What is actually being claimed
Two claims again, and they are not equally supported.
The first is that resistance training improves strength, lean mass, physical function, bone mineral density and the risk of falling. This is randomised evidence in older adults and it is good evidence. It is also the claim that matters most to most people, because loss of function is what actually ends independence.
The second is that muscle-strengthening activity independently lowers the risk of dying. This comes from cohort studies, and it is the claim we grade here at B.
The distinction is worth holding because the second claim is the one used in marketing while the first is the one with better support. A reader who trains for function is on firmer ground than one who trains for a mortality curve.
The mechanism, and how well it is established
Skeletal muscle is not only a motor. It is the body's largest site of glucose disposal, a reservoir of amino acids drawn on during illness, and an endocrine tissue that releases signalling molecules during contraction. Losing it has consequences well beyond weakness.
Age-related loss of muscle mass and strength, sarcopenia, is a recognised clinical entity with defined diagnostic approaches, and strength declines faster than mass, meaning quality falls as well as quantity. The functional consequence is a slow reduction in reserve until an ordinary insult, an infection or a fall, exceeds it.
Resistance training reverses part of this. The adaptations are well characterised: neural adaptation early, then hypertrophy, with improvements in tendon and bone loading tolerance. Older adults adapt more slowly than younger adults but they do adapt, and studies in people well into later life have shown meaningful strength gains.
The route from that to survival is indirect and plausible rather than demonstrated. Better glucose disposal, preserved bone, fewer falls, greater reserve during illness and maintained independence are each associated with better outcomes. That is a chain of associations, and every link in it is a place where the causal argument could fail.
What the human evidence shows
The randomised evidence is the strong part. Structured resistance programmes in older adults improve strength, muscle mass and measures of physical function in trials, and multicomponent exercise programmes that include strength and balance work reduce falls in older people living in the community. Falls are a clinical endpoint with serious consequences, and reducing them is a real result rather than a surrogate.[1]
The mortality evidence is observational. Large cohorts that asked participants about muscle-strengthening activity and then followed them for years have reported lower all-cause mortality among those who reported doing it, and the association generally persists after adjustment for aerobic activity, which is the key comparison. Pooled analyses across cohorts have reported the same direction.
Three cautions belong with that. First, exposure is almost always self-reported and captured in coarse categories, typically a count of sessions per week with no information about load, effort or progression. Self-reported activity is systematically over-reported and the error is not random with respect to health.
Second, the shape of the relationship at the upper end is not consistent between datasets. Some report a flattening, some report attenuation at high volumes. Whether that is a real biological ceiling, a measurement artefact, or a consequence of who reports very high volumes is unresolved.
Third, people who do resistance training differ from those who do not in ways that predict survival: they are more likely to be non-smokers, to have fewer chronic conditions, to be more affluent and to be in less physically damaging work. Adjustment reduces this and cannot remove it, for the same reason it cannot in the metformin literature.
Grip strength deserves a mention because it is often used as a shorthand. It is a convenient measure that predicts outcomes in cohort studies, but it is a proxy for general muscular and neurological status rather than a target in itself, and training grip alone has no claim to any of the associated outcomes.
The limitations that hold the grade at B
| Limitation | Why it matters for the grade |
|---|---|
| Self-reported exposure | Sessions per week captures neither load nor effort, and reporting error correlates with health and education. |
| No mortality trial | Randomised evidence stops at function and falls. Nobody has randomised adults to years of lifting and counted deaths. |
| Inconsistent dose-response at the top | Different cohorts disagree about what happens at higher volumes, which weakens the causal reading. |
| Healthy exerciser confounding | The behaviour clusters with other advantages that independently predict survival. |
| Reverse causation | Early illness reduces the ability to train before it is diagnosed. |
| Heterogeneous definitions | What counts as muscle strengthening varies between surveys, so pooling is imperfect. |
It is worth stating what would not weaken the case. The randomised evidence for function, strength and falls prevention stands on its own and does not depend on the mortality question at all. UK physical activity guidance recommends muscle-strengthening activity on at least two days a week for adults of all ages on that basis.[2]
What would change the grade
Grade A on the prognostic claim would require cohort evidence using verified or objectively measured resistance exposure rather than self-report, replicating the independent association with a consistent dose-response shape.
The interventional claim, that taking up resistance training lowers a given person's mortality, would move from its current position with randomised trials reporting hard clinical endpoints. Trials in populations at high risk of fracture or of functional decline are the most plausible route, since the event rate makes them feasible.
The grade would fall if better-measured exposure data made the independent association disappear once aerobic activity and general health behaviour were fully accounted for. That result would not affect the case for training to preserve function, which is a separate and better supported claim. For the complementary aerobic literature see VO2 max as a longevity marker.