Quantitative Ultrasound and Bone Mineral Density Are Equally Strongly Associated with Risk Factors for Osteoporosis

JOURNAL OF BONE AND MINERAL RESEARCH Volume 16, Number 2, 2001 2001 American Society for Bone and Mineral Research Quantitative Ultrasound and Bone Mineral Density Are Equally Strongly Associated with Risk Factors for Osteoporosis M.L. FROST, G.M. BLAKE, and I. FOGELMAN ABSTRACT Because resources do not allow all women to be screened for osteoporosis, clinical risk factors are often used to identify those individuals at increased risk of fracture who are then assessed by bone densitometry. The aim of this study was to compare calcaneal quantitative ultrasound (QUS) and axial bone mineral density (BMD) T and Z scores in a large group of women, some with no clinical risk factors and others with one or more risk factors for osteoporosis. The study population consisted of 1115 pre- and postmenopausal women. A subgroup of 530 women was used to construct reference data for calculating T and Z scores. A total of 786 women was found to have one or more of the following risk factors: (i) atraumatic fracture since the age of 25 years, (ii) report of X-ray osteopenia, (iii) predisposing medical condition or use of therapy known to affect bone metabolism, (iv) premature menopause before the age of 45 years or a history of amenorrhea of longer than 6 months duration, (v) family history of osteoporosis, (vi) body mass index (BMI) <20 kg/m 2, and (vii) current smoking habit. Calcaneal broadband ultrasound attenuation (BUA) and speed of sound (SOS) measurements were performed on a Hologic Sahara and a DTUone and BMD was measured at the spine and hip using dual-energy X-ray absorptiometry (DXA). The Z score decrements associated with the seven risk factors calculated using multivariate regression analysis were similar for QUS and BMD measurements. Z score decrements (mean of BMD and QUS measurements combined) associated with a history of atraumatic fracture ( 0.67), X-ray osteopenia ( 0.36), a family history of osteoporosis ( 0.23), and a low BMI ( 0.53) were all statistically significant compared with women with no risk factors. Z score decrements associated with a medical condition or use of therapy known to affect bone metabolism, a premature menopause or prolonged amenorrhea, or those who were current smokers were not significantly different from zero. As the number of risk factors present in each individual increased, the mean Z score decrements became more negative, increasing from 0.28 for women with one risk factor to 1.19 for those with four or more risk factors. QUS and BMD measurements yielded similar mean Z scores for women with one, two, three, or more than four risk factors. Using the World Health Organization (WHO) criteria to diagnose osteoporosis for BMD measurements and revised diagnostic criteria for QUS, approximately one-third of postmenopausal women aged 50 years with clinical risk factors were classified as osteoporotic compared with only 12% of women without clinical risk factors. Over two-thirds of postmenopausal women with risk factors were classified as osteopenic or osteoporotic and approximately 28% were classified as normal. The proportion of women classified into each diagnostic category was similar for BMD and QUS. In conclusion, clinical risk factors for osteoporosis affected calcaneal BUA and SOS Z score measurements to the same extent as axial BMD Z score measurements. Provided revised diagnostic criteria are adopted for QUS, similar proportions of postmenopausal women are identified as osteopenic or osteoporotic as with BMD. (J Bone Miner Res 2001;16:406 416) Key words: quantitative ultrasound, osteoporosis, risk factors, Z scores, T scores Osteoporosis Screening and Research Unit, Guy s Hospital, London, United Kingdom. 406

QUS AND RISK FACTORS FOR OSTEOPOROSIS INTRODUCTION OSTEOPOROSIS IS the most common generalized disease of the skeleton, and with the progressive aging of the world s population, the proportion of women with osteoporosis is expected to increase further. The two major risk factors for an osteoporotic fracture are age and bone mineral density (BMD). (1,2) Several large prospective studies have confirmed that for each SD decline in BMD there is an approximate doubling of fracture risk. (1 5) Because of the strong relationship between low BMD and risk of fracture and the fact that no symptoms occur before fracture, a World Health Organization (WHO) Working Party defined osteoporosis as a BMD value at the spine, hip, or forearm 2.5 or more SDs below the young adult mean, with or without the presence of a fragility fracture. (6) The most widely used technique to measure BMD is by dual X-ray absorptiometry (DXA). However, because of the increasing demand for bone densitometry services, alternative cheaper technologies such as quantitative ultrasound (QUS) have been developed over recent years. Because of the pioneering work of Langton and colleagues in 1984, (7) scientific and clinical interest has been directed at QUS as an alternative method to DXA for assessing skeletal status. A large number of studies published in the past decade have examined the utility of QUS and its potential role in the field of osteoporosis. Large prospective studies have confirmed that both broadband ultrasound attenuation (BUA) and speed of sound (SOS) measurements at the calcaneus can identify those individuals at risk of osteoporotic fracture as reliably as BMD. (8,9) Many of the risk factors associated with osteoporosis, such as estrogen deficiency and glucocorticoid use, are associated with reduced BMD. At present, resources do not permit all women to be screened for osteoporosis and therefore clinical risk factors for low BMD are used to target individuals for bone densitometry and possible therapeutic intervention. This case-finding strategy was recommended in the recent Royal College of Physicians (RCP) guidelines for the prevention and treatment of osteoporosis. (10) A recent study of 3530 women, investigating the relationship between spine and femur BMD results and the reason for referral for a DXA scan, confirmed that certain indications such as premature menopause, secondary amenorrhea, and X-ray osteopenia were associated with low BMD and an increased prevalence of osteoporosis. (11) A similar finding also was observed in another study of 929 men and