- Email: [email protected]
Interest of biochemical markers of bone turnover for long-term prediction of new vertebral fracture in postmenopausal osteoporotic women
Maturitas 44 (2003) 259 /265 www.elsevier.com/locate/maturitas
Interest of biochemical markers of bone turnover for long-term prediction of new vertebral fracture in postmenopausal osteoporotic women Olivier Bruyere a,b,c, Julien Collette c, Pierre Delmas d, Alain Rouillon e, Christian Roux f, Laurence Seidel g, Florent Richy a,b, Jean-Yves Reginster a,b,c,h,* a
WHO Collaborating Center for Public Health Aspect of Osteoarticular Disorders, Liege, Belgium b Department of Public Health and Epidemiology, University of Liege, Liege, Belgium c Bone and Cartilage Research Unit, Policlinique BRULL, University of Liege, 45 Quai G. Kurth, 4020 Liege, Belgium d Service de Rhumathologie, Hoˆpital Edouard Herriot, Lyon, France e Sanofi Recherche, Gentilly, France f Service de Rhumatologie, Hoˆpital Cochin, Universite´ Rene´ Descartes, Paris, France g Biostatistical Department, University of Liege, Liege, Belgium h Georgetown University Medical Center, Washington, DC, USA Received 19 March 2002; received in revised form 21 November 2002; accepted 7 January 2003
Abstract Objective: To analyse the interest of baseline levels and short-term (3-months) changes in serum osteocalcin (BGP), serum bone-specific alkaline phosphatase (BALP) and urinary C-telopeptide of type I collagen/creatinine ratio (UCTX) to predict 3-years changes in bone mineral density (BMD) and spinal deformity index (SDI) in postmenopausal osteoporotic women. Methods: Data were derived from a cohort of 603 osteoporotic women corresponding to the placebo arm of a 3-years prospective, double-blind study. Results: Baseline values of BALP, BGP and U-CTX were negatively and significantly correlated with baseline spinal BMD. Significant correlations were also observed between the changes in BMD observed after 36 months at the spine and baseline BALP (r /0.20, P /0.0001), BGP (r /0.09, P /0.05) and U-CTX (r / /0.11, P /0.02). At 3 years, 71 women (15.9%) showed an increase in their SDI, corresponding to the occurrence of at least one new vertebral deformity. Baseline values of the four bone turnover markers (BTM) were not significantly related to the occurrence of new vertebral deformities. However, when considering the changes in the BTM observed after 3-months of follow-up, BGP (P /0.003) and U-CTX (P /0.047) were identified as significant predictors of an increase of SDI. The associated odds ratios (95% confidence interval (CI)) were 10.922 (2.218 /53.78) for unit changes of log BGP and 1.369 (1.003 /1.867) for unit changes of log U-CTX. The relative risk (RR) (IC 95%) of having a new vertebral fracture over 36 months was 0.31 (0.15 /0.65) when being in the lowest quartile of 3-months changes in BGP as compared with the highest. Conclusion: We conclude that two sequential
* Corresponding author. Tel.: /32-4-270-3257; fax: /32-4-270-3253. E-mail address: [email protected] (J.-Y. Reginster). 0378-5122/03/$ - see front matter # 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0378-5122(03)00042-2
260
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
measurements of BGP and U-CTX performed at 3-months intervals could be of interest to identify postmenopausal osteoporotic women with the highest risk to present new vertebral deformities. # 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Osteoporosis; Bone turnover markers; Vertebral fracture; Bone mineral density
1. Introduction Bone mineral density (BMD), can now be easily assessed by means of dual X-ray absorptiomertry (DXA). Low BMD is widely considered as the best predictor of later developing an osteoporotic fracture [1,2]. However, due to the variability of the DXA technique and the relatively slow rate of BMD changes, clinically relevant variations in BMD usually request a gap of 2 /3-years between two measurements. Conversely, biochemical markers of bone remodelling provide a dynamic picture of the actual rate of bone turnover, as opposed to the static approach provided by the BMD measurement [3]. They are widely used in daily practice and clinical research. Previous studies have extensively documented their ability to predict future fracture risk in healthy postmenopausal women, to a similar extend than measurement of BMD do [4 /7]. However, relatively few prospective, long-term studies, have evaluated their interest to predict the risk of women with prevalent vertebral fractures to develop incident vertebral deformities. Based on a 3-years extensive follow up of a large cohort of untreated osteoporotic women, we evaluated the ability of three currently marketed biochemical markers of bone turnover to predict changes in lumbar and hip BMD and to identify women who are more prone to experience new vertebral deformities.
the effect of intermittent cyclical tiludronate, a third generation bisphosphonate, in the treatment of established postmenopausal osteoporosis [8]. The primary inclusion criteria were the presence of at least two ‘mild’ or one ‘moderate’ but no more than five vertebral fractures of any severity and a low BMD (T -score B/ /1.7) of at least two assessable lumbar vertebrae (L2, L3, L4) measured with DXA. Patients were advised to include at least 700 mg of calcium in their daily diet. Nevertheless, a daily calcium supplementation (Calcium Sandoz8, Sandoz, Basle, Switzerland) equivalent to 500 mg elemental calcium was provided everyday except the first 7 days of each month. 2.2. Methods
2.1. Population
2.2.1. Bone densitometry BMD (g/cm2) of the antero-posterior spine and the femoral neck was measured at baseline and after 3 years, by dual energy X-ray densitometry using Hologic (Waltham, MA) or Lunar (Lunar Corporation, Madison, WI) instruments. BMD of each subject was measured by the same instrument throughout the entire study. Quality control of densitometric equipment and data was dual-faceted. Firstly, there was daily a quality control of the Hologic phantom at each center and in case of a significant drift in the QC observation, a correction factor was applied. Secondly, each scanner was cross-calibrated with two transportable phantoms at baseline and every 6 months, during visits performed by the Center for Clinical and Basic Research, (CCBR) (Ballerup, Denmark), Quality Assurance Center team.
