| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Bone Metabolism Unit, Creighton University School of Medicine (J.C.G., P.B.R., J.R.D.), and Department of Medicine, Creighton University (G.H.), Omaha, Nebraska 68131
Address all correspondence and requests for reprints to: Dr. Prema B. Rapuri, Bone Metabolism Unit, Creighton University School of Medicine, 601 North 30th Street, Room 6718, Omaha, Nebraska 68131. E-mail: thiyyari{at}creighton.edu.
Abstract
In a 5-yr randomized prospective study we examined the treatment effect of estrogen replacement therapy/hormone replacement therapy (ERT/HRT), calcitriol, ERT/HRT and calcitriol, or placebo for 3 yr and the effect of discontinuation of therapy for 2 more yr on bone mineral density (BMD), calciotropic hormones, markers of bone remodeling, and calcium absorption in 489 elderly women. The treatment phase of the study was double-blinded. After discontinuing therapy for 2 yr, there was rapid bone loss in all 3 treatment groups, and most of the decrease in BMD occurred in the first year.
In the ERT/HRT group, spine BMD increased 5.5% in yr 3, decreased 3.2% in yr 4, and decreased 0.7% in yr 5; femoral neck BMD increased 3.7% in yr 3, decreased 2.5% in yr 4, and decreased 0.4% in yr 5; total body BMD increased 2.1% in yr 3, decreased 1.4% in yr 4, and decreased 0.6% in yr 5. In the combination group, spine BMD increased 7.1% in yr 3, decreased 4.3% in yr 4, and decreased 0.3% in yr 5; femoral neck BMD increased 4.5% in yr 3, decreased 3.0% in yr 4, and decreased 0.01% in yr 5; total body BMD increased 2.2% in yr 3, decreased 1.5% in yr 4, and decreased 0.6% in yr 5. In the calcitriol group, spine BMD increased 1.8% in yr 3, decreased 1.8% in yr 4, and showed no change in yr 5; femoral neck BMD increased 0.2% in yr 3, decreased 0.2% in yr 4, and decreased 0.6% in yr 5; total body BMD decreased 0.4% in yr 3, decreased 0.6% in yr 4, and decreased 0.4% in yr 5.
Compared with placebo, all treated groups at yr 5 had significantly higher total body BMD; only the combination group had significantly higher spine BMD (3.4%; P < 0.001) and total hip BMD (2.4%; P < 0.01.) compared with the placebo group. Compared with baseline, only spine BMD in the combination group was significantly higher (2.6%; P < 0.001) at yr 5.
The increase in calcium absorption and the decrease in serum PTH levels in the calcitriol groups were reversed after discontinuation of treatment, and the decrease in bone markers was reversed in the hormone-treated groups. These results suggest that discontinuation of ERT/HRT and/or calcitriol therapy in elderly women leads to a decrease in much of the BMD gained on treatment; however, in the combination group there was a statistically significant residual effect on spine BMD.
THERE ARE CONFLICTING findings on the effect of discontinuing osteoporosis therapy on bone mass. Both accelerated (1, 2, 3, 4) and normal rates of bone loss (5, 6) have been reported after hormone replacement therapy (HRT) or estrogen replacement therapy (ERT) withdrawal in early postmenopausal women. Eiken et al. (7) observed higher rate of bone loss at spine after HRT withdrawal compared with nontreated group, but the rate of bone loss at the forearm was same as that in the nontreated group. There has been one report (8) on the effect of withdrawing alendronate therapy, and bone loss was not accelerated. In a recent study of elderly subjects (9), the effect of withdrawing daily vitamin D (600 IU) and calcium (600 mg) was followed for 2 yr. All of the gain in spine and femoral neck bone density disappeared within 2 yr of discontinuation. Shiraki et al. (10) reported rapid bone loss and development of new fractures after discontinuation of treatment of postmenopausal women with 1
-hydroxycholecalciferol.
In a randomized, double-blind, placebo-controlled trial of 3 yr, we recently reported increases in bone mineral density (BMD) during ERT/HRT, on a combination of ERT/HRT and calcitriol, and on calcitriol alone (11) in elderly women. In a planned extension of this study, we followed the effect of discontinuation of treatment on BMD, bone markers, calcium absorption, and serum PTH for 2 yr.
