Are arm span measurements useful in the prediction of osteoporosis in postmenopausal women?

Abstract

A consensus report on osteoporosis in The Netherlands recommends general practitioners to use an arm span-height difference of at least 3 cm as one of the criteria for suspecting osteoporosis. In this study it was investigated how well this criterion discriminates between patients with established osteoporosis and controls in a group of postmenopausal women. When the Dutch general practitioners criterion of arm span-height difference exceeding 3 cm was applied in a logistic regression analysis to predict the probability of having osteoporosis, a resultant sensitivity of 58% and a specificity of 56% were found. In order to improve the predicted probability of osteoporosis, two more predictors in addition to the arm span-height difference were introduced in the logistic regression model. The first additional predictor related to information whether subject’s age was below or above 70 years, and the second whether the arm span was below or above 160 cm. All three predictors appeared to be highly significant. It was shown that the predicted probability of osteoporosis by this model could be considerably improved when the age category and the arm span category were also taken into account, leading to a sensitivity of 81% while the specificity amounts to 64%. This seems quite satisfactory for such a simple method and clearly improves on the original single criterion of arm span-height difference.