Comparative and Dose-Response Effects of Exercise Modalities on Sleep Quality in Individuals With Breast Cancer: A Bayesian Network Meta-Analysis.

Abstract

Objective: To compare the effects of exercise modalities on sleep quality in individuals with breast cancer, and to examine dose-response associations between exercise dose and sleep improvement.

Methods: PubMed, Web of Science, Embase, the Cochrane Library, and EBSCO were searched from inception to March 2026. Risk of bias was assessed using RoB 2. Bayesian multilevel network meta-analysis compared exercise modalities against active controls. Meta-regression and nonlinear dose-response analyses explored effect modifiers and dose-response patterns.

Results: Thirty-nine randomized controlled trials involving 4066 participants were included. Compared with active controls, Tai Chi/Qigong showed the largest sleep benefit (Hedges' g = -0.45, 95% CrI: -0.86 to -0.04), followed by yoga and combined exercise (COM; both g = -0.30). Aerobic exercise suggested possible benefit but with uncertainty, whereas resistance training, dance, and stretching showed no clear advantage. Dose-response analyses suggested greater benefits around potentially beneficial moderate-dose ranges, with exploratory estimated values of approximately 4.5 METs, 281.3 MET·min/session, and 910 MET·min/week. Higher baseline Pittsburgh Sleep Quality Index scores were associated with greater sleep improvement.

Conclusions: Tai Chi/Qigong, yoga, and COM may be promising exercise modalities for improving sleep quality in this population. Exercise-related sleep benefits may occur within a moderate dose window. Given the very low certainty of evidence and limited head-to-head comparisons, these findings should be interpreted cautiously and confirmed by high-quality randomized controlled trials.