The International Society for Clinical Densitometry

s 233 Journal of Clinical Densitometry Volume 8, 2005 on BMD and z-scores, particularly at younger ages. Use of variable SDs in pediatric reference ranges provide more accurate z-scores and improved assessment of skeletal health. TB BMD excluding the skull may provide a more sensitive indication of skeletal status and greater sensitivity to small BMD changes in pediatric subjects. Poster Number 5 Central DXA APPLICATION OF PRECISION ASSESSMENT RESULTS Robert Blank, MD, University of Wisconsin Osteoporosis Clinical Research Program Nellie Vallarta-Ast, William S Middleton Memorial Veteran’s Hospital; Karen Elver, Univeristy of Wisconsin Hostpial and Clinics; Brenda McCarney, University of Wisconsin Hosptial and Clinics; Diane Krueger, University of Wiscosin Osteoporosis Clinical Research Program; Mary Checovich, University of Wisconsin Osteoporosis Clinical Research Program; Xiaodan Wei, Department of Biostatistic, University of Wisconsin; Neil Binkley, University of Wisconsin Osteoporosis Clinical Research Program BMD changes over time are the foundation of clinical decision making for patients with osteoporosis and other metabolic bone diseases. Interpretation of serial BMD results depends on measurement precision and least significant change (LSC). The ISCD recommends each DXA center establish precision using patients representative of their clinical population. Here, we illustrate that LSC is a more complex and subjective concept than generally appreciated. The region being evaluated may affect precision. Vertebral body exclusion, recommended by ISCD when degenerative disease is present, decreases bone area, which should decrease precision. To determine the “penalty” arising from vertebral exclusion, we calculated precision for every subset of adjacent lumbar vertebrae in three patient populations, two were entirely female, the other largely male. Surprisingly, in men we found the L1–L3 LSC to be 0.037 g/cm2 while the L1–L4 is 0.047 g/cm2. It is notable that spinal degenerative disease is common in older men and degenerative changes are more frequent as one proceeds caudally. By contrast, in the two female samples, in whom spinal degenerative disease was less prevalent, the L1–L3 LSC was 0.045 and 0.035 g/cm2 compared with 0.042 and 0.027 g/cm2 at L1–L4. These data suggest that degenerative disease per se reduces DXA precision. These data support the ISCD recommendation that DXA precision be determined in a sample representative of the clinical population being measured. Additionally, they suggest that criteria for choosing the precision sample should be refined. Finally, these results suggest that the practice of applying a single LSC to individual patients with varying degrees of spinal degenerative disease requires further evaluation. Poster Number 6 Central DXA DXA SOFTWARE UPGRADES WHAT DO THEY REALLY MEAN? Jan M Bruder, MD, MD Associate Professor of Medicine, Division of Endocrinology, University of Texas Beatrice Cardenas, LVN, University Hospital; Glenn M Garcia, MD, Assistant Professor of Radiology, UTHSCSA At the 2001 ISCD position development conference, it was recommended that the lowest t-score of either the lumbar spine, total hip, femoral neck or trochanter be considered for the diagnosis of osteoporosis. It is unclear how often the t-score at the trochanter site is the lowest. However, in our experience the trochanter is seldom the lowest. Recently, we noted an increase in the diagnosis of osteoporosis at the trochanter in patients referred to our clinic. This prompted us to repeat some BMDs at our facility. The referral BMDs were measured on the GE Lunar Prodigy (PA + 41169) 8.10 software version. Our facility has the GE Lunar Prodigy (DF + 13520) 7.53. One example was a 31-yr-old premenopausal woman without risk factors for osteoporosis who suffered a compression fracture following a MVA. The diagnosis of osteoporosis was made based on the t-score of left hip trochanter of 2.7. The repeat BMD t-scores of the left trochanter was –1.3. No t-score at other sites fulfilled the criteria for osteoporosis. Binkley et al. recently reported at the 2004 ASBMR meeting that the software upgrade versions 7.0 to 8.6 resulted in lower t-scores than the software before 7.0 at both the trochanter and femoral neck. In response to this analysis, GE Lunar has revised the software with an upgrade patch to version 8.8. The discrepancy; however, between the t-scores in the above example, and other examples to be presented, is left unexplained since the NHANES reference data is