SEARCH

SEARCH BY CITATION

Keywords:

  • CATHEPSIN K INHIBITOR;
  • QCT;
  • DXA;
  • SPINE;
  • PROXIMAL FEMUR

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

ONO-5334 (Ono Pharmaceutical Co., Ltd., Osaka, Japan) inhibits cathepsin K and has been shown to increase areal bone mineral density (BMD) at the hip and spine in postmenopausal osteoporosis. Quantitative computed tomography (QCT) allows the study of the cortical and trabecular bone separately and provides structural information such as cortical thickness. We investigated the impact of 2 years of cathepsin K inhibition on these different bone compartments with ONO-5334. The clinical study was a randomized, double-blind, placebo, and active controlled parallel group study conducted in 13 centers in six European countries. The original study period of 12 months was extended by another 12 months. A total of 147 subjects (age 55–75 years) of the QCT substudy who participated in the extension period were included. Subjects had been randomized into one of five treatment arms: placebo; ONO-5334 50 mg twice per day (BID); ONO-5334 100 mg once daily (QD); ONO-5334 300 mg QD; or alendronate 70 mg once weekly (QW). QCT was obtained to evaluate bone structure at the lumbar spine and proximal femur. After 24 months ONO-5334 showed statistically significant increases versus placebo for integral, trabecular, and cortical BMD at the spine and the hip (for ONO-5334 300 mg QD, BMD increases were 10.5%, 7.1%, and 13.4% for integral, cortical, and trabecular BMD at the spine, respectively, and 6.2%, 3.4%, and 14.6% for integral, cortical, and trabecular total femur BMD, respectively). Changes in cortical and trabecular BMD in the spine and hip were similar for alendronate as for ONO-5334. Integral volume did not demonstrate statistically significant changes under ONO-5334 treatment, thus there was no evidence of periosteal apposition, neither at the spine nor at the femur. Cortical thickness changes were not statistically significant for ONO-5334 in the spine and hip, with exception of a 2.1% increase after month 24 in the intertrochanter for ONO-5334 300 mg QD. Over 2 years ONO-5334 showed a statistically significant and persistent increase of trabecular and integral BMD at the spine and the hip. Cortical BMD also progressively increased but at a lower rate. Changes in bone size and of periosteal apposition were not observed. © 2014 American Society for Bone and Mineral Research.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

ONO-5334 (N-((1S)-3-{(2Z)-2- [(4R)-3,4-Dimethyl-1,3-thiazolidin-2-ylidene]hydrazino}-2,3-dioxo-1-(tetrahydro-2H-pyran-4-yl)propyl) cycloheptanecarboxamide; (Ono Pharmaceutical Co., Ltd., Osaka, Japan) is a low-molecular-weight synthetic inhibitor of cathepsin K. Cathepsin K is a member of the papain cysteine proteinase superfamily[1] and has a critical role in the bone resorption process. Cathepsin K is the most highly expressed cysteine protease in osteoclasts and is excreted into resorption lacunae beneath the ruffled border of the osteoclast where it degrades the organic matrix of bone.

Humans lacking cathepsin K exhibit pycnodysostosis, which is characterized by abnormally dense bones.[2] Specific inhibition of cathepsin K has therefore been a new drug target for diseases that have elevated bone resorption such as osteoporosis. ONO-5334 has been shown to improve both bone mineral density (BMD) and bone strength in the ovariectomized monkey osteoporosis model[3] and has also shown potent suppression of biochemical markers of bone resorption in a Phase I study of healthy postmenopausal women.[4] In the recent randomized, double-blind, parallel-group Phase II study of ONO-5334 in postmenopausal women with osteoporosis (ONO-5334 cathepsin K inhibitor European [OCEAN] study), results for areal BMD (aBMD) and biochemical markers have been reported for the initial 12 months[5] and the extended 24 months of treatment.[6] A dose-dependent increase in spine and hip aBMD was observed over 24 months of treatment.

In this study additional exploratory measurements obtained from quantitative computed tomography (QCT) measurements were used to study the effect of ONO-5334 on the different bone compartments of the spine and the hip. QCT, using modern whole-body spiral CT scanners in combination with appropriate three-dimensional (3D) image analysis tools, is a truly volumetric technique. In contrast to dual-energy X-ray absorptiometry (DXA), QCT provides information on cortical and trabecular bone separately, as well as detailed information about bone geometry that is not available from DXA. In addition, QCT BMD is less affected than DXA by degenerative disc and facet disease, scoliosis, or aortic calcification.[7-9] These are frequent findings in elderly subjects, who typically participate in clinical trials in osteoporosis. aBMD results by DXA of the proximal femur are sensitive to the rotation of the feet during the image acquisition, which is one reason for the higher precision error of DXA femur compared to DXA spine measurements along with the smaller regions analyzed.[10-12]

This is the first report on QCT results from the OCEAN study. QCT was applied to a subset of subjects to specifically evaluate differential effects of ONO-5334 on trabecular and cortical BMD, to evaluate whether there were local differences of the effect of ONO-5334 within the trabecular BMD of the spine and hip, and to investigate the effect of ONO-5334 on cortical thickness.

