- Email: [email protected]
Unilateral vs Bilateral Hip Bone Mineral Density Measurement for the Diagnosis of Osteoporosis
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health, vol. 17, no. 1, 84e90, 2014 Ó Copyright 2014 by The International Society for Clinical Densitometry 1094-6950/17:84e90/$36.00 http://dx.doi.org/10.1016/j.jocd.2013.04.003
Section III: Quality Issues
Unilateral vs Bilateral Hip Bone Mineral Density Measurement for the Diagnosis of Osteoporosis Shota Ikegami,*,1 Mikio Kamimura,2 Shigeharu Uchiyama,3 Keijiro Mukaiyama,3 and Hiroyuki Kato3 1
Department of Orthopaedic Surgery, Yodakubo Hospital, Nagano, Japan; 2Center for Osteoporosis and Spinal Disorders, Kamimura Orthopaedic Clinic, Matsumoto, Japan; and 3Department of Orthopaedic Surgery, Shinshu University, Matsumoto, Japan
Abstract It has not been established whether unilateral or bilateral hip dual-energy X-ray absorptiometry (DXA) is preferable for the diagnosis of osteoporosis. We investigated the discordance in DXA measurements in bilateral hips to determine whether unilateral DXA is valid for osteoporosis diagnosis. The subjects were 2964 Japanese patients without a previous diagnosis of primary osteoporosis. We measured bilateral femoral bone mineral density (BMD) and calculated indices, related to the unilateral results, for predicting contralateral hip osteoporosis. A likelihood ratio (LR) of a negative test (LR []) of less than 0.2 was considered to exclude the diagnosis. In the normal spinal BMD group, the sensitivity of unilateral DXA for women was 27e73% and LR () was 0.28e0.73; the sensitivity for men was 0e50% and LR () was 0.51e1.00; the diagnosis of contralateral osteoporosis was not excluded. Sensitivity increased and LR () increased with worsening spinal BMD status; however, LR () did not meet the cutoff for exclusion. We could exclude unilateral hip osteoporosis, in women only, by performing contralateral femoral DXA; this necessitated lowering the T-score cutoff from 2.5 to 2.0. Unilateral femoral DXA is not useful for excluding the diagnosis of contralateral hip osteoporosis. Key Words: Bone mineral density; dual-energy X-ray absorptiometry; femur neck; osteoporosis; sensitivity and specificity.
at which BMD is measured, femoral BMD best predicts the risk of hip fracture (2,3); femoral dual-energy X-ray absorptiometry (DXA) is performed worldwide for the purposes of diagnosing osteoporosis. The International Osteoporosis Foundation and The National Osteoporosis Foundation both suggest measuring bone density at the hip for diagnosing osteoporosis in the elderly (4,5). Spinal DXA is popular in Japan, with femoral DXA not used as frequently for the diagnosis of osteoporosis as it is in Europe and North America. As the number of hip fractures in the Japanese elderly increases due to the rapid growth of the elderly population (6), preventing these fractures is coming to the forefront of clinicians’ attention. The latest revision of the Japanese guidelines for prevention and treatment of osteoporosis, published in 2011 and based on the position of the International Society for Clinical Densitometry (ISCD) (7), recommend BMD measurement at both the spine and the
Introduction There is a close relationship between declining bone mass and increasing fracture risk; the World Health Organization (WHO) working group has therefore developed an operational definition for osteoporosis based on bone mineral density (BMD). Patients are divided into 3 categories based on the deviation of their BMD from the young adult reference mean. A BMD value within 1 standard deviation below the young adult reference mean (T-score 1) is normal, a T-score between 1 and 2.5 indicates osteopenia, and a T-score 2.5 denotes osteoporosis (1). Of all the sites Received 02/23/13; Revised 03/23/13; Accepted 04/17/13. *Address correspondence to: Shota Ikegami, MD, PhD, Department of Orthopaedic Surgery, Yodakubo Hospital, 2857 Furumachi, Nagawa, Nagano 386-0603, Japan. E-mail: [email protected]
84
Should Bilateral Femoral DXA Be Performed? hip (8). Spinal DXA is not adequate to predict low bone density at the hip, and there may be many missed osteoporosis diagnoses in Japanese patients if spinal measurement alone is used (9). There is a worldwide consensus on the use of femoral DXA to diagnose osteoporosis; however, there is insufficient data to make a clear recommendation for either bilateral or unilateral measurement (10). In this study, we investigated the discordance in osteoporosis diagnosis using BMD measurements at the left or the right hip and whether unilateral DXA is valid to diagnose osteoporosis. We hypothesized that unilateral BMD measurement would not be as valid as bilateral measurement for the diagnosis of osteoporosis in Japanese patients.
Materials and Methods Subjects A total of 2964 Japanese patients first visited our institution between 2005 and 2010 out of concern for their bone health. None had previously received any examination or medical treatment for osteoporosis. An advance interview confirmed that patients did not have any radiopaque implants in the BMD measurement areas. Patients younger than 50 yr were excluded. All subjects gave verbal informed consent for study participation and underwent BMD measurement before receiving any treatment for osteoporosis. The study protocol was approved by the Ethics Committee of Shinshu University School of Medicine, Japan.
Measurement Variables BMD was measured using a DXA fan-beam bone densitometer (Lunar Prodigy; GE Healthcare, Waukesha, WI) at the L2eL4 levels of the posteroanterior spine and at the bilateral hips, including the total hip, the femoral neck, and the trochanter. All data were used for analysis, regardless of the presence of degenerative or traumatic vertebral changes. Measurement precision was calculated in 17 patients after triplicate scans of the spine and the bilateral hips, repositioning the subjects after each scan. These measurement values were used to calculate site-specific least significant change (LSC) with a 95% confidence interval (95% CI). We did not observe a significant left-right difference in the value of the coefficient of variation when evaluating triplicate BMD scans: the p values were 0.57 at the total hip, 0.61 at the femoral neck, and 0.45 at the trochanter, using the Wilcoxon signed-rank test. Therefore, we used a femoral BMD LSC that unified the right and left sides. Site-specific LSC values at our institution were 1.9% for L2eL4, 1.8% for the total hip, 3.2% for the femoral neck, and 2.4% for the trochanter.
