Strength and muscle growth

The most recent meta-regressions across the resistance-training literature model hypertrophy and strength against weekly set volume. Hypertrophy rises with volume along a decelerating curve: gains continue as volume increases, with each increment returning less than the last. Strength follows a far shallower relationship — most of the measured strength gain is present at the lower end of the volume range sampled, and added volume moves it comparatively little.

Confidence intervals widen sharply at the top of that range because few trials operate there, which is why the analyses describe a flattening rather than locating a ceiling. The umbrella review across existing systematic reviews of prescription variables reports precisely this pattern of agreement and disagreement: the reviews concur that volume is the variable most closely tied to muscle mass, and diverge on where the dose-response ends.

The Bayesian network meta-analysis supplies the complementary half. When whole prescriptions are ranked against one another rather than examined variable by variable, muscle size sorts on volume and strength sorts on load — the same divergence the load literature reports, reproduced under a different statistical model.

The systematic review of proximity-to-failure and hypertrophy pooled trials comparing sets taken to failure against sets stopped short of it. It reports a small advantage to training closer to failure for muscle growth, with considerable uncertainty around the estimate and an interaction with volume: the comparison behaves differently depending on whether total work was equated between the conditions.

The follow-up series of meta-regressions replaced that dichotomy with a continuous variable, using estimated repetitions in reserve. It reports hypertrophy increasing as sets are taken closer to failure across the range studied, and strength gain showing no meaningful relationship with proximity to failure at all.

That asymmetry matches the load findings. Size responds to how much of the muscle is working and for how long; strength responds to how heavy the work is. Load and proximity to failure emerge as two different routes toward the same recruitment state.

The frequency meta-analysis separates the two study designs deliberately. In volume-equated comparisons — the same weekly work distributed across more or fewer sessions — differences in hypertrophy between frequencies were not statistically significant. In non-equated comparisons, higher frequencies produced greater hypertrophy, which is what would be expected if frequency acts as a vehicle for volume rather than as an independent stimulus.

The recent meta-regressions, which modelled volume and frequency together across the literature, reach the same resolution: once weekly volume is accounted for, frequency contributes little additional explanatory power to either hypertrophy or strength.

Frequency findings reported without volume control are almost always volume findings wearing a different label.

The narrative review of hypertrophy mechanisms evaluates the candidates the field has proposed: mechanical tension, metabolic stress, and exercise-induced muscle damage. It reports mechanical tension as the mechanism carrying the strongest and most direct evidence; metabolic stress as plausible largely through its effect on motor-unit recruitment rather than as an independent signal; and muscle damage as an accompaniment to training that has not been shown to add to hypertrophy, and that may subtract from it by degrading subsequent sessions.

The systemic-hormone hypothesis is addressed and rejected. The transient post-exercise rises in testosterone and growth hormone that dominated an earlier generation of research show no consistent relationship with the hypertrophy those sessions produced.

This is why the load literature comes out where it does. If tension on active fibres is the operative signal, then heavy load and proximity to failure are two routes to a similar recruitment state — which is what the network meta-analysis of load reports when it finds hypertrophy broadly insensitive to load and strength distinctly sensitive to it.

The network meta-analysis of resistance-training volume in older adults draws on 151 randomised trials and evaluates four outcomes separately: physical function, lean body mass, lower-body hypertrophy and lower-body strength. It reports that they do not share a single dose-response. Strength and hypertrophy track volume in the direction the general literature predicts; physical function improved across the volume conditions compared, including the lowest of them; and lean body mass responded least clearly of the four.

That divergence is partly biological and partly a property of the instruments. Physical function in these trials is assessed with timed batteries — chair rises, gait speed, timed up-and-go — which are sensitive to the neural and coordinative gains that arrive early in training, and which run into ceiling effects among participants already functioning well at baseline.

The umbrella review of prescription variables covers physical function alongside mass and strength for the same reason, and reports the same untidiness: the reviews converge on the direction of effect and diverge on the dose, more so for function than for either tissue outcome.