In a randomized, double blind, clinical prospective trial comprising 35 women treated with either hormone replacement therapy (HRT) (cyclic estradiol/norethisterone acetate) or placebo we performed histomorphometric studies on paired bone biopsies obtained before and after 2 years of treatment. Untreated women developed a progressively more negative balance at individual bone multicellular units (BMUs) (i.e., wall thickness-erosion depth) (2.2 ± 1.7 μm vs. −5.7 ± 1.4 μm; p < 0.01), while women on HRT displayed preservation of bone balance (2.4 ± 2.4 μm vs. 2.5 ± 2.5 μm; NS). No significant differences in wall thickness between the two groups were demonstrable, but the untreated women developed a pronounced increase in erosion depth over 2 years (46.9 ± 1.8 μm vs. 52.0 ± 1.9 μm; p < 0.05), while the HRT group revealed no change (47.8 ± 2.7 μm vs. 44.6 ± 1.7 μm; NS). Furthermore, the placebo group displayed an increased osteoclastic erosion depth (17.8 ± 1.6 μm vs. 25.0 ± 1.7 μm; p < 0.001), compared with unchanged values in the HRT group (20.0 ± 1.6 μm vs. 16.9 ± 1.4 μm/day; NS). While the placebo group revealed a slight increase in volume referent resorption rate (35 ± 8% vs. 38 ± 8%; NS) the HRT group revealed a pronounced decrease (46 ± 8% vs. 28 ± 5%; p < 0.05). No significant changes in marrow star volume (an index of trabecular perforations) were demonstrable in either group. Our results demonstrate that bone remodeling in early postmenopausal women is characterized by progressive osteoclastic hyperactivity, which is reduced by cyclic HRT. This reduction of resorptive activity at the BMU level after HRT seems to precede the reduction in activation frequency demonstrated in previous studies on older postmenopausal women.
A multitude of studies have demonstrated that hormone replacement therapy (HRT) is an effective prophylaxis against postmenopausal bone loss.(1–4) The underlying mechanisms mediating its protective effect on bone are, however, still subject to debate, and only a few studies have analyzed the effects of HRT on bone remodeling by histomorphometry. In an earlier study on osteoporotic women we reported that the main effects of cyclical HRT on bone remodeling was a 50% reduction in turnover as reflected in a 50% decrease in activation frequency.(5) No significant changes in bone multicellular unit (BMU) balance were demonstrable in the HRT group. Women with senile osteoporosis are, however, characterized by osteoblastic insufficiency(6,7) and may therefore not display possible effects of HRT on osteoblast function. We therefore wanted to investigate whether the alterations in bone remodeling demonstrated in old osteoporotic women were different from the changes seen in younger women.
MATERIALS AND METHODS
Fifty women aged 45–55 years were recruited via the central registry after approval from the regional ethical committee. The women were healthy, had no diseases, and did not take any medication with known influence on calcium metabolism. Early menopause was defined as cessation of menstrual bleeding within 6–24 months prior to participation in the study and corresponding high levels of follicle-stimulating hormone in serum. They accepted to participate in a prospective, randomized double blind study, comparing placebo and a cyclic HRT regimen (Trisequence; NOVO-Nordisk A/S, Copenhagen, Denmark). Prior to treatment, the women underwent a transiliac bone biopsy after tetracycline double labeling (tetracycline, 10 day labeling interval). After 2 years, a second bone biopsy was taken from the opposite side after renewed tetracycline labeling.
The biopsies were embedded in methylmethacrylate, sectioned, and stained with Goldner-Trichrome for histomorphometric analysis. Tetracycline based indices were assessed in unstained sections. Blinded investigators performed all measurements and all calculations of histomorphometric indices were finished before the treatment code was broken. The histomorphometric indices were assessed according to American Society of Bone and Mineral Research guidelines,(8) and further indices pertaining to bone resorption and bone balance were assessed according to the methods of Eriksen et al.(9) as were the reconstructed remodeling sequences.
