Comparison of Speed of Sound Measures Assessed by Multisite Quantitative Ultrasound to Bone Mineral Density Measures Assessed by Dual-Energy X-Ray Absorptiometry in a Large Canadian Cohort: the Canadian Multicentre Osteoporosis Study (CaMos)

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health, vol. -, no. -, 1e8, 2015 Ó Copyright 2015 by The International Society for Clinical Densitometry 1094-6950/-:1e8/$36.00 http://dx.doi.org/10.1016/j.jocd.2015.04.004

Original Article

Comparison of Speed of Sound Measures Assessed by Multisite Quantitative Ultrasound to Bone Mineral Density Measures Assessed by Dual-Energy X-Ray Absorptiometry in a Large Canadian Cohort: the Canadian Multicentre Osteoporosis Study (CaMos) Wojciech P. Olszynski,*,1 Jonathon D. Adachi,2 David A. Hanley,3 Kenneth S. Davison,4 and Jacques P. Brown5 1

Department of Medicine, University of Saskatchewan, Saskatoon, SK, Canada; 2Department of Medicine, McMaster University, Hamilton, ON, Canada; 3Department of Medicine, University of Calgary, Calgary, AB, Canada; 4Department of Graduate Studies, University of Victoria, Victoria, BC, Canada; and 5Department of Medicine, Laval University, Quebec City, QC, Canada

Abstract Dual-energy X-ray absorptiometry (DXA) is an important tool for the estimate of fracture risk through the measurement of bone mineral density (BMD). Similarly, multisite quantitate ultrasound can prospectively predict future fracture through the measurement of speed of sound (SOS). This investigation compared BMD (at the femoral neck, total hip, and lumbar spine) and SOS measures (at the distal radius, tibia, and phalanx sites) in a large sample of randomly-selected and community-based individuals from the Canadian Multicentre Osteoporosis Study. Furthermore, mass, height, and age were also compared with both measures. There were 4123 patients included with an age range of 30e96.8 yr. Pearson product moment correlations between BMD and SOS measures were low (0.21e0.29; all p ! 0.001), irrespective of site. Mass was moderately correlated with BMD measures (0.40e0.58; p ! 0.001), but lowly correlated with SOS measures (0.03e0.13; p ! 0.05). BMD and SOS were negatively correlated to age ( 0.17 to 0.44; p ! 0.001). When regression analyses were performed to predict SOS measures at the 3 sites, the models predicted 20%e23% of the variance, leaving 77%e80% unaccounted for. The SOS measures in this study were found to be largely independent from BMD measures. In areas with no or limited access to DXA, the multisite quantitative ultrasound may act as a valuable tool to assess fracture risk. In locales with liberal access to DXA, the addition of SOS to BMD and other clinical risk factors may improve the identification of those patients at high risk for future fracture. Key Words: Bone mineral density; dual-energy X-ray absorptiometry; quantitative ultrasound; speed of sound. mQUS device assesses bone at distal radius (DR), tibia (TIB), and phalanx (PX) sites and provides an estimate of bone stiffness, expressed as speed of sound (SOS, in m/s). These SOS values can then be compared with normative mQUS values that have been published for both women and men (2), allowing for the identification of individuals at an increased risk for fracture. Dual-energy X-ray absorptiometry (DXA) provides an estimate of areal bone mineral density (BMD). In 1993, BMD was

Introduction The Beam-Med Omnisense multisite quantitative ultrasound (mQUS) has been proven to prospectively predict fracture risk in women older than a 5-yr follow-up (1). The Received 03/27/15; Accepted 04/20/15. *Address correspondence to: Wojciech P. Olszynski, MD, PhD, Suite 103, 39-23rd St E., Saskatoon, SK, Canada S7K 0H6. E-mail: [email protected]