Data Analysis At first, we adopted the WHO criterion for osteoporosis diagnosis for all subjects (1). We defined ‘‘unilateral hip osteoporosis’’ as the condition in which femoral BMD on only a single side was less than or equal to the osteoporotic cutoff
85 value and ‘‘bilateral hip osteoporosis’’ as the condition of both hips meeting the criterion. We calculated the prevalence of unilateral or bilateral hip osteoporosis in each age group, by decade, and assessed the relationship between aging and the prevalence of osteoporosis using ordered logistic regression modeling. In addition to the WHO criterion, we also examined the diagnostic method suggested by ISCD (10) of using the lowest T-score for BMD of 3 measured sites: spine, total hip, and femoral neck. Next, we calculated the sensitivity, specificity, and the likelihood ratio (LR) of a positive (LR [þ]) and a negative (LR []) test of unilateral femoral DXA to predict osteoporosis at the contralateral hip. An LR (þ) of more than 5 and an LR () of less than 0.2 were set to determine usefulness of a unilateral test to predict the presence of osteoporosis at the contralateral site. These indices were calculated for 3 patient groups, determined by spinal BMD status: normal, osteopenia, and osteoporosis. Statistical analyses were performed using the statistical package R, version 2.10.1 (available at http://www.r-project. org). A p value of !0.05 was considered to be statistically significant.
Results Of the 2964 patients recruited, 2477 were women, with a mean age of 68.2 8.8 yr (range: 50e94 yr), and 487 were men, with a mean age of 72.0 8.7 yr (range: 50e95 yr). As there were few patients older than 89 yr, all of these were classified with the eighth-decade group, creating the 80e99 yr age group. The baseline BMD data are shown in Table 1. Age had a negative effect on BMD, particularly in women. Men in their 60s seemed to have a tendency toward higher BMD values compared with men in their 50s; however, there were no significant differences on statistical analysis: p values were 0.15 at the spine, 0.38 at the total hip, 0.58 at the femoral neck, and 0.34 at the trochanter, using contrast test. Significant strong positive correlations between the left and right femoral BMD values were observed. The Pearson correlation coefficients were 0.94 for the total hip, 0.91 for the femoral neck, and 0.90 for the trochanter in women and 0.92 for the total hip, 0.92 for the femoral neck, and 0.88 for the trochanter in men.
Left-Right Difference in Femoral BMD Figure 1 shows the distribution of the differences between left and right femoral BMD values. A difference of about 3e5% was observed at all hip sites and in both sexes. There were many subjects for whom the left-right difference in femoral BMD exceeded the LSC value at all hip sites: 71% of women and 69% of men at the total hip, 58% of both women and men at the femoral neck, and 75% of women and 71% of men at the trochanter. In addition, 92% of all women and 91% of all men in the study had a left-right BMD difference exceeding the LSC value at more than 1 pair of hip sites. In all female subjects, the prevalence of bilateral hip osteoporosis was 12% at the total hip, 13% at the femoral neck,
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
86
Ikegami et al. Table 1 Bone Mineral Density Values by Age Group
Age group (yr) Women 50e59 60e69 70e79 80e99 50e99 Men 50e59 60e69 70e79 80e99 50e99
Number
Spine (g/cm2)
Total hip (g/cm2)
Femoral neck (g/cm2)
Trochanter (g/cm2)
470 986 750 271 2477
1.001 0.934 0.911 0.874 0.933
0.169 0.170 0.189 0.191 0.182
0.843 0.800 0.738 0.654 0.773
0.130 0.120 0.117 0.131 0.135
0.784 0.736 0.675 0.605 0.712
0.116 0.104 0.099 0.112 0.119
0.667 0.639 0.591 0.524 0.617
0.117 0.109 0.112 0.116 0.120
41 162 201 83 487
1.085 1.141 1.120 1.070 1.116
0.189 0.230 0.211 0.254 0.224
0.908 0.931 0.887 0.779 0.885
0.133 0.138 0.169 0.148 0.161
0.842 0.854 0.805 0.718 0.810
0.111 0.123 0.142 0.123 0.138
0.777 0.802 0.764 0.666 0.761
0.128 0.131 0.163 0.149 0.154
and 8% at the trochanter; the prevalence of unilateral hip osteoporosis was 6% at the total hip, 8% at the femoral neck, and 6% at the trochanter. In all male subjects, the prevalence of bilateral hip osteoporosis was 4% at the total hip, 6% at the femoral neck, and 3% at the trochanter; the prevalence of unilateral hip osteoporosis was 3% at the total hip, 6% at the femoral neck, and 4% at the trochanter. Figure 2 shows the prevalence of bilateral or unilateral hip osteoporosis for each age decade. The prevalence of both unilateral and bilateral hip osteoporosis significantly increased with aging. The odds ratio in women, per decade of increasing age, for unilateral and bilateral hip osteoporosis was 3.4 (95% CI: 3.0e3.9) at the total hip, 3.5 (95% CI: 3.1e4.0) at the femoral neck, and 2.9 (95% CI: 2.5e3.4) at the trochanter. In men, the odds ratio per decade of increasing age was 2.7 (95% CI: 1.7e4.4) at the total hip, 2.9 (95% CI: 2.0e4.2) at the femoral neck, and 2.6 (95% CI: 1.6e4.2) at the trochanter.
