SEARCH

SEARCH BY CITATION

Keywords:

  • DXA;
  • bone mineral density inaccuracies;
  • nonuniform extraosseous fat distributions;
  • marrow;
  • anthropometrics

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Nonuniform extraosseous fat is shown to raise the magnitude of inaccuracies in DXA in vivo BMD measurements into the range of 20-50% in clinically relevant cases. Hence, DXA-based bone fragility diagnoses/prognoses and evaluations of bone responsiveness to treatment can be unreliable.

Patient-specific DXA in vivo bone mineral areal density (BMD) measurements have been demonstrated to be inherently inaccurate even when extraosseous fat (F) and lean muscle tissue (L) are uniformly distributed throughout the scan region of interest (ROI). The present work extends these investigations to quantitative evaluation of the extent to which clinically realistic soft tissue inhomogeneities external to the bone within the DXA scan ROI affect patient-specific in vivo BMD measurement inaccuracies. The results are particularly relevant to patient-specific lumbar vertebral and proximal femoral sites. Norland, Hologic, and Lunar DXA scans and corresponding DXA simulation studies of the same set of 225 different phantom arrays were carried out. The phantoms were specially fabricated absorptiometric replications of bone mineral material (B), red marrow (RM), and yellow marrow (YM) mixtures, and extraosseous F and L combinations spanning the anthropometric ranges encountered clinically. The three different DXA scanners yielded BMD results that effectively coincided, were in excellent agreement with the findings of the present corresponding DXA-simulation studies in each case, and confirmed the validity of the DXA BMD inaccuracy analysis formalism. It was found that only relatively small extraosseous soft tissue inhomogeneities within the ROI of DXA BMD scans can increase substantially the already sizable BMD inaccuracies shown earlier to pertain for uniformly distributed extraosseous soft tissues. The extent of these in vivo BMD inaccuracies (%) are shown to depend on the mean extraosseous F-to-L areal density ratio and its degree of nonuniformity within the local bone scan ROI, the marrow thickness and specific composition, and the actual BMD in any given case. It was found that patient-specific DXA-measured in vivo BMD inaccuracies can, in many clinically encountered cases, be as large as 20-50%, particularly so for osteopenic, osteoporotic, and elderly patients. It is concluded that, because these DXA in vivo BMD inaccuracies are unavoidable and clinically unpredictable, diagnoses/prognoses of bone fragility and evaluations of bone responsiveness to treatment of individual patients based mainly on DXA in vivo BMD measurements can be unreliable.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

THE INCREASING INCIDENCE of osteoporosis in the vulnerable population groups (1–3) has intensified the need for effective and reliable means of diagnosis and assessment of the condition and for the evaluation of the patient-specific therapeutic efficacy of its treatments. For more than a decade, the appraisal of skeletal bone mineral status in clinical practice and research has been based mainly on noninvasive planar DXA measurements of in vivo bone mineral areal density (BMD, g/cm2).(4) However, notwithstanding the widespread clinical use of, and extensive reliance on, DXA for clinical research, diagnostics of osteoporosis, and prognostics of the risk of osteoporotic fractures,(4–13) it is particularly disquieting that there is mounting evidence and serious concern that these measurements are subject to sizable inherent patient-specific inaccuracies.

These inaccuracies derive from the known inapplicability of planar DXA methodology to bone sites comprised of more than two absortiometrically distinguishable components within the scan region of interest (ROI)—the two-component DXA limitation.(14–26) The extent of these inaccuracies is sufficient to call into question a number of aspects of osteopenic and osteoporotic diagnoses/prognoses, bone fragility, remedial bone therapy effectiveness, and other related systematic features gleaned primarily from DXA in vivo BMD measurements. Faulkner(13) has reviewed the relevant evidence and explored the prospect that measured in vivo BMD, in and of itself, may not be homologous with bone strength. Marshall et al.(10) concluded that in vivo DXA measurement of BMD is not sufficiently definitive to be relied on to identify those specific individuals who will develop a future bone fracture. The likelihood that these inherent DXA in vivo BMD inaccuracies and unsettling clinical observations(10,13, 18, 19, 24–29) are linked(20–24, 30)—the former effectively manifest in the latter—looms as a fundamental issue relevant to the intrinsic viability and reliability of patient-specific planar DXA in vivo bone densitometry and bone fragility studies in clinical practice and research.

