We studied 112 healthy men and 261 healthy women aged 18-92 years, and 34 men and 73 postmenopausal women with vertebral fractures aged 45-90 years to determine (i) whether patients with vertebral fractures have shorter stature before fracture, and (ii) whether the difference between arm span and standing or sitting height can be used to identify patients with fractures. Arm span was measured by using a calibrated extended ruler. Standing height, sitting height and leg length were measured by using a Holtain stadiometer. The results were expressed in absolute term and standard deviation (SD) or Z-scores (mean+/-SEM). Advancing age was associated with decreased sitting height (r=-0.37 to -0.41, both P<0.01) and a trend towards decreased arm span (r=-0.12 to -0.17, P=0.06 and 0.07) in healthy men and women; leg length was independent of age in both sexes (r=-0.09 to -0.12, NS). In patients with vertebral fractures, sitting height was reduced in women (Z=-0.83+/-0.14 SD, P<0.01) and men (Z=-1.37+/-0.21 SD, P<0.01) but only the women had reduced leg length (Z=-0.46+/-0.15 SD, P<0.01) and arm span (Z=-0.76+/-0.15 SD, P<0.01). Univariate and multivariate analyses suggest that the predictive ability of the difference between arm span and standing or sitting height to identify patients with vertebral fractures is limited. We concluded that women, not men, with vertebral fractures may come from a population with short stature. The difference between arm span and standing or sitting height cannot be used to predict vertebral fracture risk.