Few data concern the relationship between bone turnover and microarchitecture in men. We investigated the association between levels of biochemical markers of bone turnover (BTM) and bone microarchitecture in 1149 men aged 19 to 85 years. Bone microarchitecture was assessed by high-resolution peripheral quantitative computed tomography at the distal radius and tibia. Bone formation was assessed by serum osteocalcin, bone alkaline phosphatase, and N-terminal extension propeptide of type I collagen. Bone resorption was assessed by serum C-terminal telopeptide of type I collagen and urinary excretion of total deoxypyridinoline. BTM levels were high in young men and decreased until age 50 years. Urinary deoxypyridinoline (DPD) increased after age 70 years, whereas other BTMs remained stable. Before 50 years of age, only cortical volumetric bone mineral density (Dcort) correlated negatively with BTM levels. Between 50 and 70 years of age, Dcort and some microarchitectural parameters correlated significantly with BTM at the radius and tibia. After 70 years of age, higher BTM levels were associated with lower cortical thickness and Dcort at both the skeletal sites. At the distal radius, men in the highest BTM quartile had lower trabecular density, number (Tb.N), and thickness (Tb.Th) and more heterogeneous trabecular distribution compared with men in the lower quartiles. At the distal tibia, higher BTM levels were associated with lower Tb.N and Tb.Th in the central but not subendocortical area. Thus, in men, bone microarchitecture depends weakly on the current bone turnover rate until age 70. Thereafter, bone turnover seems to be a significant determinant of bone microarchitecture. © 2010 American Society for Bone and Mineral Research.
Osteoporosis in men is a major public health problem because 25% to 30% of osteoporotic fractures occur in men.1 However, bone mineral density (BMD) measured by dual-energy X-ray absorptiometry (DXA) does not predict fractures satisfactorily. A gender-specific BMD T-score of less than −2.5 identifies 40% to 50% of future osteoporotic fractures in women and about 20% in men.2–5 Measurement of cortical bone volume and trabecular volumetric BMD (vBMD) at the hip does not improve the prediction of fractures in men.6
Osteoporosis is characterized by low bone mass and microarchitectural deterioration.7 Microarchitectural deterioration is associated with the presence of the fragility fractures.8–10 Thus in vivo assessment of bone microarchitecture may improve fracture prediction in men. In men, an age-related decrease in trabecular number and thickness, as well as in cortical thickness and vBMD, may increase bone fragility.11 However, pathophysiologic mechanisms of the age-related deterioration of bone microarchitecture are not fully elucidated.
Levels of bone turnover markers (BTMs) are highest in young men and then decrease.12, 13 In older men, bone resorption increases slightly, whereas serum levels of bone-formation markers remain stable or even decrease slightly in very old men.13–15 This imbalance between increased bone resorption and relatively stable bone formation may underlie bone loss in older men, in line with the negative correlation between BTMs and BMD in this group.12, 13, 16, 17 In older men in the highest quartiles of BTM levels, BMD was 1.8% to 12.5% lower than in men in the lowest quartile dependent on the skeletal site.13 In men aged 40 years and over, elevated BTM levels correlated weakly negatively with trabecular volume, number, and thickness after adjustment for age.18 However, the microarchitectural basis underlying the association between BMD and BTM levels is not fully understood.
Even less is known about the association between bone turnover rate and bone mass in young adults. Men experience 42% of their total lifetime trabecular bone loss and 15% of cortical bone loss before age 50 years.19 However, in young men, correlation of the BTM levels with BMD and bone microarchitecture is weak or nonsignificant.13, 18
Thus, given paucity of data on the association between bone turnover rate and bone microarchitecture in men and their potential importance for therapeutic intervention, the aim of our study is to analyze the relationship between the BTM levels and bone microarchitecture at the distal radius and distal tibia assessed by high-resolution peripheral computed tomography (HR-pQCT) in a cohort of 1149 men aged 20 to 85 years.
Materials and Methods
The STRucture of the Aging Men's BOnes (STRAMBO) study is a single-center prospective cohort study of the skeletal fragility and its determinants in men. It is a collaboration between INSERM (National Institute of Health and Medical Research) and MTRL (Mutuelle de la Région Lyonnaise). Participants were recruited in 2006–2008 from the MTRL rolls in Lyon. Letters inviting participation into the study were sent to a randomly selected sample of men aged 20 to 85 years living in Lyon and its vicinity. The study obtained authorization of the local ethics committee and was performed in agreement with the Helsinki Declarations of 1975 and 1983. Eleven hundred and sixty-nine men agreed to participate and provided informed consent. This analysis was carried out in 1149 men who had BMD measurement, bone microarchitecture evaluation, and collection of blood and urinary samples at baseline (2006–2008). All participants replied to an epidemiologic questionnaire that was administered by an interviewer and covered lifestyle factors and health status. All men able to give informed consent, to answer the questionnaire, and to participate in the diagnostic exams were included. No specific exclusion criteria were used.
