Purpose: Osteoporosis affects approximately 200 million people worldwide. Dual-energy x-ray absorptiometry (DXA) is widely used for osteoporosis diagnosis. However, calcaneal quantitative ultrasound technology is less expensive than DXA, portable, and does not involve radiation. Our aim was to determine the sensitivity and specificity of calcaneal quantitative ultrasound relative to DXA for identifying patients who meet diagnostic criteria for osteoporosis.

Methods: We searched MEDLINE and other databases for articles that evaluated the accuracy of calcaneal quantitative ultrasound for identifying adult patients with DXA T-scores of less than -2.5 at the hip or spine; had at least 10 participants with and 10 participants without DXA-determined osteoporosis; and reported sensitivity and specificity. We used two methods to estimate sensitivity and specificity. Because most of the studies that evaluated the quantitative ultrasound index (QUI) parameter reported T-score cutoff thresholds for positive tests corresponding to each pair of sensitivity and specificity, we used random effects regression models to predict sensitivity and specificity of this parameter based on threshold. We also developed summary receiver operating characteristic (SROC) curves for the QUI and other parameters.

Results: Of 1724 potentially relevant articles, 21 met inclusion criteria. Sensitivity and specificity of the QUI parameter were modest as determined by regression models. For a QUI T-score threshold of -1, sensitivity and specificity were 80% and 58%, respectively. At a pretest probability of 22% (e.g., average risk 65 year old woman), the posttest probability of DXA-determined osteoporosis (when using a T-score cutoff threshold of -1) was 35% after a positive result and 9% after a negative result. For a T-score cutoff threshold of 0, sensitivity improved to 93%, but specificity declined to 19%. A SROC curve for the QUI parameter had an area under the curve of 0.75, and a maximum joint sensitivity and specificity of 70%. SROC curves for other quantitative ultrasound parameters revealed similar overall performance. When we compared a regression-derived ROC curve for the QUI parameter to the SROC curve for this parameter, the curves were very similar.

Conclusions: The sensitivity and specificity of calcaneal quantitative ultrasound at commonly used thresholds are too low to exclude or confirm DXA-determined osteoporosis. Until further evidence becomes available, it is premature to recommend use of calcaneal quantitative ultrasound in an evidence-based screening program for osteoporosis.