The population was constituted by a cohort of 603 Caucasian (n /602) or Asian (n /1) women, aged between 50 and 80 years, corresponding to the placebo arm of a double blind, prospective, randomised, placebo-controlled study evaluating
2.2.2. Radiography Lateral radiographs of both thoracic and lumbar spine were obtained at baseline and 36 months. Anterio-posterior views were added at baseline. The radiographs were obtained according to a
2. Patients and methods
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
standardised protocol, with the beam centred on T7 and L3 for thoracic and lumbar spine, respectively. The interpretation of all radiographs was performed by an independent consultant (Christian Roux). Both quantitative and semi-quantitative assessments of vertebral deformities were performed with all available radiographs. A semi quantitative vertebral deformity score was assigned to each individual vertebra from T4 to L4 according to the method previously published by Genant and colleagues [9]. Score ‘0’ for normal, non-fractured vertebra; score ‘1’ for a mild deformity with a 20/25% reduction in anterior, middle or posterior height (or all three); score ‘2’ for a moderate deformity with a /25 /40% reduction in any height; score ‘3’ for a severe deformity with a reduction of /40% in any height. A new fracture in a particular vertebra is defined as the progression from a grade 0 to a grade of 1 or higher. The spinal deformity index (SDI) for a subject was defined as the sum of the individual vertebral deformities scores, divided by the number of vertebrae evaluated. 2.2.3. Biochemical markers Blood and urine samples were obtained after an overnight fasting at baseline and after 3 and 36 months. Three BTM were taken into account: / Serum Osteocalcin was determined by a commercially available solid-phase ‘sandwich’ immunoradiometric assay (ELSA-Osteoporosis, CIS-Bio International, France). / Urine C-telopeptide of type I collagen was measured by a commercially available enzymelinked immuno-sorbent assay (Crosslaps8 Tm EIA, CIS-Bio International, France). / Serum bone specific alkaline phosphatase determinations were performed using a commercially available immunoradiometric assay (‘Tandem’-R Ostase Beckmann Coulter, USA). 2.2.4. Statistical analysis Quantitative variables were expressed as mean 9/S.D. Relationships between each of the bone turnover markers (BTM) and the baseline values or the long-term (36-months) changes in BMD (spine or femoral neck) were assessed by the
261
Pearson correlation coefficient, by multiple and stepwise regression. We also assessed 3-years changes in BMD in each quartile of baseline values of biochemical markers. Differences between quartiles were assessed by ANOVA. The predictive value of the BTM as explanatory variables for an increase of SDI after 36 months was assessed by a stepwise logistic regression. We also looked at the RR of having a new vertebral fracture over 36 months in the lowest quartile as compared with the highest quartile of baseline and 3-months change’s values of biochemical markers. To normalise the distribution of BTM, we performed a logarithmic transformation of the BTMs values. The results were considered significant at the 5% level (P B/0.05). Statistical calculations were carried out with the SAS software version 6.12 (SAS Institute, Cary, NC, USA).
3. Results Baseline demographics of the population are summarised in Table 1. Baseline values of serum bone-specific alkaline phosphatase (BALP), serum osteocalcin (BGP), and urinary C-telopeptide of type I collagen/creatinine ratio (U-CTX) were negatively and significantly correlated with baseline spinal BMD (r / /0.17, P B/0.0001 for BALP, r / /0.02, P B/0.0001 for BGP, r / /0.12, P /0.005 for U-CTX). However, when performing a stepwise regression, BGP remained the only marker to be significantly correlated to baseline spinal BMD (P /0.0002). For the feTable 1 Baseline demographics (mean S.D.) Variables
Mean (S.D.)
Number of patients Age (year) BMI (kg/m2) Years since menopause (year) Femoral neck BMD (g/cm2) Spine BMD (g/cm2) BGP (ng/ml) BALP (ng/ml) U-CTX (mg/mmol creatinine)
603 67.4 (6.8) 25.3 (3.8) 19.5 (8.0) 0.596 (0.087) 0.731 (0.094) 26.70 (8.28) 13.62 (5.82) 270.96 (230.1)
262
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
moral baseline BMD, negative and significant correlations were found for BALP (r / /0.09, P /0.029) and BGP (r / /0.16, P B/0.0001). The same markers remained significant after a stepwise regression (P /0.0041 and P B/0.0001 for BALP and BGP, respectively). Significant correlations were observed between the changes in spinal BMD observed after 36 months and baseline BALP (r /0.20, P B/0.0001) (Fig. 1), BGP (r /0.09, P /0.05) and U-CTX (r / /0.11, P /0.02) (Fig. 2). After stepwise regression, only BALP and U-CTX remained significant (P /0.0001 and P /0.001, respectively). When dividing our population into quartiles of baseline BALP values, patients in the lowest quartile ( B/9.7 ng/ml) experienced, after 3 years, a mean (S.D.) increase of 0.004 (0.04) g/cm2 at the spine compared with 0.031 (0.052) g/cm2 in the highest quartile (/16.2 ng/ml) (P B/0.0001). For U-
CTX, increase was 0.018 (0.053) and 0.004 (0.046) g/cm2 in the lowest (B/121.1 mg/mmol) and in the highest (/340.7 mg/mmol) quartiles, respectively (P /0.03). At the femur, no significant relationships were observed. After 3-years of follow-up, 71 women (15.9%) showed an increase in their SDI, corresponding to the occurrence of at least one new vertebral deformity. In this subgroup of patients, the baseline spine BMD values was 0.719 (0.111) compared with 0.744 (0.087) g/cm2 in the non-increased SDI subgroup (P /0.03). At the femoral neck, the BMD was not statistically significant between the two groups (P /0.71). Regarding the BTM at baseline in the increased SDI group, BALP value (14.71 (8.44) ng/ml) was statistically different compared with the non increased SDI group (13.09 (5.29) ng/ml) (P /0.03). The two other markers were not statistically different between
Fig. 1. Correlation between the changes in spinal BMD observed after 36 months and baseline BALP.