women, which reported that X-ray osteopenia, glucocorticoid therapy, and previous fractures were associated with low BMD. (12) Several studies have shown that both BUA and SOS also are decreased in individuals with risk factors for osteoporosis, such as primary hyperparathyroidism, (13 15) Crohn s disease, (16) anorexia, (17) kidney disease, (18) and glucocorticoid use. (19) However, the majority of these studies have been small. The aim of this study was to compare calcaneal QUS measurements and DXA BMD measurements in a large group of women, some with no clinical risk factors and others with one or more risk factors for osteoporosis. The objective was to show that QUS measurements respond as effectively as BMD measurements to the presence of the risk factors commonly used by general practitioners (GPs) to refer patients for bone densitometry studies. The prevalence of osteoporosis diagnosed using BMD and QUS measurements in women with and without clinical risk factors was also examined. MATERIALS AND METHODS Subjects The study population consisted of 1115 women recruited from two sources: (i) patients referred by their GP or hospital consultant for routine bone density screening by DXA and (ii) women from the general population who volunteered to participate in clinical research. All women completed a self-administered questionnaire. Out of the 1115 patients studied, a group of 530 women was used to construct reference ranges for both BMD and QUS measurement variables. Exclusion criteria for the reference data were an atraumatic fracture after the age of 25 years, a premature menopause before the age of 40 years, amenorrhea of longer than 6 months duration, and a predisposed medical condition or therapy known to affect bone metabolism. These criteria were the same as those previously described by Ryan et al. for the construction of BMD reference ranges. (20) On examination of the GP referral letter and self-administered questionnaire, 786 women were found to have one or more of the following clinical risk factors: (i) an atraumatic fracture since the age of 25 years, (ii) a report of X-ray osteopenia (patients were included in this category if there was a radiologist s report of X-ray osteopenia in the patient s hospital notes or mentioned in the GP referral letter), (iii) predisposing medical condition or use of therapy known to affect bone metabolism, (iv) a premature menopause before the age of 45 years or a history of amenorrhea of longer than 6 months duration, (v) a family history of osteoporosis, (vi) a body mass index (BMI) 20 kg/m 2, or (vii) a current smoking habit. These risk factors were chosen because they were the main seven risk factors listed in one or more of the following published guidelines for identifying individuals at risk of fracture: the RCP, (10) the European Foundation of Osteoporosis and Bone Disease, (21) and the National Osteoporosis Foundation (22) guidelines. Some women with risk factors were included in the group used to construct reference ranges because women with X-ray osteopenia, a family history of osteoporosis, low BMI, or women who currently smoked were not excluded from the reference population. Of the 1115 women, 329 pre- and postmenopausal women did not have any of the seven risk factors described previously. Women who were past or current users of estrogen or bisphosphonate therapy were not excluded from analysis. Written informed consent was obtained from all study participants and the study was approved by the Guy s Hospital Research Ethics Committee. Measurements 407 Subjects had QUS measurements at the calcaneus and DXA measurements at the lumbar spine and hip. BMD

408 FROST ET AL. measurements of the lumbar spine (L1 L4), femoral neck, and total hip were performed using a Hologic QDR4500 (Hologic, Bedford, MA, USA). QUS measurements were performed on a Hologic Sahara and DTUone (Osteometer MediTech, Hawthorne, CA, USA). However, not all women had measurements on the DXA device and both QUS devices. Hologic Sahara The Sahara Clinical Bone Sonometer (Hologic) consists of two unfocused transducers mounted coaxially on a motorized caliper. One transducer acts as a transmitter and the other as a receiver. The transducers are coupled acoustically to the heel using soft rubber pads and an oil-based coupling gel. The Sahara device measures both BUA and SOS at a fixed region of interest in the midcalcaneus and the results are combined to provide an estimate of heel BMD (Est.heel BMD) with units of grams per square centimeter using the following equation: Est.heel BMD 0.002592* BUA SOS 3.687 g/cm 2 (1). It is important to note that Est.heel BMD is inferred from a linear combination of BUA and SOS and is not an actual measurement of calcaneal BMD. Osteometer DTUone The DTUone (Osteometer MediTech) consists of two focused transducers mounted coaxially in a water bath containing a surfactant. A rectilinear scan of the calcaneus is performed yielding an image size of approximately 60 mm 80 mm with a pixel size of 0.5 mm. Both BUA and SOS are calculated at each pixel and an automatic region of interest is selected in an area with a local minimum of attenuation, located in the posterior tuberosity of the calcaneus. Data analysis Data for the study population were combined as means and SDs. The pre- and postmenopausal groups with and without risk factors were compared statistically using the Student s t-test. The values of p 0.05 were considered significant. Women were considered treated if they were past or current users of estrogen or bisphosphonates for longer than 12 months. The group of 530 premenopausal and postmenopausal women defined using the criteria of Ryan et al. (20) was used to construct local normal ranges for both BMD and QUS measurement variables for the purpose of calculating Z scores: Z score measurement value age-matched mean. (2) age-matched population SD From the group of 530 women a subgroup of 150 premenopausal women aged 20 40 years was used to estimate the young adult mean and SD for the purpose of calculating T scores: T score measurement value young adult mean. (3) young adult population SD Multivariate regression analysis was used to calculate the Z score decrements associated with each of the seven clinical risk factors using the following equation: Z score b 0 b i RF i, (4) i where b 0 is the constant and represents the