Comparative analysis of bone outcomes between quantitative ultrasound and dual-energy x-ray absorptiometry from the UK Biobank cohort

SUMMARY: This large cohort study investigated reliability and validity of heel ultrasound to estimate bone mineral density in adults. Reliability calculated between left and right heels was relatively poor and so was criterion validity assessed relative to dual-energy X-ray absorptiometry. Heel ultrasound should be used cautiously when estimating bone mineral density. INTRODUCTION: Calcaneal quantitative ultrasound (QUS) may be used as a safe, low cost, and portable means to estimate bone mineral density (BMD) in large cohorts. The purpose of this study was to quantify the reliability and validity of QUS in comparison to dual-energy X-ray absorptiometry (DXA), which is the reference method for BMD measurement and diagnoses of osteopenia and osteoporosis. METHODS: Bone outcomes measured on the large UK Biobank cohort were used. The reliability of QUS estimated BMD was quantified by comparing values obtained from the left and right heel measured in the same session. Criterion validity was assessed through agreement between QUS and DXA, quantifying correlations, and sensitivity and specificity of osteopenia and osteoporosis diagnoses. RESULTS: Reliability calculations were made using data from over 216,000 participants demonstrating similar QUS BMD values between left and right heels in the absolute scale (Sd of difference for men: 0.12 and 0.07 g·cm(−2)). However, when expressed in relative scales, including concordance of quartiles, reliability was poor. Agreement between QUS and DXA was quantified using data from 5042 participants. Low to modest correlations (r = 0.29 to 0.44) were obtained between multiple QUS variables and DXA BMD, with sensitivity identified as very poor (0.05 to 0.23) for osteoporosis, and poor (0.37 to 0.62) for osteopenia diagnoses. CONCLUSIONS: The findings of this large comparative analysis identify that whilst calcaneal QUS has the potential to produce reliable absolute BMD measurements and demonstrate modest associations with DXA BMD measures, use of that information to make relative statements about participants in the context of the larger population or to appropriately diagnose osteopenia or osteoporosis may be severely limited.