Recent Publications: Reviews and Commentaries by Christopher E. Cann, Ph.D.
Comparison of Noninvasive Bone Mineral Measurements in Assessing Age-Related Loss, Fracture Discrimination, and Diagnostic Classification
Grampp S, Genant HK, Mathur A, Lang P, Jergas M, Takada M, Gluer C-C, Lu Y, Chavez M
Over the years many comparisons of BMD measurement methods have been made to determine relative diagnostic sensitivity and specificity for osteoporosis. This latest comparison reaffirms the superiority of QCT over DXA measures for diagnosis of spinal osteoporosis, and adds several of the newer BMD modalities to broaden the comparison. Three groups of women, healthy premenopausal, healthy postmenopausal, and osteoporotic postmenopausal, were compared using 13 measurement sites and techniques (spinal QCT and DXA, hip DXA, radial DXA and pQCT, phalangeal and metacarpal RA, and calcaneal ultrasound). Osteoporosis was defined based on spinal fracture and did not include any analysis of peripheral osteoporotic fractures. Correlations among the techniques and sites ranged from 0.10-0.93. As expected for diagnosis of spinal fracture, the age-adjusted odds ratio was highest for trabecular QCT (4.3), with integral QCT (3.0), DXA LAT (2.6), and DXA PA (2.4) also being relatively high. Some peripheral BMD measures were also significantly predictive of spinal fracture (DXA TROC 2.2, DXA TOT, MID, and UD RAD 2.2, 2.1, 1.9, RA PHAL 2.0 and CALC US SOS and BUA 2.1 and 1.7). The other peripheral measures (other DXA hip sites, pQCT of radius, metacarpal RA) were not significantly predictive.
The diagnostic agreement among the different measurements was generally poor in classifying postmenopausal women as osteopenic (T-score <-2.0) or those with fractures as osteoporotic (T-score <-2.5), except for QCT, lateral DXA, and phalangeal RA. The authors conclude “Spinal QCT appeared to provide the most robust performance overall. The results derived from the various techniques were only moderately correlative, which precluded the prediction of one parameter from another parameter in the individual woman and caused disagreement in diagnostic classification in many women.”
In an accompanying editorial (Which Bone Density Measurement? by M. Kleerekoper and D. A. Nelson), their review of the article makes two primary points. First, any BMD measurement is helpful in assessing risk of fracture in groups of patients. The ROC analysis in the article shows this with values of 0.69 for pQCT and ultrasound up to 0.82 for spinal QCT; any method with a value above 0.5 in an ROC analysis is considered to provide “useful” information. The reviewers cite an abundance of studies indicating that peripheral BMD assesses spine and hip fracture risk reasonably well, with only the special case of hip BMD for prediction of hip fracture in elderly patients doing better. The data in the study, however, show an odds ratio of 4.3 for QCT, that is, a 4.3 times increase in risk of vertebral fracture for a 1 SD decrease in QCT BMD. This is even better than the hip fracture/BMD data, where the odds ratio is 2.7. The second point made is that none of the methods is sufficiently specific to identify the individual patient at highest risk for fracture (although they state “The suggestion that spinal QCT provides the best performance overall adds to the disappointment, since most clinicians cannot hope to have access to, or routinely use, this technology for the assessment of osteoporosis in most patients). They go on to say that because of the poor agreement among methods in classifying patients as osteoporotic or osteopenic, the selection of a single criterion for diagnosing osteoporosis is arbitrary and may not be appropriate for all skeletal sites and techniques, and that the primary purpose of measuring BMD should be to assess fracture risk in individual patients. The reviewers’ comment that diagnosis of vertebral fracture from BMD is a “who cares” question gently sidesteps the issue, that low QCT BMD dramatically increases the odds for osteoporosis.
While prediction of vertebral fracture risk is not the only goal, this article shows that QCT does provide the best discrimination of all the methods in detection of osteoporosis, and it also shows the most rapid change in BMD in the early postmenopausal years. Thus, it has the best chance of all the methods to assess fracture risk early in individual patients, at a time when prevention of bone loss can have a great impact on future risk of fracture.
