysis of time-to-first-event was performed using a proportional hazards model.
Results and Conclusions: The annualized rate of global index events was 1.83% in the placebo group and 1.39% in the combined raloxifene dose groups (hazard ratio [HR], 0.75; 95% CI, 0.62-0.92). Analyzing individual dose groups separately yielded the same results (HR for 60 mg/day, 0.75; 95% CI, 0.60-0.96: HR for 120 mg/day, 0.75; 95% CI, 0.59-0.95). Subgroup analyses showed no significant interactions between age or hysterectomy status and the effect of raloxifene on the global index (interaction p > 0.1), whereas the global index risk reduction seemed to be greater in obese women compared with nonobese women (interaction p = 0.03). The significant 25% reduction in global index is compatible with a favorable risk-benefit safety profile when raloxifene is used for osteoporosis treatment in postmenopausal women. These results require confirmation in ongoing clinical trials.
THE WOMEN'S HEALTH Initiative (WHI) estrogen-progestin randomized trial focused on defining the risks and benefits of conjugated equine estrogen combined with medroxyprogesterone acetate on major disease outcomes in postmenopausal women.(1) The primary efficacy outcome of the WHI trial was coronary heart disease (CHD) and the primary adverse safety outcome was invasive breast cancer, with the analysis of hormone therapy on overall health included as a central focus of the trial.(2) The global index of risks and benefits, defined to assess the risk-benefit safety profile of hormone therapy for prevention of chronic diseases, included coronary heart disease, stroke, pulmonary embolism, hip fracture, invasive breast cancer, endometrial or colorectal cancer, and deaths because of other causes.(1) The estrogen-progestin portion of the WHI trial was stopped early in July 2002, based on the finding of increased risk of invasive breast cancer and evidence of other overall health risks exceeding the benefits supported by the global index analysis.(1,3) A review and meta-analyses performed for the U.S. Preventive Services Task Force also suggested that risks with prolonged hormone therapy use outweighed benefits.(4) These results emphasize the need to assess the risk and benefit balance of osteoporosis therapies across multiple outcomes, particularly when therapies may have different effects in different tissues.
Raloxifene (Eli Lilly and Company), a selective estrogen receptor modulator (SERM) with tissue-dependent estrogen agonist or antagonist effects, is currently indicated for the prevention and treatment of osteoporosis in postmenopausal women.(5,6) The Multiple Outcomes of Raloxifene Evaluation (MORE) was an osteoporosis treatment trial that enrolled women with osteoporosis based on low BMD or the presence of vertebral fracture, in contrast to the WHI trial, which enrolled postmenopausal women regardless of osteoporosis status.(5) The present analysis was undertaken to assess the risk-benefit safety profile of 4-year treatment with raloxifene using data from the MORE trial and the global index defined in the WHI trial.(1)
MATERIALS AND METHODS
Study subjects were 7705 postmenopausal women enrolled in the MORE study, a randomized, double-blind, placebo-controlled trial conducted at 180 sites in 25 countries. Details of the study design, eligibility criteria, conduct of the trial, patient flow through the trial, and individual results for major clinical outcomes have been published.(5,7–10) Participants were at least 2 years postmenopausal and had osteoporosis defined by a BMD T score of −2.5 SD or less at the lumbar spine or femoral neck or radiographically apparent fractures. They were stratified by site and randomly assigned by a central computer program to receive raloxifene 60 or 120 mg/day or placebo. All women were provided with 500 mg of calcium and 400-600 IU of vitamin D/day. MORE consisted of a 3-year core treatment phase followed by a 1-year extension phase for 4 years of total follow-up. During the fourth year, women were permitted to take other bone-active agents, except for oral estrogen or estrogen-progestin therapy.(9) This report is based on data reported through 4 years of follow-up.
Participants, investigators, laboratory staff, and those who adjudicated clinical outcomes were blinded to treatment assignment. The predefined primary objectives of MORE were to determine the effect of raloxifene compared with placebo on the rate of new vertebral fractures and change in BMD and to assess the safety of chronic raloxifene treatment in postmenopausal women with osteoporosis. Predefined secondary objectives included assessment of the impact of raloxifene treatment on the risk of cardiovascular disease, and breast and endometrial cancer. The protocol was approved by the ethical review board at each site. All women gave written informed consent to participate in the study in accordance with the ethical principles stated in the Declaration of Helsinki.