Subjects and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Study design

The OCEAN study was initially designed as a 12-month double-blind, placebo and active controlled, parallel group Phase II multicenter study that was extended to 24 months without changes in treatment. A total of 285 postmenopausal female subjects (age 55–75 years) with a DXA T-score at the lumbar spine or total hip below or equal to –2.5 with no fragility fractures, or a T-score between –1 and –2.5 and exactly one fragility fracture were recruited at nine different sites in Europe (Fig. 1). Of the 285 subjects, 278 were included in the full analysis population.[5] Major exclusion criteria were low aBMD defined as a DXA T-score below –3.5, low bone turnover defined as urinary cross-linked C-telopeptide I (CTX-I) below 200 µg/mmol Cr, any nonvertebral fragility fractures after the age of 50 years, clinically relevant fractures 1 year prior to the randomization visit, secondary causes of osteoporosis, metabolic bone disorders or diseases affecting bone and mineral metabolism, and previous use of osteoporosis medication. All subjects gave written informed consent and the study was approved by the appropriate ethics committees and regulatory authorities. The study is a fully registered public clinical trial (http://clinicaltrials.gov/show/NCT00532337. A Multi-centre, Randomized, Double Blind, Parallel Group Study to Investigate Efficacy and Safety of ONO-5334 in Postmenopausal Women With Osteopenia or Osteoporosis). Full details of the methodology and inclusion/exclusion criteria can be found in the report by Eastell and colleagues.[5]

image

Figure 1. Study enrollment.

Download figure to PowerPoint

Subjects were randomized to one of the following five study arms: three doses of ONO-5334 (50 mg twice per day [BID], 100 mg one daily [QD], or 300 mg QD), placebo (PBO), or alendronate (ALN, 70 mg once weekly [QW]). In addition, daily doses of calcium (at least 500 mg) and vitamin D (at least 400 IU) were administered to all subjects. After the initial 12 months an optional study extension of another 12 months was added; however, only 197 of the initial 285 subjects chose to continue through month 24 (M24). Of these 197 subjects, 147 were enrolled in the exploratory QCT substudy during the optional extension. QCT results of the full analysis set (Fig. 1) are reported herein.

DXA measurements on Prodigy scanners (GE Lunar Inc, Madison, WI, USA) were performed at the lumbar spine (L1–L4) and at the proximal femur at screening and after 3, 6, 12, 18, and 24 months using standard scan protocols as specified by the manufacturer. QCT measurements were performed in a subset of the study population at the lumbar spine (L1–L2) and at the proximal femur at randomization (M0) and after 12 (M12) and 24 (M24) months. Nine different clinical CT scanners from GE, Philips, Toshiba, Hitachi, and Siemens were used in the study. The CT scans were acquired using 120 kV, 100 mAs (spine) or 170 mAs (femur), and a slice thickness of 1 or 1.25 mm, depending on the scanner type. For reconstruction, a medium kernel and a field of view of 20 cm were used, resulting in an in-plane pixel spacing of 0.4 mm. An in-scan phantom (Mindways Software Inc, Austin, TX, USA) was used for BMD calibration. The Mindways QA phantom was scanned regularly on each CT scanner and the results were used to monitor the longitudinal scanner stability. Central analysis of the DXA and QCT images was performed at Synarc (Hamburg, Germany).

QCT analysis

The QCT analysis was performed with MIAF (Medical Image Analysis Framework, Institute of Medical Physics, University of Erlangen, Germany). The QCT analysis workflow implemented in MIAF was similar for the femur and the spine: (1) 3D segmentation of periosteal and endosteal bone surfaces; (2) determination of anatomic coordinate systems (a neck coordinate system for the femur and a vertebral coordinate system for each vertebra); and (3) determination of analysis volumes of interest (VOIs), for which volume, BMD, and bone mineral content (BMC) were measured. In addition, cortical thickness was determined. Apart from a placement of a few seed points (eg, to identify the approximate center of the vertebral body) the QCT analysis was fully automated but allowed the operator to apply corrections if necessary.[13, 14] Periosteal and endosteal bone surfaces were determined locally from profiles perpendicular to the cortex using a 50% BMD criterion. For a more detailed investigation of treatment effects at the endosteal border an additional subcortical compartment was introduced between the trabecular and cortical compartments by homogenously peeling the endosteal bone surface by 2 mm. The integral compartment is the sum of cortical, subcortical, and trabecular compartments.