1

2 pronounced central to the diagnosis of osteoporosis by the World Health Organization (3) with the statement that osteoporosis was a ‘‘disease characterized by low bone mass and microarchitectural deterioration of bone tissue, leading to enhanced bone fragility and a consequent increase in fracture risk.’’ In this definition of osteoporosis, low bone mass was operationalized as a DXA BMD T-score of  2.5. These guidelines led to DXA becoming the most widely used diagnostic test to assess the presence of osteoporosis. In 2000, the National Institutes of Health (United States) shifted the focus from BMD as the sole diagnostic benchmark for osteoporosis to including other clinical risk factors with their definition of osteoporosis, ‘‘a skeletal disorder characterized by compromised bone strength predisposing to an increased risk of fracture. Bone strength reflects the integration of 2 main features: bone density and bone quality (4).’’ Soon thereafter, several publications demonstrated that although the risk for osteoporotic fracture was highest in women with a BMD T-score of  2.5, most women who actually suffered a fragility fracture possessed BMD T-scores above this cutoff (5,6). Consequently, models were constructed that integrated BMD with other clinical risk factors in an attempt to better identify those individuals who are at an elevated risk for future fracture so that they can be offered appropriate management to minimize those risks. Now, through modeling large prospective epidemiologic studies, powerful algorithms have been created by that provided better prognostic utility for the assessment of 10-yr fracture risk by combining the information gleaned from BMD and clinical risk factors (7,8). Although these new 10-yr risk assessments are an improvement from the use of BMD alone, the prognostic ability of these models can still be improved to better discriminate those who are at low, moderate, or high risk of fracture. Although both instruments (mQUS and DXA) attempt to provide information on the risk for fracture via an estimate of bone strength, they do so in very different manners: the mQUS makes an estimate of bone strength via a measure of bone stiffness through the propagation of sound waves, whereas the DXA provides an estimate of bone strength via a measure of BMD through the absorption of X-rays. Because these 2 technologies have been developed with the same outcome in mind, there have been many questions as to how they compare in their estimates of bone strength and in their identification of those at risk. Few studies have compared mQUS with DXA in the same patients to investigate if there is significant concordance between the measures, and fewer have done so with a significant sample size. The objective of this investigation was to compare DXA BMD with mQUS SOS data in a large sample of randomly selected and community-based individuals from the Canadian Multicentre Osteoporosis Study (CaMos).

Materials and Methods This investigation used a subset of participants from the CaMos cohort. CaMos is a prospective study that has the

Olszynski et al. mandate to better understand the factors that lead to osteoporosis and fractures in Canadians. The methods and objectives of the CaMos study have been previously published (9). Briefly, CaMos is an ongoing prospective cohort study involving 9423 randomly selected community-dwelling women (n 5 6539) and men (n 5 2884) aged 25 yr and older at baseline and who lived within 50 km of 9 major Canadian cities (St. John’s, Newfoundland, and Labrador; Halifax, Nova Scotia; Quebec City, Quebec; Toronto, Hamilton, and Kingston, Ontario; Saskatoon, Saskatchewan; Calgary, Alberta; and Vancouver, British Columbia). Households were randomly selected from a list of residential phone numbers, and participants were randomly selected from eligible household members using standard protocol. Of those selected, 42% agreed to participate and had a baseline interview. All research carried out in the CaMos has been approved by local university ethics boards in each of the cities the study had centers in and have satisfied the criteria of the World Medical Association Declaration of HelsinkidEthical Principles for Medical Research Involving Human Subjects. All participants provided informed consent. Data collection at baseline and each follow-up visit included an extensive and standardized intervieweradministered questionnaire and a clinical assessment. Full assessments (clinical measures and questionnaires) occurred at baseline, after 3 yr (only for participants aged 40e60 yr at baseline), after 5 yr, and after 10 yr. Clinical assessment measures included height, weight, and DXA BMD of the spine (lumbar vertebrae L1eL4), femoral neck (FN), and total hip (TH). Lateral lumbar and thoracic spine X-rays were performed in all subjects who were 50 yr and older. Vertebral deformities were assessed from X-rays by a trained technologist using digital vertebral morphometry. In years that participants did not come to a study center, a self-administered fracture questionnaire was mailed out to identify incident fractures. At the 5-yr follow-up investigation, a number of the clinical sites expanded their protocol by assessing participants with a mQUS (at the 5-yr follow-up Sunlight Omnisense MultiSite Quantitative Ultrasound 7000S and now BeamMed Omnisense MultiSite Quantitative Ultrasound 7000S, Israel), in addition to the normal CaMos assessments (Calgary, Alberta; Halifax, Nova Scotia; Hamilton, Ontario; Saskatoon, Saskatchewan; Quebec City, Quebec; and St. John’s Newfoundland and Labrador; all in Canada). Because all participants were at least of 25 yr at baseline, all the participants in this analysis were at least of 30 yr (baseline plus an additional 5 yr). mQUS measurements were obtained and assessed at the DR, TIB, and PX on the nondominant side of the participant and were recorded as SOS (m/s). The metatarsal site was not assessed in this investigation. The mQUS was equipped with 2 handheld probes specifically designed for measurements of axial SOS along the surfaces of bone: 1 probe was suitable for measurements at the DRs and TIB, whereas the other assessed the phalanx. Details regarding the standard manufacturersuggested techniques involved with bone measurement with the mQUS have been detailed previously, and these standards

