Strength training earns its place in a lever list on two claims: it lowers mortality and major disease independently of aerobic activity, and it does so at a small dose. Both come from one meta-analysis (Momma 2022) of observational cohorts, graded «very low» — so this is a real but weakly-warranted lever, held one tier below the RCT-grade BP and lipid evidence.
The effects — modest, broad, and independent of aerobic exercise
Any muscle-strengthening activity (MSA) vs none, «independent of aerobic activities»:
| Outcome | RR (95% CI) | GRADE |
|---|---|---|
| All-cause mortality | «RR 0.85; 95% CI 0.79 to 0.93» (−15%) | very low |
| Cardiovascular disease | «RR 0.83; 95% CI 0.73 to 0.93» (−17%) | very low |
| Total cancer | «0.88 (0.80 to 0.97)» (−12%) | very low |
| Diabetes | «0.83 (0.77 to 0.89)» (−17%) | low |
The independence matters: these are adjusted for aerobic activity, and combined MSA + aerobic gives a much larger benefit — «a 40% lower risk of all-cause mortality (RR 0.60; 95% CI 0.54 to 0.67)» — so strength and aerobic are additive, not substitutes. This corroborates and refines the held strength-guideline finding on Physical Activity Dose and Mortality (both-guideline HR ~0.71); it is not independent backing (same observational PA-epidemiology lineage, overlapping US cohorts), so it is a refinement, not an [E-independent] convergence.
The dose — small, and a J-shape whose upper arm should not be believed yet
The dose-response is the interesting part, and it is a worked case for The U-Shaped Association Artifact:
«J-shaped associations were found between muscle-strengthening activities and all-cause mortality, CVD and total cancer, with the maximum risk reduction (approximately 10–20%) at approximately 30–60 min/week»
The pooled nadir is «The lowest RR (RR 0.83; 95% CI 0.79 to 0.86)… at 40 min/week», with benefit persisting to ~130-140 min/week and the hazard then rising above 1.0 — the J’s upper arm, suggesting more is worse past ~2 hours/week.
Two cautions on that «40 min/week», because the point is fragile and easy to over-read
[user-challenge 2026-07-29]:
- The optimum is a wide, imprecise region, not 40 minutes. Momma’s own headline is a range («approximately 30–60 min/week»), and a key constituent study puts it higher: «the lowest risk of all-cause mortality was observed at 82 min/week of strength training» (Kamada). On very-low-GRADE, sparse, self-reported data the single pooled point (40) should be read as somewhere in the ~30–80+ min/week region, not a precise target.
- This is total weekly MINUTES, not FREQUENCY — Momma «could not directly examine the frequency of muscle-strengthening activities». So the curve says nothing about 2x vs 3x/week or session spacing; 40 min/week could be two 20-minute sessions as easily as one. In particular it does not speak to the training-physiology heuristic of stimulating each muscle group every ~48 h — that is advice for muscle/strength adaptation (hypertrophy, protein synthesis), a different outcome on a different evidence base -> Protein and Resistance Training for Muscle and Strength (Morton 2018, now held: RT is the driver, protein a modest RCT-grade adjunct up to ~1.6 g/kg/day). The mortality dose-response and the strength-adaptation question are two things, and this source answers only the first.
Do not act on the upper arm. Three reasons it is likely an artifact, not a dose feature:
- The authors disown it: «the influence of a higher volume… on all-cause mortality, CVD and total cancer is unclear, considering the observed J-shaped associations»; «a higher volume may require caution» — a hedge, not a finding.
- The evidence cannot bear it: «very low» GRADE, «only a small number of studies», self-reported exposure, sparse data at high volumes, and publication bias untestable («pooled estimates… might have been overestimated»). The upper arm is exactly the unequal-precision / reverse-causation region The U-Shaped Association Artifact warns about.
- The built-in counter-instance: for diabetes the same exposure is L-shaped, not J — «the
risk of diabetes sharply decreasing until up to 60 min/week… followed by a gradual decrease», no
upturn, and with a clear muscle-glucose mechanism. The same MSA cannot protect monotonically for
diabetes and turn harmful for mortality at high volume unless the mortality upturn is a measurement or
confounding artifact.