Ability of Unilateral Femoral DXA to Predict Contralateral Hip Osteoporosis
sites in both sexes. LR () of the total hip was relatively small, at 0.17 (95% CI: 0.13e0.22) for the right and 0.24 (95% CI: 0.20e0.29) for the left in women, and 0.35 (95% CI: 0.21e0.59) for the right and 0.19 (95% CI: 0.08e0.47) for the left in men; the upper limits of the 95% CIs were all above the cutoff value for usefulness. In the normal spinal BMD status group, in particular, LR () of unilateral femoral DXA values was insufficient at all hip sites: LR () for women was 0.28e0.73 and LR () for men was 0.51e1.00. Patients with worse spinal BMD status had lower LR () values: in the osteoporotic spinal BMD group, LR () for women was 0.15e0.27 and LR () for men was 0.16e0.38. The 95% CI of the LR () values exceed the cutoff usefulness value in all groups. Unilateral femoral DXA was therefore not useful to exclude contralateral hip osteoporosis. We considered the possibility that changing the T-score cutoff for unilateral DXA values could reduce the number of missed contralateral femoral osteoporosis diagnoses and therefore performed an additional analysis. Figure 5 shows
Figure 3 shows the sensitivity and specificity of unilateral femoral DXA for contralateral hip osteoporosis. ‘‘Spine þ total hip þ femoral neck’’ expresses the diagnostic method suggested by ISCD. With better spinal BMD status, more hip osteoporosis diagnoses were missed. In the group with normal spinal BMD, the sensitivity of unilateral femoral DXA was 27e73% in women and 0e50% in men. Even using the diagnostic method of the ISCD, the sensitivity was still insufficient: 58e63% in women and 29e50% in men. In contrast, the specificity was 87% or higher in both women and men, and there were few false positive results. Figure 4 shows LRs of unilateral femoral DXA for contralateral hip osteoporosis. LR (þ) was almost always greater than the usefulness cutoff of 5; this would indicate that unilateral femoral DXA values are useful for the definitive diagnosis of contralateral hip osteoporosis. LR (), however, was not less than the usefulness cutoff of 0.2 at most measured
Fig. 1. Left-right difference in bone mineral density at the hip.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
Should Bilateral Femoral DXA Be Performed?
87
Fig. 2. Prevalence of hip osteoporosis.
the change in LR () when we changed the T-score cutoff of unilateral femoral DXA from 2.5 to 1.0, in a stepwise fashion. As discussed previously, when the cutoff value was
defined as a T-score of 2.5, LR () was not less than 0.2 in either sex at any measurement site. In other words, many diagnoses were missed. But when the cutoff value was
Fig. 3. Sensitivity and specificity of unilateral femoral dual-energy X-ray absorptiometry for contralateral hip osteoporosis. Error bars represent 95% confidence intervals. Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
88
Ikegami et al.
Fig. 4. LRs of unilateral femoral dual-energy X-ray absorptiometry for contralateral hip osteoporosis. Error bars represent 95% confidence intervals; values in the shaded gray areas are considered useless for the diagnosis of contralateral hip osteoporosis. LR, likelihood ratio; LR (þ), LR of a positive test; LR (), LR of a negative test. changed to 2.0, LR () for women was 0.03e0.10, and it was less than 0.2 at all hip sites. Changing the cutoff value further did not lead to a remarkable decrease in LR (). Men also showed a tendency to a decreased LR () with lower T-score cutoff values, but the tendency was not statistically significant. Only in women were we able to exclude unilateral hip osteoporosis by performing contralateral femoral DXA, using a T-score cutoff of 2.0.
Discussion Our results support the hypothesis that unilateral BMD measurement is not as valid as bilateral measurement for the diagnosis of osteoporosis. We propose measuring bilateral femoral BMD in Japanese patients who are suspected of having bone loss. Our results showed that 92% of women and 91% of men had left-right differences in BMD, with about a 3e5% left-right difference at all hip sites and in both sexes. The prevalence of unilateral hip osteoporosis was significantly higher in older subjects, and unilateral femoral BMD measurement was not valid for excluding contralateral femoral osteoporosis. In patients with better spinal status, the validity of unilateral hip DXA to exclude contralateral hip osteoporosis was even lower. Even using the diagnostic method suggested by ISCD, rather than that suggested by WHO, did not produce a valid BMD measurement at the unilateral hip. Many factors are associated with fracture risk, but BMD is currently considered the most predictive. Osteoporosis may be suspected from patient history and from bone metabolism markers in the blood and the urine, but osteoporotic treatment is provided based on BMD measurements. Therefore, it is
important to assess BMD at the proper sites to avoid missing the diagnosis and the subsequent loss of treatment opportunity. Patients with osteopenia are most at risk to suffer adverse effects from insufficient site assessment. Many bone fractures occur in the osteopenia stage (11); hence, it creates clinical problems when patients are not treated because their disease has been underestimated. Left-right differences in femoral BMD have been previously reported (12). Some authors recommend unilateral femoral DXA, despite this laterality (13e15), whereas other authors insist that bilateral femoral DXA testing is necessary to reduce missed osteoporosis diagnoses (16e18). The leftright difference in femoral BMD includes both true laterality and pseudo differences. Insufficient internal rotation of the hip causes a falsely high BMD value on DXA (19), with the pseudo difference influencing the diagnosis of osteoporosis (20). A true difference between the left and the right femoral BMD often exists in patients with osteoarthritis. In Japan, the prevalence of hip osteoarthritis is reported to be between 1.0% and 2.4% (21,22). Patients with hip osteoarthritis have femoral BMD values 3% to 7% higher than normal control subjects, according to the Rotterdam Study (23). In a twin study, femoral BMD in those patients with unilateral hip osteoarthritis was found to be significantly higher than the normal control value only in the side with osteoarthritis (24). When BMD is only measured at the side with hip osteoarthritis, the validity of the study to evaluate osteoporosis is low. The limitations of the present study include the fact that that the cause of the laterality seen in femoral BMD values was not adequately examined. The cause may be, for example, lack of load/weight-bearing with a painful knee or ankle joint or the influence of footedness. In addition, the present
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
Should Bilateral Femoral DXA Be Performed?
