Intertester reliability and validity of motion assessments during lumbar spine accessory motion testing.

BACKGROUND AND PURPOSE Posterior-anterior (PA) assessment of the lumbar spine correlates with radiographic signs of instability and can guide treatment choices, yet studies of the validity of lumbar PA assessments have not been conducted in vivo. The purposes of this study were to determine the intertester reliability of the PA examination in assessing intersegmental lumbar spine motion and to evaluate the validity of this procedure in vivo with dynamic magnetic resonance imaging (MRI). SUBJECTS Twenty-nine subjects with central lumbar pain participated in this study. METHODS Two physical therapists independently identified each subject's most and least mobile lumbar segments using the PA procedure. Midsagittal lumbar images were obtained simultaneously during one examiner's assessment. Lumbar segmental mobility was quantified from magnetic resonance images as the change in the intervertebral angle between the resting position and the end range of the PA force application. For each vertebral level tested, maximal sagittal-plane segmental motion was determined. RESULTS The intertester reliability for identifying the least mobile segment was good (agreement=82.8%, kappa=.71, 95% confidence interval [CI]=.48 to .94), but it was poor for identifying the most mobile segment (kappa=.29, 95% CI=-.13 to .71), despite good agreement (79.3%). The level of agreement between the PA assessments and intervertebral motion measured by MRI was poor (kappa=.04, 95% CI=-.16 to .24, and kappa=.00, 95% CI=-.09 to .08, for the least and most mobile segments, respectively). DISCUSSION AND CONCLUSION Despite good intertester reliability for identifying the least mobile segment, PA assessments of lumbar segmental mobility did not agree with sagittal-plane motion measured by dynamic MRI. This finding calls into question the validity of the PA procedure for assessing intervertebral lumbar spine motion.