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

QUS SOS Compared to DXA BMD

3

were used in this investigation (10e14). Briefly, the mQUS emits and detects acoustic waves at a frequency of 1.25 MHz. The SOS measure acquired is the time taken for the sound wave to travel from the time of sound wave emission to its detection. Quality control measurements were performed daily after procedures recommended by the manufacturer. Intraobserver in vivo short-term precision has been reported as 0.76% for the radius, 0.47% for the tibia, and 1.54% for the phalanges and interobserver precision from 0.77% to 2.39% (15). mQUS assessments were conducted by 1 technologist at each of the 6 study centers. BMD (g/cm2) was assessed at the lumbar spine (L1eL4; LS), FN, TH, and the greater trochanter (GT) by DXA. The DXAs were cross-calibrated among centers using a common phantom, and a calibrated BMD was used for all analyses. Pearson product moment correlations were performed among measures of BMD (LS, FN, TH, and GT), SOS (DR, TIB, and PX), age, height, and weight. As BMD and SOS are known to both decline with increasing age (1,16) and that BMD is altered by mass (17), partial correlations were conducted controlling for age, mass, and height (exploratory) separately and for all 3 combined. Linear regression was also conducted with each of the 3 SOS sites as a dependent variable with all BMD sites, age, mass, and height as predictor variables. All analyses were completed on a Windows-based workstation with Dell Statistica, version 12 (Dell, Round Rock, TX). In situations where there were missing data, pairwise deletion was used as opposed to casewise deletion to retain as much data as possible. Statistical significance was considered to have occurred at an alpha of 0.05.

Results A total of 4123 patients were included in this investigation (2946 women and 1177 men). Women included in this study

were a mean 66.5 yr old (range: 30.0e96.8 yr), and men were an average of 63.7 yr old (range: 30.2e91.9 yr). Table 1 presents the demographic data for this cohort, including SOS and BMD at all sites investigated. By investigating the correlations between measures of SOS with BMD and other anthropometric measures, bivariate shared variance can be discerned. Table 2 provides a correlation matrix among all the investigated variables (separate data for women and men not shown). A visual example of the relationship between SOS and BMD is provided in Fig. 1AeD, where the scatterplots between the DR site and all 4 BMD sites are displayed. For the entire cohort, correlations between BMD and SOS were low (0.21e0.29; all p ! 0.01), irrespective of QUS site or DXA site. The women as a group also had low correlations between SOS and BMD (0.19e0.29; all p ! 0.01) and for the men, correlations were substantially lower (0.01e0.09; p ! 0.05 to not significant). Notably, the correlation between different BMD sites was high (0.64e0.94; all p ! 0.001). Mass was moderately correlated with BMD (0.40e0.58; p ! 0.001), whereas SOS had much lower direct correlations (0.03e0.13). BMD was also moderately correlated with height (0.33e0.45), and, to a smaller extent, SOS (0.17e0.31). Last, BMD and SOS were negatively correlated to age ( 0.17 to 0.43 and 0.23 to 0.44, respectively; all p ! 0.001). Table 3 provides the partial correlation matrix when controlled for age. When controlled for the influence of age, there was a consistent decrease in the correlation between SOS and BMD variables (0.11e0.21), particularly so for BMD measures from the hip. The association between SOS and mass but disappeared when controlled for age, whereas the association between BMD and mass remained constant or even strengthened. Controlling for mass did not have a dramatic impact on most correlations (data not shown), whereas controlling for height decreased the correlations

Table 1 Participant Demographics at Yr 5 of CaMos All participants Variable Age (yr) Mass (kg) Height (cm) DR SOS (m/s) TIB SOS (m/s) PX SOS (m/s) LS BMD (g/cm2) FN BMD (g/cm2) TH BMD (g/cm2) GT BMD (g/cm2)

Female cohort

Male cohort

Valid N

Mean

Std. dev.

Valid N

Mean

Std. dev.

Valid N

Mean

Std. dev.

