Bone Histomorphometry of Transiliac Paired Bone Biopsies After 6 or 12 Months of Treatment With Oral Strontium Ranelate in 387 Osteoporotic Women: Randomized Comparison to Alendronate

Authors


ABSTRACT

Preclinical studies indicate that strontium ranelate (SrRan) induces opposite effects on bone osteoblasts and osteoclasts, suggesting that SrRan may have a dual action on both formation and resorption. By contrast, alendronate (ALN) is a potent antiresorptive agent. In this multicenter, international, double-blind, controlled study conducted in 387 postmenopausal women with osteoporosis, transiliac bone biopsies were performed at baseline and after 6 or 12 months of treatment with either SrRan 2 g per day (n = 256) or alendronate 70 mg per week (n = 131). No deleterious effect on mineralization of SrRan or ALN was observed. In the intention-to-treat (ITT) population (268 patients with paired biopsy specimens), changes in static and dynamic bone formation parameters were always significantly higher with ALN compared with SrRan at month 6 (M6) and month 12 (M12). Static parameters of formation were maintained between baseline and the last value with SrRan, except for osteoblast surfaces, which decreased at M6. Significant decreases in the dynamic parameters of formation (mineralizing surface, bone formation rate, adjusted apposition rate, activation frequency) were noted at M6 and M12 in SrRan. Compared with ALN, the bone formation parameters at M6 and M12 were always significantly higher (p < 0.001) with SrRan. ALN, but not SrRan, decreased resorption parameters. Compared with the baseline paired biopsy specimens, wall thickness was significantly decreased at M6 but not at M12 and cancellous bone structure parameters (trabecular bone volume, trabecular thickness, trabecular number, number of nodes/tissue volume) were significantly decreased at M12 with SrRan; none of these changes were significantly different from ALN. In conclusion, this large controlled paired biopsy study over 1 year shows that the bone formation remains higher with a lower diminution of the bone remodeling with SrRan versus ALN. From these results, SrRan did not show a significant anabolic action on bone remodeling. © 2014 American Society for Bone and Mineral Research.

Introduction

The objective of anti-osteoporotic treatment is to prevent new fracture by improving bone strength. Bone is a composite material, and the integrity of each component controlled by the bone remodeling contributes to the bone strength.[1] Any deleterious change in the material or structural components of bone or the inability of bone remodeling to adapt these components to load produces bone fragility.[2] Postmenopausal osteoporosis results—at least in part—from an imbalance in bone resorption and formation, leading to bone loss, with trabecular thinning and loss of connectivity, along with increased cortical porosity.

Therapies for osteoporosis improve bone strength and reduce fracture risk by either inhibiting resorption (eg, with bisphosphonates) or by stimulating formation (eg, with teriparatide). Strontium ranelate (SrRan) could be a third class of agent with a dual mechanism of action capable of both stimulating bone formation and reducing bone resorption. This is supported by in vitro studies, which show stimulation of osteoblast differentiation and activity[3] and diminution of osteoclastogenesis and osteoclast resorptive activity.[4, 5] Additionally, in vivo experimental studies in different rodent models have confirmed this dual action for SrRan.[6-8] Large clinical trials have demonstrated its antifracture efficacy, associated with increased bone mineral density (BMD) with 2 g per day SrRan.[9-13] Serum bone alkaline phosphatase, a marker of bone formation, slightly but significantly increased from the third month of treatment to 3 years, and no significant difference in bone formation parameters was observed after 6 months compared with teriparatide, a bone anabolic agent, despite a higher mineralizing surface/bone surface (MS/BS) value in the teriparatide group.[14] A cross-sectional histomorphometric clinical study failed to show clear stimulation of formation or inhibition of resorption after 1 to 5 years versus placebo, but in the few paired biopsy specimens available, 3D microstructure analysis suggested an increase in bone volume over 3 years of strontium ranelate treatment.[15] The lack of large paired biopsy specimen studies may explain the difficulties in establishing a significant effect of SrRan. The purpose of the present study was to better understand the early mechanisms of action of SrRan on bone tissue. Paired biopsy specimens were analyzed by histomorphometry in a large population of postmenopausal osteoporotic women. One transiliac bone biopsy was obtained at baseline and a second after either 6 or 12 months.

Materials and Methods

Study design

This was a phase III, multicenter, international, randomized, double-blind, double-dummy study (protocol CL3-12911-025) to assess the effects of 6 or 12 months' administration of 2 g per day SrRan compared with 70 mg per week of ALN in the treatment of postmenopausal women with osteoporosis. A total of 38 centers in 13 countries were involved in this study. The randomization was unbalanced with a ratio 2:1 (SrRan/ALN) and stratified per country. A balanced randomization in each treatment group was used to determine the visit of the second biopsy (month 6 [M6] or month 12 [M12]) for each patient. The therapeutic units and the time of the post-baseline biopsy were allocated to patients by an interactive voice system (Cardinal Systems, Paris, France).

The study received ethical review board approval at all sites, and all patients gave informed written consent. The study was conducted in accordance with the principles stated in the Declaration of Helsinki, and was performed according to the rules of ICH guideline for Good Clinical Practice.