89
Fig. 5. Changing the value of LR of a negative test by changing the cutoff T-score value given by unilateral femoral dualenergy X-ray absorptiometry. Error bars represent 95% confidence intervals; values in the shaded gray areas are considered useless for the diagnosis of contralateral hip osteoporosis. LR, likelihood ratio; LR (), LR of a negative test.
study is not a population study. For healthy subjects unlikely to have osteoporosis, in other words, those with a low pretest probability, the insufficiency of unilateral femoral DXA to exclude contralateral osteoporosis may not have a great influence. Therefore, we do not recommend bilateral femoral DXA at the screening stage: we emphasize that the present study evaluated laterality in patients who were suspected of having osteoporosis and who therefore underwent diagnostic, not screening, testing. Finally, we showed that missed diagnoses of contralateral osteoporosis could be reduced by using a lower BMD T-score as the threshold for suspicion. We were able to demonstrate the advantage of this stricter threshold in women but not in men. We were not able to determine whether this discrepancy is due to inherent sex differences or merely the result of insufficient statistical power. Even when DXA is used only for diagnosis, its cost can prohibit the performance of bilateral femoral DXA in all patients. The cutoff T-score value of 2.0 is more accurate than 2.5: using a cutoff T-score of 2.0 in the measuring hip decreases the possibility of missing contralateral hip osteoporosis by 0.03e0.10 times in women. We purport that an effective reduction in missed osteoporosis diagnoses in women can only be achieved by performing additional
contralateral femoral DXA when the T-score on unilateral femoral DXA is equal to or less than 2.0. On the other hand, men require bilateral femoral DXA if they are suspected of osteoporosis, regardless of unilateral femoral DXA values. Of course, performing accurate BMD measurement is important, and White et al (25) reported that DXA at the bilateral hips is higher in accuracy than unilateral measurement. We clearly indicated that the validity of unilateral femoral DXA for excluding contralateral hip osteoporosis is insufficient in patients of all spinal BMD statuses, even when using the ISCD diagnosing method. When osteoporotic treatment is provided based on the results of femoral DXA, hip BMD measurements should be performed bilaterally. This study did not show any evidence that unilateral hip osteoporosis is predicted by contralateral femoral BMD measurement. Aging increases the risk of unilateral hip osteoporosis; we recommend performing bilateral DXA so as not to miss this diagnosis.
References 1. World Health Organization. 1994 Assessment of fracture risk and its application to screening for postmenopausal osteoporosis.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
90
2. 3.
4.
5.
6. 7. 8. 9.
10.
11. 12.
Ikegami et al. Report of a WHO Study Group. World Health Organ Tech Rep Ser 843:1e129. Marshall D, Johnell O, Wedel H. 1996 Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 312:1254e1259. Cummings SR, Black DM, Nevitt MC, et al. 1993 Bone density at various sites for prediction of hip fractures. The Study of Osteoporotic Fractures Research Group. Lancet 341: 72e75. Kanis JA, Delmas P, Burckhardt P, et al. 1997 Guidelines for diagnosis and management of osteoporosis. The European Foundation for Osteoporosis and Bone Disease. Osteoporos Int 7: 390e406. Council of the National Osteoporosis Foundation. 1996 Guidelines for the early detection of osteoporosis and prediction of fracture risk. Council of the National Osteoporosis Foundation. S Afr Med J 86:1113e1116. Yoshimura N, Suzuki T, Hosoi T, Orimo H. 2005 Epidemiology of hip fracture in Japan: incidence and risk factors. J Bone Miner Metab 23(Suppl):78e80. Leib ES, Lewiecki EM, Binkley N, et al. 2004 Official positions of the International Society for Clinical Densitometry. J Clin Densitom 7:1e6. Orimo H, ed. 2011 Japanese Guidelines for the Prevention and Treatment of Osteoporosis 2011 Edition. Tokyo, Japan: Life Science Publishing. Ikegami S, Kamimura M, Uchiyama S, et al. 2009 Bone mineral density measurement at both spine and hip for diagnosing osteoporosis in Japanese patients. J Clin Densitom 12: 337e344. Baim S, Binkley N, Bilezikian JP, et al. 2008 Official Positions of the International Society for Clinical Densitometry and executive summary of the 2007 ISCD Position Development Conference. J Clin Densitom 11:75e91. Siris ES, Chen YT, Abbott TA, et al. 2004 Bone mineral density thresholds for pharmacological intervention to prevent fractures. Arch Intern Med 164:1108e1112. Balseiro J, Fahey FH, Ziessman HA, Le TV. 1988 Comparison of bone mineral density in both hips. Radiology 167:151e153.
13. Yang RS, Chieng PU, Tsai KS, Liu TK. 1996 Symmetry of bone mineral density in the hips is not affected by age. Nucl Med Commun 17:711e716. 14. Rao AD, Reddy S, Rao DS. 2000 Is there a difference between right and left femoral bone density? J Clin Densitom 3:57e61. 15. Petley GW, Taylor PA, Murrills AJ, et al. 2000 An investigation of the diagnostic value of bilateral femoral neck bone mineral density measurements. Osteoporos Int 11:675e679. 16. Hamdy R, Kiebzak GM, Seier E, Watts NB. 2006 The prevalence of significant left-right differences in hip bone mineral density. Osteoporos Int 17:1772e1780. 17. Xu H, Gong J, Chen JX, et al. 2007 Bilateral femoral bone mineral density measurements in Chinese women and men. J Clin Densitom 10:165e169. 18. Alele JD, Kamen DL, Hermayer KL, et al. 2009 The prevalence of significant left-right hip bone mineral density differences among black and white women. Osteoporos Int 20:2079e2085. 19. Goh JC, Low SL, Bose K. 1995 Effect of femoral rotation on bone mineral density measurements with dual energy X-ray absorptiometry. Calcif Tissue Int 57:340e343. 20. Lekamwasam S, Lenora RS. 2003 Effect of leg rotation on hip bone mineral density measurements. J Clin Densitom 6: 331e336. 21. Yoshimura N, Campbell L, Hashimoto T, et al. 1998 Acetabular dysplasia and hip osteoarthritis in Britain and Japan. Br J Rheumatol 37:1193e1197. 22. Inoue K, Wicart P, Kawasaki T, et al. 2000 Prevalence of hip osteoarthritis and acetabular dysplasia in French and Japanese adults. Rheumatology (Oxford) 39:745e748. 23. Burger H, van Daele PL, Odding E, et al. 1996 Association of radiographically evident osteoarthritis with higher bone mineral density and increased bone loss with age. The Rotterdam Study. Arthritis Rheum 39:81e86. 24. Antoniades L, MacGregor AJ, Matson M, Spector TD. 2000 A cotwin control study of the relationship between hip osteoarthritis and bone mineral density. Arthritis Rheum 43:1450e1455. 25. White J, Harris SS, Dallal GE, Dawson-Hughes B. 2003 Precision of single vs bilateral hip bone mineral density scans. J Clin Densitom 6:159e162.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
