Twenty-five studies (n=6,676) were included in the review. The studies evaluated a number of parameters: broadband ultrasound attenuation (11 studies, n=2,519), speed of sound (4 studies, n=989), velocity of sound (4 studies, n=770), quantitative ultrasound index (11 studies, n=3,054) and stiffness parameter (4 studies, n=1,729).
Study quality (for studies that evaluated the quantitative ultrasound index parameter).
All of the included studies enrolled participants prospectively and 3 studies selected patients either consecutively or by random sampling. All studies had 30 or more participants with and without DXA-defined osteoporosis, and had completion rates greater than 90%. Most studies did not report the time elapsed between ultrasound and DXA, and no study reported that the ultrasound and DXA results were assessed independently.
Diagnostic accuracy.
The AUC was 0.76 (95% CI: 0.72, 0.79) for the quantitative ultrasound index parameter (11 studies), 0.77 (95% CI: 0.73, 0.81) for broadband ultrasound attenuation (11 studies), 0.74 (95% CI: 0.71, 0.77) for speed of sound and velocity of sound (4 studies), and 0.79 (95% CI: 0.71, 0.86) for stiffness (4 studies).
Subgroup analyses for the quantitative ultrasound index parameter gave an AUC of 0.76 (95% CI: 0.70, 0.82) for studies of women only, and an AUC of 0.75 (95% CI: 0.66, 0.82) for studies of postmenopausal women only.
Pre- and post-test probabilities of DXA-defined osteoporosis were tabulated by age group (50 to 59 years, 60 to 69 years, 70 to 79 years, and 80 years or older), and for three ultrasound T-score thresholds. For example: at a pre-test probability of 22% (65-year-old white woman at average risk), the post-test probability was 34% (95% CI: 26, 41) after a positive test and 10% (95% CI: 5, 12) after a negative test, using a T-score threshold of -1.