- Email: [email protected]
Comparison of Bone Mineral Density by Dual-Energy X-Ray Absorptiometry and Bone Strength by Speed-of-Sound Ultrasonography in Adults With Gaucher Disease
ARTICLE IN PRESS Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health, vol. ■, no. ■, 1–6, 2015 © Copyright 2015 by The International Society for Clinical Densitometry 1094-6950/■:1–6/$36.00 http://dx.doi.org/10.1016/j.jocd.2015.12.002
Original Article
Comparison of Bone Mineral Density by Dual-Energy X-Ray Absorptiometry and Bone Strength by Speed-of-Sound Ultrasonography in Adults With Gaucher Disease Eytan Baskin,† Tama Dinur, Ehud Lebel, Maayan Tiomkin, Deborah Elstein,* and Ari Zimran Gaucher Clinic and Department of Orthopedic Surgery, Shaare Zedek Medical Center, Affiliated with the HadassahHebrew University Medical School, Jerusalem, Israel
Abstract Patients with the lysosomal disorder Gaucher disease (GD) are at risk of osteoporosis and/or avascular necrosis, but to date, no adequate biomarkers are available to ascertain individual predilections. Bone mineral density by dual-energy X-ray absorptiometry (DXA) has traditionally been used to monitor trends. With the availability of a speed-of-sound (SOS) ultrasonography to assess bone strength/elasticity, we aimed to ascertain whether these modalities are complimentary or comparable so SOS, with no radiation risk, might be used more routinely as a potential biomarker. A prospective comparative study in adult GD patients undergoing routine follow-up of bone mineral density T- and Z-scores at forearm (FA), femoral neck, and lumbar spine, and SOS Z-scores at FA was initiated. Interpretation was by qualitative categorization of Z-scores. The kappa measure of agreement beyond chance was calculated between pairs of measurements and the McNemar test was then applied. This noninterventional trial (ClinicalTrials.gov Identifier: NCT02067247) was approved by the institutional ethics committee. There were 89 patients (ages 21–78 years, 61% female, 62% common Ashkenazi genotype, 18% splenectomized, and 18% with avascular necrosis/fractures). When comparing Z-scores at FA by DXA and SOS, only 39.3% correlated, while the remaining results were in disagreement; no trend was noted. Similarly, when comparing Z-scores at the femoral neck by DXA with those at FA by SOS, 44.9% of the results were in agreement; no trend was noted; and Z-scores at the lumbar spine by DXA with FA by SOS, 46% were in agreement and no trend was noted. DXA at the 3 sites did not track in the same direction or the same magnitude of difference with SOS at FA in adult patients with GD. Due to the fundamental differences between the 2 measurements and their clinical correlates, plus the lack of long-term follow-up to assess outcome, the potential added value of the measurements at the FA by SOS in patients with GD awaits further studies. Key Words: Avascular necrosis; bone mineral density; dual-energy X-ray absorptiometry; Gaucher disease; speed-of-sound ultrasonography.
Introduction Dual-energy X-ray absorptiometry (DXA) is a current standard of practice for assessment of bone mineral density (BMD) and derivative values such as risk of osteoporosis based on T-scores (1). Although DXA technology continues to evolve, new instruments and technologies have been introduced (2), making it necessary to document how these advances compare with earlier equipment and prior
Received 11/8/15; Revised 12/14/15; Accepted 12/16/15. *Address correspondence to: Deborah Elstein, PhD, Gaucher Clinic, Shaare Zedek Medical Center, 12 Bayit Street, Jerusalem 91031, Israel. E-mail: [email protected] † This study was performed by EB in partial fulfillment of the requirements for an MD degree at the Hadassah-Hebrew University Medical School, Ein-Karem, Jerusalem, Israel.
1
ARTICLE IN PRESS 2 densitometric evaluations. Among the newer technologies is the ultrasound-driven speed-of-sound (SOS) BeamMed system, which measures bone strength and elasticity (3) as distinct from BMD of the bony architecture. Patients with Gaucher disease (GD), one of the more common lysosomal storage diseases (4), have been shown to suffer from low BMD (5) with increased risk of osteonecrosis (5) and pathological fractures (6), but to date, no biomarker or technology has been able to accurately predict which patients are imminently at risk. Predicting skeletal damage because of osteoporosis at any site but also osteonecrosis at the femoral neck (FN) due to osteopenia or osteoporosis because of Gaucher cell infiltration in the bone marrow (4) has clinically relevant ramifications (7). Moreover, because the option of disease-specific treatment is very costly, whether with either intravenous enzyme replacement therapy (ERT) (8) or oral substrate reduction therapy (SRT) (4), its use is often predicating on the estimated risk of severe disease, particularly skeletal involvement. DXA and SOS (9) putatively measure 2 different components of the bone, and may be either complimentary or comparable. If it can be shown that SOS is correlated with GD bone parameters and is sensitive to changes in elasticity/ fragility and/or density over time because of ERT or SRT, then SOS would have a place in routine Gaucher evaluations. This is especially relevant because SOS does not involve radiation exposure and could be performed on an annual or semiannual basis in the clinic or medical office. SOS technology is based on the principle that ultrasound waves travel faster through bone than through soft tissue. The bone sonometer measures time elapsed between axially transmitted sound and its reception after traveling through the selected bone. With greater density and elasticity, speed of propagation is greater (10). BMD measured by DXA has long been accepted as the most valid predictor of fracture risk in the future and consequently of all the attendant morbidity and mortality that that may entail. All recent clinical trials of new management options for GD have included DXA as an exploratory or tertiary endpoint; both current patient registries include DXA evaluations for treated and untreated patients. Moreover, low bone density has been associated with other Gaucher-specific parameters of disease severity (5,6). Nonetheless, to date, no studies have proven whether BMD as reflective of calcium content and trabeculation and measured by DXA is the equivalent feature of bone that is evaluated in SOS measurements. Additionally, there have been no studies using the SOS modality in GD. Thus, the present study is a comparison of 2 methodologies that as yet have not been applied in tandem in patients with GD with the underlying premise that, should there be any correlations in the findings that implicate risk of osteoporosis and/or avascular necrosis (AVN) in GD patients, one could envision using sonographic SOS on a routine basis in GD to assess what has heretofore been assessable with DXA, the latter only at biannual intervals because of the radiation involved.
Baskin et al.
