Effect of sarcopenia and paraspinal muscle morphology on conservative treatment outcomes in lumbar spinal stenosis: a prospective cohort study

Abstract

Introduction The impact of sarcopenia on postoperative outcomes in lumbar spinal stenosis (LSS) is well documented, but its role in conservative treatment remains unclear. In this prospective cohort study, we examined the association between sarcopenia and outcomes after conservative treatment in elderly LSS patients and explored whether paraspinal cross-sectional area (CSA) and fat percentage mediate this relationship. Methods Elderly LSS patients undergoing conservative treatment at Sichuan Provincial Orthopedic Hospital (June 2023–May 2025) were enrolled. Sarcopenia was diagnosed according to the criteria of the Asian Working Group for Sarcopenia. The primary outcome was >50% improvement in the Japanese Orthopedic Association (JOA) score at discharge. Missing data were imputed with random forest. Candidate predictors were screened using univariate logistic regression (p < 0.1) and selected using the least absolute shrinkage and selection operator (LASSO) regression. Sensitivity analyses comprised Firth’s penalized correction for sparse data, subgroup analyses, and examination of individual sarcopenia components [appendicular skeletal muscle mass index (ASMI), grip strength, and gait speed] in relation to treatment outcome. Mediation analysis with bootstrap resampling evaluated total paraspinal CSA and mean fat percentage as mediators. Results Among 463 patients, 89 (19.2%) had sarcopenia. Sarcopenic patients had significantly smaller paraspinal CSA and greater fat infiltration than non-sarcopenic patients. After adjustment for sex, age, and albumin, sarcopenia remained independently associated with poor discharge outcome (OR = 0.21, 95% CI: 0.12–0.38, p < 0.001). Subgroup analyses showed consistent effects across age, sex, body mass index (BMI), albumin, and stenosis severity strata (all p for interaction >0.05). Among individual components, ASMI and grip strength independently predicted poor outcomes, whereas gait speed did not. Mediation analysis demonstrated that total paraspinal CSA mediated 74.64% of the effect (p = 0.012), whereas mean fat percentage showed no significant mediation (p = 0.710). Sensitivity analyses confirmed the robustness of these findings against missing data, model specification, and sparse data bias. Conclusion Sarcopenia is independently associated with worse short-term outcomes after conservative treatment in older LSS patients, with this association statistically mediated by reduced paraspinal CSA.