Study population

Ambulatory osteoporotic women aged >50 years, postmenopausal for at least 3 years, were included if a T-score measured with dual-energy X-ray absorptiometry (DXA) at the lumbar L1 to L4 and/or femoral neck level was ≤−2.5 SD, or if a T-score at the lumbar L1 to L4 and/or femoral neck level was ≤−1 SD with at least one prevalent low-trauma fracture. Patients were excluded if they had a history of disease affecting bone metabolism other than osteoporosis; had undergone a previous bone biopsy within 1 year; had a previous double-labeling with tetracycline or tetracycline treatment within 1 year; had a history of progressive cancer during the past 5 years; had a contra-indication for bone biopsy (coagulation abnormality, anticoagulant medication, hip prosthesis, or severe obesity); had significant and progressive hyperthyroidism diagnosed within the previous 1 year; had severe malabsorption or severe and documented liver insufficiency; had significantly impaired renal function (creatinine clearance <30 mL/min); had vitamin D insufficiency defined as 25(OH) vitamin D <15 ng/mL; or had a history of severe alcohol abuse. Patients with a personal past history or an increased risk for venous thromboembolic events were not selected. Patients treated with glucocorticoids (>2 g cumulative dose) within the year prior to selection, calcitriol, bisphosphonates, or any drugs developed for bone diseases or interfering with bone metabolism were also excluded.

A total of 387 patients aged 50 to 84 years were randomized; 256 patients received 2 g/day SrRan and 131 patients received 70 mg/week ALN as follows: one sachet of 2 g/day SrRan at bedtime and one capsule of placebo ALN once a week upon arising, or one sachet of placebo SrRan at bedtime and one capsule of ALN once a week upon arising. All patients were supplemented with 1000 mg elemental calcium and 800 IU vitamin D per day. Therapeutic units were supplied by Les Laboratoires Servier Industrie (Gidy, France).

Bone histomorphometry

A transiliac bone biopsy was performed with a 7.5-mm inner diameter trephine in all patients at baseline. A second bone biopsy specimen was collected randomly after 6 or 12 months of treatment on the opposite side. Before bone biopsy, patients received a double tetracycline labeling, 600 mg per day demeclocycline hydrochloride before the first biopsy and 1 g per day tetracycline hydrochloride before the second according to the following schedule: 2 days on, 10 days off, 2 days on. Biopsies were performed within 5 to 7 days of the last dose of tetracycline. The bone biopsy specimens were stored and transported in 70% ethanol, in the dark to prevent the labels from fading, to the central histomorphometry laboratory (INSERM UMR 1033, Lyon, France) for processing, reading, and interpretation of the results.

After fixation in 70% ethanol, dehydration in 100% ethanol, specimens were embedded in methylmetacrylate. Three sets of 8-µm-thick sections were cut 200 µm apart in the central part of the sample. In each set, sections were stained with modified Goldner's trichrome, solochrome cyanin R, toluidine blue, or May-Grünwald-Giemsa. Some sections were left unstained for the measurement of the tetracycline labels under fluorescence.[16]

Each bone specimen underwent a qualitative analysis to assess mineralization impairment, osteomalacia, the presence of any abnormalities of bone marrow (eg, the presence of lymphoid nodes, fibrosis, or metastases), and the type of bone (eg, woven or lamellar bone).

The quantitative analysis was performed on all complete and unbroken samples. The histomorphometry measurements were done on the whole tissue including the cancellous (Cn), endocortical (Ec), and cortical (Ct) envelopes of three sections (one per set) with a total cancellous bone tissue volume ≥20 mm2. For all the analysis, the investigators were blinded to treatment allocation.

The parameters of bone structure and connectivity (strut analysis after skeletization) were measured with an automatic image analyzer (MorphoExpert, Explora Nova, La Rochelle, France). The static parameters reflecting resorption, formation, and the dynamic parameters of bone formation and mineralization were measured by using a semiautomatic image analyzer (Tablet'Measure, Explora Nova, La Rochelle, France). The abbreviations of the bone histomorphometric parameters used were recommended by the ASBMR Histomorphometric Nomenclature Committee.[17] All measured thicknesses (except cortical thickness) were multiplied by π/4. Structural parameters were cortical thickness (Ct.Th, µm) and porosity (Ct.Po, %), and cancellous bone volume (Cn-BV/TV, %). The parameters of microarchitecture (trabecular thickness [Tb.Th, µm], number [Tb.N, #/mm2], and separation [Tb.Sp, µm]) were derived from area and perimeter measurements according to the Parfitt's formulae.[18] Bone resorption was assessed with measurements of eroded surfaces (ES/BS, %), osteoclast number (Oc.N/BS, /mm) and surface (Oc.S/BS, %), mean and maximum erosion depth (mean and maxE.De, µm), and eroded volume (EV/BV, %). Static bone formation was reflected by osteoblast surfaces (Ob.S/BS, %), osteoid surfaces (OS/BS, %), volume (OV/BV, %), and thickness (O.Th, µm). Osteoid seams with a minimum width of 2.5 µm were measured. All these parameters were measured on Goldner-stained sections. The mineral apposition rate (MAR, µm/day) and the ratio of mineralizing surface to bone surface (MS/BS, % calculated as double plus half of single-labeled surfaces) were analyzed on unstained sections under ultraviolet light. The mean wall thickness (W.Th, µm) was measured on Solochrome cyanin R-stained sections, under polarized light. Bone formation rate (BFR/BS [µm3/µm2/day] = (MS/BS) × MAR), adjusted apposition rate (Aj.AR [µm/day] = BFR/OS), formation period (FP [days] = W.Th/Aj.AR), mineralization lag time (Mlt [days] = O.Th/Aj.AR), activation frequency (Ac.f [per year] = (BFR/BS)/W.Th), and osteoid maturation time (OMT [days] = O.Th/MAR) were calculated. The number of nodes/tissue volume (N.Nd/TV, #/mm2) was obtained after strut analysis. All parameters were measured in the cancellous area. MAR was also measured in cortical bone.

