The validity and reliability of noninvasive methods for measuring kyphosis.

STUDY

DESIGN

This is a cross-sectional study.

OBJECTIVE

To examine the interrater and intrarater reliability and validity of tools for measuring thoracic kyphosis, namely the digital inclinometer and the flexicurve.

SUMMARY OF BACKGROUND DATA: Various methods are used to measure kyphosis, and each has its own advantages and disadvantages. Radiography is common, highly reliable, and valid but unsuitable for regular use because of radiation exposure and cost. Other clinical methods allow safe and rapid assessment of spinal curvature. The validity and reliability of some nclinometers have been confirmed; however, there are no data that compare inclinometers and flexicurves using radiography for measuring thoracic kyphosis.

METHODS

We enrolled 105 patients with hyperkyphosis, aged between 10 and 80 years. The Cobb angle was measured radiographically by a spine specialist. Two other examiners, blinded to the Cobb angles, measured thoracic kyphosis using the flexicurve and digital inclinometer.

RESULTS

Comparing the kyphosis angle with the radiographic Cobb angle, as a gold standard, revealed that the digital inclinometer was reasonably valid for patients aged below 30 and above 50 years, whereas the validity of the flexicurve in both age ranges was poor. On the intrarater evaluation, the digital inclinometer provided a high reliability in patients aged below 30 and above 50 years; however, the flexicurve was also acceptable in this regard. In contrast, on the interrater evaluation, the digital inclinometer provided a high reliability in subjects aged below 30 and above 50 years. The flexicurve provided a poor interrater reliability in subjects aged below 30 years and an acceptable level of reliability in those aged above 50 years.

CONCLUSIONS

The digital inclinometer is a valid and reliable instrument for measuring thoracic kyphosis and can be used for regular screening. In contrast, the flexicurve has poor interrater and intrarater reliability.