Methods The present study was planned as an observational prospective comparative study of modalities with a predefined cohort of adult GD patients comparing 2 methodologies not previously assessed in this manner in this population. The population included nonpregnant adult (>18 years) patients with GD who undergo routine annual monitoring at the Gaucher Clinic, which also includes DXA evaluation of BMD approximately once every 1 to 2 years. DXA was assessed using a standard Hologic DXA unit (Hologic, Bedford, MA). The routine DXA evaluation includes 3 sites: forearm (FA), FN, and lumbar spine (LS). In otherwise healthy populations, including postmenopausal women and the aged, LS is the standard site for assessment of T- and Z-scores of BMD and is the basis for recommendations for the prevention of osteoporosis and/ or treatment options. The FN is also usually assessed in otherwise healthy populations but is especially of interest in GD because the hip joint is the site of the most frequent cases of AVN in GD. The FA is only rarely informative relative to the LS and FN in GD. SOS was performed using a Sunlight machine (BeamMed Ltd., Sunlight Omnisense 8000 series, Petach Tikva, Israel). The SOS portion was described to the patients as a research study and that participation was optional; therefore, only those patients who volunteered for the SOS in addition to undergoing the routine requisite DXA evaluation on the same day were included in the study. The SOS site chosen was the FA only, although some other sites such as the tibia can be evaluated. The choice of FA only was to minimize the inconvenience to the patient with the rationale that, should this modality be incorporated as a routine evaluation, patient convenience would be an important consideration. Each modality (DXA and SOS) was performed by a single technician (MT and TD, respectively) for all the patients to eliminate issues of reproducibility because of operator-dependent variables. Daily calibration was performed for each modality; for DXA a standardized calibrating phantom was provided by the company. SOS was determined as the mean of 3 scans statistically similar as calculated by the software. Institutional Ethical Review Board (Helsinki Committee) approval was received and the study was registered (January 28, 2014) as a clinical trial (NCT02067247).
Statistical Analysis Raw data of BMD and T- and Z-scores are imputed by the respective modality software. Direct comparison of the absolute measurements and/ or of the T- and Z-scores using standard statistical approaches did not yield meaningful meta-analyses (not shown). This approach also disallowed division into subpopulations such as those with the common Ashkenazi Jewish genotype (N370S/N370S), which is predictive of
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015
ARTICLE IN PRESS BMD and SOS in Gaucher Adults milder disease and lower risk of osteoporosis relative to those patients with other genotypes (5); or those who had undergone splenectomy and were therefore at greater risk for bone disease (5) compared to those with intact spleen; or those receiving ERT for GD compared to patients whose disease was considered mild enough and did not require ERT. Therefore, upon consideration of the above, interpretation of the results was based on qualitative categorization into categories of normal, nearing risk, and at risk, approximating the World Health Organization T-scores but using the quantitative results of Z-scores. This choice was predicated on the fact that, to date, Z-scores have significance in GD patients by virtue of the impact of the disease on the bone mineral content as well as density that is usually more dramatically conveyed in Z-scores. In addition, there is a growing acceptance of the hypothesis that skeletal maturation is slower among patients with GD as evinced by delayed bone age and delayed eruption of teeth (4). Hence, T-scores that are predicated on peak bone mass at the chronological age of 25–30 years would conceivably not be accurate for patients with GD whose achievement of peak bone mass is delayed relative to the general population. Based on the empirical data, the percent of agreement between the technologies and/or between the measurements at the 3 different sites using the DXA method and the SOS at the FA only were calculated. The kappa measure of agreement beyond chance was employed to test the agreement between the pairs of measurements. The McNemar test was then applied to test whether in cases of disagreement there exists a trend.
Results There were 89 patients who signed informed consent and participated in the study. Table 1 summarizes the baseline characteristics of the patients. The age range was across 5 decades; there were more females; 62% of the patients were homozygous for the common Ashkenazi Jewish N370S mutation; 18% had
Table 1 Characteristics of the Patients (n = 89) Median age (years, range) Males : females, n (%) N370S/N370S, N370S/other, missing; n (%) Splenectomy, n (%) Avascular necrosis or fractures, n (%) Bone-building medications, n (%) Enzyme replacement therapy, n (%)
51 (21–78) 35 (39.3):54 (60.7) 55 (61.8), 32 (36), 2 (2.2) 16 (18) 20 (22.5) 3 (3.4) 59 (66.3)
3 undergone splenectomy; and the majority did not have a history of AVN and/or pathological fractures. Table 2 presents the distribution of the categorized Z-scores for the SOS at the FA with comparison to the Z-scores at the 3 anatomical sites using DXA: DXA measurements at each site were always compared to the SOS measurements at the FA, the only site measured by SOS. The table shows the percent of the entire cohort with BMD and SOS Z-score by categories of normal range, approaching at risk and at risk (≤−2.5) by anatomical site while recognizing that this nomenclature is comparable to that of T-scores (normal, osteopenia, and osteoporosis), it is utilized for Z-scores in an attempt to categorize the findings in a clinically meaningful way (see above) specific to GD. Table 2 shows the Z-scores divided into normal range, approaching at risk, and at risk by anatomical site using DXA and at the FA using SOS. Accordingly, at the FA, DXA Z-scores were in the normal range for 44% of the cohort, but SOS predicted that 58% of the cohort values were in the normal range. This dichotomy was further revealed in that the DXA Z-scores at the FA were in the approaching at risk, and at risk range for ~48% and ~8%, respectively, of the cohort but only in ~30% and ~11%, respectively, of the cohort using SOS Z-scores. Table 2 also shows that, overall, there were slightly higher percentages of approaching at risk, and at risk at the LS than at the FN as expected (5,7). In Table 2, the last 3 columns indicate comparison of the Z-scores by the 2 modalities. There was a range of 39%– 46% agreement between the DXA Z-score at the 3 sites relative to the SOS Z-score at the FA. This means that more than half of the time, the SOS FA Z-score was dissimilar to the DXA Z-score at any of the 3 sites including the FAto-FA comparison. When comparing FA to FA, 26% of the time, the DXA Z-score was better than the SOS and 35% of the time it was worse, so that ~60% of the time there was no comparable Z-score at the FA. For the FN using DXA relative to the FA using SOS, the percentages were approximately opposite with 35% better with DXA and 19% worse than DXA. At the LS, comparison to FA SOS showed 46% comparable values but the remainder of the comparisons were half of the time (27%) higher by DXA and half of the time (27%) lower by DXA. Table 3 presents the results of comparisons among the entire cohort and various subcohorts based on demographic characteristics as described by actual kappa scores and p values of the McNemar test. There were only 3 statistically significant results with regard to a (disagreement) trend, all comparing SOS Z-scores at the FA to the DXA Z-scores at the FN: (1) among patients who are compound heterozygotes, (2) in those with a history of AVN or pathological fractures, and (3) in patients receiving ERT. In all of these subcohorts, the DXA Z-scores were better relative to the SOS Z-scores. Admittedly, there may have been a slight effect of the need to join 2 groups of SOS Z-scores to perform the McNemar test. However, despite this, were one to compare the 3 subcohorts above, without
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015
ARTICLE IN PRESS
Table 3 FA by SOS Compared to Each of the 3 Sites Evaluated by DXA, Sequentially FA, FN, and LS: Results of Kappa Scores and p Values of the McNemar Test With Boldfaced Values Are Statistically Significant Note: Standard ranges for normal (>−1), approaching at risk (between −1 and −2.5), and at risk (≤−2.5) were employed for both DXA and SOS. Abbr: BMD, bone mineral density; DXA, dual-energy X-ray absorptiometry; SOS, speed of sound.
