More than 11 million postmenopausal women in the United States were estimated to be on hormone therapy (HT; ie, estrogen, either alone or with the addition of a progestin)1 before the Women's Health Initiative results reported in July 2002.2 Although HT is often initiated for the relief of menopausal symptoms, particularly hot flushes, the presumed overall multisystem benefits of HT, including prevention of heart disease, osteoporosis, and possibly other chronic conditions, have been the rationale for continued treatment in many postmenopausal women. However, recent data from the 16,608-participant combined estrogen–progestin arm of the Women's Health Initiative study show net harm rather than benefit.2 As a result, organizations such as the National Institutes of Health (Kirschstein R. Menopausal hormone therapy: summary of a scientific workshop [abstract]. Ann Intern Med 2003;138:361–4), the U.S. Preventive Services Task Force,3 the American College of Obstetricians and Gynecologists,4 the North American Menopause Society,5 and the U.S. Food and Drug Administration recommend that use of HT be limited to short-term treatment of menopausal symptoms using the lowest effective dose. Since the results from the Women's Health Initiative became available, millions of women have discontinued postmenopausal HT (Brody JE. Options for protecting bones after menopause. New York Times. April 22, 2003;Science Times:F6).
As shown in a large number of studies, estrogen use initially increases, and later maintains, bone mass.6–12 Although observational studies of HT use have suggested an antifracture benefit,6,12 the Women's Health Initiative data provided the first confirmation from a randomized trial that estrogen use reduces the incidence of fractures, including hip fractures, even in a population not selected for low bone mass.2 However, discontinuation of estrogen use results in accelerated bone loss, as shown in most observational studies6–8,12 and several clinical trials.9,11 It thus seems likely that the antifracture efficacy of estrogen might dissipate after estrogen withdrawal, although the extent and time course of any loss of benefit would not necessarily correlate precisely with bone mineral density loss. This possibility is supported by a recent prospective study showing increased fracture rates among past hormone users when compared with current users.12
We present here results from the National Osteoporosis Risk Assessment, a large longitudinal study of postmenopausal women in the United States on the association between the cessation of postmenopausal estrogen therapy and hip fracture risk. We focus on hip fractures because they are the most devastating, costly, and reliably reported events. Results on other fractures (wrist, forearm, rib, spine, and hip combined) have been published.12
MATERIALS AND METHODS
The National Osteoporosis Risk Assessment, begun in 1997, is a longitudinal observational study of osteoporosis among postmenopausal women throughout the United States. The study protocol and consent documents were approved by a national institutional review board, and all participants provided informed consent. Details of the study design have been published.13–15 Briefly, all participants were postmenopausal women, aged at least 50 years with no previous diagnosis of osteoporosis, no bone mineral density testing in the preceding 12 months, and no use of osteoporosis-specific medications. Estrogen use was permitted. A total of 200,160 women were recruited through the practices of 4,236 primary care physicians located in the 34 states and the District of Columbia. Each office provided randomly chosen names of eligible women who received letters of invitation from their physicians to participate in NORA. In most practices, 40 to 100 women were enrolled after giving informed consent.
Each participant had bone mineral density measured at a single peripheral site (heel, forearm, or finger) under standardized quality control, as previously described.13–15 It has been shown that peripheral bone mineral density is predictive of hip fractures in the National Osteoporosis Risk Assessment13,15 and in other cohorts.16,17 Information collected by standard self-administered questionnaires and used in the current analyses included age, ethnicity, education, self-rated health status, height and weight, number of years since menopause, personal and family history of fracture, maternal history of osteoporosis, medication use (thyroid hormone, cortisone, or diuretics), current and past postmenopausal estrogen use, use of calcium supplements, cigarette smoking, exercise, and use of caffeine and alcohol. Body mass index (BMI) was calculated from reported weight and height. Duration and recency of postmenopausal estrogen use were based on answers to 5 questions that addressed: 1) whether postmenopausal estrogen had ever been used, 2) the number of years of use, 3) the age at which estrogen use began, 4) whether estrogen was being used currently, and 5) if estrogen had been discontinued, how many years since last use.
Approximately 12 months after the baseline evaluation, participants were asked to complete a follow-up questionnaire that included information on the occurrence and site(s) of new fractures (wrist, forearm, rib, spine, and hip). Women who did not respond to the follow-up survey (36,181) were excluded from this study. An additional 23,395 women with missing responses to questions about estrogen recency or duration of use were also excluded, leaving a sample of 140,584 women for the current study. Incident fractures were defined as having occurred between the date of bone mineral density measurement and the date of completion of the follow-up survey. Participants reporting 4 or more incident fractures at follow-up were excluded from the analysis because multiple fractures are likely to have been the result of trauma.