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
263
Fig. 2. Correlation between the changes in spinal BMD observed after 36 months and baseline U-CTX.
the two groups (P /0.27 for BGP and P /0.32 for U-CTX). Using a stepwise analysis regression, the baseline values of the three BTM were not significantly related to an increase in the SDI (P /0.2 for BALP, P /0.14 for BGP, P /0.96 for U-CTX). When considering the changes in the BTM observed after 3-months of follow-up, BGP (P / 0.003) and U-CTX (P /0.047) but not BALP were identified as significant predictors of an increase of SDI after 3 years. The associated odds ratios (95% confidence interval (CI)) were 10.92 (2.22 /53.78) for unit changes of log BGP and 1.37 (1.003 /1.87) for unit changes of log U-CTX. The RR (95% CI) of having a new vertebral fracture over 36 months when being in the lowest compared with the highest quartile was 0.66 (0.37/1.17) for BALP, 1.12 (0.62/2.04) for BGP, 0.88 (0.50/1.53) for UCTX, 0.89 (0.51/1.54) for 3-months changes in BALP, 0.60 (0.34 /1.09) for 3-month changes in U-
CTX and 0.31 (0.15/0.65) for 3-month changes in BGP.
4. Discussion As previously reported [6,10], baseline biochemical markers of bone formation (BALP, BGP) and bone resorption (U-CTX) were negatively correlated with bone density at the spine and at the hip. High levels of bone remodelling are associated with a higher risk of low BMD. However, because of the low magnitude of the explained variance (5.7%), BTM offer little practical information for estimating BMD levels in an individual woman and can not be used as a surrogate measure to predict bone mass. Since a rapid decrease in bone mass is a major risk factor for fracture, in addition to baseline BMD [11], it is important to identify patients whose BMD will show the highest de-
264
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
crease over time. While a single BMD measure by DXA is considered accurate and precise [12], the BMD variability for repeated measurements is in the range of 3 /4%, hence within the same magnitude as the physiological changes expected over 2/ 4 years [3]. Therefore, a duration of 2/3 years is actually needed before an observed change in BMD can be regarded as clinically relevant. One of the goals of BTM could thus be to serve as a predictor for BMD changes, in order to assess the future fracture risk of an individual subject. At the spine, our study showed a positive and significant correlation between baseline biochemical markers of bone formation (BALP, BGP) and the 3-years changes in BMD. U-CTX was negatively correlated with 3-years changes in spinal BMD. When considering 3-months changes in BTM, only BALP and BGP were predictive of BMD changes. The absence of a clear association, also found in other [7,13], but not all studies [1,4], could be explained by an absence of bone loss at the lumbar spine in our elderly population (/ 2.16% in our study). This could be related to the intake of 500 mg elemental calcium daily or by the high incidence of spinal osteoarthritis, osteophytes, sclerosis or aortic calcification in this age range. In this view, the standard anterio-posterior measurement of vertebral BMD may not be an accurate reflection of vertebral skeletal integrity in this age group [14,15]. While we did not assess femoral neck BMD in the present study, previous publication that showed weak [4,16 /18] or no significant relationship [7,13,19] between BTM at baseline and the 3years changes in femoral neck BMD. Nevertheless, two of these studies [13,19] found a better correlation with total hip. The lower correlation at the femoral neck may be at least partly due to the poorer precision of this BMD measurement [20]. However, results with spine and hip are not always consistent. In fact, biochemical markers reflect the whole body rate of bone formation and resorption and are likely to reflect changes in the number of bone remodelling sites. Therefore, they may provide a more representative index of the overall skeletal bone loss than would be obtained by measuring the rates of change in BMD at specific skeletal sites containing different ratios of cancel-
lous to cortical components with different metabolic rates [21]. Bone mass is important to assess because of its high relation with fracture risk [22]. Nevertheless, BMD is only one of a number of fracture risk factors [23] and its assessment by DXA cannot systematically identify all individuals who will experience a fracture [2]. So, BTM could also be used to improve the evaluation of the risk of future fractures. In our population of relatively old women with at least one prevalent vertebral fracture, one bone formation marker (BGP) and one bone resorption marker (U-CTX) were reported to predict new vertebral deformities over 3 years. In the literature, it remains unclear whether bone formation marker levels are related to fracture risk. Some studies found a significant relation with spine fractures [24], or all site fractures [1,5,24] but other investigations failed to do so [6,10]. In our study, we showed that patients with the highest decrease in BGP over 3 months had a decreased risk of having new vertebral deformities over a period of 3 years. We also found a significant relationship between the bone resorption marker U-CTX and the occurrence of new vertebral deformities. In other studies, vertebral and non-vertebral fractures were associated with high baseline or high changes in UCTX [5,6,24] or in other bone resorption markers [1,5,6]. In the EPIDOS [6] and OFELY [5] study, the RR of fracture associated with an increase in BTMs is consistently around 2. Moreover, bone resorption markers predict osteoporotic fractures independently of BMD measured at several sites. However, in order to increase the predictive power of BTM to predict osteoporotic fractures, they should be analysed separately for the hip, vertebral and other sites, because of the heterogeneity in pathogenesis and potentially in underlying bone turnover abnormalities [3]. Nevertheless, in the light of our results, we support the idea that an increase in bone remodelling, assessed by biochemical markers, is associated with an increased risk of fractures. In conclusion, the measurement of BGP and U-CTX help at identifying patients with the highest risk to present, in the next 3 years, new vertebral deformities.