The principal objective of the present work was to extend the scope of these previous quantitative studies of inherent DXA in vivo BMD measurement inaccuracies by systematic investigation of the extent to which nonuniform distributions of extraosseous fat within the scan ROI further contribute to and affect these BMD measurement inaccuracies. It is noted that such a systematic and comprehensive study would not be feasible or practical using actual patients or in situ cadavers, because in neither could the true BMD and/or the requisite details of soft tissue composition and distribution within the scan ROI be ascertained(20) [the “true” value of BMD, (BMD)true, being that BMD value that would be obtained from a DXA in vivo measurement free of attendant inaccuracies]. Furthermore, in neither case could the investigator determine or preselect for interrogation any of these pertinent anthropometric and X-ray absorptiometric particulars to an extent sufficient to elucidate the underlying causative relationships fully. Thus, the use of realistic phantom arrays that closely replicate these soft tissue anthropometrics becomes essential in any thorough quantitative endeavor to establish and understand the fundamental causes of these inaccuracies and the extent to which they can affect patient-specific DXA in vivo BMD measurements. The present work differs from earlier, ostensibly similar, efforts (6, 14–19, 25, 26, 31–33) in that it develops directly from the underlying source of these inaccuracies and does so across the entire practical clinical range of BMD and soft tissue anthropometrics.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Phantom and DXA scans

Phantom materials formulated(23) as X-ray absorptiometric equivalents of bone material (B), red and yellow marrows (RM and YM, respectively), lean muscle tissue (L), and fat (F) were assembled into 225 different “anthropometric” arrays, as schematically represented in Fig. 1. These arrays spanned the full range of soft tissue anthropometrics encountered clinically, and as well, a broad range of nonuniform distributions of extraosseous fat and lean muscle tissue. The fabrication of these phantom tissues was based on the essentials of the loaded-epoxy resin method of White et al.(34) All phantom materials were identical in composition and overall dimensions to those used in our earlier publication.(23)

thumbnail image

Figure FIG. 1.. Schematic representation typical of the 225 different phantom assemblies used in the present work, indicating the bone position, marrow composition, direction of incident DXA scan X-rays (simulated and actual), and scan raster. Each extraosseous phantom block was a 4-cm cube; the overall A/P thickness of the represented “torso” was 20 cm. The F-to-L areal density ratio, q, designates the composition of the extraosseous soft tissue lateral to the “bone,” Q, that in the scan region in which bone material intercepts the X-ray flux, and y and x denote linear dimensional measures along X-ray paths in these regions, respectively. Bone material, yellow marrow (YM) and red marrow (RM) were 4 × 4-cm slabs of requisite thicknesses.

Download figure to PowerPoint

In each such phantom assembly, the extraosseous “soft tissue” material lateral to the “bone” was the absorptiometric equivalent of a specific preselected homogeneous F-to-L areal density ratio q = (ρFyFLyL); y is the linear anterior-posterior (A/P) dimension (cm) of the given extraosseous “tissue” along X-ray paths lateral to bone, and q is the lateral extraosseous soft tissue descriptor relevant to the extraction of BMD from DXA scans.(20–23) Along X-ray paths traversing “bone material” in the phantom array, the linear extent of the extraosseous soft tissue and its F-to-L areal density ratio are denoted by x and Q, respectively, so that Q = (ρFxFLxL). However, in this study, q and Q were generally not the same, and y was not equal to x.(23) The various separate values of q and Q used (0.0, 0.1, 0.3, 0.6, and 1.0) spanned the full range encountered clinically,(35) ranging from 0% to 50% F by weight in the scan ROI.

The thicknesses of phantom bone material inserted into the “bone site” of each phantom array were selected to be (BMD)true values of 0.6, 0.8, and 1.4 g/cm2, ranging(9, 36) from that typical of the osteoporotic, through osteopenic, to high BMD values, respectively. The remaining bone-site space was completely filled with various combinations of RM and YM phantom slabs. The fraction of the marrow volume that was yellow in each case, g = (YM)/(YM + RM) = 0.6, 0.8, and 1.0 (40% RM, 20% RM, and wholly YM, respectively), spanned the marrow constitutions typical of mature and elderly patients.(37, 38) In this way, phantom arrays replicating 225 different combinations of BMD and intra-/extraosseous soft tissue anthropometrics and absorptiometrics were assembled and DXA scanned.

As in our earlier work,(23) Norland XR-26 (Norland, Fort Atkinson, WI, USA), Hologic QDR-1000 (Hologic, Waltham, MA, USA), and Lunar DPX-α (Lunar, Madison, WI, USA) densitometers were used. All particulars of these DXA scans were as detailed earlier.(23)

DXA simulation studies of phantom arrays

As dual poly-energetic DXA simulated BMD values (after beam-hardening corrections were made) were in excellent quantitative agreement (< ∼0.2%) with actual DXA poly-energetic BMD measurement results and with dual mono-energetic DXA BMD simulation findings (intrinsically free of beam-hardening) for similar 20-cm-thick phantom arrays,(23) dual mono-energetic DXA BMD simulations of each of the phantom arrays were carried out.(20–22)

The BMD values of all simulated scans, (BMD)sim, were evaluated (as are the BMD values obtained in standard clinical DXA scans) using the analytic DXA equation,(20)

  • equation image(1)

where ln(J1/J01) and ln(I1/I01) represent the natural logarithms of the fraction of incident low-energy X-ray photon spectral region transmitted through the specimen lateral to and through the “bone,” respectively, and ln(J2/J02) and ln(I2/I02) are the corresponding values for the high-energy X-ray region, with λ1B and λ2B being, respectively, the low- and high-energy X-ray mass attenuation coefficients (cm2/g) of bone material at the two energies.