in both software versions. Many questions have thus been raised regarding various software versions especially as they may result in over diagnosis and treatment of patients. In addition, the marked difference in the t-scores on two GE Lunar Prodigy machines with the same reference database raises much concern. Poster Number 7 Central DXA CHARACTERISTICS OF MEN WITH VFA-DETECTED FRACTURES Selected as Best Abstract by a Technologist Nellie Vallarta-Ast, RT(R), CDT, William S. Middleton Memorial Veterans Hospital Diane Krueger, University of Wisconsin Osteoprosis Clinical Research Center; Neil Binkley, University of Wisconsin Osteoprosis Clinical Research Center Densitometric vertebral fracture assessment (VFA) technology is useful in men. However, a substantial minority of men with VFA identified fractures have normal BMD. This report depicts characteristics of such a group. The study population was selected from 361 male veterans referred for routine clinical bone mass measurement at the Middleton VAMC. Of these men, 25 were identified who had grade-2 or -3 vertebral compression fractures using the Genant VSQ system but normal BMD at all three sites routinely imaged (L1-4 spine, proximal femur and midradius). Based on history obtained by the DXA technologist, these men were divided into three groups; prior low-trauma fracture (12), prior high-trauma fracture (five) and no known prior fracture (eight). Of the men without a history of hightrauma fracture, 80% (16/20) had historical risk factors contributing to skeletal fragility or a history of fracture with falling. The most common fracture risk factor was chronic corticosteroid use (6/16) followed by other medications or toxins associated with bone loss (e.g., heparin, chemotherapeutics, alcohol abuse or antiepileptics) in 5/16. An additional 5/16 were frequent fallers. Other noteworthy historical conditions included ulcerative colitis, hyperparathyroidism and hypovitaminosis D. In only four of 20 were no fracture risk factor identified by history. In conclusion, medical history detects fracture risk factors in the vast majority of men with VFA-identified fractures. The high prevalence of secondary causes of bone loss suggests that a metabolic bone disease evaluation may be prudent in men with VFA demonstrated fractures and no history of high trauma fracture, even when the BMD is normal. Poster Number 8 Central DXA FOREARM BONE DENSITY IN PATIENTS WITH HAND OSTEOARTHRITIS AG Stern, MD, Rheumatologist, McGuire Veterans Affairs Medical Center VI Petkov, MD, Research Assistant, McGuire Research Institute, Richmond VA; TPS Rao, MD, Chief Rheumatology, McGuire Veterans Affairs Medical Center, Richmond VA; D Disler, MD, Radiologist, Virginia Commonwealth University, Richmond VA Commonwealth Radiology, Richmond VA; P Carlson, PhD, Laboratory Scientist, McGuire Veterans Affairs Medical Center, Richmond VA; RA Adler, MD, Chief Endocrinology, McGuire Veterans Affairs Medical Center, Richmond VA and Professor of Internal and Preventive Medicine, Virginia Commonwealth University, Richmond VA Several studies have suggested that erosive hand osteoarthritis (EOA) may be a distinct clinical entity. The objective of this study is to determine if there is a difference in forearm (FA) BMD in the two OA groups and to investigate independent predictors of FA BMD in OA subjects. Study sample consisted of 61 Caucasian subjects (16 with EOA, 45 non-EOA). OA classification was based on hand radiographs. BMD was measured using a Hologic Delphi densitometer. We use student t-test to compare groups, ANCOVA to account for possible confounding, and multiple regression to a build predictive model. We studied as potential independent predictors age, gender, height, weight, OA group/radiographic OA-score, and Creactive protein. The mean age of EOA subjects was 70 (+ 7.3) yr, 81% female; for non-EOA the mean age was 68 (+ 8.5), 67% female. There was no significant difference in FA BMD in the two groups: 0.559 gm/cm2 (+ 0.099) in the EOA group and 0.567 (+ 0.108) in nonEOA subjects. Adjusting for covariates did not change the means. In the EOA group 6.3% had osteoporosis, 31.3% osteopenia, and 62.5% had normal t-score in the total forearm. The proportional distribution in nonEOA group was 17.8%, 22.2%, and 60.0%, respectively. Independent predictors of FA BMD were gender, age and radiographic OA-score, accounting for 56% of the variance.Thus although hand EOA has distinctive clinical features it is not associated with an increased prevalence of low bone mass in the forearm. 13_Abstracts 4/28/05 10:34 PM Page 233