Subjects and Methods
Four hundred eighty-nine elderly women, aged 65–77 yr, were enrolled into a prospective double-blind, randomized, placebo-controlled clinical trial (STOP IT) to test the efficacy of conjugated equine estrogens and calcitriol, given individually or in combination, on bone loss. Women were excluded if they had severe chronic illness (chronic liver or kidney disease, severe chronic obstructive pulmonary disease, severe rheumatoid arthritis, or serious heart failure), had primary hyperparathyroidism or active renal stone disease, or had taken certain medications, such as bisphosphonates, anticonvulsants, estrogen, fluoride, or thiazide diuretics, in the previous 6 months. There were 470 white, 13 black, 4 Hispanic, 1 Asian, and 1 woman of mixed race.
In the original treatment phase of the trial, women were randomized to one of four groups: conjugated equine estrogens [0.625 mg/daily; Premarin; medroxyprogesterone acetate (2.5 mg/daily; Provera) was added if the woman had a uterus], calcitriol (0.25 µg, twice daily), and the combination of both or placebo. At the end of 3 yr of treatment, therapy was discontinued, and the women were asked to volunteer for 2 yr of follow-up. One woman who developed Paget’s disease was excluded from the analyses. Of the remaining 488 women, 415 came in for final visit at 3 yr, and of this group 337 were adherent to the medication. If at the end of 3 yr a woman’s BMD value fell below –2.5 SD for her age a recommendation was made for her to receive therapy. Women were also given the option to take calcium and vitamin D supplements. Sixty-three women were placed on estrogen or alendronate by their private physicians after 3 yr of study. Two hundred thirty-three women completed a fourth year of no treatment, and 178 completed a fifth year of no treatment. The distribution of subjects into different treatment groups at the end of the treatment period (yr 3) and the follow-up phases (yr 4 and 5) is shown in Fig. 1
. The analyses were however performed only on the 178 women who came for the fifth year of the study. The Creighton University human institutional review board approved both phases of the trial, and all patients signed an informed consent form.
|
The same questionnaires were used throughout the 5 yr for the medical history, medication record, hospitalizations and illnesses, dietary records, smoking history, alcohol use, and physical activity record. The 7-d dietary records (Food Processor II Plus, version 5.1, Esha Research, Salem, OR) were obtained at baseline and 36 and 60 months. Smoking status and alcohol use were recorded at baseline and 36, 48, and 60 months. Height (Harpenden stapediometer) and weight (digital scale) were measured regularly. The BMD of the spine (L2–L4), proximal femur (neck, trochanter, and total hip), and total body was measured by dual energy absorptiometry (model DPX-L, Lunar Corp., Madison, WI) at 6-month intervals during the treatment phase and at 12-month intervals in the nontreatment phase. The hip and spine scans were performed in duplicate, and the mean was used for the analysis. The coefficients of variation for BMD measurements were 1.8% for femoral neck, 2.4% for spine, and 0.7% for total body. Further, an independent phantom was scanned every week as a quality control measure for the scanner. At baseline and yr 3, 4, and 5, blood and 24-h urine were collected while fasting for measurements of serum 25-hydroxyvitamin D (25OHD; only baseline and yr 3 and 5), PTH, serum osteocalcin, urinary creatinine and urinary N-telopeptides. Serum 25OHD was assayed with a competitive protein binding assay (12) after prepurification of serum on Sep-Pak cartridges (Waters Corp., Milford, MA) (13). The limit of detection for the assay was 12.5 mmol/liter (5 µg/liter), and the interassay coefficient of variation was 5%. Serum intact PTH was measured with an Allegro immunoradiometric assay (Nichols Institute Diagnostics, San Juan Capistrano, CA). The interassay coefficient of variation was 5%, and the limit of detection for the assay was 1 ng/liter (1 pg/ml). Serum concentrations of osteocalcin were determined by RIA (INCSTAR Corp., Stillwater, MN). The limit of detection was 0.78 µg/liter, and the interassay coefficient of variation was 5%. Urinary collagen cross-links were measured by ELISA (Osteomark International, Seattle, WA) as N-telopeptides, a marker for bone type I collagen. The lower limit of detection was 20 nmol bone collagen equivalents, and the interassay coefficient of variation was 6%. The data are expressed as nanomoles of bone collagen equivalents per millimoles of creatinine. Urinary creatinine was measured on a Nova Nucleus (Nova Nucleus Chemistry Analyzer, Waltham, MA). Calcium absorption was measured in a fasting state at baseline, at the end of treatment phase (36 month), and at the end of the follow-up period (60 months). 45Ca (18.5 x 104 Bq, 5 µCi; Amersham Pharmacia Biotech, Arlington Heights, IL) in 100 mg CaCl2 carrier was administered in a total volume of 250 ml distilled water (14). Blood samples were collected at 1, 2, and 3 h after the oral dose. 45Ca activity was counted in 2 ml serum with a 1900 CA Tri-Carb liquid scintillation analyzer (Packard Instruments, Meriden, CT). Calcium absorption was expressed as percentage of the actual dose per liter of blood after 3 h and corrected for body size.
