Lumbar Muscle Cross-Sectional Areas Do Not Predict Clinical Outcomes in Adults With Spinal Stenosis: A Longitudinal Study

Background Minimal longitudinal data exist regarding the role of lumbar musculature in predicting back pain and function. In cross-sectional study designs, there is often atrophy of the segmental multifidus muscle in subjects with low back pain compared with matched controls. However, the cross-sectional design of these studies prevents drawing conclusions regarding whether lumbar muscle characteristics predict or modify future back pain or function. Objective The primary objective of this study is to determine whether the cross-sectional area (CSA) of lumbar muscles predict functional status or back pain at 6- or 12-month follow-up in older adults with spinal degeneration. The secondary objective is to evaluate whether these muscle characteristics improve outcome prediction above and beyond the prognostic information conferred by demographic and psychosocial variables. Design Secondary analysis of a randomized controlled trial. Participants A total of 209 adults aged 50 years and older with clinical and radiographic spinal stenosis from the Lumbar Epidural steroid injection for Spinal Stenosis (LESS) trial. Methods Using baseline magnetic resonance images, we calculated CSAs of the lumbar multifidus, psoas, and quadratus lumborum muscles using a standardized protocol by manually tracing the borders of each of the muscles. The relationship between lumbar muscle CSAs and baseline measures was assessed with Pearson or Spearman correlation coefficients. The relationship between lumbar muscle characteristics and 6- and 12-month Roland Morris Disability Questionnaire (RDQ) and back pain Numeric Rating Scale (NRS) responses was further evaluated with multivariate linear regression. A hierarchical approach to the regression was performed: a basic model with factors of conceptual importance including age, gender, BMI, and baseline RDQ score formed the first step. The second and third steps evaluated whether psychosocial variables or muscle measures conferred additional prognostic information to the basic model. Main Outcome Measures Function as measured by the RDQ and back pain as measured by the NRS at 6- and 12-month follow-up. Results Lumbar muscle CSA was not a significant predictor of 6- or 12-month RDQ or pain score in multivariate analyses. Conclusions Cross-sectional areas of lumbar muscles do not predict function or pain at medium- and long-term follow-up in adults with lumbar spinal stenosis. Level of Evidence III