58.4 30.3 11.2 23/89 (25.8%) 35/89 (39.3%) 31/89 (34.8%) Not applicable Not applicable Not applicable 32/89 (35.9%) 40/89 (44.9%) 17/89 (19.1%) Not applicable Not applicable Not applicable 24/89 (26.9%) 41/89 (46%) 24/89 (26.9%) 7.9 0 6.7 48.3 29.2 33.7 43.8 70.8 59.6 Forearm Femoral neck Lumbar spine
Nearing risk Z-scores (%) Normal Nearing risk At risk Z-scores Z-scores Z-scores (%) (%) (%)
DXA
Normal Z-scores (%)
SOS
At risk Z-scores (%)
DXA better than SOS
DXA same as SOS
DXA worse than SOS
Baskin et al. Table 2 Percent of Cohort With BMD and SOS Z-Score Results by Categories of Normal Range, Nearing Risk, and at Risk by Anatomical Site; the Latter 3 Columns Compare DXA Z-Scores at Each of 3 Sites to SOS Z-Scores at the Forearm Only
4
All
Males (n = 35)
Females (n = 54)
N370S/ N370S (n = 55)
N370S/other (n = 32)
Intact spleen (n = 73)
Splenectomy (n = 16)
No bone disease (n = 69)
Avascular necrosis/ fractures (n = 20) Untreated (n = 30)
ERT (n = 59)
FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS
Kappa score
p Value
−0.031 −0.09 0.005 −0.079 −0.13 −0.016 0.004 −0.06 0.021 −0.113 0.04 −0.02 0.029 −0.24 −0.04 −0.042 −0.12 0.04 −0.073 0.04 −0.19 0.005 −0.02 0.072 −0.221 −0.14 −0.124 −0.118 −0.11 −0.007 −0.022 −0.06 0
0.095 0.101* 0.328 0.325 0.059* 0.196 0.157 0.564* 0.278 0.163 0.532* 0.572 0.469 0.05* 0.225* 0.253 0.25* 0.727 0.209 0.157* 0.206* 0.132 0.724* 0.542 0.695 0.013* 0.175 0.300 0.593* 0.611 0.330 0.02* 0.306
Abbr: ERT, enzyme replacement therapy; FA, forearm; FN, femoral neck; LS, lumbar spine; SOS, SOS, speed of sound. *SOS Z-scores were joined for the McNemar test (see Results section).
joining 2 groups of SOS Z-scores to perform the McNemar tests, the results would nonetheless not be statistically significant.
Discussion Only in 40% of patients did DXA Z-scores and SOS Z-scores coincide at the FA, and the remainder of the time they were not in a consistent pattern; that is, the DXA
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015
ARTICLE IN PRESS BMD and SOS in Gaucher Adults Z-scores were not consistently higher (closer to the normal range) or lower (closer to the ≤−2.5 range) than the SOS Z-scores. Moreover, this was despite the fact that the SOS Z-scores were in the normal range for the majority of patients tested, and this was considerably higher than the percentage based on DXA Z-scores. In less than half of the patients, there were Z-scores by DXA at the FN and at the LS that were in the same range as SOS Z-scores at the FA. Of note, there also was no overt pattern of the FN and LS DXA Z-scores relative to the SOS Z-scores at the FA, that is, consistently better or worse. The FN Z-scores by DXA being consistently better (fewer in the range between normal and, in this series, none in the at-risk range) is as has been reported in the literature regardless of the presence of ERT and other variables except splenectomy, which seems to increase the risk of AVN and particularly at the hip joint in GD (4,5). Further studies including other subpopulation analyses, and especially by age, might highlight some associations that were not evident in the current study. However, a recent study with a similar design in a (Malaysian) pediatric cohort found similar results (11). Comparison of the change in cortical bone thickness in conjunction with assessment of BMD was among the earliest attempts to understand the impact of diseasespecific therapy on skeletal parameters in GD (12). Partly, the rationale for the measurement of both appendicular and trabecular bones reflects the character of GD skeletal pathology that is putatively a result of bone marrow lipid storage, so cortical bone thickness is considered a separate compartment that might not be impacted to the same extent as the architectural trabeculation in the marrow. Yet, measurement of cortical bone thickness was imprecise because the relative position at the midshaft was not comparable at different ages and in different body types and because of the relative changes in bone length over many years of follow-up for documentation (12). It was also clear that in GD, as in otherwise healthy individuals, although osteoporosis (as ascertained by DXA) virtually always causes bone fragility, bone fragility may nonetheless be present in persons with normal BMD T-scores (13), whereas those with good BMD T- and Z-scores may experience AVN or fractures. Decreasing the risk of pathological fractures is in fact one of the goals of GD-specific interventions (14), and hence accurate prediction of those at risk would conceivably make those patients candidates for GD-specific therapy. Yet, to date there is no such biomarker. Therefore, the use of BMD T-scores as correlated with risk for fractures is also used in GD (5) as in otherwise healthy persons (15). Primarily, the risk prediction reflects abnormal values at the LS. In GD, however, there is the additional risk of developing AVN at the joints, particularly at the hip joint (16), and hence this is an additional and uniquely important site in GD. It is for this reason that both of these anatomical sites are of importance in GD. To date, DXA of LS and FN and the T- and Z-scores
5 are employed as part of the follow-up regimen (approximately every 1–2 years) of good clinical practice in GD (17). For monitoring the natural disease trajectory of the visceral features of GD and any changes with interventions, that is, the change in spleen and liver volumes before and after the advent of ERT or SRT, although computed tomography or magnetic resonance imaging (MRI) may be more accurate, ultrasonography is used on a more routine basis, mainly because it does not have radiation risk and is cheap and convenient for the patient and the operator. This is the kind of thinking that has led to interest in SOS technology for monitoring bone disease, which is included in the therapeutic goals of GD management (14) and which has been used successfully clinically in comparison to DXA at other sites (18) with good sensitivity and specificity for diagnosing osteopenia (19). Others have used FA DXA BMD for long-term followup in the general population of fracture risk (20) and specifically the incidence of hip fractures (21). DXA BMD and Z-scores at the FA show good correlation of results over time (22) and in some cases with more clinically relevant results than at other sites (23). Thus, although the FA is not the classic site to assess in patients with GD, there is reasonable experience that the FA can be informative with regard to fracture risk, thereby reducing radiation exposure time. Conversely, it has been suggested that SOS provides an assessment complimentary to that of BMD (9). The approach of DXA at the “classic” anatomical sites in conjunction with SOS at the FA would then parallel the earlier approach of measuring BMD and cortical bone thickness in GD to give added value to the concept that GD bone fragility/fracture risk and AVN are not pathologies that can be attributed to and measured by a single compartment in GD. The lack of a pattern or trends to the multiple comparisons that were performed in the current study does not allow for a clear understanding of the clinical positioning of SOS relative to DXA’s long use in GD as documented in the literature. One might view this as a qualitative deficiency in comparing bone strength/elasticity by SOS at a site that is not considered to be at risk in GD to bone density by DXA at sites that are at risk; but there also was a negative kappa score when comparing Z-scores at the FA only. It is important to mention that BMD is not the only modality for assessing the impact of GD on bones. The use of MRI-based data has long been recognized (24–27): quantitative chemical shift imaging (24,25) and the bone marrow burden scores derived from magnetic resonance images (26,27) provide clinically relevant information about the skeleton in GD by assessing the fat fraction, and both of these can be used in monitoring changes with therapeutic interventions. One may posit that bone elasticity as measured by SOS might be either complementary or more comparable in predictive value to bone marrow packed with Gaucher cells as evaluated by MRI.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015
ARTICLE IN PRESS 6 In conclusion, the present study highlighted the lack of statistical correlation between DXA outcome measures and SOS outcome measures, which may be interpreted as implying that these tools measure discrete compartments of bone in patients with GD. This is underscored by a lack of association when comparing the 2 modalities at the FA. However, because DXA BMD scoring has not been a perfect biomarker for risk of skeletal involvement in GD, further comparative studies over time are warranted using other modalities such as MRI and possibly in more patients with vertebral fractures and/or AVN.
