Estrogen maintains bone density in postmenopausal women by inhibiting bone resorption,1–3 thus reducing the risk of osteoporotic fractures. Long-term therapy is required for maximal protection against osteoporotic fractures.4 To prevent the development of endometrial hyperplasia and carcinoma, progestogens are often given sequentially or cyclically with estrogens to women with intact uteri. Many women find the menseslike bleeding that accompanies sequential regimens unacceptable.
Hormone replacement therapy (HRT) regimens that increase the chances of women's adherence to therapy have the greatest potential for long-term benefits for their bones. Continuous, combined HRT (a daily dose of estrogen and progestogen) has been used to induce atrophic endometria and prevent bleeding.5,6 Eiken et al7 reported increased adherence to continuous, combined therapy compared with sequential therapy over 10 years. The most commonly used dose of estrone sulfate in Australia is 1.25 mg, shown by Harris et al8 and Genant et al9 to be the bone-sparing dose for spine and hip.
The primary objective of this study was to establish the optimum oral dose of micronized medroxyprogesterone acetate (Provera; Pharmacia & Upjohn, Kalamazoo, MI) given daily, in combination with a fixed dose of piperazine estrone sulfate (Ogen; Pharmacia & Upjohn), as HRT in postmenopausal women. The endpoints for acceptable therapy were cessation of vaginal bleeding, protection of endometrium determined by histopathologic examination, maintenance of acceptable blood lipid profile, maintenance of bone density, and improvement of vasomotor and other menopausal symptoms.
We report the results of bone density changes over the 2-year study. The results of bleeding patterns and endometrial changes10 and symptoms and side effects11 were already reported.
Subjects and Methods
The details of study design and methodology were reported in a previous paper.10 This multicenter, parallel-group, randomized, double-blind study compared three daily oral doses of medroxyprogesterone acetate tablets combined with a fixed daily dose of estrone sulfate as HRT. Nine centers took part in this clinical trial between 1990 and 1994. All eligible subjects received estrone sulfate, 1.25 mg, daily and were randomized to receive medroxyprogesterone acetate 2.5, 5, or 10 mg daily. Approval from the institutional ethics committees of each center was obtained, and subjects gave written informed consent before enrollment. Healthy women who were 1–10 years postmenopausal or women receiving HRT who had previously had amenorrhea for 12 months and had ceased HRT at least 6 weeks before entry were enrolled. Other entry criteria were FSH greater than 40 IU/L, intact uterus, and evidence of menopausal symptoms. The study was double-masked by encapsulation of medroxyprogesterone acetate tablets, but was open-label for estrone sulfate. Five hundred sixty-eight women were enrolled. Compliance with medication was checked using subject diaries and pill counts of returned medication at each 3-month follow-up visit.
Bone density was measured with dual-energy x-ray absorptiometry at entry and 12 and 24 months. Three machine types were used at the various centers, Lunar DPX (Lunar Corporation, Madison, WI), Hologic (Hologic Inc., Waltham, MA), and Norland (Norland Corporation, Fort Atkinson, WI). Quality assurance tests were run on each machine daily, and scanning of phantoms was done weekly at each center. Repeat bone-density measurements on subjects in each center were made on the same bone densitometer throughout the study. The two sites chosen for investigation were the lumbar spine and hip. Subjects were not masked to their bone-density results. There were no recommendations for dietary modifications other than a recommended dose of 1.5 g/day of calcium either by diet or supplements. Calcium intake was recorded on a frequency questionnaire of certain foods over a 1-week period (before visit) at baseline and during each 3-month follow-up visit. Exercise, alcohol intake, and smoking were not controlled for, and no recordings of those were made once treatment started. For analysis, the average bone density (g/cm2) of the second to fourth lumbar vertebrae was used for the lumbar spine, and femoral neck measurements were used for the hip. Bone-density measurements were standardized over the three machine types using the method described by Genant et al.12 Bone density was compared across treatment groups at baseline and 12 and 24 months in all analyzable subjects, who were those who met all the eligibility criteria, took at least 50% of their prescribed medication between each visit (by actual tablet-capsule count), and attended each of the scheduled visits within a 45-day window. Seventy-three percent took at least 80% of the study medications (74%, 71%, and 75% for the 2.5-, 5.0-, and 10-mg groups, respectively). Baseline data were analyzed on an intent-to-treat basis. Bone density changes were measured on analyzable subjects because an assessment of the different effects of dosages among subjects who adhered to treatment was desired. That approach could introduce bias because it does not adjust for differential adherence rates, but because the withdrawal rate was even across treatment groups (49, 54, and 49 subjects in the 2.5-, 5-, and 10-mg groups, respectively), that was not considered to be a problem.
