Relationship between risk factors and QUS in a European Population: The OPUS study

Bone 39 (2006) 609 – 615 www.elsevier.com/locate/bone

Relationship between risk factors and QUS in a European Population: The OPUS study A. Stewart a,⁎, D. Felsenberg b , R. Eastell c , C. Roux d , C.C. Glüer e , D.M. Reid a a

d e

Osteoporosis Research Unit, Department of Medicine and Therapeutics, University of Aberdeen, Aberdeen, UK b Diagnostische Radiologie, Klinikums Benjamin Franklin der Freien Universität Berlin, Germany c Division of Clinical Sciences, University of Sheffield, Sheffield, UK Centre d'Evaluation des Maladies Osseuses, Service de Rhumatologie, Hôpital Cochin, Université René Descartes, Paris, France Medizinische Physik, Klinik für Diagnostische Radiologie, Universitätsklinikum Schleswig-Holstein, Campus Kiel, Kiel, Germany Received 7 November 2005; revised 7 February 2006; accepted 21 February 2006 Available online 27 April 2006

Abstract There are many risk factors associated with low bone mineral density. Quantitative ultrasound (QUS) is a generally accepted method for measurement of bone and has been shown to be strongly associated with future fracture risk. The Osteoporosis and Ultrasound Study (OPUS) is a multi-centre European wide study examining 5 different QUS scanners (4 calcaneal, 1 finger device). The aim of this paper was to examine the relationship between risk factors (as assessed by questionnaire) and QUS measurements. 449 younger women (aged 20 to 39 years) and 2283 older women (aged 55 to 79 years) were included in this analysis. As expected, those with a self-reported previous fracture had lower QUS measurements than those without (P < 0.001). However, no significant difference was seen between those reporting a maternal hip fracture and those who did not report such an event. Differences were found for smokers vs. non-smokers for SOS but not for BUA measurements. Weight was positively correlated with all BUA variables but only with some SOS variables. We determined which risk factors were most strongly associated with QUS measurements by using step-wise multiple regression. Models for each QUS measurement were calculated, and the R2 values ranged from 0.18 to 0.28 for SOS, 0.27 to 0.32 for BUA and 0.31 to 0.42 for the finger QUS device. The most common risk factors across all models were age, use of hormone replacement therapy, self-reported previous fracture, self-reported diagnosis of osteoporosis, current weight, pulse rate and self-reported estimated height at age 20 years. We analysed relationships across the 5 centres and detected some geographical differences in the prevalence of the risk factors. In conclusion, similar relationships are seen with QUS measurements as are found for bone mineral density. However, the strength of the association is dependent on the type of QUS device and variable measured. © 2006 Elsevier Inc. All rights reserved. Keywords: Risk factors; Osteoporosis; Quantitative ultrasound

Introduction There have been many risk factors associated with low bone mineral density, osteoporosis and fragility fractures. Generally accepted are the 4 major risk factors put forward by the Study of Osteoporotic Fractures (SOF) group [1]: previous fragility fracture, low body weight, smoking and maternal hip fracture. These have been shown to be in similar prevalence in a UK population previously [2]. Quantitative ultrasound (QUS) has emerged as another possible technique to estimate future fracture risk. There are ⁎ Corresponding author. E-mail address: [email protected] (A. Stewart). 8756-3282/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2006.02.072

many commercially available QUS devices, which measure in different ways and at different sites. The relationship between these different QUS devices and the major risk factors has not been compared alongside each other before. In this study, we have examined a large-population-based cohort of postmenopausal women and a smaller-populationbased younger cohort in 5 centres across Europe. We have analysed risk factors to determine the association with quantitative ultrasound (QUS) parameters. Methods and materials Five European centres participated in the OPUS study: Aberdeen (ABE), Berlin (BER), Kiel (KIE), Paris (PAR), and Sheffield (SHE). The Medical Physics Research Group in Kiel, Germany coordinated the study. All

610

A. Stewart et al. / Bone 39 (2006) 609–615

Table 1 Data for younger and older groups of women (P < 0.001 for all measurements)

Age (years) Height (cm) Weight (kg) Body mass index (BMI) (kg/m2) Years since menopause (years) BUA Achilles DTU-one UBIS QUS2 SOS Achilles DTU-one UBIS Ad-SOS IGEA BTT IGEA UBPI IGEA Stiffness index Achilles DXA spine (g/cm2) DXA femoral neck (g/cm2) DXA total hip (g/cm2) a

Younger women n = 449

Older women n = 2283

Mean

Mean

31.3 165.9 65.9 23.9 3.2 a 116.4 49.8 68.7 89.7 1569.0 1554.4 1523.1 2079.0 1.58 0.68 96.7 1.163 0.910 0.759

