6
Review Recommendations for monitoring antiresorptive therapies in postmenopausal osteoporosis Christian Roux *, Patrick Garnero, Thierry Thomas, Jean-Pierre Sabatier, Philippe Orcel, Maurice Audran, pour le Comité Scientifique du GRIO 1 Groupe de Recherche et d’Information sur les Ostéoporoses (Task Force for Research and Information on Osteoporosis), Hôpital Cochin, 27, rue du Faubourg St Jacques, 75014, Paris, France Received 3 February 2004; accepted 1 July 2004 Available online 11 September 2004 Abstract Antiresorptive agents are effective in preventing and treating postmenopausal osteoporosis, provided they are taken as directed. Regular physical examinations including height measurements may fail to ensure optimal compliance. Bone mineral density (BMD) measurement is indispensable for determining whether treatment is warranted. The measurement can be repeated after 2 years at least, provided quality- control procedures are adequate. BMD changes over time should be compared to the least significant change calculated from the in vivo BMD reproducibility at the measurement center. However, BMD changes are not correlated with the fracture risk reduction induced by antiresorp- tive treatment. Biochemical markers for bone turnover can be monitored after only 3–6 months provided steps are taken to control for intraindividual variability. They are useful when patient compliance is poor or the treatment response inadequate. © 2004 Published by Elsevier SAS. Keywords: Postmenopausal osteoporosis; Bone markers; Bone densitometry; Estrogens; SERMS; Bisphosphonates 1. Introduction Fracture prevention is the goal of bone resorption inhibi- tor therapy in patients with osteoporosis. However, a decrease in the fracture rate cannot be used to monitor the effective- ness of treatment in the individual patient. Absence of frac- tures during the first year is not proof that the treatment is working, and a patient with major bone fragility may con- tinue to experience fractures for several months after starting an effective drug regimen. The need for long-term use, con- straints related to administration modalities, and potential side effects of antiresorptive agents can result in poor compliance or in drug discontinuation by the patient. Thus, improving compliance is essential to ensure optimal effectiveness. Providing patients with evidence that their treatment works when taken as directed is likely to improve compliance. Thus, surrogate effectiveness criteria such as bone mineral density (BMD) and bone turnover markers may be useful. This article presents a literature review on the advantages and limitations of these criteria. We will then suggest recommendations based on the opinion of a panel of experts. 2. Physician visits Regular physician visits are invaluable for gathering data on compliance with the drug dosage and administration modalities and for detecting side effects. In addition, height measurement is useful in postmenopausal women. Although nonspecific, a reduction in height can indicate a vertebral frac- ture, disk space narrowing, exaggeration of the physiological spinal curvatures, or scoliosis. Standing height should be mea- sured using a stadiometer after checking that the line connect- ing the tragus to the nose is parallel with the floor. Standing height decreases normally during the day, by about 0.6– 1.0 cm over 12 h [1–3]. The reproducibility of * Corresponding author. Fax: +33-1-44-07-01-07. E-mail address: [email protected] (C. Roux). 1 GRIO Scientific Committee members: E. Attlan, M. Audran, B. Basse- Cathalinat, C.L. Benhamou, C. Bergot, B. Cortet, P. Dargent-Molina, E. Drapier-Faure, P. Fardellone, C. Jeandel, P.O. Kotzki, M.A. Limouzin- Lamothe, X. Marchandise,Y. Maugars, P.J. Meunier, P. Orcel, J. Puget, C. Ribot, C. Roux, J.P. Sabatier, B. Sutter, F. Trémollières, Th. Thomas, M.C. de Vernejoul, G. Weryha. Joint Bone Spine 72 (2005) 26–31 http://france.elsevier.com/direct/BONSOI/ 1297-319X/$ - see front matter © 2004 Published by Elsevier SAS. doi:10.1016/j.jbspin.2004.07.003

Recommendations for monitoring antiresorptive therapies in postmenopausal osteoporosis

Embed Size (px)

Citation preview

Page 1: Recommendations for monitoring antiresorptive therapies in postmenopausal osteoporosis