In the spine L1 and L2 were scanned. In each vertebra, integral, cortical, and trabecular compartments of the total vertebral body VOI were analyzed (Fig. 2). In the spine the analysis of the subcortical compartment was limited to the mid-portion of the vertebral body, which is less affected by osteophytes and related endosteal sclerosis. Results were always averaged over L1 and L2. Sub-VOIs in the hip are shown in Fig. 3, which for comparison also includes the analysis regions of interest (ROI) of a DXA scan. The DXA intertrochanteric ROI extends further into the shaft compared to the QCT VOI. In QCT, the neck VOI covers the complete neck, whereas in DXA only a 15-mm-thick neck box is analyzed. For both QCT and DXA the total femur VOI is defined as the sum of neck, trochanteric, and intertrochanteric VOIs.

image

Figure 2. Cortical (white), subcortical (light gray), and trabecular (dark gray) compartments of the hip and the spine. The integral compartment is the sum of the three. The extended cortical compartment is the combination of the cortical and the subcortical compartments.

Download figure to PowerPoint

image

Figure 3. Analysis VOIs or the proximal femur: QCT left, DXA right.

Download figure to PowerPoint

Statistics

The primary endpoint for the original study was based on the percentage change of spinal DXA from baseline at M12.[5] As QCT was used as exploratory endpoint, observed data instead of a last observation carried forward (LOCF) approach were used. Each dose of ONO-5334 and alendronate was compared to placebo using an analysis of covariance (ANCOVA) with treatment group and country as factors in the model. Baseline values for age, body mass index (BMI), urinary CTX-I at screening and the baseline value of the variable were used as covariates. In order to compare QCT and DXA data in this report, for DXA the observed data set was also used. Here the same statistical model was employed as for QCT. In the publication from Eastell and colleagues,[6] DXA numerical results are slightly different because the primary a priori analysis was based on LOCF. BMD data are shown as least squares (LS) mean and standard error (SEM) herein. For some QCT analyses not only the percentage change in BMD and BMC, but also absolute changes were examined.

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

The full analysis set (FAS) of the QCT substudy of the 12-month extended treatment phase included 147 subjects with DXA scans, 123 with QCT hip scans, and 120 with QCT spine scans (Fig. 1). Demographic and baseline characteristics are shown in Table 1.

Table 1. BMD Baseline Characteristics
 Placebo (n = 24)ONO-5334 50 mg BID (n = 27)ONO-5334 100 mg QD (n = 34)ONO-5334 300 mg QD (n = 32)Alendronate 70 mg QW (n = 30)
  1. Values are least squares mean ± SEM.

  2. BMD = bone mineral density; BID = two times per day; QD = once daily; QW = once weekly; DXA = dual-energy X-ray absorptiometry; QCT = quantitative computed tomography; LS = lumbar spine (DXA L1–L4, QCT L1–L2); Int = integral compartment; Trab = trabecular compartment; Cort = cortical compartment.

  3. a

    DXA values are given for the total femur ROI.

  4. b

    QCT values are given for the total vertebral body and total femur VOIs.

Age (years)65.0 ± 4.365.3 ± 5.165.7 ± 5.064.6 ± 5.265.7 ± 4.4
Weight (kg)64.7 ± 9.365.8 ± 8.765.5 ± 11.666.6 ± 9.964.5 ± 9.4
DXAa
L1–L4 (LS) BMD (g/cm2)0.84 ± 0.0130.85 ± 0.0120.87 ± 0.0110.87 ± 0.0120.85 ± 0.012
Hip BMD (g/cm2)0.78 ± 0.0170.79 ± 0.0160.77 ± 0.0140.77 ± 0.0150.77 ± 0.015
QCTb
Spine L1–L2 (LS), n2022272724
Int BMD (mg/cm3)157 ± 4.0150 ± 3.8152 ± 3.4162 ± 3.5154 ± 3.8
Trab BMD (mg/cm3)84 ± 3.584 ± 3.382 ± 3.091 ± 3.185 ± 3.3
Cort BMD (mg/cm3)320 ± 7.5308 ± 6.9316 ± 6.4331 ± 6.5318 ± 7.0
Hip, n2022282825
Int BMD (mg/cm3)235 ± 7.1243 ± 6.6233 ± 5.9237 ± 6.1245 ± 6.5
Trab BMD (mg/cm3)78 ± 5.586 ± 5.180 ± 4.677 ± 4.785 ± 5.0
Cort BMD (mg/cm3)524 ± 11.1535 ± 10.3516 ± 9.2542 ± 9.5541 ± 10.1

