Frontiers in Physiology · Oct 30, 2026
Comparative effects of different exercise modalities and doses on low-density lipoprotein cholesterol in postmenopausal women: a systematic review and pairwise, network, and dose–response meta-analysis
By Sheng Zhang, Lin Wang, Shijie Liu et al.
Objective: To compare how different exercise modalities affect low-density lipoprotein cholesterol (LDL-C) in postmenopausal women and determine their relative effects. The analysis also examined whether LDL-C changes varied nonlinearly across exercise doses, providing evidence relevant to the design of exercise programs for this population. Methods: We searched PubMed, MEDLINE, EMBASE, Cochrane CENTRAL, and Web of Science for RCTs evaluating exercise interventions of at least 4 weeks. Treatment effects and relative rankings across exercise modalities were synthesized using Bayesian random-effects pairwise meta-analysis and NMA. Exercise exposure was expressed as MET-min/week and incorporated into a dose–response NMA. Study-level risk of bias was judged with RoB 2, while confidence in the network estimates was appraised using CINeMA. Results: Seventy RCTs involving 3,323 postmenopausal women were included. The NMA showed that mind–body exercise (MB), aquatic exercise (AQ), high-intensity interval training (HIIT), resistance training (RT), and aerobic exercise (AE) were associated with reductions in LDL-C. MB yielded the largest estimated effect and ranked highest according to SUCRA (MD=−0.56 mmol/L; 95% CrI [−0.83, −0.29]). No clear improvement was observed with multi-component exercise (MC), and differences between most exercise modalities remained uncertain. The relationship between exercise dose and LDL-C change was nonlinear. The largest model-predicted effect occurred at approximately 810 MET-min/week (MD=−0.28 mmol/L; 95% CrI [−0.38, −0.18]). The dose range associated with achieving the minimum clinically important difference was approximately 350–1,200 MET-min/week. Conclusion: Several exercise modalities may reduce LDL-C in postmenopausal women, with MB, AQ, and HIIT showing relatively larger effect estimates. Exercise dose showed a nonlinear relationship with LDL-C lowering, with the model predicting the greatest reduction at approximately 810 MET-min/week. However, evidence for several comparisons remains limited; therefore, both the relative ordering of exercise modalities and the estimated dose–response pattern warrant cautious interpretation and should not be regarded as fixed prescriptions for individual patients.