Of the 140,584 women included in these analyses, 364 reported an incident hip fracture. Subsequently telephone contact was successful for 344 (94.5%) women. Among these 344 women, 260 (78%) confirmed a hip fracture, and a diagnosis other than hip fracture was indicated by 75 (22%). Self-report of hip fracture has shown to be reliable compared to radiographs or medical records.18,19 We report here data using the telephone-verified hip fractures. The conclusions were not altered when analyses were repeated based on all self-reported hip fractures.
Analyses were conducted by using SAS 8 software (SAS Institute, Inc, Cary, NC). Baseline number of years of estrogen use (duration) was categorized as 0 to 5 years, 6 to 10 years, and more than 10 years. Baseline recency of use was categorized as current, quit within past 5 years, and quit more than 5 years ago. The relationship of recency and duration with incident hip fracture and covariates was examined in bivariate analysis by using Cochran-Mantel-Haenszel test for trend.20 We did not examine change in HT during the 1-year interval between baseline hormone use assessment and incident fractures. Statistical significance was tested by using χ2 statistics for categorical variables and analysis of variance models for continuous variables.
Hip fracture rates per 1,000 person years were calculated by summing new fractures reported in follow-up divided by person-years of follow-up. Logistic regression models were used to estimate the effects of estrogen duration and recency on incident hip fractures compared with never users while controlling for potentially confounding variables. These variables were selected based on the existing literature, significance in the bivariate analysis, and stepwise selection in the multivariable logistic regression models. All reported P values are from 2-tailed tests of statistical significance.
Of the 140,584 women included in this report, 48% were on HT at baseline and an additional 14% had used postmenopausal estrogen in the past; 128,976 (92%) were white, 4,719 (3.4%) were African American, 3,522 (2.5%) were Hispanic, 1,092 (0.8%) were Asian, 995 (0.7%) were Native American, and 1,280 (0.9%) were of other race. The mean age, BMI, and bone mineral density T score were 63.8 years (standard deviation [SD] 8.97 years), 27.7 kg/m2 (SD 5.9 kg/m2), and −0.82 (SD 1.13), respectively. Tables 1 and 2 present the association of demographic and clinical factors with duration and recency of estrogen use. According to Cochran-Mantel-Haenszel test for trend age, T-score, BMI, health status, prior fracture, maternal history of fracture, and cortisone use were each significantly (P < .01) associated with both duration and recency of estrogen use.
The association of demographic and clinical factors with hip fracture status is shown in Table 3. Women who had a hip fracture were older, weighed less, more apt to have reported poor health status, and more likely to have lower T-scores, prior fracture, and maternal history of fracture and cortisone use.
Table 4 shows the crude and adjusted association between duration and recency of estrogen use and incident hip fracture. The unadjusted incidence rate (per 1,000 person years) of hip fracture was similar in never users of estrogen (incidence rate 2.24; 95% confidence interval [CI] 1.87, 2.59) and in women who had discontinued estrogen, whether within the last 5 years (incidence rate 2.17; 95% CI 1.28, 3.05), or more than 5 years earlier (incidence rate 2.51; 95% CI 1.63, 3.39). Only current users of estrogen reported a lower incidence of hip fracture (incidence rate 0.81; 95% CI 0.61, 1.00). In multivariate analysis accounting for effects of age, BMI, prior fracture, health status, maternal history of fracture, and cortisone use, women who reported current use of estrogen at baseline had a highly significant (P < .001) reduced odds of hip fracture compared to never users (odds ratio 0.60; 95% CI 0.44, 0.82). Women who had stopped using estrogen more than 5 years earlier had an adjusted hip fracture risk similar to those who never used estrogen. However, women who had discontinued estrogen within the prior 5 years had an increased hip fracture odds (odds ratio 1.65; 95% CI 1.05, 2.59). There were no significant associations between duration of estrogen use and fracture risk within the current and prior user cohorts (Table 4).
Consistent with the data from Women's Health Initiative, the current study showed a 40% decrease in incidence of hip fracture among women currently taking HT relative to never users, and the antifracture benefit of current HT use was apparent within the first 5 years of treatment. These data support the concept that antiresorptive treatment of a general population of postmenopausal women not known to have osteoporosis may be associated with a substantial decrease in hip fracture risk.
Based on the compelling data from Women's Health Initiative, as well as data from the Heart and Estrogen/Progestin Replacement Study,21,22 several organizations (Kirschstein R. Menopausal Hormone Therapy: Summary of a Scientific Workshop [abstract]. Ann Intern Med 2003;138:361–4)3–5 and the U.S. Food and Drug Administration strongly recommend that women should not take any type of estrogen for the prevention of chronic diseases. As a result, there has been a substantial decline in the number of women prescribed estrogen (Brody JE. New York Times. April 22, 2003;Science Times:F6). With increasing awareness and revised product labeling, it is likely that many more women will discontinue estrogen or decide not to use it at all.