Journal of Bone and Mineral Research 1997; 12:697-711
The effects of standardization and reference values on patient classification for spine and femur dual-energy x-ray absorptiometry.
Simmons A, Simpson DE, O’Doherty MJ, Barrington S, Coakley AJ
With the recent adoption of the WHO panel recommendation that “osteoporosis” be defined quantitatively, the importance of accurate reference population data has increased. Each of the major manufacturers of DXA equipment has, over the years, developed reference databases according to specific criteria, not all of which have been the same among manufacturers. This article provides a quantitative comparison of two methods for cross-comparison among manufacturers, with the aim of determining if a diagnosis of osteoporosis made on one machine (T-score less than -2.5) would also be made on other machines once cross-calibrations have been done.
The answer is maybe for the spine, and no for the femoral neck.
Cross-comparisons were made using the method of the International DXA Standardization Committee and the European Community's COMAC-BME program. 2497 women 22-90 years old were scanned using a Norland XR-26 Mark II for the study. Classifications by T-score were done according to the WHO recommendations, T-score -1 or greater = normal, -1 to -2.5 = osteopenia, and less than -2.5 = osteoporosis. For the spine measurements, the percentage of patients classified identically by the different cross-comparison methods and the different instruments was 69-98%, while for the femoral neck 35-98% identical classifications were obtained. For the spine, 9.6-21.1% of the patients were classified as osteoporotic, while the range for the femoral neck was 2.3–13.8%, excluding the recently-changed Hologic femoral neck measure.
In this study there is a correlation among methods in classifying patients as osteoporotic by spine DXA and by hip DXA (r=0.84). However, about 10% more patients are classified as osteoporotic by spine DXA than by hip DXA, whatever calibration method is used. No information on the distribution of ages of the patients is given in this article, so it is impossible to tell if this “bias” toward spinal osteoporosis is simply due to a preponderance of menopausal women, in whom spine BMD is lost faster than hip, or if this is a real “misclassification” effect. The common practice of scanning a patient by DXA at both the spine and hip is called into question by these data, especially for the younger population. The lifetime risk of osteoporotic fracture in women age 50 years is estimated by Melton to be about 40%, which suggests that the hip DXA results may not be very sensitive in the diagnosis of osteoporosis, except in the elderly population where the spine results are often compromised. Recently, the major DXA manufacturers have agreed to standardize at least one measurement in the proximal femur (the “total hip” measurement), which may help to give better agreement across instruments. However, this does not address the needs of the large number of clinicians who are trying to help their patients decide if they should be on therapy, and for whom the 2-4 fold disparity in “diagnosis” among the DXA methods analyzed here leaves them wondering of what value are the BMD measurements. As the authors state, “The creation of standardized reference data must be an important priority in order to harmonize patient management (but) the choice of standardization technique must be addressed in light of the results presented here.” Clearly more work needs to be done in order to give the clinician a BMD measurement which can be used in clinical decision-making.
Osteoporosis International; 7: 200-206, 1997
Impact of Spinal Degenerative Changes on the Evaluation of Bone Mineral Density with Dual Energy X-Ray Absorptiometry (DXA)
Rand Th, Seidl G, Kainberger F, Resch A, Hittmair K, Schneider B, Gluer CC, Imhof H
Spinal degenerative changes cause artifically elevated BMD values when DXA is used to measure the spine in many older patients, but the prevalence of this effect is not well documented. These authors studied 144 postmenopausal women aged 40-84 years referred for routine evaluation for suspected postmenopausal osteoporosis. After excluding women with compression fractures, 59% of the population had at least one degenerative factor; osteophytes, vascular calcifications, scoliosis, or osteochondrosis, with osteophytes in 66 of the 122 women without fractures. Probability for osteophytes increased from 20% at age 50 to 70% by age 75; probability for any degenerative change reached 90% by age 75. The authors estimated that presence of osteophytes increased BMD values by about 10% (or +1.0 T-score value). In addition, because osteophytes can show rapid growth, the authors recommended caution in interpreting followup exams in patients with degenerative changes. They also recommend longitudinal spinal radiographs in women over 60 or scanning the proximal femur instead of the spine in order to obtain useful results in these women.
Calcified Tissue International: (1997) 60:430-433