Review

Recommendations for monitoring antiresorptive therapiesin postmenopausal osteoporosis

Christian Roux *, Patrick Garnero, Thierry Thomas, Jean-Pierre Sabatier, Philippe Orcel,Maurice Audran, pour le Comité Scientifique du GRIO 1

Groupe de Recherche et d’Information sur les Ostéoporoses (Task Force for Research and Information on Osteoporosis),Hôpital Cochin, 27, rue du Faubourg St Jacques, 75014, Paris, France

Received 3 February 2004; accepted 1 July 2004

Available online 11 September 2004

Abstract

Antiresorptive agents are effective in preventing and treating postmenopausal osteoporosis, provided they are taken as directed. Regularphysical examinations including height measurements may fail to ensure optimal compliance. Bone mineral density (BMD) measurement isindispensable for determining whether treatment is warranted. The measurement can be repeated after 2 years at least, provided quality-control procedures are adequate. BMD changes over time should be compared to the least significant change calculated from the in vivo BMDreproducibility at the measurement center. However, BMD changes are not correlated with the fracture risk reduction induced by antiresorp-tive treatment. Biochemical markers for bone turnover can be monitored after only 3–6 months provided steps are taken to control forintraindividual variability. They are useful when patient compliance is poor or the treatment response inadequate.© 2004 Published by Elsevier SAS.

Keywords: Postmenopausal osteoporosis; Bone markers; Bone densitometry; Estrogens; SERMS; Bisphosphonates

1. Introduction

Fracture prevention is the goal of bone resorption inhibi-tor therapy in patients with osteoporosis. However, a decreasein the fracture rate cannot be used to monitor the effective-ness of treatment in the individual patient. Absence of frac-tures during the first year is not proof that the treatment isworking, and a patient with major bone fragility may con-tinue to experience fractures for several months after startingan effective drug regimen. The need for long-term use, con-straints related to administration modalities, and potential sideeffects of antiresorptive agents can result in poor complianceor in drug discontinuation by the patient. Thus, improvingcompliance is essential to ensure optimal effectiveness.

Providing patients with evidence that their treatment workswhen taken as directed is likely to improve compliance. Thus,surrogate effectiveness criteria such as bone mineral density(BMD) and bone turnover markers may be useful. This articlepresents a literature review on the advantages and limitationsof these criteria. We will then suggest recommendations basedon the opinion of a panel of experts.

2. Physician visits

Regular physician visits are invaluable for gathering dataon compliance with the drug dosage and administrationmodalities and for detecting side effects. In addition, heightmeasurement is useful in postmenopausal women. Althoughnonspecific, a reduction in height can indicate a vertebral frac-ture, disk space narrowing, exaggeration of the physiologicalspinal curvatures, or scoliosis. Standing height should be mea-sured using a stadiometer after checking that the line connect-ing the tragus to the nose is parallel with the floor. Standingheight decreases normally during the day, by about 0.6–1.0 cm over 12 h [1–3]. The reproducibility of

* Corresponding author. Fax: +33-1-44-07-01-07.E-mail address: [email protected] (C. Roux).

1 GRIO Scientific Committee members: E. Attlan, M. Audran, B. Basse-Cathalinat, C.L. Benhamou, C. Bergot, B. Cortet, P. Dargent-Molina, E.Drapier-Faure, P. Fardellone, C. Jeandel, P.O. Kotzki, M.A. Limouzin-Lamothe, X. Marchandise, Y. Maugars, P.J. Meunier, P. Orcel, J. Puget, C.Ribot, C. Roux, J.P. Sabatier, B. Sutter, F. Trémollières, Th. Thomas, M.C.de Vernejoul, G. Weryha.

Joint Bone Spine 72 (2005) 26–31

http://france.elsevier.com/direct/BONSOI/

1297-319X/$ - see front matter © 2004 Published by Elsevier SAS.doi:10.1016/j.jbspin.2004.07.003

Page 2: Recommendations for monitoring antiresorptive therapies in postmenopausal osteoporosis

height measurements depends on the instrument used. Thestandard deviation of the mean ranged from 0.9 to 1.7 mm ina study that used Harpenden stadiometers [3]. When twodifferent observers did each measurement twice, the standarddeviation ranged from 2.4 to 3.2 mm [4]. Most patients do notknow their exact height, and a measurement should be ob-tained around 50 years of age to serve as a reference. Adecrease of more than 3 cm compared to this reference valuemay warrant investigations, most notably radiographs of thespine [5]. Other cutoffs can be used to monitor individualpatients.

