Resistance training and muscle aging: what the trials show

Strength goes up reliably at any age; mass and function move less. Three large syntheses, and where each one stops.

goodWell supported, with the usual caveats about effect size.Revised 28 Sept 2026

Muscle aging has a large literature and a small number of questions that matter. Do older adults still gain strength from lifting? Does that strength turn into better daily function? And what training details actually change the outcome? Three syntheses answer those in order, and each one leaves a gap worth knowing about.

The definition underneath

The revised European consensus (EWGSOP2, 2019) made low muscle strength the primary criterion for sarcopenia, ahead of low muscle mass. Strength is what predicts falls, disability and mortality best, and it is also the thing most responsive to training. That ordering shapes everything below: the trials measure strength first, mass second, function third, and the results follow exactly that ranking.

Strength: the settled part

Peterson and colleagues pooled 47 studies with 1,079 participants aged 50 and over. Every strength outcome improved. Absolute gains ranged from 9.8 to 31.6 kg across the four standard tests, and the percentage changes were 29, 24, 33 and 25 percent for the leg press, chest press, knee extension and lat pull respectively. The predictor that mattered was intensity: harder training produced larger gains, including in the oldest participants.

The Cochrane review by Liu and Latham, at 121 trials and 6,700 participants, found the same thing on a larger scale: a large positive effect on muscle strength across 73 trials and 3,059 participants (standardised mean difference 0.84). In most included trials people trained two to three times a week at high intensity. Strength responding to loading in old age is about as settled as exercise science gets.

Function: real, but smaller

This is where the numbers shrink, and honestly reporting the shrinkage is the point. In the same Cochrane review, the improvement in overall physical ability across 33 trials and 2,172 participants was small (SMD 0.14). Gait speed rose by 0.08 metres per second across 24 trials — a genuine but modest change. The bright spot was chair-rise performance, with a moderate-to-large effect across 11 trials and 384 participants. People with osteoarthritis also reported less pain (SMD -0.30 across 6 trials).

So the honest summary is: training makes older adults stronger by a lot and more capable by a little, with standing up from a chair as the clearest functional win. Anyone selling the idea that strength gains transfer one to one into daily life is not reading the same review.

Mass: the smallest response

Borde and colleagues restricted themselves to randomised trials in adults with a mean age of 65 and over — 25 studies — and separated strength from muscle morphology. Strength improved substantially (mean SMD 1.57). Muscle size, measured as cross-sectional area, volume or thickness, barely moved (mean SMD 0.42 across nine studies). Older muscle gets much better at producing force and only slightly bigger. Neural adaptation does most of the early work, and the hypertrophy literature in the very old stays thin.

Training details: what the meta-regression favours

The same review asked which training variables predict larger strength gains, and the answers are concrete enough to plan around. The largest effects came from training twice a week, two to three sets per exercise, seven to nine repetitions per set, at 70 to 79 percent of the one-repetition maximum, with longer programmes outperforming shorter ones. Frequency above twice a week did not add strength effects in these data. Treat that as a starting template from old, mostly healthy adults — not a prescription, and not tested in frail or recently hospitalised people.

What the data do not show

Three limits, stated plainly. First, adverse events were poorly recorded across the Cochrane trials; musculoskeletal complaints such as joint pain and soreness appeared wherever they were monitored, serious events were rare, and none were judged directly related to the programme — but “poorly recorded” means the safety picture is thinner than the efficacy picture. Second, the participants were mostly healthy community-dwelling adults, so transfer to frail, multimorbid or institutionalised populations needs caution, which the Cochrane authors say themselves. Third, almost everything here measures months, not years: that strength persists only while the loading continues is physiology, not pessimism, but no trial in these reviews demonstrates decade-long adherence or decade-long benefit.