In the analysis, for biopsy specimens missing label in the cancellous and endocortical bone analyzed, the value of MS/BS was 0 and the parameters derived, MAR and BFR/BS were missing values.[19] When only single labels were present, MAR and BFR/BS were considered as missing values.[19]

Statistical analysis

The primary endpoint was cancellous MS/BS. The sample size was estimated on the change of MS/BS from baseline to the value at the 6-month visit for patients having the second biopsy at 6 months, or 12 months for patients having the second biopsy at 12 months. Taking into account that a maximum effect is observed very rapidly with ALN, as soon as 6 months, data from ALN group biopsy specimens at 6 and 12 months were pooled.

All efficacy analyses were performed in intention-to-treat (ITT) patients having taken at least one dose of study medication, with a baseline and post-baseline evaluation of primary criterion cancellous MS/BS. The main characteristics of patients, including demography, prognostic factors, baseline values of endpoints, as well as status of patients and concomitant treatment intake, were described by treatment group and overall. Changes in both SrRan groups were compared with the change in ALN group using a general linear model using Dunnett's multiple comparison procedure with baseline and country as covariates. Adjusted difference of change between two-group comparison was estimated, and standard error and the 95% confidence interval (CI) have been computed. Intragroup comparison within both SrRan groups and ALN group was performed using a two-sided Student's t test for paired samples. Sensitivity analyses were conducted with a nonparametric approach: covariance analysis with the same model, Wald test with Bonferroni correction for intergroup comparison, and Wilcoxon signed-rank test for intragroup comparison. A p < 0.05 was considered as significant.

The following descriptive statistics were provided, depending on the nature of variables: Quantitative variable is the number of observed values, mean and standard deviation. Qualitative or ordinal variable is the number and percentage per class. Safety was analyzed using descriptive statistics.

Results

A total of 387 patients were included and randomized in the study (n = 256 SrRan; n = 131 ALN). Overall, 49 patients (12.7%) withdrew from the study, mainly because of adverse events (8.6% in the SrRan group versus 3.1% in the ALN group) or nonmedical reasons (6.6% versus 3.1%, respectively), and 1 patient was lost to follow-up. A total of 337 patients completed the study: 215 patients (84.0%) in the SrRan group and 122 patients (93.1%) in the ALN group (Fig. 1). The full analysis set comprised 268 patients (n = 90 in SrRan M6, n = 89 in SrRan M12, and n = 89 in ALN), ie, 69.3% of the randomized population. Most exclusions from the full analysis set were attributable to the absence of baseline and/or post-baseline evaluation of the primary endpoint MS/BS because of the quality of the biopsy ie, broken or incomplete sample with a cancellous bone area lower than 20 mm2. Main baseline characteristics of patients are described on the full analysis set (Table 1). Patients' ages ranged from 50 to 84 years with a mean of 63.7 ± 7.1 years. Hypertension was the most frequently reported medical history (33.2%), followed by osteoarthritis (32.1%). Consistently with the medical history, the most commonly concomitant treatments at inclusion were nonsteroidal anti-inflammatory drugs and anti-rheumatic products (25.7%) and agents acting on the renin-angiotensin system (20.1%). All patients presented primary postmenopausal osteoporosis. Mean time since diagnosis was 22.3 ± 37.4 months. Sixty-five patients (24.3%) reported at least one prevalent osteoporotic fracture. A total of 118 patients (44%) reported at least one previous treatment for osteoporosis, mainly mineral supplement (calcium) (29.9%), calcium combinations with other drugs (12.7%), and vitamin D and analogues (15.3%). At baseline, there were no relevant differences in patient characteristics in the full analysis set (Table 1).

Figure 1.

Flow diagram of the progress through the phases of the study.

Table 1. Baseline Characteristics of Patients in the FAS
 Strontium ranelate M6 (n = 90)Strontium ranelate M12 (n = 89)Alendronate M6 + M12 (n = 89)
  1. BMI = body mass index; BMD = bone mineral density.
  2. Values are means ± SD or n (%).
Age (years)64.0 ± 7.362.9 ± 7.264.2 ± 6.7
Ethnic origin (n [%])
White77 (86.5)77 (86.5)71 (79.8)
Black3 (3.4)3 (3.4)2 (2.2)
Asian1 (1.1)1 (1.1)0 (0.0)
Other8 (9.0)8 (9.0)16 (18.0)
Unknown100
BMI (kg/m2)25.34 ± 3.6525.90 ± 3.8825.48 ± 3.78
Time since osteoporosis diagnosis (months)22.50 ± 37.7121.10 ± 42.1923.15 ± 31.84
Time since menopause (years)16.84 ± 8.4415.02 ± 8.1915.95 ± 6.80
Lumbar L1 to L4 BMD (mg/cm2)777.69 ± 82.04789.56 ± 85.00773.97 ± 88.33
T-score−2.95 ± 0.69−2.85 ± 0.72−2.98 ± 0.75
Femoral neck BMD (mg/cm2)662.20 ± 74.1688.94 ± 78.6681.83 ± 86.99
T-score−2.31 ± 0.61−2.09 ± 0.65−2.15 ± 0.72
Total hip BMD (mg/cm2)731.23 ± 89.41758.92 ± 91.25742.08 ± 98.66
T-score−1.83 ± 0.73−1.60 ± 0.741.74 ± 0.80