3. Bone mineral density measurement

BMD measurement is needed to select patients for os-teoporosis treatment, as there is no proof that drugs forosteoporosis (other than hormone replacement therapy[HRT]) are beneficial in women with clinical risk factors forfractures but normal BMD values. The use of BMD measure-ment as a tool for monitoring osteoporosis treatment raisesboth metrological and clinical issues.

3.1. Metrological properties of bone mineral densitymeasurement

Reproducibility is far better for BMD measurement thanfor most laboratory tests. Reproducibility is usually 1–2% atthe spine on anteroposterior images and 2–3% at the proxi-mal femur in individuals with normal BMD values; the dif-ference between the two sites is ascribable to greater difficul-ties with repositioning and examining the femur, ascompared to the spine. However, these data obtained undernearly experimental conditions may not apply to everydayclinical practice. Reproducibility depends heavily on quality-assurance factors, including tests to control the quality andperformance of the machine, as well as the experience of theoperator. Assessment of machine performance requires dailyscanning of a phantom (which may be anthropomorphic ornot), followed by calculation of the in vitro coefficient ofvariation (CV), which serves to evaluate short-term and long-term performance and to detect drift in measurement accu-racy. These in vitro data, however, do not necessarily reflectin vivo reproducibility, which should be evaluated at eachmeasurement center. Measurements are obtained either threetimes in each of 15 patients or twice in each of 30 patients,and the CV (m/r) is calculated from the mean (m) andstandard deviation (r) of these repeated measurements. TheCV is expressed as a percentage and depends on mean BMDvalues [6]. The standard deviation reflects measurement er-ror, which is a characteristic of machine performance and isindependent from the value measured. It can be used tocompute the least significant change (LSC) that characterizesa given measurement center; with a confidence interval of95%, the LSC is calculated as 2.77r. Thus, in practice, twoabsolute values (g/cm2) can be compared [7], rather than two

percentages. When serial measurements are obtained in apatient, only changes greater than the LSC (in g/cm2) can beascribed to treatment effects. Smaller changes may be relatedto measurement error.

The time interval between two measurements in the samepatient must be long enough to allow occurrence of a changegreater than the LSC. Therefore, it depends on the expectedrate of change, which varies according to whether the mea-surement site is composed predominantly of trabecular or ofcortical bone. At the femoral neck for instance, the LSC isusually about 2.77 × 0.011 g/cm2, i.e., about 30 mg/cm2, andconsequently treatment-related changes cannot be evaluatedin the short-term. At the lumbar spine on an anteroposteriorview, the LSC is usually about 2.77 × 0.009 g/cm2, i.e., about25 mg/cm2, so that a treatment-induced BMD increase can bedetected after 2 years. The small changes induced by ralox-ifene therapy cannot be detected in individual patients. Al-though the spine may not be the best site for the diagnosis ofosteoporosis given the high prevalence of spinal degenerativedisease, it is the most sensitive site for detecting changes overtime.

Computing the LSC provides a means of accounting forregression toward the mean, which is a common property ofvariation in biochemical parameters [8]. Regression towardthe mean is a statistical phenomenon in which values ob-tained over time tend to move closer to the mean, as com-pared to initial values. For instance, within a patient popula-tion, those individuals with the largest BMD gains after1 year will tend to have the smallest gains after 2 years, andvice versa. However, use of the LSC shows that these indi-vidual variations are related to fluctuations in measurementerror rather than to genuine biological variation [7].