Statistical methods
At baseline, subjects were assigned to four treatment groups using simple randomization stratified on hysterectomy status. The design was a four-group rather than a factorial one because it was a comparison between placebo and the other three treatment groups. The statistical analyses were accomplished by SAS version 8.2 (SAS Instruments, Inc., Cary, NC). The percentage change in the variables was calculated as the difference between follow-up measurements (yr 3, 4, or 5) and baseline or between the follow-up (yr 4 and 5) and the 3 yr measurement divided by baseline measurements, and multiplied by 100. The hypothesis for the discontinuation phase of the study was that bone loss would occur in all groups after discontinuation of treatment, and that bone loss would occur most rapidly in the calcitriol group, followed by the estrogen group, then the combination group, and slowest in the placebo group.
Primary analyses
The percentage change in BMD at multiple skeletal sites, calcium absorption, serum 25OHD, serum PTH, osteocalcin, and urinary N-telopeptides/creatinine ratio were compared in the four treatment groups by a general linear mixed effects model with repeated measures on factor time, as implemented in the SAS version 8.2 procedure MIXED in the 178 women who completed the fifth year of the study. Potential first order interactions were also included in the initial model and were kept in it if they were significant in the type 3 F test. The pairwise comparisons were made through the least squares means option in procedure MIXED. The data were adjusted for age, height, weight, hysterectomy status, smoking, and alcohol consumption status. The percentage changes in BMD presented for yr 3, 4, and 5 in the tables, figures, and the text are raw unadjusted means. The statistical significance, however, was calculated based on the adjusted data. The significance level was kept at 5% in all tests, including the overall level when multiple pairwise comparisons were performed.
Secondary analyses
The above-mentioned analyses were also repeated in the 233 women who came for the fourth year of study using a modified intent to treat analysis. The missing values for 55 women in the fifth year were imputed according to a last observation carried forward algorithm.
A further analyses examined the changes in BMD when results were combined for the two hormone groups. The percentage change in BMD at multiple skeletal sites, calcium absorption, serum 25OHD, serum PTH, osteocalcin, and urinary N-telopeptides/creatinine ratio were compared in the placebo, calcitriol, and combined hormone groups.
Results
Primary analyses
Characteristics of the study population.
The characteristics of the study population at baseline and follow-up are given in Table 1. There were no significant differences in any of the characteristics between the women assigned to various treatment groups at baseline and yr 5. Although there was an increase in the mean dietary calcium intake in all treatment groups at yr 5 compared with baseline, it was not significantly different from respective baseline measures.
|
and Figs. 2 and 3).Spine. In the placebo group, there were no statistically significant changes in mean spine BMD compared with baseline during the treatment and follow-up phases.
|
|
|
On the combination of ERT/HRT and calcitriol, spine BMD increased 7.1% at yr 3; after discontinuation it decreased by 4.3% (P < 0.001) during yr 4 and by another 0.3% in year 5. Mean spine BMD at the end of yr 5 was significantly higher than that with placebo (2.6% vs. –0.8%; P < 0.01) and also was significantly (P < 0.01) higher than the baseline.