The calculation of BMU-referent and volume referent balance of cancellous bone was based on measurements of mean wall thickness and erosion depth. The first was measured according to standard methodology,(8) the latter measured by counting number of lamellae eroded away at resorptive sites.(9) BMU-referent balance was taken as the difference between mean wall thickness and erosion depth.(9) Volume referent indices of resorption and formation were calculated as previously reported as the product of surface, erosion, or formation rates and surface-to-volume ratios.(9) These calculations demand the presence of a steady state between resorption and formation.(9)
Marrow star volume, which reflects the degree of trabecular disintegration caused by perforating resorption, was measured according to the method of Vesterby et al.(10)
In 17 women from the HRT group and 18 from the placebo group, we succeeded in getting evaluable pre- and post-treatment biopsies. The subsequent analysis is based on these biopsies only.
Differences between pre- and post-treatment results and differences between groups were tested using Student's t-test.
Primary histomorphometric indices
The primary indices recorded in the two groups before and after 2 years of treatment with either HRT or placebo are shown in Table 1. The two groups had similar baseline values. Erosion surface and osteoclastic erosion surface (an indirect index of osteoclast number) increased significantly in the placebo group (p < 0.01), while no significant differences over time were seen in the HRT group. The increase in erosion surface in placebo-treated individuals was paralleled by an increase in final and osteoclastic erosion depth (p < 0.01), while erosion depth remained unchanged and osteoclastic erosion depth was reduced in the HRT group. Wall thickness remained unchanged over time in both groups.
Table Table 1.. Primary Histomorphometric Indices (Mean ± SEM) in 17 Women Treated with HRT and 18 Women Treated with Placebo, in Whom Both Pre- and Posttreatment Biopsies Were Obtained
The derived parameters (Table 2) revealed more pronounced differences between women on HRT and women treated with placebo (pretreatment values (mentioned first) compared with post-treatment values). No significant differences with respect to baseline indices obtained in the two groups were demonstrable. Osteoclastic resorption rate was reduced in the HRT group. Untreated women developed a progressively more negative bone balance (2.2 ± 1.7 μm vs. −5.7 ± 1.4 μm; p < 0.01), while women on HRT displayed preservation of bone balance (2.4 ± 2.4 μm vs. 2.5 ± 2.5 μm; NS). No significant differences in wall thickness between the two groups were demonstrable. The volume-based remodeling indices corroborated these changes in BMU activity. The volume referent bone balance also became negative in the placebo group (0 ± 2% vs. –5 ± 2%; p < 0.05), while it remained unchanged in the HRT group (0 ± 2% vs. 1.4 ± 2%; NS). Both groups displayed a nonsignificant reduction in volume referent formation rates. However, while the placebo group revealed a slight increase in resorption rate (35 ± 8% vs. 38 ± 8%; NS) the HRT group revealed a pronounced decrease (46 ± 8% vs. 28 ± 5%; p < 0.05), which explains the differences in bone balance detected. Activation of new remodeling cycles was reduced by 8% in the HRT group versus 5% in the placebo group. Resorption and formation periods, as well as mineralization lag time, remained unchanged in both groups.
Table Table 2.. Key Remodeling Indices in 17 Women treated with HRT and 18 Women Treated with Placebo for 2 Years
No significant changes in marrow star volume (an index of trabecular perforations) were demonstrable in either group (Table 2).
Reconstructed remodeling sequences
The reconstructed remodeling sequences at individual BMUs further highlight the differences in balance in the two groups. While the untreated women developed a negative balance, the women treated with HRT revealed preservation of balance between resorption and formation (Fig. 1). The reconstructed sequences also revealed similar initial mineralization lag time, resorption, formation, and remodeling periods in the two groups.
Our data show distinctly different cancellous bone remodeling sequences in early postmenopausal women, depending on whether they are treated with hormones or left untreated. Our findings support the notion that estrogen exerts its action on bone remodeling in early postmenopausal women via its antiosteoclastic effects.