3.2. Clinical usefulness of bone mineral densitymeasurement

The clinical usefulness of serial BMD measurement de-serves discussion, in the light of several factors. Given thesize of the LSC, serial measurements must be spaced at least2 years apart to detect variations that are larger than thoseproduced by measurement error. With the exception of HRT,treatment dosages cannot be adjusted on the basis of BMDchanges. There is no proof that repeating BMD measure-ments improves compliance, as most patients discontinueantiresorptive medications after a few months because ofadministration constraints, side effects, or lack of interest.Above all, BMD is used as a surrogate marker for the fracturerisk, yet BMD increases do not reliably reflect a reduction inthe fracture risk. Although bisphosphonates, raloxifene, andHRT have not been compared in the same study, they seem toproduce comparable reductions in the risk of vertebral frac-tures, of about 30–50%, whereas BMD changes differ mark-edly across medications. Studies have shown that BMD gainsexplain only a small proportion of the vertebral fracture riskreduction: 28% with risedronate [9], 16% with alendronate[10], and 4% with raloxifene [11]. It has been suggested that

27C. Roux et al. / Joint Bone Spine 72 (2005) 26–31

Page 3: Recommendations for monitoring antiresorptive therapies in postmenopausal osteoporosis

the percentage of BMD change may be related to the changein the relative risk of fracture [12]. In one study, a linearrelationship was found between these two parameters, but a1% increase in spinal BMD was associated with an only 3%decrease in the relative risk of vertebral fracture [10]. Forperipheral fractures, in contrast, the risk reduction is clearlyrelated to the BMD gain [13]. Preliminary data suggestingthat the fracture risk may decrease despite a reduction inBMD have not been confirmed [14]. Common sense indi-cates that a BMD increase during treatment should be pref-erable over a BMD decrease. However, a prospective studyfound that the fracture risk was more heavily dependent onBMD at baseline than on BMD changes during treatment[15].

3.3. Conclusion

Serial BMD measurements can be used to monitor currentantiresorptive treatments (bisphosphonates, raloxifene, andHRT) provided a number of conditions are met. Adequatequality-control procedures must be used. At each measure-ment center, the LSC must be calculated from in vivo repro-ducibility data. The LSC is used to estimate the significanceof observed changes, in absolute values. BMD measure-ments must be spaced at least 2 years apart. The main goal ofserial BMD measurement is to check that no further boneloss has occurred; quantitation of BMD gains is the second-

ary objective. This should be explained to the patients, manyof whom expect to recover normal BMD values. Finally,BMD should be measured at completion of a treatment se-quence, in order to obtain new reference values.

4. Biochemical markers for bone turnover

Table 1 lists the main bone turnover markers. At present,routine assays of bone turnover markers are not recom-mended as a means of selecting patients for antiresorptivetreatment. One study found that the efficacy of antiresorptivetreatment was independent from the baseline bone turnoverlevel [16]. When using these markers as treatment monitor-ing tools, intraindividual variability should be taken intoaccount (Table 2).

4.1. Variability in bone turnover markers

Intraindividual variability limits the usefulness of boneturnover markers. However, its impact can be minimized by acareful examination of preanalytical and analytical sourcesof variability. Manual and/or automated assays are availablefor most bone turnover markers. Within-run and run-to-runreproducibilities range from 3% to 10% [17]. Several factorsresponsible for preanalytical variability cannot be controlled,such as age, sex, hormonal status, renal and hepatic function,

Table 1Biochemical markers for bone turnover. Markers that can be used in clinical practice are in bold type

Bone formation Bone resorptionSerum Plasma/serumOsteocalcin Tartrate-resistant acid phosphatase (TRAP 5b)Total alkaline phosphatase and bone alkaline phosphatase N-terminal (NTX) and C-terminal (CTX) telopeptides of type I collagenProcollagen I C-terminal and N-terminal propeptides (PICP and PINP) Urine

Free pyridinoline (PYD) et deoxypyridinoline (DPD)Procollagen I N-terminal (NTX) and C-terminal (CTX) propeptidesHelical peptide of chain alpha I of collagen type IUrinary calcium excretionUrinary hydroxyproline excretionGalactosyl-hydroxylysine

Table 2Intraindividual variability (as the coefficient of variation, CV%) in biochemical markers for bone turnover in untreated postmenopausal women (adapted fromHannon and Eastell [30])

