Estrogen deficiency after menopause causes an increase in the rate of bone loss and a corresponding increase in the risk of fractures.1 Immediately after menopause, bone mass decreases by 3–5% per year, and even after age 65, bone mass can continue to decrease at a rate of 0.5–1% per year.2,3 Increased bone turnover is associated with increased bone loss and risk of fractures.4,5
Estrogen replacement therapy (ERT) reduces the rate of bone loss, maintains bone mass, and reduces the risk of hip fracture associated with reductions in bone mass after the menopause.6–8 Bone mineral density increased 5–10% over 2 years and the fracture rate decreased by 50% with the use of antiresorptive drugs such as estrogen.1
It is generally accepted that oral estrogen at a dose of 1 mg of estradiol (E2) or 0.625 mg of equine estrogen is needed to prevent postmenopausal bone loss.9–11 However, recent studies have indicated that lower doses of estrogen may be effective for maintaining bone mass.12,13 This study was done to investigate the efficacy of four doses of a 7-day transdermal 17β-E2 delivery system, including a dose of 0.025 mg/day, for the prevention of bone loss in postmenopausal women. The safety and efficacy of this product in the treatment of vasomotor symptoms have been reported previous.14
Materials & Methods
We conducted a multicenter, double-masked, randomized, placebo-controlled study at ten centers in the United States. The study protocol was approved by an appropriate institutional review board. Each participant gave written informed consent before entry. Because subjects assigned to active treatment were to receive unopposed estrogen for 2 years, only women without uteri were eligible. Women who had undergone hysterectomies without oophorectomies were required to be at least 45 years old and have ovarian failure, as evidenced by vasomotor symptoms for at least 1–5 years before enrollment. Women who had also undergone oophorectomies had to be at least 40 years old and 4 weeks to 5 years postoophorectomy. All women were required to have a serum E2 concentration of at most 20 pg/mL, an FSH level of at least 50 U/L, and fasting serum concentrations of cholesterol of at most 300 mg/dL, triglycerides of at most 300 mg/dL, and glucose of at most 140 mg/dL. Baseline bone mineral density of L2–L4 measured by dual-energy x-ray absorptiometry (Lunar Corporation, Madison, WI or Hologic Inc., Waltham, MA) had to be at least 0.09 g/cm2 (Lunar) or at least 0.086 g/cm2 (Hologic).
Exclusion criteria included known or suspected bone disease, hypo- or hypercalcemia, vitamin D deficiency, bone fracture within 6 months, immobilization for 2 or more of the preceding 6 months, hot flashes requiring hormone therapy, or a history of skin irritation caused by transdermal drug-delivery systems. Women also were excluded if they had ever received bisphosphonates, fluoride, or calcitonin; were receiving chronic treatment with corticosteroids or agents that affect bone metabolism; had had recent ERT or treatment with lipid-lowering drugs; or had participated in another clinical trial within 3 months.
Eligible women were assigned randomly in a ratio of 3:2:2:2:2 to receive placebo or one of four doses of a 17β-E2 transdermal system (Climara; Berlex Laboratories, Wayne, NJ): 0.025,0.05, 0.06, and 0.1 mg/day of E2 delivered in patches of 6.5,12.5,15, and 25 cm2, respectively. The block size at each center was 11 (ie, 3 + 2 + 2 + 2 + 2). Each subject wore two patches (an active and a different-sized inactive patch for the study groups; two different-sized inactive patches for the placebo group). One of a class of Latin square designs masked the assignment of four different patch sizes to the 11 combinations of patch pairs worn in each block. A permuted-blocks randomization for the five treatment groups was centrally computed independently for each investigative site, and the drug was packaged and shipped to each site accordingly. A seeded pseudorandom number generator was programmed with SAS software (SAS, Inc., Gary, NC). Subjects were sequentially assigned random numbers at each site. Personnel conducting the study remained masked to the treatment assignments until after the database was locked. Subjects applied both patches to the anterior trunk and replaced them weekly. Treatment was continued for 26 4-week cycles. Calcium supplementation was provided to achieve a total daily calcium intake of 1500 mg.
