Fluoroscopic video to identify aberrant lumbar motion

• Published in: Spine (Philadelphia, Pa. 1976) Pa. 1976), 2007-04, Vol.32 (7), p.E220-E229
• Main Authors: Teyhen, Deydre S, Flynn, Timothy W, Childs, John D, Kuklo, Timothy R, Rosner, Michael K, Polly, David W, Abraham, Lawrence D
• Format: Article
• Language: English
• Subjects: Adult; Biomechanical Phenomena; Female; Fluoroscopy - methods; Humans; Low Back Pain - diagnostic imaging; Low Back Pain - physiopathology; Lumbar Vertebrae - diagnostic imaging; Lumbar Vertebrae - physiopathology; Male; Middle Aged; Models, Biological; Movement - physiology; Posture - physiology; Prone Position - physiology; Prospective Studies; Range of Motion, Articular - physiology; Reproducibility of Results; Video Recording - methods
• ISSN: 0362-2436; 1528-1159
• DOI: 10.1097/01.brs.0000259206.38946.cb

A prospective, case-control design. To develop a kinematic model that characterizes frequently observed movement patterns in patients with low back pain (LBP). Understanding arthrokinematics of lumbar motion in those with LBP may provide further understanding of this condition. Digital fluoroscopic video (DFV) was used to quantify the magnitude and rate of attainment of sagittal plane intersegmental angular and linear displacement from 20 individuals with LBP and 20 healthy control subjects during lumbar flexion and extension. Three fellowship-trained spine surgeons subsequently qualitatively analyzed the DFVs to determine normality of movement. Final classification was based on agreement between their symptom and motion status (11 with LBP and aberrant motion and 14 healthy controls without aberrant motion). Independent t tests, receiver operator characteristic curves, and accuracy statistics were calculated to determine the most parsimonious set of kinematic variables able to distinguish patients with LBP. Eight kinematic variables had a positive likelihood ratio > or = 2.5 and entered the model. Six of the variables described a disruption in the rate of attainment of angular or linear displacement during midrange postures. When 4 or more of these variables were present, the positive likelihood ratio was 14.0 (confidence interval 3.2-78.5), resulting in accurately identifying 96% of participants. DFV was useful for discriminating between individuals with and without LBP based on kinematic parameters. Disruptions in how the motion occurred during midrange motions were more diagnostic for LBP than range of motion variables. Cross validation of the model is required.