Bone formation Bone resorptionShort-term CV%(1–5 weeks)

Long-term CV%(3 months–3 years)

Short-term CV%(1–5 weeks)

Long-term CV%(3 months–3 years)

Serum SerumOsteocalcin 7–13 8 NTX 3.3 7.5Bone AP 7–13 9 CTX 8 (fasting) 9–18

14 (nonfasting)PINP – 7.5 UrinePICP 10.6 8.6 Total DPD (HPLC) 12–24 17–63

Free DPD (ELISA) 12 9–13CTX – 18–24NTX 10–18 16–25

PAO: bone alkaline phosphatase; PINP: N-terminal propeptide of collagen type I; PIcP: C-terminal propeptide of collagen type I; DPD: deoxypyridinoline;NTX: N-terminal telopeptide of collagen type I; CTX : C-terminal telopeptide of collagen type I

28 C. Roux et al. / Joint Bone Spine 72 (2005) 26–31

Page 4: Recommendations for monitoring antiresorptive therapies in postmenopausal osteoporosis

comorbidities, a recent history of fracture, and current treat-ments. These factors should be sought routinely and takeninto account when interpreting marker values [18]. Control-lable sources of preanalytical variability include samplingtime, sample preservation procedures, and food intake. Se-rum levels of the carboxy-terminal peptide of collagen I(CTX), for instance, vary over the 24-h cycle and are influ-enced by meals (Fig. 1). This was shown by collecting bloodsamples at 3-h intervals over two 24-h periods in 11 post-menopausal women. The women fasted throughout one 24-hperiod and ate normally during the other. Mean variability inCTX values over 24-h was 13.6% without and 34% with food[19]. Therefore, to ensure satisfactory reproducibility of se-rial measurements in a given patient, serum and urinesamples must be collected in the morning after an overnightfast; serum samples should be obtained before 9 A.M., ifpossible, and urine samples at the first or second voiding inthe morning. Urinary values should be normalized for theurinary creatinine level. When these rules are followed, in-traindividual variability in untreated postmenopausal womenranges from about 5% to 15% for serum bone formation andresorption markers and from 12% to 25% for urinary boneresorption markers. The LSC of values obtained under opti-mal conditions can be calculated for each marker. In afollow-up study of over 2000 women given bisphosphonatetherapy for osteoporosis, the proportion of patients with adecrease in urinary amino-terminal peptide of collagen I(NTX) exceeding the LSC was 65% after 3 months and 84%after 6 months.

4.2. Effect of treatments on biochemical markers

After the initiation of antiresorptive therapy, bone resorp-tion markers decline gradually until they reach a plateau after3–6 months, whereas bone formation markers decrease laterand reached a plateau after 6–12 months. The magnitude ofthe decreases varies according to the potency of the medica-tion (with bisphosphonates producing the largest reductions),the marker, and the treatment regimen (route of administra-tion of HRT or continuous vs. cyclic administration of bis-

