| dc.description.abstract | Introduction: Trochanteric soft tissue thickness (TSTT), which attenuates the force applied to the hip upon impact, has emerged as a potential factor that may improve the assessment of the risk of a hip fracture. However, a gold standard technique has not been established for the measurement of TSTT. Different measurement techniques using ultrasound (US) and dual energy x-ray absorptiometry (DXA) have been used, but the accuracy and reliability of these techniques have not been extensively assessed.
Objectives and Hypotheses: The first objective was to determine the concordance validity between US and DXA measurements of TSTT. The second objective was to determine if there were significant differences in TSTT between standing, supine, and side-lying measurements, as well as between internal hip rotation of 25°, 0° rotation, external hip rotation of 25°, and possible interactions. The third objective was to determine the intra and inter-rater reliability across DXA and US measurements made in a standing, supine, and side-lying position. The corresponding hypotheses for the first and third objectives are that the intraclass correlation will be strong (ICC > 0.8), and that the coefficient of repeatability (i.e. the 95% confidence interval of the differences) will be below a clinical threshold of 0.96 cm (CR < 0.96 cm). The corresponding hypotheses for the second objective were that there will be a significant difference between the TSTT collected across: a) postures including standing, supine, and side-lying positions; b) hip rotation angles of 25° internal, 0°, and 25° external, and c) that there would be no significant interaction between body position and hip rotation (p > 0.05).
Methods: Forty-five community-dwelling older adults (20 males, 25 females) were recruited for this study. Mean (SD) age was 70.2 (10.8) years, and BMI was 27.5 (4.3) kg/m2. TSTT were obtained using a QDR Discovery™ DXA Scanner (Hologic Inc., MD, USA) and by using a curvilinear ultrasound transducer (C60x, 2-5MHz) in combination with an M-Turbo Ultrasound Unit 1.0.6™ (Sonosite Inc., WA, USA). Ultrasound measurements were made with each participant in a standing, supine, and side-lying position. Within each position, the hip was rotated internally at 25°, at 0°, or externally at 25° by using standardized foam triangles. Repeat measurements, as well as measurements by another investigator, were taken for reliability analyses. The investigators were blinded to the TSTT value during the collection, and the protocol of landmarking the greater trochanter was repeated for every measurement. The concordance validity between US and DXA, and the different reliabilities were tested by calculating two-way random ICCs and CRs. Also, Bland-Altman plots were used to visualize the results. A two-factor repeated measures ANOVA (α=0.05) was used to determine the main effects of body position and hip rotation on TSTT and a potential interaction.
Results: The ICC (2,1) of TSTT measurements between US and DXA was 0.898, but the CR was 2.15 cm and the mean bias of the differences was 0.46 cm. There was a main effect of body position (p < 0.001) and hip rotation (p < 0.001) on TSTT. Specifically, mean (SD) standing TSTT was 4.33 (2.1) cm, the supine TSTT was 5.57 (2.8) cm, the side-lying TSTT was 3.29 (1.7) cm. Mean (SD) TSTT for the 25° internal hip rotation was 4.17 (2.4) cm, no rotation was 4.33 (2.4) cm, and the 25° external rotation was 4.69 (2.5) cm. There was also a significant ordinal interaction between body position and hip rotation (p = 0.018). The US intra-rater ICC (2,1) was 0.980, 0.972, and 0.977 for the standing, supine, and side-lying measurements, and their respective CR was 0.87 cm, 1.32 cm, and 0.69 cm. The US inter-rater ICC (2,1) was 0.970, 0.939, and 0.977 for the standing, supine, and side-lying measurements, and their respective CR was 1.17 cm, 2.08 cm, and 1.10 cm. The DXA intra-rater and inter-rater reliability analyses produced ICCs (2,1) of 0.995 and 0.995, and CRs of 0.45 cm and 0.38 cm. For all of the reliability analyses, the mean bias of the differences was under 0.2 cm.
Discussion/Conclusion: When compared to US, DXA underestimates TSTT, which goes against common belief in the literature. As main effects of body position and hip rotation were found, future researchers should strongly consider tightly controlling / standardizing these factors when measuring TSTT. From a reliability perspective, the side-lying US and DXA analyses were superior to US measurement in standing or supine postures. This thesis provides important information about TSTT characteristics across measurement modalities and body postures. Future research should be conducted to determine what approach for measuring TSTT is most effective as part of models that predict hip fracture risk. | en |