However, the DXA equation (Eq. 1) will yield the true value of BMD if, and only if, proper account can be taken of the presence of marrow (and its composition) and any differences that pertain between the makeup of the extraosseous fat/lean tissue in the X-ray “shadow” of the bone and lateral to it (q,Q differences).(20) As DXA cannot, (BMD)meas may differ markedly from (BMD)true, with DXA methodology unable to estimate the extent of these BMD inaccuracies. However, these inaccuracies can be quantitatively assessed in the present study, because in both the actual DXA scans and the corresponding simulation studies of each phantom array, all anthropometric and X-ray absorptiometric particulars, no matter how disparate, are known a priori.

BMD calibration

Fifteen additional similar phantom arrays were also assembled with the same standard intra- and extraosseous dimensions, each constituted with q = Q, and all marrow slabs replaced by material identical in composition to that of the extraosseous phantom tissue, q, of the same array. Because these assemblies were comprised of only two absortiometrically distinct materials (bone material plus a single intra-/extraosseous soft tissue equivalent), the two-component DXA limitation was satisfied fully, allowing DXA-measured BMD to be determined correctly (i.e., without inaccuracy). These 15 phantom assemblies were DXA scanned using all five q = Q values for each of the three selected phantom BMD values. Scans of these q = Q = g arrays established (calibrated) the three true (actual) BMD values, (BMD)true, of phantom bone material used in the other 225 phantom arrays. In this way, the corresponding overall percentage inaccuracies in both DXA-measured and normalized DXA-simulated BMD [(BMD)meas and (BMD)sim, respectively] were obtained for each of the 225 phantom array cases:

  • equation image(2)

Estimates of q, Q differences in DXA in vivo vertebral and proximal femoral scans

To obtain reasonable estimates of the extent to which the extraosseous soft tissue areal density ratios along X-ray paths passing lateral to (q) and those intersecting bone (Q) in planar DXA in vivo BMD scans actually differ within the scan ROI of lumbar vertebral and proximal femoral bone sites for most patients encountered clinically, a number of anatomical atlases were consulted. These depicted actual cross-sectional anatomical cadaveric slices,(39) computerized tomographic,(40–41) and magnetic resonance imaging(41) sections in which extraosseous adipose tissues (both subcutaneous and intraperitoneal) are distinguishable from lean muscle tissue and internal organs.

Careful measurements were made of the separate overall extraosseous fat and lean tissue dimensions along a large number of individual notional A/P X-ray paths passing lateral to and through the bone in each depicted contiguous transverse anatomical section image through the lumbar spinal and proximal femoral bone sites, as exemplified in Norland, Hologic, and Lunar DXA manuals and various texts.(42) In addition, the selected soft tissue ROI boundaries laterally adjacent to the bone were varied in breadth, and the new soft tissue anthropometrics encompassed were again evaluated. In each case, the separate average fat and lean tissue A/P dimensions were weighted by their respective volumetric densities,(23) ρF and ρL, to obtain the mean F-to-L areal density ratios, q and Q, and the (qQ) difference (positive and/or negative). In doing so, all nonfat tissues (internal organs, muscle tissue, etc.) present in the scan ROI were assigned the same volumetric density and linear X-ray attenuation coefficients of lean muscle tissue.(23) In this way, the difference between the average extraosseous F-to-L areal density ratio lateral to the bone and that along each notional X-ray path intersecting bone mineral material in the ROI (qQ) was obtained.

Because no transverse sections of particularly obese or exceedingly slender persons were depicted in the above noted anatomical atlases, the overall A/P torso and hip thicknesses of the full range of patients presenting for DXA BMD scans were not represented. To obtain the (qQ) value for any given q value that might reasonably be anticipated clinically, various A/P uniform fat layers were notionally added or removed from the depicted “normal” cross-sectional anatomical atlas slices, and new overall resultant q, Q, and (qQ) values were simply derived from them in each case. Because adult weight gain (loss) largely results from the accretion (loss) of subcutaneous (and/or visceral) fat, the thickness of the fat added (lost) seems to be spread relatively uniformly across the local scan ROI when viewed in A/P projection DXA BMD measurements of supine patients. On this basis, extrapolations from the “normal” 20-cm-thick cross-sectional atlas slices were transformed into reasonably realistic representations of A/P torso thicknesses ranging from 16 to 28 cm, effectively covering the full range of F-to-L areal density ratios and A/P torso thicknesses encountered clinically. These estimates of inhomogeneous fat distributions in the lateral and bone-shadow segments of DXA scans of these bone sites were found to be contained within the limits −0.10 ≤ (qQ) ≤ +0.03 in most cases, for all q values pertaining to A/P lumbar spinal and proximal femoral DXA scan regions of most patients presenting for BMD measurement.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The percentage inaccuracies found in the present DXA-measured BMD scans using Norland XR-26, Hologic QDR-1000, and Lunar DPX-α densitometers were effectively indistinguishable for each of the “anthropometrically” different 225 phantom arrays, as was also the case in our earlier phantom studies.(23) The results of the present investigation need therefore be displayed for any single densitometer (arbitrarily selected here for our Norland data only). Because the full span of the (qQ) differences DXA scanned and simulated in the present study may not be encountered clinically, the percentage BMD inaccuracies (Eq. 2) displayed in Figs. 2, 3, and 4 are largely restricted to those data and results corresponding to anthropometric factors [q, Q, (qQ), g, and (BMD)true] anticipated to be clinically relevant for adult patients (some results outside these ranges are also included to elucidate the trends and extent of inherent DXA BMD percentage inaccuracies that could result from whatever inhomogeneities in extraosseous soft tissue values, (qQ), might pertain, or be introduced, in some research or clinical studies). These data also make clear the sensitivity of the percentage BMD inaccuracies to inhomogeneous anthropometric and X-ray absorptiometric aspects of the various intra- and extraosseous soft tissues.

