Evaluation of calcaneus bone densitometry against hip and spine for diagnosis of osteoporosis.

Dual X-ray absorptiometry (DXA) is now widely available as a method for assessing bone density. However, the place of peripheral bone densitometry in clinical practice for diagnosis of osteoporosis is not yet clear. To examine the potential use in our district general hospital setting, we compared calcaneus measurements with conventional DXA of the hip and spine in 100 patients referred for assessment following identification of risk factors for osteoporosis. Measurements were made on both heels and the results were found to be similar but not completely interchangeable. Use of receiver operating characteristic curves confirmed that a threshold T-score of -1.6 could be used to identify many of the high risk subjects. However, there was only moderate agreement between fracture risk classifications derived from heel T-scores, and diagnostic classification (osteoporosis/osteopenia/normal) derived from axial DXA. The specificity of heel measurements was high, but sensitivity was poorer. Heel measurements could therefore be valuable in some circumstances for finding patients for whom treatment of osteoporosis would be appropriate, such as in a population with a low prevalence of osteoporosis. They may also be of value in a population with a high prevalence of disease, particularly if there were no alternative means of bone densitometry. However, with an intermediate prevalence, the relatively high risk of false negative values would mean that false reassurance could be given to many of those classed as "low risk". This could be a major drawback in clinical practice if heel densitometry were used as the initial investigation and axial measurements were also available, since they would give conflicting results for a substantial proportion of these patients.