Qualitative analysis

Most patients had a complete biopsy specimen: 248 patients (97.3%) and 127 (96.2%) at baseline and 188 patients (90.8%) and 101 patients (87.1%) at the study end in the SrRan and ALN groups, respectively. Fourteen biopsy specimens had only one complete cortex with a sufficient cancellous bone area. Four samples had only cancellous bone measurable. All biopsy specimens had a normal lamellar texture. Microcallus, made of woven bone and corresponding to a microfracture repair, was observed in seven biopsy specimens at baseline (three SrRan and four ALN) and five biopsy specimens (two SrRan and three ALN) post-baseline. There was no evidence for osteomalacia, Paget's disease, or metastasis in any of the biopsy specimens. Benign small isolated lymphoid nodes were observed in five specimens at baseline (three SrRan and two ALN) and in 21 post-treatment biopsy specimens (19 [9.7%] SrRan and one [0.9%] ALN). For one patient in the SrRan group, an abnormal marrow cell proliferation of lymphocytes was observed at baseline and was confirmed at 12 months with the presence of numerous lymphoid nodes. At baseline, one biopsy specimen had no label. In SrRan groups, four biopsy specimens had no label (three M6 and one M12) and two had only single labels (two M6). In ALN groups, 31 biopsy specimens had no label (14 M6 and 17 M12), five had only single labels (two M6 and three M12), and six had only double labels (four M6 and two M12).

Quantitative bone histomorphometry

A total of 268 paired biopsy specimens were complete and assessable for bone histomorphometry, 90 and 89 in the SrRan M6 and M12 groups, respectively, and 43 and 46 in the ALN M6 and M12, respectively. Because no difference was observed between M6 and M12 in the ALN group (Table 2), the results were pooled.

Table 2. Histomorphometric Parameters in Postmenopausal Women Treated With Alendronate for 6 or 12 Months and With Paired Evaluable Bone Biopsy Specimens
 ALN-M6 (n = 43)ALN-M12 (n = 46)
BaselineM6BaselineM12
  • Cn = cancellous; BV = bone volume; Tb = trabecular; Th = thickness; Sp = separation; MWT = mean wall thickness; ES = eroded surface; Oc.S = osteoclast surface; Oc.N = osteoclast number; EV = eroded volume; E.De = erosion depth; Ob.S = osteoblast surface; OS = osteoid surface; OV = osteoid volume; MS = mineralizing surface; BFR/BS = bone formation rate; Aj.AR = adjusted apposition rate; Ac.f = activation frequency.
  • Values are mean ± SD. Within-group analysis: mean comparison between baseline and M6/M12 visit using Wilcoxon T test.
  • *p < 0.05;
  • **p < 0.01;
  • ***p < 0.001.
Cn-BV/TV (%)16.1 ± 6.315.7 ± 6.315.4 ± 5.416.6 ± 6.7
Tb.Th (µm)122.1 ± 33.2123.5 ± 37.9122.8 ± 33.2128.3 ± 45.1
Tb.N (/mm2)1.29 ± 0.281.26 ± 0.301.24 ± 0.321.29 ± 0.31
Tb.Sp (µm)695.8 ± 242.8733.4 ± 288.4742.9 ± 301.0700.1 ± 263.5
Cn-W.Th (µm)29.7 ± 3.928.9 ± 3.730.2 ± 3.230.0 ± 3.1
Cn-ES/BS (%)2.5 ± 1.51.6 ± 1.6***1.9 ± 0.91.6 ± 1.2
Cn-Oc.N/BS (/mm)0.02 ± 0.020.01 ± 0.04**0.02 ± 0.020.01 ± 0.01**
Cn-Oc.S/BS (%)0.08 ± 0.130.07 ± 0.26*0.08 ± 0.100.05 ± 0.09
Cn-EV/BV (%)0.5 ± 0.40.3 ± 0.3***0.4 ± 0.20.3 ± 0.2
Cn-E.DeMax (µm)15.4 ± 3.414.4 ± 4.314.1 ± 2.614.1 ± 3.7
Cn-E.DeMean (µm)9.6 ± 2.39.1 ± 2.78.7 ± 1.58.7 ± 2.4
Cn-Ob.S/BS (%)1.7 ± 1.90.09 ± 0.2***1.7 ± 1.60.05 ± 0.14***
Cn-OS/BS (%)9.03 ± 6.35.8 ± 6.8**9.0 ± 4.93.2 ± 3.2***
Cn-OV/BV (%)2.2 ± 1.70.9 ± 0.9***2.3 ± 1.50.5 ± 0.6***
Cn-O.Th (µm)10.4 ± 2.28.3 ± 1.3***11.7 ± 2.28.4 ± 1.2***
Cn-MAR (µm/d)0.59 ± 0.090.53 ± 0.120.61 ± 0.100.58 ± 0.09
Cn-MS/BS (%)4.6 ± 3.50.2 ± 0.3***6.1 ± 3.60.3 ± 0.6***
Cn-BFR/BS (µm3/µm2/d)0.027 ± 0.0220.002 ± 0.002***0.036 ± 0.0220.004 ± 0.003**
Cn-Aj.AR (µm/d)0.32 ± 0.180.10 ± 0.12**0.43 ± 0.150.13 ± 0.13***
Cn-Ac.f (/year)0.34 ± 0.260.03 ± 0.04***0.43 ± 0.240.04 ± 0.03***

Bone safety

In the SrRan groups, primary mineralization was normal as shown by the absence of diminution of MAR in M6 and M12 biopsy specimens compared with baseline values. Indeed, MAR was higher in cortical than in endosteal bone. There was no sign of osteoid accumulation because O.Th significantly decreased at M6 and M12 (Fig. 2). Mlt slightly increased (p < 0.05), although Omt did not change compared with baseline. In the ALN group, MAR (cortical and cancellous), O.Th, and Omt significantly decreased and Mlt significantly increased after treatment compared with baseline (p < 0.05 to 0.001). Compared with the SrRan M6 and M12 groups, parameters reflecting bone mineralization were significantly lower in the ALN group (p < 0.01 to 0.001), except Omt at M6 and endosteal MAR.