Nonfasting serum and urine were collected at 1:00 p.m. and stored at −80°C until assayed. Serum total osteocalcin (OC), N-terminal extension propeptide of type I collagen (PINP), and β-isomerized C-terminal telopeptide of type I collagen (CTX) were measured by human-specific two-site immunochemiluminescence assay (ELECSYS, Roche, Indianapolis, IN, USA). For OC, the detection limit is 0.5 ng/mL. The interassay coefficient of variation (CV) is 11.1% for 21.4 ng/mL, 11.9% for 27.1 ng/mL, 12.6% for 109.1 ng/mL, and 14.6% for 220.5 ng/mL. For PINP, the detection limit is 5 ng/mL. The interassay CV is 4.7% for 41 ng/mL, 4.5% for 74.7 ng/mL, 5.1% for 381.5 ng/mL, and 5.9% for 766.7 ng/mL. For CTX, the detection limit is 0.01 ng/mL. the interassay CV is 3.7% for 0.3 ng/mL, 2.3% for 0.36 ng/mL, 3.8% for 0.69 ng/mL, and 5% for 2.8 ng/mL. Bone-specific alkaline phosphatase (BAP) was measured by enzymatic immunoassay using a monoclonal antibody against BAP (MetraBAP, Quidel, San Diego, CA, USA). The detection limit is 0.7 µmol/L. The interassay CV is 5.9% for 7.5 µmol/L, 5% for 15.7 µmol/L, 3.4% for 20.7 µmol/L, 5.6% for 26.1 µmol/L, and 2% for 49.4 µmol/L. Urinary deoxypyridinoline (DPD) was measured after acid hydrolysis by ELISA (Metra Total DPD, Quidel). The detection limit is 0.5 nmol/L. The interassay CV is 10.2% for 1.2 nmol/L, 12.8% for 2.7 nmol/L, 14.2% for 4.3 nmol/L, 4.4% for 9.7 nmol/L, and 11.3% for 19.2 nmol/L.
BMD and bone microarchitecture measurement
Volumetric BMD (vBMD) and microarchitecture were assessed at the nondominant distal radius and right distal tibia by HR-pQCT (XtremeCT, Scanco Medical, Brüttisellen, Switzerland). The arm or leg of the patient was immobilized in a carbon fiber shell. An anteroposterior scout view was used to define the measured volume of interest (VOI).20 At each site, a stack of 110 parallel CT slices with an isotropic voxel size of 82 µm was obtained, thus delivering a 3D representation of approximately 9 mm in the axial direction. The most distal CT slice was placed 9.5 and 22.5 mm proximal to the endplate of the radius and tibia, respectively. Quality control was performed by daily scans of a phantom containing rods of hydroxyapatite (densities of 0, 100, 200, 400, and 800 mg/cm3) embedded in a soft-tissue equivalent resin (QRM, Moehrendorf, Germany). The CV for the densities varied from 0.7% to 1.5%.
The VOI was separated into a cortical and trabecular region using a threshold-based algorithm. This threshold was set to one-third the cortical vBMD (Dcort). Cortical thickness (C.Th) was defined as the mean cortical volume divided by the outer bone surface. Trabecular vBMD (Dtrab) in milligrams of hydroxyapatite (HA) per cubic centimeter was computed as the average vBMD within the trabecular VOI. Trabecular bone volume (BV) fraction [BV/trabecular volume (TV), %] was derived from trabecular vBMD assuming fully mineralized bone to have a mineral density of 1.2 g HA/cm3, that is, BV/TV (%) = 100 × [Dtrab (mg HA/cm3)/1.2 g HA/cm3]. Trabecular elements were identified by the midaxis transformation method, and the distance between them was assessed three-dimensionally by the distance-transform method. Trabecular number (Tb.N, mm−1) was defined as the inverse of the mean spacing of the midaxes. Trabecular thickness (Tb.Th, µm) and separation (Tb.Sp, µm) were derived from BV/TV and Tb.N, that is, Tb.Th = (BV/TV)/Tb.N and Tb.Sp = (1 – BV/TV)/Tb.N. Intraindividual distribution of separation (Tb.NSD, µm) was quantified by standard deviation of Tb.Sp, a parameter reflecting the heterogeneity of the trabecular network.
