Relationship of Drug Associated Change in Bone Mineral Density to Fracture Risk

1
Relationship of Drug Associated Change in Bone
Mineral Density to Fracture Risk

• Marc C. Hochberg, MD, MPH
• FDA Endocrinologic and Metabolic Drugs
  Advisory Committee
• 25 September 2002

2
Osteoporosis Definition(NIH Consensus
Conference, 2000)

• A skeletal disorder characterized by
  compromised bone strength predisposing to an
  increased risk of fracture
• Bone strength reflects bone mass and bone quality
• Bone mass is estimated by bone mineral density
• Bone quality refers to microarchitecture,
  turnover, damage accumulation (eg,
  microfractures), and degree of mineralization

NIH Consensus Development Panel on Osteoporosis
Prevention, Diagnosis, and Therapy. JAMA.
2001285785795.
3
Bone Mineral Density and Bone Turnover

• Components of the new definition of
  osteoporosis
• Are changes in bone mineral density and bone
  turnover with antiresorptive therapy important in
  explaining the antifracture efficacy of these
  agents?
• Vertebral fractures
• Nonvertebral fractures

4
Pharmacologic Treatment of Osteoporosis

• Antiresorptive agents
• Bisphosphonates
• Alendronate, Risedronate
• Calcitonin
• Estrogen (not FDA approved for treatment)
• Selective estrogen receptor modulators
• Anabolic agents
• Fluoride (not FDA approved)
• Parathyroid hormone (not FDA approved)

5
Antiresorptive Agents

• Inhibit bone resorption
• Fewer resorption sites improves bone
  microarchitecture (fewer cavities)
• Shallower resorption sites improves bone balance
  (amount of formation greater than resorption)
• Slower turnover rate allows better
  mineralization of bone
• Result
• Increased bone mass
• Improved bone strength and bone quality
• Out of proportion to increase in bone mass

6
Vertebral Fractures
7
Changes in Bone Density and Antifracture Efficacy
of Antiresorptive Agents

• Methods
• Identified 13 randomized placebo-controlled
  trials of antiresorptive agents that reported
  both vertebral fracture incidence and changes in
  BMD
• Used Poisson regression
• Related ?BMD and fracture risk reduction
• Pooled data to obtain best fit
• Sensitivity analysis performed

Wasnich RD, Miller PD. J Clin Endocrinol Metab.
200085231236.
8
Greater Increase in Bone Mineral Density (BMD)
and Larger Reduction in Fracture Risk
1.6
T
A alendronate
C calcitonin
1.4
E etidronate
H hormone replacement
1.2
R
R raloxifene
T tiludronate
1.0
T
22
C
C
54 risk reduction
E
Relative Risk Reduction ()
Relative Risk of Vertebral Fracture
0.8
T
A
T
A
C
R
H
0.6
A
R
A
A
A
E
0.4
C
0.2
0.0
Percent Change in Spine BMD (vs Placebo)
Wasnich RD, Miller PD. J Clin Endocrinol Metab.
200085231236.
9
Predicting the Effect of Antiresorptive
Treatments on Vertebral Fractures

• Results (Method 1)
• Identified 12 randomized, placebo-controlled
  trials of antiresorptive agents that lasted ?2
  years and reported both vertebral fracture
  incidence (?5 fractures per treatment group) and
  changes in spine BMD
• Used weighted regression models to estimate the
  association of change in spine BMD and reduction
  in vertebral fracture risk
• Expected RR 0.75 - 0.03 X increase in LSBMD

BMD Bone mineral density. Cummings SR et al. Am
J Med 2002112281-9.
10
Reduction in Fracture Risk Predicted From ?
Lumbar Spine Bone Mineral Density (BMD)
Wasnich RD, Miller PD. J Clin Endocrinol Metab.
200085231236. Cummings SR et al. Am J Med.
2002112281-9.
11
Predicting the Effect of Antiresorptive
Treatments on Vertebral Fractures

• Results (Method 2)
• Relative risk reduction in vertebral fractures
  exceeded that estimated from the regression model
  using change in lumbar spine BMD.
• Observed RR 0.83 X Expected RR - 0.11
• Concluded that observed changes in lumbar spine
  BMD explain only a small proportion of the actual
  reduction in risk of vertebral fractures.

