Mineralization density distribution of postmenopausal osteoporotic bone is restored to normal after long-term alendronate treatment: qBEI and sSAXS data from the fracture intervention trial long-term extension (FLEX)

Abstract

Long-term treatment studies showed that the therapeutic effects of alendronate (ALN) were sustained over a 10-year treatment period. However, data on the effects on intrinsic bone material properties by long-term reduction of bone turnover are still sparse. We analyzed transiliacal bone biopsies of a subgroup of 30 Fracture Intervention Trial Long-Term Extension (FLEX) participants (n = 6 were treated for 10 years with ALN at dose of 10 mg/day, n = 10 were treated for 10 years with ALN at dose of 5 mg/day, and n = 14 were treated for 5 years with ALN plus a further 5 years with placebo) by quantitative backscattered electron imaging (qBEI) and scanning small-angle X-ray scattering (sSAXS) to determine the bone mineralization density distribution (BMDD) and the mineral particle thickness parameter T. BMDD data from these FLEX participants were compared with those from a previously published healthy population (n = 52). Compared with 5 years of ALN plus 5 years of placebo 10 years of ALN treatment (independent of the dose given) did not produce any difference in any of the BMDD parameters: The weighted mean (Ca_mean), the typical calcium concentration (Ca_peak), the heterogeneity of mineralization (Ca_width), the percentage of low-mineralized bone areas (Ca_low), and the portion of highly mineralized areas (Ca_high) were not different for the patients who continued ALN from those who stopped ALN after 5 years. Moreover, no significant differences for any of the BMDD parameters between the FLEX participants and the healthy population could be observed. In none of the investigated cases were abnormally high mineralization or changes in mineral particle thickness observed (Ca_high and T were both in the normal range). The findings of this study support the recommendation that antiresorptive treatment with ALN should be maintained for 5 years. Even with longer treatment durations of up to 10 years, though, no negative effects on bone matrix mineralization were observed. Copyright © 2010 American Society for Bone and Mineral Research

Introduction

At present, bisphosphonates are the most commonly used therapeutic agents for antiresorptive treatment of postmenopausal osteoporosis. For example, randomized clinical trials with alendronate (ALN) have demonstrated its efficacy in reducing vertebral fractures by about 50%.1, 2 The effect on nonvertebral fractures was consistent with a reduction of about 45% to 55%.3, 4 An increase in bone mineral density (BMD) and a decrease in bone turnover owing to the treatment are typical and are both likely contributors to the observed decrease in fracture incidence.5 However, BMD is not the sole predictor of fracture risk, and the increase in the BMD during bisphosphonate treatment only partially explains the decrease in an individual's fracture risk.6 Deterioration of the trabecular architecture (connectivity) and alterations in the bone material such as altered collagen cross-linking patterns,7 altered bone matrix mineralization status,8–10 and potential changes in accumulation and morphology of microdamage11 have to be considered in relation to bone fragility. Hence beneficial changes in bone quality parameters may play an important role in restoring the mechanical integrity of bone.12 Indeed, it was found that the decrease in fracture risk resulting from the bisphosphonate-induced reduction in bone turnover also was accompanied by significant changes in bone mineralization status, as first shown by microradiography13, 14 and later by quantitative backscattered electron imaging (qBEI) and synchrotron radiation micro-CT (SR-µCT).8, 10, 15

In general, the degree of bone matrix mineralization is a key determinant for the stiffness, hardness, and toughness of bone material16, 17 and can be described by the bone mineralization density distribution (BMDD). It can be measured in bone biopsies by techniques such as qBEI, microradiography, and SR-µCT,13, 15, 18–20 as mentioned earlier. The BMDD is the frequency distribution of the local calcium concentration values throughout the heterogeneously mineralized bone tissue. Two processes, bone turnover and the kinetics of mineralization,19, 21 are responsible for the shape of the BMDD. It has been shown that in the normal case of adult healthy individuals, the BMDD has a specific shape, independent of age, gender, ethnicity, or skeletal site.19, 22 Therefore, the BMDD is a sensitive tool for the diagnosis of alterations in pathologic cases and/or for the evaluation of treatment effects on bone mineralization. As has been described previously, a typical feature of osteoporotic bone material is its shift of the BMDD to lower calcium concentrations, indicating hypomineralization owing to the pathologically increased bone turnover8, 10 and potential changes in mineralization kinetics.19 Bisphosphonate therapy for up to 2 years' duration caused a characteristic narrowing of the BMDD, which reflected an increase in homogeneity of bone matrix mineralization and a slight shift of the BMDD toward higher calcium concentrations.8, 10, 13, 14 These two treatment effects on the mineralization status were ascribed to the sudden reduction of bone turnover causing a prolonged secondary mineralization period and to a lower percentage of newly formed bone undergoing primary mineralization.