Baseline examinations included complete histories, physical examinations including pelvic examinations, and general laboratory tests including hematology, blood chemistry, liver function tests, lipid profile, E2, FSH, and TSH. Serum osteocalcin and urinary concentrations of pyridinoline and deoxypyridinoline cross-links were measured in second morning voided specimens. Subjects wore placebo patches for 1 week to assess tolerance of the transdermal system.
Bone mineral density was measured at four anatomic sites (lumbar spine, nondominant radius, ipsilateral total hip, and femoral neck), and study medication was dispensed. Participants returned for nine follow-up assessments at approximately 3-month intervals. Bone mineral density measurements were repeated at 6, 12, 18, and 24 months. Bone mineral density data were analyzed at a single quality-assurance center, which also provided technical training to all investigative sites and standardization of all equipment used. At these same time points, assays of serum osteocalcin15 and urinary concentrations of pyridinoline and deoxypyridinoline crosslinks (Pyrilinks-D; Metra Biosciences, Mountain View, CA)16,17 were performed in a central laboratory, and physical and pelvic examinations, vital signs, and laboratory tests were repeated. Mammograms were done at baseline and at 1- or 2-year intervals, depending on age.
The primary efficacy variable was the percentage change from baseline in spine bone mineral density in the anteroposterior view of lumbar vertebrae L2–L4. Secondary efficacy variables were the percentage changes from baseline in bone mineral density of the midshaft of the radius, the proximal femur, and the total hip. The percentage change from baseline in serum osteocalcin and the change and percentage change in urinary deoxypyridinoline-creatinine and pyridinoline-creatinine ratios were determined. Safety evaluations were based on physical examinations, mammography, fracture incidence, vital signs, weight, and laboratory tests. Adverse experiences and study investigators' observations of abnormalities or changes in physical or biochemical characteristics were recorded.
Efficacy was determined in the intent-to-treat population, defined as all randomized subjects who took at least one dose of study medication. An analysis using data from the last observation carried forward was done also to minimize bias that may result should responders tend to stay in the study and nonresponders tend to drop out. Analysis of continuous variables was done using two-way analysis of variance with treatment, center, and treatment-by-center interaction as covariates. Pairwise comparisons were performed using least-square-means testing of active-dose groups versus placebo. Paired t tests were used to compare visit and baseline means within each group. When required, nonparametric analyses were done with these same linear models to compare overall and pairwise differences in treatment, using ranks of the original data. Treatment and center effects were retained additively in the model after demonstrating negligible treatment-by-center effects. Although the statistical error rate was preset at α = 0.05, for between-group analyses of efficacy, P values were adjusted by Hochberg method18 of multiple comparisons before being considered significant.
The primary objective was to compare the response to treatment of each dose with placebo. For this analysis, sample size was calculated to be 176 subjects based upon detecting a difference of 4% (±3% standard deviation) in the percentage change in bone mineral density between placebo and active treatment, which corresponds to α = 0.0022 with a power of 60%. The study was not large enough to show efficacy differences between treatment groups, nor could it reliably detect infrequent side effects. A treatment dose was considered effective if that dose resulted in a mean change in bone mineral density that was not less than zero. The fraction of subjects who did not lose bone at each dose was calculated and compared using the extended Cochran-Mantel-Haenszel test. This overall test was followed by pairwise comparisons between placebo and each active dose using Cochran-Mantel-Haenszel tests. The P values from these methods for proportions were corrected by the Hochberg method.18
A total of 175 women were enrolled, and 97 (55%) completed 2 years of treatment. Seventy-eight women did not complete the study; one withdrew for lack of efficacy (vasomotor symptoms), four (2.3%) had protocol deviations, 19 (10.9%) withdrew consent, 20 (11.4%) had adverse events, and 34 (19.4%) were lost to follow-up or relocated or withdrew for other reasons. There were no significant differences in baseline demographic or clinical characteristics among the groups (Table 1).