Section III: Quality Issues
Unilateral vs Bilateral Hip Bone Mineral Density Measurement for the Diagnosis of Osteoporosis Shota Ikegami,*,1 Mikio Kamimura,2 Shigeharu Uchiyama,3 Keijiro Mukaiyama,3 and Hiroyuki Kato3 1
Department of Orthopaedic Surgery, Yodakubo Hospital, Nagano, Japan; 2Center for Osteoporosis and Spinal Disorders, Kamimura Orthopaedic Clinic, Matsumoto, Japan; and 3Department of Orthopaedic Surgery, Shinshu University, Matsumoto, Japan
Abstract It has not been established whether unilateral or bilateral hip dual-energy X-ray absorptiometry (DXA) is preferable for the diagnosis of osteoporosis. We investigated the discordance in DXA measurements in bilateral hips to determine whether unilateral DXA is valid for osteoporosis diagnosis. The subjects were 2964 Japanese patients without a previous diagnosis of primary osteoporosis. We measured bilateral femoral bone mineral density (BMD) and calculated indices, related to the unilateral results, for predicting contralateral hip osteoporosis. A likelihood ratio (LR) of a negative test (LR []) of less than 0.2 was considered to exclude the diagnosis. In the normal spinal BMD group, the sensitivity of unilateral DXA for women was 27e73% and LR () was 0.28e0.73; the sensitivity for men was 0e50% and LR () was 0.51e1.00; the diagnosis of contralateral osteoporosis was not excluded. Sensitivity increased and LR () increased with worsening spinal BMD status; however, LR () did not meet the cutoff for exclusion. We could exclude unilateral hip osteoporosis, in women only, by performing contralateral femoral DXA; this necessitated lowering the T-score cutoff from 2.5 to 2.0. Unilateral femoral DXA is not useful for excluding the diagnosis of contralateral hip osteoporosis. Key Words: Bone mineral density; dual-energy X-ray absorptiometry; femur neck; osteoporosis; sensitivity and specificity.
at which BMD is measured, femoral BMD best predicts the risk of hip fracture (2,3); femoral dual-energy X-ray absorptiometry (DXA) is performed worldwide for the purposes of diagnosing osteoporosis. The International Osteoporosis Foundation and The National Osteoporosis Foundation both suggest measuring bone density at the hip for diagnosing osteoporosis in the elderly (4,5). Spinal DXA is popular in Japan, with femoral DXA not used as frequently for the diagnosis of osteoporosis as it is in Europe and North America. As the number of hip fractures in the Japanese elderly increases due to the rapid growth of the elderly population (6), preventing these fractures is coming to the forefront of clinicians’ attention. The latest revision of the Japanese guidelines for prevention and treatment of osteoporosis, published in 2011 and based on the position of the International Society for Clinical Densitometry (ISCD) (7), recommend BMD measurement at both the spine and the
Introduction There is a close relationship between declining bone mass and increasing fracture risk; the World Health Organization (WHO) working group has therefore developed an operational definition for osteoporosis based on bone mineral density (BMD). Patients are divided into 3 categories based on the deviation of their BMD from the young adult reference mean. A BMD value within 1 standard deviation below the young adult reference mean (T-score 1) is normal, a T-score between 1 and 2.5 indicates osteopenia, and a T-score 2.5 denotes osteoporosis (1). Of all the sites Received 02/23/13; Revised 03/23/13; Accepted 04/17/13. *Address correspondence to: Shota Ikegami, MD, PhD, Department of Orthopaedic Surgery, Yodakubo Hospital, 2857 Furumachi, Nagawa, Nagano 386-0603, Japan. E-mail: [email protected]
84
Should Bilateral Femoral DXA Be Performed? hip (8). Spinal DXA is not adequate to predict low bone density at the hip, and there may be many missed osteoporosis diagnoses in Japanese patients if spinal measurement alone is used (9). There is a worldwide consensus on the use of femoral DXA to diagnose osteoporosis; however, there is insufficient data to make a clear recommendation for either bilateral or unilateral measurement (10). In this study, we investigated the discordance in osteoporosis diagnosis using BMD measurements at the left or the right hip and whether unilateral DXA is valid to diagnose osteoporosis. We hypothesized that unilateral BMD measurement would not be as valid as bilateral measurement for the diagnosis of osteoporosis in Japanese patients.
Materials and Methods Subjects A total of 2964 Japanese patients first visited our institution between 2005 and 2010 out of concern for their bone health. None had previously received any examination or medical treatment for osteoporosis. An advance interview confirmed that patients did not have any radiopaque implants in the BMD measurement areas. Patients younger than 50 yr were excluded. All subjects gave verbal informed consent for study participation and underwent BMD measurement before receiving any treatment for osteoporosis. The study protocol was approved by the Ethics Committee of Shinshu University School of Medicine, Japan.
Measurement Variables BMD was measured using a DXA fan-beam bone densitometer (Lunar Prodigy; GE Healthcare, Waukesha, WI) at the L2eL4 levels of the posteroanterior spine and at the bilateral hips, including the total hip, the femoral neck, and the trochanter. All data were used for analysis, regardless of the presence of degenerative or traumatic vertebral changes. Measurement precision was calculated in 17 patients after triplicate scans of the spine and the bilateral hips, repositioning the subjects after each scan. These measurement values were used to calculate site-specific least significant change (LSC) with a 95% confidence interval (95% CI). We did not observe a significant left-right difference in the value of the coefficient of variation when evaluating triplicate BMD scans: the p values were 0.57 at the total hip, 0.61 at the femoral neck, and 0.45 at the trochanter, using the Wilcoxon signed-rank test. Therefore, we used a femoral BMD LSC that unified the right and left sides. Site-specific LSC values at our institution were 1.9% for L2eL4, 1.8% for the total hip, 3.2% for the femoral neck, and 2.4% for the trochanter.
