Effects of Intravenous Zoledronic Acid Once Yearly on Bone Remodeling and Bone Structure

Authors


  • Dr Recker is a consultant for Novartis Pharmaceuticals Corp, Merck & Co, Inc, Roche Laboratories, Amgen Inc, Procter & Gamble Pharmaceuticals, Eli Lilly & Co, GlaxoSmithKline, NPS Allelix, and Wyeth Pharmaceuticals, Inc. Dr Delmas has received research grants from Procter & Gamble Pharmaceuticals, Eli Lilly & Co, and Amgen Inc. He has consulted and/or received speaker fees from Acceleron Pharma, Amgen Inc, Eli Lilly & Co, GlaxoSmithKline, Merck, Sharpe & Dohme, Novartis Pharmaceuticals Corp, Nycomed Group, Organon USA Inc, Pfizer Inc, Procter & Gamble Pharmaceuticals, Roche Laboratories, Sanofi-Aventis, Servier Laboratories, Wyeth Pharmaceuticals, and Zelos Therapeutics, Inc. Dr Reid is a consultant for Novartis Pharmaceuticals Corp, Merck & Co, Inc, and Sanofi-Aventis. Dr Lewiecki has received grants/research support from Novartis Pharmaceuticals Corp, Merck & Co, Inc, Procter & Gamble Pharmaceuticals, Eli Lilly & Co, Roche Laboratories, GlaxoSmithKline, Wyeth Pharmaceuticals, and Pfizer Inc. He is a consultant for Novartis Pharmaceuticals Corp, Merck & Co, Inc, Procter & Gamble Pharmaceuticals, Eli Lilly & Co, Roche Laboratories, GlaxoSmithKline, and Wyeth Pharmaceuticals. He owns stock in General Electric and Procter & Gamble Pharmaceuticals. Dr Miller has received scientific grants from Procter & Gamble Pharmaceuticals, Sanofi-Aventis, Roche Laboratories, Eli Lilly & Co, Merck & Co, Inc, Novartis Pharmaceuticals Corp, and Amgen Inc. He is on speaker boards, advisory boards, or is a consultant for Procter & Gamble Pharmaceuticals, Sanofi-Aventis, Merck & Co, Inc, Eli Lilly & Co, Amgen Inc, NPS Pharmaceuticals, Novartis Pharmaceuticals Corp, Roche Laboratories, and GlaxoSmithKline. Dr Hu is an employee of Novartis Pharmaceuticals, East Hanover, NJ, USA. Dr Mesenbrink owns stock, restricted stock, exercisable options, and tradable options in Novartis Pharmaceuticals Corp of East Hanover, NJ, USA, of which he is an employee. Dr Hartl is an employee of F. Hoffmann-La Roche AG, Basel, Switzerland, and a former employee of Novartis Pharma AG, Basel, Switzerland. Dr Gasser has a corporate appointment with Novartis Pharma AG, Basel, Switzerland. He also owns stock that may cause a conflict of interest. Dr Eriksen is an employee of Novartis Pharma AG, Basel, Switzerland. All other authors state that they have no conflicts of interest.

Abstract

In a substudy of the HORIZON pivotal fracture trial, in which yearly intravenous zoledronic acid 5 mg was found to significantly reduce risk of various fracture types in patients with postmenopausal osteoporosis, 152 patients underwent bone biopsy. Zoledronic acid reduced bone turnover by 63% and preserved bone structure and volume, with evidence of ongoing bone remodeling in 99% of biopsies obtained.

Introduction: In the HORIZON pivotal fracture trial (PFT), enrolling 7736 women with postmenopausal osteoporosis, three annual intravenous infusions of the bisphosphonate zoledronic acid (5 mg) significantly reduced morphometric vertebral, clinical vertebral, hip, and nonvertebral fractures by 70%, 77%, 41%, and 25%, respectively. Whereas 79% of patients received zoledronic acid/placebo only (stratum I, n = 6113), 21% received concomitant treatment with other antiresorptive drugs, excluding other bisphosphonates, PTH, and strontium (stratum II, n = 1652).

Materials and Methods: To determine effects on bone remodeling and bone architecture, iliac crest bone biopsies were obtained in 152 patients on active treatment or placebo at 3 yr after double tetracycline labeling. In five patients, only qualitative histology was performed, leaving 147 biopsy cores (79 on active treatment and 68 on placebo) for νCT analysis and histomorphometry.