All four treatment groups showed increases from baseline in mean bone mineral density of the lumbar spine (P < .05 versus placebo at all time points; Figure 1). The mean percentage change in bone mineral density at month 24 ranged from +2.37% in the group treated with the 0.025-mg dose to +4.70% with the 0.10-mg dose, compared with −2.49% in the placebo group. Using data from the last observation carried forward to 24 months, the differences between active treatment and placebo in the mean change in lumbar spine bone mineral density ranged from 4.65% to 7.53% (Table 2).
Mean bone mineral density of the total hip increased from baseline in all four treatment groups (P < .05 versus placebo at all time points; Figure 2). The mean percentage change in bone mineral density at month 24 ranged from +0.26% with the 0.025-mg dose to +2.03% with the 0.10-mg dose, compared with −2.04% in the placebo group. At the 24-month assessment, the differences in the means for hip bone mineral density compared with placebo were 2.31%, 4.07%, 4.27%, and 3.65% for the 0.025-, 0.05-, 0.06-, and 0.10-mg doses, respectively (Table 2). The mean percentage changes from baseline in bone mineral density of the radius and femoral neck compared with placebo were not statistically significant except at 18 and 24 months in the 0.10-mg dose group.
Bone mineral density was maintained at the lumbar spine and in the total hip in most women at all doses and time points. The proportions of women who did not lose bone at 24 months were 90% in the lumbar spine (Table 3) and 76% in the hip at the 0.10-mg dose (Table 4). Although between-dose comparisons were not done, a trend analysis found a significant (P < .05) dose-response effect for bone mineral density of the lumbar spine and total hip at 6, 12, 18, and 24 months; for the radius at 6, 18, and 24 months; for the femoral neck at 18 and 24 months; and for the total hip at 6, 12, 18, and 24 months.
Serum osteocalcin levels decreased from baseline at all time points in all treatment groups but increased in the placebo group. The mean percentage change from baseline in serum osteocalcin was statistically significantly greater than for placebo at 12,18, and 24 months in all treatment groups except for the 0.05-mg and 0.06-mg doses at 24 months (Figure 3). The mean percentage change in urinary deoxypyridinoline/creatinine ranged from −12.3% to −22.1% in the four treatment groups at 24 months. These differences were not statistically significant compared with placebo because of a −8.0% change in the placebo group. The results were similar for urinary pyridinoline/creatinine.
Of the 175 subjects enrolled, 97 (55.4%) reported one or more adverse events during the 2-year treatment. Overall, 20 women (11.4%) withdrew from the study because of adverse events. The most common causes were application-site reactions in six (3.4%), emotional lability in three (1.7%), breast pain in three (1.7%), hot flashes in four (2.3%), headache in two (1.1%), and depression in two (1.1%). The incidence of adverse events with active treatment was comparable to or lower than that with placebo. There were no spontaneous fractures, but four women experienced traumatic fractures during the study, one each in the placebo, 0.025-mg, 0.06-mg, and 0.10-mg groups. There were no clinically relevant or statistically significant changes in routine laboratory results, diastolic blood pressure, heart rate, or body weight.
This study reaffirms the efficacy of transdermal delivery of E2 for preventing postmenopausal bone loss19 and confirms recent reports of decreased bone turnover at doses as low as 0.025 mg/day.12 Compared with placebo, there were significant increases in lumbar spine bone mineral density at 6 months, which were maintained or extended among women who were followed to 24 months. Similar efficacy was seen for maintenance of bone mineral density in the total hip. Consistent decreases in serum osteocalcin and urinary deoxypyridinoline/creatinine, indicative of decreased bone turnover, supported the positive bone mineral density findings.
The most widely prescribed doses of estrogen (0.625 mg of conjugated estrogen, 0.05 mg of E2 transdermally, and 1 mg of E2 orally) have been believed to be at or near the minimum effective dose for preventing postmenopausal bone loss. However, 0.5 mg of E2 administered orally showed preservation of bone mineral density at the lumbar spine at 18 months, although at 24 months the results were not significantly different from placebo.20 The small sample and low precision of measurement in that study probably are responsible for this result. More recently, 0.3 mg of esterified estrogens was shown to preserve bone mineral density.13 Our current study has demonstrated in a 2-year prospective design that doses of transdermal E2 as low as 0.025 mg/day prevent osteoporosis in postmenopausal women.