DXA results

Figure 4 and Table 2 show the DXA results for lumbar spine (Fig. 4A) and proximal femur (Fig. 4C) for the 147 subjects included in the QCT substudy of the extension study. For comparison, the DXA results of the primary LOCF analysis[6] are shown in Fig. 4B and D. DXA results between the two groups did not differ significantly.

image

Figure 4. DXA aBMD least squares mean changes ± SEM versus baseline. (A, B) Lumbar spine L1–L4. (C, D) Total femur. (A, C) Total of 147 subjects Full Analysis Set (FAS) observed data of the QCT substudy; (B, D) FAS of the extension study, LOCF. Statistical analyses were performed with t tests from the ANCOVA model, difference from placebo treatment arm, *p < 0.05, ***p < 0.001 (shown for 24 months only).

Download figure to PowerPoint

Table 2. DXA and QCT BMD Percentage Changes Versus Placebo at Time Periods Specified
 Time periodONO-5334 50 mg BID (n = 27)ONO-5334 100 mg QD (n = 34)ONO-5334 300 mg QD (n = 32)Alendronate 70 mg QW (n = 30)
  • Values are least squares means and 95% confidence intervals. Results are shown for the total hip (Tot hip) and the total vertebral body VOIs.

  • DXA = dual-energy X-ray absorptiometry; QCT = quantitative computed tomography; BMD = bone mineral density; BID = two times per day; QD = once daily; QW = once weekly; M12 = 12 months; M24 = 24 months; int = integral compartment, trab = trabecular compartment; cort = cortical compartment; VOI = volume of interest.

  • Difference from placebo treatment arm, t test:

  • #

    p < 0.1, p > 0.05;

  • *

    p < 0.05;

  • **

    p < 0.01;

  • ***

    p < 0.001.

DXA
SpineM124.0 (2.1 to 5.9)***3.2 (1.4 to 5.0) ***5.3 (3.4 to 7.1)***5.7 (3.9 to 7.6)***
 M244.3 (1.9 to 6.6)***4.1 (1.8 to 6.4)***7.4 (5.1 to 9.6)***7.6 (5.3 to 9.9)***
HipM122.1 (0.7 to 3.5)**0.9 (−0.5 to 2.2)2.4 (1.1 to 3.8)***2.6 (1.3 to 4.0)***
 M243.4 (1.8 to 5.0)***1.7 (0.2 to 3.3)#3.8 (2.2 to 5.3)***4.2 (2.7 to 5.8)***
QCT
Spine intM127.3 (5.1 to 9.4)***5.3 (3.3 to 7.4)***8.1 (6.0 to 10.1)***7.4 (5.2 to 9.5)***
 M248.3 (4.8 to 11.8)***8.1 (4.8 to 11.4)***10.5 (7.3 to 13.8)***9.2 (5.9 to 12.6)***
Spine trabM129.4 (5.9 to 13.0)***7.6 (4.2 to 10.9)***8.7 (5.3 to 12.1)***8.8 (5.3 to 12.2)***
 M2412.8 (7.4 to 18.2)***12.7 (7.5 to 17.8)***13.4 (8.2 to 18.5)***13.0 (7.7 to 18.2)***
Spine cortM124.7 (2.1 to 7.4)***3.1 (0.6 to 5.6)*6.7 (4.2 to 9.2)***5.3 (2.8 to 7.9)***
 M245.7 (2.6 to 8.7)***4.5 (1.7 to 7.4)**7.1 (4.3 to 9.9)***7.1 (4.1 to 10.0)***
Tot hip intM123.9 (2.5 to 5.4)***2.2 (0.8 to 3.6)**4.3 (2.9 to 5.7)***4.3 (2.8 to 5.7)***
 M245.5 (3.3 to 7.7)***3.6 (1.5 to 5.6)***6.2 (4.2 to 8.3)***5.5 (3.4 to 7.6)***
Tot hip trabM127.5 (2.9 to 12.1)**4.2 (−0.1 to 8.5)#8.6 (4.3 to 12.8)***6.6 (2.2 to 11.0)**
 M2412.9 (7.1 to 18.7)***9.8 (4.3 to 15.3)***14.6 (9.2 to 20.1)***11.5 (5.8 to 17.1)***
Tot hip cortM122.9 (1.5 to 4.3)***1.5 (0.1 to 2.8)*3.1 (1.8 to 4.4)***3.4 (2.0 to 4.7)***
 M242.9 (1.0 to 4.7)**1.3 (−0.5 to 3.1)3.4 (1.7 to 5.2)***2.5 (0.7 to 4.3)**