thumbnail image

Figure FIG. 2.. Results of the present series of quantitative simulation studies (solid lines) and actual Norland DXA scans (individual data point symbols) of a set of 75 different phantom arrays of the type depicted in Fig. 1 for a true BMD value of 0.6 g/cm2 displayed over the clinically relevant range of intra- and extraosseous soft tissues of adult patients. The upper segment of this figure relates to bone marrow with 60% yellow marrow, the middle section to marrow with 80% yellow marrow, and the lower section is the case of fully yellow marrow. The fractional BMD inaccuracies (%) inherent in DXA measured in vivo BMD are plotted as a function of Q, the extraosseous F-to-L areal density ratio within that section of the scan ROI section in which bone material is present. These are shown for representative extraosseous F-to-L areal density ratio values lateral to the bone, (A) q = 0, (B) q = 0.1, (C) q = 0.3, (D) q = 0.6, and (E) q = 1.0, with the actual measured DXA scan BMD inaccuracies designated by open squares, solid squares, open diamond shapes, solid diamond shapes, and open triangles, respectively. Demarcations of the limits −0.10 ≤ Q ≤ +0.03 found in the present work to represent the effective range of extraosseous soft tissue areal density ratio differences (qQ) encountered in most A/P DXA lumbar spine and proximal femoral BMD scans of typical patients are delineated by the upper and lower dotted curves, while the central dashed curve denotes the DXA BMD inaccuracies pertaining for the case of uniformly composed extraosseous tissue (qQ) = 0 within the scan ROI. Positive % inaccuracies overestimate the true BMD; negative BMD % inaccuracies are underestimates.

Download figure to PowerPoint

thumbnail image

Figure FIG. 3.. DXA BMD inaccuracy results (%) obtained for a set of 75 different phantom arrays similar to those of Fig. 2, but for the case of true BMD = 0.8 g/cm2. (Other relevant descriptive details given in caption of Fig. 2.)

Download figure to PowerPoint

thumbnail image

Figure FIG. 4.. DXA BMD inaccuracy results (%) obtained for a set of 75 different phantom arrays similar to those of Figs. 2 and 3, but for the case of true BMD = 1.4 g/cm2. (Other relevant descriptive details given in caption of Fig. 2.)

Download figure to PowerPoint

Also displayed in Figs. 2, 3, and 4 are those demarcations of the limits [−0.10 ≤ (qQ) ≤ +0.03] found in the present investigation to reasonably represent the range of extraosseous soft tissue areal density ratio differences (qQ) encountered in actual A/P DXA lumbar spine and proximal femoral BMD scans of typical patients. For example, in the particular case of an osteopenic individual [(BMD)true = 0.8 g/cm2] with a marrow composition 80% yellow (g = 0.8) and a lateral F-to-L areal density ratio q = 0.3, a DXA-measured BMD value may well underestimate the true BMD by between ∼14% [(BMD)meas ≈ 0.69 g/cm2] and ∼23% [(BMD)meas ≈ 0.62 g/cm2], leading to possible misdiagnosis of this patient as osteoporotic. For an osteoporotic individual, (BMD)true = 0.6 g/cm2, with the same soft tissue compositions (Fig. 2), the underestimates of the true BMD value range between ∼19% [(BMD)meas ≈ 0.49 g/cm2] and ∼31% [(BMD)meau ≈ 0.41 g/cm2].