phosphonates). After 6 months of oral treatment with17 beta-estradiol (2 mg/d), urinary CTX and NTX excretiondecreased by about 60% and free urinary deoxypyridinolineby about 30% [20]. Similar variations were noted when17 beta-estradiol was given percutaneously, transdermally, orintranasally [21,22]. During the first 3 months of treatment,changes in bone formation markers may vary with the routeof administration. Thus, serum osteocalcin levels decreasedby 30% in women given oral estrogen therapy but showed nochange with estrogen therapy given percutaneously or intra-nasally. This difference has been ascribed to a hepatic first-pass effect, which may modify the biological effect of insulingrowth factor-1 [23]. With continuous oral bisphosphonatetherapy, decreases are largest for type 1 collagen peptidemarkers (CTX, NTX, and helical peptide), intermediate fortotal pyridinoline and deoxypyridinoline excretion, andsmall or absent for urinary excretion of free pyridinoline anddeoxypyridinoline. These differences in response may beascribable to differences across bisphosphonates regardingeffects on the enzymes involved in collagen breakdown andon the renal metabolism of pyridinoline peptides. With oralalendronate, 10 mg/day, urinary CTX and NTX decreased byabout 70% and total deoxypyridinoline by about 50%. Oralrisedronate, 5 mg/day, decreased peptide pyridinoline mark-ers (urinary NTX and serum CTX) by 35–50%. Reductionsin bone resorption markers are similar with daily vs. weeklyadministration of alendronate or risedronate. After intrave-nous ibandronate or pamidronate, bone resorption markersdrop sharply then increase slowly. The time to, and magni-tude of, the increase vary with the dosage and frequency ofthe injections. Zoledronate induces long-lasting changes inbone resorption markers. Discontinuation of alendronatetherapy is followed by an increase in bone remodeling thatdenotes resumption of bone cell activity. When alendronatewas stopped after 5 years, the bone turnover increase wasabout 20%, so that the therapeutic gains were not entirely lost[24]. This phenomenon should be taken into account whendesigning treatment strategies. Changes in bone remodelingmarkers after discontinuation of risedronate are being evalu-ated. As compared to bisphosphonates or HRT, oral ralox-ifene induces smaller decreases in bone remodeling markers,of about 30–40% for urinary CTX and 20–30% for boneformation markers [25]. Calcium and vitamin D supplementsin physiological dosages induce small but significant de-creases in bone resorption markers, of about 10–20% [26].

4.3. Usefulness of markers in clinical practice

In prospective studies, changes in bone remodeling mark-ers were linked to changes in BMD, although the correlationswere weak. It has been suggested that markers may be usefulfor identifying patients whose BMD fails to increase undertreatment. Predicting a BMD gain greater than 3% with 90%specificity requires a 45% decrease in a urinary resorptionmarker (CTX or NTX) with HRT and a 65% decrease withalendronate. In practice, markers are useful mainly for pre-

Fig. 1. Circadian variations in serum levels of carboxy-terminal peptide ofcollagen I (CTX) and effects of food intake (from Quest et al. [18]).

29C. Roux et al. / Joint Bone Spine 72 (2005) 26–31

Page 5: Recommendations for monitoring antiresorptive therapies in postmenopausal osteoporosis

dicting the fracture-preventing effect of treatment. After6 months of raloxifene therapy, changes in serum osteocalcinor bone alkaline phosphatase were associated with the risk ofsubsequent vertebral fracture, whereas changes in BMDwere not [27]. During risedronate therapy, changes in urinaryCTX and NTX after 3 and 6 months explained 50–70% of thereduction in vertebral fractures and 54–74% of the reductionin nonvertebral fractures [28]. In this study, the relationbetween the magnitude of the marker decrease and the risk ofvertebral fractures was not linear (Fig. 2); no additionalfracture-preventing effect was obtained beyond a 60% reduc-tion in urinary CTX or a 40% reduction in urinary NTX [28].This finding of considerable practical relevance indicates aneed for similar studies with other antiresorptive medica-tions. A study of alendronate reported in abstract form foundthat the risk of vertebral fractures over nearly 4 years wassignificantly correlated with 1-year decreases in bone alka-line phosphatase, procollagen type 1 N-terminal propeptide(PINP), and serum CTX [29].

4.4. Conclusion

Biochemical markers can be used to monitor antiresorp-tive treatments (bisphosphonates, raloxifene, or HRT) in pa-tients with osteoporosis. Only changes greater than the LSCfor the marker of interest indicate treatment-related effects.The marker most relevant to the medication used should beselected for monitoring. Assays of resorption markers shouldnot be done before 3–6 months and those of formation mark-ers before 6 months. Assays of bone remodeling markers arenot recommended for routine monitoring but may be usefulwhen poor compliance with the dosage or administrationinstructions is suspected or when the treatment is ineffective,in order to assist in the selection of the second-line treatment.