Ascertainment of clinical outcomes
All events were collected either as serious adverse events reported by the investigator or solicited by direct questioning at each visit. Serious adverse events included those that required in-patient hospitalization, were life-threatening or permanently disabling, were cancer, or resulted in death, and were required by the trial protocol to be reported immediately. Cardiovascular outcomes, cases of breast and endometrial cancer, and hip fracture were ascertained and evaluated as described in detail previously.(5,8–11) Cardiovascular events were collected by solicited questioning as well as by recording unsolicited reports of serious adverse cardiovascular events. Case summaries prepared by investigators for each cardiovascular event were subsequently adjudicated by a cardiologist not associated with the trial and blinded to treatment assignment. Criteria for myocardial infarction included symptoms and documentation of either positive cardiac enzymes or positive electrocardiogram findings compatible with myocardial infarction or documentation of myocardial infarction as a hospital discharge diagnosis or in a physician's statement in the medical record. CHD included nonfatal myocardial infarction, coronary death, or silent myocardial infarction determined by serial electrocardiogram. Coronary death included fatal myocardial infarction, sudden death, unwitnessed death in the absence of other likely noncoronary etiologies, and death related to a coronary artery procedure. Silent myocardial infarction was identified from changes in serial electrocardiograms, coded by the Epidemiological Cardiology Research Center (EPICARE) at Wake Forest University. We used the NOVACODE serial classification system,(12) which required an increase in Q wave score of at least 25 points (grade 2) or at least 15 points (grade 1) with clinically significant ST-T wave evolution from the baseline ECG (change in ST-T score of at least 10). Criteria for stroke included symptoms and documentation of either a positive CT scan or persistent neurological deficit (>24 h) or documentation of stroke as a hospital discharge diagnosis or in a physician's statement in the medical record. For ascertainment of breast cancer, mammography was performed at years 2, 3, and 4, and women were asked every 6 months about any breast cancer, breast biopsies, or breast surgeries that had occurred since the last visit. All breast cancer diagnoses were adjudicated by an independent oncology review board consisting of five physician specialists in breast cancer.(7,8) Cases of endometrial cancer were confirmed by an independent board of gynecologists, based on blinded review of all available clinical data, including centrally read histopathology reports.(7,8) The incidence of osteoporotic nonvertebral fractures, including hip fracture, were determined by direct questioning at each clinic visit and validated by X-rays and medical records.(5) Cases of pulmonary embolism and colorectal cancer were identified by review of serious adverse event reports and were confirmed by physicians blinded to treatment assignment, who independently reviewed each case. Pulmonary embolism events were classified as definite, probable, or possible following the guidelines of the World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception.(13) Colorectal cancer reports supported by both histopathology reports and documentation of surgery or other treatment options were classified as definite; reports supported by either histopathology reports or documentation of surgery or other treatment options were classified as probable. Only cases of pulmonary embolism and colorectal cancer classified as definite or probable were included in these analyses. All deaths were included in the analysis either as deaths because of one of the outcomes included in the global index or as death because of other causes.
A global index of potential benefits and risks was defined using WHI criteria as the first occurrence of coronary heart disease, stroke, pulmonary embolism, invasive breast cancer, endometrial cancer, colorectal cancer, hip fracture, or death because of other causes.(1) Survival analysis of time-to-first event was performed using the intention-to-treat principle through 4 years of follow-up. To provide results based on analyses similar to those used for the global index as originally applied in the WHI trial,(1) events reported after a patient discontinued from the trial were censored from the analysis. However, similar results were obtained when all known events were included in the analysis whether or not the subject was in the study at the time the event was reported. Results also were similar for each dose group compared with placebo. Thus, unless otherwise noted, we present the data most closely resembling the analysis as performed in the WHI trial, and the 60 and 120 mg/day raloxifene dose groups were pooled. Comparisons of the placebo event rates in WHI and MORE were carried out using exact binomial tests with no adjustments for covariates.
Hazard ratios (HRs) and 95% CIs were calculated from Cox proportional hazards analyses. A global index hazard ratio <1 is consistent with net risk reduction and a favorable risk-benefit profile, whereas a hazard ratio >1 is consistent with a net risk increase and an unfavorable risk-benefit profile.
Because women in MORE tended to be older and leaner than women enrolled in WHI and could have had a hysterectomy (whereas all women in the WHI estrogen-progestin trial had an intact uterus), subgroup analyses were performed to test for interaction (p < 0.1) between treatment effect and age, body mass index (BMI), and hysterectomy status.
Baseline characteristics of MORE participants have been published.(8) On average, women were 67 years old, 19 years postmenopausal, primarily white (96%), and had a mean BMI of 25 kg/m2; 29% reported prior estrogen therapy, 17% were current smokers, 23% had had a hysterectomy, and 12% had at least one first-degree relative with breast cancer. There were no significant differences in these characteristics across treatment groups.