It is also seen (Figs. 2, 3, and 4) that the red/yellow marrow compositional effect (g-dependence) on DXA-measured BMD inaccuracies still pertains, no matter the value of q, Q, (qQ), or (BMD)true.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The extent of these inaccuracies have been quantitatively delineated on a foundation of anatomically realistic phantom representation of in vivo soft tissue and bone anthropometrics and X-ray absorptiometrics relevant to the broad range of patients presenting for clinical DXA BMD scan procedures. Of particular importance, the present study has extended the earlier quantitative expositions(20–24) of DXA in vivo BMD inaccuracies to those more anatomically realistic cases in which the extraosseous fat and lean soft tissue mix is not uniform throughout the scan ROI. In particular, it has been shown here that the values of the extraosseous fat-to-lean soft tissue areal density ratios lateral to and in the bone-shadow regions, and their differences, (qQ), in the local bone site DXA scan ROI are anthropometric parameters directly relevant to and affecting the extent of BMD inaccuracies manifest in DXA-derived in vivo measurements. This is of some clinical importance, because it is clear that fat and lean tissues are not distributed uniformly throughout the body, and that the general distribution of soft tissues and accretion/loss of extraosseous fat is significantly different for children and adults, men and women, and pre and postmenopausal women.(43) Furthermore, the specific local bone-site F-to-L areal density ratio in the scan ROI may vary, sometimes considerably, as a function of age, physical condition, bone site, and other external factors such as the specific boundaries selected for the scan ROI. If the lateral segments of the ROI defining the extraosseous soft tissue areal density, q, can be varied by the DXA operator (or are DXA instrument dependent), somewhat different values for q and (qQ) could pertain for scans of a given bone site, thereby altering the effective BMD inaccuracy and the measured value of BMD. As such, measures of patient whole body fat-mass/total body-mass/body mass index, although indicative, may not necessarily be reliable or dependable descriptors of particular local bone site F-to-L areal density ratios within the scan ROI. Just as q is bone-site dependent, so too can be the quantity (qQ). In addition to other scan particulars, the relevant q and (qQ) values can also differ in A/P, lateral, and other DXA scan projections of the same bone site, leading to different BMD inaccuracies and, therefore, different measured BMD values. (It is noted here that the present study, by design, was unaffected by bone-edge detection concerns or uncertainties related to bone alignment relative to incident X-ray paths.)

It is also seen that within reasonable limits of extraosseous soft tissue disparity and composition (Figs. 2, 3, and 4), the inherent (BMD)meas inaccuracies in patient-specific DXA in vivo measurements can readily be expected to exceed 20% in clinically realistic cases. Particularly for osteoporotic patients (low true BMD), for postmenopausal and elderly individuals (marrow tending to higher yellow content in the lumbar vertebrae and proximal femoral regions),(37, 38) and for nonobese and lean patients (moderate and low relative fat content within and surrounding the lean tissues within the DXA scan ROI), the inaccuracies in DXA-derived BMD measurements can be quite large (30–50%). As a result, patient-specific DXA-measured in vivo BMD values can be seriously in error, more often than not considerably underestimating the true BMD of the very patients whose BMD it is most important for DXA to assess accurately. Given the reported 2-fold increase in population-based fracture risk per each SD decrement in BMD,(44) these sizable inaccuracies can be seen to insinuate confounding aspects into the clinical assessments of patient-specific fracture risk and diagnoses of osteoporosis based on the simple T-score criterion (T-score < −2.5).

As a consequence of the inability of planar DXA methodology to assess the absorptiometrics of both the extraosseous F-to-L areal density ratio in the bone-shadow region, Q, and the bone marrow, g, (one of the most labile of body tissues),(36–38, 45–47) the inaccuracies in DXA-measured in vivo BMD values may be both large and unpredictable in any given patient case. This is as expected, because the presence of marrow, irrespective of its specific red/yellow mix, g, effectively modifies the overall soft tissue areal density ratio (combined intra- and extraosseous tissues) along all X-ray paths traversing marrow within any given scanned bone. The marrow composition (largely affected through hematopoietic activity) can vary with patient age, disease, drug therapy, immobilization, physical activity regimen, bed confinement, infection, etc. This is the case particularly in the vertebrae and proximal extremities of appendicular bones where hematopoietic activity can be quite variable.(18, 37, 38, 45, 46) Similarly, the extraosseous fat-to-lean muscle tissue areal density ratio in the ROI of DXA BMD scans of patients can separately alter as a result of dietary regimens, hormonal changes, drug therapy, exercise programs, illness, etc. Consequently, changes in the soft tissue parameters, with or without accompanying true changes in bone mineral material, can be reflected in changes in DXA-measured BMD that may mislead interpretations of prospective BMD measurements of individual patients. Indeed, all DXA-measured in vivo BMD values will reflect this variability to some unknown extent, and as a result, may lead to interpretations and evaluations of diagnostic, prognostic, longitudinal, or cross-sectional BMD measurements that are seriously flawed.