mean Z score for women with no clinical risk factors, the regression coefficient b i for the ith independent variable (RF i ) represents the mean Z score decrement associated with that risk factor, and RF i equals one if the risk factor is present and zero otherwise. The number of risk factors present for each individual was calculated. The risk factors used were those stated previously and therefore an individual could have a maximum of seven possible risk factors for osteoporosis. The mean T and Z scores for women with one, two, three, or more than four risk factors were calculated. A two-sample Student s t-test was used to assess whether the mean T and Z scores were significantly different from those obtained for women with no clinical risk factors. Using spine and femur BMD T scores women were classified into three groups as defined by the WHO Working Party (6) : Normal. A T score 1 Osteopenia (low bone mass). A T score 1 but 2.5 Osteoporosis. A T score 2.5 Because the WHO criteria are not applicable to QUS, women were also classified into three groups using revised criteria for QUS that have been shown to be applicable to the Sahara and DTUone devices (23) : Normal. A T score 0.5 Osteopenia (low bone mass). A T score 0.5 but 1.8 Osteoporosis. A T score 1.8 The percentage of women classified as normal, osteopenic, or osteoporotic was calculated for women with and without clinical risk factors, using the WHO criteria for BMD and the revised criteria for QUS. RESULTS Patient characteristics Patient characteristics for the pre- and postmenopausal women with and without clinical risk factors are shown in Table 1. Both pre- and postmenopausal women with clinical risk factors for osteoporosis had significantly lower BMD and QUS variables compared with women without risk factors. All the BMD and QUS measurement variables were significantly lower in postmenopausal women compared

QUS AND RISK FACTORS FOR OSTEOPOROSIS 409 TABLE 1. PATIENT CHARACTERISTICS Premenopausal with risk factors Premenopausal without risk factors Postmenopausal with risk factors Postmenopausal without risk factors n 237 183 549 146 Age (years) 40.0 (8.5)* 38.3 (9.1) 60.9 (9.3) 59.9 (8.4) YSM (years) N/A N/A 14.4 (9.4) 10.2 (8.2) Height (cm) 164 (6.3) 165 (6.0) 160 (6.9) 161 (5.9) Weight (kg) 63.7 (12.9)* 66.9 (9.9) 64.8 (11.8) 66.1 (9.8) BMI (kg/m 2 ) 23.6 (4.7)* 24.7 (3.6) 25.2 (4.3) 25.5 (3.5) Lumbar spine BMD (g/cm 2 ) 1.025 (0.120)* 1.064 (0.112) 0.871 (0.164) 0.946 (0.145) Femoral neck BMD (g/cm 2 ) 0.838 (0.116)* 0.881 (0.114) 0.706 (0.131) 0.773 (0.126) Total hip BMD (g/cm 2 ) 0.937 (0.115)* 0.974 (0.099) 0.813 (0.150) 0.897 (0.127) Sahara BUA (db/mhz) 73.8 (13.9)* 77.7 (13.817) 63.7 (16.7) 72.8 (13.9) Sahara SOS (m/s) 1553.2 (29.4)* 1560.8 (29.9) 1531.2 (31.8) 1548.0 (27.9) Sahara BMD (g/cm 2 ) 0.531 (0.108)* 0.561 (0.108) 0.448 (0.121) 0.515 (0.103) DTU BUA (db/mhz) 52.5 (6.2)* 54.8 (6.0) 47.9 (9.0) 52.3 (6.2) DTU SOS (m/s) 1549.8 (11.3)* 1552.6 (11.0) 1542.7 (11.4) 1548.7 (9.5) % Treated a 6.30% 0% 35.70% 0% N/A, not applicable; YSM, years since menopause. a Women were considered treated if they were currently taking or had previously taken estrogen or bisphosphonates for longer than 12 months. * p 0.05 versus premenopausal women without risk factors; p 0.05 versus postmenopausal women without risk factors. TABLE 2. PATIENT CHARACTERISTICS AND NUMBER OF SUBJECTS FOR EACH CLINICAL RISK FACTOR Risk factor Age mean (SD) % Treated Number of subjects (% of total) Total DXA SAH DTU History of atraumatic fracture 62.8 (11.1) 34.7 236 214 (90.7) 221 (93.6) 178 (75.4) X-ray osteopenia 65.8 (10.1) 41.5 53 53 (100.0) 52 (98.1) 49 (92.5) Predisposing medical condition or therapy use 57.4 (13.4) 38.2 225 183 (81.3) 190 (84.4) 158 (70.2) Premature menopause or prolonged amenorrhea 55.3 (12.2) 31.9 254 216 (85.0) 232 (91.3) 189 (74.4) Family history of osteoporosis 54.0 (11.9) 24.0 333 331 (99.4) 318 (95.5) 268 (80.5) BMI 20 49.3 (14.9) 22.6 106 75 (70.8) 90 (84.9) 63 (59.4) Current smoking 53.0 (13.5) 22.6 146 146 (100.0) 138 (94.5) 113 (77.4) No clinical risk factors 47.9 (13.8) 0 329 246 (74.8) 277 (84.2) 203 (61.7) SAH, Hologic Sahara. with the premenopausal women. Six percent of the premenopausal women and 36% of the postmenopausal women were past or current users of estrogen or bisphosphonate therapy. The mean length of estrogen and bisphosphonate use was 5.0 years and 2.6 years, respectively. Table 2 shows the patient characteristics for subjects with each of the seven clinical risk factors and the number of subjects measured by DXA and the two QUS devices. Women with a history of atraumatic fracture and X-ray osteopenia were older compared with women with other clinical risk factors. Women with a BMI 20 kg/m 2 and women with no clinical risk factors were younger compared with women with other clinical risk factors. Approximately one-third of the women with a history of atraumatic fracture, X-ray osteopenia, a medical condition, or use of therapy known to affect bone metabolism and women who had a premature menopause or prolonged amenorrhea were past or current users of estrogen or bisphosphonate therapy. Approximately 25% of the women with a family history of osteoporosis, a BMI 20 kg/m 2, and current smokers were past or current users of estrogen or bisphosphonate therapy. Table 3 shows the local reference ranges for both BMD and QUS measurement variables. These reference data were used to calculate T and Z scores. The BMD and QUS variables peaked in the third and fourth decade of life and declined thereafter. Lumbar spine BMD appeared to increase in women after the age of 70 years, which probably was because of the well-documented effects of osteophytes and aortic calcification on lumbar spine BMD and therefore these data were not used in the calculation of lumbar spine Z scores. Z scores Table 4 shows the Z score decrements associated with each of the different clinical risk factors calculated from the regression coefficients b i in Eq. (4). Also included in Table

