Heidel · The Journal of sports medicine and physical fitness 2022 · systematic review and meta-analysis · n=?

Machines and free weight exercises: a systematic review and meta-analysis comparing changes in muscle size, strength, and power.

Cited 24 times in the scientific literature.

Level 1 - systematic review of randomized trials

Systematic review and meta-analysis of comparative trials

PubMed 34609100 · doi:10.23736/S0022-4707.21.12929-9 · record verified 2026-08-29

What was done

A systematic review and meta-analysis searched PubMed, Scopus, and Web of Science through September 15, 2020, to compare the effects of free-weight versus machine-based resistance training on muscle size, dynamic strength, and power. Effect sizes were calculated using a three-level random-effects meta-analytic model.

What was found

When dynamic strength was tested using a free-weight exercise, free-weight training resulted in greater strength increases than machine training (effect size [ES]: 0.655; 95% CI: 0.269 to 1.041). Conversely, when strength was tested using a machine exercise included in the training program, machine training led to greater increases (ES: -0.784; 95% CI: -1.223 to -0.344). When strength was assessed using a neutral device, gains were similar between modalities (ES: 0.128; 95% CI: -0.303 to 0.559). There were no significant differences between exercise modalities for muscle hypertrophy (ES: -0.01; 95% CI: -0.525 to 0.545) or power (ES: -0.049; 95% CI: -0.557 to 0.460).

Why it matters

This review demonstrates that muscle hypertrophy and power adaptations are modality-independent, and that strength gains follow the principle of test specificity. Practitioners can select either modality based on personal preference or specific athletic testing requirements.

Limits

The abstract does not report the total number of included studies, total participant sample size, participant demographics (such as sex or training status), training volumes, or intervention durations. Potential heterogeneity between study protocols and risk of bias across included studies were not detailed in the abstract.

Cited by