In the present study, women who had a diagnosis of osteoporosis or a bone mineral density measurement in the 12 months before study start were excluded. It is therefore unlikely that skeletal protection was the primary motivation for estrogen use in these women. Based on a recent search within a large pharmacy benefits manager database, only 6% of women discontinuing estrogen subsequently initiated a bone-specific therapy, such as alendronate, risedronate, raloxifene, or calcitonin (unpublished data, Merck Health Solutions database, 2003). Perhaps one reason for this low initiation rate is the assumption by patients and physicians that prior estrogen use preserves bone and prevents fractures for years after estrogen discontinuation. However, recent clinical trials have shown that rapid bone loss follows estrogen discontinuation, such that bone mineral density approached the pretreatment level within 2 years after estrogen withdrawal.9–11,23
Rapid decreases in bone mineral density are known to be associated with deterioration in the microarchitectural structure of bone because excessive osteoclastic resorption leads to loss of many trabecular cross-connections, resulting in profound reductions in bone strength.24 Consistent with these biomechanical considerations, there is evidence that, for any given bone mineral density, patients with high rates of bone turnover (reflected by increased biochemical markers), are at higher risk of fracture than those with lower bone turnover.25 In addition, because during the early phase of estrogen withdrawal, the rate of bone resorption substantially exceeds that of bone formation, not only are individual trabeculae at high risk of loss from osteoclastic resorption, but repair of partly transected trabeculae may be incomplete, allowing either physical stresses or subsequent waves of resorptive activity to result in loss of connectivity.
We have previously reported that National Osteoporosis Risk Assessment women who were current users of HT had bone mineral density values that were approximately .70 SD higher than levels in women who either never took HT after menopause or last used HT more than 5 years previously, whereas those who had more recently discontinued HT had intermediate bone mineral density values.12 Based on those findings, if bone mineral density were the sole predictor of fracture risk, it would be expected that recent HT users would retain some degree of fracture risk reduction for several years after HT withdrawal. However, the present National Osteoporosis Risk Assessment follow-up study suggests that hip fracture risk reduction is completely lost, and may be reversed, within 5 years after estrogen cessation. Thus, the crude incidence of hip fractures in women who had discontinued estrogen was very similar to that in never users. After appropriate adjustment for baseline risk factors, the women who had discontinued estrogen within the last 5 years actually had a higher risk of hip fracture than never users, whereas those who had discontinued estrogen more than 5 years previously were indistinguishable from never users. However, whether the risk of hip fracture during the first few years after HT cessation actually exceeds that in never users of HT requires confirmation. Follow-up of Women's Health Initiative women, for whom HT cessation was strongly recommended, and their control group of placebo-treated women, should be very informative and could shed additional light on the effect of HT discontinuation on subsequent fracture risk.
Despite the advantages of large size and broad geographic and ethnic participation, the National Osteoporosis Risk Assessment is limited in its generalizability to the whole U.S. population of postmenopausal women because women who had been diagnosed as having osteoporosis, were under treatment for osteoporosis, or had recent bone mineral density testing were ineligible for participation. Furthermore, women who responded to the follow-up survey tended to be younger, better educated, more health conscious, and more likely to be current or past estrogen users than the nonresponders.
This study suggests that after estrogen withdrawal women have a risk for hip fracture that is as high as, or higher than, that in women who never used estrogen. A large number of women discontinued HT since the Women's Health Initiative results were published and many more will do so over the next several years, resulting in increased risk for fracture. Postestrogen evaluation should include consideration of the need for intervention to prevent bone loss and fractures.
2.Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ, Kooperberg C, Stefanick ML, et al.; Writing Group for the Women's Health Initiative Investigators. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results From the Women's Health Initiative randomized controlled trial. JAMA 2002;288:321–33.
3.U.S. Preventive Services Task Force. Postmenopausal hormone replacement therapy for primary prevention of chronic conditions: recommendations and rationale. Ann Intern Med 2002;137:834–9.
4.American College of Obstetricians and Gynecologists. Guidelines for women's health care. 2nd ed. Washington, DC: American College of Obstetricians and Gynecologists; 2002.
5.North American Menopause Society. Estrogen and progestogen use in peri- and postmenopausal women: September 2003 position statement of the North American Menopause Society. Available at: http://www.menopause.org/HTpositionstatement.pdf
. Retrieved January 28, 2004.
6.Cauley JA, Seeley DG, Ensrud K, Ettinger B, Black D, Cummings SR. Estrogen replacement therapy and fractures in older women Study of Osteoporotic Fractures Research Group. Ann Intern Med 1995;122:9–16.
7.Schneider DL, Barrett-Connor EL, Morton DJ. Timing of postmenopausal estrogen for optimal bone mineral density: The Rancho Bernardo Study. JAMA 1997;277:543–7.