QCT results

Standard BMD measurements

Figure 5 shows M12 and M24 QCT BMD percentage changes relative to baseline and Table 2 shows BMD percentage changes versus placebo for integral, trabecular, and cortical BMD of the total vertebral body VOI. For all three ONO-5334 doses, and for alendronate BMD percentage increases versus baseline and against placebo, the values were statistically significant.

image

Figure 5. QCT: BMD least squares mean changes ± SEM versus baseline for integral, trabecular, and cortical BMD of the total vertebral body VOI (L1 and L2), observed data, t test, difference from placebo, *p < 0.05, **p < 0.01, ***p < 0.001.

Download figure to PowerPoint

The corresponding data for the total femur VOI are presented in Fig. 6 and Table 2. Integral and cortical percentage BMD increases and differences versus placebo were smaller compared to the spine but percentage increases in trabecular BMD were similar. More detailed results for integral and trabecular BMD in the sub-VOIs in the femur are shown in Table 3, where for briefness the M12 results are omitted. Within the femur, trabecular BMD percentage changes versus placebo were higher in the trochanter and intertrochanter compared to those in the neck. Integral BMD percentage changes were also higher in the trochanter than in the neck.

image

Figure 6. QCT: BMD least squares mean changes ± SEM versus baseline for integral, trabecular, and cortical BMD of the total hip VOI, observed data, t test, difference from placebo, #p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001.

Download figure to PowerPoint

Table 3. QCT Integral and Trabecular BMD Percentage Changes Versus Placebo at M24: Results for Sub-VOIs in the Femur
 ONO-5334 50 mg BIDONO-5334 100 mg QDONO-5334 300 mg QDAlendronate 70 mg QW
  • Values are least squares means and 95% confidence intervals.

  • QCT = quantitative computed tomography; BMD = bone mineral density; M24 = 24 months; VOI = volume of interest; BID = two times per day; QD = once daily; QW = once weekly; int = integral compartment; Troch = trochanter; IT = intertrochanter; trab = trabecular compartment.

  • Difference from placebo treatment arm, t test:

  • #

    p > 0.05, p < 0.1;

  • *

    p < 0.05;

  • **

    p < 0.01;

  • ***

    p < 0.001.

Total hip int5.5 (3.3 to 7.7)***3.6 (1.5 to 5.6)***6.2 (4.2 to 8.3)***5.5 (3.4 to 7.6)***
Neck int4.8 (2.5 to 7.1)***3.6 (1.4 to 5.8)**5.5 (3.3 to 7.7)***4.7 (2.5 to 7.0)***
Troch int6.9 (4.1 to 9.8)***4.3 (1.6 to 7.0)**8.0 (5.3 to 10.6)***6.5 (3.7 to 9.3)***
IT int4.8 (1.9 to 7.7)**2.6 (−0.2 to 5.4)#4.2 (1.5 to 7.0)**4.1 (1.3 to 6.9)**
Total hip trab12.9 (7.1 to 18.7)***9.8 (4.3 to 15.3)***14.6 (9.2 to 20.1)***11.5 (5.8 to 17.1)***
Neck trab9.8 (3.9 to 15.7)**5.9 (0.3 to 11.6)*10.4 (4.8 to 16.0)***8.4 (2.7 to 14.1)**
Troch trab15.6 (7.0 to 24.2)***11.9 (3.6 to 20.2)**16.4 (8.2 to 24.5)***12.3 (3.9 to 20.7)**
IT trab13.9 (7.4 to 20.4)***11.8 (5.6 to 18.0)***16.9 (10.7 to 23.0)***13.4 (7.1 to 19.8)***
Detailed cortical BMD measurements

In the spine, cortical BMD percentage changes versus placebo were smaller than trabecular BMD changes (Table 2). A more detailed analysis involving the subcortical compartment of the mid-section of the vertebral bodies (Table 4) showed that at M24 BMD percentage changes versus placebo in the subcortical compartment were slightly higher than in the cortical compartment, whereas BMD changes in the extended cortical compartment were more comparable to the cortical compartment.

Table 4. Cortical BMD Percentage Changes Versus Placebo at M24: Results for the Mid-Vertebral Body VOI
 ONO-5334 50 mg BIDONO-5334 100 mg QDONO-5334 300 mg QDAlendronate 70 mg QW
  • Values are least squares means and 95% confidence intervals.