5. Conclusion

When used with discernment, absorptiometry and assaysof markers for bone remodeling are helpful for monitoring

antiresorptive treatments in patients with postmenopausalosteoporosis. These tests have a number of limitations. Forinstance, they provide no information on changes in bonemicroarchitecture, which are currently a focus of method-ological and technological research. However, osteoporosisis a chronic disease for which burdensome treatments aregiven for many years, making intermediate evaluation oftreatment effectiveness crucial. BMD and bone remodelingmarkers will generate even greater interest when anabolicagents become available, as these medications can inducemajor BMD changes in the short-term and act by stimulatingbone remodeling within a few months.

References

[1] Renier JC, Audran M, Seret P, Secher V. Contribution à la connais-sance de l’évolution naturelle de l’ostéoporose: 30 malades avec unrecul moyen de 11 ans. Rev Rhum Mal Osteoart 1986;53:451–7.

[2] Tyrrell AR, Reilly T, Troup JDG. Circadian variation in stature and theeffects of spinal loading. Spine 1985;10:161–4.

[3] Coles RJ, Clements DG, Evans WD. Measurement of height: practicalconsiderations for the study of osteoporosis. Osteoporos Int 1994;4:353–6.

[4] Marks GC, Habricht JP, Mueller WH. Reliability, dependability andprecision of anthropometric measurements. Am J Epidemiol 1989;130:578–87.

[5] Gunnes M, Lehmann EH, Mellström D, Johnell O. The relationshipbetween anthropometric measurements and fractures in women. Bone1996;19:407–13.

[6] Glüer CC, Blake G, Lu Y, Blunt BA, Jergas M, Genank HK. Accurateassessment of precision errors: how to measure the reproducibility ofbone densitometry techniques. Osteoporos Int 1995;5:262–70.

[7] Ravaud P, Reny JL, Giraudeau B, Porcher R, Dougados M, Roux C.Smallest detectable difference in individual bone mineral densitymeasurements. J Bone Miner Res 1999;14:1449–56.

[8] Cummings SR, Palermo L, Browner W, Marcus R, Wallace R, Pear-son J, et al. Monitoring osteoporosis therapy with bone densitometry.Misleading changes and regression to the mean. JAMA 2000;283:1318–21.

[9] Li Z, Meredith MP, Hoseyni MS. A method to assess the proportion oftreatment effect explained by a surrogate end-point. Statist Med 2001;20:3175–88.

Fig. 2. Use of bone resorption markers (urinary amino- and carboxy-terminal peptides of collagen I, NTX and CTX) for predicting the effectiveness of oralrisedronate in decreasing incidence vertebral fractures (from Eastell et al. [28]).

30 C. Roux et al. / Joint Bone Spine 72 (2005) 26–31

Page 6: Recommendations for monitoring antiresorptive therapies in postmenopausal osteoporosis

[10] Cummings SR, Karpf DB, Harris F, Genant HK, Ensrud K, Lac-roixAZ, et al. Improvement in spine bone density and reduction in riskof vertebral fractures during treatment with antiresorptive drugs. Am JMed 2002;112:281–9.

[11] Sarkar S, Mitlak BH, Wong M, Stock JL, Black DM, Harper KD.Relationship between bone mineral density and incident vertebralfracture risk raloxifene therapy. J Bone Miner Res 2002;17:1–10.

[12] Wasnich RD, Miller P. Antifracture efficacy of antiresorptive agentsare related to changes in bone density. J Clin Endocrinol Metab2000;85:231–6.

[13] Hochberg MC, Greenspan S, Wasnich RD, Miller P, Thompson DE,Ross PD. Changes in bone density and turnover explain the reductionsin incidence of non-vertebral fractures that occur during treatementwith antiresorptive agents. J Clin Endocrinol Metab 2002;87:1586–92.

[14] Cummings SR, Palermo L, Ensrud KE, Hochberg MC. Are non-responders responding? J Bone Miner Res 2000;15(suppl):144[abstract].

[15] Hochberg MC, Ross PD, Black D, Cummings SR, Genant HK, Nev-itt MC, et al. Larger increases in bone mineral density during alendr-onate therapy are associated with a lower risk of new vertebral frac-tures in women with postmenopausal osteoporosis. Arthr Rheum1999;42:1246–54.