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
References [1] Melton LJ, III, Khosla S, Atkinson EJ, O’Fallon WM, Riggs BL. Relationship of bone turnover to bone density and fractures. J Bone Miner Res 1997;12:1083 /91. [2] Marshall D, Johnell O, Wedel H. Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. Br Med J 1996;312:1254 /9. [3] Delmas PD. The role of markers of bone turnover in the assessment of fracture risk in postmenopausal women, Osteoporesis Int. 1998;8(Suppl. 1):S32 /6. [4] Garnero P, Sornay-Rendu E, Chapuy M-C, Delmas PD. Increased bone turnover in late postmenopausal women is a major determinant of osteoporosis. J Bone Miner Res 1996;11:337 /49. [5] Garnero P, Sornay-Rendu E, Claustrat B, Delmas PD. Biochemical markers of bone turnover, endogenous hormones and risk of fracture in postmenopausal women: the OFELY study, J Bone Miner Res 2000;15:1526 /36. [6] Garnero P, Hausherr E, Chapuy M-C, Marcelli C, Grandjean H, Muller C, et al. Markers of bone resoprtion predict hip fracture in elderly women: the EPIDOS prospective study. J Bone Miner Res 1996;11:1531 /8. [7] Keen RW, Nguyen T, Sobnack R, Perry LA, Thompson PW, Spector TD. Can biochemical markers predict bone loss at the hip and spine: a 4-years prospective study of 141 early postmenopausal women. Osteoporesis Int 1996;6:399 /406. [8] Reginster JY, Christiansen C, Roux C, Fechtenbaum J, Rouillon A, Tou KP. Intermittent cyclic tiludronate in the treatment of osteoporosis. Osteoporosis Int 2001;12:169 / 77. [9] Genant HK, Wu CY, Van Kuijk C, Nevitt MC. Vertebral fracture assessment using a semi-quantitative technique. J Bone Miner Res 1993;8:984 /96. [10] Tromp AM, Ooms ME, Popp-Snijders C, Roos JC, Lips P. Predictors of fractures in elderly women. Osteoporosis Int 2000;11:134 /40. [11] Riis BJ, Hansen MA, Jensen AM, Overgaard K, Christiansen C. Low bone mass and fast rate of bone loss at menopause: equal risk factors for future fracture: a 15years follow-up study. Bone 1996;19:9 /12. [12] Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K, et al. Bone density at various sites for prediction of hip fractures. Lancet 1993;341:72 /5. [13] Dresner-Pollak R, Parker RA, Poku M, Thompson J, Seibel MJ, Greespan SL. Biochemical markers of bone
[14]
[15]
[16]
[17]
[18]
[19]
[20] [21]
[22]
[23]
[24]
265
turnover reflect femoral bone loss in elderly women. Calcif Tissue Int 1996;59:328 /33. Steiger P, Cummings SR, Black DM, Spencer NE, Genant HK. Age-related decrements in bone mineral density in women over 65. J Bone Miner Res 1992;7:625 /32. Dawson-Hughes B, Dallal GE. Effect of radiographic abnormalities on rate of bone loss from the spine. Calcif Tissue Int 1990;46:280 /1. Marcus R, Holloway L, Wells B, Greendale G, James MK, Wasilauskas C, et al. The relationship of biochemical markers of bone turnover to bone density changes in postmenopausal women: results from the postmenopausal estrogen/progestin interventions (PEPI) trial. J Bone Miner Res 1999;14:1583 /95. Melton LJ, III, Atkinson EJ, O’Connor MK, O’Fallon WM, Riggs BL. Determinants of bone loss from the femoral neck in women of different ages. J Bone Miner Res 2000;15:24 /31. Rosen CJ, Chesnut CH, III, Mallinak NJS. The predictive value of biochemical markers of bone turnover for bone mineral density in early postmenopausal women treated with hormone replacement or calcium supplementation. J Clin Endocrinol Metab 1997;82:1904 /10. The Study of Osteoporotic Fractures Research Group, Bauer DC, Sklarin PM, Stone KL, Black DM, Nevitt MC, Ensrud KE, et al. Biochemical markers of bone turnover and prediction of hip bone loss in older women: the study of osteoporotic fractures. J Bone Miner Res 1999;14:1404 / 10. Reginster JY. Osteoporose postmenopausique, Traitement prophylactique, 1993, Paris, Masson. Delmas PD, Eastell R, Garnero P, Seibel MJ, Stepan J, for the Committee of Scientific Advisor of the International Osteoporosis Foundation. The use of biochemical markers of bone turnover in osteoporosis, Osteoporosis Int. 2000;11(Suppl. 6): 2 /17. Jergas M, Genant HK. Contributions of bone mass measurements by densitometry in the definition and diagnosis of osteoporosis. In: Meunier PJ, editor. Osteoporosis: diagnosis and management. Martin Dunitz, 1998. Cummings SR, Nevitt MC, Browner WS, Stone D, Fox KM, Ensrud KE, et al. Risk factors for hip fracture in white women. New Engl J Med 1995;332:767 /73. Ross PD, Kress BC, Parson RE, Wasnich RD, Armour KA, Mizrahi IA. Serum bone alkaline phosphatase and calcaneus bone density predict fractures: a prospective study. Osteoporosis Int 2000;11:76 /82.