Data Analysis At first, we adopted the WHO criterion for osteoporosis diagnosis for all subjects (1). We defined ‘‘unilateral hip osteoporosis’’ as the condition in which femoral BMD on only a single side was less than or equal to the osteoporotic cutoff
85 value and ‘‘bilateral hip osteoporosis’’ as the condition of both hips meeting the criterion. We calculated the prevalence of unilateral or bilateral hip osteoporosis in each age group, by decade, and assessed the relationship between aging and the prevalence of osteoporosis using ordered logistic regression modeling. In addition to the WHO criterion, we also examined the diagnostic method suggested by ISCD (10) of using the lowest T-score for BMD of 3 measured sites: spine, total hip, and femoral neck. Next, we calculated the sensitivity, specificity, and the likelihood ratio (LR) of a positive (LR [þ]) and a negative (LR []) test of unilateral femoral DXA to predict osteoporosis at the contralateral hip. An LR (þ) of more than 5 and an LR () of less than 0.2 were set to determine usefulness of a unilateral test to predict the presence of osteoporosis at the contralateral site. These indices were calculated for 3 patient groups, determined by spinal BMD status: normal, osteopenia, and osteoporosis. Statistical analyses were performed using the statistical package R, version 2.10.1 (available at http://www.r-project. org). A p value of !0.05 was considered to be statistically significant.
Results Of the 2964 patients recruited, 2477 were women, with a mean age of 68.2 8.8 yr (range: 50e94 yr), and 487 were men, with a mean age of 72.0 8.7 yr (range: 50e95 yr). As there were few patients older than 89 yr, all of these were classified with the eighth-decade group, creating the 80e99 yr age group. The baseline BMD data are shown in Table 1. Age had a negative effect on BMD, particularly in women. Men in their 60s seemed to have a tendency toward higher BMD values compared with men in their 50s; however, there were no significant differences on statistical analysis: p values were 0.15 at the spine, 0.38 at the total hip, 0.58 at the femoral neck, and 0.34 at the trochanter, using contrast test. Significant strong positive correlations between the left and right femoral BMD values were observed. The Pearson correlation coefficients were 0.94 for the total hip, 0.91 for the femoral neck, and 0.90 for the trochanter in women and 0.92 for the total hip, 0.92 for the femoral neck, and 0.88 for the trochanter in men.
Left-Right Difference in Femoral BMD Figure 1 shows the distribution of the differences between left and right femoral BMD values. A difference of about 3e5% was observed at all hip sites and in both sexes. There were many subjects for whom the left-right difference in femoral BMD exceeded the LSC value at all hip sites: 71% of women and 69% of men at the total hip, 58% of both women and men at the femoral neck, and 75% of women and 71% of men at the trochanter. In addition, 92% of all women and 91% of all men in the study had a left-right BMD difference exceeding the LSC value at more than 1 pair of hip sites. In all female subjects, the prevalence of bilateral hip osteoporosis was 12% at the total hip, 13% at the femoral neck,
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
86
Ikegami et al. Table 1 Bone Mineral Density Values by Age Group
Age group (yr) Women 50e59 60e69 70e79 80e99 50e99 Men 50e59 60e69 70e79 80e99 50e99
Number
Spine (g/cm2)
Total hip (g/cm2)
Femoral neck (g/cm2)
Trochanter (g/cm2)
470 986 750 271 2477
1.001 0.934 0.911 0.874 0.933
0.169 0.170 0.189 0.191 0.182
0.843 0.800 0.738 0.654 0.773
0.130 0.120 0.117 0.131 0.135
0.784 0.736 0.675 0.605 0.712
0.116 0.104 0.099 0.112 0.119
0.667 0.639 0.591 0.524 0.617
0.117 0.109 0.112 0.116 0.120
41 162 201 83 487
1.085 1.141 1.120 1.070 1.116
0.189 0.230 0.211 0.254 0.224
0.908 0.931 0.887 0.779 0.885
0.133 0.138 0.169 0.148 0.161
0.842 0.854 0.805 0.718 0.810
0.111 0.123 0.142 0.123 0.138
0.777 0.802 0.764 0.666 0.761
0.128 0.131 0.163 0.149 0.154
and 8% at the trochanter; the prevalence of unilateral hip osteoporosis was 6% at the total hip, 8% at the femoral neck, and 6% at the trochanter. In all male subjects, the prevalence of bilateral hip osteoporosis was 4% at the total hip, 6% at the femoral neck, and 3% at the trochanter; the prevalence of unilateral hip osteoporosis was 3% at the total hip, 6% at the femoral neck, and 4% at the trochanter. Figure 2 shows the prevalence of bilateral or unilateral hip osteoporosis for each age decade. The prevalence of both unilateral and bilateral hip osteoporosis significantly increased with aging. The odds ratio in women, per decade of increasing age, for unilateral and bilateral hip osteoporosis was 3.4 (95% CI: 3.0e3.9) at the total hip, 3.5 (95% CI: 3.1e4.0) at the femoral neck, and 2.9 (95% CI: 2.5e3.4) at the trochanter. In men, the odds ratio per decade of increasing age was 2.7 (95% CI: 1.7e4.4) at the total hip, 2.9 (95% CI: 2.0e4.2) at the femoral neck, and 2.6 (95% CI: 1.6e4.2) at the trochanter.