In the calcitriol group, spine BMD increased 1.8% from baseline at yr 3 (P < 0.01), decreased 1.8% in yr 4, and showed no further change in yr 5. At yr 5 the mean change in spine BMD was not significantly higher than that the placebo or baseline value.
Total body. In the placebo group, total body BMD decreased significantly by 2.3% (P < 0.001) during the first 3 yr and continued to decrease during the follow-up phase.
During ERT/HRT, total body BMD increased 2.1% at yr 3; after discontinuation it decreased by 1.4% (P < 0.001) during yr 4 and 0.6% in yr 5. Mean total body BMD at the end of yr 5 (0.1%) was significantly higher (P < 0.01) than that with placebo (-2.8%), but was not significantly different compared with baseline.
With the combination of ERT/HRT and calcitriol, total body BMD increased by 2.2% at 3 yr (P < 0.0001); after discontinuation, total body BMD decreased by 1.5% in yr 4 and 0.6% in yr 5. At the end of yr 5, the mean change in total body BMD was significantly (P < 0.05) higher than that with placebo, but was not different from baseline.
In the calcitriol group, total body BMD decreased –0.4% at yr 3; after discontinuation it decreased –0.6% in yr 4 and –0.4% in yr 5. The mean change in total body BMD was significantly (P < 0.05) higher than that with placebo at yr 5 (-1.5% vs. -2.8%).
Total femur. In the placebo group, total hip BMD decreased significantly during the first 3 yr and continued to decrease during the follow-up phase.
In the ERT/HRT group, total hip BMD increased significantly by 3.6% at yr 3 (P < 0.0001), After discontinuation of treatment, BMD decreased 3.3% in yr 4 and 0.7%, in yr 5. The mean change in BMD at yr 5 was not significantly different from that with placebo or at baseline.
In the combination group, total hip BMD increased by 5.4% at yr 3 and decreased by 3.6% in yr 4 and by 1.0% by yr 5. At the end of yr 5, the mean change in total hip BMD in the combination group was significantly higher than that in the placebo group (0.8% vs. –1.5%; P < 0.05), but not compared with baseline.
In the calcitriol group, total hip BMD increased 0.3% at yr 3, was unchanged in yr 4, and decreased 0.5% in yr 5. Compared with placebo and baseline values, the changes at yr 5 were not significant.
The changes in femoral neck and trochanter were similar to those in total hip, except that bone loss from the trochanter on ERT/HRT was marked. At the end of yr 5, the BMD of femoral neck and trochanter was not significantly different from the placebo and baseline values in any treatment group.
Calcium absorption and serum 25OHD, PTH, and bone markers (Table 3 and Figs. 4 and 5).
At the end of 3 yr there was a statistically significant (P < 0.0001) decrease in urinary N-telopeptides and serum osteocalcin in the hormone-treated groups (Table 3 and Fig. 5). After discontinuation of therapy, both serum osteocalcin and urinary N-telopeptides returned to pretreatment values. At yr 5, serum osteocalcin was not significantly different from the placebo value, but urinary telopeptides were significantly higher than those in the placebo group in the combination group (Table 3 and Fig. 5).
|
|
|
and Fig. 4). At the end of 3 yr, there was an increase in mean calcium absorption over baseline in the two calcitriol groups. After discontinuation of therapy, calcium absorption in the two calcitriol arms returned to pretreatment values and was not significantly different from the placebo value at yr 5 (Table 3 and Fig. 4). Serum 25OHD levels decreased significantly over baseline in all treatment groups at yr 3, and at the end of discontinuation phase they returned to pretreatment levels and were not significantly different among the groups (Table 3 and Fig. 4). Secondary analyses
When the analyses was repeated by including the data for the 55 women who came for the 4 yr of study, the overall results remained similar to those of the primary analyses (data not presented).
When the two hormone groups were combined (data not presented), spine BMD increased 6.3% from baseline at yr 3 (P < 0.0001); after discontinuation it decreased by 3.8% in yr 4 and by 0.5% in yr 5. At the end of yr 5, the mean change in spine BMD in the combined hormone groups was significantly (P < 0.05) higher than that in the placebo group (2.1% vs. –0.80%) and the corresponding baseline measure (P < 0.05). Similar observations were made with regard to total body and total femur BMD, where discontinuation of treatment resulted in a rapid loss of bone (data not presented). The mean change in total body BMD at yr 5 remained significantly (P < 0.001) higher than that in the placebo group, but not over the baseline value. The mean change in total femur BMD at yr 5 was not different from the placebo or baseline values.