Results: Analysis of bone structure by νCT revealed higher trabecular bone volume (BV/TV) in the zoledronic acid group (median, 16.6% versus 12.8%; p = 0.020). In addition, patients treated with zoledronic acid exhibited higher trabecular numbers (p = 0.008), decreased trabecular separation (p = 0.011), and a trend toward improvement in connectivity density (p = 0.062), all indicating better preservation of trabecular structure after treatment with zoledronic acid. Qualitative analysis revealed presence of tetracycline label in 81 of 82 biopsies from patients on zoledronic acid and all 70 biopsies from placebo patients, indicative of continued bone remodeling. No bone pathology was observed. Zoledronic acid induced a 63% median (71% mean) reduction of the activation frequency (Ac.f; p < 0.0001) and reduced mineralizing surface (MS/BS; p < 0.0001) and volume referent bone formation rate (BFR/BV) versus placebo, indicating reduced bone turnover. Mineral appositional rate was higher in the zoledronic acid group (p = 0.0002), suggesting improved osteoblast function compared with placebo. Mineralization lag time was similar in the two groups, whereas osteoid volume (OV/BV; p < 0.0001) and osteoid thickness (O. Th; p = 0.0094) were lower in zoledronic acid-treated patients, indicating normal osteoid formation and mineralization of newly formed bone. Concomitant administration of other antiresorptive osteoporosis therapies (e.g., raloxifene, tamoxifen, tibolone, ipriflavone) did not significantly alter the tissue level response to zoledronic acid.

Conclusions: Annual dosing for 3 yr with zoledronic acid 5 mg intravenously resulted in a median 63% (mean, 71%) reduction of bone turnover and preservation of bone structure and mass without any signs of adynamic bone. Concomitant treatment with other osteoporosis therapies did not significantly affect the bone response to zoledronic acid.

INTRODUCTION

Previous studies of bone biopsies obtained from osteoporotic patients have shown that the main defect is incomplete refilling of resorption lacunae leading to a negative balance at the level of individual bone multicellular units (BMUs).(1) Over time, this imbalance leads to a gradual thinning or perforation of trabeculae, impaired trabecular connectivity, and deterioration of bone microarchitecture. In postmenopausal osteoporosis, activation frequency is generally increased; however, values may vary widely.(2)

Treatment with first-generation bisphosphonates such as etidronate raised concerns over mineralization defects.(3) No such problems were encountered in long-term clinical studies with newer, potent amino-bisphosphonates such as alendronate, risedronate, and ibandronate.(4-6) These studies also reported total absence of other qualitative abnormalities (woven bone, marrow fibrosis, or signs of cellular toxicity) in newly formed bone. As expected for this class of drug, bone turnover was markedly reduced, but without any evidence of osteomalacia.

Zoledronic acid is a third-generation aminobisphosphonate displaying the highest inhibition of farnesyl disphosphate synthase (FPP synthase) and greatest affinity for bone mineral to date. In ovariectomized rats, zoledronic acid prevented estrogen-deficient bone loss in vertebrae and showed a prominent protective effect on trabecular thinning.(7) In a 12-mo phase II study, zoledronic acid 4 mg given intravenously was shown to be a potent inhibitor of bone resorption. Increases in BMD and reductions in bone markers 1 yr after a single dose were similar to those of approved bisphosphonates with daily oral treatment.(8) This prompted the initiation of a 3-yr phase III trial. Because of the gradual increase in bone turnover within the postmenopausal range during the time from the nadir with once-yearly dosing seen at 12 mo in the phase II study, as well as the emergence of two negative bisphosphonate studies in which underdosing was suspected as the main cause for lack of efficacy,(9,10) a higher dose of 5 mg intravenous once yearly was chosen to secure sufficient reduction of bone turnover throughout the year in all patients. The phase III study (HORIZON Pivotal Fracture Trial [PFT]) enrolled 7736 postmenopausal women with osteoporosis and showed pronounced antifracture efficacy of yearly 5 mg infusions of zoledronic acid, with significant reductions of morphometric vertebral, clinical vertebral, hip, and nonvertebral fractures by 70%, 77%, 41%, and 25%, respectively, over a 3-yr period.(11) To assess bone safety and effects on remodeling and bone structure of this treatment regimen, bone biopsies obtained from a subgroup of 152 women were analyzed using νCT, qualitative histology, and quantitative histomorphometry.