  • BMD = bone mineral density; M24 = 24 months; VOI = volume of interest; BID = two times per day; QD = once daily; QW = once weekly; mid = mid-vertebral; cort = cortical compartment; subcort = subcortical compartment; ext cort = extended cortical compartment.

  • Difference from placebo treatment arm, t test:

  • #

    p > 0.05, p < 0.1;

  • *

    p < 0.05;

  • **

    p < 0.01;

  • ***

    p < 0.001.

Spine mid cort3.5 (0.9 to 6.1)**3.5 (1.0 to 5.9)**5.4 (2.9 to 7.9)***6.4 (3.8 to 8.9)***
Spine mid subcort5.6 (1.6 to 9.5)**3.7 (0.0 to 7.4)#4.8 (1.1 to 8.5)*7.4 (3.6 to 11.2)***
Spine mid ext cort4.1 (1.4 to 6.8)**3.7 (1.1 to 6.3)**5.6 (3.0 to 8.2)***6.9 (4.2 to 9.5)***

In the femur, cortical BMD percentage changes versus placebo were also smaller than trabecular BMD changes (Table 2). In the intertrochanter, cortical BMD percentage changes versus placebo were statistically either not significant (ONO-5334 50 mg BID and 100 mg QD) or borderline significant (ONO-5334 300 mg QD and alendronate). Percentage increases in the extended cortical compartment of the intertrochanter did not attain statistical significance (Table 5). In contrast, the trochanter cortical BMD percentage increases differed significantly from placebo and were statistically higher than those in the total femur VOI.

Table 5. QCT Cortical BMD Percentage Change Versus Placebo at M24: Sub-VOIs in the Femur
 ONO-5334 50 mg BIDONO-5334 100 mg QDONO-5334 300 mg QDAlendronate 70 mg QW
  • Values are least squares means and 95% confidence intervals.

  • QCT = quantitative computed tomography; BMD = bone mineral density; M24 = 24 months; VOI = volume of interest; BID = two times daily; QD = once daily; QW = once weekly; cort = cortical compartment; ext cort = extended cortical compartment; Troch = trochanter; IT = intertrochanter.

  • Difference from placebo treatment arm, t test:

  • #

    p > 0.05, p < 0.1;

  • *

    p < 0.05;

  • **

    p < 0.01;

  • ***

    p < 0.001.

Total hip cort2.9 [1.0 to 4.7]**1.3 [−0.5 to 3.1]3.4 [1.7 to 5.2]***2.5 [0.7 to 4.3]**
Total hip ext cort3.5 [1.8 to 5.3]***2.1 [0.4 to 3.8]*4.1 [2.5 to 5.8]***3.3 [1.5 to 5.0]***
Troch cort4.5 [2.2 to 6.9]***2.3 [0.0 to 4.5]*5.6 [3.4 to 7.8]***4.5 [2.3 to 6.8]***
Troch ext cort5.6 [3.2 to 8.0]***3.3 [1.0 to 5.6]**6.5 [4.2 to 8.7]***5.3 [3.0 to 7.6]***
IT cort1.5 [−0.8 to 3.7]−0.7 [−2.9 to 1.5]0.7 [−1.5 to 2.8]0.9 [−1.3 to 3.1]
IT ext cort2.2 [−0.2 to 4.5]#0.0 [−2.3 to 2.2]1.5 [−0.7 to 3.7]1.4 [−0.9 to 3.7]

In the total femur VOI, percentage changes from baseline were higher in the extended cortical than in the cortical compartment alone and percentage changes were highest in the subcortical compartment. The subcortical compartment was only evaluated for the total femur VOI (Table 5).

Volume and cortical thickness measurements

Integral volume showed no statistically significant changes under treatment, thus no periosteal apposition, neither at the spine nor at the femur was observed. Cortical thickness changes (Table 6) were not statistically significant for ONO-5334 in the spine and hip with exception of a 2.1% increase after M24 in the intertrochanter for ONO-5334 300 mg QD that was paralleled by a 3.4% increase for alendronate.

Table 6. QCT Cortical Thickness Percentage Change Versus Placebo at M24: Sub-VOIs in the Spine and Femur
 ONO-5334 50 mg BIDONO-5334 100 mg QDONO-5334 300 mg QDAlendronate 70 mg QW
  • Values are least squares means and 95% confidence intervals.

  • QCT = quantitative computed tomography; M24 = month 24; VOI = volume of interest; BID = two times daily; QD = once daily; QW = once weekly.

  • a

    Mid-portion of the vertebral body.