Interest of biochemical markers of bone turnover for long-term prediction of new vertebral fracture in postmenopausal osteoporotic women Olivier Bruyere a,b,c, Julien Collette c, Pierre Delmas d, Alain Rouillon e, Christian Roux f, Laurence Seidel g, Florent Richy a,b, Jean-Yves Reginster a,b,c,h,* a
WHO Collaborating Center for Public Health Aspect of Osteoarticular Disorders, Liege, Belgium b Department of Public Health and Epidemiology, University of Liege, Liege, Belgium c Bone and Cartilage Research Unit, Policlinique BRULL, University of Liege, 45 Quai G. Kurth, 4020 Liege, Belgium d Service de Rhumathologie, Hoˆpital Edouard Herriot, Lyon, France e Sanofi Recherche, Gentilly, France f Service de Rhumatologie, Hoˆpital Cochin, Universite´ Rene´ Descartes, Paris, France g Biostatistical Department, University of Liege, Liege, Belgium h Georgetown University Medical Center, Washington, DC, USA Received 19 March 2002; received in revised form 21 November 2002; accepted 7 January 2003
Abstract Objective: To analyse the interest of baseline levels and short-term (3-months) changes in serum osteocalcin (BGP), serum bone-specific alkaline phosphatase (BALP) and urinary C-telopeptide of type I collagen/creatinine ratio (UCTX) to predict 3-years changes in bone mineral density (BMD) and spinal deformity index (SDI) in postmenopausal osteoporotic women. Methods: Data were derived from a cohort of 603 osteoporotic women corresponding to the placebo arm of a 3-years prospective, double-blind study. Results: Baseline values of BALP, BGP and U-CTX were negatively and significantly correlated with baseline spinal BMD. Significant correlations were also observed between the changes in BMD observed after 36 months at the spine and baseline BALP (r /0.20, P /0.0001), BGP (r /0.09, P /0.05) and U-CTX (r / /0.11, P /0.02). At 3 years, 71 women (15.9%) showed an increase in their SDI, corresponding to the occurrence of at least one new vertebral deformity. Baseline values of the four bone turnover markers (BTM) were not significantly related to the occurrence of new vertebral deformities. However, when considering the changes in the BTM observed after 3-months of follow-up, BGP (P /0.003) and U-CTX (P /0.047) were identified as significant predictors of an increase of SDI. The associated odds ratios (95% confidence interval (CI)) were 10.922 (2.218 /53.78) for unit changes of log BGP and 1.369 (1.003 /1.867) for unit changes of log U-CTX. The relative risk (RR) (IC 95%) of having a new vertebral fracture over 36 months was 0.31 (0.15 /0.65) when being in the lowest quartile of 3-months changes in BGP as compared with the highest. Conclusion: We conclude that two sequential
* Corresponding author. Tel.: /32-4-270-3257; fax: /32-4-270-3253. E-mail address: [email protected] (J.-Y. Reginster). 0378-5122/03/$ - see front matter # 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0378-5122(03)00042-2
260
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
measurements of BGP and U-CTX performed at 3-months intervals could be of interest to identify postmenopausal osteoporotic women with the highest risk to present new vertebral deformities. # 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Osteoporosis; Bone turnover markers; Vertebral fracture; Bone mineral density
1. Introduction Bone mineral density (BMD), can now be easily assessed by means of dual X-ray absorptiomertry (DXA). Low BMD is widely considered as the best predictor of later developing an osteoporotic fracture [1,2]. However, due to the variability of the DXA technique and the relatively slow rate of BMD changes, clinically relevant variations in BMD usually request a gap of 2 /3-years between two measurements. Conversely, biochemical markers of bone remodelling provide a dynamic picture of the actual rate of bone turnover, as opposed to the static approach provided by the BMD measurement [3]. They are widely used in daily practice and clinical research. Previous studies have extensively documented their ability to predict future fracture risk in healthy postmenopausal women, to a similar extend than measurement of BMD do [4 /7]. However, relatively few prospective, long-term studies, have evaluated their interest to predict the risk of women with prevalent vertebral fractures to develop incident vertebral deformities. Based on a 3-years extensive follow up of a large cohort of untreated osteoporotic women, we evaluated the ability of three currently marketed biochemical markers of bone turnover to predict changes in lumbar and hip BMD and to identify women who are more prone to experience new vertebral deformities.
the effect of intermittent cyclical tiludronate, a third generation bisphosphonate, in the treatment of established postmenopausal osteoporosis [8]. The primary inclusion criteria were the presence of at least two ‘mild’ or one ‘moderate’ but no more than five vertebral fractures of any severity and a low BMD (T -score B/ /1.7) of at least two assessable lumbar vertebrae (L2, L3, L4) measured with DXA. Patients were advised to include at least 700 mg of calcium in their daily diet. Nevertheless, a daily calcium supplementation (Calcium Sandoz8, Sandoz, Basle, Switzerland) equivalent to 500 mg elemental calcium was provided everyday except the first 7 days of each month. 2.2. Methods
2.1. Population
2.2.1. Bone densitometry BMD (g/cm2) of the antero-posterior spine and the femoral neck was measured at baseline and after 3 years, by dual energy X-ray densitometry using Hologic (Waltham, MA) or Lunar (Lunar Corporation, Madison, WI) instruments. BMD of each subject was measured by the same instrument throughout the entire study. Quality control of densitometric equipment and data was dual-faceted. Firstly, there was daily a quality control of the Hologic phantom at each center and in case of a significant drift in the QC observation, a correction factor was applied. Secondly, each scanner was cross-calibrated with two transportable phantoms at baseline and every 6 months, during visits performed by the Center for Clinical and Basic Research, (CCBR) (Ballerup, Denmark), Quality Assurance Center team.