4123 3680 3681 3600 3610 3876 4031 3995 3954 3995

65.704 73.203 163.473 4039.648 3865.354 3813.885 0.972 0.724 0.887 0.670

12.0219 15.4814 9.3842 150.6041 145.2289 215.5046 0.1723 0.1281 0.1577 0.1332

2946 2639 2640 2521 2514 2766 2888 2853 2822 2853

66.501 69.286 159.492 4026.177 3835.183 3785.125 0.941 0.695 0.843 0.631

11.4943 14.2240 6.8511 157.8130 145.5448 219.3230 0.1657 0.1201 0.1417 0.1159

1177 1041 1041 1079 1096 1110 1143 1142 1132 1142

63.709 83.132 173.569 4071.123 3934.560 3885.552 1.050 0.794 0.995 0.767

13.0438 14.0373 7.0883 126.8430 118.3895 187.4584 0.1639 0.1199 0.1425 0.1236

Abbr: BMD, bone mineral density; CaMos, Canadian Multicentre Osteoporosis Study; DR, distal radius site; FN, femoral neck site; GT, greater trochanter site; LS, lumbar spine site; PX, phalanx site; SOS, speed of sound; Std. dev., standard deviation; TH, total hip site; TIB, tibia site. Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

4

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Table 2 Pearson Product Moment Correlation Matrix (Full Cohort) Variable Age TIB SOS DR SOS PX SOS LS BMD FN BMD TH BMD GT BMD Mass Height

Age

TIB

DR

PX

LS

FN

TH

GT

Mass

Height

1.0000 N 5 4123 0.2254 N 5 3610 0.3990 N 5 3600 0.4383 N 5 3876 0.1718 N 5 4031 0.4335 N 5 3995 0.3634 N 5 3954 0.2636 N 5 3995 0.1733 N 5 3680 0.2681 N 5 3681

1.0000 N 5 3610 0.4238 N 5 3435 0.3673 N 5 3378 0.2483 N 5 3543 0.2509 N 5 3511 0.2697 N 5 3480 0.2657 N 5 3511 0.0349* N 5 3208 0.3094 N 5 3209

1.0000 N 5 3600 0.4853 N 5 3368 0.2148 N 5 3528 0.2684 N 5 3493 0.2619 N 5 3464 0.2371 N 5 3493 0.0425* N 5 3215 0.1731 N 5 3216

1.0000 N 5 3876 0.2294 N 5 3791 0.2877 N 5 3755 0.2846 N 5 3715 0.2362 N 5 3755 0.1333 N 5 3520 0.2420 N 5 3521

1.0000 N 5 4031 0.6422 N 5 3979 0.6830 N 5 3940 0.6688 N 5 3979 0.4011 N 5 3632 0.3259 N 5 3635

1.0000 N 5 3995 0.9020 N 5 3954 0.8254 N 5 3995 0.4818 N 5 3600 0.4164 N 5 3603

1.0000 N 5 3954 0.9397 N 5 3954 0.5789 N 5 3560 0.4508 N 5 3563

1.0000 N 5 3995 0.5381 N 5 3600 0.4546 N 5 3603

1.0000 N 5 3680 0.5428 N 5 3676

1.0000 N 5 3681

Note: All correlations significant at p ! 0.001 with the exception of those demarked with *, which are significant at p ! 0.05. Abbr: BMD, bone mineral density; DR, distal radius site; FN, femoral neck site; GT, greater trochanter site; LS, lumbar spine site; PX, phalanx site; SOS, speed of sound; TH, total hip site; TIB, tibia site.

-,

2015

Olszynski et al.

Volume

QUS SOS Compared to DXA BMD

5

Fig. 1. Scatterplots of distal radius (A) SOS vs lumbar spine BMD, (B) femoral neck BMD, (C) total hip BMD, and (D) greater trochanter BMD for the entire cohort. BMD, bone mineral density; SOS, speed of sound. between BMD and SOS (0.14e0.22; all p ! 0.01; data not shown). Correction for age, mass, and height further decreased the associations between BMD and SOS (0.06e0.21; all p ! 0.01; data not shown). Table 4 presents the results of the regression analyses, which were performed for the combined cohort for each of the 3 SOS sites. At the DR site, the model accounted for 20% of the variance, with age accounting for the largest variance, with LS BMD, FN BMD, mass, and height all accounting for smaller proportions. At the TB site, the model also accounted for 20% of the variance with mass, height, and TH BMD accounting for the largest variances, with age and the other BMD sites accounting for smaller proportions. At the PX site, the model accounted for 23% of the variance, with age and TH BMD accounting for the largest variance, with the other BMD sites, mass, and height accounting for smaller proportions.