  • Difference from placebo treatment arm, t test:

  • #

    p > 0.05, p < 0.1;

  • *

    p < 0.05;

  • **

    p < 0.01.

Spinea1.5 (−14 to 17)5.4 (−9.5 to 20)5.1 (−3.6 to 26)5.5 (−12 to 19)
Neck−0.1 (−3.7 to 3.6)−0.1 (−3.6 to 3.4)1.5 (−1.9 to 5.0)4.1 (0.5 to 7.6)*
Trochanter−0.7 (−3.2 to 1.7)−1.5 (−3.8 to 0.9)−0.2 (−2.5 to 2.2)1.5 (−0.9 to 3.9)
Intertrochanter2.1 (0.0 to 4.2)#0.0 (−2.1 to 2.0)2.1 (0.1 to 4.1)*3.4 (1.3 to 5.4)**

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

In this study, QCT was used to better understand treatment effects of the ONO-5334 on BMD in the spine and the proximal femur. Three groups of subjects receiving different doses of ONO-5334 and alendronate were compared with a group receiving placebo. Alendronate served as an active reference. Subjects with low bone turnover, defined as urinary CTX-I below 200 µg/mmol Cr, were not included in the study. We speculated that drugs that work primarily by the inhibition of bone turnover would not be very effective if the baseline bone turnover was low, so we felt the optimal design for our study was to exclude such patients. One of the likely consequences would be to observe a greater BMD response than studies that do not select in this way. This fact needs to be taken into account in the interpretation of the results.

M12 and M24 results for the primary efficacy endpoint, spine aBMD as measured by DXA, have been presented and discussed in separate publications.[5, 6] Here, DXA results presented by Eastell and colleagues[6] were only shown for comparison in order to demonstrate that the DXA data of the QCT substudy were very similar to those of the full analysis set. In the vertebrae of the spine and in the total femur, integral, trabecular, and cortical BMD changes as measured by QCT from baseline (Figs. 5 and 6) and versus placebo (Table 2) were statistically significant for all doses of ONO-5334 with a few exceptions for cortical BMD.

It is interesting that in the total hip, integral BMD percentage changes versus placebo were numerically higher than those measured by DXA of the total hip, although in contrast to the spine no degenerative hypertrophy or overlying affect or bones interfere with the projectional measurement. Part of this effect may be explained by slight differences in the definition of the total femur VOIs between QCT and DXA (Fig. 3). Because the integral BMD percentage increases are slightly larger in the neck than in the intertrochanter (Table 3) the larger neck and smaller intertrochanteric VOIs used in QCT, relative to DXA, could result in higher BMD increases compared to DXA. Nevertheless, slightly higher increases in QCT of the total femur were previously observed for oral ibandronate, (12-month BMD increase versus placebo: QCT 2.2%; DXA 2.0%),[15] zoledronic acid (36-month BMD increase versus placebo: QCT 6.0%; DXA 5.0.%),[16] and teriparatide (12-month BMD increase versus placebo: QCT 4.1%; DXA 2.7%) and alendronate (12-month BMD increase versus placebo: QCT 2.7%; DXA 2.2%).[17]

The most prominent effect of ONO-5334 was the large percentage increase of trabecular BMD. Trabecular BMD changes in the hip paralleled those in the spine, but at M24 the differentiation among treatment groups was larger in the hip. In the hip, BMD percentage increases versus placebo were higher in the trochanter and intertrochanter VOIs compared to the neck VOI (Table 3). However, when measuring cortical, trabecular, and integral compartments separately, percentage changes have to be interpreted in the context of absolute values because differences in percentage changes often can be largely explained by a numerator-denominator effect. For example, for 300 mg QD ONO-5334, trabecular BMD increased from baseline values of 98 mg/cm3, 61 mg/cm3, and 82 mg/cm3 in the neck, trochanter, and intertrochanter, respectively, by 12.7 mg/cm3, 10.5 mg/cm3, and 13.9 mg/cm3, respectively. Thus on an absolute scale the M24 numeric increase in trochanter trabecular BMD was actually slightly smaller than trabecular BMD increases in the neck and intertrochanter. Although the trochanter benefitted most in terms of relative change, absolute increases of trabecular BMD were rather homogenous in the femur.

In general, as a result of the numerator-denominator effect, relative trabecular BMD changes are expected to be considerable higher than those in integral or cortical BMD. Accordingly, for ONO-5334, percentage increases of cortical BMD were much smaller than of trabecular BMD. For example, increases versus placebo were 14.6% versus 3.4% at the total femur VOI for ONO-5334 300 mg QD after 24 months of treatment. However, in absolute terms cortical BMD increased from 542 mg/cm3 at baseline by 18.5 mg/cm3 and trabecular BMD from 77 mg/cm3 by 11.4 mg/cm3. In other words, for a given volume more BMC was added to the cortical than to the trabecular compartment.