BMD Bone mineral density. Cummings SR et al. Am
J Med 2002112281-9.
12
Extended New Analysis

• Repeated Wasnich Miller analysis excluding
  trials of non-approved medications and adding
  risedronate VERT studies total of 13 trials.
• Results largely unchanged
• Change in LSBMD remains significantly associated
  with reduction in risk of VFx
• RR 0.90 (95 CI 0.83, 0.97) per 1 D in BMD
• Independent effect of treatment even without any
  increase in LSBMD
• RR 0.81 (95 CI 0.66, 1.00)

13
Summary Change in BMD and Reduction in Risk of
Vertebral Fractures

• Greater increase in BMD is associated with
  greater reduction in fracture risk
• Models predict that an 8 increase in BMD at the
  spine (or 5 increase in BMD at the hip) would
  decrease risk by 50
• Significant treatment effect independent of ?BMD
• Models predict that an antiresorptive agent that
  does not increase BMD would decrease risk by 25
• Probably due to ? bone turnover (resorption)

BMDBone mineral density.
14
Different Effects of Antiresorptive Therapies on
Vertebral and Nonvertebral Fractures

• Some agents (eg, calcitonin, raloxifene) decrease
  vertebral fracture risk, but have not been shown
  to reduce risk of nonvertebral fractures
• Smaller changes in BMD and bone turnover were
  also observed for these agents compared to
  aminobisphosphonates

BMD Bone mineral density.
15
Nonvertebral Fractures
16
Changes in Bone Density and Turnover Reductions
in Incidence of Nonvertebral Fractures

• Methods
• Pooled all randomized, double-blind,
  placebo-controlled clinical trials with data on
  changes in BMD and/or BCM of bone turnover and
  incidence of nonvertebral fractures
• Larger studies given greater weight
• Related ?BMD and ?BCM over 1 year to reduction in
  fracture risk over duration of trial

BMD Bone mineral density. BCM Biochemical
markers. Hochberg MC et al. J Clin Endocrinol
Metab. 2002871586-92.
17
Changes in Bone Density and TurnoverReductions
in Incidence of Nonvertebral Fractures

• Results
• 18 trials identified
• 30 active antiresorptive treatment groups
• 69,369 woman-years of follow-up
• 92 in the 8 largest studies
• 2,415 women with new nonvertebral fractures
• 90 in the 8 largest studies

Hochberg MC et al. J Clin Endocrinol Metab.
2002871586-92.
18
Relative Risk (RR) of Nonvertebral Fracture
BMD Bone mineral density. Hochberg MC et al. J
Clin Endocrinol Metab. 2002871586-92.
19
Relative Risk (RR) of Nonvertebral Fracture
BMD Bone mineral density. Hochberg MC et al. J
Clin Endocrinol Metab. 2002871586-92.
20
Relative Risk (RR) of Nonvertebral Fracture
Hochberg MC et al. J Clin Endocrinol Metab.
2002871586-92.
21
Relative Risk (RR) of Nonvertebral Fracture
Hochberg MC et al. J Clin Endocrinol Metab.
2002871586-92.
22
Correlation of ?BMD or ?BCM With Reduction in
Nonvertebral Fracture Risk

• Increases in BMD and decreases in BCM at 1 year
  were both significantly associated with reduction
  in risk of new nonvertebral fractures
• Spine BMD 0.08 per 1 ? (P 0.02)
• Hip BMD 0.27 per 1 ? (P 0.006)
• Resorption BCM 0.07 per 10 ? (P lt 0.05)
• Formation BCM 0.14 per 10 ? (P 0.01)

BMD Bone mineral density. BCM Biochemical
markers. Hochberg MC et al. J Clin Endocrinol
Metab. 2002871586-92.
23
Sensitivity Analyses

• Results generally robust to removal of both
  individual trials as well as all trials of
  individual agents

24
Extended New Analysis

• Repeated analysis excluding trials of
  non-approved medications total of 15 trials.
• Results largely unchanged
• Change in BMD at 1 year remains significantly
  associated with reduction in risk of non-VFx
• RR 0.92 (95 CI 0.86, 0.99) per 1 D in LSBMD
• RR 0.78 (95 CI 0.64, 0.95) per 1 D in HBMD
• No significant independent effect of treatment
  without any increase in LSBMD

25
Summary Nonvertebral Fractures

• Greater increase in BMD or decline in BCM at 1
  year is associated with greater reduction in
  fracture risk
• Models predict that a 6 ? in BMD at spine or 3
  ? BMD at the hip, or 70 ? in BCM of resorption
  or 50 ? in BCM of bone formation would decrease
  risk of nonvertebral fractures by 40 to 45
• No significant treatment effect independent of
  ?BMD or ?BCM

BMD Bone mineral density. BCM Biochemical
markers. Hochberg MC et al. J Clin Endocrinol
Metab. 2002871586-92.
26
Conclusions

• ? BMD and ? BCM are important indicators of
  antifracture efficacy of antiresorptive drugs
• Both for vertebral and nonvertebral fracture
• ? BMD (and large ? BCM) appear to be necessary to
  ? risk of nonvertebral fractures
• Is there a threshold effect for vertebral
  fractures?
• Consistent with hypothesis of Riggs Melton
  (2002)
• Further analyses of existing data needed
• Results cannot be extrapolated to anabolic agents
  such as fluoride or PTH