[16] Seibel MJ, Naganathan V, Barton I, Grauer A. Relationship betweenpretreatment bone resorption and vertebral fracture incidence in post-menopausal osteoporotic women treated with risedronate. J BoneMiner Res 2004;19:394–401.

[17] Garnero P, Bianchi F, Carlier MC, Genty V, Jacob N, Kamel S, et al.Biochemical markers of bone remodeling: pre-analytical variationsand guidelines for their use. Ann Bio Clin (Paris) 2000;58:683–704.

[18] Qvist P, Christgau S, Pedersen BJ, Schlemmer A, Christiansen C.Circadian variation in the serum C-terminal telopeptide of type Icollagen (Serum CTx): effects of gender, age, menopausal status,posture, daylight, serum cortisol and fasting. Bone 2002;31:57–61.

[19] Delmas PD, Garnero P. Utility of biochemical markers of bone turn-over in osteoporosis. In: Marcus R, Feldman D, Kelsey J, editors.Osteoporosis. New York: Academic Press; 1996.p. 1075–85.

[20] Cooper C, Stakkestad JA, Radowicki S, Hardy P, Pilate C, Dain MP,et al. Matrix delivery transdermal 17ß-estradiol for the prevention ofbone loss in postmenopausal women. Osteoporos Int 1999;9:358–66.

[21] Garnero P, Tsouderos Y, Marton I, Pelissier C, Varin C, Delmas PD.Effects of intranasal 17ß-estradiol on bone turnover and seruminsulin-like growth factor I in postmenopausal women. J Clin Endo-crinol Metab 1999;84:2390–7.

[22] Ho KKY, Weissenberg AJ. Impact of short-term estrogen administra-tion on growth hormone secretion and action: distinct route-dependent effects on connective and bone tissue metabolism. J BoneMiner Res 1992;7:821–7.

[23] Bellantoni MF, Harman DM, Cho DE, Blackman MR. Effects ofprogestin-opposed transdermal estrogen administration on growthhormone and insulin-like growth factor-I in postmenopausal womenof different ages. J Clin Endocrinol Metab 1991;72:172–8.

[24] Tonino RP, Meunier PJ, Emkey R, Rodriguez-Portales JA, Men-kes CJ, Wasnich RD, et al. Skeletal benefits of alendronate: 7-yeartreatment of postmenopausal osteoporotic women. Phase IIIOsteoporosis Treatment Study Group. J Clin Endocrinol Metab 2000;85:3109–15.

[25] Delmas PD, Bjarnason NH, Mitlak BH, Ravoux AC, Shah AS,Huster WJ, et al. Effects of raloxifène on bone mineral density, serumcholesterol concentrations, and uterine endometrium in postmeno-pausal women. N Engl J Med 1997;337:1641–7.

[26] Grados F, Brazier M, Kamel S, Mathieu M, Hurtebize N, Maamer M,et al. Prediction of bone density variation by bone remodeling markersin post menopausal women with vitamin D insufficiency treated withcalcium and vitamin D supplementation. J Clin Endocrinol Metab2003;88:5175–9.

[27] Bjarnason NH, Christiansen C, Sarkar S, Mitlak B, Knickerbocker R,Delmas P, Cummings S, for the MORE Study Group. 6-monthchanges in biochemical markers predict 3-year response in vertebralfracture rate in postmenopausal, osteoporotic women: results from theMORE study. Osteoporos Int 2001;12:922–30.

[28] Eastell T, Barton I, Hannon RA, Chines A, Garnero P, Delmas PD.Relationship of early changes in bone resorption to the reduction infracture risk with risedronate. J Bone Miner 2003;18:1051–6.

[29] Bauer DC, Black DM, Garnero P, Hochberg M, Ott S, Schneider DL,et al. Reduction in bone turnover predicts hip, non-spine, and vertebralfracture in alendronate treated women: the fracture intervention trial. JBone Miner Res 2002;17(suppl 1):S187.

[30] Hannon R, Eastell R. Preanalytical variability of biochemical markersof bone turnover. Osteoporosis Int 2000;11(suppl 6):S30–S44.

31C. Roux et al. / Joint Bone Spine 72 (2005) 26–31