In contrast, there is less information on the effects on bone material quality of long-term reduction of bone turnover by bisphophonates. Recently, it has been reported that there was no association between parameters of mineralization or Vicker's hardness and the duration of bisphosphonate therapy and bone remodeling in postmenopausal osteoporotic women treated with ALN or risedronate for 3 to 12 years.23 The Fracture Intervention Trial Long-Term Extension (FLEX) study demonstrated that fracture risk was similar after 10 years continuation of ALN compared with 5 years of ALN and thereafter discontinuation with 5 years of placebo treatment. Interestingly, 10-year ALN-treated patients revealed a lower fracture risk of clinically recognized vertebral fractures than those who stopped ALN after 5 years.24 The gain in BMD was greater at all measured sites, and bone turnover markers remained at lower levels for the 10-year treated patients compared with those who stopped ALN after 5 years.24–26 Potential detrimental effects at the bone material level after 10 years of bisphosphonate treatment remained unclear and are still under investigation. For instance, based on theoretical considerations, it is debated whether suppression of bone turnover for prolonged periods might cause hypermineralization of the bone matrix, leading to a more brittle bone and/or an accumulation of microdamage.27 Damage to bone at the material level could result in decreased bone strength, as shown in canine bone.28, 29 While Stepan et al. suggested greater microdamage accumulation in bisphosphonate-treated patients compared with normal individuals,30 others did not observe such an increase in microcrack frequency in treated patients.31 Histomorphometric and histologic data did not reveal any detrimental effects of long-term ALN therapy.24, 32 However, data on intrinsic bone composite material properties after long-term bisphosphonate therapy are still lacking. In this study we analyzed the BMDD of transiliac bone biopsies from a subgroup of 30 patients from the FLEX study (5 and 10 years of ALN treatment), as well as the mineral particle-thickness parameter T and the mineral platelet width W derived from the study.33 Moreover, we also compared the material properties data from the FLEX participants to a previously published normal reference group.22

Materials and Methods

Patients

A total of 1099 women with postmenopausal osteoporosis from the previous Fracture Intervention Trial (FIT) study2 (aged 55 to 81 years) who completed at least 3 years of ALN treatment (mean time on ALN was 5 years) were enrolled for the FLEX study.24 Participants were randomly allocated to receive ALN 10 mg/day, ALN 5 mg/day, or placebo for 5 years. Additionally, all participants received 500 mg calcium and 250 U vitamin D per day. The mean age (± SD) of the FLEX participants was 72.9 ± 5.5 years in the 10 mg/d for 10 years group, 72.7 ± 5.7 years in the 5 mg/d for 10 years group, and 73.7 ± 5.9 years in the placebo group at baseline. Total hip, femoral neck, and lumbar spine were decreased on average (T-scores: –1.9, –2.2, and –1.3, respectively). Further details about participating patients, treatment, and follow-up of the FLEX study have been published elsewhere.24

Bone biopsies

From a subgroup of 30 patients of the FLEX study, transiliac biopsies were obtained (n = 6 referred to the ALN 10 mg/day group, n = 10 to the ALN 5 mg/day group, and n = 14 to placebo) and were analyzed for bone mineralization density distribution (BMDD) by quantitative backscattered electron imaging (qBEI)19, 34 and for average mineral particle thickness parameter T obtained from scanning small-angle X-ray scattering (sSAXS) analysis.35 All biopsies were embedded in PMMA, and block samples with plane parallel surfaces were prepared for qBEI analysis. Sections 200 µm thick were cut from these blocks for sSAXS analysis.