Our primary histomorphometric indices are similar to those reported by Recker et al.(11) The same study reported variations in volume based bone formation rate (BFR/BV) from 0.7% to 28%/year. Our values were somewhat higher (mean 37%), but this discrepancy is mainly explained by the fact that Recker et al. only used a double-labeled surface to calculate BFR/BV. We used both double and single labels for our calculations of double-labeled surface.
The natural history of bone remodeling in early menopause is displayed in the placebo group. The results obtained show a progression toward an ever more negative bone balance, at least over the 2-year period in this study. These women are characterized by continued osteoclastic hyperactivity as reflected in the high osteoclastic erosion rate and increased erosion surface and erosion depth. Thus, our observations corroborate the hypothesis on “killer osteoclasts” being responsible for postmenopausal bone loss, initially put forward by Parfitt.(12) The findings are, however, also consistent with the concept that estrogen deficiency leads to delayed osteoclast apoptosis.(13,14) Delayed apoptosis would cause increased life span for osteoclasts and consequently increases resorptive activity.
While the estrogen-deficient placebo group exhibited a progressive increase in erosion depth over the observation period, the women treated with HRT showed no further increase in resorptive activity, as reflected in the unchanged erosion depth. These changes were demonstrable at both the BMU and volume level. The main effect of HRT seems to be a direct inhibition of the osteoclastic hyperactivity, as reflected in the pronounced reduction in osteoclastic resorption rate and prevention of deepening of resorption lacunae. Our data suggest that the effects on osteoclastic resorption seem to precede the antiactivation effects of HRT. In a previous study,(5) we reported a 50% reduction in activation frequency, but in the present study only a minor reduction in activation frequency was demonstrable in the HRT group (see Table 2). Elias et al. reported similar effects in early postmenopausal women.(15) Despite using both calcium kinetic and histomorphometric methods, they were unable to demonstrate significant changes in bone balance, despite bone loss being detectable on repeated BMD measurements. Our results corroborate these findings.
HRT did not seem to affect osteoblastic matrix formation. Wall thickness did not change significantly in either group, and the tetracycline-based indices reflecting osteoblastic synthesis rates (MAR, BFR/BS, and BFR/BV also remained similar in both groups after treatment. Thus, no signs of HRT-induced osteoblastic stimulation were demonstrable. We were also unable to demonstrate any significant reduction in osteoblastic activity over time in the placebo group. This is in apparent contrast to some aspects of a recent paper on a unitary model for involutional osteoporosis, where reduced osteoblastic activity causing decreased bone formation was cited as an important factor behind the remodeling imbalance occurring during early menopause.(16) A relative osteoblastic insufficiency was, however, present in the placebo group, because osteoblastic bone formation was unable to keep up with the increase in bone resorption.
Despite the pronounced differences in remodeling indices demonstrated, we were unable to demonstrate significant differences in bone structure, as reflected in marrow star volume. As seen in Table 2, this index displays a huge interindividual variation. This fact combined with the relatively short observation period for the detection of structural changes may explain the absence of significant changes. Vedi et al. were also unable to demonstrate significant changes in cancellous bone structure after 2 years of HRT using three different structural indices.(17) In a recent study by Boyce et al. on changes in bone structure in ovariectomized pigs, the marrow star volume was shown to be relatively insensitive to the pronounced increase in trabecular perforations demonstrated with other techniques.(18)
In conclusion, our results demonstrate that bone remodeling in the estrogen-deficient early postmenopausal woman is characterized by progressive osteoclastic hyperactivity. Cyclic HRT is able to preserve bone balance by reducing osteoclastic hyperresorption. Our results suggest that the reduction of resorptive activity at the BMU level may precede the reduction in activation frequency previously reported in studies on older postmenopausal women. In this study, HRT exerted no demonstrable stimulatory effect on bone formation.
We appreciate the preparation of biopsy specimens by Anette Baattrup and Rita Ullerup. This study was supported by NOVO-Nordisk A/S, and The Center for Growth and Regeneration, Aarhus University.