Multivariate analyses (repeated measures models13) on absolute and percentage changes of bone density, at each site, were done with treatment group and duration (12 or 24 months) as predictors. Those were adjusted for covariates baseline bone density, age, number of years since menopause (less than or more than 2 years), body mass index (BMI), previous HRT, calcium intake, parity, smoking status, alcohol intake, and center. As the results for absolute changes, percentage changes, and changes in measurements before standardization for machine types were similar, only the percentage changes of machine-standardized results are reported. Type III sums of squares14 were used for the statistical tests, at a 5% level of significance.
Categoric variables were compared across the three treatment groups using χ2 tests for differences of proportions among the groups for nominal categories and Kendall's tau-b15 for ordinal categories. The Kruskal-Wallis test was used to compare the three treatment groups when the variables were not distributed normally; analysis of variance was used when they were.
Five hundred sixty-eight women were recruited for the 2-year study, of whom 566 were considered eligible for bone-density evaluation at baseline (188, 189, and 189 in the 2.5-, 5-, and 10-mg groups, respectively). Four hundred fourteen women completed the study. Eight subjects completed other aspects of the study but did not have their 24-month bone-density measurements.
One hundred forty-six women withdrew from the study, 77 for medical reasons and 69 for nonmedical reasons. Details on subjects who discontinued from the study have already been reported.10
Table 1 shows baseline characteristics of subjects by treatment group. The overall mean age was 52.8 years. No significant differences were noted in any baseline characteristics between the three groups. Approximately 98% of women were white. Over the 24 months, there was a significant increase (P < .001) in the mean calcium intake in the groups (300 mg, 236 mg, and 227 mg per day in the 2.5-, 5-, and 10-mg groups, respectively) with no significant differences between groups (P = .23).
In the study population, there was a significant difference from baseline to 24 months in lumbar spine (4.0% ± 0.27%) and femoral neck (3.2% ± 0.28%) bone density, with no significant differences between groups (Table 2). In the lumbar spine, all three groups had significant rises in mean bone density from baseline to month 12 and from month 12 to 24; at the femoral neck, a significant rise was seen only in the first 12 months. From baseline to month 24, there was an average increase in lumbar spine bone density of 3.8%, 4.9%, and 3.3% in the 2.5-, 5-, and 10-mg groups, respectively; at the femoral neck, the corresponding increases were 2.4%, 3.9%, and 3.2%. In the repeated-measures analysis, there was a significant treatment effect on percentage changes at 12 and 24 months (P = .03) at the lumbar spine (2.5 mg and 5 mg gave greater increases than 10 mg), but not at the femoral neck (P = .92). Subjects in the lowest tertile of baseline lumbar spine bone density (0.63–1.03 g/cm2) had an average increase of 6%; the middle tertile (1.03–1.16 g/cm2), 3.8%; and the upper tertile (1.16–1.79 g/cm2), 2.5%, over 24 months. Subjects in the lowest tertile of baseline femoral neck bone density (0.55–0.80 g/cm2) had an average increase of 4.9%, the middle tertile (0.80–0.90 g/cm2) 2.6%, and the upper tertile (0.90–1.29 g/cm2) 1.9%, over 24 months. Those trends are shown in Figures 1 and 2. At the femoral neck, the 2.5- and 10-mg groups showed a similar pattern as for lumbar spine bone density, whereas the 5-mg group showed little pattern of association (Figure 2).
