The relative effects of different exercise modes on physical and metabolic health in older adults: a network meta-analysis

Multiple exercise modalities are recommended for older adults, yet their comparative effectiveness remains uncertain. We conducted a Bayesian network meta-analysis of randomized trials to compare common exercise modes on cardiorespiratory fitness and metabolic health in adults aged ≥55 years. Trials randomized participants to high-intensity interval training (HIIT), aerobic training (AT), resistance training (RT), combined aerobic-resistance training (CART), or non-exercise control. The primary outcome was maximal/peak oxygen uptake (VO2max/VO2peak). Secondary outcomes included BMI, body fat percentage, fat-free mass, systolic/diastolic blood pressure, and blood lipids. We fitted random-effects Bayesian network meta-analysis models and summarized ranking probabilities using the surface under the cumulative ranking curve. No exercise modality showed a clear advantage for VO2max/VO2peak; credible intervals were wide for most between-modality comparisons. Versus control, RT increased fat-free mass, CART reduced body fat percentage and systolic blood pressure, and HIIT reduced BMI and triglycerides. For total cholesterol, LDL-C, HDL-C, and diastolic blood pressure, credible intervals generally included the null. Heterogeneity was moderate, and formal inconsistency assessment was limited by sparse networks. Current evidence does not identify a single "best" exercise modality for improving VO2max/VO2peak in older adults. Modality selection may be better guided by the primary goal (eg, RT for lean mass, CART for adiposity and systolic blood pressure, HIIT for BMI and triglycerides), while considering feasibility and safety. Larger, well-reported head-to-head trials are needed to strengthen comparative estimates.