410 FROST ET AL. TABLE 3. REFERENCE DATA FOR BMD AND QUS MEASUREMENT VARIABLES Age n LS BMD FN BMD THIP BMD SAH BUA SAH SOS SAH BMD DTU BUA DTU SOS 20 40 a 150 1.062 (0.117) 0.893 (0.113) 0.988 (0.099) 77.2 (13.0) 1561.8 (26.5) 0.562 (0.097) 54.0 (6.0) 1553.4 (8.4) 20 24 21 1.059 (0.125) 0.914 (0.172) 1.003 (0.085) 74.7 (10.9) 1557.3 (28.0) 0.544 (0.093) 53.7 (7.1) 1552.7 (11.9) 25 29 41 1.062 (0.107) 0.914 (0.099) 0.993 (0.099) 75.1 (10.2) 1561.5 (23.4) 0.556 (0.081) 53.4 (4.7) 1554.5 (4.8) 30 34 35 1.060 (0.142) 0.901 (0.104) 0.987 (0.117) 79.4 (15.0) 1562.0 (26.2) 0.568 (0.102) 54.3 (6.7) 1555.7 (9.0) 35 39 43 1.063 (0.118) 0.859 (0.111) 0.975 (0.110) 76.9 (14.4) 1561.7 (29.4) 0.561 (0.109) 54.4 (6.6) 1550.6 (8.7) 40 44 47 1.069 (0.124) 0.869 (0.148) 0.951 (0.138) 76.8 (13.5) 1558.7 (32.7) 0.553 (0.115) 52.5 (6.4) 1550.0 (11.6) 45 49 95 1.021 (0.102) 0.833 (0.091) 0.936 (0.092) 77.5 (15.7) 1556.5 (30.0) 0.549 (0.115) 53.9 (6.0) 1549.5 (10.6) 50 54 88 0.975 (0.145) 0.810 (0.125) 0.923 (0.121) 72.6 (12.8) 1548.1 (28.5) 0.515 (0.104) 53.1 (6.9) 1550.3 (10.7) 55 59 48 0.935 (0.135) 0.781 (0.095) 0.891 (0.094) 74.9 (14.9) 1549.6 (33.5) 0.525 (0.120) 53.6 (5.4) 1547.0 (11.2) 60 64 52 0.922 (0.152) 0.741 (0.108) 0.883 (0.127) 70.6 (12.0) 1547.1 (26.8) 0.507 (0.095) 51.5 (5.5) 1548.0 (10.7) 65 69 32 0.853 (0.151) 0.686 (0.125) 0.805 (0.127) 64.4 (13.5) 1532.6 (27.0) 0.454 (0.100) 51.4 (6.3) 1544.2 (7.4) 70 74 17 0.927 (0.161) 0.701 (0.111) 0.830 (0.105) 65.4 (13.7) 1537.0 (28.5) 0.467 (0.106) 47.4 (5.6) 1542.4 (7.1) 75 79 11 0.910 (0.167) 0.662 (0.113) 0.795 (0.119) 62.2 (15.3) 1532.0 (35.7) 0.446 (0.126) 47.7 (7.3) 1542.2 (15.7) Data are expressed as mean (SD). LS, lumbar spine; FN, femoral neck; THIP, total hip; SAH, Hologic Sahara. a Young adult group used to calculate T scores. TABLE 4. MEAN ZSCORE DECREMENTS a FOR EACH CLINICAL RISK FACTOR FOR OSTEOPOROSIS Z score decrements Mean Z score Risk factor History of fracture X-ray osteopenia Medical or therapy Early menoor amenorrhea Family history BMI 20 Current smoking No risk factors LS BMD 0.50 (0.08) 0.27 (0.15)* 0.02 (0.09) 0.23 (0.08)* 0.14 (0.07)* 0.50 (0.12) 0.06 (0.09) 0.15 (0.05) FN BMD 0.51 (0.08) 0.34 (0.13)* 0.03 (0.08) 0.03 (0.07) 0.23 (0.06) 0.68 (0.11) 0.01 (0.08) 0.08 (0.05) THIP BMD 0.75 (0.09) 0.56 (0.16) 0.03 (0.09) 0.06 (0.08) 0.19 (0.07)* 0.92 (0.13) 0.06 (0.10) 0.08 (0.05) SAH BUA 0.71 (0.08) 0.39 (0.15)* 0.04 (0.09) 0.14 (0.08) 0.25 (0.07) 0.48 (0.11) 0.07 (0.10) 0.10 (0.05) SAH SOS 0.66 (0.08) 0.33 (0.14)* 0.08 (0.08) 0.18 (0.07)* 0.25 (0.07) 0.21 (0.10)* 0.04 (0.09) 0.10 (0.05) SAH BMD 0.71 (0.08) 0.37 (0.15)* 0.07 (0.08) 0.18 (0.08)* 0.26 (0.07) 0.31 (0.11)* 0.05 (0.09) 0.11 (0.05) DTU BUA 0.85 (0.10) 0.48 (0.17)* 0.13 (0.10) 0.02 (0.09) 0.25 (0.08)* 0.87 (0.15) 0.09 (0.11) 0.09 (0.06) DTU SOS 0.68 (0.09) 0.16 (0.15) 0.09 (0.09) 0.02 (0.08) 0.28 (0.08) 0.24 (0.13)* 0.08 (0.10) 0.11 (0.06) LS, lumbar spine; FN, femoral neck; THIP, total hip; SAH, Hologic Sahara. a Each decrement represents the difference in Z scores between those with the specified risk factor and those with no clinical risk factors. * p 0.05 versus subjects with no clinical risk factors; p 0.001 versus subjects with no clinical risk factors. 4 are the mean Z scores for women with no clinical risk factors calculated from the constant b 0 in Eq. (4). Women with a history of fracture, X-ray osteopenia, a family history of osteoporosis, or a low BMI had significantly lower Z scores than women with no risk factors. Lumbar spine BMD and both Sahara SOS and Sahara BMD Z scores were significantly lower in women who reported an early menopause or prolonged amenorrhea. Women with predisposing medical conditions or use of therapy and women who currently smoke yielded essentially normal Z scores for both BMD and QUS measurement variables. The Z score decrements for each clinical risk factor were similar for BMD and QUS measurements, although total hip and DTU BUA Z scores were more negative for women with a history of atraumatic fracture, X-ray osteopenia, or a low BMI compared with lumbar spine and DTU SOS measurements, respectively. The number of risk factors present for each individual was estimated and the mean Z score for women with one, two, three, or more than four risk factors for osteoporosis was calculated (Fig. 1). All mean Z scores were significantly lower than the mean Z score for women with no clinical risk factors. As the number of risk factors present in each individual increased, the mean Z score offset compared with subjects with no risk factors became more negative, increasing from approximately 0.28 for women with one risk factor to 1.19 for women with four or more risk factors. BMD and QUS measurements yielded similar Z scores for women with one, two, three, or more than four risk factors. T scores The mean T scores calculated for women with one, two, three, or more than four risk factors for osteoporosis are

QUS AND RISK FACTORS FOR OSTEOPOROSIS 411 FIG. 1. The mean T scores for women with one, two, three, or more than four clinical risk factors for osteoporosis (error bars represent the SEM). shown in Fig. 2. The mean T score for women with no risk factors was approximately 0.58 and 0.22 for BMD and QUS measurements, respectively. All mean T scores for women with clinical risk factors were significantly lower than those obtained for women with no risk factors and became more negative as the number of risk factors present increased. Mean QUS T scores decreased from approximately 0.6 for women with one risk factor to approximately 2.0 for women with more than four risk factors. Corresponding figures for BMD were 1.1 and 2.7, respectively. In all cases QUS measurements yielded less negative T scores than BMD measurements, with the offset between BMD and QUS T scores being approximately 0.5. Prevalence of osteopenia and osteoporosis FIG. 2. The mean Z scores for women with one, two, three, or more than four clinical risk factors for osteoporosis (error bars represent the SEM). The percentage of women classified as normal, osteopenic, or osteoporotic was calculated for postmenopausal women aged 50 years with and without clinical risk factors (Figs. 3A and 3B). Approximately one-third of postmenopausal women with clinical risk factors were classified as osteoporotic using either BMD or QUS measurements, compared with 12% of women with no clinical risk factors. Approximately one-half of women with no clinical risk factors were classified as normal while over two-thirds of women with clinical risk factors were classified as osteopenic or osteoporotic. Provided the revised WHO criteria (21) are applied to the QUS measurements, the proportion of postmenopausal women classified as normal, osteopenic, or osteoporotic was similar when using either BMD or QUS measurements. DISCUSSION A recent survey by the United Kingdom National Osteoporosis Society revealed that only one-third of GPs had direct access to bone densitometry services and, in addition, two-thirds of patients with osteoporosis were not diagnosed until they had suffered a fracture. (24) This indicates that there is a need to increase the availability of bone densitometry services and increase awareness of how indications