In repeated-measures analysis, the percentage changes in bone density were related significantly to baseline bone density at both lumbar spine (P < .001) and femoral neck (P < .001) and to treatment duration at both sites (lumbar spine P < .001 and femoral neck P = .001). At the lumbar spine only, previous HRT (P = .008), BMI (P = .001), years since menopause (P = .027), and smoking (P < .001) also were significant predictors of bone-density changes. The changes at both sites were not related to age, parity, alcohol intake, or calcium intake. Center-to-center variation was corrected for (lumbar spine, P = .006, femoral neck, P < .001). At the lumbar spine, smaller bone-density increases were observed in subjects who had used HRT previously, had smaller BMIs, were smokers, or were within 2 years of menopause. Moderate smokers (1–10 cigarettes/day) showed smaller increases in lumbar spine bone density than nonsmokers; heavy smokers (more than 10 cigarettes/day) had even smaller increases than moderate smokers. Within-subject correlation between bone-density changes at 12 and 24 months was 0.63 in both sites.
This study showed the beneficial effect on bone density, in the spine and hip, of 1.25 mg of continuous daily estrone sulfate with 2.5, 5, or 10 mg medroxyprogesterone acetate over 2 years. The increase in the spine was significantly greater in the 2.5- and 5-mg groups than in the 10-mg group. The concern about whether 10 mg of medroxyprogesterone acetate had a negative impact on the estrogen-induced increases in spine bone density has to be raised. In a clinical setting, the importance of that finding must be weighed against the other attributes of progestin dose, including bleeding frequency and endometrial protection. Unlike norethisterone,16 medroxyprogesterone acetate does not prevent bone loss. This study did not show any advantage to increasing doses of medroxyprogesterone acetate on bone density.
The increase in bone density was greatest in the first 12 months of treatment. A similar initial significant rise in lumbar bone density was shown by other groups using unopposed estrogens or estrogens with various combinations of progestogens.17 Women who had used HRT recently gained bone density in the lumbar spines and hips at significantly lower rates compared with women who had not. It is reasonable to assume they had already gained bone with their previous treatment. Once estrogen therapy is withdrawn, bone loss ensues, and the rate of loss seems to be similar to that of untreated postmenopausal women.18 Therefore, to have maximum benefit from HRT, long-term estrogen therapy is desirable. Approximately 80% of women became amenorrheic by month 6 on those regimens,10 so it is expected that longer-term treatment would be more acceptable to women than sequential therapy, which has a high rate of withdrawal bleeding. In a 10-year study of 151 women, Eiken et al7 used oral HRT in continuous, combined, or sequential manner for two-thirds of the women, and the remaining one-third used a placebo. At the end of the study, 38% of those receiving continuous combined HRT, 22% of those receiving sequential HRT, and 49% of the untreated group remained on therapy. The HRT groups increased lumbar spine bone density by 14.5%; the placebo group decreased by 4.7%; femoral neck bone density was maintained in the HRT group; the placebo group decreased by 17%. They also showed continued increases in lumbar spines between 5 and 10 years of HRT use. Our study and the Postmenopausal Estrogen/Progestin Interventions study17 showed ongoing increases in lumbar spines after the first year of treatment. Although all HRT groups increased bone density significantly in both sites, the effect was greater in the spine than the hip. The bone remodeling rate is higher in the spine, and there are differential distributions of cortical and trabecular bone in the two regions, so the two sites have different rates of loss and gain.19
The factors that influenced the bone density changes at both sites were baseline bone density and treatment duration. Women with low bone density had significantly greater increases than those with normal bone density. Smokers did not increase their spine bone density as much as nonsmokers. Unlike previous reports,20,21 we were unable to show the effect of smoking on the hip. Higher BMI and greater number of years since menopause were associated with significantly greater increases in spine bone density.
There were two important limitations of this study. It does not address fully the issue of whether adding medroxyprogesterone acetate to 1.25 mg of estrone sulfate affects the skeletal response because there is no estrogen-only arm. Also, our conclusions apply only to one dose of estrone sulfate. At lower, perhaps suboptimal doses of estrogen, an effect of adding medroxyprogesterone acetate might be seen, as shown by Gallagher et al.22 Harris et al8 also showed prevention of spinal bone loss in postmenopausal women given 0.625 or 1.25 mg of cyclical estrone sulfate, but not with 0.3 mg, after 24 months of treatment.
Bone density of the spine and femoral neck increased significantly in all three treatment groups, greater in those with lower baseline bone density. The response was blunted at the spine in women who smoked. A slight but significantly smaller increase in spine bone density was noted in the 10-mg group compared with the others when important covariates had been corrected for.