SD

SD

5.2 6.7 12.9 4.4

66.9 160.5 68.6 26.6

7.0 6.3 12.4 4.5

2.3

18.6

9.2

9.3 6.5 11.5 13.9 28.3 9.6 25.8 70.8 0.19 0.14 12.6 0.133 0.137 0.119

107.3 46.7 61.1 76.7 1530.6 1545.6 1499.5 1890.3 1.37 0.33 80.0 1.034 0.756 0.679

9.6 7.9 13.9 16.6 28.5 10.6 29.3 127.5 0.25 0.20 13.4 0.182 0.130 0.129

n = 15, all others were premenopausal.

investigations were conducted in accordance with the Declaration of Helsinki and were approved by the appropriate institutional human research committee at each participating centre. The study details have been published elsewhere [3], but, in brief, participants of the OPUS study were recruited from random population samples between April 1999 and April 2001. We included women of two different age segments: age 20–39 (“younger women”) or age 55–79 (“older women”). Exclusion criteria were limited to disorders that precluded valid QUS measurements (i.e. bilateral fractures of the calcaneus, bilateral hip prostheses, disorder of the hand), general inability to undergo the specified examinations, and cognitive limitations that preclude filling out self-administered questionnaires. Pregnant women were excluded because of potential risks associated with X-ray exposure. Details of each examination and questionnaires completed are detailed elsewhere [3], but, in this study, we concentrated on QUS measurements as obtained on five different QUS devices. For measurements at the calcaneus, we used the Achilles+ (GE Lunar, Madison, WI, USA), UBIS 5000 (Diagnostic Medical Systems, Montpellier, France), DTU-one (OSI/Osteometer Meditech, Hawthorn, CA, USA) and QUS-2 (Quidel/Metra, San Diego, CA, USA), and for measurements at the finger phalanges, we used the DBM Sonic BP (IGEA, Carpi, Italy). The following QUS variables were evaluated: Speed of Sound (SOS) on the Achilles+, the UBIS 5000, and the DTU-one; Broadband Ultrasound Attenuation (BUA) on all four calcaneus devices; stiffness index as linear combination of BUA and SOS on the Achilles+; Amplitude-Dependent Speed of Sound (Ad-SOS) and, as secondary variables, Bone Transmission Time (BTT) and Ultrasound Bone Profile Index (UBPI) on the DBM Sonic BP device. Each measurement was performed twice on each device with interim repositioning of the subject. If the two measurements deviated by more than a predefined device-specific threshold, a third final measurement was obtained. To obtain the final result for any variable of a given patient, we averaged two results — if a third measurement had been taken, the two closest results were averaged. Dual energy X-ray absorptiometry (DXA) scans were also performed on all subjects, in all centres, at the spine and hip. In 3 centres a Hologic QDR-4500 (Hologic, MA, USA) scanner was used and in the other 2 centres a Lunar Expert (GE Lunar, WI, USA) device was used. Each participant also filled out a number of questionnaires. The “OPUS risk factor questionnaire”, a modified version of the EVOS risk factor questionnaire

of the European Vertebral Osteoporosis Study [4], was administered in interview fashion. It includes biographical questions, aspects of family history of osteoporosis, self-reported medical history (with a focus on fractures and falls), medications known to affect skeletal metabolism, nutrition and lifestyle aspects, etc. Univariate analyses of all 98 risk factors were calculated individually for the 11 QUS measurements. Those variables which were statistically significant were then entered into a step-wise regression model. The step-wise regression model included a constant and excluded cases listwise. t tests and analysis of variance (ANOVA) were used to calculate differences in those with/without fractures, smoking status, weight and maternal hip fracture. Statistics were carried out using SPSS v 12 software (SPSS Inc, Illinois, Chicago, USA).

Results We present data for 449 younger women, aged 20 to 40 years, and for 2283 older postmenopausal women aged between 55 and 80 years. Baseline data are shown in Table 1. Step-wise regression Forward step-wise models for each of the 11 individual QUS variables were calculated (Table 2), and R2 values ranged from 0.18 for DTU-SOS to 0.42 for IGEA SOS. Number of risk factors entered into the model varied from 7 for DTU-SOS, IGEA SOS, IGEA UBPI to 12 for Ach BUA. In order to identify if different risk factors are associated with BUA/SOS/Stiffness index, we used the individual models to ascertain the most common risk factors within BUA and SOS. The 3 IGEA variables were also assessed separately due to the different technology and site used in obtaining measurements. The results are shown in Table 3. Relationship between 4 common risk factors and QUS Here, we examined the relationship between QUS and 4 common risk factors, for hip fracture, of previous personal history of fracture, maternal hip fracture, smoking and weight. If we examine the relationship between QUS and previous fracture, we find that, for the younger group only, BTT, IGEA Ad-SOS and DTU-SOS were significantly lower in the fracture group. However, in the older group, we found that all QUS parameters were significantly different in the fracture group (Table 4). To compare with DXA, we also analysed both spine Table 2 Results from step-wise regression for each QUS measurement