- What the muscle-glucose mechanism IS (Challenge #24, average adult): skeletal muscle is the
principal site of insulin-stimulated glucose disposal, so building and using muscle enlarges the
body’s glucose «sink» and improves glucose tolerance. This makes resistance training a glycaemic
lever, not only a function/mortality one — decision-relevant specifically for the dysglycaemic /
metabolic-syndrome stratum (the drifting-median adult), where the held diabetes L-shape is the
outcome this mechanism produces.
[INFERRED — the accepted physiology; the diabetes L-shape outcome it explains is [EXTRACTED (Momma - Muscle-Strengthening Activities and Mortality 2022)]; a dedicated glucose-disposal/insulin-sensitivity RCT source is not held -> a named acquirable gap.]The exposure whose mechanism is clear (diabetes) shows no upturn; the outcomes whose upturn lacks a mechanism (mortality/CVD/cancer) show one — the diagnostic signature of an artifactual arm.
- What the muscle-glucose mechanism IS (Challenge #24, average adult): skeletal muscle is the
principal site of insulin-stimulated glucose disposal, so building and using muscle enlarges the
body’s glucose «sink» and improves glucose tolerance. This makes resistance training a glycaemic
lever, not only a function/mortality one — decision-relevant specifically for the dysglycaemic /
metabolic-syndrome stratum (the drifting-median adult), where the held diabetes L-shape is the
outcome this mechanism produces.
What survives: the lower arm — a real, modest benefit reached by a modest weekly volume (a wide, imprecise ~30-80+ min/week region; the standard «at least 2 days/week» sits comfortably inside it). What does not: any reason to fear or to chase high MSA volume for mortality — nor any reading of an optimal frequency, which this source cannot supply.
Decision relevance
- Add strength training — a modest weekly volume reaches most of the mortality benefit (~30-80+ min/week, imprecise; the standard «at least 2 days/week» is inside it), and it adds benefit on top of aerobic (the two combined are the best-evidenced activity state). The effective volume is small; the optimal frequency / spacing (e.g. every ~48 h) is a strength-adaptation question this mortality evidence does not address.
- But weight the warrant honestly. This is observational, self-reported, very-low-GRADE evidence — a plausible, directionally-robust association, not an RCT-grade effect. It belongs below the BP and aerobic-activity levers in certainty, not above.
- For fitness (VO2max) it is complementary, not the same lever. Strength training raises strength and lean mass; aerobic/interval work raises CRF (Measuring and Raising Cardiorespiratory Fitness). Both belong in an untrained person’s program.
The healthspan case is stronger than the mortality case — it runs through function and falls (Challenge #17)
The very-low certainty above is for one endpoint — muscle-strengthening -> all-cause mortality (Momma, self-reported, observational). It is not the verdict on strength training’s healthspan value, which runs through better-evidenced channels:
- Function and falls are held, patient-important, and stronger. Physical activity in older adults reduces falls, injurious falls, and probably fractures — outcomes people care about directly, not surrogates, and the strongest outcome class the annex carries for this stratum (Physical Activity Dose and Mortality, WHO PA annex). Avoiding a fall/fracture and keeping physical function is the healthspan payoff, and it is graded above the mortality signal. (Caveat: the held evidence is for exercise broadly; falls-prevention specifically combines resistance + balance training, which the wiki holds at the activity level, not RT-specifically -> a gap.)
- Grip strength and muscle mass are strong mortality PREDICTORS — but un-ingested, and predictors, not proof. Low grip strength and sarcopenia track mortality, falls and disability (the maintainer’s point), and the association is robust — but the wiki has not ingested it, and it is the same predictor-vs-cause gap as fitness (Cardiorespiratory Fitness and Mortality): a strong predictor does not prove the intervention (resistance training) moves the outcome. -> acquire.
So strength training’s weak card is the mortality evidence; its strong card is function/falls (held) plus the muscle-mass/strength predictor literature (unheld). For an older or sarcopenia-risk stratum the case is stronger than the very-low mortality number alone implies — the lever’s value is in keeping function and staying off the floor, not (on current evidence) a proven mortality reduction.
Limits
- One meta-analysis, observational, «very low» GRADE; mostly US cohorts; self-reported exposure; high heterogeneity (I2 up to 83%); publication bias untestable.
- The dose-response upper arm is not usable (above); the diabetes L-shape is the more trustworthy dose curve.
- Frequency vs total-time could not be separated; the target-dose is a region, not a number.