MATERIAL AND METHODS

Study population

Overall, 7765 patients were enrolled in the HORIZON-PFT trial (Fig. 1). Patients were categorized into two strata according to whether they were receiving other medications for osteoporosis: Stratum I patients were not allowed any concomitant osteoporosis medication and received either yearly doses of active drug or placebo plus daily calcium (1.0-1.5 g) and vitamin D (400-1200 IU); Stratum II patients received zoledronic acid or placebo, daily elemental calcium (1.0-1.5 g), and vitamin D (400-1200 IU), plus other therapies used in usual care that included the following medications: hormone replacement therapy (HRT), selective estrogen receptor modulators (SERMs; raloxifene, tamoxifen), calcitonin, tibolone, dehydroepiandrosterone (DHEA), ipriflavone, and medroxyprogesterone. Patients previously treated with PTH, fluoride, or strontium were excluded. Bisphosphonates were either prohibited or needed to have adequate washout before study entry (2 yr [if used for >48 wk]; 1 yr [if used for >8 but <48 wk]; 6 mo [if used for >2 but ≤8 wk]). All concomitant osteoporosis medications and therapies had to be documented during the screening period. A total of 6084 (78.7%) patients from stratum I and 1652 (21.3%) from stratum II were included in the intent-to-treat population, of which 3875 were randomized to zoledronic acid and 3861 were randomized to placebo. Of those randomized, 6517 (84.2%) patients completed the study and 1219 (15.8%) patients discontinued from the study.

Figure Fig. 1..

Patient disposition according to stratum and test type.

Patients who were willing to participate in the bone biopsy procedures and signed an informed consent had a bone biopsy performed at any time between month 33 and month 36 (visit 7). By the end of the study on June 15, 2006, 152 patients had been recruited from 24 centers in 10 countries. A total of 143 biopsy cores (76 from zoledronic acid recipients and 67 from the placebo group) had at least one νCT imaging parameter measured. A total of 152 biopsy cores (82 zoledronic acid and 70 placebo) underwent qualitative histological assessment at the histomorphometry laboratory. Quantitative histomorphometry could be performed on 111 biopsies (59 zoledronic acid and 52 placebo), and all planned histomorphometric indices were assessed in 99 biopsies. Forty-one core samples (23 zoledronic acid and 18 placebo) were not readable because of technical reasons (insufficient tissue for histomorphometry, fragmented biopsy, artifacts, etc.).

Bone biopsy procedures and assessments

Labeling

The labeling schedule was as follows: 3 days of tetracycline, 14 days without tetracycline, 3 additional days of tetracycline, and at least a 5-day interval before the biopsy was performed. Tetracycline was used for the first label and dimethylchlorotetracycline for the second label. Where dimethylchlorotetracycline was not available, tetracycline was used for both labeling sequences. If tetracycline was not available, doxycycline (100 mg, twice daily) was used. The dose of tetracycline hydrochloride was 250 mg, four times daily taken at least 30 min after a meal, and the dose of dimethylchlorotetracycline was 150 mg, three times daily. Compliance with tetracycline intake was verified by assessment of tetracycline in urine.

Collection of bone biopsy samples

At the time the bone biopsy was obtained, the specimen was inspected for quality. Only those samples that contained both the inner and outer cortex, and intact trabecular bone that was not crushed, were suitable for analysis of all structural and histomorphometric parameters. Crushed samples were included for qualitative label analysis, however. In cases where the first specimen obtained contained only one cortex, or was otherwise inadequate, a second biopsy from a site adjacent to that of the first was taken to acquire an acceptable sample of bone. The samples were stored in 70% ethanol in 20-ml scintillation vials.

Mailing and handling

All samples were handled by Covance Central Laboratory Services. Urine samples were collected to confirm the presence of tetracycline in all subjects. Bone biopsy samples were shipped to Novartis Basel AG for νCT analysis and later sent to the Osteoporosis Research Center Histomorphometry Laboratory at Creighton University for histomorphometry analysis. All analyses described in subsequent sections were performed according to a prespecified analysis plan approved by the HORIZON-FPT Steering Committee.

νCT analysis of transiliac bone biopsies

Before the histomorphometric analysis, the biopsy cores were subjected to 3D analysis of overall morphology and quality by νCT in the νCT Laboratory (Dr JA Gasser) at Novartis, Basel, Switzerland. Suitable biopsy cores were subjected to assessment of bone structure using a VivaCT40 MicroCT (Scanco Medical, Bassersdorf, Switzerland) with a resolution of 15 νm in all three dimensions.