Figure 2.

Histomorphometric parameters of bone mineralization in postmenopausal women treated with strontium ranelate or alendronate for 6 and 12 months with paired evaluable bone biopsy specimens. Values are mean ± SD. Within-group analysis: mean comparison between baseline and M6/M12 visit using two-sided Student's T test for paired sample. *p < 0.05; **p < 0.01; ***p < 0.001. Between-group analysis using a general linear model. The p value = multiple comparison p value (Dunnett's test).

Bone formation

In the SrRan M6 group, the primary endpoint MS/BS decreased from 5.53% ± 4.54% to 2.94% ± 3.72%, and in the SrRan M12 group, MS/BS decreased from 6.63% ± 4.25% to 4.91% ± 4.15%, whereas in the pooled alendronate group, the MS/BS decreased from 5.40% to 0.24%. The mean changes in the primary endpoint MS/BS from baseline to last value were significantly lower in the SrRan than ALN groups (p < 0.001 for the three intragroup comparisons with a significant between-group difference in favor of SrRan) (Table 3).

Table 3. Histomorphometric Parameters of Bone Formation in Postmenopausal Women Treated With Strontium Ranelate or Alendronate for 6 or 12 Months and With Paired Evaluable Bone Biopsy Specimens
 SrRan-M6SrRan-M12ALNBetween-group difference SrRan M6 versus ALNBetween-group difference SrRan M12 versus ALN
BaselineM6BaselineM12BaselineM6 + M12Estimate (SE)p valueEstimate (SE)p Value
  • Cn = cancellous; Ob.S = osteoblast surface; OS = osteoid surface; OV = osteoid volume; MS = mineralizing surface; BFR/BS = bone formation rate; Aj.AR = adjusted apposition rate; Ac.f = activation frequency.
  • Values are mean ± SD. Within-group analysis: mean comparison between baseline and M6/M12 visit using two-sided Student's t test for paired sample.
  • *p < 0.05;
  • **p < 0.01;
  • ***p < 0.001.
  • Between-group analysis: Estimate (standard error) of adjusted means difference - SrRan minus ALN (baseline and country as fixed effects) using a general linear model. p value = multiple comparison p value (Dunnett's test).
Cn-Ob.S/BS (n)(90)(89)(89)    
(%)1.8 ± 1.591.23 ± 1.41**1.89 ± 1.661.71 ± 1.811.74 ± 1.740.07 ± 0.17***1.18 (0.20)<0.0011.68 (0.20)<0.001
Cn-OS/BS (n)(90)(89)(89)    
(%)9.31 ± 5.8910.27 ± 8.3811.10 ± 6.2810.05 ± 6.499.10 ± 5.604.46 ± 5.46***5.80 (1.03)<0.0015.24 (1.05)<0.001
Cn-OV/BV (n)(90)(89)(89)    
(%)2.25 ± 1.402.04 ± 1.732.61 ± 1.642.24 ± 1.652.24 ± 1.600.69 ± 0.81***1.35 (0.22)<0.0011.51 (0.22)<0.001
Cn-MS/BS (n)(90)(89)(89)    
(%)5.53 ± 4.542.94 ± 3.73***6.63 ± 4.254.91 ± 4.15***5.40 ± 3.590.24 ± 0.46***2.73 (0.48)<0.0014.65 (0.49)<0.001
Cn-BFR/BS (n)(76)(84)(37)    
(µm3/µm2/d)0.04 ± 0.030.02 ± 0.02***0.04 ± 0.030.03 ± 0.03*0.03 ± 0.020.003 ± 0.003***0.02 (0.00)<0.0010.03 (0.00)<0.001
Cn-Aj.AR (n)(75)(84)(37)    
(µm/day)0.38 ± 0.170.23 ± 0.17***0.38 ± 0.170.31 ± 0.15**0.38 ± 0.190.12 ± 0.12***0.10 (0.03)0.0030.19 (0.03)<0.001
Cn-Ac.f (n)(76)(84)(37)    
(/year)0.43 ± 0.320.26 ± 0.30***0.49 ± 0.300.40 ± 0.32*0.38 ± 0.250.04 ± 0.03***0.21 (0.06)<0.0010.35 (0.06)<0.001

In the SrRan group, a decrease in MS/BS from baseline to last value was observed in 67.6% of the patients in the SrRan group, whereas 32.4% had an increase in MS/BS. The decrease in the mean cancellous MS/BS was significantly lower in the SrRan M12 group than in the SrRan M6 group (p < 0.001). Similar results were observed in the SrRan groups concerning other dynamic parameters of bone formation, whereas static parameters of formation did not change between baseline and end with SrRan, except for Ob.S/BS, which decreased at M6. In a subgroup of patients with low baseline remodeling, ie, baseline MS/BS lower than the median value of 5%, MS/BS, BFR/BS, and Ac.f significantly increased (p < 0.05) after 12 months versus patients with high baseline remodeling.

In the ALN group, all parameters of bone formation significantly decreased in cancellous and endosteal areas (0.001 ≤ p ≤ 0.0001), reflecting a marked diminution of bone turnover. A decrease in MS/BS from baseline was observed in all patients but one, and Ac.f was 10-fold lower after ALN treatment (Table 3). During this study, the bone formation parameters decreased to a greater extent with ALN than with SrRan (p < 0.001 for all comparisons).

Similar results were observed in the endosteal area.

Bone resorption

Treatment with SrRan did not significantly modify resorption parameters, regardless of bone compartment. After ALN treatment, ES/BS, EV/BV Oc.S/BS, and Oc.N/BS significantly decreased in cancellous and endosteal bone (p < 0.001) (Table 4). Compared with SrRan, significant between-group differences were observed for most resorption parameters.