BMD Bone mineral density, BCM Biochemical
markers, PTH Parathyroid hormone. Hochberg MC
et al. J Clin Endocrinol Metab. 2002871586-92.
27
Extended New Analysis

• Repeated both analyses including data from
  pivotal phase III trial of teripartide (Neer et
  al, 2001)
• Results largely unchanged for both VFx and
  non-VFx
• Change in BMD remains significantly associated
  with reduction in risk of VFx
• Change in BMD at 1 year remains significantly
  associated with reduction in risk of non-VFx
• Significant independent effect of treatment
  without any increase in BMD for VFx but not for
  non-VFx

28
Revised Extended Conclusions

• ? BMD is an important indicator of antifracture
  efficacy for both antiresorptive and anabolic
  drugs (eg., PTH)
• Both for vertebral and nonvertebral fracture
• Larger ?s in BMD appear to be necessary to ? risk
  of nonvertebral fractures
• Reductions in BCM of bone turnover are
  independent predictors of reductions in risk of
  vertebral but not for nonvertebral fractures

BMD Bone mineral density, BCM Biochemical
markers, PTH Parathyroid hormone. Hochberg MC
et al. J Clin Endocrinol Metab. 2002871586-92.
29
Back-ups
30
Osteoporosis Definition(Consensus Conference,
1993)

• A systemic skeletal disorder characterized by
  low bone mass and microarchitectural
  deterioration of bony tissue, with a consequent
  increase in bone fragility and susceptibility to
  fractures.

31
Osteoporosis Definition(World Health
Organization, 1994)

• Bone mineral density that is more than 2.5
  standard deviations below the young average value
  (T-score gt -2.5)

32
Low Bone Mineral DensityRisk Factor for Fracture

• 60 to 80 of bone strength is related to bone
  mineral density1
• Low bone mineral density predicts increased
  fracture risk2,3
• Increases in bone mineral density are associated
  with a reduction in fracture risk4

1. Curry JD. Engineering in Medicine.
198615153154. 2. Cummings SR et al. Lancet.
19933417275. 3. Ettinger B et al. JAMA.
1999282637645. 4. Hochberg MC et al. J Clin
Endocrinol Metab. 2002871586-92.
33
Fracture Rates and Relative Risk by BMD Category
3.5
3
Normal
Adjusted relative risk Normal 1.0 Osteopenia 1.7
(1.6, 1.9) Osteoporosis 2.7 (2.4, 3.1)
Osteopenia
2.5
Osteoporosis
2
Rate per 100 Person-Years
1.5
1
0.5
0
BMD Category
Data From National Osteoporosis Risk Assessment
(NORA) Siris ES et al JAMA. 200128628152822.
34
High Bone Turnover Risk Factor for Fracture

• Higher levels of bone turnover are associated
  with increased rates of bone loss in untreated
  postmenopausal women
• Higher levels of bone turnover are associated
  with increased risk of hip fracture in
  postmenopausal women

35
Bone Mineral Density and Bone Turnover Risk
Factors for Hip Fracture
BMD Bone mineral density. CTX
CTelopeptide. D-Pyr Deoxypyridinoline Garnero
P et al. J Bone Miner Res. 19961115311538.
36
Bone Mineral Density, Quality and Turnover
Independent Risk Factors for Fracture
2.9
3
2.6
2.5
2
Risk of Hip Fracture
(Odds Ratio)
1
0
BMD
Heel BUA
Urinary CTX
BMD Bone mineral density. BUA Broadband
Ultrasonographic Attenuation. CTX
CTelopeptide. Garnero P et al. Osteoporos Int.
19988563569.

37
Antiresorptive Agents Should Prevent Fractures
Through Effects on Bone Strength
Reduce biochemical markers ofbone turnover to
premenopausal levels Increase both cortical
and trabecular bone mass (bone mineral density)
and improve bone quality Decrease fracture risk
38
Reduction in Biochemical MarkersPredicts
Increase in Bone Mineral Density (BMD)

• Changes in markers occur before changes in BMD
• Percent changes in bone turnover markers are much
  greater than the percent change in BMD
• Significant, albeit modest, correlations exist
  between changes in bone turnover markers and
  changes in BMD within individual studies
• Significant strong correlations exist between
  changes in bone turnover markers and changes in
  BMD across studies

39
Change in Resorption Markers and BMD
Hochberg MC et al. J Clin Endocrinol Metab.
2002871586-92.
40
Change in Formation Markers and BMD
Hochberg MC et al. J Clin Endocrinol Metab.
200287In press.
41
Assessing the Effects of Antiresorptive
Therapies in Reducing Osteoporotic Fracture

• At present BMD measurement is the best
  surrogate clinical marker of therapeutic
  efficacy.