412 FROST ET AL. FIG. 3. The prevalence of osteopenia and osteoporosis in (A) women with no clinical risk factors and (B) women with one or more clinical risk factors using BMD and QUS measurements. for osteoporosis can be used to select individuals at risk for future fracture. The recent guidelines for the prevention and treatment of osteoporosis by the RCP recommended that certain clinical risk factors should be used to identify individuals who would benefit from a bone density examination and possible prophylactic treatment. (10) QUS has been introduced as a novel method of assessing skeletal status and because of the advantages of low cost and absence of ionizing radiation it has become an alternative tool in identifying individuals at risk of fracture. It is clear from the evidence in the literature that DXA-based measurements of BMD are significantly lower in individuals with clinical risk factors for osteoporosis. (11,12) The aim of this study was to determine whether calcaneal QUS variables are affected by clinical risk factors to the same extent as axial BMD measurements obtained using DXA. This is particularly important if QUS is to be used alone without the need for additional densitometry examinations because it has been reported that this may be the most cost-effective approach of using QUS. (25) The mean T scores for both BMD and QUS measurements were significantly lower for women with one, two, three, or four plus clinical risk factors than those obtained for women with no clinical risk factors (Fig. 2). As the number of risk factors present in each individual increased, the mean T scores became more negative, partly because of older patients having more risk factors but also because of the effect of the number of risk factors on Z scores (Fig. 1). A comparison of the BMD and QUS mean T scores for each group revealed that the mean T scores for BMD measurements were more negative than those obtained using QUS measurements by approximately 0.5 T score units. This is consistent with studies that reported that the age-related decline in T scores for QUS measurements is approximately half that seen for BMD measurements performed by DXA. (23,26) The slower rate of age-related decline in T scores observed for QUS compared with DXA has implications when using the current WHO criteria to diagnose osteoporosis using QUS because very few individuals would be classified as osteoporotic. Revised diagnostic criteria were used for QUS in the present study as described by Frost et al., who used BMD-equivalent prevalence rates of osteoporosis. (23) A postmenopausal woman was classified as osteoporotic if she had a T score 1.80 and classified as normal if she had a T score 0.50. The prevalence of women classified as normal, osteopenic, and osteoporotic was similar for BMD and QUS measurements (Fig. 3). This indicates that the revised criteria used in the present study for QUS can be applied to Sahara and DTUone results with confidence, without the concern that a large percentage of women will be classified differently using either QUS or BMD measurements. However, these criteria may be device specific, and therefore modified criteria may be appropriate for other QUS devices. Although the revised diagnostic criteria for QUS was adequate in the present study, it has been suggested that population-based sampling studies of osteoporosis and fracture prevalence rates would best address the issue of establishing appropriate diagnostic criteria for QUS. (25) The use of Z scores removes the confounding factors of chronological age and the influence of differences in the age-related decline of T scores between different technologies. The Z score decrements compared with women with no clinical risk factors were highly statistically significant for women with a history of atraumatic fracture, evidence of X-ray osteopenia, a family history of osteoporosis, and a low BMI. Z score decrements for lumbar spine BMD and both Sahara SOS and Sahara BMD also were significantly reduced in women reporting a premature menopause or prolonged amenorrhea. When expressed as Z scores, the reductions in QUS measurements were similar to that observed for axial BMD measurements (Table 4). The mean Z score decrement obtained for women with a history of fracture or evidence of X-ray osteopenia for BMD and QUS measurements combined was 0.67 and 0.36, respectively. Many previous studies have reported that QUS measurements are significantly reduced in women with osteoporotic fractures. (27 29) In addition, several large prospective studies confirmed that QUS could identify women at risk of future fracture (8,9,30 33) and that the ability of QUS to estimate fracture risk was comparable with that seen for BMD measurements. (8,9,33) Therefore, women with a his-

QUS AND RISK FACTORS FOR OSTEOPOROSIS tory of atraumatic fracture would have an increased risk of future fracture compared with women with no clinical risk factors, regardless of whether QUS or BMD variables were assessed. As expected, the Z score decrement for spinal BMD was significantly reduced in women with X-ray osteopenia, as reported by Bainbridge and Eastell (12) and Ahmed et al. (11) Hip BMD and QUS measurements were also significantly reduced. To date, this is the first study to report significantly reduced BUA and SOS measurements in women with X-ray osteopenia, although this may not be surprising because it has been reported that X-ray osteopenia can only reliably be detected in patients with BMD 0.73 g/cm 2. (34) X-ray osteopenia requires prior assessment by radiography whereas a knowledge of the other risk factors can be acquired by relatively simple means. However, X-ray osteopenia was included as a clinical risk factor because it was mentioned in the RCP and European Foundation for Osteoporosis and Bone Disease guidelines, (10,21) and it was a frequent reason for referral