| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Correspondence |
We welcome this expression of interest in our paper, but take issue with the opening sentence. We did not say that heel densitometry cannot be used in diagnosis of osteoporosis, but have advised caution and selectivity in its use.
Our study assessed one instrument for heel densitometry against axial bone densitometry. While osteoporosis can only be diagnosed definitively from a tissue specimen, measurements of bone density are used to detect low bone density, with a definition of osteoporosis based on it [1] as a practical expediency. However, it is now apparent that there is only moderate agreement [2] between results of bone densitometry at axial and peripheral sites (as Dr Kullenberg notes) and that different physical methods, such as ultrasound and X-ray absorptiometry [3, 4], can give different results at the same site. We therefore expect that use of measurements from different sites, or using different methods, can place an individual subject in different diagnostic categories.
Dr Kullenberg quotes the paper by Siris et al [5], which also notes that bone densitometry at the hip is the best predictor of hip fracture; in view of the serious morbidity and the mortality associated with hip fracture in elderly people, we and others [1] place particular emphasis on estimating the risk of fracture at this site. We disagree to some extent with the inference drawn from that paper, that it showed prediction of hip fracture from heel densitometry to be as good as that from hip measurements. That paper actually states that for heel densitometry the area under the receiver operating characteristic (ROC) curve for a fracture was comparable with that observed in a previous study using hip bone density [6]. Similar areas under ROC curves do not necessarily imply that the diagnostic tests to which they relate will perform equally, because the shape of the curves may be different and the operating point (also referred to as threshold or cut-off) selected on each curve may also be different, resulting in different values for sensitivity and specificity in practice.
The data we have obtained and published concern evaluation of particular items of equipment from one manufacturer, and this is clear in the details given in our paper. Since our heel measurements used dual rather than single X-ray absorbtiometry, it is possible that their relationship to fracture risk could be different from that described by Siris et al. While we are aware of the limitations to generalization from our findings, there is evidence from similar studies (as noted above) with other equipment that has shown broadly similar results to ours.
In considering the significance of our results, we have examined not just the overall relationship between results from two instruments, but have also addressed the impact of testing on individuals and those advising on their healthcare. For them, discrepant results from different instruments apparently measuring a similar quantity are likely to be confusing. As we note, since peripheral measurements are considered inappropriate for monitoring treatment, and monitoring treatment is a common practice, this confusion is likely to occur when an axial scan follows an abnormal peripheral measurement. We have also addressed appropriate targeting of investigation methods (when there is a choice), taking into account disease prevalence in the population studied, and note the applicability of a peripheral, lower cost and relatively rapid investigation for a population with low prevalence of osteoporosis, or for those for whom investigation is not otherwise available. Those with an intermediate prevalence of the condition, for whom long-term and relatively expensive treatment is being considered, are less suitable for this type of approach.
Yours etc.,
Sunderland Royal Hospital, Sunderland, UK
Received for publication June 11, 2003. Accepted for publication July 17, 2003.
References
Related articles in BJR:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| BJR | DMFR | IMAGING | ALL BIR JOURNALS |