The CV values for parameters of radius and tibia, respectively, were as follows: total vBMD: 0.9% and 1.3%; Dcort: 0.7% and 0.9%; C.Th: 1.2% and 0.9%; Dtrab: 1.0% and 1.5%; Tb.N: 3.0% and 3.8%; Tb.Th: 3.2% and 4.4%; Tb.Sp: 2.8% and 4.3%, and Tb.NSD: 2.5% and 3.3%. The inner (central) and outer (subendocortical) areas were defined as follows: Trabecular contour was generated from the total contour by subtracting the cortical area. The remaining area was defined as the trabecular area (Tb.Area). This area is progressively peeled by one pixel on the entire contour, and the remaining area is measured. When it is below 60% of the initial area, the procedure is stopped. The remaining area in the center is defined as the inner area, and the peeled area is defined as the outer area.
Statistical analyses were performed using SPSS 15.0 software (SPSS, Chicago, IL, USA) and SAS 9.1 software (SAS, Cary, NC, USA). Data are presented as mean ± SD and as median and interquartile (IQ) range. Correlation was assessed by partial Pearson's correlation coefficient adjusted for age, weight, and height. Variables with non-Gaussian distribution were log-transformed. Age-related changes in BTM levels and microarchitectural parameters were modeled by PROC LOESS using the option Automatic Smoothing Parameter Selection. This method performs robust fitting in the presence of outliers in the data.
Association between BTM quartiles and bone microarchitectural parameters was assessed by analysis of covariance adjusted for age, weight, and height. Then we calculated the models in the three lower BTM quartiles. If the dependent variable differed significantly in three lower quartiles, we calculated trend across the four quartiles and difference between the adjusted mean in the fourth quartile minus adjusted mean in the first quartile, which was expressed in percentage and number of SDs. If the dependent variable did not differed in the three lower BTM quartiles, we calculated the difference between the adjusted mean in the highest quartile minus the adjusted mean in the collapsed three lower quartiles. The differences were expressed in percentage and number of SDs. Given multiple comparisons of the interdependent variables, p < .01 was considered significant. In order to check the validity of the selected thresholds, we carried out sensitivity analysis. We changed the thresholds by up to 5 years and repeated the same analyses of covariance. Given the nonhomogeneous distribution of men in age groups, we repeated the analyses in a randomly selected age-stratified sample of 338 men.
We show that early-afternoon nonfasting BTM levels are high in young men and decrease until age 50. Then bone-formation markers and serum CTX are stable, whereas DPD increases again after age 70. Before age 50, Dcort, but not other parameters, correlates negatively with BTM levels (adjusted for age, weight, and height). Between 50 and 70 years of age, BTM levels correlate with Dcort and more weakly with some other parameters. After age 70, the highest BTM levels are associated with lower Dcort and C.Th values at the distal radius and tibia. They are associated with lower Dtrab, Tb.N, and Tb.Th values in the inner region at both skeletal sites. In the outer region, this association is weak at the distal radius and nonsignificant at the distal tibia.
The age-related changes in BTM levels are similar to those found in studies where fasting blood samples for serum BTMs and first-morning or 24-hour urine analyses for urinary BTMs were used.13, 14, 16 Thus standardized collection of the nonfasting and urinary samples in the early-afternoon avoids circadian variability in BTM levels and can provide a valid alternative for the morning fasting samples in the assessment of bone turnover in men, provided that blood samples are collected under standardized comparable conditions.21, 22
The negative correlation between BTM levels and Dcort in men aged less than 50 years suggests an active bone turnover on the intracortical surface. Zebaze and colleagues showed that intracortical bone remodeling contributes substantially to bone turnover in women.23 Lower vBMD values in men with highest BTM levels reflects the higher number of bone-remodeling units, which may lead to higher cortical porosity, the higher fraction of young, partly mineralized bone around cortical canals, and the subendocortical layer of cortical bone. In young men, coexistence of the highest BTM levels and lowest cortical vBMD values may reflect the late phase of formation of peak cortical vBMD, whereas in men aged 50 to 70 years, this coexistence may reflect the early phase of bone loss.
Lack of correlation of BTMs with trabecular parameters before age 50 is astonishing because prospective and cross-sectional studies report a substantial trabecular bone loss before age 50 in men.11, 18, 19 Thus, in this age range, the impact of the current bone turnover rate on the variability of Dtrab is weak because Dtrab still may strongly depend on the peak bone mass acquired during growth. A purported association also may be obscured by the high variability of BTM levels and their strong correlation with age in this age group.