Discussion In this large randomly selected cohort, we have demonstrated that the mQUS SOS measurements are largely

independent from the DXA BMD measurements with unadjusted shared variance ranging from 4% to 8% (r2) for the combined cohort. When adjusted for height or age through partial correlation, the shared variance between SOS and BMD measures decreased further, with BMD and SOS measures only sharing 1%e5% (r2) of their variance. Multiple regression analyses further demonstrated the relative independence of mQUS SOS measures from DXA BMD with no BMD measure accounting for more than 7% of the model variance. Thus, through multiple analyses, it can be said that DXA BMD and mQUS SOS have under 10% shared variance at best. Although BMD is a powerful predictor of fracture risk (18e22), with an inverse relationship between BMD and fracture risk (5,6), most women who suffer a fragility fracture have a BMD T-score that is better than that which would be deemed osteoporotic (5,6). With this fact in mind, 10-yr fracture risk assessment models have been generated that combine the predictive power of BMD with that of other risk factors to better capture the variance of fracture risk (7,8,23). The mQUS used in these analyses has already been demonstrated by our group to be a significant predictor of fracture risk over a 5-yr follow-up period when used in

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

6

Olszynski et al. Table 3 Partial Correlation Matrix of Full Cohort (Controlled for Age)

Variable

TIB

DR

PX

LS

FN

TH

GT

Mass

Height

TIB SOS

1.0000 N 5 3610 0.3738 N 5 3435 0.3067 N 5 3378 0.2183 N 5 3543 0.1744 N 5 3511 0.2069 N 5 3480 0.2196 N 5 3511 0.0043NS N 5 3208 0.2653 N 5 3209

1.0000 N 5 3600 0.3767 N 5 3368 0.1619 N 5 3528 0.1155 N 5 3493 0.1368 N 5 3464 0.1492 N 5 3493 0.0295NS N 5 3215 0.0748 N 5 3216

1.0000 N 5 3876 0.1740 N 5 3791 0.1206 N 5 3755 0.1496 N 5 3715 0.1392 N 5 3755 0.0647 N 5 3520 0.1438 N 5 3521

1.0000 N 5 4031 0.6395 N 5 3979 0.6762 N 5 3940 0.6562 N 5 3979 0.3827 N 5 3632 0.2949 N 5 3635

1.0000 N 5 3995 0.8868 N 5 3954 0.8181 N 5 3995 0.4582 N 5 3600 0.3457 N 5 3603

1.0000 N 5 3954 0.9390 N 5 3954 0.5623 N 5 3560 0.3937 N 5 3563

1.0000 N 5 3995 0.5183 N 5 3600 0.4131 N 5 3603

1.0000 N 5 3680 0.5231 N 5 3676

1.0000 N 5 3681

DR SOS PX SOS LS BMD FN BMD TH BMD GT BMD Mass Height

Note: All correlations significant at p ! 0.001 with the exception of those demarked with NS, which are not significant at p ! 0.05. Abbr: BMD, bone mineral density; DR, distal radius site; FN, femoral neck site; GT, greater trochanter site; LS, lumbar spine site; NS, not significant; PX, phalanx site; SOS, speed of sound; TH, total hip site; TIB, tibia site.

isolation (without consideration of any other risk factors) and when added to all the clinical risk factors and FN BMD that are used as inputs in the FRAX 10-yr fracture risk assessment (1). The mQUS values were able to identify a significant proportion of the unaccounted for variance of fracture risk above that of all the variables entered into the FRAX model. It is the relative independence of mQUS SOS from DXA BMD and other anthropometric variables, as demonstrated in this investigation, and its ability to independently predict fracture that make it a valuable tool. If mQUS SOS and DXA BMD were found to be highly correlated, then the addition of mQUS to fracture prediction models that already have BMD as an input would lead to no meaningful change in fracture risk prediction and would then unnecessarily complicate the model and be a waste of time and resources. Given this, the combination of mQUS and DXA should be better than either one alone in predicting fractures. As the incidence of fragility fracture increases with age, it was reassuring to see that one of the most consistent and powerful correlates of mQUS was age. In this investigation, DXA BMD values between sites were very high, particularly between hip regions, demonstrating that adding 2 or more BMD sites to a fracture risk model may add little to the ability to predict fracture. Furthermore, the high correlation between mass and BMD was not observed between mass and SOS in this investigation, further implicating mQUS as an important independent variable for addition to 10-yr fracture risk models.