Ach BUA DTU-one BUA UBIS BUA QUS2 BUA Ach SOS DTU-one SOS UBIS SOS IGEA Ad-SOS IGEA BTT IGEA UBPI Ach Stiffness index

n

R2 value

Number of risk factors in model

496 730 477 510 596 492 561 468 586 462 501

0.306 0.270 0.322 0.277 0.253 0.181 0.224 0.418 0.311 0.366 0.297

12 9 9 8 10 7 9 7 9 7 10

A. Stewart et al. / Bone 39 (2006) 609–615 Table 3 R2 values from step-wise regression when most common risk factors are entered Measurement Most common risk factors entered into regression

R2

Ach BUA DTU-one BUA UBIS BUA QUS2 BUA

Age, current weight, self-reported diagnosis of rheumatoid arthritis, self-reported hip fracture, self-reported use of hormone replacement therapy in last year, self-reported diagnosis of osteoporosis, ever used hormones around time of menopause, self-reported use of thyroxine in last year

0.298 0.214

Ach SOS DTU-one SOS UBIS SOS

Age, current weight, self-reported diagnosis 0.292 0.175 of rheumatoid arthritis, self-reported use of hormone replacement therapy in last year, self-reported diagnosis of osteoporosis, ever used hormones around time of 0.197 menopause, current smoking status, pulse rate, height, self-reported use of steroids in last year

IGEA Ad-SOS IGEA BTT IGEA UBPI

Age, self-reported diagnosis of rheumatoid arthritis, self-reported use of hormone replacement therapy in last year, study centre, number of years in full-time education, self-reported fracture other than spine, hip, wrist or rib, self-reported hysterectomy

Ach Stiffness Age, current weight, self-reported use of hormone index replacement therapy in last year, time spent doing sports since leaving school, self-reported diagnosis of rheumatoid arthritis, self-reported diagnosis of osteoporosis, self-reported diagnosis of hyperthyroidism, pulse rate

0.259 0.266

0.533 0.264 0.564

0.356

and total hip BMD. Neither spine nor total hip BMD were associated with previous fracture in the younger group, but both were highly significant lower in previous fracture cases (P < 0.001) for the older group. Of the younger group, only 2 (0.4%) indicated that their mother had a hip fracture, therefore no statistics were calculated for the younger group due to low prevalence of maternal hip fracture. In the older group, 242 (10.4%) indicated that their mother had a hip fracture. However, there was no significant difference in any of the QUS parameters for those who indicated a maternal hip fracture compared to those who did not report it. For BMD, we find significant differences in the older group (maternal hip fracture mean = 0.841, SD 0.133; no maternal hip fracture mean = 0.869, SD 0.145; P = 0.003) for total hip BMD,

611

but no significant differences for spine BMD (maternal hip fracture mean = 1.035, SD 0.183; no maternal hip fracture mean = 1.028, SD 0.182; P = 0.541). With regards smoking, women were categorised as ever smoked or never smoked. For the younger women, 212 (46.1%) had ever smoked (current and past), while for the older population, 1016 (43.2%) had ever smoked. We found significant differences between the ever and never smoked categories for the younger population in IGEA Ad-SOS (mean ever smoker = 2068.6 (SD 78.9); mean never smoker = 2088.1 (SD 62.4); P < 0.004) only. IGEA UBPI showed a trend, but this did not reach significance (P = 0.078). When we subdivided the ever smoker category into current and past smoker and compared to never smoker, IGEA Ad-SOS remained significant (P = 0.011), while the other QUS parameters remained insignificant. No significant associations were found with smoking and spine and total hip BMD in either the younger or older age groups. For the older population, BUA (for either Achilles, DTUone, UBIS or QUS-2) did not show any significant differences between ever and never smokers. SOS however showed significant differences regardless of the scanner it was measured on (Table 5). IGEA UBPI was also significant (P < 0.001), while IGEA BTT was showing a non-significant trend (P = 0.057). When we subdivided the ever smoker category, as above, only DTU BUA became significant (P = 0.014), while all the SOS measurements remained significant, except for Ach SOS (P = 0.131). IGEA UBPI remained significant (P = 0.003), while IGEA BTT was now non-significant (P = 0.118). We investigated whether there was a relationship between smoking and weight since the effect of smoking may simply be a surrogate for weight. In the older population, we observed a significant difference (P < 0.001) in weight across the smoking categories, with the current smokers having the lowest weight (mean = 66.2; SD 12.5), while past smokers had the highest weight (mean = 70.5; SD 12.6), and never smokers had an intermediate value (mean = 68.2; SD 12.1). Using a multivariate analysis to adjust for any effect of weight on smoking, there was still a significant relationship (Table 5) between smoking and SOS and UBPI, indicating that the relationship is independent of weight.