of patients to our clinic for bone densitometry assessment. (11) Exclusion of women with X-ray osteopenia from multivariate regression analysis led to very small and nonsignificant decreases in mean Z score decrements of just 0.01 U. Clinical studies of BMD in families with osteoporosis have shown evidence of a genetic contribution to osteoporosis (35 37) and large prospective studies have shown that women who report a maternal history of hip fracture are at increased risk of fracture themselves. (1,38) It also has been shown that calcaneal QUS measurements are influenced by genetic factors. A classical twin study of 500 female twins yielded heritability estimates of 0.53 and 0.61 for BUA and SOS measurements, respectively. (35) A study of 207 mother-daughter pairs reported a heritability estimate of 0.53 for BUA in postmenopausal daughters, which was similar in magnitude to that obtained for calcaneal BMD. (36) A study of 93 mono- and dizygotic twins revealed that up to 80% of the variance in QUS was attributable to genetic factors, making QUS an equally heritable trait as BMD. (39) The study by Howard et al. also revealed that only about 30% of the genetic variance of QUS was attributable to genetic factors specific to femoral neck BMD, indicating that separate genetic factors influence QUS. (39) In the present study, the mean Z score decrement for women with a family history of osteoporosis averaged 0.26 for all the QUS measurements, which was similar to that obtained for BMD measurements. Women with a BMI 20 had significantly reduced BMD and QUS Z scores, which were of similar magnitude to those obtained for women with a history of atraumatic fracture. Body weight has been shown to be a strong determinant of BMD (40,41) and risk of fracture. (1) These facts, along with the evidence here, show that body weight should be considered an important factor when evaluating patients need for bone densitometry assessment. Women who reported an early menopause or prolonged amenorrhea, a medical condition or use of therapy known to affect bone metabolism, or a current smoking habit yielded essentially normal Z scores. This was a somewhat surprising finding because these clinical risk factors have been associated with low BMD. (11,12,42,43) Only lumbar spine BMD and Sahara QUS measurements were significantly reduced 413 in women reporting an early menopause or prolonged amenorrhea (Table 4). The reason for this finding may be explained by a combination of factors: women who had undergone hysterectomy without oophorectomy were included and it is not clear if this caused or accelerated ovarian failure; over two-thirds of women who had undergone both hysterectomy and oophorectomy were being treated (mean length of treatment 7.5 years); and, finally, the mean duration of amenorrhea was only 16 months. Of the women with a predisposing medical condition or use of therapy known to affect bone metabolism, 37% reported taking glucocorticoids, known to cause considerable decreases in bone mass, particularly at sites rich in trabecular bone. (44) On examination of this subgroup, 45% of the women were being treated with estrogen or bisphosphonates and 50% had a BMI 25 kg/m 2 (and were therefore classified as overweight or obese). These factors may explain, at least in part, why the Z score decrements for these women were not reduced to a greater extent. Although tobacco smoking is associated with reduced BMD (43) and a greater rate of postmenopausal bone loss, (45) BMD and QUS Z scores were not significantly reduced. On examination of the data, BMD and QUS measurements showed a negative correlation with the number of years spent as a smoker (r 0.10), indicating that the duration of smoking may be more important in terms of the effects on bone mass. The exclusion of women with those risk factors that were not associated with significantly reduced mean Z scores did not significantly change the mean T and Z score decrements obtained for women with other clinical risk factors. Although the Z scores for the women referred to previously were not significantly reduced, the important fact is that Z scores for QUS measurements were almost identical to those obtained for BMD measurements. This indicates that both are affected to the same magnitude by clinical risk factors for osteoporosis. Women with multiple risk factors for osteoporosis had significantly lower BMD and QUS measurements than women with no or only one risk factor (Fig. 1). This finding also has been observed by Cummings et al. who reported that a large proportion of women with more than five risk factors for hip fracture had calcaneal BMD in the lowest third for their age and these women also had a greater rate of hip fracture. (1) It would appear that these clinical risk factors have an additive effect on bone mass whereby, as the number of risk factors present in each individual increases, BMD progressively decreases. This is consistent with our statistical model [Eq. (4)], which assumes that the effects of risk factors on Z scores are additive. Women with four or more clinical risk factors for reduced BMD had a mean Z score for both BMD and QUS variables of approximately 1.2, which, according to the evidence from prospective fracture studies, would double their risk of future fracture. (1 4) At present, there is a widely accepted convention to report BMD results in terms of T scores. The prevalence of osteoporosis in women with and without risk factors was similar for QUS and BMD indicating that as long as appropriately modified diagnostic T score thresholds are used for QUS, a similar proportion of individuals will be classified into the