In this context, it is important to stress that it is not the overall thickness or the difference in the linear thicknesses of fat alone or of lean tissues alone (or the linear extent of the body section alone) in either or both segments of the scan ROI (lateral to the bone and in the bone-shadow) that are relevant to the magnitude of inherent inaccuracies in DXA in vivo BMD measurements. Rather, it is the overall F-to-L areal density ratio, q, and the effective difference between the F-to-L areal density ratios (q and Q) along X-ray paths through the scan ROI that is of consequence. Thus, the local bone-site ratio of the F-to-L masses along the X-ray paths, and not the overall patient fat mass or lean mass or body mass singly, is the relevant soft tissue anthropometric parameter affecting DXA-measured in vivo BMD values. It is on these grounds that previous conflicting reports(48–61) specifying patient fat mass rather than body lean mass (and vice versa) to be the soft tissue parameter correlated strongly with DXA-measured BMD may be resolved, because neither one of these anthropometric quantities alone underpins the inherent DXA in vivo BMD inaccuracies from which these ostensible BMD correlations arise.

It is clear that when true BMD decreases, the bone must be more fragile to some extent (effectively the definition of “osteoporosis”). However, the present results have shown equally clearly that because of the sizable inherent, patient-specific DXA in vivo BMD measurement inaccuracies demonstrated here and in earlier works,(14–24, 30) a decrease (increase) in measured BMD may not necessarily signal any change in the bone fragility for a given individual. As a consequence of demonstrated large inherent systematic patient-specific DXA in vivo BMD inaccuracies, it is of utmost importance that it be recognized that DXA-measured BMD and true BMD are not necessarily synonymous. It is also difficult not to conclude from this and past investigations that introspective review and critical assessment of those aspects of present consensual knowledge of in vivo bone fragility and bone responsiveness to treatment based primarily on DXA BMD measurements are very much warranted.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