The finding of beneficial effects on bone mineral density with a 6.5-cm2 patch delivering E2 at 0.025 mg/day contrasts with the results of some older studies that showed no benefit of low estrogen doses for preventing postmenopausal bone loss.10,11 These differences may relate to the lower sensitivity of older instrumentation, which thus required larger populations to demonstrate efficacy. The most recent report showing no benefit from low-dose estrogen was a case-control study, the results of which may be limited by confounding bias and low statistical power in subgroup analyses.9 More recent studies have shown efficacy of low doses of estrogen.
The magnitude of changes in bone mineral density in our study are consistent with those reported in a double-masked, randomized, placebo-controlled study of oral esterified estrogens at doses as low as 0.3 mg/day.13 The effects of the 0.025-mg/day dose on bone mineral density can also be compared with recent findings using the selective estrogen-receptor modulator raloxifene.21 Results at 24 months with the 0.025-mg/day dose (+2.31% relative to placebo) are similar to those obtained with 60 mg/day of raloxifene (+2.4% versus placebo) for total hip bone mineral density and are almost double that of raloxifene relative to placebo for the lumbar spine (+4.65% versus +2.4% with raloxifene).
Mean bone mineral density values at the spine and hip with the 0.025-mg dose were somewhat lower at 24 months than at 18 months. Although the differences were not statistically significant, this observation may have clinical implications. Nevertheless, most women had preservation of bone mass at the hip and spine at 18 and 24 months. Longer studies will be needed to determine whether these findings are the result of intrasubject variability or evidence that resumption of age-related bone loss, albeit at a slower rate, can be seen as early as 2 years after the start of ERT.
The positive effects of ERT on signs and symptoms of postmenopausal estrogen deficiency, and the long-term effects of reducing morbidity and mortality from cardiovascular disease and osteoporotic fractures, are well known.22–24 However, most postmenopausal women do not initiate ERT, and of those who do, most discontinue therapy within 6 months, depriving themselves of potential long-term benefits.25 Among the reasons cited for not starting or for discontinuing ERT are the risk of cancer, adverse events, and continued cyclic bleeding.26 The challenge is to overcome the objections of women to continuing long-term therapy with ERT. A low-dose transdermal E2 delivery system that prevents bone loss is an important option that may address the problem of long-term compliance by reducing dose-related adverse events. It remains to be shown that low-dose transdermal estrogen therapy also conveys the presumptive cardiovascular protection of higher-dose therapy.
The potential clinical implications of our results are that women may be offered higher doses of transdermal E2 early after surgical menopause to control symptoms and to provide initial protection against the cardiovascular and metabolic effects of estrogen deficiency. As symptoms decrease, the dose can be reduced progressively and substantially to maintain symptomatic relief while still providing protection against osteoporotic bone loss.
1. Eastell R. Treatment of postmenopausal osteoporosis. N Engl J Med 1998;338:736–45.
2. Greenspan SL, Maitland LA, Myers ER, Krasnow MB, Kido TH. Femoral bone loss progresses with age: A longitudinal study in women over age 65. J Bone Miner Res 1994;9:1959–65.
3. Raisz LG, Priestwood KM. Estrogen and the risk of fracture—new data, new questions. N Engl J Med 1998;334:767–8.
4. Riggs BL, Melton LJ III, O'Fallon WM. Drug therapy for vertebral fractures in osteoporosis: Evidence that decreases in bone turnover and increases in bone mass both determine antifracture efficacy. Bone 1996;18(suppl):197S–201S.
5. Garnero P, Sornay-Rendu E, Chapuy MC, Delmas PD. Increased bone turnover in late postmenopausal women is a major determinant of osteoporosis. J Bone Miner Res 1996;11:337–49.
6. Grodstein F, Stampfer MJ, Colditz GA, Willett WC, Manson JE, Joife M, et al. Postmenopausal hormone therapy and mortality. N Engl J Med 1997;336:1769–75.
7. Ettinger B, Friedman GD, Bush T, Quesenberry CP Jr. Reduced mortality associated with long-term postmenopausal estrogen therapy. Obstet Gynecol 1996;87:6–12.