414 FROST ET AL. different diagnostic categories. Approximately one-third of postmenopausal women with clinical risk factors were classified as osteoporotic using either BMD or QUS measurements, which is significantly greater than the 12% of women with no risk factors classified as osteoporotic (Figs. 3A and 3B). Over two-thirds of women with clinical risk factors were classified as osteopenic or osteoporotic. This shows that women with the clinical risk factors for osteoporosis suggested by the RCP, (10) the European Foundation of Osteoporosis and Bone Disease, (21) and the National Osteoporosis Foundation (22) for identifying individuals at risk of fracture have a higher prevalence of low bone mass and osteoporosis. This approach of identifying women for therapeutic intervention using risk factors and BMD assessment has been shown to be more cost-effective than treating all women with clinical risk factors. (21) This study has shown that when expressed as Z scores, QUS measurements are affected to the same extent as BMD measurements by clinical risk factors for osteoporosis. Therefore, QUS could provide an effective alternative method of assessing skeletal status to aid in a decision regarding treatment. However, it must be noted that at present there is a lack of evidence on whether QUS can be used to monitor disease progression or treatment efficacy, and so patients may still need to be monitored using DXA. (46) There were several limitations to this study. There are many different commercially available QUS devices in use worldwide, but only two were examined in this study. Other QUS devices, especially those that measure skeletal sites other than the calcaneus, may perform differently. Women who had taken estrogen or bisphosphonates for longer than 12 months were not excluded from the risk factor groups. The effects of antiresorptive therapy on bone mass are often more pronounced at the spine (47,48) and to date it appears that QUS measurements are affected to a lesser degree. (49 51) Consequently, spinal BMD may have been affected by therapeutic use to a greater extent than hip or calcaneal measurements. However, exclusion of women on estrogen or bisphosphonate therapy led to a small decrease in mean Z scores of just 0.09 U and 0.06 U for BMD and QUS measurement variables, respectively, and these changes did not change the statistical significance of any of the study results. Medical records were unavailable and, therefore, information regarding indications for osteoporosis was obtained using the GP referral letter and self-reported information from an osteoporosis questionnaire. This means that data such as name of medication, medication dosage, and reasons for secondary amenorrhea, for example, could not be confirmed. Another limitation of the present study was that T and Z scores were calculated using local reference data generated using the methods described by Ryan et al. (20) This may explain why the mean Z scores for women with no clinical risk factors were positive for both BMD and QUS variables (Table 4). However, the offsets were small (approximately 0.1) and did not affect the validity of the conclusions from the statistical analysis. A genuine population sample would have provided more reliable reference data, but even that would not exclude all risk factors such as a low BMI, X-ray osteopenia, or a current smoking habit. The aim of this study was not to assess the appropriateness of criteria to construct reference ranges but rather to assess whether QUS variables respond to clinical risk factors for osteoporosis. Therefore, this consistency of using the same reference population to calculate T and Z scores is advantageous as any differences observed between BMD and QUS measurements are caused by different technologies and not discrepancies in reference data. Nevertheless, the local reference data used in this study were similar to those provided by the manufacturers for both BMD and QUS measurement variables, with the young adult mean being within 2% of the manufacturer-provided values and the young adult SD being within 7%. In conclusion, indications for osteoporosis affected calcaneal BUA and SOS measurements to the same extent as axial BMD measurements. Therefore, calcaneal QUS could become an alternative tool for identifying individuals with low bone mass and possible candidates for therapeutic intervention. However, at present, criteria for the diagnosis of osteoporosis and strategies for using QUS for fracture risk assessment need to be developed before QUS can be used more effectively in clinical practice. ACKNOWLEDGMENTS We are very grateful to the women who participated in this study and we also thank A. Jefferies and F. Crane for performing the DXA scans. REFERENCES 1. Cummings SR, Nevitt MC, Browner WS, Stone K, Fox KM, Ensrud KE, Cauley J, Black D, Vogt TM, for the Study of Osteoporotic Fractures Research Group 1995 Risk factors for hip fracture in white women. N Engl J Med 332:767 773. 2. Hui SL, Slemenda CW, Johnston CC 1988 Age and bone mass as predictors of fracture in a prospective study. J Clin Invest 81:1804 1809. 3. Black DM, Cummings SR, Genant HK, Nevitt MC, Palermo L, Browner W 1992 Axial and appendicular bone density predicts fractures in older women. J Bone Miner Res 7:633 638. 4. Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K, Genant HK, Palermo L, Scott J, Vogt TM 1993 Bone density at various sites for prediction of hip fractures. Lancet 341:72 75. 5. Ross PD, Davis JW, Epstein RS, Wasnich RD 1991 Preexisting fractures and bone mass predict vertebral fracture incidence in women. Ann Intern Med 114:919 923. 6. The World Health Organization (WHO) Study Group 1994 Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. Technical Report Series 843, WHO, Geneva, Switzerland. 7. Langton CM, Palmer SB, Porter RW 1984 The measurement of broadband ultrasound attenuation in cancellous bone. Med Eng Phys 13:89 91. 8. Bauer DC, Glüer CC, Cauley JA, Vogt TM, Ensrud KE, Genant HK, Black DM, for the Study of Osteoporotic Fractures Research Group 1997 Broadband ultrasound attenuation predicts fractures strongly and independently of densitometry in older women. Arch Intern Med 157:629 634. 9. Hans D, Dargent-Molina P, Schott AM, Sebert JL, Cormier C, Kotzki PO, Delmas PD, Pouilles JM, Breart G, Meunier PJ 1996 Ultrasonographic heel measurements to predict hip frac-