After an average of 3.4 years of follow-up, 160 of the 2576 women assigned to placebo had experienced a global index event (183/10,000 person-years), as had 244 of the 5129 women assigned to raloxifene (139/10,000 person-years), representing a significant 25% global index risk reduction associated with raloxifene use (absolute rate difference of 44 events/10,000 person-years; HR, 0.75; 95% CI, 0.62-0.92; Table 1). Results were the same when the raloxifene dose groups were analyzed separately for the global index (HR, 0.75; 95% CI, 0.60-0.96 for 60 mg/day and HR, 0.75; 95% CI, 0.59-0.95 for 120 mg/day), and for all of the individual components of the global index with the exception of total mortality. The total mortality hazard ratio for the 60 mg/day group compared with placebo (HR, 0.61, 95% CI, 0.36-1.03) was significantly lower than that for the 120 mg/day group compared with placebo (HR, 1.09, 95% CI, 0.69-1.71).
The reduction in global risk in the raloxifene groups was associated with a statistically significant reduction in the incidence of invasive breast cancer, with nonsignificant trends for reductions in the incidence of cardiovascular events (particularly stroke) and death and an increase in pulmonary embolism. No significant effects were observed on the incidences of endometrial cancer, colorectal cancer, or hip fracture (Table 1; Fig. 1).
Subgroup analyses revealed no significant interactions between the effect of raloxifene on the global index with age or hysterectomy status. When the clinical categorical definition of obesity (BMI > 30 kg/cm2, n = 887) was used, the interaction p value was statistically significant (p = 0.03), with obese women having a greater risk reduction than nonobese women (HR, 0.42; 95% CI, 0.24-0.72 and HR, 0.82; 95% CI, 0.66-1.02, respectively).
In this posthoc analysis, the overall safety of raloxifene was assessed using a global index of clinical events developed previously for the WHI estrogen-progestin randomized trial and applied to postmenopausal women who participated in the MORE osteoporosis treatment trial. During 4 years of treatment, women assigned to raloxifene had significantly fewer global index events compared with women assigned to placebo. In absolute terms, the risk reduction compared with placebo during the MORE trial corresponded to 150 fewer global index events per 10,000 women taking raloxifene for an average of 3.4 years. These findings suggest a favorable risk-benefit profile in postmenopausal women using raloxifene for the treatment of osteoporosis.
The WHI and MORE trials enrolled postmenopausal women with somewhat different characteristics, in that MORE enrolled women with osteoporosis defined by low BMD or prior vertebral fracture,(5) whereas >90% of the women in the WHI trial who had a BMD measurement did not have osteoporosis at baseline.(14) Women enrolled in MORE also were an average of 4 years older and were more likely to be current smokers or to have had a hysterectomy, differences that could have contributed to the higher rate of global index events observed in the MORE placebo group compared with the WHI placebo group (Table 2). On the other hand, women enrolled in WHI were more likely to be overweight (mean BMI in placebo groups of WHI and MORE was 28.5 and 25.2 kg/m2, respectively). Despite these differences and the much greater sample size in the WHI trial (N = 16,608) compared with MORE (N = 7,705), the rates of major clinical outcomes included in the global index observed in the placebo groups were qualitatively similar between the two trials, except for fractures and strokes (Table 2). The higher fracture rate is expected, because MORE was designed as an osteoporosis treatment trial. The higher rate of stroke in the MORE versus WHI placebo groups may be related to age, tobacco use, or a chance observation, but is interesting in light of other studies suggesting linkage between risks of osteoporosis and cardiovascular disease, particularly stroke, in postmenopausal women.(15–17)
During the 5.2 years of follow-up in the WHI estrogen-progestin trial, women assigned to combination conjugated equine estrogen (0.625 mg) and medroxyprogesterone (2.5 mg) experienced ∼100 more global index events per 10,000 women compared with placebo (HR, 1.15; 95% CI, 1.03-1.28).(1) Contributing to this risk-benefit safety profile were significant increases in the risk of invasive breast cancer, CHD events, stroke, and pulmonary embolism, and significant decreases in the risks of hip fracture and colorectal cancer.(1) An increased risk of venous thromboembolism (VTE), including pulmonary embolism, by a magnitude similar to that of estrogen-progestin therapy, has been previously observed in postmenopausal women treated with raloxifene.(8) However, there was no increase in the risk of invasive breast cancer in postmenopausal women treated with raloxifene compared with raloxifene in the MORE trial; indeed, in this trial, women assigned to raloxifene compared with placebo experienced a significant reduction in the risk of invasive breast cancer.(8) Also, in the MORE trial, there was no evidence of an increased incidence of CHD events or stroke associated with raloxifene therapy, and in the subgroup of women who were determined retrospectively to be at increased CHD risk, those treated with raloxifene compared with placebo had a significantly lower incidence of cardiovascular events.(10)
The effect of raloxifene on global index risk was consistent between subgroups of age and was not influenced by hysterectomy status. The evidence that the reduction in risk of global index events was greater among obese women compared with nonobese women is compatible with the greater risk of breast cancer in postmenopausal women with higher endogenous estradiol levels(18) and the antiestrogen effect of raloxifene,(19,20) although this finding may have been because of chance.