This research was supported in part by Grant 202011 of the National Health and Medical Research Council (Australia) and the Research Fund of Tampere University Hospital (Finland).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Cooper C, Campion C, Melton LJ III 1992 Hip fractures in the elderly: A worldwide projection. Ostoeporos Int 2:285289.
  • 2
    Poulsen LW, Qvesel D, Brixsen K, Verterby A, Boldsen JL 2001 Low bone mineral density in the femoral neck of medieval women: A result of multiparity? Bone 28:454458.
  • 3
    Kannus P, Niemi S, Parkkari J, Palvanen M, Vuori I, Järvinen M 1999 Hip fractures in Finland between 1970 and 1997 and predictions for the future. Lancet 353:802805.
  • 4
    Genant HK, Engelke K, Fuerst T, Glueer C-C, Grampp S, Harris S, Jergas M, Lang T, Lu Y, Majundar S, Mathur A, Takada M 1992 Noninvasive assessment of bone mineral and structure: State of the art. J Bone Miner Res 11:707730.
  • 5
    Dunn L, Wahner HW 1994 Commercial instruments available for bone mineral measurements. In: WahnerHW, FogelmanI (eds.) The Evaluation of Osteoporosis: Dual Energy X-ray Absorptiometry in Clinical Practice. Martin Dunitz, Ltd., London, UK, pp. 3548.
  • 6
    Wahner HW, Dunn WL, Brown ML, Morin RL, Riggs BL 1988 Comparison of dual-energy X-ray absorptiometry for bone mineral measurements of the lumbar spine. Mayo Clin Proc 63:10751084.
  • 7
    Jergas M, Breitneseher M, Gluer C-C, Yu W, Genant HK 1995 Estimates of volumetric bone density from projectional measurements improve the discriminatory capability of dual X-ray absorptiometry. J Bone Miner Res 10:11011110.
  • 8
    Hansen MA, Overgaard K, Riis BJ, Christiansen C 1991 Role of peak bone mass and bone loss in postmenopausal osteoporosis: 12 year study. BMJ 30:961964.
  • 9
    Hui SL, Slemenda CW, Johnston CC 1989 Baseline measurement of bone mass predicts fracture in white women. Ann Intern Med 114:355361.
  • 10
    Marshall D, Johnell O, Wedel H 1996 Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fracture. BMJ 312:12541259.
  • 11
    Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K, Genant HK, Palermo L, Scott J, Vogt TM for the Study of Osteoporotic Fractures Research Group 1993 Bone density at various sites for prediction of hip fracture. Lancet 341:7275.
  • 12
    Hui SL, Gao S, Zhou X-H, Johnston CC, Lu Y, Gluer C-C, Grampp S, Genant HK 1997 Universal standardization of bone density measurements: A method with optimal properties for calibration among several instruments. J Bone Miner Res 12:14631470.
  • 13
    Faulkner KG 2000 Bone matters: Are density increases necessary to reduce fracture risk? J Bone Miner Res 15:183187.
  • 14
    Webber CE 1987 The effect of fat on bone mineral measurements in normal subjects with recommended values of bone, muscle, and fat attenuation coefficients. Clin Phys Physiol Meas 8:143148.
  • 15
    Sorenson JA 1990 Effects of nonmineral tissues on measurements of bone mineral content by dual-photon absorptiometry. Med Phys 17:905912.
  • 16
    Tothill P, Pye DW 1992 Errors due to non-uniform distribution of fat in dual X-ray absorptiometry of the lumbar spine. Br J Radiol 65:807813.
  • 17
    Tothill P, Avenell A, Reid RM 1994 Precision and accuracy of measurements of whole-body bone mineral: Comparisons between Hologic, Lunar and Norland dual energy absorptiometers. Br J Radiol 67:12101217.
  • 18
    Kuiper JW, Van Kuijk C, Grashuis JL, Ederveen AGH, Schutte HE 1996 Accuracy and influence of marrow fat on quantitative CT and dual-energy X-ray absorptiometry measurements of the femoral neck in vitro. Osteoporos Int 6:2530.
  • 19
    Antonacci MD, Hanson DS, Hegeness MH 1996 Pitfalls in the measurement of bone mineral density by dual energy X-ray absorptiometry. Spine 21:8791.
  • 20
    Bolotin HH 1998 Analytic and quantitative exposition of patient-specific systematic inaccuracies inherent in planar DXA-derived in vivo BMD measurements. Med Phys 25:139151.
  • 21
    Bolotin HH 1998 A new perspective on the causal influence of soft tissue composition on DXA-measured in vivo bone mineral density. J Bone Miner Res 13:17391746.
  • 22
    Bolotin HH 1998 A new perspective on the effects of body weight and body fat mass on DXA-measured in vivo bone mineral density. Osteoporos Int 8:514.
  • 23
    Bolotin HH, Sievänen H, Grashuis JL, Kuiper JW, Järvinen 2001 Inaccuracies inherent in patient-specific dual-energy X-ray absorptiometry bone mineral density measurements: Comprehensive phantom-based evaluation. J Bone Miner Res 16:417426.
  • 24
    Bolotin HH, Sievänen 2001 Inaccuracies inherent in dual-energy X-ray absorptiometry in vivo bone mineral density can seriously mislead diagnostic/prognostic interpretation of patient-specific bone fragility. J Bone Miner Res 16:799805.
  • 25
    Formica C, Loro M-L, Gilsanz V, Seeman E 1995 Inhomogeneity in body fat distribution may result in inaccuracy in the measurement of vertebral bone mass. J Bone Miner Res 10:15041511.
  • 26
    Farrell TJ, Webber CE 1989 The error due to fat inhomogeneity in lumbar spine bone mineral measurements. Clin Phys Physiol Meas 10:5764.
  • 27
    Lochmüller E-M, Eckstein F, Kaiser D, Zeller JB, Landgraf J, Puts R, Steldinger R 1998 Prediction of vertebral failure loads from spinal and femoral dual-energy X-ray absorptiometry and calcaneal ultrasound: An in situ analysis with intact soft tissue. Bone 23:417424.
  • 28
    Svendsen OL, Hassager C, Skodt V, Christiansen C 1995 Impact of soft tissue on in vivo accuracy of bone mineral measurements in the spine, hip, and forearm: A human cadaver study. J Bone Miner Res 10:868873.
  • 29