BMD Bone mineral density. Chesnut CH, Rosen CJ,
et al. J Bone Miner Res. 200116216372.
42
Relationship of Change in Bone Turnover to
Reduction in Risk of Vertebral Fractures

• Secondary analysis of data from a subset of 2,622
  participants in MORE trial who had measurement of
  bone markers
• 2-part analysis
• Association of tertiles of change in bone markers
  with risk of vertebral fracture
• Multivariate analysis of change in bone markers
  as continuous variables adjusting for confounders

MORE Multiple Outcomes of Raloxifene
Evaluation. Bjarnason NH et al. Osteoporos Int.
200112922930.
43
Changes in Biochemical Markers Predict Vertebral
Fracture Risk Reduction During Raloxifene Therapy
1.05
0.97
1
1
0.89
0.74
Relative Risk of Spine Fracture
Relative Risk of Spine Fracture
0.38
0.35
0
0
gt8.7
8.7?3.1
gt3.1
gt5.7
5.7?1.9
gt1.9
Bone Alkaline Phosphatase (µg/L)
Osteocalcin (µg/L)
Plt0.05. Bjarnason et al. Osteoporos Int.
200112922930.
44
Relationship of Change in Bone Turnover to
Reduction in Risk of Vertebral Fractures

• In patients treated with raloxifene, reduction in
  bone formation at 6 months was significantly
  related to reduction in vertebral fracture risk
  after adjustment for age, baseline BMD, BMI,
  smoking status, and prevalent vertebral fracture
• OR 0.76 (0.61, 0.96) for 8.9 µg/L ? in serum OC
• OR 0.63 (0.50, 0.80) for 5.5 µg/L ? in serum BAP

BMD Bone mineral density. BMI Body mass
index. Bjarnason NH et al. Osteoporos Int.
200112922930.
45
Relationship of Change in Bone Turnover to
Reduction in Risk of Vertebral Fractures

• Secondary analysis of data from a subset of 2,442
  participants in VERT study who had measurement of
  bone markers.
• In patients treated with risedronate, reduction
  in bone resorption at 3 months was significantly
  related to reduction in vertebral fracture risk
• 30 ? in urinary NTX/Cr 17 ? in VFx risk

VERT Vertebral Efficacy and Risedronate
Therapy. NTX/Cr N-telopeptide/creatinine. Eastel
l R et al. J Bone Miner Res. 2001Abstract 1107.
46
Reduction in Bone Turnover Predicts Fracture in
Alendronate Treated Women The FIT Study

• DC Bauer et al, for the FIT Research Group
• Presented at IOF WCO, 14 May 2002

47
Research Question

• Among alendronate and placebo treated women, are
  baseline values and changes in biochemical
  markers of bone turnover associated with
• Vertebral fracture?
• If so, are effects of bone turnover independent
  of BMD?

48
Methods Design and Subjects

• Prospective cohort study
• Fracture Intervention Trial (FIT)
• 6459 postmenopausal women
• FNBMD lt 0.68 gm/cm2
• Randomized to ALN (5 mg/d) or placebo
• ALN increased after 2 years to 10 mg/d
• Calcium/vit D supplements if needed (gt80)

49
Methods Measurements

• Bone turnover at baseline and 1 year (non-fasting
  archived serum)
• BAP (Hybritech)
• sCTX (Crosslaps)
• P1NP (Orion)
• BMD at baseline and 1 year (Hologic QDR-2000)
• Fracture outcomes (3.6 yr follow-up)
• Central review of paired lateral spine
  radiographs

50
Results Placebo Group

• Markers fell 8-31 over 1 year
• Neither baseline marker level nor change over 1
  year were associated with vertebral fracture

51
Results Alendronate Group

• Baseline marker not associated with fracture
• Greater reductions in one or more markers were
  associated with fewer vertebral fractures

52
Alendronate Group Change in BAP and Risk of
Spine Fracture
29
22
20
28
19
RR per SD 0.74 (0.63, 0.87)
Odds Ratio (CI)
Change in BAP (mg/dl)
53
Alendronate Group Change in Spine BMD and Risk
of Spine Fracture
RR per SD 0.92 (0.76, 1.11)
Odds Ratio (CI)
Change in BMD ()
54
Additional Analyses

• Change in marker results similar after
• Adjustment for baseline BMD or vertebral fx
• Limiting analysis to women with osteoporosis
• low BMD and/or existing VFx
• Further adjustment for change in spine BMD did
  not alter results

55
Conclusions

• In the Fracture Intervention Trial
• Baseline markers not associated with fracture
• Change in one or more marker was associated with
  subsequent vertebral fracture in alendronate
  treated women.
• Effect independent of BMD
• Reduction in bone turnover clearly plays an
  important role in the efficacy of alendronate
  treatment

56
Antiresorptive AgentsMechanisms of Fracture
Risk Reduction
?