A possible explanation for the discordant findings between instruments may be that they are in fact both estimating the risk of the primary clinical outcome (fracture risk) but are measuring different components of the ultimate strength that would lend to a fragile or strong bone. A recent investigation demonstrated that there was a significant negative correlation between QUS SOS and cortical porosity (24). Investigations with high-resolution peripheral quantitative computerized tomography have demonstrated that cortical porosity is a very powerful predictor of bone strength and future fracture (25e28). DXA is not able to assess cortical porosity. Although the power of this sample allowed for statistically significant results to be found between the methods, it can be safely assumed that they are measuring different attributes of bone, which may have important implications for fracture prediction that need to be further assessed with long-term prospective trials, particularly with mQUS. In conclusion, the mQUS device SOS values were found to be largely independent from DXA BMD measures. mQUS SOS measures have been previously demonstrated by our group to be able to prospectively predict fragility fracture either when used as a solitary risk factor or when added to the variables included in 10-yr fracture risk assessment models. Thus, in areas with no or limited access to DXA, the mQUS may act as a valuable tool to assess fracture risk. In locales with liberal access to DXA, the addition of mQUS SOS to BMD and other clinical risk factors may improve the identification of those patients at high risk for future fracture.

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

QUS SOS Compared to DXA BMD

7

Table 4 Linear Regression Analyses Results for Entire Cohort Regression summary for distal radius SOS R 5 0.45; R2 5 0.20; adjusted R2 5 0.20; F(7,3207) 5 117.09, p ! 0.0000; std. err. of the estimate: 134.55 b coefficient Intercept Age LS BMD FN BMD TH BMD GT BMD Mass Height

0.360886 0.117660 0.081746 0.108862 0.090230 0.183647 0.081565

Std. err. of b coeff.

Raw b coeff.

Std. err. of raw b coeff.

0.018375 0.021976 0.038596 0.064602 0.048665 0.021342 0.019691

4062.538 4.521 102.833 96.135 103.935 102.041 1.787 1.309

52.81404 0.23019 19.20661 45.38923 61.67776 55.03528 0.20761 0.31601

t(3207) 76.9216 19.6401 5.3540 2.1180 1.6851 1.8541 8.6050 4.1424

p Value

Valid N

0.000000 0.000000 0.000000 0.034250 0.092062 0.063817 0.000000 0.000035

4123 4031 3995 3954 3995 3680 3681

Regression summary for tibia SOS R 5 0.45; R2 5 0.20; adjusted R2 5 0.20; F(7,3200) 5 113.05, p ! 0.0000; std. err. of estimate: 130.18 b coefficient Intercept Age LS BMD FN BMD TH BMD GT BMD Mass Height

0.123422 0.135171 0.132963 0.265915 0.026675 0.325238 0.332174

Std. err. of b coeff.

Raw b coeff.

Std. err. of raw b coeff.

0.018456 0.022073 0.038766 0.064887 0.048880 0.021436 0.019777

3108.106 1.491 113.921 150.786 244.817 29.090 3.051 5.141

51.15368 0.22295 18.60280 43.96229 59.73874 53.30509 0.20109 0.30607

t(3200) 60.7602 6.6873 6.1238 3.4299 4.0981 0.5457 15.1725 16.7957

p Value 0.000000 0.000000 0.000000 0.000611 0.000043 0.585296 0.000000 0.000000

Valid N 4123 4031 3995 3954 3995 3680 3681

Regression summary for phalanx SOS R 5 0.48, R2 5 0.23, adjusted R2 5 0.23; F(7,3512) 5 150.16, p ! 0.0000; std. err. of estimate: 189.25 b coefficient Intercept Age LS BMD FN BMD TH BMD GT BMD Mass Height

0.382166 0.127290 0.105699 0.241449 0.099375 0.087812 0.126069

Std. err. of b coeff.

Raw b coeff.

Std. err. of raw b coeff.