Acknowledgments The Sunlight Omnisense 8000 machine for the measurements of the speed of sound was donated by BeamMed Ltd. via Mr. Gilad Zamir. We also gratefully acknowledge the diligence and patience of our biostatistician, Ms. Tali Bdolah-Abram (Hadassah-Hebrew University School of Medicine).
References 1. Baim S, Wilson C, Lewiecki EM, et al. 2006 Precision assessment and radiation safety for dual-energy X-ray absorptiometry: position paper of the International Society for Clinical Densitometry. J Clin Densitom 8:371–378. 2. Shepherd JA, Lu Y, Wilson K, et al. 2006 Cross-calibration and minimum precision standards for dual-energy X-ray absorptiometry: the 2005 ISCD Official Positions. J Clin Densitom 9(1):31–36. 3. Goossens L, Vanderoost J, Jaecques S, et al. 2008 The correlation between the SOS in trabecular bone and stiffness and density studied by finite-element analysis. IEEE Trans Ultrason Ferroelectr Freq Control 55(6):1234–1242. 4. Zimran A, Elstein D. 2010 Lipid storage diseases. In: Williams Hematology. 8th ed. Lichtman MA, Kipps T, Seligsohn U, et al., eds. New York: McGraw-Hill, 1065–1071. 5. Khan A, Hangartner T, Weinreb NJ, et al. 2012 Risk factors for fractures and avascular osteonecrosis in type 1 Gaucher disease—a study from the International Collaborative Gaucher Group (ICGG) Gaucher Registry. J Bone Miner Res 27(8):1839–1848. 6. Javier RM, Hachulla E, Rose C, et al. 2011 Vertebral fractures in Gaucher disease type I: data from the French “Observatoire” on Gaucher disease (FROG). Osteoporos Int 22(4):1255–1261. 7. Mistry PK, Weinreb NJ, Kaplan P, et al. 2011 Osteopenia in Gaucher disease develops early in life: response to imiglucerase enzyme therapy in children, adolescents and adults. Blood Cells Mol Dis 46(1):66–72. 8. Barton NW, Brady RO, Dambrosia JM, et al. 1991 Replacement therapy for inherited enzyme deficiency: macrophagetargeted glucocerebrosidase for Gaucher’s disease. N Engl J Med 324:1464–1470. 9. Olszynski WP, Brown JP, Adachi JD, et al. 2013 Multisite quantitative ultrasound for the prediction of fractures over five years of follow-up: the Canadian Multicentre Osteoporosis Study. J Bone Miner Res 28(9):2027–2034. 10. Ashman RB, Cowin SC, Van Buskirk WC, Rice JC. 1984 A continuous wave technique for the measurement of the elastic properties of cortical bone. J Biomech 17:349–361.
Baskin et al. 11. Chong KH, Poh BK, Jamil NA, et al. 2015 Radial quantitative ultrasound and dual energy x-ray absorptiometry: intermethod agreement for bone status assessment in children. Biomed Res Int 2015:232876. 12. Rosenthal DI, Doppelt SH, Mankin HJ, et al. 1995 Enzyme replacement therapy for Gaucher disease: skeletal responses to macrophage-targeted glucocerebrosidase. Pediatrics 96(4 Pt 1):629–637. 13. Fonseca H, Moreira-Gonçalves D, Coriolano HJ, Duarte JA. 2014 Bone quality: the determinants of bone strength and fragility. Sports Med 44(1):37–53. 14. Pastores GM, Weinreb NJ, Aerts H, et al. 2004 Therapeutic goals in the treatment of Gaucher disease. Semin Hematol 41(4 Suppl 5):4–14. 15. Warriner AH, Patkar NM, Curtis JR, et al. 2011 Which fractures are most attributable to osteoporosis? J Clin Epidemiol 64(1):46–53. 16. Katz K, Horev G, Grunebaum M, Yosipovitch Z. 1996 The natural history of osteonecrosis of the femoral head in children and adolescents who have Gaucher disease. J Bone Joint Surg Am 78(1):14–19. 17. Martins AM, Valadares ER, Porta G, et al. 2009 Recommendations on diagnosis, treatment, and monitoring for Gaucher disease. J Pediatr 155(4 Suppl):S10–S18. 18. Flöter M, Bittar CK, Zabeu JL, Carneiro AC. 2011 Review of comparative studies between bone densitometry and quantitative ultrasound of the calcaneus in osteoporosis. Acta Reumatol Port 36(4):327–335. 19. De Schepper J, Roggen I, Van Biervliet S, et al. 2012 Comparative bone status assessment by dual energy X-ray absorptiometry, peripheral quantitative computed tomography and quantitative ultrasound in adolescents and young adults with cystic fibrosis. J Cyst Fibros 11(2):119–124. 20. El Hayek J, Pronovost A, Morin S, et al. 2012 Forearm bone mineral density varies as a function of adiposity in Inuit women 40-90 years of age during the vitamin D-synthesizing period. Calcif Tissue Int 90(5):384–395. 21. Rosengren BE, Ahlborg HG, Gärdsell P, et al. 2012 Forearm bone mineral density and incidence of hip fractures in Swedish urban and rural men 1987-2002. Scand J Public Health 40(1):102–108. 22. Costa ML, Krupa FG, Rehder PM, et al. 2012 Forearm bone mineral density changes during postpartum and the effects of breastfeeding, amenorrhea, body mass index and contraceptive use. Osteoporos Int 23(6):1691–1698. 23. Wood K, Dhital S, Chen H, Sippel RS. 2012 What is the utility of distal forearm DXA in primary hyperparathyroidism? Oncologist 17(3):322–325. 24. Johnson LA, Hoppel BE, Gerard EL, et al. 1992 Quantitative chemical shift imaging of vertebral bone marrow in patients with disease. Radiology 182(2):451–455. 25. Hollak C, Maas M, Akkerman E, et al. 2001 Dixon quantitative chemical shift imaging is a sensitive tool for the evaluation of bone marrow responses to individualized doses of enzyme supplementation therapy in type 1 Gaucher disease. Blood Cells Mol Dis 27(6):1005–1012. 26. Maas M, van Kuijk C, Stoker J, et al. 2003 Quantification of bone involvement in Gaucher disease: MR imaging bone marrow burden score as an alternative to Dixon quantitative chemical shift MR imaging—initial experience. Radiology 229(2):554–561. 27. Robertson PL, Maas M, Goldblatt J. 2007 Semiquantitative assessment of skeletal response to enzyme replacement therapy for Gaucher’s disease using the bone marrow burden score. AJR Am J Roentgenol 188(6):1521–1528.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015
Original Article
Comparison of Bone Mineral Density by Dual-Energy X-Ray Absorptiometry and Bone Strength by Speed-of-Sound Ultrasonography in Adults With Gaucher Disease Eytan Baskin,† Tama Dinur, Ehud Lebel, Maayan Tiomkin, Deborah Elstein,* and Ari Zimran Gaucher Clinic and Department of Orthopedic Surgery, Shaare Zedek Medical Center, Affiliated with the HadassahHebrew University Medical School, Jerusalem, Israel
Abstract Patients with the lysosomal disorder Gaucher disease (GD) are at risk of osteoporosis and/or avascular necrosis, but to date, no adequate biomarkers are available to ascertain individual predilections. Bone mineral density by dual-energy X-ray absorptiometry (DXA) has traditionally been used to monitor trends. With the availability of a speed-of-sound (SOS) ultrasonography to assess bone strength/elasticity, we aimed to ascertain whether these modalities are complimentary or comparable so SOS, with no radiation risk, might be used more routinely as a potential biomarker. A prospective comparative study in adult GD patients undergoing routine follow-up of bone mineral density T- and Z-scores at forearm (FA), femoral neck, and lumbar spine, and SOS Z-scores at FA was initiated. Interpretation was by qualitative categorization of Z-scores. The kappa measure of agreement beyond chance was calculated between pairs of measurements and the McNemar test was then applied. This noninterventional trial (ClinicalTrials.gov Identifier: NCT02067247) was approved by the institutional ethics committee. There were 89 patients (ages 21–78 years, 61% female, 62% common Ashkenazi genotype, 18% splenectomized, and 18% with avascular necrosis/fractures). When comparing Z-scores at FA by DXA and SOS, only 39.3% correlated, while the remaining results were in disagreement; no trend was noted. Similarly, when comparing Z-scores at the femoral neck by DXA with those at FA by SOS, 44.9% of the results were in agreement; no trend was noted; and Z-scores at the lumbar spine by DXA with FA by SOS, 46% were in agreement and no trend was noted. DXA at the 3 sites did not track in the same direction or the same magnitude of difference with SOS at FA in adult patients with GD. Due to the fundamental differences between the 2 measurements and their clinical correlates, plus the lack of long-term follow-up to assess outcome, the potential added value of the measurements at the FA by SOS in patients with GD awaits further studies. Key Words: Avascular necrosis; bone mineral density; dual-energy X-ray absorptiometry; Gaucher disease; speed-of-sound ultrasonography.