The segmentation was carried out in the cancellous bone area starting at a distance of 1 mm away from the two cortical shells. Also, the segmentation excluded the outer 1 mm of the bone cylinders to avoid interference of bone debris with the measurements. The complete spongiosa in the biopsy as specified above was evaluated to obtain the numerical values, thereby completely avoiding the sampling errors caused by random deviations of a single section. The resulting grayscale images were segmented using a low-pass filter to remove noise, and a fixed threshold to extract the mineralized bone phase. From the binary images, structural indices were assessed with recently developed 3D techniques without model assumptions for the appearance of trabecular bone.

Material density (MATDEN), which reflects average degree of mineralization of bone matrix in the biopsy core as a whole, was determined after calibration of the VivaCT40 with a five-step hydroxyapatite phantom provided by the manufacturer.

Calculated 3D morphometric endpoints

Relative bone volume (BV/TV, %), trabecular number (Tb.N, 1/mm), trabecular thickness (Tb.Th, νm), and trabecular separation (Tb.Sp, νm) were calculated by measuring 3D distances(9) directly in the trabecular network, and taking the mean over all voxels. The diameter of spheres filling the structure were taken as Tb.Th, the thickness of the marrow spaces as Tb.Sp, and the inverse of the mean distances of the skeletonized structure was calculated as Tb.N. Bone surface (BS/TV, mm2) was calculated using a tetrahedron meshing technique generated with the Marching Cube method. Degree of anisotropy (DA) was calculated by projecting the surfaces of the triangles onto an ellipse.(10) By displacing the surface of the structure in infinitesimal amounts, the structure model index (SMI)(9) was calculated as

equation image

The SMI quantifies the plate versus rod characteristics of trabecular bone, where an SMI of 0 reflects a purely plate-shaped bone, an SMI of 3 indicates a purely rodlike bone, and values in between represent mixtures of plates and rods. Furthermore, connectivity density (Conn. D, 1/mm3) based on Euler number(11) was determined. Conn. D determines how many connections each trabecular element has with the surrounding structures: the higher the value, the more connections between the trabecular elements.

The CV for all reported parameters was <1.02%, except for the SMI, for which it was 1.64%.

Specimen preparation and histomorphometric assessment

After νCT assessment, the complete biopsy cores in ethanol were shipped to RRR for qualitative and histomorphometric analysis. Specimens were dehydrated and embedded into methyl methacrylate and then sectioned. Sections were mounted unstained for analysis of tetracycline labels or stained with Goldner-Trichrome or Toluidine blue for qualitative and histomorphometric analysis. Histomorphometry variables were analyzed in both trabecular bone and cortical bone by semiautomated image analysis.

Qualitative histological analysis

Qualitative histological analysis was performed on all biopsies to detect bone mineralization abnormalities (i.e., osteomalacia), woven bone, marrow fibrosis, defects in cellular components, absence of tetracycline labels, or other abnormalities. Assessment of tetracycline label was performed on two consecutive 7-νm sections. If label was totally absent in cortex or trabecular bone, four additional sections were examined. This procedure was only necessary for two patients (one zoledronic acid and one placebo).

Quantitative histomorphometric analysis

Quantitative histomorphometry was performed on Goldner-Trichrome-stained sections, and unstained sections were used for calculation of tetracycline-based indices. The parameters measured were reported according to the ASBMR nomenclature.(12)

Statistics

The analyses performed compared bone quality and structure parameters between zoledronic acid and placebo treatment groups in stratum I, stratum II, and overall.

Descriptive summary statistics (mean, median, SE, minimum and maximum) are presented for the quantitative assessments by treatment group and stratum. Between-treatment differences for each quantitative measure were evaluated using a Wilcoxon rank-sum test.

The need for patient consent may have biased the biopsy population slightly (e.g., no patients in Asia participated in the bone biopsy substudy), although baseline demographics were not significantly different from the study population as a whole. The analyses and p values for the bone biopsy and νCT data are reported assuming that similar results would be obtained if the samples were strictly random.

RESULTS

The biopsy subpopulation was well matched with the overall trial population in terms of age, BMI, prevalent fractures, distribution between strata, and previous bisphosphonate use (Table 1). The geographic distribution differed slightly, with more patients being from the Americas in the biopsy cohort. Moreover, no biopsies were obtained from Asia. Hip BMD was slightly higher in the biopsy cohort, whereas spine BMD was similar for both groups.

