The predictive value of pain drawings in lumbar spinal fusion surgery.

Abstract

Pain drawings have been used extensively in spine surgery. It has been associated with inferior outcome after disc and stenosis surgery. Results regarding the predictive value in fusion surgery have been conflicting. To evaluate the predictive value of pain drawings in relation to outcome after lumbar spinal fusion. To investigate if there are differences between spondylolisthesis patients and patients with degenerative disease as well as between patients with or without radicular pain. Prospective clinical cohort with a minimum of 1-year follow-up. One hundred thirty-five patients undergoing lumbar spinal fusion. Fifty-seven men and 78 women, mean age 44 years (range 21-59 years). Dallas Pain Questionnaire (DPQ), Low Back Pain Rating Scale (LBPRS) pain index and patient satisfaction. Minimal clinical important difference was defined for the LBPRS score. Pain drawings were classified, using the visual inspection method, as organic or nonorganic and correlated to outcomes. Multivariate adjustment for several possible confounding variables was done using logistic regression analysis. Thirty-three percent of the drawings were classified as nonorganic. Nonorganic drawings were associated with significantly higher DPQ and LBPRS scores preoperatively and at follow-up. Differences between organic and nonorganic drawings were larger in spondylolisthesis patients than in patients with degenerative disorders. Nonorganic pain drawings were associated with poorer outcome in patients with low back pain and radicular symptoms, however, not in patients without radicular symptoms. A nonorganic pain drawing predicted negative patient satisfaction with odds ratio (OR) 3.01 (95% confidence interval (CI): 1.14-8.55, p=.027) but had no significant predictive value with respect to improvement in the LBPRS pain index OR 1.92 (95% CI: 0.82-4.47, p=.132). A nonorganic pain drawing was a significant risk factor for inferior outcome after spinal fusion surgery. The predictive value did not allow for patient selection.