Table 4 Previous fracture history and QUS (mean with SD in parentheses) QUS parameter

Machine

BUA

Achilles DTU-one UBIS QUS2 Achilles DTU-one UBIS IGEA IGEA IGEA Achilles

SOS

Ad-SOS BTT UBPI Stiffness index

Younger age group

Older age group

Fracture n = 149 (32.4%)

No fracture n = 311(67.6%)

P value

Fracture n = 1016 (43.4%)

No fracture n = 1323 (56.6%)

P value

116.2 (9.2) 49.3 (6.2) 67.7 (11.0) 90.0 (14.4) 1567.3 (26.4) 1552.6 (8.5) 1519.2 (22.9) 2068.3 (68.5) 1.55 (0.18) 0.67 (0.15) 96.1 (12.0)

116.4 (9.3) 50.0 (6.7) 69.1 (11.7) 89.5 (13.7) 1569.8 (29.2) 1555.2 (10.0) 1525.1 (27.0) 2084.3 (71.6) 1.59 (0.20) 0.69 (0.13) 97.0 (12.9)

0.836 0.335 0.267 0.736 0.428 0.008 0.034 0.026 0.038 0.161 0.550

106.4 (9.5) 45.5 (7.9) 58.7 (13.8) 74.6 (16.2) 1526.1 (27.2) 1543.8 (10.0) 1494.0 (28.0) 1869.8 (123.8) 1.33 (0.24) 0.30 (0.18) 78.1 (13.1)

108.1 (9.6) 47.6 (7.8) 62.9 (13.6) 78.4 (16.6) 1534.5 (28.9) 1547.1 (10.9) 1503.8 (29.6) 1907.1 (128.4) 1.39 (0.24) 0.36 (0.20) 81.6 (13.3)

<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

612

A. Stewart et al. / Bone 39 (2006) 609–615

Table 5 Smoking and SOS in the older population only SOS variable

Ach DTU UBIS IGEA a BTT b UBPI c a b c

Ever smokers n = 1016

Never smoked n = 1336

Mean

SD

Mean

SD

1529.2 1544.8 1496.6 1877.8 1.35 0.32

29.0 10.6 29.4 129.6 0.25 0.19

1531.8 1546.2 1501.3 1898.9 1.37 0.35

28.1 10.5 28.9 125.4 0.24 0.20

P value

Current smokers adjusted for weight

Past smokers adjusted for weight

Never smoked adjusted for weight

Mean

SE

Mean

SE

Mean

SE

0.054 0.002 <0.001 <0.001 0.057 0.001

1528.9 1544.3 1494.7 1879.0 1.344 0.321

1.6 0.6 1.6 7.0 0.014 0.011

1529.4 1544.9 1497.5 1878.2 1.359 0.317

1.2 0.4 1.2 5.0 0.010 0.008

1531.8 1546.2 1501.4 1989.5 1.373 0.346

0.9 0.3 0.8 3.5 0.007 0.005

P value

0.151 0.002 <0.001 0.001 0.115 0.005

IGEA Ad-SOS. IGEA BTT. IGEA UBPI.

Weight was related to QUS parameters by examining the correlation between weight and QUS. We found that in the younger population weight was positively correlated to QUS parameters for Ach BUA (r = 0.350, P < 0.001), DTU BUA (r = 0.295, P < 0.001), UBIS BUA (r = 0.360, P < 0.001), QUS2 BUA (r = 0.266, P < 0.001), DTU SOS (r = 0.099, P = 0.037) and Ach Stiffness index (r = 0.145, P = 0.006). Weight was negatively correlated with IGEA Ad-SOS (r = −0.420, P < 0.001) and IGEA UBPI (r = −0.268, P < 0.001). Ach SOS, UBIS SOS and IGEA BTT did not show significant correlations with weight. The correlation between weight and BMD in the younger group was 0.343 for spine BMD and 0.463 for total hip BMD, both of which were highly significant (P < 0.001). In the older population, we found significant correlations for all QUS parameters except IGEA BTT. Positive correlations were found for Ach BUA (r = 0.325, P < 0.001), DTU BUA (r = 0.321, P < 0.001), UBIS BUA (r = 0.374, P < 0.001), QUS2 BUA (r = 0.330, P < 0.001), Ach SOS (r = 0.091, P < 0.001), DTU SOS (r = 0.123, P < 0.001), UBIS SOS (r = 0.070, P = 0.001) and Ach Stiffness index (r = 0.209, P < 0.001). Significant negative correlations were found for IGEA Ad-SOS (r = −0.122, P < 0.001) and IGEA UBPI (r = −0.057, P = 0.006). The correlation between weight and BMD in the older group was 0.331 for spine BMD and 0.456 for