Quantitative backscattered electron imaging (qBEI)

The BMDD of transiliac trabecular bone was determined by qBEI.19 Details of the digital scanning electron microscope (DSM 962, Zeiss, Oberkochen, Germany) used for qBEI and details of transformation of the gray levels to calcium concentration values, as well as the precision of the technique, have been described in detail in previous publications.34, 36 Digital images from the samples were taken using the following microscope settings: an accelerating voltage of 20 kV, a probe current of 110 pA, and a working distance of 15 mm. The entire trabecular bone tissue area was recorded by images of the same size (2 × 2.5 mm) at × 50 nominal magnification (corresponding to a resolution of 4 µm/pixel) using a scan speed of 100 seconds per frame, resulting in calibrated gray-level images as shown in Fig. 1A. These images were used for evaluation of the gray-level histograms, which were further transformed to weight-percent Ca histograms (BMDD), as shown in Fig. 1B. Five variables were evaluated to characterize the BMDD (introduced in refs. 19 and 22): Ca_mean, the weighted mean Ca concentration of the bone area; Ca_peak, the peak position of the histogram, which indicates the most frequently occurring (typical) calcium concentration of the studied bone area; Ca_width, the full width at half maximum of the distribution, describing the variation in mineralization density; Ca_low, the percentage of bone area with a calcium concentration of less than the 5th percentile of the reference BMDD (<17.68 wt% Ca), which reveals the amount of bone area undergoing primary mineralization; and Ca_high, the portion of bone area with a calcium concentration higher than the 95th percentile (>25.30 wt% Ca) of the reference BMDD (predominantly interstitial bone). The reference BMDD was obtained from trabecular bone from a healthy population described in detail in ref. 22. Trabecular bone from different skeletal sites (e.g., transiliac bone, vertebrae, femoral neck and head, and patella) of adult individuals (aged 25 to 90 years) of different ethnicity (African and Caucasian Americans) and gender was measured for its BMDD. The results revealed that none of these biologic factors caused significant differences in the BMDD. Thus a reference BMDD, together with reference BMDD parameters for human adult trabecular bone, including 52 individual BMDDs with varying aspects of biologic factors, could be established. Even small deviations from this BMDD were shown to have biologic meaning and were detected in cases of metabolic diseases or after treatment.19 In this study, we compared the BMDD from alendronate-treated bone with the reference BMDD in order to examine if and which deviations from normal BMDD exist after long-term bisphosphonate treatment.

Figure 1. (A) Backscattered electron image of a cross section of a transiliac bone biopsy from a participant of the FLEX study after 10 years of therapy with 10 mg of ALN. (B) At the bottom, the corresponding BMDD of trabecular bone from the biopsy (solid black line) is shown together with the 95% confidence interval of the normal reference BMDD (gray area) from a previous study.22

Scanning small-angle X-ray scattering

For the mineral particle-thickness parameter T, scanning small-angle x-ray scattering (sSAXS)35, 37 was performed on 200-µm-thick sections that were cut from the original blocks following qBEI measurements. sSAXS spectra were collected from the trabecular bone compartment using a Nanostar (Bruker AXS, Karlsruhe, Germany). A monochromatic beam of wavelength λ = 0.154 nm (Cu Kα) and of approximately 100 µm in diameter at the sample position was used. The x-ray scattering pattern was recorded by gas-filled area detector of 1024 × 1024 pixels. The resulting sSAXS spectra were reduced and analyzed using the Bruker software package and custom-made routines. Each frame was integrated radially and corrected for background and transmission. The sSAXS parameter T was determined from the sSAXS spectra according to procedures described earlier.35 Using the widely accepted assumption that the mineral particles have a platelike shape, the mineral platelet width W was derived from T by W = T/[2(1 − Φ)], where Φ is the volume fraction of mineral in the bone material, as described elsewhere.33 The Φ values were obtained from the corresponding Ca_mean values measured by qBEI.

Statistical analysis

Statistical analysis was performed using SigmaStat for Windows version 2.03 (SPSS, Inc.). Comparison between the 5-year ALN plus 5-year placebo group, the 10-year ALN 10 mg/day group, and the 10-year ALN 5 mg/day group was done using ANOVA (or ANOVA on ranks when appropriate). The differences of these three FLEX study groups from the normal population were tested for significance using t tests (or rank-sum tests when appropriate). In Table 1, normally distributed data are presented by mean (SD) and nonnormally distributed data by median (25%, 75%). Differences in mineral platelet width W of the FLEX participants were tested by ANOVA. Relationship between mineral volume fraction Φ and mineral platelet width W was tested by linear regression.

