Page 1 Currently Available Biochemical Markers of Bone Turnover - - PDF document

Page 1

Bone Turnover Markers in Clinical Practice: Ready for Prime Time?

Douglas C. Bauer, MD University of California San Francisco, CA No disclosures

2

Bone Remodeling Sequence

3

Currently Available Biochemical Markers

• f Bone Turnover
• Resorption (urine and serum)

– Pyridinoline and deoxypyridinoline – N-telopeptides of type 1 collagen (uNTX) – C-telopeptides of type 1 collagen (uCTX, sCTX*)

*Recommended for clinical use by IOF-IFCC Working Group

4

Free PYD and DPD (40%)

(OH) NH2 COOH OH COOH NH2 NH2 COOH C C N N 1 1 2 Hyl Hyl Bone matrix NTX COOH NH2 COOH CTX NH2

C-telopeptides N-telopeptides Crosslinked C and N-telopeptides (60%)

PYD = pyridinoline; CTX = C-telopeptides of type I collagen DPD = deoxypyridinoline; NTX = N-telopeptides of type I collagen

Markers of Bone Resorption: Type I Collagen Crosslinks

Osteoclastic bone resorption

Page 2

5

Currently Available Biochemical Markers

• f Bone Turnover
• Formation (serum)

– Osteocalcin (OC) – Bone alkaline phosphatase (Bone ALP) – N-terminal propeptide type I procollagen (PINP)*^

*Recommended for clinical use by IOF-IFCC Working Group ^Automated assay platform not FDA approved or available in US

6 C‐Terminal pro‐peptid (PICP)

+

N‐Terminal pro‐peptid (PINP) Procollagen type I Collagen type I Intact PINP (trimer ‐ liver) PINP (monomer – kidneys)

Markers of Bone Formation: PINP

7

Possible Clinical Uses of Biochemical Markers

• Predict fracture risk in untreated individuals
• Promote compliance among treated

individuals

• Identify non-compliance among treated

individuals

• Assess treatment efficacy during and after

drug therapy

8

BTMs Predicting Fracture in Untreated Individuals

• Published studies; some positive and some

negative

• Meta-analysis of 6 prospective cohort studies

without adjustment for BMD

– PINP per SD increase: RR=1.2 (CI: 1.1-1.4) – sCTX per SD increase: RR=1.2 (CI: 1.1-1.3)

• Weaker than hip BMD alone (typically RR=2-3

per SD decrease)

Johansson H et al, Calcif Tissue Int, 2014

Page 3

9

If BMD Known, Does Marker Help?

• Fracture risk calculators (e.g. FRAX) do not include
• BTMs. Should they?
• Combination of tests best assessed by AUC in ROC

analyses

• Only one study (EPIDOS) has reported ROC for

fracture prediction:

– AUC with high uCTX = 0.58 – AUC with low BMD = 0.63 – AUC with both = 0.66

• Suggests little benefit from BTM if BMD known

Garnero P et al. Osteo Int 1998

10

Possible Clinical Uses of Biochemical Markers

• Predict fracture risk in untreated individuals
• Promote compliance among treated

individuals

• Identify non-compliance among treated

individuals

• Assess treatment efficacy during and after

drug therapy

11

Use of Bone Turnover to Enhance Compliance

• BMD testing increases likelihood of

medication use

• Most medication discontinuation occurs

during the first year

• Does routine use of bone turnover shortly

after initiation of therapy enhance compliance?

12

BTM Feedback and Compliance

• 3 RCTs: women starting treatment, randomized

to BTM monitoring/feedback or no BTMs

• No improvement in compliance or persistence

with BTM feedback

• The good news: clinician visits 3, 6 and 8 mo.

after initiation increased adherence by 57%!

Silverman S et al, Osteo Int 2012 Delmas P et al, JCEM 2007 Clowes JA et al, JCEM 2004

Page 4

13

Possible Clinical Uses of Biochemical Markers

• Predict fracture risk in untreated individuals
• Promote compliance among treated

individuals

• Identify non-compliance among treated

individuals (applies to oral agents)

• Assess treatment efficacy during and after

drug therapy

14

ALN vs. PBO in FIT: Overlap of Change in BAP

Bauer DC et al, J Bone Miner Res 2004

15

Identification of Non-Adherent Individuals in IMPACT (Risedraonte)

Adherence >30% Drop in sCTX <30% Drop in sCTX

>80% 279 (77%) 84 (23%) <80% 1353 (84%) 262 (16%)

Eastell et al, JBMR 2011

16

Possible Clinical Uses of Biochemical Markers

• Predict fracture risk in untreated individuals
• Promote compliance among treated

individuals

• Identify non-compliance among treated

individuals

• Assess treatment efficacy during and after

drug therapy (particularly bisphosphonates)

Page 5

17

Biochemical Markers of Bone Turnover are Decreased by Antiresorptive Therapy*

Reduction (%) Years

1 2 3 –10 –20 –30 –40 –50 –60 –70

NTX Bone ALP

Reduction (%)

1 2 3 –10 –20 –30 –40 –50 –60 –70

Years

*Results from Fracture Intervention Trial (FIT) Placebo Alendronate

18

What is Optimal Reduction in Turnover for Fracture Protection?

• Meta-analyses of RCTs: greater reduction in

turnover associated with greater reduction in fracture risk.