Table Table 1.. Demographic Characteristics of Zoledronic Acid Treatment and Placebo Groups in the Bone Biopsy Population Compared With the Total Population Studied in HORIZON-PFT [Mean (±SD)]
original image

νCT

Full evaluation of all parameters for bone structure using νCT was possible in a total of 99 cores (Table 2). The analysis revealed higher BV/TV in the zoledronic acid group (median, 16.6% versus 12.8%; p = 0.020). Patients on zoledronic acid also exhibited increased median Tb.N, decreased Tb.Sp, and a strong trend toward higher Conn. D compared with placebo. Tb.Th, as well as the structure model index (SMI; an index of trabecular morphology), was similar between groups. The median cortical thickness was 0.72 mm in patients treated with zoledronic acid versus 0.63 mm in placebo (p = 0.122). Representative νCT images of biopsies from patients on zoledronic acid and placebo are shown in Fig. 2. Between-treatment differences in structural indices were larger in stratum II than in stratum I and reached significance for Tb.N, Tb.Sp, and cortical thickness (Table 3). The νCT analysis also provided estimates of average degree of bone matrix mineralization as reflected in material density (MATDEN). The two groups revealed a similar degree of matrix mineralization as reflected in MATDEN (909 [900-924] mg/cm3 for the treatment group and 902 [887-920] mg/cm3 for placebo; p = 0.446; Table 2).

Table Table 2.. Bone Structural Parameters Assessed by νCT [Median (95% CI)]
original image
Table Table 3.. Static, Dynamic, and Structural Parameters Assessed by Quantitative Histomorphometry [Median (95% CI)]
original image
Figure Fig. 2..

Representative biopsies obtained from one patient treated with zoledronic acid 5 mg and placebo. The figure shows a νCT rendition of the whole biopsy core and a thick (middle panel) and a thin section (left panel). Note the better preservation of trabecular bone architecture in the zoledronic acid-treated patient.

Qualitative histology

Eighty-one of 82 biopsies obtained from patients treated with zoledronic acid and all 70 biopsies from placebo patients showed detectable tetracycline label in either trabecular or cortical bone. The single biopsy without label was fragmented, had only one cortex, and contained insufficient tissue for a more detailed measurement. Thus, no quantitative histomorphometric data were available for this subject. Qualitative analysis also revealed that all biopsies obtained from patients treated with zoledronic acid contained lamellar bone without evidence of marrow fibrosis, woven bone, or cellular toxicity.

Quantitative histomorphometry

Static and dynamic histomorphometric analyses of evaluable biopsies are presented in Table 3. A total of 86 patients of stratum I and stratum II had sufficient tetracycline double labels within trabecular bone to permit assessment of tetracycline-based indices and therefore underwent full quantitative histomorphometric analysis.

For 25 patients (4 placebo and 21 zoledronic acid patients), only partial quantitative data were available because the amount of single or double label present in trabecular bone was deemed insufficient for reliable assessment of tetracycline-based indices. Details for these 25 patients are indicated below:

  • Placebo: two patients had double label in cortical bone only. In one placebo patient, triple labels were present. Finally, in one placebo patient, the amount of label in trabecular bone was insufficient for reliable estimation of labeling indices.

  • Zoledronic acid: 16 patients had double label within the cortical bone but not in the trabecular bone. Three patients revealed single label in cortex or trabecular bone only. Two patients were not evaluable because of the presence of prior tetracycline label.

For these aforementioned patients, mineralization surface was recorded as a value of zero in the data set because of the inability to reliably estimate tetracycline parameters.

The effect of zoledronic acid on bone turnover was reflected in the following parameters: a significant reduction in activation frequency (Ac.f; median reduction, 63%; mean reduction, 71%; p < 0.0001), mineralizing surface (MS/BS), and volume-referent bone formation rate (BFR/BV) in the zoledronic acid group versus placebo. Eroded surface (ES/BS) trended lower in biopsies obtained from zoledronic acid-treated patients (Table 3).

The following osteoid parameters used to assess mineralization of newly formed bone were all significantly lower in the zoledronic acid group versus placebo: osteoid thickness (p = 0.009), median osteoid surface (OS/BS; p < 0.0001), and osteoid volume (OV/BV; p < 0.0001). A small but significant increase in mineral apposition rate (MAR; 0.60 versus 0.53 νm/d; p < 0.001) was observed with zoledronic acid treatment. Mineralization lag time (Mlt: interval between initiation of osteoid formation and initiation of mineralization) tended to be longer with zoledronic acid treatment but was not significantly different from placebo (p = 0.096). Remodeling and resorption periods were both significantly prolonged by zoledronic acid (p < 0.0001; Table 3).

2D assessment of BV/TV also revealed significantly higher values in the zoledronic acid treatment group (p = 0.046), corroborating the νCT findings. Wall thickness and cortical thickness exhibited higher numerical values in patients on zoledronic acid, but these differences were not statistically significant (Table 3).