Table 1. The BMDD Parameters of the FLEX Participants Compared with Those of a Healthy Population

	5ALN/5PLA, n = 14	10ALN5mg, n = 10	10ALN10mg, n = 6	p Value	Normal reference from ref. 22
Note: No statistical significance among the three FLEX study groups could be detected (p value given). Comparison with normal reference data revealed no statistical significance. Source: From Roschger et al.22
Camean (wt%)	22.14 (0.58)	22.23 (0.59)	22.45 (0.49)	0.54	22.20 (0.45)
Capeak (wt%)	22.69 (0.59)	22.82 (0.55)	22.96 (0.45)	0.59	22.96 (22.70, 23.14)
Cawidth (wt%)	3.47 (3.29, 3.81)	3.62 (0.44)	3.23 (0.28)	0.14	3.29 (3.12, 3.47)
Calow (%)	4.41 (0.99)	4.61 (0.99)	3.88 (0.75)	0.34	4.52 (3.87, 5.79)
Cahigh (%)	6.21 (4.12)	6.33 (3.92)	6.97 (4.38)	0.93	4.62 (3.52, 6.48)

For all analyses, p < .05 was considered significantly different.

Results

We analyzed the mineralization status and mineral particle thickness parameter of trabecular bone of transiliac bone biopsies from a subgroup of participants of the FLEX study. A typical qBEI image of a sectioned bone biopsy area together with its corresponding BMDD from a patient treated for 10 years with 10 mg ALN is shown in Fig. 1.

• 1.
  There was no significant difference in any of the BMDD parameters between the group of patients who stopped ALN after 5 years and were treated with placebo for 5 more years (5ALN/5PLA) and the two groups of patients who continued ALN at 5 or 10 mg (10ALN5mg and 10ALN10mg). The weighted mean (Ca_mean) and typical calcium concentration (Ca_peak), the heterogeneity of mineralization (Ca_width), and the percentage of low (Ca_low) and high (Ca_high) mineralized bone areas were similar for the three study groups.
• 2.
  Comparison with normal reference data from healthy adult individuals (published in ref. 22; data see Table 1 and gray bars in Fig. 2) showed that all BMDD parameters of 10-year ALN groups, as well as for the group that stopped ALN after 5 years, were not significantly different from normal.
• 3.
  The mineral particle-thickness parameter T was found in the normal range (3.60 to 4.23 nm) for all studied bone biopsies from the FLEX study independent of treatment (Fig. 3A). Comparisons of the platelet width W showed no significant difference (p = .088) among the FLEX study groups. A linear relationship (r = .46, p = .01) was observed between W and the mineral volume fraction Φ (see Fig. 3B).

Figure 2. BMDD parameters Ca_mean, Ca_peak, Ca_width, Ca_low, and Ca_high from the three FLEX study groups: 5 years of ALN followed by 5 years of placebo (5ALN/5PLA) indicated by white, 10 years of 5 mg/day ALN (10ALN5mg), indicated by gray, and 10 years of 10 mg/day ALN (10ALN10mg), indicated by black symbols. Data shown are mean (SD) or median (IQR). Normal reference BMDD parameters (from ref. 22) are indicated by mean or median values (dotted lines) together with SD or interquartile range (gray areas). The levels A and B indicate the corresponding values after 2 to 3 years of placebo (A) and 2 to 3 years of ALN (B) treatment (mean data from the previously published phase III ALN study8).

Figure 3. The mineral particle-thickness parameter T as measured by sSAXS (A) and the mineral platelet width W (B) versus the volume fraction of mineral Φ, as measured by qBEI. White symbols indicate 5ALN/5PLA, gray and black symbols indicate 10ALN (5 and 10 mg doses, respectively). The solid line shows the linear regression of W versus Φ.

Discussion

For the first time, BMDD and the thickness of mineral particles were analyzed after long-term treatment with a bisphosphonate in transiliac biopsies from postmenopausal osteoporotic women. Remarkably, a subgroup of participants in the FLEX study showed that continuation of ALN for 10 years did not further change the BMDD compared with stopping ALN after 5 years and subsequent 5 years of placebo treatment. Since we used exactly the same methodical approach to measure the BMDD in the previously published phase III ALN study,8 we are able to compare the BMDD data from these two studies. Although the two study populations were not similar, the distinct differences in the BMDD are likely to show the changes in the mineralization pattern with treatment duration (from 2 to 3 years of ALN in the phase III ALN study compared with 5 and 10 years of ALN in the FLEX study). The patients from the FLEX study revealed normal (as defined by the BMDD parameters of a healthy population22) mean and typical calcium concentrations (Ca_mean and Ca_peak, respectively), whereas the placebo- and ALN-treated patients of the phase III ALN trial both showed lowered mineralization densities compared with normal (see levels of mineralization indicated by the dashed lines in Fig. 2). As observed in other studies, hypomineralization (shift toward lower mineralization densities19) of bone seems to be a common finding in untreated postmenopausal osteoporosis patients,8, 10 which reflects the pathologically high bone turnover38 and/or a certain deficiency in calcium and vitamin D levels.10 Both effects are likely major contributors to the increased fracture risk of these patients. Recently, highly significant correlations between the degree of mineralization and the microhardness of single bone packets from postmenopausal osteoporotic patients have been shown, demonstrating the importance of the mineralization density for mechanical performance of the bone material.17