    Bjarnason K, Hassager C, Svendsen OL, Stang H, Christiansen C 1996 Anteroposterior and lateral spinal DXA for the assessment of vertebral body strength: Comparison with hip and forearm measurement. Osteoporos Int 6:3742.
  • 30
    Bolotin HH 2001 Inaccuracies inherent in dual-energy X-ray absorptiometry in vivo bone mineral densitometry may flaw osteopenic/osteoporotic interpretations and mislead assessment of antiresorptive therapy effectiveness. Bone 28:548555.
  • 31
    Hangartner TN, Johnston CC 1990 Influence of fat on bone measurements with dual-energy absorptiometry. Bone Miner 9:7181.
  • 32
    Cullum ID, Ell PJ, Ryder JP 1989 X-ray dual-photon absorptiometry: A new method for measurement of bone density. Br J Radiol 62:587592.
  • 33
    Hassager C, Jensen SB, Gotfredsen A, Christiansen C 1991 The impact of measurement errors on the diagnostic value of bone mass measurements: Theoretical consideration. Osteoporos Int 1:250256.
  • 34
    White DR, Martin RJ, Darlison R 1977 Epoxy resin tissue substitutes. Br J Radiol 50:814421.
  • 35
    Wang J, Heymsfield SB, Aulet M, Thornton JC, Pierson RN Jr 1989 Body fat from body density: Underwater weighing vs. dual-photon absorptiometry. Am J Physiol 256:E829E834.
  • 36
    Dunnill MS, Anderson JA, Whitehead R 1967 Quantitative histological studies on age change in bone. J Pathol Bacteriol 94:275291.
  • 37
    Custer 1974 An Atlas of the Blood and Bone Marrow. Saunders, Philadelphia, PA, USA, pp. 3342.
  • 38
    Glowacki J 1995 Influence of age on human marrow. Calcif Tissue Int 35(Suppl 1):S50S51.
  • 39
    Spitzer VM, Whitlock DG 1998 National Library of Medicine Atlas of the Visible Human Male; Reverse Engineering of the Human Body. Jones and Bartlett Publishers, Sudbury, MA, USA.
  • 40
    Ellis H, Logan B, Dixon A 1991 Human Cross-Sectional Anatomy; Atlas of Body Sections and CT Images. Butterworth-Heinemann Ltd., Oxford, UK.
  • 41
    Cahill DR, Orland MJ, Miller GM, Bland JM, Altman DG 1995 Atlas of Human Cross-Sectional Anatomy with CT and MR Images, 3rd ed. Wiley-Liss, New York, NY, USA.
  • 42
    Wahner HW, Steiger P, von Setten E 1994 Instruments and measurements techniques. In: WahnerHW, FogelmanI (eds.) The Evaluation of Osteoporosis: Dual Energy X-Ray Absorptiometry in Clinical Practice. Martin Dunitz, Ltd., London, UK, pp. 1434.
  • 43
    Ley CJ, Lees B, Stevenson JC 1992 Sex- and menopause-associated changes in body-fat distribution. Am J Clin Nutr 55:950954.
  • 44
    Riggs BL, Melton LJ III 1992 The prevention and treatment of osteoporosis. N Engl J Med 327:620627.
  • 45
    Ricci C, Cova M, Kang YS, Yang A, Rahmouni A, Scot WW, Zerhouni EA 1990 Normal age-related pattern of cellular and fatty bone marrow distribution in the axial skeleton MR Imaging study. Radiology 177:8388.
  • 46
    Mazess RB 1983 Errors in measuring trabecular bone by computed tomography due to marrow and bone composition. Calcif Tissue Int 35:148152.
  • 47
    Minaire P, Edouard C, Arlot M, Meunier PJ 1984 Marrow changes in paraplegic patient. Calcif Tissue Int 36:338340.
  • 48
    Compston JE, Laskey MA, Croucher PL, Coxon A, Kreitzman S 1992 Effect of diet-induced weight loss on total body bone mass. Clin Sci (Lond) 82:429432.
  • 49
    Harris SS, Dawson-Hughes B 1996 Weight, body composition, and bone density in postmenopausal women. Calcif Tissue Int 59:428432.
  • 50
    Hystup L, Andersen T, McNair P, Breum L, Trasbol I 1993 Bone metabolism in obesity: Changes to severe overweight and dietary weight reduction. Acta Endocrinol 129:393398.
  • 51
    Ravn P, Cizza G, Bjarnason H, Thompson F, Daley M, Wasnich RD, McClung M, Hosking D, Yates AJ, Christiansen C for the Early Postmenopausal Intervention Cohort (EPIC) Study Group 1999 Low body mass index is an important risk factor for low bone mass and increased bone loss in early postmenopausal women. J Bone Miner Res 14:16221627.
  • 52
    Jensen LB, Quaade F, Sorensen OH 1994 Bone loss accompanying voluntary weight loss in obese humans. J Bone Miner Res 9:459461.
  • 53
    Reid IR, Evans MC, Ames RW 1994 Volumetric bone density of the lumbar spine is related to fat mass but not lean mass in normal postmenopausal women. Osteoporos Int 4:362367.
  • 54
    Bevier WC, Wiswell RA, Pyka G, Kozak KC, Newhall KM, Marcus R 1989 Muscle strength, physical fitness and weight, but not age predict femoral neck bone mass. J Bone Miner Res 4:441448.
  • 55
    Chen Z, Lohman TG, Stini WA, Ritenbaugh C, Aickin M 1997 Fat or lean tissue mass: Which is the major determinant of bone mineral mass in healthy postmenopausal women? J Bone Miner Res 12:144152.
  • 56
    Glauber HS, Vollmer WM, Nevitt MC, Ensrud KE, Orwoll ES 1995 Body weight versus body fat distribution, adiposity, and frame size as predictors of bone density. J Clin Endocrinol Metab 80:11181123.
  • 57
    Liel Y, Edwards J, Shary J, Spicer KM, Gordon L, Bell NH 1988 The effects of race and body habitus on bone mineral density of the radius, hip and spine in premenopausal women. J Clin Endocrinol Metab 66:12471250.
  • 58
    Pritchard JE, Nowson CA, Wark JD 1996 Bone loss accompanying dietary induced or exercise induced weight loss: A randomized controlled study. Int J Obes Relat Metab Disord 20:513523.
  • 59
    Reid IR, Plank LD, Evans MC 1992 Fat mass is an important determinant of whole body bone density in premenopausal women but not in men. J Clin Endocrinol Metab 75:779782.
  • 60
    Slemenda CW 1995 Body composition and skeletal density: Mechanical loading or something more? J Clin Endocrinol Metab 80:17611763.
  • 61
    Revilla M, Vill LF, Sanchez-Atrios A, Hernandez ER, Rico H 1997 Influences of body mass index on the age-related slope of total regional bone mineral content. Calcif Tissue Int 61:13451348.