Discussion
In the present study we examined for the first time in elderly women the effect of discontinuation of hormone replacement therapy (ERT/HRT), calcitriol therapy, or the combination of both on BMD at several skeletal sites and on various calciotropic indexes. Discontinuation of ERT/HRT and the combination treatment (ERT/HRT plus calcitriol) resulted in rapid bone loss at all skeletal sites, and much of the loss occurred during the first year with very little change in the second year. Spine BMD in the estrogen-treated groups remained higher than baseline, but was only statistically significant on the combination treatment, but total body BMD remained significantly higher than placebo in all treatment groups, which is consistent with preservation of cortical bone. The increase in urinary N-telopeptides after discontinuation of all three therapies and the increase in serum osteocalcin after discontinuation of estrogen treatment indicate that the decrease in BMD was due to the increase in bone remodeling as a result of uncoupling of bone remodeling; however, resorption exceeds formation, which is usually seen in postmenopausal women. The decreased calcium absorption and increased serum PTH after discontinuation of calcitriol therapy would also have contributed to the increased bone loss.
There have been differences in the pattern of bone loss reported after estrogen withdrawal. Several studies reported accelerated bone loss after discontinuation of estrogen. Lindsay et al. (1), observed that 14 oophorectomized women treated with estrogen for 4 yr lost all of the bone gained in metacarpal over the next 4 yr. Davis et al. (15) also reported an increased rate of bone loss at calcaneus and radial sites in Japanese-American late postmenopausal women who stopped estrogen treatment compared with nonusers. In two retrospective studies, Saarto et al. (16) studied the impact of previous HRT use on bone loss in postmenopausal women with breast cancer. They reported higher bone loss at the spine and femoral neck in recent HRT users compared with that in non-HRT users at the end of 5 yr of follow-up. Tremollieres et al. (4) measured changes in vertebral BMD after cessation of HRT in early postmenopausal women, suggesting accelerated bone loss in the first 2 yr similar to that seen in the first 2 yr in untreated women. In contrast, normal rates of bone loss have been reported after discontinuation of estrogen. Christiansen et al. (5) reported that the annual reduction in radial bone mass after hormone withdrawal was identical to that in the placebo group. More recently, Greendale et al. (6) reported that discontinuation of HRT in early postmenopausal women resulted in the same decrease in BMD of the hip and spine as that seen in the placebo group. However, it was reported in an earlier paper from the same study that women in the placebo group who previously used hormones before enrollment in the study lost BMD more rapidly at the spine (17); the reason for the discrepancy in these results is not clear.
In the present study the mean BMD at yr 5 at all sites was 1.2–3.5% higher in the hormone-treated groups and 0.3–1.4% higher in the calcitriol-treated group compared with that in the placebo group. The residual effect on BMD was always larger in the combination treatment group. Other studies have seen small residual effects on BMD. In younger postmenopausal women treated with HRT for 3 yr, radial BMD remained 8% higher in women previously treated with HRT compared with women who were taking placebo (5). In the French study (4), women had significantly higher spine BMD 3 yr after cessation of HRT treatment than at baseline. Even in the Rancho Bernardo study, Schneider et al. (18) reported that past estrogen use whether started earlier (before age 60 yr) or later (after age 60 yr) was associated with higher BMD than that in women who never used estrogen. This effect may be clinically important, because some studies (19, 20, 21, 22), but not all (23), have shown a reduction in fractures in women who were past users of estrogen. Further, in the Framingham study past estrogen use was shown to offer some degree of protection against hip fracture (24). Michaelsson et al. (25) also found that fracture protection decreased according to the time since last estrogen use, although it was present 5 yr after ERT/HRT had been discontinued. Probably the length of treatment with therapy is the most important determinant of long-term efficacy along with the age at which treatment is started. Protection against fracture may be due to a combination of higher BMD and/or an effect on bone quality.