When strata I and II cohorts were analyzed separately, median values for important histomorphometric indices were similar between strata (Table 4). Differences between indices, however, revealed somewhat smaller changes in stratum II. Activation frequency was reduced by 63% and 45% after zoledronic acid treatment in strata I and II, respectively. There was no treatment-by-stratum interaction, indicating the effect across strata was different (p = 0.402), whereas MS/BS and OS/BS revealed similar reductions in the two strata. Moreover, the increased MAR in the zoledronic acid treatment groups was present, irrespective of stratum.

Table Table 4.. Key Structural and Histomorphometry Variables in Strata I and II [Median (95% CI)]
original image

DISCUSSION

The analysis of bone structure by νCT and bone remodeling by qualitative and quantitative methods revealed that treatment with once-yearly doses of zoledronic acid 5 mg intravenous for 3 yr reduced bone turnover by a median of 63% (mean, 71%) and preserved trabecular bone structure in the iliac crest. The bone formed during treatment with zoledronic acid was without evidence of marrow fibrosis, woven bone, or osteomalacia, further supporting the bone safety of this new bisphosphonate.

3D assessment by νCT revealed higher BV/TV and trabecular number (Tb.N) and decreased Tb.Sp in biopsies obtained from osteoporotic women treated with zoledronic acid. In addition, a trend (p = 0.062) toward improved connectivity density was seen. These data are in accordance with a preservation of trabecular bone structure in the zoledronic acid group, which may partly explain the antifracture efficacy of this agent. As reported for other bisphosphonates, no change in cortical thickness was demonstrable in the combined strata I and II.(4-6) However, in stratum II taken alone, cortical thickness was increased in patients treated with zoledronic acid. 2D histomorphometric analysis of bone structure such as Tb.N, Tb.Th, and wall thickness (W.Th) also revealed trends toward preservation, but significant changes were only shown for bone volume and Tb.N. Analysis of bone structure based on 2D indices, such as Tb.N, Tb.Th, and Tb.Sp, demands many assumptions regarding the shape of bone elements and entails examination of very limited volumes of bone. Such analyses are consequently less sensitive than 3D analyses with νCT, where the entire biopsy core is measured.(11) Thus, the apparent discrepancies between the two methodologies with respect to bone structure were to be expected. No prior study on bisphosphonates(4-6) has revealed significantly higher bone volume in the treatment group. In a previous νCT study on risedronate using paired biopsies, significant improvements in bone structure over placebo were only shown for patients with high turnover at baseline.(13) However, Recker et al.(14) showed similar increases when comparing BV/TV assessed by both νCT and histomorphometry in alendronate-treated versus placebo patients in a study comprising a total of 88 specimens. The significantly thicker cortex after zoledronic acid treatment in stratum II is noteworthy. It could represent a combined positive effect of other antiresorptive osteoporosis therapies and zoledronic acid therapy on this envelope. However, with the limited number of specimens available for this subgroup, a type II error cannot be excluded.

Since the introduction of bisphosphonates for the treatment of osteoporosis, concerns were raised that these compounds could lead to adynamic bone disease, resulting in increased fracture incidence with prolonged use.(15-18) The 152 biopsies obtained in this study were available for qualitative analysis of tetracycline label, and all but one from a patient treated with zoledronic acid contained label indicating preserved remodeling capacity. The biopsy without label was incomplete and fragmented, which may have limited the ability to detect label. Combined analyses of bone biopsies from two alendronate studies (71 alendronate and 88 placebo) showed the absence of tetracycline label in only one patient on alendronate and one on placebo.(4) The histomorphometric analysis from the risedronate pivotal study revealed the presence of label in all biopsies.(6) More biopsies obtained from patients treated with zoledronic acid (n = 21) had insufficient amounts of double label for quantitative assessment of bone formation rates compared with biopsies obtained from the placebo group (n = 4). Based on past histomorphometric analyses in patients treated with bisphosphonates, the increased abundance of single label in patients treated with zoledronic acid is not unexpected. As shown in Table 3, zoledronic acid, as with other bisphosphonates, prolongs the remodeling cycle.(4,6,19) This implies that more remodeling sites will be in the later stages of remodeling, in which more on-off phenomena take place, and which would create more single labels.(20)

The findings obtained in this study, together with the fact that absence of tetracycline label is occasionally observed even in untreated patients, are therefore consistent with active, ongoing bone remodeling during treatment with zoledronic acid. This notion is further corroborated by the biomarker profile obtained in this study.(21) Between each dosing, markers of bone resorption and formation all exhibited increases, in keeping with preservation of remodeling capacity in bone.