– May not apply to individual patients

• Utility for individuals: analysis of change in

marker and fracture outcomes in the active treatment group of RCT

Hochberg M et al, J Clin Endocrinol Metab

19

Reduction in uCTX and New Vertebral Fracture: Risedronate

Risedronate 5 mg

3 and 6 month change in a CTX (%) Incidence %

25 15 10 5

• 70 -65
• 60 -55 -55
• 50 -45 -40
• 35 -30
• 25
• 20 -15 -10
• 5

5

0-3 year vertebral fracture incidence

n=358 risedronate-treated postmenopausal women

Eastell R et al. J Bone Miner Res. 2003

20

n=3105 alendronate-treated postmenopausal women

Bauer et al. J Bone Miner Res. 2004

Probability of non-spine fracture

• 100
• 50

50 100 0.05 0.10 0.15 0.20 0–4 year non-spine fracture incidence

1-year change in bone ALP (%)

Reduction in Bone ALP and Non-spine Fracture: Alendronate

Alendronate 5-10 mg

Page 6

21

Reduction in Bone ALP and New Vertebral Fracture: Raloxifene

1-year change in bone ALP (%) Risk of new vertebral fracture (risk ± 95% CI)

Raloxifene

• 60
• 40
• 20

20 40 60 0.03 0.06 0.09 0.12 0.15 0.18 0.21 0.24 0.27 0.30

• 80

0-3 year vertebral fracture incidence

n=601 raloxifene-treated post menopausal women

Reginster JY et al. Bone. 2004

22

What Threshold Indicates an Acceptable Clinical Response?

• Observed change > Least Significant Change?

– LSC = change unlikely due to chance (p<0.05) – LSC could be less than optimal change for a given treatment

• On-treatment BTM < premenopausal range+2SD?

– Why not 1SD? 3SD? – Is premenopausal turnover optimal?

• Base decision on fracture risk among treated

women at specific marker thresholds…

23

Placebo Group ALN Group BAP<30%† ALN Group BAP>30% Vertebral 7.3% 4.3%* 3.8%* Non-spine 9.8% 8.7% 6.8%* Hip 1.0% 0.8% 0.2%*

FIT Fracture Rates (Mean F/U 3.6 Years)

Alendronate: Fracture Rates With and Without “Good” BTM Response

†44% of ALN-treated women *p<0.001 compared to PBO group

Bauer DC et al. J Bone Miner Res. 2004.  

24

Placebo Group RIS Group sCTX<30%† RIS Group sCTX>30% Vertebral NA NA NA Non-spine NA 4.3% 1.7%* Hip NA NA NA

IMPACT Fracture Rates (Mean F/U 1 Year)

Risedronate: Fracture Rate With and Without “Good” Marker Response

†17% of RIS-treated women *p=0.002 compared to <30% group

Eastell et al. J Bone Miner Res. 2011

Page 7

25

What Happens When BTMs Indicate Suboptimal Response to Therapy?

• Little known about causes of

suboptimal marker response – Compliance? Absorption? Genetics? 2° causes?

• What should clinicians do?

– Reassess compliance? Not likely to help – Change dosing or treatment? May help but no data – Repeat test? Delay tactic, increases anxiety and costs Can sub-optimal response be improved? No data…

26

What About BTM Monitoring After Bisphosphonate Discontinuation?

• Drug Holiday: serial changes in BTMs might

indicate those at greater risk of fracture

• Only trial data is from FLEX placebo group:

– 437 postmenopausal women previously treated with alendronate for about 5 years – BTMs measured at time of discontinuation and

• ne year later

– 82 women fractured during next 5 years

Bauer DC et al, Jama Intern Med 2014

27

Fractures in FLEX Placebo Group by Tertile of 1-year Change in BTM

Bauer DC et al, Jama Intern Med 2014

28

One Other Big Uncertainty: Commercial Lab Reproducibility

• Before routine use, BTMs must be reliably

measured in clinical practice

• Greatly improved: pre-analytic factors (fasting

status, etc) and automated assays

• Laboratory performance not disseminated
• Pooled sera sent to 6 US commercial labs:

reproducibility of serum BAP and urine NTX

– Same day CV: 2-17% (NTX), 0-15% (BAP) – Longitudinal CV: 5-38% (NTX), 3-24% (BAP)

Schafer A et al, Osteo Inter 2009

Page 8

29

Does Lab Reproducibility Matter?

• Model effects of observed lab variability

(CVs) on reporting of BTM results

• Example: if a clinician orders a baseline and

follow-up BTM using the same lab, what are 95% CI for a known 50% reduction?

– Plausible range of reported results for a true 50% decrease in BTM

30

Plausible Results For Known 50% Decrease in BAP or NTX/Cr

• Plausible results for known -50% change in BAP

– Esoterix: -56% to -43% – Labcorp: -89 to +7% (i.e. could be reported as a 89% decrease or a 7% increase)

• Plausible results for known -50% change in NTX/Cr

– ARUP: -61% to -37% – Specialty: -143% to +100% Note: even worse if baseline and follow-up measurements sent to different labs!

31

Summary: Clinical Utility of Bone Turnover Markers

• Continued advances and optimism, but not

quite ready for prime time…

• Most promising clinical use: monitoring therapy

– Some believe data adequate now

• Several areas clearly need improvement

– Optimal BTM cutpoints – Improved laboratory reproducibility

• Can we show that use of BTM improves clinical
• utcomes?

32

Acknowledgements

• Investigators and staff

at San Francisco Coordinating Center

• Research funding from

NIAMS and NIA