Usefulness of quantitative heel ultrasound compared with dual-energy X-ray absorptiometry in determining bone mineral density in chronic haemodialysis patients.

BACKGROUND

Reduced bone mineral density (BMD) is associated with renal osteodystrophy and osteoporosis in end-stage renal failure patients. Dual-energy X-ray absorptiometry (DXA) is the standard non-invasive method to assess BMD, but is not always widely available. Quantitative heel ultrasound (QUS) is a mobile, relatively inexpensive, easy to perform and radiation-free method which can predict fractures to the same extent as DXA. This study assessed the usefulness of QUS vs DXA in determining BMD in chronic haemodialysis patients.

METHODS

Patients had their BMD at the hip and spine measured by DXA (Lunar Expert). QUS of the left heel (McCue CubaClinical II machine) measured broadband ultrasound attenuation (BUA) and velocity of sound (VOS). Correlations between DXA and QUS parameters were calculated. Receiver operator characteristic (ROC) curves were plotted for BUA and VOS and used to define cut-off points for calculating sensitivities and specificities for BUA and VOS. Femoral neck BMD was applied as the standard for diagnosing osteoporosis (T< or =-2.5) and osteopaenia (T>-2.5 and < or =-1) by WHO criteria.

RESULTS

Eighty eight patients (45.5% women), mean age 58+/-17 years, were studied. A total of 19% and 49% had femoral neck BMDs in the 'osteoporosis' and 'osteopaenia' ranges, respectively. There were good correlations between hip BMD and QUS parameters (r=0.68-0.79, P<0.001). Areas under the ROC curves for BUA and VOS in diagnosing 'osteoporosis' were 0.86 and 0.80, respectively. BUA and VOS had sensitivities of 76 and 71% and specificities of 80 and 69%, respectively, for diagnosing 'osteoporosis'. The positive predictive values for BUA and VOS were 48 and 35%, respectively, and the negative predictive values were 93 and 91% respectively.

CONCLUSIONS

DXA and QUS parameters were significantly correlated. However, sensitivities and specificities of QUS parameters were not sufficiently high for QUS to be used simply as an alternative to DXA. The relatively high negative predictive values suggest that QUS may reliably screen out patients unlikely to have a BMD in the osteoporotic range. The relatively low positive predictive values, however, mean that subjects classified as osteoporotic using QUS require further investigations such as DXA to confirm the diagnosis.