Ability of Unilateral Femoral DXA to Predict Contralateral Hip Osteoporosis
sites in both sexes. LR () of the total hip was relatively small, at 0.17 (95% CI: 0.13e0.22) for the right and 0.24 (95% CI: 0.20e0.29) for the left in women, and 0.35 (95% CI: 0.21e0.59) for the right and 0.19 (95% CI: 0.08e0.47) for the left in men; the upper limits of the 95% CIs were all above the cutoff value for usefulness. In the normal spinal BMD status group, in particular, LR () of unilateral femoral DXA values was insufficient at all hip sites: LR () for women was 0.28e0.73 and LR () for men was 0.51e1.00. Patients with worse spinal BMD status had lower LR () values: in the osteoporotic spinal BMD group, LR () for women was 0.15e0.27 and LR () for men was 0.16e0.38. The 95% CI of the LR () values exceed the cutoff usefulness value in all groups. Unilateral femoral DXA was therefore not useful to exclude contralateral hip osteoporosis. We considered the possibility that changing the T-score cutoff for unilateral DXA values could reduce the number of missed contralateral femoral osteoporosis diagnoses and therefore performed an additional analysis. Figure 5 shows
Figure 3 shows the sensitivity and specificity of unilateral femoral DXA for contralateral hip osteoporosis. ‘‘Spine þ total hip þ femoral neck’’ expresses the diagnostic method suggested by ISCD. With better spinal BMD status, more hip osteoporosis diagnoses were missed. In the group with normal spinal BMD, the sensitivity of unilateral femoral DXA was 27e73% in women and 0e50% in men. Even using the diagnostic method of the ISCD, the sensitivity was still insufficient: 58e63% in women and 29e50% in men. In contrast, the specificity was 87% or higher in both women and men, and there were few false positive results. Figure 4 shows LRs of unilateral femoral DXA for contralateral hip osteoporosis. LR (þ) was almost always greater than the usefulness cutoff of 5; this would indicate that unilateral femoral DXA values are useful for the definitive diagnosis of contralateral hip osteoporosis. LR (), however, was not less than the usefulness cutoff of 0.2 at most measured
Fig. 1. Left-right difference in bone mineral density at the hip.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
Should Bilateral Femoral DXA Be Performed?
87
Fig. 2. Prevalence of hip osteoporosis.
the change in LR () when we changed the T-score cutoff of unilateral femoral DXA from 2.5 to 1.0, in a stepwise fashion. As discussed previously, when the cutoff value was
defined as a T-score of 2.5, LR () was not less than 0.2 in either sex at any measurement site. In other words, many diagnoses were missed. But when the cutoff value was
Fig. 3. Sensitivity and specificity of unilateral femoral dual-energy X-ray absorptiometry for contralateral hip osteoporosis. Error bars represent 95% confidence intervals. Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
88
Ikegami et al.
Fig. 4. LRs of unilateral femoral dual-energy X-ray absorptiometry for contralateral hip osteoporosis. Error bars represent 95% confidence intervals; values in the shaded gray areas are considered useless for the diagnosis of contralateral hip osteoporosis. LR, likelihood ratio; LR (þ), LR of a positive test; LR (), LR of a negative test. changed to 2.0, LR () for women was 0.03e0.10, and it was less than 0.2 at all hip sites. Changing the cutoff value further did not lead to a remarkable decrease in LR (). Men also showed a tendency to a decreased LR () with lower T-score cutoff values, but the tendency was not statistically significant. Only in women were we able to exclude unilateral hip osteoporosis by performing contralateral femoral DXA, using a T-score cutoff of 2.0.
Discussion Our results support the hypothesis that unilateral BMD measurement is not as valid as bilateral measurement for the diagnosis of osteoporosis. We propose measuring bilateral femoral BMD in Japanese patients who are suspected of having bone loss. Our results showed that 92% of women and 91% of men had left-right differences in BMD, with about a 3e5% left-right difference at all hip sites and in both sexes. The prevalence of unilateral hip osteoporosis was significantly higher in older subjects, and unilateral femoral BMD measurement was not valid for excluding contralateral femoral osteoporosis. In patients with better spinal status, the validity of unilateral hip DXA to exclude contralateral hip osteoporosis was even lower. Even using the diagnostic method suggested by ISCD, rather than that suggested by WHO, did not produce a valid BMD measurement at the unilateral hip. Many factors are associated with fracture risk, but BMD is currently considered the most predictive. Osteoporosis may be suspected from patient history and from bone metabolism markers in the blood and the urine, but osteoporotic treatment is provided based on BMD measurements. Therefore, it is
important to assess BMD at the proper sites to avoid missing the diagnosis and the subsequent loss of treatment opportunity. Patients with osteopenia are most at risk to suffer adverse effects from insufficient site assessment. Many bone fractures occur in the osteopenia stage (11); hence, it creates clinical problems when patients are not treated because their disease has been underestimated. Left-right differences in femoral BMD have been previously reported (12). Some authors recommend unilateral femoral DXA, despite this laterality (13e15), whereas other authors insist that bilateral femoral DXA testing is necessary to reduce missed osteoporosis diagnoses (16e18). The leftright difference in femoral BMD includes both true laterality and pseudo differences. Insufficient internal rotation of the hip causes a falsely high BMD value on DXA (19), with the pseudo difference influencing the diagnosis of osteoporosis (20). A true difference between the left and the right femoral BMD often exists in patients with osteoarthritis. In Japan, the prevalence of hip osteoarthritis is reported to be between 1.0% and 2.4% (21,22). Patients with hip osteoarthritis have femoral BMD values 3% to 7% higher than normal control subjects, according to the Rotterdam Study (23). In a twin study, femoral BMD in those patients with unilateral hip osteoarthritis was found to be significantly higher than the normal control value only in the side with osteoarthritis (24). When BMD is only measured at the side with hip osteoarthritis, the validity of the study to evaluate osteoporosis is low. The limitations of the present study include the fact that that the cause of the laterality seen in femoral BMD values was not adequately examined. The cause may be, for example, lack of load/weight-bearing with a painful knee or ankle joint or the influence of footedness. In addition, the present
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
Should Bilateral Femoral DXA Be Performed?