Several limitations to our study deserve comment. The global index as applied originally to the WHI trial was limited to life-threatening or otherwise serious outcomes and did not include other clinical outcomes affected by raloxifene. Thus, the global index remains an analysis of safety for raloxifene in postmenopausal women with osteoporosis. From an efficacy standpoint, raloxifene significantly reduced the risk of vertebral fractures, the primary endpoint of the MORE trial, by ∼40% at 4 years.(9) This effect was not included in the current analyses because vertebral fracture was not included as an outcome in the WHI global index.(1) The global index also did not address health-related quality of life issues. Although the occurrence of hot flushes is increased in some women treated with raloxifene and estrogen-progestin is an established treatment for hot flushes, neither has been shown to have a substantial effect on overall health-related quality of life in relatively asymptomatic postmenopausal women.(21,22) Because MORE was an osteoporosis treatment trial, all women had low BMD or a previous vertebral fracture. This differs from the WHI trial, which enrolled women without regard to osteoporosis status. Whether the risk-benefit safety profile observed in this trial may be extrapolated to postmenopausal women without osteoporosis or with different characteristics is unknown. The information available for adjudication of some events was limited to information provided by investigators as serious adverse events reports. Because some events were collected as adverse events, under- or overestimation of the actual event rates is possible dependent on regional differences in reporting or monitoring practices, but should have affected reporting across the treatment groups equally for a given event. That said, the rates of breast and colon cancers observed in placebo group of MORE were consistent with expectations for white women over the age of 65,(8,23) and the rate of coronary events was similar to those reported previously for postmenopausal women enrolled in primary cardiovascular prevention studies.(24–26) Assessing risks of cardiovascular disease, breast cancer, nonvertebral fracture, and endometrial cancer were secondary objectives of the trial, but assessing risk of colorectal cancer was not a predefined objective of the trial. MORE was not designed to determine a global effect of raloxifene on major clinical outcomes and the number of global index events observed in MORE (404 events in 7705 women) was much less than that from WHI (1374 in 16,608 women). Although the rates of nonfracture events were comparable in the WHI and MORE placebo groups, the rates of stroke, hip fracture, and global index events overall were higher in MORE (Table 2), consistent with differences in the patient populations enrolled in these trials. These results were based on posthoc analyses and require confirmation. The Raloxifene Use for the Heart (RUTH) trial is assessing the effect of raloxifene on risk of cardiovascular events and invasive breast cancer in 10,101 postmenopausal women at high risk for or with established CHD.(27) The effect of raloxifene on risk of invasive breast cancer is also being studied in ∼4000 women in the Continuing Outcomes Relevant to Evista (CORE) trial and in ∼19,000 women in the Study of Tamoxifen And Raloxifene (STAR).(28)
In conclusion, raloxifene is currently indicated for the prevention and treatment of osteoporosis, but its effects on other clinical outcomes contribute to the overall assessment of safety of the drug. The global index safety analysis in the MORE osteoporosis treatment trial was consistent with a favorable safety profile for 4 years of raloxifene therapy in postmenopausal women with osteoporosis. Data from ongoing trials will enhance our understanding of the risks and benefits of raloxifene across multiple diseases that can impact the health of postmenopausal women.
The authors thank Yiyong Fu for statistical programming and analysis support and all of the MORE trial investigators who made this study possible (a complete list has been published previously: JAMA 1999;282:644-645). This study was supported by a grant from Eli Lilly and Company. Data were analyzed at Lilly Research Laboratories, Eli Lilly and Company. All authors had full access to the data and analyses.
- 121992 NOVACODE serial ECG classification system for clinical trials and epidemiologic studies. J Electrocardiol 24 (Suppl): 179–187., ,
- 13World Health Organization 1995 A multinational case-control study of cardiovascular disease and steroid hormone contraceptives. Description and validation of methods. World Health Organization Collaborative Study of Cardiovascular and Steroid Hormone Contraception. J Clin Epidemiol 48: 1513–1547.
- 23National Cancer Institute 2000 Surveillance, epidemiology, and end result (SEER) cancer statistics review, 1975–2000. Available online at http://seer.cancer.gov/csr/1975_2000. Accessed on January 6, 2004.