0.017261 0.020644 0.036256 0.060686 0.045715 0.020048 0.018497

3669.501 6.851 159.192 177.874 329.862 160.813 1.222 2.895

70.99253 0.30942 25.81749 61.01212 82.90713 73.97833 0.27907 0.42478

t(3512) 51.6885 22.1404 6.1660 2.9154 3.9787 2.1738 4.3801 6.8157

p Value 0.000000 0.000000 0.000000 0.003575 0.000071 0.029788 0.000012 0.000000

Valid N 4123 4031 3995 3954 3995 3680 3681

Abbr: BMD, bone mineral density; coeff., coefficient; FN, femoral neck site; GT, greater trochanter site; LS, lumbar spine site; SOS, speed of sound; std. err., standard error; TH, total hip site.

Acknowledgments CaMos Research Group David Goltzman (co-principal investigator, McGill University), Nancy Kreiger (co-principal investigator, Cancer Care Ontario), Alan Tenenhouse (principal investigator emeritus, Toronto); McGill University, Montreal, Quebec: Elham Rahme

(biostatistician), J. Brent Richards (investigator), Suzanne N. Morin (investigator). CaMos Coordinating Center: Suzanne Godmaire (research assistant), Silvia Dumont (research assistant), Claudie Berger (study statistician), Lisa Langsetmo (fellow); Memorial University, St. John’s Newfoundland: Carol Joyce (director), Christopher S. Kovacs (co-director), Minnie

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

8 Parsons (coordinator); Dalhousie University, Halifax, Nova Scotia: Susan Kirkland, Stephanie M. Kaiser (co-directors), Barbara Stanfield (coordinator); Laval University, Quebec City, Quebec: Jacques P. Brown (director), Louis Bessette (co-director), GRMO, Jeanette Dumont (coordinator), Martin Despres (imaging IT technician).; Queen’s University, Kingston, Ontario: Tassos P. Anastassiades (director), Tanveer Towheed (co-director), Wilma M. Hopman (investigator), Karen J. Rees-Milton (coordinator); University of Toronto, Toronto, Ontario: Robert G. Josse (director), Sophie A. Jamal (co-director), Angela M. Cheung (investigator), Barbara Gardner-Bray (coordinator); McMaster University, Hamilton, Ontario: Jonathan D. Adachi (director), Alexandra Papaioannou (co-director), Laura Pickard (coordinator); University of Saskatchewan, Saskatoon, Saskatchewan: Wojciech P. Olszynski (director), K. Shawn Davison (co-director), Jola Thingvold (coordinator); University of Calgary, Calgary, Alberta: David A. Hanley (director), Steven K. Boyd (co-director), Jane Allan (coordinator); University of British Columbia, Vancouver, British Columbia: Jerilynn C. Prior (director), Millan S. Patel (co-director), Brian Lentle (investigator/radiologist), Nerkeza Andjelic (coordinator); University of Alberta, Edmonton, Alberta: Stuart D. Jackson (medical physicist); University of Manitoba, Winnipeg, Manitoba: William D. Leslie (investigator/nuclear medicine physician). CaMos is currently supported by Canadian Institutes of Health Research, Amgen Canada Inc., Merck Canada Inc., Eli Lilly and Company, Actavis Specialty Pharmaceuticals Co. (formerly Warner Chilcott Canada Co.), and Hologic Inc. We also acknowledge our incredible CaMos participants who made this study possible through their unwavering support of this investigation.

References 1. Olszynski WP, Brown JP, Adachi JD, et al. 2013 Multisite quantitative ultrasound for the prediction of fractures over five years of follow-up: the Canadian Multicentre Osteoporosis Study. J Bone Miner Res 28:2027e2034. 2. Olszynski WP, Brown JP, Adachi JD, et al. 2013 Normative data for multisite quantitative ultrasound: the Canadian Multicenter Osteoporosis Study. J Clin Densitom 17:534e540. 3. Consensus development conference: diagnosis, prophylaxis, and treatment of osteoporosis. 1993 Am J Med 94:646e650. 4. Osteoporosis prevention, diagnosis and therapy. NIH consensus statement. 2000. Report No.: 17(1). 5. Siris ES, Chen YT, Abbott TA, et al. 2004 Bone mineral density thresholds for pharmacological intervention to prevent fractures. Arch Intern Med 164:1108e1112. 6. Cranney A, Jamal SA, Tsang JF, et al. 2007 Low bone mineral density and fracture burden in postmenopausal women. CMAJ 177:575e580. 7. Kanis JA, McCloskey EV, Johansson H, et al. 2010 Development and use of FRAX in osteoporosis. Osteoporos Int 21(2 Suppl): S407eS413. 8. Leslie WD, Berger C, Langsetmo L, et al. 2011 Construction and validation of a simplified fracture risk assessment tool for Canadian women and men: results from the CaMos and Manitoba cohorts. Osteoporos Int 22:1873e1883.