total hip BMD, both of which were highly significant (P < 0.001). From the step-wise multiple regression, pulse rate was found to be significant in the models for SOS and the IGEA parameters. If we calculate correlation coefficients, the QUS parameters, except DTU-BUA, were highly inversely correlated with pulse rate (r = −0.047, P = 0.02 for UBIS BUA to r = −0.177, P < 0.001 for IGEA Ad-SOS) (Table 6). Age (r = 0.100, P < 0.001), height (r = −0.104, P < 0.001) and weight (r = 0.121, P < 0.001) are all highly correlated with pulse rate, therefore we used partial correlation coefficients to adjust for these parameters. Similar results are found with DTU-BUA being the only parameter not statistically significant (Table 6). No significant relationship between spine and total hip BMD was found with pulse rate. We found differences in prevalence of the 4 major risk factors (low body weight, previous fracture, maternal hip fracture and smoking) across the centres. The differences may relate to the differing attitudes towards smoking across Europe. Indeed, we did find a significant difference in smoking habits across Europe, with the UK having the highest number of smokers compared to Germany and France. When we just examined 3 common risk factors, excluding smoking, there is no longer a significant difference (data not shown), indicating that the major difference is being driven by the influence of smoking habits.

Table 6 Correlation coefficients and partial correlation coefficients between pulse rate and QUS (P value in parentheses) QUS parameter

BUA

SOS

Ad-SOS BTT UBPI Stiffness index a

Total population

Achilles DTU-one UBIS QUS2 Achilles DTU-one UBIS IGEA IGEA IGEA Achilles

Younger population only

Older population only

R value

R value for partial correlations a

R value for partial correlations a

R value for partial correlations a

−0.082 (<0.001) −0.029 (0.133) −0.047 (0.020) −0.077 (0.001) −0.171 (<0.001) −0.117 (<0.001) −0.140 (<0.001) −0.177 (<0.001) −0.144 (<0.001) −0.153 (<0.001) −0.141 (<0.001)

−0.065 (0.003) −0.030 (0.119) −0.052 (0.011) −0.072 (0.002) −0.134 (<0.001) −0.091 (<0.001) −0.118 (<0.001) −0.101 (<0.001) −0.080 (<0.001) −0.073 (<0.001) −0.112 (<0.001)

−0.048 (0.373) −0.001 (0.976) −0.056 (0.280) −0.088 (0.130) −0.174 (0.001) −0.073 (0.127) −0.110 (0.032) 0.051 (0.287) 0.031 (0.513) 0.074 (0.121) −0.132 (0.014)

−0.061 (0.010) −0.024 (0.256) −0.039 (0.075) −0.051 (0.050) −0.119 (<0.001) −0.085 (<0.001) −0.109 (<0.001) −0.110 (<0.001) −0.088 (<0.001) −0.086 (<0.001) −0.100 (<0.001)

Partial correlation coefficients adjusted for age, height and weight at baseline.

A. Stewart et al. / Bone 39 (2006) 609–615

Geographic differences for risk factors by centre We examined risk factors individually and found that for previous fracture there was a significant difference (Chi-square P = 0.05) across the centres, with PAR showing the lowest fracture rate with 34.5%, BER 38.7%, KIE 39.2%, ABE 39.8% and SHE with the highest rate of 43.9%. Report of maternal hip fracture also differed significantly across the centres (Chi-square P = 0.007), with BER showing the lowest prevalence (6.7%) while PAR showing the highest prevalence (12.1%). The other centres had intermediate prevalences of maternal hip fracture, i.e. SHE 6.9%, KIE 8.1% and ABE 10.4%. Weight was highly significant (ANOVA P < 0.001) across the centres. SHE showed the highest mean weight (mean = 70.2 kg, SD 13.6), while PAR were the lightest in weight (mean = 62.8, SD 11.5). The other 3 centres were similar: ABE 68.1 (SD 12.3), BER 69.2 (SD 12.0), KIE 69.4 (SD 11.8). Since there was a difference across the centres, the quartiles of weight were calculated per centre. Smoking differed significantly across the centres also (Chisquare P < 0.001), with PAR showing the lowest prevalence of ever smoking (28.3%), while SHE showed the highest prevalence (53.5%). The other 3 centres showed intermediate prevalence rates: ABE 46.4%, BER 43.6% and KIE 43.6%. Discussion The step-wise regression models showed that between 18% and 42% of the variance can be accounted for by the various risk factors. The most common risk factors are age, self-reported diagnosis of rheumatoid arthritis, use of hormone replacement therapy, current weight and self-reported diagnosis of osteoporosis. Since bone measurements show an age-related loss, it is no surprise to find that age is an important determinant. Rheumatoid arthritis has been shown previously to be associated with bone loss, particularly in periarticular regions [5–10]. HRT is known to increase bone mineral density and has been shown previously to also increase QUS parameters [11– 13]. Weight has also been shown to be highly associated with QUS and is discussed later. In our study, we found an inverse relationship between pulse rate and QUS parameters. Pulse rate has been found to be associated with fracture in a previous study [1,14] and has been associated with bone mineral density [14] but not QUS previously. It has been suggested that pulse rate is a marker for general poor health, overall physical fitness or is indicating undiagnosed hyperthyroidism [1,14]. When we examined the relationship between QUS and risk factors, we found differences between the younger and older cohorts. For previous personal fracture, all the QUS parameters were significantly reduced in those who fractured in the older cohort, but in the younger cohort, we found that only a small number of the parameters were significant. It is not surprising to find an association between all QUS parameters and previous fractures in the older cohort as QUS has been strongly associated with fracture many times previously [15–19], particularly in the