Between 50 and 70 years of age, BTMS and microarchitectural parameters were stable in the entire subgroup. However, in this age range, bone loss continues.19 This may explain the weak, but significant for some parameters, association between microarchitectural parameters and BTM levels. This relationship is stronger at the radius than at the tibia, where weight can have a protective effect. The associations were stronger for the parameters that have lower precision error (eg, Dcort). They were weaker for the trabecular parameters (eg, Tb.N and Tb.Th), which depend more on the partial-volume effect and whose reproducibility is poor. They were stronger for bone-formation markers than for bone-resorption markers. However, DPD measured in the early afternoon urine and CTX measured in the serum collected after a nonstandardized lunch may be not representative of the overall bone-resorption activity.
In older men, microarchitectural parameters, which decrease with age, were negatively associated with BTM levels. Thus current bone turnover rate is a significant determinant of their variability. This is consistent with data showing that, in older men, elevated BTM levels are associated with lower BMD values in cross-sectional studies13 and with faster bone loss in prospective studies.24, 25
In aging men, cortical bone loss is determined by a parallel decrease in Dcort and C.Th.26 At both skeletal sites, Dcort and C.Th decreased, whereas Tb.Area increased and total CSA did not vary across the increasing quartiles of BTM levels. This suggests that the accelerated bone turnover on the endocortical and intracortical surfaces is a major determinant of age-related cortical bone loss in men.
In aging men, trabecular bone loss is determined mainly by the decrease in Tb.Th and, to a lesser extent, the decrease in Tb.N.27 In our study, both Tb.N and Tb.Th decreased with increasing BTM levels. At the radius, the association between the microarchitectural parameters and BTM levels was significant in the inner and outer areas but slightly stronger in the inner area. At the tibia, association between BTM levels and trabecular parameters was weaker and driven by the significant association in the inner area. Mechanical strain (weight and cyclic internal and external rotational load during gait at the tibia and manual activity at the radius) may have a stronger effect on bone remodeling in the outer area, which is submitted to greater bending stress. The protective action of mechanical load may reduce the effect of general factors on bone turnover (eg, hormonal deficits, nutritional deficits, or smoking). Around the neutral axis, low bending stress has a low impact on bone remodeling, although this part is submitted to compressive stress (eg, during standing). This difference may explain the stronger association between BTM levels and trabecular parameters in this area. In contrast to the tibia, the distal radius was submitted only to the low mechanical load of urban-living older men. Therefore, bone turnover at the radius can depend more on the general factors. This may explain why the inverse association between BTM levels and trabecular microarchitecture was stronger at the radius than at the tibia and why it was stronger in the inner area than in the outer area.
The strengths of our study are the large cohort covering a large age range and representing various professional and social groups, assessment of bone microarchitecture at the weight-bearing tibia and non-weight-bearing radius, and a large panel of BTMs. Our study also has limitations. Our cohort may be not representative of the French population. The men were recruited from the rolls of a complementary insurance company. Thus social groups with lower income levels and poorer health status may be underreprestented in our cohort. Volunteers accepted to participate in a research study are often healthier among older men and in poorer health among younger men. The cross-sectional design limits inference on the cause and effect. With HR-pQCT, trabecular thickness is calculated. Despite high resolution, partial-volume effects still exist and contribute to erroneous estimations of Dcort and C.Th in men with the lowest C.Th values and thus mainly in the oldest men. In men with very thin trabeculae, it also may result in underestimation of Tb.N. We collected blood samples in the standardized manner, but in the early afternoon and after a nonstandardized lunch. Our data are consistent with those from the fasting blood samples. However, minor differences cannot be excluded. Nonfasting conditions affect serum CTX concentration. Food influences serum CTX level, but it is not clear if serum CTX depends on the food quantity or on specific nutrients, nor if the impact of food varies with age. Thus data concerning CTX and its association with bone microarchitecture should be interpreted cautiously. Since bone turnover is regulated by sex steroids, the age-related decrease in testicular secretion may influence the circadian rhythm of BTM levels in men.
In summary, in men aged less than 50 years, bone turnover rate is associated with Dcort but not other architectural parameters. Thus, in this age range, bone microarchitecture depends only weakly on the current rate of bone turnover and is determined mainly probably by peak bone mass acquired during growth. Between 50 and 70 years of age, bone turnover was negatively correlated with Dcort and, more weakly, with some trabecular parameters. After age 70, bone turnover is significantly associated with the cortical and trabecular microarchitecture at both skeletal sites. However, the observed differences were moderate in absolute terms. This is consistent with weak or moderate associations of BTM levels with BMD and bone loss found previously in two large cohorts of older men.13, 24, 28 Prospective studies are needed to establish whether the high bone remodeling is associated with an accelerated deterioration of bone microarchitecture.
This work was supported by grants from the Roche Pharmaceutical Company (Basle, Switzerland), from the Agence Nationale de la Recherche, and from Hospices Civils de Lyon.