Bisphosphonates such as ALN cause a distinct reduction in bone turnover and subsequently also a reduction in the generation of new bone packets, which exhibit a lower mineralization status (primary mineralization state). Additionally, already formed bone packets continue to mineralize. These effects together are responsible for the observed increase in mean and typical Ca content, for the narrowing of the BMDD peak (reduction in Ca_width), and for the reduction in the amount of low-mineralized bone areas after 2 to 3 years of ALN.8 This increase in homogeneity in bone matrix mineralization seen after 2 to 3 years of bisphosphonate therapy mirrors the sudden bone turnover reduction that already can be observed by a reduction in bone turnover markers after 3 to 6 months of treatment. Computed modeling of the action of antiresorptive therapy on the BMDD confirmed this transient nature of the narrowing of the BMDD peak, as observed during the first 3 years of therapy.21 Moreover, the model also suggests that after long-term therapy, the BMDD peak width returns to normal values.39 This phenomenon was observed in the previously published analysis from the risedronate (VERT-NA) study10, 15 but was not well understood at the time owing to the fact that the aforementioned computer model was not available. The relative amount of low-mineralized bone (areas undergoing primary mineralization, Ca_low) was found to be increased in postmenopausal osteoporosis10, 19 and is expected to parallel the transient decrease in Ca_width (describing the homogeneity of mineralization), as mentioned earlier. At the time of the phase III ALN study, Ca_low had not yet been defined; thus it was not calculated.8 However, a transient numerical decrease after 3 years of bisphosphonate in Ca_low was observed in the risedronate VERT-NA study.10 After 5 as well as 10 years of ALN (FLEX study), Ca_low reaches normal values. Additionally, the portion of high-mineralized bone Ca_high is in the upper-normal range after 5 years of ALN treatment and not statistically higher than normal even after the longer treatment period of 10 years. It has to be mentioned that although not statistically significant, Ca_high values from all three treatment groups in the present study were numerically higher than normal. This may indicate that there is a tendency toward a greater portion of fully mineralized bone packets in this new bone turnover equilibrium state after long-term bisphosphonate therapy. However, even when both doses of the 10-year ALN-treated groups were pooled, the difference from normal did not reach statistical significance, obviously owing to a relatively large variation in this parameter in the reference group as well.

It is interesting to analyze the relationship of data among BMD, serum markers, and BMDD in the case of discontinuation and continuation of ALN treatment. The outcomes of the clinical trials were that the gain in BMD was somewhat higher and the bone turnover markers lower in patients who continued ALN for 10 years compared with 5 years of ALN.24–26 On the contrary, the BMDD data did not show any significant changes between these two treatment times. Taking into account that the changes in BMD reflect changes in bone volume and mineralization density of the bone matrix,40 this implies that the patients who continued ALN exhibited small increases in bone volume, indicating a continuous positive net balance between bone formation and resorption within the further 5 years of ALN treatment. These increases in bone volume were likely too small to be detected by histomorphometry.24 Histomorphometry has been described to require large sample numbers for the detection of small changes in bone volume.41 Indeed, strong indication for increases in bone volume by anticatabolic treatment with bisphosphonates has been reported recently for osteoporotic patients treated with risedronate.40 This was concluded using a novel method combining BMD data of the spine with BMDD data of the iliac crest. Remarkably, in the FLEX study, the discontinuation of ALN after 5 years was followed only by a decline in hip BMD compared with untreated patients of the same age. One possible explanation for the latter finding and for similar BMDD data in the present study independent of stopping or continuing treatment might be that the bisphosphonate that had been incorporated into the bone mineral during treatment becomes “recycled” owing to bone resorption and could act for longer time after stopping treatment. This hypothesis is also consistent with the maintenance of spine BMD data of the placebo-treated patients in the FLEX study. However, hip BMD declined rather rapidly with discontinuation of bisphosphonate treatment. This is not in line with the “recycle” hypothesis but may indicate that stopping bisphosphonates has a greater effect during the first 5 years on cortical bone from the hip than on trabecular compartments in the spine, as well as in the iliac crest, as shown by our BMDD data.