QUS AND RISK FACTORS FOR OSTEOPOROSIS ture in elderly women: The EPIDOS Prospective Study. Lancet 348:511 514. 10. Royal College of Physicians 1999 Strategic considerations. In: Osteoporosis: Clinical Guidelines for Prevention and Treatment. Royal College of Physicians, London, UK. 11. Ahmed AIH, Ilic D, Blake GM, Rymer JM, Fogelman I 1996 Review of 3530 referrals for bone density measurements of spine and femur: Evidence that radiographic osteopenia predicts low bone mass. Radiology 207:619 624. 12. Bainbridge PR, Eastell R 1998 Indications for bone densitometry: Do they identify patients with low bone mineral density? Current research in osteoporosis and bone mineral measurement. Br I Radiol 5:5 6. 13. Gómez Acotto C, Schott AM, Hans D, Niepomniszcze H, Mautalen CA, Meunier PJ 1998 Hyperparathyroidism influences ultrasound bone measurement on the os calcis. Osteoporos Int 8:455 459. 14. Ingle BM, Thomas WEG, Eastell R 1998 Differential effects on ultrasound properties of bone in primary hyperparathyroidism. Osteoporos Int 8(Suppl 3):S56. 15. Minisola S, Rosso R, Scarda A, Pacitti MT, Romagnoli E, Mazzuoli G 1995 Quantitative ultrasound assessment of bone in patients with primary hyperparathyroidism. Calcif Tissue Int 56:526 528. 16. Robinson RJ, Carr I, Iqbal SJ, al-azzawi F, Abrams K, Mayberry JF 1998 Screening for osteoporosis in Crohn s disease. A detailed evaluation of calcaneal ultrasound. Eur J Gastroenterol Hepatol 10:137 140. 17. Newrkla S, Weiss P, Grampp S, Resch H 1998 Bone ultrasonometry in patients with anorexia nervosa comparison of three methods of densitometry. Osteoporos Int 8(Suppl 3):S50. 18. Wittich A, Vega E, Casco C, Marini A, Forlano C, Segovia F, Nadal M, Mautalen C 1998 Ultrasound velocity of the tibia in patients on haemodialysis. J Clin Densitometry 1:157 163. 19. Blanckaert F, Cortet B, Coquerelle B, Flipo RM, Duquesnoy B, Matchandise X, Delcambre B 1997 Contribution of calcaneal ultrasonic assessment to the evaluation of postmenopausal and glucocorticoid-induced osteoporosis. Rev Rhum 64:305 313. 20. Ryan PJ, Spector TP, Blake GM, Doyle DV, Fogelman I 1993 A comparison of reference bone mineral density measurements derived from two sources: Referred and population based. Br J Radiol 66:1138 1141. 21. Kanis JA, Delmas P, Burckhardt P, Cooper C, Torgerson D, on behalf of the European Foundation for Osteoporosis and Bone Disease 1997 Guidelines for diagnosis and management of osteoporosis. Osteoporos Int 7:390 406. 22. National Osteoporosis Foundation 1998 Risk assessment. In: Physician s Guide to Prevention and Treatment of Osteoporosis. National Osteoporosis Foundation, Washington DC, USA. 23. Frost ML, Blake GM, Fogelman I 2000 Can the WHO criteria for diagnosing osteoporosis be applied to calcaneal quantitative ultrasound? Osteoporos Int 11:321 330. 24. Rowe R 1999 The management of osteoporosis in general practice: Results of a national survey. Osteoporos Rev 7:1 3. 25. Glüer CC, Hans D 1999 How to use ultrasound for risk assessment: A need for defining strategies. Osteoporos Int 9:193 195. 26. Faulkner KG, von Stetten E, Miller P 1999 Discordance in patient classification using T-scores. J Clin Densitometry 2:343 350. 27. Ross P, Huang C, Davis J, Imose K, Yates J, Vogel J, Wasnich R 1995 Predicting vertebral deformity using bone densitometry at various skeletal sites and calcaneus ultrasound. Bone 16:325 332. 28. Stewart A, Reid DM, Porter RW 1994 Broadband ultrasound attenuation and dual energy x-ray absorptiometry in patients with hip fracture: Which technique discriminates fracture risk? Calcif Tissue Int 54:466 469. 415 29. Thompson P, Taylor J, Fisher A, Oliver R 1998 Quantitative heel ultrasound in 3180 women between 45 and 75 years of age: Compliance, normal ranges and relationship to fracture history. Osteoporos Int 8:211 214. 30. Huang C, Ross PD, Yates AJ, Walker RE, Imose K, Emi K, Wasnich RD 1998 Prediction of fracture risk by radiographic absorptiometry and quantitative ultrasound: A prospective study. Calcif Tissue Int 63:380 384. 31. Pluijm SMF, Graafmans WC, Bouter LM, Lips P 1999 Ultrasound measurements for the prediction of osteoporotic fractures in elderly people. Osteoporos Int 9:550 556. 32. Porter RW, Miller CG, Grainger D, Palmer SB 1990 Prediction of hip fracture in elderly women: A prospective study. BMJ 301:638 641. 33. Stewart A, Torgerson DJ, Reid DM 1996 Prediction of fractures in perimenopausal women: A comparison of dual energy x-ray absorptiometry and broadband ultrasound attenuation. Ann Rheum Dis 55:140 142. 34. Jergas M, Uffman M, Escher H 1994 Interobserver variation in the detection of osteopenia and comparison with dual X-ray absorptiometry of the lumbar spine. Skeletal Radiol 23:193 199. 35. Arden NK, Baker J, Hogg C, Baan K, Spector T 1996 The heritability of bone mineral density, ultrasound of the calcaneus and hip axis length: A study of postmenopausal twins. J Bone Miner Res 11:530 534. 36. Danielson ME, Cauley JA, Baker CE, Newman AB, Dorman JS, Towers JD, Kuller LH 1999 Familial resemblance of bone mineral density (BMD) and calcaneal ultrasound attenuation: The BMD in mothers and daughters study. J Bone Miner Res 14:102 110. 37. Grainge MJ, Coupland CAC, Cliffe SJ, Chilvers CED, Hosking DJ 1999 Association between a family history of fractures and bone mineral density in early postmenopausal women. Bone 24:507 512. 38. Fox KM, Cummings SR, Powell-Threets, Stone K, for the Study of Osteoporotic Fractures Research Group 1998 Family history and risk of osteoporotic fracture. Osteoporos Int 8:557 562. 39. Howard GM, Nguyen TV, Harris M, Kelly PJ, Eisman JA 1998 Genetic and environmental contributions to the association between quantitative ultrasound and bone mineral density measurements: A twin study. J Bone Miner Res 13:1318 1327. 40. Bauer DC, Browner WS, Cauley JA, Orwoll ES, Scott JC, Black DM, Tao JL, Cummings SR 1993 Factors associated with appendicular bone mass in older women. Ann Intern Med 118:657 665. 41. Ravn P, Cizza G, Bjarnason NH, Thompson D, Daley M, Wasnich RD, McClung M, Hosking D, Yates AJ, Christiansen C, for the Early Postmenopausal Intervention Cohort (EPIC) Study Group 1999 Low body mass index is an important risk factor for low bone mass and increased bone loss in early postmenopausal women. J Bone Miner Res 14:1622 1627. 42. Ross PD 1996 Osteoporosis. Frequency, consequences, and risk factors. Arch Intern Med 156:1399 1411. 43. Hopper JL, Seeman E 1994 The bone density of female twins discordant for tobacco use. N Engl J Med 330:387 392. 44. Reid IR 1998 Glucocorticoid-induced osteoporosis. Assessment and treatment. J Clin Densitometry 1:65 73. 45. Krall EA, Dawson-Hughes B 1999 Smoking increases bone loss and decreases intestinal calcium absorption. J Bone Miner Res 14:215 220. 46. Glüer C-C, for the International Quantitative Ultrasound Consensus Group 1997 Quantitative ultrasound techniques for the assessment of osteoporosis: Expert agreement of current status. J Bone Miner Res 12:1280 1288. 47. Eastell R 1998 Treatment of postmenopausal osteoporosis. N Engl J Med 338:736 746.