Table 4. Histomorphometric Parameters of Bone Resorption in Postmenopausal Women Treated With Strontium Ranelate or Alendronate for 6 or 12 Months With Paired Evaluable Bone Biopsy Specimens
 SrRan-M6SrRan-M12ALNBetween-group difference SrRan M6 versus ALNBetween-group difference SrRan M12 versus ALN
BaselineM6BaselineM12BaselineM6 + M12Estimate (SE)p valueEstimate (SE)p value
  • Cn = cancellous; ES = eroded surface; Oc.S = osteoclast surface; Oc.N = osteoclast number; EV = eroded volume; E.De = erosion depth.
  • Values are mean ± SD. Within-group analysis: mean comparison between baseline and M6/M12 visit using two-sided Student's t test for paired sample.
  • *p < 0.05; **p < 0.01;
  • ***p < 0.001.
  • Between-group analysis: Estimate (standard error) of adjusted means difference - SrRan minus ALN (baseline and country as fixed effects) using a general linear model. p value = multiple comparison p value (Dunnett's test).
Cn-ES/BS (n)(90)(89)(89)    
(%)2.29 ± 1.312.58 ± 1.822.23 ± 1.202.45 ± 1.602.20 ± 1.251.58 ± 1.37***0.99 (0.23)<0.0010.91 (0.24)<0.001
Cn-Oc.S/BS (n)(90)(89)(89)    
(%)0.09 ± 0.160.09 ± 0.150.08 ± 0.120.11 ± 0.160.08 ± 0.110.06 ± 0.200.03 (0.03)0.3700.06 (0.03)0.040
Cn-Oc.N/BS (n)(90)(89)(89)    
(/mm)0.02 ± 0.030.02 ± 0.030.02 ± 0.020.02 ± 0.030.02 ± 0.020.01 ± 0.030.00 (0.00)0.0710.01 (0.00)0.009
Cn-EV/BV (n)(90)(89)(89)    
(%)0.46 ± 0.300.56 ± 0.49*0.43 ± 0.300.52 ± 0.400.44 ± 0.320.31 ± 0.28***0.26 (0.06)<0.0010.22 (0.06)<0.001
Cn-E.DeMax (n)(90)(89)(89)    
(µm)15.00 ± 3.3215.51 ± 3.4515.62 ± 3.5915.27 ± 2.8114.80 ± 3.0314.26 ± 4.001.26 (0.51)0.0281.08 (0.52)0.075
Cn-E.DeMean (n)(90)(89)(89)    
(µm)9.37 ± 2.0810.01 ± 2.379.61 ± 2.149.45 ± 1.799.17 ± 1.958.95 ± 2.531.04 (0.34)0.0050.50 (0.34)0.255

Bone structure

For trabecular parameters, no statistical significant between-group difference was observed between SrRan and ALN groups, but there was a decrease over time within the SrRan M12 group in BV/TV (p = 0.011), Tb.Th (p = 0.024), and Nd.N/TV (p < 0.001), a decrease in wall thickness in the SrRan M6 group (p = 0.007), and an increase in Tb.Sp in both SrRan groups (p = 0.035 and p = 0.034 in M6 and M12 groups, respectively). Ct.Po was significantly decreased in the ALN group with a significant between-group difference compared with the SrRan groups (p < 0.05 for both SrRan groups) (Table 5).

Table 5. Histomorphometric Parameters of Bone Structure in Postmenopausal Women Treated With Strontium Ranelate or Alendronate for 6 or 12 Months With Paired Evaluable Bone Biopsy Specimens
 SrRan-M6SrRan-M12ALNBetween-group difference SrRan M6 versus ALNBetween-group difference SrRan M12 versus ALN
BaselineM6BaselineM12BaselineM6 + M12Estimate (SE)p valueEstimate (SE)p value
  • Ct.Th = cortical thickness; Ct.Po = cortical porosity; BV/TV = bone volume; Tb.Th = trabecular thickness; Tb.N = trabecular number; Tb.Sp = trabecular separation; NdN/TV = number of nodes/tissue volume; W.Th = wall thickness.
  • Values are mean ± SD. Within-group analysis: mean comparison between baseline and M6/M12 visit using two-sided Student's t test for paired sample.
  • *p < 0.05;
  • **p < 0.01;
  • ***p < 0.001.
  • Between-group analysis: Estimate (standard error) of adjusted means difference - SrRan minus ALN (baseline and country as fixed effects) using a general linear model. p value = multiple comparison p value (Dunnett's test).
Ct.Th (n)(88)(88)(88)    
(µm)645.53 ± 225.61679.48 ± 265.68678.78 ± 246.77681.96 ± 260.76686.71 ± 230.71674.61 ± 233.8426.45 (32.62)0.63121.68 (32.90)0.735
Ct.Po (n)(88)(88)(88)    
(%)5.00 ± 2.465.27 ± 3.865.53 ± 3.195.32 ± 3.595.21 ± 2.364.17 ± 2.78**1.18 (0.50)0.0341.17 (0.50)0.039
BV/TV (n)(90)(89)(89)    
(%)16.39 ± 4.9816.16 ± 6.5217.32 ± 5.7715.57 ± 5.73*15.75 ± 5.8215.97 ± 6.19−0.07 (0.83)0.994−0.84 (0.85)0.504
Tb.Th (n)(90)(89)(89)    
(µm)123.73 ± 28.70123.32 ± 30.61129.25 ± 34.10120.23 ± 29.60*122.49 ± 33.16124.74 ± 40.15−2.01 (4.83)0.882−6.43 (4.91)0.321
Tb.N (n)(90)(89)(89)    
(/mm2)1.33 ± 0.301.29 ± 0.351.34 ± 0.261.29 ± 0.131.27 ± 0.301.27 ± 0.03−0.01 (0.04)0.902−0.02 (0.038)0.847
Tb.Sp (n)(90)(89)(89)    
(µm)675.71 ± 218.09730.19 ± 336.69*658.97 ± 231.94712.63 ± 245.89*720.70 ± 275.01719.51 ± 274.8338.14 (34.97)0.44522.32 (35.52)0.752
Nd.N/TV (n)(90)(89)(89)    
(/mm2)1.05 ± 0.500.99 ± 0.631.13 ± 0.480.93 ± 0.54***0.96 ± 0.520.93 ± 0.470.03 (0.07)0.855−0.05 (0.08)0.757
W.Th (n)(90)(89)(89)    
(µm)30.45 ± 3.5029.00 ± 3.38**30.53 ± 3.4929.67 ± 2.8729.87 ± 3.4829.44 ± 3.41−0.50 (0.47)0.4590.37 (0.47)0.650