Potential hypermineralization and/or mineralization defects of the bone matrix, changes in collagen structure,42 retarded fracture healing,43 and accumulation of microdamage44 have been discussed in connection with bisphosphonate treatment in general and specifically in the context with long-term treatment. Nevertheless, we observed that the mineral content of bone packets did not exceed the values for normal, fully mineralized bone matrix (interstitial bone), and the percentage of fully mineralized bone areas was not statistically different from that of normal/healthy adults. Boivin et al.23 reported recently a declining trend in mean mineralization densities with long-term bisphosphonate treatment that is in apparent disagreement with our data. This discrepancy may be attributable to either methodologic differences19 and/or to the fact that both our baseline and current measurements were performed in cancellous bone exclusively. In addition, differences in analogous patient populations and variable compliances could contribute to these differences.

In agreement with previous observations on bisphosphonate therapy,8 we found the average mineral particle-thickness parameter T to be within the normal range. A more detailed analysis gave the mineral particle width W, which was found to increase linearly with the mineral volume fraction Φ. This can be interpreted in such a way that the increase in mineral volume fraction during secondary mineralization is due primarily to an increase in platelet width. The particle-thickness parameters showed no significant differences between the FLEX study groups.

All these findings indicate that long-term ALN treatment did not specifically change the mineral phase of the collagen/mineral composite material of bone. Recent data on microcrack frequency in ALN-treated bone from osteoporotic subjects are somewhat controversial. Microdamage accumulation parameters, such as crack surface density, crack density, and crack length, were correlated with low bone mineral density and increased age in ALN-treated patients but not in controls, giving evidence for higher levels of microdamage in bisphosphonate-treated patients.30 In another study, microcrack frequency was found to be low and not significantly increased after long-term bisphosphonates.31 Although we did not study microcracks in the present work, our finding of normal Ca_high after long-term treatment indicates that there is no increase in the amount of highly mineralized bone areas, which are usually considered to be the sites where microdamage preferentially occurs.45 However, animal studies with partly detrimental effects of bisphosphonates on mechanical integrity of bone material have been reported as well.28, 29 Histologically and histomorphometrically, no evidence for abnormalities or mineralization defects were observed after 3 years32 and 5 to 10 years of ALN treatment,24 as well as after 3 years of zoledronic acid therapy.46

The limitations of our study are relatively small sample sizes within the study groups. However, it has to be emphasized that the technical variance of the BMDD parameters is 10 times lower than the biologic variance of these parameters. Ca_mean, for instance, has a technical variance of 0.3%34 compared with the variance of 2.2% of this parameter within the 10ALN10mg group (the other two study groups reveal similar values). For Ca_high, which is the parameter with the largest technical variance (11.5%), biologic variance is nearly sixfold within the study groups. Another limitation is that the temporal development of BMDD with treatment duration could not be studied within the same population because for each time point only different study groups were available.

In conclusion, the FLEX study showed that 10 years of treatment with ALN was well tolerated, and the therapeutic effects were sustained over the whole treatment period.25 Our data on the material properties show that 5 years of ALN treatment restores the altered BMDD in postmenopausal osteoporosis to normal values. This normalization of mineralization density distribution lasts even for longer treatment durations of up to 10 years without any evidence for hypermineralization or other detrimental effects. Our data therefore support the view that antiresorptive treatment with ALN should be maintained for up to 5 years and is safe even for longer treatment periods.

Disclosures

All the authors state that they have no conflicts of interest.

Acknowledgements

We thank G Dinst, D Gabriel, P Messmer, and S Thon for careful sample preparation and qBEI measurements at the bone material laboratory of the Ludwig Boltzmann Institute of Osteology, Vienna, Austria. This study was supported by the AUVA (Austrian Social Insurance for Occupational Risk), the WGKK (Social Health Insurance Vienna), and FWF (Austrian Science Fund) Project No. P19009-N20. Financial support for this study was provided by Merck. Part of this study was presented at the 29th ASBMR Meeting, Honululu, HI, USA, September 16–19, 2007.