8.Beck TJ, Stone KL, Oreskovic TL, Hochberg MC, Nevitt MC, Genant HK, et al. Effects of current and discontinued estrogen replacement therapy on hip structural geometry: the study of osteoporotic fractures. J Bone Miner Res 2001;16:2103–10.
9.Greenspan SL, Emkey RD, Bone HG, Weiss SR, Bell NH, Downs RW, et al. Significant differential effects of alendronate, estrogen, or combination therapy on the rate of bone loss after discontinuation of treatment of postmenopausal osteoporosis: a randomized, double-blind, placebo-controlled trial. Ann Intern Med 2002;137:875–83.
10.Gallagher JC, Rapuri PB, Haynatzki G, Detter JR. Effect of discontinuation of estrogen, calcitriol, and the combination of both on bone density and bone markers. J Clin Endocrinol Metab 2002;87:4914–23.
11.Greendale GA, Espeland M, Slone S, Marcus R, Barrett-Connor E; PEPI Safety Follow-Up Study (PSFS) Investigators. Bone mass response to discontinuation of long-term hormone replacement therapy: results from the Postmenopausal Estrogen/Progestin Interventions (PEPI) Safety Follow-up Study. Arch Intern Med 2002;162:665–72.
12.Barrett-Connor E, Wehren LE, Siris ES, Miller PD, Chen YT, Abbott TA, et al. Recency and duration of postmenopausal hormone therapy: effects on bone mineral density and fracture risk in the National Osteoporosis Risk Assessment (NORA) Study. Menopause 2003;10:412–9.
13.Siris ES, Miller PD, Barrett-Connor E, Faulkner KG, Wehren LE, Abbott TA, et al. Identification and fracture outcomes of undiagnosed low bone mineral density in postmenopausal women: results from the National Osteoporosis Risk Assessment. JAMA 2001;286:2815–22.
14.Siris E, Miller P, Barrett-Connor E, Abbott T, Sherwood L, Berger M. Design of NORA, the National Osteoporosis Risk Assessment Program: a longitudinal US. registry of postmenopausal women. Osteoporos Int 1998;8(suppl 1):S62–9.
15.Miller PD, Siris ES, Barrett-Connor E, Faulkner KG, Wehren LE, Abbott TA, et al. Prediction of fracture risk in postmenopausal women with peripheral bone densitometry: evidence from the National Osteoporosis Risk Assessment (NORA). J Bone Miner Res 2002;17:2222–30.
16.Marshall D, Johnell O, Wedel H. Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fracture. BMJ 1996;312:1254–9.
17.Black DM, Cummings SR, Genant HK, Nevitt MC, Palermo L, Browner W. Axial and appendicular bone density predict fractures in older women. J Bone Miner Res 1992;7:633–8.
18.Ismail AA, O'Neill TW, Cockerill W, Finn JD, Cannata JB, Hoszowski K, et al. Validity of self-report of fractures: results from a prospective study in men and women across Europe. Osteoporos Int 2000;11:248–54.
19.Nevitt MC, Cummings SR, Browner WS, Seeley DG, Cauley JA, Vogt TM, et al. The accuracy of self-report of fractures in elderly women: evidence from a prospective study. Am J Epidemiol 1992;135:490–9.
20.Landis RJ, Heyman ER, Koch GG. Average partial association in three-way contingency tables: a review and discussion of alternative test. Int Stat Rev 1978;46:237–54.
21.Grady D, Herrington D, Bittner V, Blumenthal R, Davidson M, Hlatky M, et al. Cardiovascular disease outcomes during 6.8 years of hormone therapy: Heart and Estrogen/Progestin Replacement Study follow-up (HERS II). JAMA 2002;288:49–57.
22.Hulley S, Furberg C, Barrett-Connor E, Cauley J, Grady D, Haskell W, et al. Noncardiovascular disease outcomes during 6.8 years of hormone therapy: Heart and Estrogen/Progestin Replacement Study Follow-up (HERS II). JAMA 2002;288:58–66.
23.Ascott-Evans BH, Guanabens N, Kivinen S, Stuckey BG, Magaril CH, Vandormael K, et al. Alendronate prevents loss of bone density associated with discontinuation of hormone replacement therapy. Arch Intern Med 2003;163:789–94.
24.Dempster DW, Birchman R, Xu R, Lindsay R, Shen V. Temporal changes in cancellous bone structure of rats immediately after ovariectomy. Bone 1995;16:157–61.
25.Garnero P, Sornay-Rendu E, Claustrat B, Delmas PD. Biochemical markers of bone turnover, endogenous hormones and the risk of fractures in postmenopausal women: the OFELY study. J Bone Miner Res 2000;15:1526–36.
© 2004 The American College of Obstetricians and Gynecologists