The population was constituted by a cohort of 603 Caucasian (n /602) or Asian (n /1) women, aged between 50 and 80 years, corresponding to the placebo arm of a double blind, prospective, randomised, placebo-controlled study evaluating
2.2.2. Radiography Lateral radiographs of both thoracic and lumbar spine were obtained at baseline and 36 months. Anterio-posterior views were added at baseline. The radiographs were obtained according to a
2. Patients and methods
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
standardised protocol, with the beam centred on T7 and L3 for thoracic and lumbar spine, respectively. The interpretation of all radiographs was performed by an independent consultant (Christian Roux). Both quantitative and semi-quantitative assessments of vertebral deformities were performed with all available radiographs. A semi quantitative vertebral deformity score was assigned to each individual vertebra from T4 to L4 according to the method previously published by Genant and colleagues [9]. Score ‘0’ for normal, non-fractured vertebra; score ‘1’ for a mild deformity with a 20/25% reduction in anterior, middle or posterior height (or all three); score ‘2’ for a moderate deformity with a /25 /40% reduction in any height; score ‘3’ for a severe deformity with a reduction of /40% in any height. A new fracture in a particular vertebra is defined as the progression from a grade 0 to a grade of 1 or higher. The spinal deformity index (SDI) for a subject was defined as the sum of the individual vertebral deformities scores, divided by the number of vertebrae evaluated. 2.2.3. Biochemical markers Blood and urine samples were obtained after an overnight fasting at baseline and after 3 and 36 months. Three BTM were taken into account: / Serum Osteocalcin was determined by a commercially available solid-phase ‘sandwich’ immunoradiometric assay (ELSA-Osteoporosis, CIS-Bio International, France). / Urine C-telopeptide of type I collagen was measured by a commercially available enzymelinked immuno-sorbent assay (Crosslaps8 Tm EIA, CIS-Bio International, France). / Serum bone specific alkaline phosphatase determinations were performed using a commercially available immunoradiometric assay (‘Tandem’-R Ostase Beckmann Coulter, USA). 2.2.4. Statistical analysis Quantitative variables were expressed as mean 9/S.D. Relationships between each of the bone turnover markers (BTM) and the baseline values or the long-term (36-months) changes in BMD (spine or femoral neck) were assessed by the
261
Pearson correlation coefficient, by multiple and stepwise regression. We also assessed 3-years changes in BMD in each quartile of baseline values of biochemical markers. Differences between quartiles were assessed by ANOVA. The predictive value of the BTM as explanatory variables for an increase of SDI after 36 months was assessed by a stepwise logistic regression. We also looked at the RR of having a new vertebral fracture over 36 months in the lowest quartile as compared with the highest quartile of baseline and 3-months change’s values of biochemical markers. To normalise the distribution of BTM, we performed a logarithmic transformation of the BTMs values. The results were considered significant at the 5% level (P B/0.05). Statistical calculations were carried out with the SAS software version 6.12 (SAS Institute, Cary, NC, USA).
3. Results Baseline demographics of the population are summarised in Table 1. Baseline values of serum bone-specific alkaline phosphatase (BALP), serum osteocalcin (BGP), and urinary C-telopeptide of type I collagen/creatinine ratio (U-CTX) were negatively and significantly correlated with baseline spinal BMD (r / /0.17, P B/0.0001 for BALP, r / /0.02, P B/0.0001 for BGP, r / /0.12, P /0.005 for U-CTX). However, when performing a stepwise regression, BGP remained the only marker to be significantly correlated to baseline spinal BMD (P /0.0002). For the feTable 1 Baseline demographics (mean S.D.) Variables
Mean (S.D.)
Number of patients Age (year) BMI (kg/m2) Years since menopause (year) Femoral neck BMD (g/cm2) Spine BMD (g/cm2) BGP (ng/ml) BALP (ng/ml) U-CTX (mg/mmol creatinine)
603 67.4 (6.8) 25.3 (3.8) 19.5 (8.0) 0.596 (0.087) 0.731 (0.094) 26.70 (8.28) 13.62 (5.82) 270.96 (230.1)
262
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
moral baseline BMD, negative and significant correlations were found for BALP (r / /0.09, P /0.029) and BGP (r / /0.16, P B/0.0001). The same markers remained significant after a stepwise regression (P /0.0041 and P B/0.0001 for BALP and BGP, respectively). Significant correlations were observed between the changes in spinal BMD observed after 36 months and baseline BALP (r /0.20, P B/0.0001) (Fig. 1), BGP (r /0.09, P /0.05) and U-CTX (r / /0.11, P /0.02) (Fig. 2). After stepwise regression, only BALP and U-CTX remained significant (P /0.0001 and P /0.001, respectively). When dividing our population into quartiles of baseline BALP values, patients in the lowest quartile ( B/9.7 ng/ml) experienced, after 3 years, a mean (S.D.) increase of 0.004 (0.04) g/cm2 at the spine compared with 0.031 (0.052) g/cm2 in the highest quartile (/16.2 ng/ml) (P B/0.0001). For U-
CTX, increase was 0.018 (0.053) and 0.004 (0.046) g/cm2 in the lowest (B/121.1 mg/mmol) and in the highest (/340.7 mg/mmol) quartiles, respectively (P /0.03). At the femur, no significant relationships were observed. After 3-years of follow-up, 71 women (15.9%) showed an increase in their SDI, corresponding to the occurrence of at least one new vertebral deformity. In this subgroup of patients, the baseline spine BMD values was 0.719 (0.111) compared with 0.744 (0.087) g/cm2 in the non-increased SDI subgroup (P /0.03). At the femoral neck, the BMD was not statistically significant between the two groups (P /0.71). Regarding the BTM at baseline in the increased SDI group, BALP value (14.71 (8.44) ng/ml) was statistically different compared with the non increased SDI group (13.09 (5.29) ng/ml) (P /0.03). The two other markers were not statistically different between
Fig. 1. Correlation between the changes in spinal BMD observed after 36 months and baseline BALP.