Introduction Dual-energy X-ray absorptiometry (DXA) is a current standard of practice for assessment of bone mineral density (BMD) and derivative values such as risk of osteoporosis based on T-scores (1). Although DXA technology continues to evolve, new instruments and technologies have been introduced (2), making it necessary to document how these advances compare with earlier equipment and prior
Received 11/8/15; Revised 12/14/15; Accepted 12/16/15. *Address correspondence to: Deborah Elstein, PhD, Gaucher Clinic, Shaare Zedek Medical Center, 12 Bayit Street, Jerusalem 91031, Israel. E-mail: [email protected] † This study was performed by EB in partial fulfillment of the requirements for an MD degree at the Hadassah-Hebrew University Medical School, Ein-Karem, Jerusalem, Israel.
1
ARTICLE IN PRESS 2 densitometric evaluations. Among the newer technologies is the ultrasound-driven speed-of-sound (SOS) BeamMed system, which measures bone strength and elasticity (3) as distinct from BMD of the bony architecture. Patients with Gaucher disease (GD), one of the more common lysosomal storage diseases (4), have been shown to suffer from low BMD (5) with increased risk of osteonecrosis (5) and pathological fractures (6), but to date, no biomarker or technology has been able to accurately predict which patients are imminently at risk. Predicting skeletal damage because of osteoporosis at any site but also osteonecrosis at the femoral neck (FN) due to osteopenia or osteoporosis because of Gaucher cell infiltration in the bone marrow (4) has clinically relevant ramifications (7). Moreover, because the option of disease-specific treatment is very costly, whether with either intravenous enzyme replacement therapy (ERT) (8) or oral substrate reduction therapy (SRT) (4), its use is often predicating on the estimated risk of severe disease, particularly skeletal involvement. DXA and SOS (9) putatively measure 2 different components of the bone, and may be either complimentary or comparable. If it can be shown that SOS is correlated with GD bone parameters and is sensitive to changes in elasticity/ fragility and/or density over time because of ERT or SRT, then SOS would have a place in routine Gaucher evaluations. This is especially relevant because SOS does not involve radiation exposure and could be performed on an annual or semiannual basis in the clinic or medical office. SOS technology is based on the principle that ultrasound waves travel faster through bone than through soft tissue. The bone sonometer measures time elapsed between axially transmitted sound and its reception after traveling through the selected bone. With greater density and elasticity, speed of propagation is greater (10). BMD measured by DXA has long been accepted as the most valid predictor of fracture risk in the future and consequently of all the attendant morbidity and mortality that that may entail. All recent clinical trials of new management options for GD have included DXA as an exploratory or tertiary endpoint; both current patient registries include DXA evaluations for treated and untreated patients. Moreover, low bone density has been associated with other Gaucher-specific parameters of disease severity (5,6). Nonetheless, to date, no studies have proven whether BMD as reflective of calcium content and trabeculation and measured by DXA is the equivalent feature of bone that is evaluated in SOS measurements. Additionally, there have been no studies using the SOS modality in GD. Thus, the present study is a comparison of 2 methodologies that as yet have not been applied in tandem in patients with GD with the underlying premise that, should there be any correlations in the findings that implicate risk of osteoporosis and/or avascular necrosis (AVN) in GD patients, one could envision using sonographic SOS on a routine basis in GD to assess what has heretofore been assessable with DXA, the latter only at biannual intervals because of the radiation involved.
Baskin et al.
Methods The present study was planned as an observational prospective comparative study of modalities with a predefined cohort of adult GD patients comparing 2 methodologies not previously assessed in this manner in this population. The population included nonpregnant adult (>18 years) patients with GD who undergo routine annual monitoring at the Gaucher Clinic, which also includes DXA evaluation of BMD approximately once every 1 to 2 years. DXA was assessed using a standard Hologic DXA unit (Hologic, Bedford, MA). The routine DXA evaluation includes 3 sites: forearm (FA), FN, and lumbar spine (LS). In otherwise healthy populations, including postmenopausal women and the aged, LS is the standard site for assessment of T- and Z-scores of BMD and is the basis for recommendations for the prevention of osteoporosis and/ or treatment options. The FN is also usually assessed in otherwise healthy populations but is especially of interest in GD because the hip joint is the site of the most frequent cases of AVN in GD. The FA is only rarely informative relative to the LS and FN in GD. SOS was performed using a Sunlight machine (BeamMed Ltd., Sunlight Omnisense 8000 series, Petach Tikva, Israel). The SOS portion was described to the patients as a research study and that participation was optional; therefore, only those patients who volunteered for the SOS in addition to undergoing the routine requisite DXA evaluation on the same day were included in the study. The SOS site chosen was the FA only, although some other sites such as the tibia can be evaluated. The choice of FA only was to minimize the inconvenience to the patient with the rationale that, should this modality be incorporated as a routine evaluation, patient convenience would be an important consideration. Each modality (DXA and SOS) was performed by a single technician (MT and TD, respectively) for all the patients to eliminate issues of reproducibility because of operator-dependent variables. Daily calibration was performed for each modality; for DXA a standardized calibrating phantom was provided by the company. SOS was determined as the mean of 3 scans statistically similar as calculated by the software. Institutional Ethical Review Board (Helsinki Committee) approval was received and the study was registered (January 28, 2014) as a clinical trial (NCT02067247).