Regarding the implication of the discrepancy of absolute versus relative BMD changes, for example, it is not yet clear whether it is more important to achieve high relative or high absolute BMD changes, in particular with respect to bone strength and future fracture prediction. However, the absolute BMD increase clearly demonstrates that ONO-5334 has an impact on cortical BMD despite small relative increases of cortical BMD.

In the femur, integral and cortical volume did not change, and, with the exception of the intertrochanter, cortical thickness did not show statistically significant changes either. Thus the analyzed cortical compartments were rather identical at baseline and follow-up visits and the changes in cortical BMD were not caused by larger differences in the location of the cortical compartment relative to the bone. This is a result and not a constraint of the analysis because baseline and follow-up datasets were analyzed independently.

In the spine, cortical measurements were only obtained in the mid-portion of the vertebral body because the inferior and superior parts are often affected by degenerative changes. In contrast to the femur, in the spine the increase in subcortical BMD was more comparable to the trabecular increase because in the spine the cortex is very thin, perhaps about 0.5 mm, and therefore more heavily affected by partial volume artifacts caused by the limited spatial resolution of the CT scanner. As a consequence, the cortical compartment in the spine already contains a higher proportion of subcortical bone than the cortical compartment in the hip.

One interesting aspect of this study was the use of multiple ONO-5334 dose regimens and the use of alendronate. Although the study was not powered to detect differences among the active treatment arms and this was not part of the a priori statistical analysis plan, ONO-5334 300 mg QD showed a consistent pattern of larger numerical increases from baseline and versus placebo than ONO-5334 100 mg QD with a similar but less consistent result compared to ONO-5334 50 mg BID. The comparison between ONO-5334 300 mg QD and alendronate after 24 months indicated very similar treatment effectiveness, perhaps with slightly higher numerical increases for ONO-5334 300 mg QD in the trabecular compartment of the hip. It is important to bear in mind that in all cases, confidence intervals overlapped and statistically significant differences among treatment arms were not observed; thus, any interpretation of dose differences remains speculative. Also, recent data from another cathepsin K formulation suggest that our study was probably too short to pick up potential differences between ONO-5334 and alendronate. Under cathepsin K treatment BMD persistently increased over 5 years,[18] contrasting a typical leveling off effect for BMD increases after 2 years of bisphosphonate treatment.[19]

One limitation of the current study was the small number of patients for whom QCT was performed. After excluding those patients in whom both L1 and L2 could not be analyzed, less than 30 subjects remained per group in which a successful QCT analysis was conducted. The QCT analysis was performed on observed data and was limited to the subset of those patients included in the extension study (ie, a completers analysis), although the DXA results did not show differences between the full QCT cohort and the smaller subset included in the extension study.

In summary, over 2 years, ONO-5334 showed a statistically significant and persistent increase of trabecular and integral BMD at the spine and the hip. Cortical BMD also progressively increased but at a lower rate. The percentage increases seen for ONO-5334 100 mg QD were consistently smaller than for ONO-5334 50 mg BID and ONO-5334 300 mg QD, confirming earlier DXA results.[5] The inspection of absolute BMD showed that ONO-5334 added as much or even more new bone to the cortical as to the trabecular compartment, although percentage increases of trabecular were higher than cortical BMD. The impact of this observation on bone strength needs further attention. Changes in bone size and therefore periosteal apposition were not observed under ONO-5334 treatment. The analysis presented here shows the versatility of QCT to investigate drug changes in BMD in various locations in the spine and the hip and the usefulness to separately assess cortical and trabecular compartments.

Disclosures

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

SN, TK and SD are employees of ONO PHARMA UK LTD or Ono Pharmaceutical Co., Ltd.; SN and TK are stockholders of Ono Pharmaceutical Co., Ltd.; KE and TF are employees of Synarc INC and have served on advisory boards of ONO PHARMA UK LTD. RE and HKG have served on advisory boards of ONO PHARMA UK LTD.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

This work was supported by ONO Pharmaceutical Co., Ltd., Japan.Authors' roles: Study design: SH, TK, SD, RE. Study management: KE, SN, TF, SD. Data collection: KE, SN, TF. Statistical Analysis: SN, MS, TK. Data Interpretation: KE, SN, TF, SD, RE, HKG. Manuscript Draft: KE. Manuscript Revision: SN, TF, SD, RE, HKG Manuscript Approval: all.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References