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
263
Fig. 2. Correlation between the changes in spinal BMD observed after 36 months and baseline U-CTX.
the two groups (P /0.27 for BGP and P /0.32 for U-CTX). Using a stepwise analysis regression, the baseline values of the three BTM were not significantly related to an increase in the SDI (P /0.2 for BALP, P /0.14 for BGP, P /0.96 for U-CTX). When considering the changes in the BTM observed after 3-months of follow-up, BGP (P / 0.003) and U-CTX (P /0.047) but not BALP were identified as significant predictors of an increase of SDI after 3 years. The associated odds ratios (95% confidence interval (CI)) were 10.92 (2.22 /53.78) for unit changes of log BGP and 1.37 (1.003 /1.87) for unit changes of log U-CTX. The RR (95% CI) of having a new vertebral fracture over 36 months when being in the lowest compared with the highest quartile was 0.66 (0.37/1.17) for BALP, 1.12 (0.62/2.04) for BGP, 0.88 (0.50/1.53) for UCTX, 0.89 (0.51/1.54) for 3-months changes in BALP, 0.60 (0.34 /1.09) for 3-month changes in U-
CTX and 0.31 (0.15/0.65) for 3-month changes in BGP.
4. Discussion As previously reported [6,10], baseline biochemical markers of bone formation (BALP, BGP) and bone resorption (U-CTX) were negatively correlated with bone density at the spine and at the hip. High levels of bone remodelling are associated with a higher risk of low BMD. However, because of the low magnitude of the explained variance (5.7%), BTM offer little practical information for estimating BMD levels in an individual woman and can not be used as a surrogate measure to predict bone mass. Since a rapid decrease in bone mass is a major risk factor for fracture, in addition to baseline BMD [11], it is important to identify patients whose BMD will show the highest de-
264
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
crease over time. While a single BMD measure by DXA is considered accurate and precise [12], the BMD variability for repeated measurements is in the range of 3 /4%, hence within the same magnitude as the physiological changes expected over 2/ 4 years [3]. Therefore, a duration of 2/3 years is actually needed before an observed change in BMD can be regarded as clinically relevant. One of the goals of BTM could thus be to serve as a predictor for BMD changes, in order to assess the future fracture risk of an individual subject. At the spine, our study showed a positive and significant correlation between baseline biochemical markers of bone formation (BALP, BGP) and the 3-years changes in BMD. U-CTX was negatively correlated with 3-years changes in spinal BMD. When considering 3-months changes in BTM, only BALP and BGP were predictive of BMD changes. The absence of a clear association, also found in other [7,13], but not all studies [1,4], could be explained by an absence of bone loss at the lumbar spine in our elderly population (/ 2.16% in our study). This could be related to the intake of 500 mg elemental calcium daily or by the high incidence of spinal osteoarthritis, osteophytes, sclerosis or aortic calcification in this age range. In this view, the standard anterio-posterior measurement of vertebral BMD may not be an accurate reflection of vertebral skeletal integrity in this age group [14,15]. While we did not assess femoral neck BMD in the present study, previous publication that showed weak [4,16 /18] or no significant relationship [7,13,19] between BTM at baseline and the 3years changes in femoral neck BMD. Nevertheless, two of these studies [13,19] found a better correlation with total hip. The lower correlation at the femoral neck may be at least partly due to the poorer precision of this BMD measurement [20]. However, results with spine and hip are not always consistent. In fact, biochemical markers reflect the whole body rate of bone formation and resorption and are likely to reflect changes in the number of bone remodelling sites. Therefore, they may provide a more representative index of the overall skeletal bone loss than would be obtained by measuring the rates of change in BMD at specific skeletal sites containing different ratios of cancel-
lous to cortical components with different metabolic rates [21]. Bone mass is important to assess because of its high relation with fracture risk [22]. Nevertheless, BMD is only one of a number of fracture risk factors [23] and its assessment by DXA cannot systematically identify all individuals who will experience a fracture [2]. So, BTM could also be used to improve the evaluation of the risk of future fractures. In our population of relatively old women with at least one prevalent vertebral fracture, one bone formation marker (BGP) and one bone resorption marker (U-CTX) were reported to predict new vertebral deformities over 3 years. In the literature, it remains unclear whether bone formation marker levels are related to fracture risk. Some studies found a significant relation with spine fractures [24], or all site fractures [1,5,24] but other investigations failed to do so [6,10]. In our study, we showed that patients with the highest decrease in BGP over 3 months had a decreased risk of having new vertebral deformities over a period of 3 years. We also found a significant relationship between the bone resorption marker U-CTX and the occurrence of new vertebral deformities. In other studies, vertebral and non-vertebral fractures were associated with high baseline or high changes in UCTX [5,6,24] or in other bone resorption markers [1,5,6]. In the EPIDOS [6] and OFELY [5] study, the RR of fracture associated with an increase in BTMs is consistently around 2. Moreover, bone resorption markers predict osteoporotic fractures independently of BMD measured at several sites. However, in order to increase the predictive power of BTM to predict osteoporotic fractures, they should be analysed separately for the hip, vertebral and other sites, because of the heterogeneity in pathogenesis and potentially in underlying bone turnover abnormalities [3]. Nevertheless, in the light of our results, we support the idea that an increase in bone remodelling, assessed by biochemical markers, is associated with an increased risk of fractures. In conclusion, the measurement of BGP and U-CTX help at identifying patients with the highest risk to present, in the next 3 years, new vertebral deformities.
