Test-Retest Reliability of 3D Ultrasound Measurements of the Thoracic Spine

• Published in: PM & R 2012-05, Vol.4 (5), p.335-341
• Main Authors: Fölsch, Christian, MD, Schlögel, Stefanie, Lakemeier, Stefan, MD, Wolf, Udo, MSc, Timmesfeld, Nina, Skwara, Adrian, MD
• Format: Article
• Language: English
• Subjects: Adult; Female; Follow-Up Studies; Humans; Imaging, Three-Dimensional - methods; Kyphosis - diagnostic imaging; Male; Physical Medicine and Rehabilitation; Reference Values; Reproducibility of Results; Thoracic Vertebrae - diagnostic imaging; Ultrasonography
• ISSN: 1934-1482; 1934-1563
• DOI: 10.1016/j.pmrj.2012.01.009

Objective To explore the reliability of the Zebris CMS 20 ultrasound analysis system with pointer application for measuring end-range flexion, end-range extension, and neutral kyphosis angle of the thoracic spine. Setting The study was performed within the School of Physiotherapy in cooperation with the Orthopedic Department at a University Hospital. Participants The thoracic spines of 28 healthy subjects were measured. Methods Measurements for neutral kyphosis angle, end-range flexion, and end-range extension were taken once at each time point. The bone landmarks were palpated by one examiner and marked with a pointer containing 2 transmitters using a frequency of 40 kHz. A third transmitter was fixed to the pelvis, and 3 microphones were used as receiver. The real angle was calculated by the software. Bland-Altman plots with 95% limits of agreement, intraclass correlations (ICC), standard deviations of mean measurements, and standard error of measurements were used for statistical analyses. The test-retest reliability in this study was measured within a 24-hour interval. Main Outcome Measurements Statistical parameters were used to judge reliability. Results The mean kyphosis angle was 44.8° with a standard deviation of 17.3° at the first measurement and a mean of 45.8° with a standard deviation of 16.2° the following day. The ICC was high at 0.95 for the neutral kyphosis angle, and the Bland-Altman 95% limits of agreement were within clinical acceptable margins. The ICC was 0.71 for end-range flexion and 0.34 for end-range extension, whereas the Bland-Altman 95% limits of agreement were wider than with the static measurement of kyphosis. Compared with static measurements, the analysis of motion with 3-dimensional ultrasound showed an increased standard deviation for test-retest measurements. Conclusions The test-retest reliability of ultrasound measuring of the neutral kyphosis angle of the thoracic spine was demonstrated within 24 hours. Bland-Altman 95% limits of agreement and the standard deviation of differences did not appear to be clinically acceptable for measuring flexion and extension.