The effects of exercise therapy on lumbar muscle structure in low back pain: A systematic review and meta-analysis

Abstract

Background: Changes in muscle structure are observed in people with low back pain (LBP). Although exercise therapy is arguably one of the most commonly used methods to treat LBP, evidence regarding its effects on muscle structure is still lacking. Objectives: To answer the following questions: (1) What are the effects of exercise therapy on lumbar muscle structure in people with LBP and (2) which type of exercise intervention has the greatest effects? Methods: Six electronic databases were systematically searched. The RoB 2 tool and the ROBINS-I tool were used to blindly assess the Risk of bias (RoB), and the RevMan 5 tool was used for the meta-analysis. Due to heterogeneity, the various exercise interventions were classified into 4 groups. Interventions that could not be classified into 1 of these 4 groups were not included in the meta-analyses and were qualitatively analysed. Results: In total, 984 records were retrieved, of which 34 articles were included. The meta-analyses showed that when considering the overall effect size per muscle structural outcome in chronic LBP (cLBP), exercise therapy showed significant positive effects on Lumbar Multifidus (LM) muscle thickness (MT) and LM crosssectional area (CSA), but not on Lumbar Erector Spinae (LES) CSA and Quadratus Lumborum CSA. Considering the different intervention groups, significant positive effects were observed for: (1) "Motor control + stabilization" with small effect on LM MT, (2) "Stabilization" with medium effect on LM MT, and large effect on LM CSA, and (3) "Motor control + Stabilization + Lumbar strengthening" with large effect on LM MT and LES CSA. There were no significant effects observed for "Lumbar strengthening". Conclusions: Overall, exercise therapy has a positive effect on lumbar muscle structural properties in a cLBP population, with a combination of Motor control + Stabilization + Lumbar strengthening giving the best results. Database registration: This systematic review was registered on PROSPERO (CRD42021232583). (c) 2025 Elsevier Masson SAS. All rights are reserved, including those for text and data mining, AI training, and similar technologies.