DISCRIMINATION OF BIOMECHANICAL BACK PAIN PATIENT SUBGROUPS FROM CONTINUOUS INTER-VERTEBRAL MOTION DATA – A PROTOCOL

Background and purpose Investigating inter-vertebral biomechanics in vivo using end-of-range imaging is difficult due to high intra subject variation, measurement errors and insufficient data. Quantitative fluoroscopy (QF) can reliably measure continuous motion but may suffer from contamination from uncontrolled loading and muscle contraction which compromises comparisons between studies and limits interpretation of results. This study presents the methods used to overcome these limitations. Methods and results Forty chronic, non-specific low back pain (CNSLPB) patients and 40 matched controls underwent QF using a passive recumbent protocol which standardised the rate and range of trunk rotation, thus reducing intra-subject variation and excluding loading and muscle contraction factors. Left, right, flexion and extension were recorded from L2-5 and vertebral motion registered using image processing algorithms, Resultant continuous inter-vertebral rotation data were normalised to produce proportional contributions of each segment throughout the trunk bend The expected continuous proportional contributions at each level and direction were determined by calculating reference intervals (mean +/− 2SD) from controls. Prevalence of patients exceeding these ranges was determined and the association with CNSLBP calculated using Chi-squared analysis. Additionally the variance of the normalised data throughout the continuous motion for each direction was determined and summed to produce an combined number. This was used to measure the difference between patients and controls and entered into ROC curve analysis to investigate discrimination between patients and controls. Conclusion A methodology for assessment of the differences between the continuous in vivo spine kinematics of CNSLBP patients and healthy controls has been developed and will be presented.