File
Spencer L1, McKenna L1, Fary R1, Ho R1, Briffa K1
1Curtin University, School of Physiotherapy and Exercise Science, Perth, Australia
Background: Non-radiological clinical measurement of thoracic kyphosis (TK) is required to assess TK without cost and exposure to radiation. The Flexicurve ruler is a non-radiological method of TK measurement that has moderately-strong correlations with the "gold-standard" radiological Cobb angle method, but consistently underestimates the TK angle. Cobb angles can include measurement errors that may explain this poor agreement, particularly in older populations. For example, degenerative morphology of T1 and T12 can lead to overestimation of TK, as these vertebral end plates are used for the Cobb angle measurement. The radiological centroid method has been advocated, but not examined, as a better alternative for radiographic comparison with flexicurve ruler in older populations.
Purpose: To examine and compare the validity of the flexicurve method using two separate “gold standard” radiological measurements of TK in postmenopausal women.
Methods: This was a cross-sectional study of postmenopausal women. Thoracic kyphosis was measured in each participant using
1) a flexicurve ruler and
2) a lateral thoracic x-ray.
Flexicurve kyphosis angles were calculated using a mathematical formula and flexicurve dimensions. Cobb and centroid angles were measured from a single x-ray of each participant. Correlations (r) between methods were calculated and Paired-samples t-tests identified differences between the angles obtained from each method. Using a Bland and Altman analysis, each method was paired with another and assessed for agreement. Linear regression was used to assess if the differences in measurement between methods were significantly predicted by age (years), body mass index (Kg/m2) and bone mineral density (g/cm2).
Results: 117 postmenopausal women (mean age: 61.43yrs, SD 6.99yrs) participated in this study. Thoracic vertebral fractures were identified in 17 women. The flexicurve kyphosis angles were more strongly correlated with the centroid angles (r=0.61, p .005) than with Cobb angles (r=0.55, p .001). The Bland and Altman analysis showed a fixed bias in the plots between flexicurve method and both the Cobb and centroid methods. The flexicurve method recorded smaller angles than the Cobb method (MD: 18.02°, 95% CI: 16.35-19.69°) and the centroid method (18.57°, MD: 95%CI: 16.99-20.17°). In relation to the Cobb method, the differences increased proportionally and progressively as the average values increased. The differences between radiological and flexicurve methods increased with increasing age which explained up to 6-8% of this variance.
Conclusion(s): Thoracic kyphosis measured in healthy postmenopausal women using the flexicurve method shows poor agreement with radiological methods regardless of whether the Cobb or centroid angles are used for comparison. Inaccuracy of the flexicurve increased with increasing angle of kyphosis and increasing age.
Implications: The magnitude of the differences and thus the potential for errors in measuring thoracic kyphosis using the flexicurve as compared to radiological methods could be considered unacceptable across a range of clinical and research settings. Caution is suggested when interpreting thoracic kyphosis angles that are determined using the flexicurve in postmenopausal women.
Keywords: Thoracic kyphosis, Radiological methods, Flexicurve
Funding acknowledgements: Primary researcher (LS) was supported by funding from an Australian Government Research Training Program Scholarship and Curtin University Research Scholarship.
Purpose: To examine and compare the validity of the flexicurve method using two separate “gold standard” radiological measurements of TK in postmenopausal women.
Methods: This was a cross-sectional study of postmenopausal women. Thoracic kyphosis was measured in each participant using
1) a flexicurve ruler and
2) a lateral thoracic x-ray.
Flexicurve kyphosis angles were calculated using a mathematical formula and flexicurve dimensions. Cobb and centroid angles were measured from a single x-ray of each participant. Correlations (r) between methods were calculated and Paired-samples t-tests identified differences between the angles obtained from each method. Using a Bland and Altman analysis, each method was paired with another and assessed for agreement. Linear regression was used to assess if the differences in measurement between methods were significantly predicted by age (years), body mass index (Kg/m2) and bone mineral density (g/cm2).
Results: 117 postmenopausal women (mean age: 61.43yrs, SD 6.99yrs) participated in this study. Thoracic vertebral fractures were identified in 17 women. The flexicurve kyphosis angles were more strongly correlated with the centroid angles (r=0.61, p .005) than with Cobb angles (r=0.55, p .001). The Bland and Altman analysis showed a fixed bias in the plots between flexicurve method and both the Cobb and centroid methods. The flexicurve method recorded smaller angles than the Cobb method (MD: 18.02°, 95% CI: 16.35-19.69°) and the centroid method (18.57°, MD: 95%CI: 16.99-20.17°). In relation to the Cobb method, the differences increased proportionally and progressively as the average values increased. The differences between radiological and flexicurve methods increased with increasing age which explained up to 6-8% of this variance.
Conclusion(s): Thoracic kyphosis measured in healthy postmenopausal women using the flexicurve method shows poor agreement with radiological methods regardless of whether the Cobb or centroid angles are used for comparison. Inaccuracy of the flexicurve increased with increasing angle of kyphosis and increasing age.
Implications: The magnitude of the differences and thus the potential for errors in measuring thoracic kyphosis using the flexicurve as compared to radiological methods could be considered unacceptable across a range of clinical and research settings. Caution is suggested when interpreting thoracic kyphosis angles that are determined using the flexicurve in postmenopausal women.
Keywords: Thoracic kyphosis, Radiological methods, Flexicurve
Funding acknowledgements: Primary researcher (LS) was supported by funding from an Australian Government Research Training Program Scholarship and Curtin University Research Scholarship.
Topic: Musculoskeletal: spine; Outcome measurement; Musculoskeletal
Ethics approval required: Yes
Institution: Curtin University
Ethics committee: Human Research Ethics Committee
Ethics number: RDHS-267-15
All authors, affiliations and abstracts have been published as submitted.