89
Fig. 5. Changing the value of LR of a negative test by changing the cutoff T-score value given by unilateral femoral dualenergy X-ray absorptiometry. Error bars represent 95% confidence intervals; values in the shaded gray areas are considered useless for the diagnosis of contralateral hip osteoporosis. LR, likelihood ratio; LR (), LR of a negative test.
study is not a population study. For healthy subjects unlikely to have osteoporosis, in other words, those with a low pretest probability, the insufficiency of unilateral femoral DXA to exclude contralateral osteoporosis may not have a great influence. Therefore, we do not recommend bilateral femoral DXA at the screening stage: we emphasize that the present study evaluated laterality in patients who were suspected of having osteoporosis and who therefore underwent diagnostic, not screening, testing. Finally, we showed that missed diagnoses of contralateral osteoporosis could be reduced by using a lower BMD T-score as the threshold for suspicion. We were able to demonstrate the advantage of this stricter threshold in women but not in men. We were not able to determine whether this discrepancy is due to inherent sex differences or merely the result of insufficient statistical power. Even when DXA is used only for diagnosis, its cost can prohibit the performance of bilateral femoral DXA in all patients. The cutoff T-score value of 2.0 is more accurate than 2.5: using a cutoff T-score of 2.0 in the measuring hip decreases the possibility of missing contralateral hip osteoporosis by 0.03e0.10 times in women. We purport that an effective reduction in missed osteoporosis diagnoses in women can only be achieved by performing additional
contralateral femoral DXA when the T-score on unilateral femoral DXA is equal to or less than 2.0. On the other hand, men require bilateral femoral DXA if they are suspected of osteoporosis, regardless of unilateral femoral DXA values. Of course, performing accurate BMD measurement is important, and White et al (25) reported that DXA at the bilateral hips is higher in accuracy than unilateral measurement. We clearly indicated that the validity of unilateral femoral DXA for excluding contralateral hip osteoporosis is insufficient in patients of all spinal BMD statuses, even when using the ISCD diagnosing method. When osteoporotic treatment is provided based on the results of femoral DXA, hip BMD measurements should be performed bilaterally. This study did not show any evidence that unilateral hip osteoporosis is predicted by contralateral femoral BMD measurement. Aging increases the risk of unilateral hip osteoporosis; we recommend performing bilateral DXA so as not to miss this diagnosis.
References 1. World Health Organization. 1994 Assessment of fracture risk and its application to screening for postmenopausal osteoporosis.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014
90
2. 3.
4.
5.
6. 7. 8. 9.
10.
11. 12.
Ikegami et al. Report of a WHO Study Group. World Health Organ Tech Rep Ser 843:1e129. Marshall D, Johnell O, Wedel H. 1996 Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 312:1254e1259. Cummings SR, Black DM, Nevitt MC, et al. 1993 Bone density at various sites for prediction of hip fractures. The Study of Osteoporotic Fractures Research Group. Lancet 341: 72e75. Kanis JA, Delmas P, Burckhardt P, et al. 1997 Guidelines for diagnosis and management of osteoporosis. The European Foundation for Osteoporosis and Bone Disease. Osteoporos Int 7: 390e406. Council of the National Osteoporosis Foundation. 1996 Guidelines for the early detection of osteoporosis and prediction of fracture risk. Council of the National Osteoporosis Foundation. S Afr Med J 86:1113e1116. Yoshimura N, Suzuki T, Hosoi T, Orimo H. 2005 Epidemiology of hip fracture in Japan: incidence and risk factors. J Bone Miner Metab 23(Suppl):78e80. Leib ES, Lewiecki EM, Binkley N, et al. 2004 Official positions of the International Society for Clinical Densitometry. J Clin Densitom 7:1e6. Orimo H, ed. 2011 Japanese Guidelines for the Prevention and Treatment of Osteoporosis 2011 Edition. Tokyo, Japan: Life Science Publishing. Ikegami S, Kamimura M, Uchiyama S, et al. 2009 Bone mineral density measurement at both spine and hip for diagnosing osteoporosis in Japanese patients. J Clin Densitom 12: 337e344. Baim S, Binkley N, Bilezikian JP, et al. 2008 Official Positions of the International Society for Clinical Densitometry and executive summary of the 2007 ISCD Position Development Conference. J Clin Densitom 11:75e91. Siris ES, Chen YT, Abbott TA, et al. 2004 Bone mineral density thresholds for pharmacological intervention to prevent fractures. Arch Intern Med 164:1108e1112. Balseiro J, Fahey FH, Ziessman HA, Le TV. 1988 Comparison of bone mineral density in both hips. Radiology 167:151e153.
13. Yang RS, Chieng PU, Tsai KS, Liu TK. 1996 Symmetry of bone mineral density in the hips is not affected by age. Nucl Med Commun 17:711e716. 14. Rao AD, Reddy S, Rao DS. 2000 Is there a difference between right and left femoral bone density? J Clin Densitom 3:57e61. 15. Petley GW, Taylor PA, Murrills AJ, et al. 2000 An investigation of the diagnostic value of bilateral femoral neck bone mineral density measurements. Osteoporos Int 11:675e679. 16. Hamdy R, Kiebzak GM, Seier E, Watts NB. 2006 The prevalence of significant left-right differences in hip bone mineral density. Osteoporos Int 17:1772e1780. 17. Xu H, Gong J, Chen JX, et al. 2007 Bilateral femoral bone mineral density measurements in Chinese women and men. J Clin Densitom 10:165e169. 18. Alele JD, Kamen DL, Hermayer KL, et al. 2009 The prevalence of significant left-right hip bone mineral density differences among black and white women. Osteoporos Int 20:2079e2085. 19. Goh JC, Low SL, Bose K. 1995 Effect of femoral rotation on bone mineral density measurements with dual energy X-ray absorptiometry. Calcif Tissue Int 57:340e343. 20. Lekamwasam S, Lenora RS. 2003 Effect of leg rotation on hip bone mineral density measurements. J Clin Densitom 6: 331e336. 21. Yoshimura N, Campbell L, Hashimoto T, et al. 1998 Acetabular dysplasia and hip osteoarthritis in Britain and Japan. Br J Rheumatol 37:1193e1197. 22. Inoue K, Wicart P, Kawasaki T, et al. 2000 Prevalence of hip osteoarthritis and acetabular dysplasia in French and Japanese adults. Rheumatology (Oxford) 39:745e748. 23. Burger H, van Daele PL, Odding E, et al. 1996 Association of radiographically evident osteoarthritis with higher bone mineral density and increased bone loss with age. The Rotterdam Study. Arthritis Rheum 39:81e86. 24. Antoniades L, MacGregor AJ, Matson M, Spector TD. 2000 A cotwin control study of the relationship between hip osteoarthritis and bone mineral density. Arthritis Rheum 43:1450e1455. 25. White J, Harris SS, Dallal GE, Dawson-Hughes B. 2003 Precision of single vs bilateral hip bone mineral density scans. J Clin Densitom 6:159e162.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume 17, 2014