Statistical Analysis Raw data of BMD and T- and Z-scores are imputed by the respective modality software. Direct comparison of the absolute measurements and/ or of the T- and Z-scores using standard statistical approaches did not yield meaningful meta-analyses (not shown). This approach also disallowed division into subpopulations such as those with the common Ashkenazi Jewish genotype (N370S/N370S), which is predictive of
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015
ARTICLE IN PRESS BMD and SOS in Gaucher Adults milder disease and lower risk of osteoporosis relative to those patients with other genotypes (5); or those who had undergone splenectomy and were therefore at greater risk for bone disease (5) compared to those with intact spleen; or those receiving ERT for GD compared to patients whose disease was considered mild enough and did not require ERT. Therefore, upon consideration of the above, interpretation of the results was based on qualitative categorization into categories of normal, nearing risk, and at risk, approximating the World Health Organization T-scores but using the quantitative results of Z-scores. This choice was predicated on the fact that, to date, Z-scores have significance in GD patients by virtue of the impact of the disease on the bone mineral content as well as density that is usually more dramatically conveyed in Z-scores. In addition, there is a growing acceptance of the hypothesis that skeletal maturation is slower among patients with GD as evinced by delayed bone age and delayed eruption of teeth (4). Hence, T-scores that are predicated on peak bone mass at the chronological age of 25–30 years would conceivably not be accurate for patients with GD whose achievement of peak bone mass is delayed relative to the general population. Based on the empirical data, the percent of agreement between the technologies and/or between the measurements at the 3 different sites using the DXA method and the SOS at the FA only were calculated. The kappa measure of agreement beyond chance was employed to test the agreement between the pairs of measurements. The McNemar test was then applied to test whether in cases of disagreement there exists a trend.
Results There were 89 patients who signed informed consent and participated in the study. Table 1 summarizes the baseline characteristics of the patients. The age range was across 5 decades; there were more females; 62% of the patients were homozygous for the common Ashkenazi Jewish N370S mutation; 18% had
Table 1 Characteristics of the Patients (n = 89) Median age (years, range) Males : females, n (%) N370S/N370S, N370S/other, missing; n (%) Splenectomy, n (%) Avascular necrosis or fractures, n (%) Bone-building medications, n (%) Enzyme replacement therapy, n (%)
51 (21–78) 35 (39.3):54 (60.7) 55 (61.8), 32 (36), 2 (2.2) 16 (18) 20 (22.5) 3 (3.4) 59 (66.3)
3 undergone splenectomy; and the majority did not have a history of AVN and/or pathological fractures. Table 2 presents the distribution of the categorized Z-scores for the SOS at the FA with comparison to the Z-scores at the 3 anatomical sites using DXA: DXA measurements at each site were always compared to the SOS measurements at the FA, the only site measured by SOS. The table shows the percent of the entire cohort with BMD and SOS Z-score by categories of normal range, approaching at risk and at risk (≤−2.5) by anatomical site while recognizing that this nomenclature is comparable to that of T-scores (normal, osteopenia, and osteoporosis), it is utilized for Z-scores in an attempt to categorize the findings in a clinically meaningful way (see above) specific to GD. Table 2 shows the Z-scores divided into normal range, approaching at risk, and at risk by anatomical site using DXA and at the FA using SOS. Accordingly, at the FA, DXA Z-scores were in the normal range for 44% of the cohort, but SOS predicted that 58% of the cohort values were in the normal range. This dichotomy was further revealed in that the DXA Z-scores at the FA were in the approaching at risk, and at risk range for ~48% and ~8%, respectively, of the cohort but only in ~30% and ~11%, respectively, of the cohort using SOS Z-scores. Table 2 also shows that, overall, there were slightly higher percentages of approaching at risk, and at risk at the LS than at the FN as expected (5,7). In Table 2, the last 3 columns indicate comparison of the Z-scores by the 2 modalities. There was a range of 39%– 46% agreement between the DXA Z-score at the 3 sites relative to the SOS Z-score at the FA. This means that more than half of the time, the SOS FA Z-score was dissimilar to the DXA Z-score at any of the 3 sites including the FAto-FA comparison. When comparing FA to FA, 26% of the time, the DXA Z-score was better than the SOS and 35% of the time it was worse, so that ~60% of the time there was no comparable Z-score at the FA. For the FN using DXA relative to the FA using SOS, the percentages were approximately opposite with 35% better with DXA and 19% worse than DXA. At the LS, comparison to FA SOS showed 46% comparable values but the remainder of the comparisons were half of the time (27%) higher by DXA and half of the time (27%) lower by DXA. Table 3 presents the results of comparisons among the entire cohort and various subcohorts based on demographic characteristics as described by actual kappa scores and p values of the McNemar test. There were only 3 statistically significant results with regard to a (disagreement) trend, all comparing SOS Z-scores at the FA to the DXA Z-scores at the FN: (1) among patients who are compound heterozygotes, (2) in those with a history of AVN or pathological fractures, and (3) in patients receiving ERT. In all of these subcohorts, the DXA Z-scores were better relative to the SOS Z-scores. Admittedly, there may have been a slight effect of the need to join 2 groups of SOS Z-scores to perform the McNemar test. However, despite this, were one to compare the 3 subcohorts above, without
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015
ARTICLE IN PRESS
Table 3 FA by SOS Compared to Each of the 3 Sites Evaluated by DXA, Sequentially FA, FN, and LS: Results of Kappa Scores and p Values of the McNemar Test With Boldfaced Values Are Statistically Significant Note: Standard ranges for normal (>−1), approaching at risk (between −1 and −2.5), and at risk (≤−2.5) were employed for both DXA and SOS. Abbr: BMD, bone mineral density; DXA, dual-energy X-ray absorptiometry; SOS, speed of sound.