Olszynski et al. 9. Kreiger N, Tenenhouse A, Joseph L, et al. 1999 The Canadian Multicentre Osteoporosis Study (CaMos): Background, rationale, methods. Can J Aging 18:376e387. 10. Weiss M, Ben-Shlomo AB, Hagag P, Rapoport M. 2000 Reference database for bone speed of sound measurement by a novel quantitative multi-site ultrasound device. Osteoporos Int 11:688e696. 11. Njeh CF, Hans D, Wu C, et al. 1999 An in vitro investigation of the dependence on sample thickness of the speed of sound along the specimen. Med Eng Phys 21:651e659. 12. Njeh CF, Saeed I, Grigorian M, et al. 2001 Assessment of bone status using speed of sound at multiple anatomical sites. Ultrasound Med Biol 27:1337e1345. 13. Knapp KM, Blake GM, Spector TD, Fogelman I. 2001 Multisite quantitative ultrasound: precision, age- and menopause-related changes, fracture discrimination, and T-score equivalence with dual-energy X-ray absorptiometry. Osteoporos Int 12:456e464. 14. Knapp KM, Blake GM, Fogelman I, et al. 2002 Multisite quantitative ultrasound: Colles’ fracture discrimination in postmenopausal women. Osteoporos Int 13:474e479. 15. Damilakis J, Papadokostakis G, Vrahoriti H, et al. 2003 Ultrasound velocity through the cortex of phalanges, radius, and tibia in normal and osteoporotic postmenopausal women using a new multisite quantitative ultrasound device. Invest Radiol 38:207e211. 16. Berger C, Goltzman D, Langsetmo L, et al. 2010 Peak bone mass from longitudinal data: implications for the prevalence, pathophysiology, and diagnosis of osteoporosis. J Bone Miner Res 25:1948e1957. 17. Ho-Pham LT, Nguyen UD, Nguyen TV. 2014 Association between lean mass, fat mass, and bone mineral density: a metaanalysis. J Clin Endocrinol Metab 99:30e38. 18. Berger C, Langsetmo L, Joseph L, et al. 2009 Association between change in BMD and fragility fracture in women and men. J Bone Miner Res 24:361e370. 19. Johnell O, Kanis JA, Oden A, et al. 2005 Predictive value of BMD for hip and other fractures. J Bone Miner Res 20:1185e1194. 20. Kanis JA, Johnell O, Oden A, et al. 2001 Ten year probabilities of osteoporotic fractures according to BMD and diagnostic thresholds. Osteoporos Int 12:989e995. 21. Langsetmo L, Goltzman D, Kovacs CS, et al. 2009 Repeat lowtrauma fractures occur frequently among men and women who have osteopenic BMD. J Bone Miner Res 24:1515e1522. 22. Sarkar S, Reginster JY, Crans GG, et al. 2004 Relationship between changes in biochemical markers of bone turnover and BMD to predict vertebral fracture risk. J Bone Miner Res 19:394e401. 23. Kanis JA, Oden A, Johnell O, et al. 2007 The use of clinical risk factors enhances the performance of BMD in the prediction of hip and osteoporotic fractures in men and women. Osteoporos Int 18:1033e1046. 24. Mathieu V, Chappard C, Vayron R, et al. 2013 Radial anatomic variation of ultrasonic velocity in human cortical bone. Ultrasound Med Biol 39:2185e2193. 25. Bjornerem A, Bui QM, Ghasem-Zadeh A, et al. 2013 Fracture risk and height: an association partly accounted for by cortical porosity of relatively thinner cortices. J Bone Miner Res 28: 2017e2026. 26. Patsch JM, Burghardt AJ, Yap SP, et al. 2013 Increased cortical porosity in type 2 diabetic postmenopausal women with fragility fractures. J Bone Miner Res 28:313e324. 27. S Y T, Vashishth D. 2011 The relative contributions of nonenzymatic glycation and cortical porosity on the fracture toughness of aging bone. J Biomech 44:330e336. 28. Macdonald HM, Nishiyama KK, Kang J, et al. 2011 Age-related patterns of trabecular and cortical bone loss differ between sexes and skeletal sites: a population-based HR-pQCT study. J Bone Miner Res 26:50e62.

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015