The aims of this study were to ascertain vertebral deformity prevalence in elderly men and women and to describe the association between bone mineral density (BMD) at the lumbar spine and femoral neck, severity of spinal degenerative disease and vertebral deformity prevalence. We performed standardized spinal radiographs in a random sample of 300 elderly men and women participating in the Dubbo Osteoporosis Epidemiology Study, a population-based study of fracture risk factors. Radiographs were read independently by masked observers for the prevalence of vertebral deformity and severity of osteophytosis. BMD was measured by dual-energy X-ray absorptiometry. The prevalence of vertebral deformities was critically dependent on the criterion used. The less strict criteria seemed to overestimate deformities at either end of the spine region analysed. However, irrespective of the criterion used, prevalence of deformity was higher in men than in women (25% vs 20% for the 3 SD criterion, 17% vs 12% for the 4 SD criterion and 27% vs 25% for the 25% criterion). Femoral neck BMD was more strongly associated with vertebral deformities than spinal BMD for the 25% criterion (OR/SD change in BMD 1.39 (p=0.02) vs 1.20 (p=0.19)), 3 SD criterion (OR/SD change in BMD 1.45 (p=0.01) vs 1.10 (p=0.34)) and 4 SD criterion (OR/SD change in BMD 1.98 (p=0.0002) vs 1.68 (p=0.008)). BMD was also more strongly associated with biconcave deformities than either wedge or crush deformities and more so in men than in women. Severity of spinal osteophytosis was not associated with vertebral deformity. In conclusion, femoral neck BMD is at least equivalent to the lumbar spine BMD in strength of association with prevalent vertebral fractures. Spinal osteophytosis falsely elevates BMD without a concomitant decrease in fracture risk, indicating that any interpretation of spinal BMD needs to be adjusted for osteophytosis. These findings support the use of femoral neck bone densitometry in older men and women. Moreover, these data indicate that current criteria for radiological assessment of vertebral deformity are sufficiently loose to include a substantial proportion of non-fractures in the elderly, with important implications for the design of clinical trials. However, irrespective of the criterion used, vertebral deformities in men are at least as common, if not more so, than in women, suggesting that vertebral osteoporotic fractures are overlooked in men.