613

elderly [20–22]. We did not find a significant association in the younger cohort between BUA and fractures perhaps due to a smaller percentage of individuals reported having sustained a fracture previously and these fractures would not be at classic osteoporotic sites since the cohort has not reached the age at which we see “osteoporotic” fractures. Smoking has been negatively associated with BMD or fractures in many studies [23–25], but its association with QUS is less clear. Two studies have previously shown a strong negative association with smoking and QUS [26,27], but other studies using a variety of QUS devices showed no significant association [28–30]. In our study, the younger cohort showed very little relationship between smoking and QUS, whereas the older cohort shows strong relationships, although this may be partly mediated through smokers tending to be lighter in weight. If we compare our results to those previously published, there are similarities. In the study [26] which does show a significant relationship with smoking, the age group is between 49 and 76 years, all the women are postmenopausal and the study numbers are large (n = 6592). This group equates to our older cohort, which did show significant relationships between smoking and QUS. Lin et al.'s study [27] had over 9000 women, but they were from a large age range of 14 to 92 years. Two of the other studies [28,29] which did not show a significant relationship had relatively small numbers (393 and 200 women respectively) and consisting of mainly premenopausal women or had a large age range of women included, but the majority were under 45 years of age. Therefore, these studies equate more to our younger cohort, which also show no significant relationship between QUS and smoking. There have been mixed results previously relating weight and QUS parameters. In some studies [29,31], BUA has been significantly associated with weight, while in other studies, no significant association was found [28]. Similarly, it is not clear from previous studies the relationship between SOS and weight [28,31]. Weight appeared to have similar relationships to BUA in both our younger and older cohorts. The relationship between BUA and weight was strong regardless of age and which QUS device it was measured on. However, the relationship between SOS and weight is not so clear. In the younger cohort, SOS as measured by the heel devices showed no relationship with weight, however, amplitude-dependent SOS as measured at the phalanges is significantly related to weight. This is unusual since the phalanges are not a weight-bearing site, and we might expect non-weight-bearing sites not to be related to weight, but at the phalanges, it may be the amount of soft tissue which is having an influence. There was a significant relationship in the older population regardless of site/technique used. Maternal hip fracture was only analysed in the older cohort since the prevalence in the younger cohort was very low. This is probably due to the age of the cohort, meaning that their mothers are still too young to have reached the age when most hip fractures occur. When examining the older cohort, we found no significant relationship between QUS and report of maternal hip fracture. This is contrary to what has been shown previously with maternal (or grand maternal) hip fracture associated with low BMD [1,32] and fracture [33–37]. However, not all studies show

614

A. Stewart et al. / Bone 39 (2006) 609–615

the same results, with some agreeing with our current study. A study investigating European men and women selected in a similar way to our population found an association in with vertebral fracture in men, but no such association was found in women [38], and in a study conducted in Finland similarly, no association with maternal hip fracture and fracture was found [39]. In conclusion, there does appear to be similar relationships generally between QUS and the 4 major risk factors found, in previous studies, associated with future hip fracture. However, the relationship for the individual risk factors depends on the age group that is measured and on the type of QUS device used as well as the parameter measured.