O. Bruyere et al. / Maturitas 44 (2003) 259 /265
References [1] Melton LJ, III, Khosla S, Atkinson EJ, O’Fallon WM, Riggs BL. Relationship of bone turnover to bone density and fractures. J Bone Miner Res 1997;12:1083 /91. [2] Marshall D, Johnell O, Wedel H. Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. Br Med J 1996;312:1254 /9. [3] Delmas PD. The role of markers of bone turnover in the assessment of fracture risk in postmenopausal women, Osteoporesis Int. 1998;8(Suppl. 1):S32 /6. [4] Garnero P, Sornay-Rendu E, Chapuy M-C, Delmas PD. Increased bone turnover in late postmenopausal women is a major determinant of osteoporosis. J Bone Miner Res 1996;11:337 /49. [5] Garnero P, Sornay-Rendu E, Claustrat B, Delmas PD. Biochemical markers of bone turnover, endogenous hormones and risk of fracture in postmenopausal women: the OFELY study, J Bone Miner Res 2000;15:1526 /36. [6] Garnero P, Hausherr E, Chapuy M-C, Marcelli C, Grandjean H, Muller C, et al. Markers of bone resoprtion predict hip fracture in elderly women: the EPIDOS prospective study. J Bone Miner Res 1996;11:1531 /8. [7] Keen RW, Nguyen T, Sobnack R, Perry LA, Thompson PW, Spector TD. Can biochemical markers predict bone loss at the hip and spine: a 4-years prospective study of 141 early postmenopausal women. Osteoporesis Int 1996;6:399 /406. [8] Reginster JY, Christiansen C, Roux C, Fechtenbaum J, Rouillon A, Tou KP. Intermittent cyclic tiludronate in the treatment of osteoporosis. Osteoporosis Int 2001;12:169 / 77. [9] Genant HK, Wu CY, Van Kuijk C, Nevitt MC. Vertebral fracture assessment using a semi-quantitative technique. J Bone Miner Res 1993;8:984 /96. [10] Tromp AM, Ooms ME, Popp-Snijders C, Roos JC, Lips P. Predictors of fractures in elderly women. Osteoporosis Int 2000;11:134 /40. [11] Riis BJ, Hansen MA, Jensen AM, Overgaard K, Christiansen C. Low bone mass and fast rate of bone loss at menopause: equal risk factors for future fracture: a 15years follow-up study. Bone 1996;19:9 /12. [12] Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K, et al. Bone density at various sites for prediction of hip fractures. Lancet 1993;341:72 /5. [13] Dresner-Pollak R, Parker RA, Poku M, Thompson J, Seibel MJ, Greespan SL. Biochemical markers of bone
[14]
[15]
[16]
[17]
[18]
[19]
[20] [21]
[22]
[23]
[24]
265
turnover reflect femoral bone loss in elderly women. Calcif Tissue Int 1996;59:328 /33. Steiger P, Cummings SR, Black DM, Spencer NE, Genant HK. Age-related decrements in bone mineral density in women over 65. J Bone Miner Res 1992;7:625 /32. Dawson-Hughes B, Dallal GE. Effect of radiographic abnormalities on rate of bone loss from the spine. Calcif Tissue Int 1990;46:280 /1. Marcus R, Holloway L, Wells B, Greendale G, James MK, Wasilauskas C, et al. The relationship of biochemical markers of bone turnover to bone density changes in postmenopausal women: results from the postmenopausal estrogen/progestin interventions (PEPI) trial. J Bone Miner Res 1999;14:1583 /95. Melton LJ, III, Atkinson EJ, O’Connor MK, O’Fallon WM, Riggs BL. Determinants of bone loss from the femoral neck in women of different ages. J Bone Miner Res 2000;15:24 /31. Rosen CJ, Chesnut CH, III, Mallinak NJS. The predictive value of biochemical markers of bone turnover for bone mineral density in early postmenopausal women treated with hormone replacement or calcium supplementation. J Clin Endocrinol Metab 1997;82:1904 /10. The Study of Osteoporotic Fractures Research Group, Bauer DC, Sklarin PM, Stone KL, Black DM, Nevitt MC, Ensrud KE, et al. Biochemical markers of bone turnover and prediction of hip bone loss in older women: the study of osteoporotic fractures. J Bone Miner Res 1999;14:1404 / 10. Reginster JY. Osteoporose postmenopausique, Traitement prophylactique, 1993, Paris, Masson. Delmas PD, Eastell R, Garnero P, Seibel MJ, Stepan J, for the Committee of Scientific Advisor of the International Osteoporosis Foundation. The use of biochemical markers of bone turnover in osteoporosis, Osteoporosis Int. 2000;11(Suppl. 6): 2 /17. Jergas M, Genant HK. Contributions of bone mass measurements by densitometry in the definition and diagnosis of osteoporosis. In: Meunier PJ, editor. Osteoporosis: diagnosis and management. Martin Dunitz, 1998. Cummings SR, Nevitt MC, Browner WS, Stone D, Fox KM, Ensrud KE, et al. Risk factors for hip fracture in white women. New Engl J Med 1995;332:767 /73. Ross PD, Kress BC, Parson RE, Wasnich RD, Armour KA, Mizrahi IA. Serum bone alkaline phosphatase and calcaneus bone density predict fractures: a prospective study. Osteoporosis Int 2000;11:76 /82.