1. Cauley JA, Seeley DO, 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.
2. Session DR, Kelly AC, Jewelewicz R. Current concepts in estrogen replacement therapy in the menopause. Fertil Steril 1993;59:277–84.
3. Lindsay R, Cosman F. Estrogen in prevention and treatment of osteoporosis. Ann N Y Acad Sci 1990;592:326–33.
4. Riggs BL, Melton LJ III. Involutional osteoporosis. N Engl J Med 1986;314:1676–86.
5. Magos AL, Brincat M, Studd JW, Wardle P, Schlesinger P, O'Dowd T. Amenorrhea and endometrial atrophy with continuous oral oestrogen and progestogen therapy in postmenopausal women. Obstet Gynecol 1985;65:496–9.
6. Weinstein L, Bewtra C, Gallagher JC. Evaluation of continuous low dose regimen of estrogen progestin for treatment of the menopausal patient. Am J Obstet Gynecol 1990;162:1534–42.
7. Eiken P, Kolthoff N, Nielsen SP. Effect of 10 years' hormone replacement therapy on bone mineral content in postmenopausal women. Bone 1996;19:191S–3S.
8. Harris ST, Genant HK, Baylink DJ, Gallagher JC, Karp SK, McConnell MA, et al. The effects of estrone (Ogen) on spinal bone density of postmenopausal women. Arch Intern Med 1991;151:1980–4.
9. Genant HK, Baylink D, Gallagher JC, Harris ST, Steiger P, Herber M. Effect of estrone sulphate on postmenopausal bone loss. Obstet Gynecol 1990;76:579–84.
10. Nand SL, Webster MA, Baber R, O'Connor V. Bleeding pattern and endometrial changes during continuous combined hormone replacement therapy. The Ogen/Provera Study Group. Obstet Gynecol 1998;91:678–84.
11. Nand SL, Webster MA, Baber R, Heller GZ. Menopausal symptom control and side effects on continuous estrone sulfate and three doses of medroxyprogesterone acetate. The Ogen-Provera Study Group. Climacteric 1998;1:1–8.
12. Genant HK, Grampp S, Gluer CC, Faulkner KG, Jergas M, Engelke K, et al. Universal standardization for dual X-ray absorptiometry: Patient and phantom cross-calibration results. J Bone Miner Res 1994;9:1503–14.
13. Goodman LA, Kruskal WH. Measures of association for cross-classification. IV. J Am Stat Assoc 1972;67:415–21.
14. Diggle PJ, Liang KY, Zeger SL. Analysis of longitudinal data. Oxford: Oxford University Press, 1994.
15. Milliken GA, Johnson DE. Analysis of messy data. Vol. I: Designed experiments. New York: Van Nostrand Reinhold Co, 1984.
16. Abdalla H, Hart DM, Lindsay R, Leggate I, Hooke A. Prevention of bone mineral loss in postmenopausal women by norethisterone. Obstet Gynecol 1985;6:789–92.
17. The Writing Group for the PEPI Trial. Effects of hormone therapy on bone mineral density. JAMA 1996;276:1389–96.
18. Christiansen C, Christiansen MS, McNair P, Hagen C, Stocklund KE, Transbol I. Prevention of early postmenopausal bone loss: Controlled 2-year study in 315 normal females. Eur J Clin Invest 1980;10:273–9.
19. Christiansen C, Riis BJ. Hormonal replacement therapy and the skeletal system. Maturitas 1990;12:247–57.
20. Jensen J, Christiansen C, Rotbro P. Cigarette smoking, serum estrogens, and bone loss during hormone replacement therapy early after menopause. N Engl J Med 1985;313:973–5.
21. Keil DP, Baron JA, Anderson JJ, Hannan MT, Felson DT. Smoking eliminates the protective effect of oral estrogens on the risk of hip fracture among women. Ann Intern Med 1992;116:716–21.
22. Gallagher CJ, Kable WT, Goldgar D. Effect of progestin therapy on cortical and trabecular bone: Comparison with estrogen. Am J Med 1991;90:171–8.
23. Cleveland WS. Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc 1979;74:829–36.