8. Grady D, Rubin SM, Petitti DB, Fox CS, Black D, Ettinger B, et al. Hormone therapy to prevent disease and prolong life in postmenopausal women. Ann Intern Med 1992;117:1016–37.
9. Michaelsson K, Baron JA, Farahmand BY, Johnell O, Magnusson C, Persson P-G, et al. Hormone replacement therapy and risk of hip fracture: Population based case-control study. BMJ 1998;316:1858–63.
10. Horsman A, Jones M, Francis R, Nordin C. The effect of estrogen dose on postmenopausal bone loss. N Engl J Med 1983;309:1405–7.
11. Lindsay R, Hart DM, Clark DM. The minimum effective dose of estrogen for prevention of postmenopausal bone loss. Obstet Gynecol 1984;63:759–63.
12. Sharp CA, Evans SF, Risteli L, Risteli J, Worsfold M, Davie MJW. Effects of low- and conventional-dose transcutaneous HRT over 2 years on bone metabolism in younger and older postmenopausal women. Eur J Clin Invest 1996;26:763–71.
13. Genant HK, Lucas J, Weiss S, Akin M, Emkey R, McNaney-Flint H, et al. Low-dose esterified estrogen therapy: Effects on bone, plasma estradiol concentrations, endometrium, and lipid levels. Arch Intern Med 1997;157:2609–15.
14. Gordon SF, Thompson KA, Ruoff GE, Imig JR, Lane PJ, Schwenker CE, et al. Efficacy and safety of a seven-day, transdermal estradiol drug-delivery system: Comparison with conjugated estrogens and placebo. Int J Fertil 1995;40:126–34.
15. Garnero P, Grimaux M, Demiaux B, Preaudat C, Sequin P, Delmas PD. Measurement of serum osteocalcin with a human-specific two-site immunoradiometric assay. J Bone Miner Res 1992;7:1389–98.
16. Pratt DA, Daniloff Y, Duncan A, Robins SP. Automated analysis of the pyridinium crosslinks of collagen in tissue and urine using solid-phase extraction and reversed-phase high-performance liquid chromatography. Anal Biochem 1992;207:168–75.
17. Robins SP, Black D, Paterson CR, Reid DM, Duncan A, Seibel MJ. Evaluation of urinary hydroxypyridinium crosslink measurements as resorption markers in metabolic bone diseases. Eur J Clin Invest 1991;21:310–5.
18. Hochberg Y. A sharper Bonferroni procedure for multiple tests of significance. Biometrika 1988;75:800–2.
19. Field CS, Dry SJ, Wahner HW, Herrmann RR, Judd HL, Riggs BL. Preventive effects of transdermal 17 beta-estradiol on osteoporotic changes after surgical menopause: A two-year placebo-controlled trial. Am J Obstet Gynecol 1993;168:114–21.
20. Ettinger B. Use of low-dosage 17 beta-estradiol for the prevention of osteoporosis. Clin Ther 1993;15:950–62.
21. Delmas PD, Bjarnason NH, Mitlak BH, Ravoux AC, Shah AS, Huster WJ, et al. Effects of raloxifene on bone mineral density, serum cholesterol concentrations, and uterine endometrium in postmenopausal women. N Engl J Med 1997;337:1641–7.
22. Writing Group for the PEPI Trial. Effects of hormone therapy on bone mineral density: Results from the Postmenopausal Estrogen/Progestin Interventions (PEPI) trial. JAMA 1996;276:1389–96.
23. Schairer C, Adami H-O, Hoover R, Persson I. Cause-specific mortality in women receiving hormone replacement therapy. Epidemiology 1997;8:59–65.
24. Kiel DP, Felson DT, Anderson JJ, Wilson PWF, Moskowitz MA. Hip fracture and the use of estrogens in postmenopausal women: The Framingham Study. N Engl J Med 1987;317:1169–74.
25. Griffing GT, Allen SH. Estrogen replacement therapy at menopause. How benefits outweigh risks. Postgrad Med 1994;96:131–40.
26. Belchetz PE. Hormonal treatment of postmenopausal women. N Engl J Med 1994;330:1062–71.