[9]

[10]

[11]

[12]

[13]

Acknowledgments We would like to acknowledge the contributions of the other members of the OPUS teams at the five participating centres: Rosie Reid, Lana Gibson in Aberdeen; Antonia Gerwinn, Dr. Maren Glüer, Roswitha John, Roswitha Marunde-Ott, Monika Mohr, Regina Schlenger, Pia Zschoche, Dr. Reinhard Barkmann, Dr. Carsten Liess, Carsten Rose and Wolfram Timm in Kiel; Therese Kolta, Nathalie Delfau in Paris; Margaret Paggiosi, Nicky Peel, Diane Shutt, Anne Stapleton, Debbie Swindell in Sheffield. This project would not have been possible without the financial support of our sponsors: Aventis, Eli Lilly, Novartis, Procter and Gamble Pharmaceuticals, and Roche. We also would like to thank for the support of the equipment manufacturers: DMS, IGEA, OSI/Osteometer Meditech, Quidel/Metra. Alison Stewart is an Arthritis Research Campaign (arc) Non-Clinical Career Development Fellow. References [1] Cummings SR, Nevitt MC, Browner WS, et al, Study of Osteoporotic Fractures Research Group. Risk factors for hip fracture in white women. N Engl J Med 1995;332(12):767–73. [2] Stewart A, Calder LD, Torgerson DJ, et al. Prevalence of hip fracture risk factors in women aged 70 years and over. Q J Med 2000;93:677–80. [3] Gluer CC, Eastell R, Reid DM, et al. Association of five quantitative ultrasound devices and bone densitometry with osteoporotic vertebral fractures in a population-based sample: the OPUS Study. J Bone Miner Res 2004;19(5):782–93. [4] O'Neill TW, Cooper C, Cannata JB, et al. Reproducibility of a questionnaire on risk factors for osteoporosis in a multicentre prevalence study: the European Vertebral osteoporosis study. Int J Epidemiol 1994;23:559–65. [5] Alenfeld FE, Diessel E, Brezger M, Sieper J, Felsenberg D, Braun J. Detailed analyses of periarticular osteoporosis in rheumatoid arthritis. Osteoporos Int 2000;11(5):400–7. [6] Roben P, Barkmann R, Ullrich S, Gause A, Heller M, Gluer CC. Assessment of phalangeal bone loss in patients with rheumatoid arthritis by quantitative ultrasound. Ann Rheum Dis 2001;60 (7):670–7. [7] Ardicoglu O, Ozgocmen S, Kamanli A, Pekkutucu I. Relationship between bone mineral density and radiologic scores of hands in rheumatoid arthritis. J Clin Densitom 2001;4(3):263–9. [8] Haugeberg G, Rstavik ORE, Uhlig T, Falch JA, Halse JI, Kvien TK. Comparison of ultrasound and X-ray absorptiometry bone measurements in a case control study of female rheumatoid arthritis patients and

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22] [23]

[24]

[25]

[26]

[27]

[28]

[29]