58.4 30.3 11.2 23/89 (25.8%) 35/89 (39.3%) 31/89 (34.8%) Not applicable Not applicable Not applicable 32/89 (35.9%) 40/89 (44.9%) 17/89 (19.1%) Not applicable Not applicable Not applicable 24/89 (26.9%) 41/89 (46%) 24/89 (26.9%) 7.9 0 6.7 48.3 29.2 33.7 43.8 70.8 59.6 Forearm Femoral neck Lumbar spine
Nearing risk Z-scores (%) Normal Nearing risk At risk Z-scores Z-scores Z-scores (%) (%) (%)
DXA
Normal Z-scores (%)
SOS
At risk Z-scores (%)
DXA better than SOS
DXA same as SOS
DXA worse than SOS
Baskin et al. Table 2 Percent of Cohort With BMD and SOS Z-Score Results by Categories of Normal Range, Nearing Risk, and at Risk by Anatomical Site; the Latter 3 Columns Compare DXA Z-Scores at Each of 3 Sites to SOS Z-Scores at the Forearm Only
4
All
Males (n = 35)
Females (n = 54)
N370S/ N370S (n = 55)
N370S/other (n = 32)
Intact spleen (n = 73)
Splenectomy (n = 16)
No bone disease (n = 69)
Avascular necrosis/ fractures (n = 20) Untreated (n = 30)
ERT (n = 59)
FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS FA vs FA FA vs FN FA vs LS
Kappa score
p Value
−0.031 −0.09 0.005 −0.079 −0.13 −0.016 0.004 −0.06 0.021 −0.113 0.04 −0.02 0.029 −0.24 −0.04 −0.042 −0.12 0.04 −0.073 0.04 −0.19 0.005 −0.02 0.072 −0.221 −0.14 −0.124 −0.118 −0.11 −0.007 −0.022 −0.06 0
0.095 0.101* 0.328 0.325 0.059* 0.196 0.157 0.564* 0.278 0.163 0.532* 0.572 0.469 0.05* 0.225* 0.253 0.25* 0.727 0.209 0.157* 0.206* 0.132 0.724* 0.542 0.695 0.013* 0.175 0.300 0.593* 0.611 0.330 0.02* 0.306
Abbr: ERT, enzyme replacement therapy; FA, forearm; FN, femoral neck; LS, lumbar spine; SOS, SOS, speed of sound. *SOS Z-scores were joined for the McNemar test (see Results section).
joining 2 groups of SOS Z-scores to perform the McNemar tests, the results would nonetheless not be statistically significant.
Discussion Only in 40% of patients did DXA Z-scores and SOS Z-scores coincide at the FA, and the remainder of the time they were not in a consistent pattern; that is, the DXA
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015
ARTICLE IN PRESS BMD and SOS in Gaucher Adults Z-scores were not consistently higher (closer to the normal range) or lower (closer to the ≤−2.5 range) than the SOS Z-scores. Moreover, this was despite the fact that the SOS Z-scores were in the normal range for the majority of patients tested, and this was considerably higher than the percentage based on DXA Z-scores. In less than half of the patients, there were Z-scores by DXA at the FN and at the LS that were in the same range as SOS Z-scores at the FA. Of note, there also was no overt pattern of the FN and LS DXA Z-scores relative to the SOS Z-scores at the FA, that is, consistently better or worse. The FN Z-scores by DXA being consistently better (fewer in the range between normal and, in this series, none in the at-risk range) is as has been reported in the literature regardless of the presence of ERT and other variables except splenectomy, which seems to increase the risk of AVN and particularly at the hip joint in GD (4,5). Further studies including other subpopulation analyses, and especially by age, might highlight some associations that were not evident in the current study. However, a recent study with a similar design in a (Malaysian) pediatric cohort found similar results (11). Comparison of the change in cortical bone thickness in conjunction with assessment of BMD was among the earliest attempts to understand the impact of diseasespecific therapy on skeletal parameters in GD (12). Partly, the rationale for the measurement of both appendicular and trabecular bones reflects the character of GD skeletal pathology that is putatively a result of bone marrow lipid storage, so cortical bone thickness is considered a separate compartment that might not be impacted to the same extent as the architectural trabeculation in the marrow. Yet, measurement of cortical bone thickness was imprecise because the relative position at the midshaft was not comparable at different ages and in different body types and because of the relative changes in bone length over many years of follow-up for documentation (12). It was also clear that in GD, as in otherwise healthy individuals, although osteoporosis (as ascertained by DXA) virtually always causes bone fragility, bone fragility may nonetheless be present in persons with normal BMD T-scores (13), whereas those with good BMD T- and Z-scores may experience AVN or fractures. Decreasing the risk of pathological fractures is in fact one of the goals of GD-specific interventions (14), and hence accurate prediction of those at risk would conceivably make those patients candidates for GD-specific therapy. Yet, to date there is no such biomarker. Therefore, the use of BMD T-scores as correlated with risk for fractures is also used in GD (5) as in otherwise healthy persons (15). Primarily, the risk prediction reflects abnormal values at the LS. In GD, however, there is the additional risk of developing AVN at the joints, particularly at the hip joint (16), and hence this is an additional and uniquely important site in GD. It is for this reason that both of these anatomical sites are of importance in GD. To date, DXA of LS and FN and the T- and Z-scores
5 are employed as part of the follow-up regimen (approximately every 1–2 years) of good clinical practice in GD (17). For monitoring the natural disease trajectory of the visceral features of GD and any changes with interventions, that is, the change in spleen and liver volumes before and after the advent of ERT or SRT, although computed tomography or magnetic resonance imaging (MRI) may be more accurate, ultrasonography is used on a more routine basis, mainly because it does not have radiation risk and is cheap and convenient for the patient and the operator. This is the kind of thinking that has led to interest in SOS technology for monitoring bone disease, which is included in the therapeutic goals of GD management (14) and which has been used successfully clinically in comparison to DXA at other sites (18) with good sensitivity and specificity for diagnosing osteopenia (19). Others have used FA DXA BMD for long-term followup in the general population of fracture risk (20) and specifically the incidence of hip fractures (21). DXA BMD and Z-scores at the FA show good correlation of results over time (22) and in some cases with more clinically relevant results than at other sites (23). Thus, although the FA is not the classic site to assess in patients with GD, there is reasonable experience that the FA can be informative with regard to fracture risk, thereby reducing radiation exposure time. Conversely, it has been suggested that SOS provides an assessment complimentary to that of BMD (9). The approach of DXA at the “classic” anatomical sites in conjunction with SOS at the FA would then parallel the earlier approach of measuring BMD and cortical bone thickness in GD to give added value to the concept that GD bone fragility/fracture risk and AVN are not pathologies that can be attributed to and measured by a single compartment in GD. The lack of a pattern or trends to the multiple comparisons that were performed in the current study does not allow for a clear understanding of the clinical positioning of SOS relative to DXA’s long use in GD as documented in the literature. One might view this as a qualitative deficiency in comparing bone strength/elasticity by SOS at a site that is not considered to be at risk in GD to bone density by DXA at sites that are at risk; but there also was a negative kappa score when comparing Z-scores at the FA only. It is important to mention that BMD is not the only modality for assessing the impact of GD on bones. The use of MRI-based data has long been recognized (24–27): quantitative chemical shift imaging (24,25) and the bone marrow burden scores derived from magnetic resonance images (26,27) provide clinically relevant information about the skeleton in GD by assessing the fat fraction, and both of these can be used in monitoring changes with therapeutic interventions. One may posit that bone elasticity as measured by SOS might be either complementary or more comparable in predictive value to bone marrow packed with Gaucher cells as evaluated by MRI.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015
ARTICLE IN PRESS 6 In conclusion, the present study highlighted the lack of statistical correlation between DXA outcome measures and SOS outcome measures, which may be interpreted as implying that these tools measure discrete compartments of bone in patients with GD. This is underscored by a lack of association when comparing the 2 modalities at the FA. However, because DXA BMD scoring has not been a perfect biomarker for risk of skeletal involvement in GD, further comparative studies over time are warranted using other modalities such as MRI and possibly in more patients with vertebral fractures and/or AVN.