Safety

The frequency of patients who reported at least one emergent adverse event was close in both treatment groups: 175 patients (68.6%) in the SrRan group and 86 patients (65.2%) in the ALN group. The most frequently affected system organ classes were infections and infestations (28.2% of the patients in the SrRan group versus 25.8% in the ALN group) and gastrointestinal disorders (23.5% versus 21.2%, respectively). The rate of treatment-related emergent adverse events was similar in both groups (18.8% with SrRan and 20.5% with ALN), and these events were most frequently related to gastrointestinal disorders (11.8% and 13.6%, respectively).

Discussion

The present histomorphometric analysis performed on a large population of postmenopausal osteoporotic patients showed that the bone formed during SrRan or ALN treatment was lamellar without evidence of mineralization defect. The static and dynamic parameters of bone formation were significantly higher after 6 or 12 months of SrRan treatment than ALN. The analysis of paired transiliac bone biopsy specimens within SrRan groups, however, showed that at 6 months, Ob.S/BS, MS/BS, BFR/BS, Aj.AR, and Ac.f were significantly lower than at baseline. At 12 months, only the dynamic parameters (MS/BS, BFR/BS, Aj.AR, and Ac.f) remained decreased compared with their paired baseline values and the decrease in MS/BS was significantly lower than at 6 months. However, the relative change of MS/BS from baseline to end of treatment was six- to sevenfold lower with SrRan than with ALN. ALN induced a marked diminution of the bone remodeling from 6 months, confirming those previously observed after 2 and 3 years.[20]

The antifracture efficacy of SrRan has been largely documented,[9-13] reporting a 41% reduction in vertebral fracture risk[9] and a 16% reduction in nonvertebral fracture risk.[10] These reductions in relative risk of fracture are associated with increases in BMD of +14% and 10% at vertebral and hip level, respectively,[9, 10] suggesting that SrRan may act on the bone remodeling by enhancing a positive bone balance, after accounting for the role of the strontium atom itself in the variation in BMD. In vitro data indicate an effect of SrRan on both osteoblast and osteoclast cultures, ie, SrRan decreased osteoclastogenesis and osteoclast activity[4, 5, 21] and increased preosteoblast replication and differentiation.[3, 22] These in vitro findings fueled the hypothesis that the clinical efficacy of SrRan was linked to a “dual action” on bone tissue, by both stimulating formation and inhibiting resorption. The action of SrRan on bone remodeling has also been confirmed by in vivo animal studies, ie, intact rats and senescent mice.[6, 8] On the other hand, in ovariectomized rats, bone formation was only maintained and bone resorption decreased.[7]

In postmenopausal women, the modifications of biochemical markers of bone remodeling in response to SrRan indicated some discrepancies,[23-25] although the augmentation of bone alkaline phosphatase after 3 months of treatment suggested an early effect on bone formation.[23] In previous histomorphometric studies performed in postmenopausal osteoporotic women after 1 to 5 years of SrRan treatment, only Cn-MAR and Cn-Ob.S/BS were increased versus controls, without any change in parameters reflecting bone remodeling. These results suggest a stimulation of bone formation but were not associated with an increase in bone volume.[15] Compared with teriparatide, a bone anabolic agent, the effects of SrRan on bone remodeling parameters were not significantly different, although P1NP (a biochemical marker of bone formation) slightly but significantly decreased with SrRan whereas it increased with teriparatide.[14] The biopsy specimens analyzed in these two studies were not paired except for 5 patients in the study by Arlot and colleagues, where 3D microstructure analysis with µCT suggested an increase in bone volume over 3 years of SrRan treatment.[15] Our results do not confirm the previous findings of in vitro and animal studies.[3, 6, 8, 22] The decreases in MS/BS and BFR/BS after 12 months were associated with reductions of bone volume, trabecular thickness, and node number, and an augmentation of the trabecular separation. These microarchitecture changes resulted from a diminution of the formation parameters with a sustained resorption. In contrast, in ALN groups, despite a strong decrease of the formation, the bone mass and microarchitecture remained unchanged because the resorption also decreased. However, the relative reduction from baseline of MS/BS was 6.8-fold higher with ALN than with SrRan (−82% and −12% at M6 and M12, respectively). On the other hand, the reduction in fracture risk has been reported to be similar in patients with various levels of bone turnover before treatment.[26]