randomly selected subjects in the population. Osteoporos Int 2003;14 (4):312–9. Birkett V, Ring EF, Elvins DM, Taylor G, Bhalla AK. A comparison of bone loss in early and late rheumatoid arthritis using quantitative phalangeal ultrasound. Clin Rheumatol 2003;22(3):203–7. Forslind K, Keller C, Svensson B, Hafstrom I. Reduced bone mineral density in early rheumatoid arthritis is associated with radiological joint damage at baseline and after 2 years in women. J Rheumatol 2003;30 (12):2590–6. Weiss M, Ben Shlomo A, Hagag P, Rapoport M, Ish-Shalom S. Effect of estrogen replacement therapy on speed of sound at multiple skeletal sites. Maturitas 2000;35(3):237–43. Balikian P, Burbank K, Houde J, et al. Bone mineral density and broadband ultrasound attenuation with estrogen treatment of postmenopausal women. J Clin Densitom 1998;1(1):19–26. Knapp KM, Blake GM, Spector TD, Fogelman I. Differential effects of hormone replacement therapy on bone mineral density and axial transmission ultrasound measurements in cortical bone. Osteoporos Int 2003;14(4):289–94. Kado DM, Lui LL, Cummings SR, the Study of Osteoporotic Fractures Research Group. Rapid resting heart rate: a simple and powerful predictor of osteoporotic fractures and mortality in older women. J Am Geriatr Soc 2002;50(3):455–60. Stewart A, Torgerson DJ, Reid DM. Prediction of fractures in perimenopausal women: a comparison of dual energy X-ray absorptiometry (DXA) and broadband ultrasound attenuation (BUA). Ann Rheum Dis 1996;55:140–2. Gluer CC, Cummings SR, Bauer DC, Stone K, Pressman A, Genant HK. Associations between quantitative ultrasound and recent fractures. J Bone Miner Res 1994;9(Suppl 10):S153. Gluer CC, Cummings SR, Bauer DC, et al. Osteoporosis: association of recent fractures with quantitative ultrasound findings. Radiology 1996;199. Thompson P, Taylor J, Fisher A, Oliver R. Quantitative heel ultrasound in 3180 women between 45 and 75 years of age: compliance, normal ranges and relationship to fracture history. Osteoporos Int 1998;8:211–4. Greenspan SL, Bouxsein ML, Melton ME, et al. Precision and discriminatory ability of calcaneal bone assessment technologies. J Bone Miner Res 1997;12(8):1303–13. Hans D, Dargent-Molina P, Schott AM, et al. Ultrasonographic heel measurements to predict hip fracture in elderly women: the EPIDOS prospective study. Lancet 1996;348:511–4. Porter RW, Johnson K, McCutchan JDS. Wrist fracture, heel bone density and thoracic kyphosis: a case control study. Bone 1990;11:211–4. Porter RW, Miller CG, Grainger D, Palmer SB. Prediction of hip fracture in elderly women: a prospective study. Br Med J 1990;301:638–41. Williams AR, Weiss NS, Ure CL, Ballard J, Daling JR. Effect of weight smoking and estrogen use on the risk of hip and forearm fractures in postmenopausal women. Obstet Gynecol 1982;60(6):695–9. Torgerson DJ, Campbell MK, Reid DM. Lifestyle, environmental and medical factors influencing peak bone mass in women. Br J Rheumatol 1995;34. Baron JA, Farahmand BY, Weiderpass E, et al. Cigarette smoking, alcohol consumption, and risk of hip fracture in women. Arch Intern Med 2001;161(7):983–8. Adami S, Giannini S, Giorgino R, et al. The effect of age, weight, and lifestyle factors on calcaneal quantitative ultrasound: the ESOPO study. Osteoporos Int 2003;14(3):198–207. Lin JD, Chen JF, Chang HY, Ho C. Evaluation of bone mineral density by quantitative ultrasound of bone in 16,862 subjects during routine health examination. Br J Radiol 2001;74(883):602–6. Gregg EW, Kriska AM, Salamone LM, et al. Correlates of quantitative ultrasound in the Women's healthy lifestyle project. Osteoporos Int 1999;10(5):416–24. Cheng S, Fan B, Wang L, et al. Factors affecting broadband ultrasound attenuation results of the calcaneus using a gel-coupled quantitative ultrasound scanning system. Osteoporos Int 1999;10(6):495–504.

A. Stewart et al. / Bone 39 (2006) 609–615 [30] Frost ML, Blake GM, Fogelman I. Quantitative ultrasound and bone mineral density are equally strongly associated with risk factors for osteoporosis. J Bone Miner Res 2001;16(2):406–16. [31] Hans D, Schott AM, Arlot ME, Sornay E, Delmas PD, Meunier PJ. Influence of anthropometric parameters on ultrasound measurements of os calcis. Osteoporos Int 1995;5:371–6. [32] Barros ER, Kasamatsu TS, Ramalho AC, Hauache OM, Vieira JG, Lazaretti-Castro M. Bone mineral density in young women of the city of Sao Paulo, Brazil: correlation with both collagen type I alpha 1 gene polymorphism and clinical aspects. Braz J Med Biol Res 2002;35 (8):885–93. [33] Fox KM, Cummings SR, Powell-Threets K, Stone K, Study of Osteoporotic Fractures Research Group. Family history and risk of osteoporotic fracture. Osteoporos Int 1998;8(6):557–62. [34] Lee SH, Dargent-Molina P, Breart G. Risk factors for fractures of the proximal humerus: results from the EPIDOS prospective study. J Bone Miner Res 2002;17(5):817–25.

615

[35] Torgerson DJ, Campbell MK, Thomas RE, Reid DM. Prediction of perimenopausal fractures by bone mineral density and other risk factors. J Bone Miner Res 1996;11:293–7. [36] Lopez-Rodriguez F, Mezquita-Raya P, Luna Jde D, Escobar-Jimenez F, Munoz-Torres M. Performance of quantitative ultrasound in the discrimination of prevalent osteoporotic fractures in a bone metabolic unit. Bone 2003;32(5):571–8. [37] Albrand G, Munoz F, Sornay-Rendu E, DuBoeuf F, Delmas PD. Independent predictors of all osteoporosis-related fractures in healthy postmenopausal women: the OFELY study. Bone 2003;32 (1):78–85. [38] Diaz MN, O'Neill TW, Silman AJ. The influence of family history of hip fracture on the risk of vertebral deformity in men and women: the European Vertebral Osteoporosis Study. Bone 1997;20(2):145–9. [39] Huopio J, Kroger H, Honkanen R, Saarikoski S, Alhava E. Risk factors for perimenopausal fractures: a prospective study. Osteoporos Int 2000; 11:219–27.