Are paraspinal muscle morphology and composition associated with lumbar spinal stenosis? A systematic review.

Background: There has been increasing interest in associations between paraspinal muscle phenotypes and common spinal disorders, including lumbar spinal stenosis (LSS). However, the relation of paraspinal muscle morphology and composition with LSS is unclear.
Objectives: To provide a systematic overview and synthesis of current evidence on the association of paraspinal muscle morphology and composition with LSS presence, pain and disability or function.
Design: Systematic review METHODS: Relevant studies were identified from a search of EMBASE, PubMed, SPORTDiscuss, Cinahl, Web of Science and PEDro. Relevant studies were selected, data were extracted, and risk of bias was assessed by two independent reviewers prior to conducting a narrative synthesis to summarize evidence using GRADE considerations. The study protocol was prospectively registered (Prospero: CRD42021246492).
Results: Of 33 studies included in the review, multifidus (n=22) was most frequently studied, followed by psoas (n=10). Fatty infiltration (20 studies), CSA (n=11) and relative CSA (n=10) were the most studied paraspinal phenotypes. Most studies (n=31) had low risk of bias, with selection and confounding bias (n=13) most common. High variability in approaches used to measure muscle morphology and composition and LSS disability and pain, and variable control of confounding factors, created challenges in synthesizing findings. The only association identified for which there was moderate certainty of evidence was between increased multifidus fatty infiltration and the presence of LSS (4 of 6 studies). All other associations studied with high or moderate strength of evidence demonstrated no association, including no association between multifidus fatty infiltration and pain severity (8 studies), multifidus CSA and LSS disability (7 studies), psoas morphology and LSS anatomical severity (4 studies), or psoas composition and LSS pain severity (4 studies) for which there was high certainty. There was moderate certainty of no association between multifidus morphology and LSS pain severity (5 of 7 studies), and erector spinae morphology and pain severity (3 studies), and erector spinae composition and LSS pain severity (3 studies) and function (3 studies). Other associations remain more unclear.
Conclusion: Largely consistent results suggest multifidus fatty infiltration is associated with the presence of LSS. However, other specific associations examined with high or moderate certainty of evidence demonstrated no association, and others remain largely undetermined. Although interest has grown in paraspinal muscle morphology and composition for improved phenotyping, prognosis and treatment of common spinal disorders, the clinical utility of paraspinal muscle imaging in LSS remains unclear.
Competing Interests: Declaration of competing interest The authors declare that they have no competing interests.
(Copyright © 2025. Published by Elsevier Inc.)