The effect of discontinuation of calcitriol therapy on BMD has not been widely studied. Shiraki et al. (10) using an analog of vitamin D, 1-hydroxyvitamin D3, reported a rapid decrease in radial mineral content after its discontinuation and development of new fractures. In another recent study Dawson-Hughes et al. (9) reported that the effect of 3 yr of treatment using vitamin D and calcium supplements on spine, femoral neck, and total body was lost within 2 yr of discontinuation in women. However, they did observe some lasting beneficial effect on total body BMD in men.
The effect of estrogen withdrawal on bone markers resulted in a return to baseline of both urine N-telopeptides and serum osteocalcin, confirming that there was an increase in bone resorption. Thomsen et al. (26) also observed that the withdrawal of ERT in early postmenopausal women resulted in reversal of decreased bone turnover markers toward the baseline values within 3 months. Calcitriol withdrawal also led to an increase in urinary N-telopeptides. In addition, serum PTH that had been suppressed by calcitriol therapy returned to baseline, and calcium absorption that increased during calcitriol treatment returned to baseline.
In conclusion, our results suggest that withdrawal of ERT/HRT and/or calcitriol treatment in late postmenopausal women results in increased bone resorption and accelerated bone loss. Most of the loss occurs in the first year of withdrawal of therapy. However, even 2 yr after discontinuation of treatment there was a small residual effect on bone mass, and this was greater in the combination treatment group.
Acknowledgments
We thank Karen A. Rafferty for her help with food dairy data collection and analysis. We also thank Kurt E. Balhorn for the laboratory analysis.
Footnotes
This work was supported by NIH Research Grants UO1-AG-10373 and RO1-AG-10358. These data were presented as an abstract at the 23rd Annual Meeting of the American Society for Bone and Mineral Research, Phoenix, Arizona, October 12–16, 2001.
Abbreviations: BMD, Bone mineral density; ERT, estrogen replacement therapy; HRT, hormone replacement therapy; 25OHD, 25-hydroxyvitamin D.
Received May 9, 2002.
Accepted July 31, 2002.
References
hydroxycholecalciferol (1-OHD3) and 1,24 dihydroxycholecalciferol (1,24(OH)2D3). Endocrinol Jpn 32:305–315[Medline]
This article has been cited by other articles:
 |
|
 |
|
  E. M. Lewiecki Denosumab in postmenopausal osteoporosis: what the clinician needs to know Therapeutic Advances in Musculoskeletal Diseases, October 1, 2009; 1(1): 13 - 26. [Abstract] [PDF]
|
|
|
|
 |
|
 A. Stern, G. A Laughlin, J. Bergstrom, and E. Barrett-Connor The sex-specific association of serum osteoprotegerin and receptor activator of nuclear factor {kappa}B legend with bone mineral density in older adults: the Rancho Bernardo Study Eur. J. Endocrinol., May 1, 2007; 156(5): 555 - 562. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. G. Bone, D. Hosking, J.-P. Devogelaer, J. R. Tucci, R. D. Emkey, R. P. Tonino, J. A. Rodriguez-Portales, R. W. Downs, J. Gupta, A. C. Santora, et al. Ten Years' Experience with Alendronate for Osteoporosis in Postmenopausal Women N. Engl. J. Med., March 18, 2004; 350(12): 1189 - 1199. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. H. Amin, C. L. Kuhle, and L. A. Fitzpatrick Comprehensive Evaluation of the Older Woman Mayo Clin. Proc., September 1, 2003; 78(9): 1157 - 1185. [Abstract] [PDF]
|
|
|
|
 |
|
 Changes in Bone Density After Stopping HRT Journal Watch Dermatology, December 31, 2002; 2002(1231): 10 - 10. [Full Text]
|
|
|
|
 |
|
 Changes in Bone Density After Stopping HT Journal Watch Women's Health, December 17, 2002; 2002(1217): 5 - 5. [Full Text]
|
|
|
|
 |
|
 Changes in Bone Density After Stopping HRT Journal Watch (General), December 10, 2002; 2002(1210): 1 - 1. [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
–––
------
——
, P < 0.05, compared with placebo group.
