of BB on AP hip radiographs are more prevalent in patients with a hip fracture. The investigational group (+hip fracture) was identified and a control group (- hip fracture) was matched by age. Inclusion criteria included: a) postmenopausal Caucasian women aged 65 or older, and b) presence of an AP radiograph of the hip taken for reasons not associated with high-energy trauma. Exclusion criteria included: a) previous medical/surgical conditions known to be associated with osteopenia or fracture, b) individuals who had ever received medications known to affect bone mineral density and c) previous hip fracture of any kind. The hip radiographs were manipulated so that only one side of the hip was visible so that the reader was blinded to the contralateral hip, hiding the fractured hip if present. The control and investigational group radiographs were intermixed and random- ized and independently evaluated by two experienced musculoskeletal radiolo- gists. The presence of BB at the intertrochanteric region of the proximal femur was scored as present or not present. There were 98 individuals in the control group and 92 in the investigational group. The mean age was 79.8 Æ 6.4 years and 79.9 Æ 6.6 years in the control and investigational groups, respectively. The prevalence of intertrochanteric BB in the control group was 38% - 39% and 53% - 75% in the investigational group. Regardless of the reader, BB are seen in a significantly higher percentage of women with a fracture (75% vs. 39%, p!0.001 or 53% vs. 38%, p50.041) as compared to those without a frac- ture. BB are seen similarly in patients with femoral neck (intracapsular) and in- tertrochanteric (extracapsular) fractures. The kappa statistic was used to assess agreement on the presence of bone bars between readers. The kappa statistic was 0.41 and p-value was !0.001 which is considered moderate agreement. BB are significantly predicted in the presence of hip fracture. The presence of bone bars on a hip radiograph should be considered a potential radiographic pre- dictor for hip fracture risk. 121 Assessment of Fracture Risk A HELPING OF HIP AND A SPRINKLING OF SPINE: THE BLENDED T-SCORE IS EQUIVALENT TO THE OFFSET ENHANCEMENT FOR FRACTURE PREDICTION WITH FRAX William Leslie, University of Manitoba Lisa Lix, University of Saskatchewan; Helena Johansson, Consulting statistician, Gothenburg; Eugene McCloskey, Osteoporosis Centre, Northern General Hospital, Sheffield; John Kanis, Collaborating Centre for Metabolic Bone Diseases, Sheffield Background: The FRAX Ò tool estimates 10-year probability of fracture based upon multiple clinical risk factors and an optional BMD measurement from the femoral neck (FN). Discordance between lumbar spine (LS) and FN T-scores is a source of confusion to some clinicians since the LS measurement is not an input variable for FRAX. A procedure for using the LS to enhance fracture risk assess- ment under the FRAX system adjusts FRAX probability based upon the T-score difference between the LS and FN (termed ‘‘offset’’). We examined an alternative approach using a ‘‘blended’’ T-score input to FRAX calculated as the weighted mean of the FN and LS T-scores (reflecting the relative proportion of non-vertebral to vertebral fractures as 3:1). Methods: The Manitoba BMD database was used to identify men and women over age 50 with valid LS T-score (with exclusions), FN T-score, and FRAX prob- abilities (N536,141). Major fracture probabilities were calculated as: FRAX (clin- ical), no BMD input; FRAX (hip): FN T-score; FRAX (blended): 0.75ÂFN T- score plus 0.25ÂLS T-score; FRAX (offset): FRAX (hip) with subsequent offset adjustment (10% per SD difference between LS and FN). Fracture outcomes were assessed from population-based administrative data (N52,316 with a major osteoporotic fracture). Results: ROC area under the curve (AUC) analysis was performed for predic- tion of major osteoporotic fracture. The AUC for FRAX (clinical) was 0.666 (95% CI 0.654-0.677) and increased for FRAX (hip) 0.696 (0.685-0.707), FRAX (blended) 0.697 (0.686-0.708), and FRAX (offset) 0.698 (0.687-0.709). There was a statistically significant increase in the integrated discrimination improve- ment with FRAX (blended) versus FRAX (hip), p!.001 but no difference be- tween FRAX (blended) and FRAX (offset), p50.946. There was extremely close agreement (r50.999) between probabilities obtained from FRAX (blended) and FRAX (offset) which closely approximated the line of identity: slope50.99, Y-intercept50.0 (Figure). Based upon 3 risk categories (!10%, 10-20%, O20%), the reclassification rates were: FRAX (femoral neck vs blended) 6.6%, FRAX (femoral neck vs offset) 6.9%, and FRAX (blended vs offset) 1.4%. FRAX (femoral neck) and FRAX (blended) showed equivalent calibration (grouped as risk quintiles). Conclusions: Major osteoporotic fracture probability from FRAX (blended) is quantitatively equivalent to FRAX (offset). Both approaches result in a small im- provement in fracture prediction compared with FRAX (hip). 122 Assessment of Fracture Risk PREVALENCE OF LOW BMDAND VITAMIN D DEFICIENCY IN PATIENTS WITH DEVELOPMENTAL DISABILITIES Philip May, Hunterdon Developmental Center Harpreet Chahal, Hunterdon Developmental Center; Anoja Warusawithana, UMDNJ-RWJMS; Robin May, Hunterdon Developmental Center; Sunil Wimalawansa, UMDNJ-RWJMS Introduction: The New Jersey Department of Health statistics indicate much lower hip fracture rates in the state for those older than 50 years than for patients with de- velopmental disabilities (DD) who reside in the state’s seven DD centers. In compar- ison with ambulatory healthy people, residents of the DD centers had approximately 30-fold increased incidence of hip fractures. The age-adjusted prevalence of osteo- porosis in New Jersey residents age 50 years and older is about 13% [confidence in- terval (CI), 12e14], in comparison with residents older than 85 years, which is 23% (CI, 19e26) http://nj.gov/health/senior/osteo/] (2005, BRFSS); For patients with DD in our center, the age-adjusted prevalence is approximately 66%. Thus, persons with DD have a much higher incidence of osteoporosis and fragility fractures. Results: Results from our study at the Hunterdon Developmental Center (HDC; Clinton, NJ), one of seven state-operated facilities for adults with DD, revealed only 110 of the 580 (19%) residents were receiving vitamin D supplements. The doses they were receiving were between 400 and 600 IU per day, and very few had undergone measurement of serum 25(OH)D levels, despite that low bone mineral density (BMD) and fractures are common in these residents. In our study, 80% of the 400 residents who had dual-energy X-ray absorptiometry (DXA) BMD evaluation were shown to have low BMD, with fragility fracture rates varying from 4.6% to 6.9% per annum. To alleviate these issues, we con- ducted a series of face-to-face and online continuing medical education (CME) programs in collaboration with UMDNJ-CCOE. Initially, a series of three online programs were offered to the staff physicians. Before the lecture series, 9 of 22 (41%) of the patients on one unit were receiving vitamin D supplements, and none (0%) had ever had serum 25(OH)D levels measured. After the CME series, serum 25(OH)D was determined in 100% (22/22). These educational activities at HDC led to additional vitamin D testing, initiation of vitamin D therapy, or in- creased dosages of vitamin D in 100% of this institutionalized patient population. Conclusion: Data suggest that vitamin D deficiency is very common in patients with developmental disabilities. The incidence of low bone mass is approximately 80%, and fracture rates are about 6% in this population. Direct teaching programs for physicians who care for this highly vulnerable population can be successful in changing their behavior. Annual Meeting Abstracts 2011 161 Journal of Clinical Densitometry: Assessment of Skeletal Health Volume 14, 2011