Acknowledgments The Sunlight Omnisense 8000 machine for the measurements of the speed of sound was donated by BeamMed Ltd. via Mr. Gilad Zamir. We also gratefully acknowledge the diligence and patience of our biostatistician, Ms. Tali Bdolah-Abram (Hadassah-Hebrew University School of Medicine).
References 1. Baim S, Wilson C, Lewiecki EM, et al. 2006 Precision assessment and radiation safety for dual-energy X-ray absorptiometry: position paper of the International Society for Clinical Densitometry. J Clin Densitom 8:371–378. 2. Shepherd JA, Lu Y, Wilson K, et al. 2006 Cross-calibration and minimum precision standards for dual-energy X-ray absorptiometry: the 2005 ISCD Official Positions. J Clin Densitom 9(1):31–36. 3. Goossens L, Vanderoost J, Jaecques S, et al. 2008 The correlation between the SOS in trabecular bone and stiffness and density studied by finite-element analysis. IEEE Trans Ultrason Ferroelectr Freq Control 55(6):1234–1242. 4. Zimran A, Elstein D. 2010 Lipid storage diseases. In: Williams Hematology. 8th ed. Lichtman MA, Kipps T, Seligsohn U, et al., eds. New York: McGraw-Hill, 1065–1071. 5. Khan A, Hangartner T, Weinreb NJ, et al. 2012 Risk factors for fractures and avascular osteonecrosis in type 1 Gaucher disease—a study from the International Collaborative Gaucher Group (ICGG) Gaucher Registry. J Bone Miner Res 27(8):1839–1848. 6. Javier RM, Hachulla E, Rose C, et al. 2011 Vertebral fractures in Gaucher disease type I: data from the French “Observatoire” on Gaucher disease (FROG). Osteoporos Int 22(4):1255–1261. 7. Mistry PK, Weinreb NJ, Kaplan P, et al. 2011 Osteopenia in Gaucher disease develops early in life: response to imiglucerase enzyme therapy in children, adolescents and adults. Blood Cells Mol Dis 46(1):66–72. 8. Barton NW, Brady RO, Dambrosia JM, et al. 1991 Replacement therapy for inherited enzyme deficiency: macrophagetargeted glucocerebrosidase for Gaucher’s disease. N Engl J Med 324:1464–1470. 9. Olszynski WP, Brown JP, Adachi JD, et al. 2013 Multisite quantitative ultrasound for the prediction of fractures over five years of follow-up: the Canadian Multicentre Osteoporosis Study. J Bone Miner Res 28(9):2027–2034. 10. Ashman RB, Cowin SC, Van Buskirk WC, Rice JC. 1984 A continuous wave technique for the measurement of the elastic properties of cortical bone. J Biomech 17:349–361.
Baskin et al. 11. Chong KH, Poh BK, Jamil NA, et al. 2015 Radial quantitative ultrasound and dual energy x-ray absorptiometry: intermethod agreement for bone status assessment in children. Biomed Res Int 2015:232876. 12. Rosenthal DI, Doppelt SH, Mankin HJ, et al. 1995 Enzyme replacement therapy for Gaucher disease: skeletal responses to macrophage-targeted glucocerebrosidase. Pediatrics 96(4 Pt 1):629–637. 13. Fonseca H, Moreira-Gonçalves D, Coriolano HJ, Duarte JA. 2014 Bone quality: the determinants of bone strength and fragility. Sports Med 44(1):37–53. 14. Pastores GM, Weinreb NJ, Aerts H, et al. 2004 Therapeutic goals in the treatment of Gaucher disease. Semin Hematol 41(4 Suppl 5):4–14. 15. Warriner AH, Patkar NM, Curtis JR, et al. 2011 Which fractures are most attributable to osteoporosis? J Clin Epidemiol 64(1):46–53. 16. Katz K, Horev G, Grunebaum M, Yosipovitch Z. 1996 The natural history of osteonecrosis of the femoral head in children and adolescents who have Gaucher disease. J Bone Joint Surg Am 78(1):14–19. 17. Martins AM, Valadares ER, Porta G, et al. 2009 Recommendations on diagnosis, treatment, and monitoring for Gaucher disease. J Pediatr 155(4 Suppl):S10–S18. 18. Flöter M, Bittar CK, Zabeu JL, Carneiro AC. 2011 Review of comparative studies between bone densitometry and quantitative ultrasound of the calcaneus in osteoporosis. Acta Reumatol Port 36(4):327–335. 19. De Schepper J, Roggen I, Van Biervliet S, et al. 2012 Comparative bone status assessment by dual energy X-ray absorptiometry, peripheral quantitative computed tomography and quantitative ultrasound in adolescents and young adults with cystic fibrosis. J Cyst Fibros 11(2):119–124. 20. El Hayek J, Pronovost A, Morin S, et al. 2012 Forearm bone mineral density varies as a function of adiposity in Inuit women 40-90 years of age during the vitamin D-synthesizing period. Calcif Tissue Int 90(5):384–395. 21. Rosengren BE, Ahlborg HG, Gärdsell P, et al. 2012 Forearm bone mineral density and incidence of hip fractures in Swedish urban and rural men 1987-2002. Scand J Public Health 40(1):102–108. 22. Costa ML, Krupa FG, Rehder PM, et al. 2012 Forearm bone mineral density changes during postpartum and the effects of breastfeeding, amenorrhea, body mass index and contraceptive use. Osteoporos Int 23(6):1691–1698. 23. Wood K, Dhital S, Chen H, Sippel RS. 2012 What is the utility of distal forearm DXA in primary hyperparathyroidism? Oncologist 17(3):322–325. 24. Johnson LA, Hoppel BE, Gerard EL, et al. 1992 Quantitative chemical shift imaging of vertebral bone marrow in patients with disease. Radiology 182(2):451–455. 25. Hollak C, Maas M, Akkerman E, et al. 2001 Dixon quantitative chemical shift imaging is a sensitive tool for the evaluation of bone marrow responses to individualized doses of enzyme supplementation therapy in type 1 Gaucher disease. Blood Cells Mol Dis 27(6):1005–1012. 26. Maas M, van Kuijk C, Stoker J, et al. 2003 Quantification of bone involvement in Gaucher disease: MR imaging bone marrow burden score as an alternative to Dixon quantitative chemical shift MR imaging—initial experience. Radiology 229(2):554–561. 27. Robertson PL, Maas M, Goldblatt J. 2007 Semiquantitative assessment of skeletal response to enzyme replacement therapy for Gaucher’s disease using the bone marrow burden score. AJR Am J Roentgenol 188(6):1521–1528.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume ■, 2015