In the Fracture Intervention Trial Long-Term Extension (FLEX) study,(22) ∼1000 women from the alendronate Fracture Intervention Trial (FIT) trial were randomized to either 5 or 10 yr of treatment with alendronate, which reduced bone turnover by 87-92% based on histomorphometric analysis.(4) No increase in fracture rate was observed over time, and the patients treated for 10 yr actually exhibited a lower number of clinical vertebral fractures than patients treated for 5 yr, further dispelling the notion of oversuppression as a side effect of long-term bisphosphonate treatment. The “frozen bone” hypothesis implies that extremely low bone turnover causes accumulation of microdamage, which subsequently leads to fractures.(15,17,18) A recent histological analysis of microcrack accumulation in human iliac crest biopsies with and without treatment with alendronate(23) further challenges this hypothesis, suggesting that oversuppression and resulting microcrack accumulation is of minor significance in human bone. In that study, using well-validated methods for microcrack identification, the investigators found that the number of microcracks in human bone was very small. Moreover, an average of 6 yr of alendronate treatment did not result in any significant difference in microcrack density between bisphosphonate-treated and untreated women.(23) In another recent study by Stepan et al.,(24) bone biopsies obtained from 38 postmenopausal women treated with daily or weekly alendronate at standard doses for a mean duration exceeding 5 yr were compared with biopsies obtained from 28 treatment-naive postmenopausal women with osteoporosis (mean age, 68.0 yr; mean BMD T-score, −1.7 at total hip and −2.8 at lumbar spine; 62% with prevalent fractures). In this study, no differences in crack surface density or crack density were demonstrable between groups (although low femoral neck density was found to be associated with increased microcrack accumulation after alendronate treatment after adjustment for potential confounders). It should be noted, however, that the paucity of microcracks in iliac crest biopsies may not accurately reflect their number in weight-bearing bones such as vertebrae.

Quantitative histomorphometric analysis after tetracycline double labeling was possible in a total of 86 biopsies. Postmenopausal osteoporosis is characterized by a cancellous bone deficit with poor trabecular connectivity, which leads to a reduced bone surface for histomorphometric examination. This reduced bone surface limits the number of active bone remodeling units that can be studied within a given section of a biopsy. As a consequence, limited tetracycline labeling, or even absence of labeling, may be observed in a bone biopsy, particularly in situations involving poor specimen quality, low remodeling rates, or a combination of these factors.

Changes in bone turnover at the tissue level, as measured by mean reduction in Ac.f, were well within the expected range for bisphosphonates, and brought bone turnover down into the range for premenopausal women previously reported by Recker et al.(25) Specifically, the 71% mean reduction of Ac.f observed in this study is comparable to the 75% mean reduction in Ac.f with oral ibandronate after 22 mo,(5) less than the mean 93% reduction in Ac.f seen after 3 yr of alendronate 10 mg once daily treatment,(4) and more pronounced than the 47% decrease in Ac.f with risedronate 5 mg once daily after a treatment period of 1 yr.(6) In the alendronate paper by Chavassieux et al.,(4) both trabecular and endocortical surfaces were assessed, whereas only trabecular bone turnover was measured in other reports.(5,6) Based on the data reported in this paper, however, these differences do not have major influence on the comparisons between drugs, because the reduction in bone turnover at the trabecular and endocortical envelopes was of similar magnitude.

The total remodeling and resorption periods were both significantly prolonged in the zoledronic acid group, an effect that has also been described for other bisphosphonates(4,6) and is in accordance with the known action of these drugs. For alendronate, the decrease in turnover and prolonged bone formation periods has been implicated in the observed increase in average mean degree of mineralization of bone (MDMB).(26) The increased mineralization has also been invoked as an important factor improving the biomechanical strength of bone after alendronate treatment. The data regarding changes in matrix mineralization after treatment with risedronate are conflicting, with one study that used Fourier transform infrared imaging showing no increase compared with placebo,(27) and another study that used νCT with synchrotron radiation reporting a 4.7% increase in mineralization.(28) Our results do not show any increase in mean degree of matrix mineralization after 3 yr of treatment with zoledronic acid. Further studies are needed to elucidate whether the discrepancies pertaining to mineralization after bisphosphonate treatment are real or caused by differences in methodology. The highly significant antifracture efficacy shown across all fractures in this study would denote that changes in matrix mineralization maybe play a lesser role in improvements in biomechanical strength than anticipated after the first alendronate studies.

