Centralization as a predictor of provocation discography results in chronic low back pain, and the influence of disability and distress on diagnostic power

Abstract

Background context: The "centralization phenomenon" (CP) is the progressive retreat of referred pain towards the spinal midline in response to repeated movement testing (a McKenzie evaluation). A previous study suggested that it may have utility in the clinical diagnosis of discogenic pain and may assist patient selection for discography and specific treatments for disc pain. Purpose: Estimation of the diagnostic predictive power of centralization and the influence of disability and patient distress on diagnostic performance, using provocation discography as a criterion standard for diagnosis, in chronic low back pain patients. Study design/setting: This study was a prospective, blinded, concurrent, reference standard-related validity design carried out in a private radiology clinic specializing in diagnosis of chronic spinal pain. Patient sample: Consecutive patients with persistent low back pain were referred to the study clinic by orthopedists and other medical specialists for interventional radiological diagnostic procedures. Patients were typically disabled and displayed high levels of psychosocial distress. The sample included patients with previous lumbar surgery, and most had unsuccessful conservative therapies previously. Outcome measures: Diagnosis: results of provocation discography. Index test: The CP. Psychometric evaluation: Roland-Morris, Zung, Modified Somatic Perception questionnaires, Distress Risk Assessment Method, and 100-mm visual analog scales for pain intensity. Methods: Patients received a single physical therapy examination, followed by lumbar provocation discography. Sensitivity, specificity, and likelihood ratios of the CP were estimated in the group as a whole and in subgroups defined by psychometric measures. Results: A total of 107 patients received the clinical examination and discography at two or more levels and post-discography computed tomography. Thirty-eight could not tolerate a full physical examination and were excluded from the main analysis. Disability and pain intensity ratings were high, and distress was common. Sensitivity, specificity, and positive likelihood ratios for centralization observed during repeated movement testing for pain distribution and intensity changes were 40%, 94%, and 6.9 respectively. In the presence of severe disability, sensitivity, specificity, and positive likelihood ratios were 46%, 80%, 3.2 and for distress, 45%, 89%, 4.1. In the subgroups with moderate, minimal, or no disability, sensitivity and specificity were 37%, 100% and for no or minimal distress 35%, 100%. Conclusions: Centralization is highly specific to positive discography but specificity is reduced in the presence of severe disability or psychosocial distress. © 2005 Elsevier Inc. All rights reserved.