All clinical studies have demonstrated the antifracture efficacy of SrRan on vertebral, nonvertebral, and hip fracture.[9-13] These effects cannot be explained exclusively by the augmentation of BMD because strontium incorporated into hydroxyapatite in place of calcium influences the measurement of BMD. It has been estimated that, after 2 to 3 years of treatment, the strontium present in bone may overestimate the BMD by 10%.[27] Strontium is present in newly formed bone[28, 29] and preserves some determinants of the bone material quality.[29] However, the absorption of strontium onto the surface of the mineral crystal may influence the intrinsic properties of bone tissue.[28] A beneficial effect of SrRan on bone resistance measured by nanoindentation has been shown in rats,[7, 30, 31] supporting the hypothesis that strontium improves bone quality by modifying its intrinsic properties. Indeed, strontium could form sacrificial bonds[32, 33] either at the mineral-organic interface or between polar groups of the organic matrix,[34] thereby increasing bone stiffness, hardness, and toughness.[35] In osteoporotic women, the biomechanical properties of the distal tibia determined by finite element analysis and the 3D microstructure appeared improved as early as 3 months with a more pronounced effect at 1 and 2 years after SrRan treatment in contrast to ALN treatment.[36] Consistent with this hypothesis, preliminary data obtained with nanoindentation from a subset of bone biopsy specimens from the present study have shown an improvement of bone intrinsic properties with SrRan.[37]

The main strength of this study is its longitudinal design with large number of paired-biopsy specimens. However, quantitative structure analysis using histomorphometry on transiliac bone biopsy specimens also has some limitations: the 2D approach of this method and the fact that iliac bone is not a weight-bearing bone or site of mechanical pressure. Bone metabolism and remodeling could differ from site to site, and bone mecano-stimulation and pharmacological intervention could produce different results depending notably on bone loading.[38] Hildebrand and colleagues[39] and others[40] reported large intersite trabecular differences between biopsy specimens from different anatomical sites (iliac crest, femoral head, lumbar spine). Of note, it has been recently demonstrated in vitro that the mechanism of action of SrRan could also be linked to its direct effect on osteocytes, the most abundant bone cells responsible for the mechanical sensitivity of bone tissue.[41] In addition, the duration of this study could be too limited for a profound evaluation of the potential effects on bone structure and matrix properties. Indeed, in previous studies, the effect of SrRan on bone markers, microarchitecture, and resistance was shown to be much higher after 2 years of treatments.[36] The subgroup analysis showing a different effect of SrRan according to the baseline level of bone turnover is exploratory and may reflect the influence of regression toward the mean to some degree.

In conclusion, this large controlled paired biopsy study shows that the bone formation remains higher after SrRan than ALN treatment, but could not demonstrate a significant anabolic action of SrRan on bone remodeling and did not provide an explanation for the beneficial effect of SrRan on bone strength.

Disclosures

PC, JPR, NPM, and MP state that they have no conflicts of interest. Over the last 3 years, PJM has received consulting fees and honoraria for conferences from Servier, and RC has received consulting fees and/or research grants and/or honoraria for conferences from Servier, Lilly, Amgen, Merck, Chugai, Roche, Novartis, UCB, and Pfizer.

Acknowledgments

This study was funded by Les Laboratoires Servier. We thank Gaëlle Martin and Brigitte Burt-Pichat for their technical assistance.

The following investigators contributed to the biopsy specimens collection: Dr A Balogh (Debrecen, Hungary), Pr K Brixen (Odense, Denmark), Pr E Czerwinski (Krakow, Poland), Dr CA de Freitas Zerbini (Sao Paulo, Brazil), Dr M De La Pen-Rodrigues (Jalisco, Mexico), Dr NR de Melo (Sao Paulo, Brazil), Dr PA Garcia Hernandez (Monterrey, Mexico), Dr T Hala (Pardubice, Czech Republic), Dr JA Hernandez Bueno (Mexico City, Mexico), Dr P Keszthelyi (Gyula, Hungary), Dr P Lakatos (Budapest, Hungary), Dr BL Langdahl (Aarhus, Denmark), Pr V Lo Cascio (Verona, Italy), Dr AJ Lowy (Footscray, Australia), Dr LE Maffei (Buenos Aires, Argentina), Dr CH Magaril (Buenos Aires, Argentina), Dr Z Man (Buenos Aires, Argentina), Dr OD Messina (Buenos Aires, Argentina), Dr JL Morales Torres (Leon, Mexico), Dr P Novosad (Zlin, Czech Republic), Dr RM Rodrigues Pereira (Sao Paulo, Brazil), Dr LG Ste-Marie (Montreal, Canada), Pr PN Sambrook (Saint Leonards, Australia), Dr A Sawicki (Warszawa, Poland), Dr J Slesinger (Brno, Czech Republic), Dr P Somogyi (Budapest, Hungary), Dr VL Szejnfeld (Sao Paulo, Brazil), Dr I Szombati (Budapest, Hungary), Dr JA Tamayo y Orozco (Mexico City, Mexico), Dr E Toth (Kistarcsa, Hungary), Pr I Valter (Tallinn, Estonia), Dr V Vyskocil (Plzen, Czech Republic), Pr J Wark (Parkville, Australia), Dr JR Zanchetta (Buenos Aires, Argentina), Dr V Zikan (Praha, Czech Republic).

Authors' roles: Study design: PC, PJM, and RC. Study conduct: PC and RC. Data collection: PC, JPR, NPM, and MP. Data analysis: PC. Data interpretation: PC, PJM, and RC. Drafting manuscript: PC, PJM, and RC. Revising manuscript content: PC, PJM, and RC. Approving final version of manuscript: PC, PJM, JPR, NPM, MP, and RC. PC and RC take responsibility for the integrity of the data analysis.

Ancillary