In patients treated with zoledronic acid, mineralization of newly formed bone was normal, with no evidence of osteomalacia. Osteoid thickness (O. Th), OS/BS, and OV/BV were significantly lower in zoledronic acid-treated patients, which, together with the increase in MAR, argues against the development of osteomalacia in these patients.

It is noteworthy that the MAR was significantly increased in the zoledronic acid group compared with placebo and that this trend was also observed for the subgroup analyses of strata I and II. Such a difference was not shown in previous histomorphometric analyses of bisphosphonates.(4-6) MAR reflects the bone-forming capacity of individual teams of osteoblasts at the BMU level. This result refutes the presence of osteoblast suppression by zoledronic acid. Rather, osteoblast activity was improved with zoledronic acid over placebo.

The fact that this difference was shown irrespective of concomitant osteoporosis therapy (Table 4) indicates that the improvement was caused by zoledronic acid and not adjunctive therapies. Previous histomorphometric studies did not reveal any consistent changes in MAR. In the histomorphometric studies on risedronate(6) and alendronate,(4) MAR was found to be similar in the treatment and placebo groups for the marketed doses of the drug. In the alendronate study,(4) the group receiving 20 mg daily followed by a later shift to 5 mg daily exhibited a marginal increase after 36 mo. No difference, however, was demonstrable after 24 mo. For ibandronate,(5) in contrast, a marginally significant increase in MAR was reported for the group receiving intermittent dosing, whereas no difference was demonstrable in the group receiving daily dosing. In accordance with previous histomorphometric analyses on bisphosphonates,(4-6) no specific missing value was assigned for MAR in patients with missing double label in this study. Potentially, this could have biased the MAR average in the zoledronic acid group. However, the same bias would have affected MAR estimates in other bisphosphonates trials, which makes comparison with the latter studies reasonable.

With the caveat that bisphosphonate doses used in cell culture studies may be excessive, several studies have reported that zoledronic acid stimulates bone marrow stromal cell proliferation(29,30) and increases mineralization of osteoblast cultures(30) and osteoblastic synthesis of bone sialoprotein and osteoprotegerin.(31,32) These findings are consistent with stimulatory effects on osteoblastic matrix synthesis and mineralization, both of which could contribute to an increase in MAR.

Concomitant treatment with non-bisphosphonate therapies (excluding strontium and PTH) did not significantly affect the tissue level response to zoledronic acid effects on bone remodeling or bone quality. When comparing reduction in Ac.f between the two strata, the reduction in stratum II seemed less than in stratum I. This difference was, however, not significant. Stratum II exhibited a somewhat lower turnover rate at baseline (Ac.f = 0.24) than stratum I (Ac.f. = 0.27), which is in accordance with the concomitant antiresorptive therapy. Other indices reflecting bone turnover (e.g., MS/BS, OS/BS) showed more pronounced reductions in stratum II. The reduction in BV/TV was also greater in stratum II, but other indices reflecting changes in trabecular structure (e.g., Tb.N and Ct. Th) did not follow the same trend. In summary, there was no consistent trend with respect to differences between strata I and II patients.

This study is the largest nonpaired histomorphometric bone biopsy analysis undertaken in a bisphosphonate-treated population and was conducted using state-of-the-art techniques. Limitations include the fact that patients who consented to participate in this substudy had slightly higher hip BMD values (spine BMD was, however, comparable). In addition, paired biopsies (before and after treatment) were not available, and analyses were based solely on samples derived from the iliac crest, which may not be representative of bone status at other locations. However, comparing even nonpaired bone biopsies between such large sample sizes of treated and nontreated groups has advantages in terms of statistical power(33) and offers a great deal of scientific evidence for showing the safety, as well as the mechanism of action, of zoledronic acid in the postmenopausal population.

In conclusion, 3 yr of treatment with intravenous zoledronic acid 5 mg once yearly yielded a significant 63% median reduction in bone turnover, resulting in significant decreases in surfaces undergoing active remodeling and preservation of bone architecture. Moreover, zoledronic acid increased osteoblastic MAR. No evidence for adverse effects on bone safety in the form of excessive reduction of bone turnover, woven bone formation, or osteomalacia were demonstrable in patients treated with zoledronic acid. Concomitant treatment with other non-bisphosphonate osteoporosis therapies did not significantly alter the bone response to zoledronic acid.

Acknowledgements

The authors thank Novartis Pharma AG, Basel, Switzerland, for financial support of this study and provision of the study drug. We also thank BioScience Communications of New York, NY, USA, for editorial assistance in the preparation of this manuscript.

Ancillary