Muscle & Physical Function: What the Primary Evidence Supports

Function beats mass. The muscle-longevity link is real but its load-bearing endpoint is strength and physical function, not muscle quantity. Grip strength inversely predicts all-cause and cardiovascular mortality, each 5 kg lower grip raises all-cause death ~16% and is a stronger predictor than systolic blood pressure [1]. In the same cohort where it was measured, muscle strength predicted mortality while muscle mass did not [2], and appendicular lean mass index (ALMI) was not independently associated with mortality once grip and gait were accounted for [3]. A test as simple as a failed 10-second one-legged stance nearly doubles all-cause mortality [4]. These are observational associations partly reflecting general vitality, so the interventional case rests on the training and nutrition trials below, not the epidemiology alone.

ALMI defines sarcopenia; it does not drive the outcome. Appendicular lean mass normalized to height is the foundational quantitative definition of sarcopenia [5], and weakness-anchored cutpoints refine it [6]; current consensus uses low strength as the primary criterion with low mass as confirmation [7]. ALMI is well-validated to track muscle quantity but is weak-to-null as a standalone mortality marker once function is considered.

The aging-muscle mechanism was reversed by the primary evidence. The classical model assumed aged muscle suffers insufficient mTORC1 signaling; the primary data show the opposite. In aged muscle, mTORC1 is chronically hyperactivated, driving GDF15/STAT3 signaling, oxidative damage, and progressive fiber degeneration, demonstrated in TSC1-knockout and naturally-aged mice with the same pattern in human aged-muscle biopsies, and rescued by rapamycin [8]. Sustained mTORC1 also blocks autophagy via Ulk1 inhibition, causing late-onset myopathy, again rapamycin-reversible [9]. This is mouse-and-tissue mechanism; no human RCT of rapamycin on muscle outcomes exists.

Resistance training is the single most powerful lever, and it is now independently well-anchored. A meta-analysis of 49 studies (n=1,328, adults ≥50) found progressive resistance training adds a pooled +1.1 kg lean body mass versus non-exercising controls, with a confirmed dose-response [10]. The Cochrane review of 121 RCTs (~6,700 participants, aged 60, 89+) found a large effect on strength (SMD 0.84), improved gait speed (+0.08 m/s) and chair-rise ability, with serious adverse events rare even in those with comorbidities [11]. Strength training is an independent mortality lever: muscle-strengthening activity lowers all-cause mortality, CVD, cancer, and diabetes ~10-17%, maximal at roughly 30, 60 min/week [12]. High-load resistance training is safe in older women with low bone density and raises lumbar BMD ~4% [13], and balance-and-functional exercise cuts the rate of falls ~24% with high certainty [14].

Protein converges on ~1.6 g/kg for muscle outcomes. Benefit for muscle plateaus near 1.6 g/kg/day in the canonical resistance-training meta-analysis [15]. In older adults, total daily quantity above the RDA (≥1.5 vs 0.8 g/kg) improves whole-body net protein balance primarily by raising synthesis [16]. The per-meal anabolic response is larger than old "ceilings" implied: 100 g of post-exercise protein produces a greater and more prolonged (>12 h) anabolic response than 25 g, with no demonstrated upper limit in magnitude or duration [17]. Pre-sleep casein (40 g) raises overnight muscle protein synthesis [18], and chronic pre-sleep protein increases muscle mass and strength over a 12-week training block [18]. A 12-month RCT found a higher-protein diet (1.6 vs 0.8 g/kg) preserved more lean mass and drove ~22% more fat loss during weight loss [19]. (The specific daily-target recommendations above 1.6 g/kg, Attia 2.0, Lyon 2.2, are expert opinion; see Contested.)

Creatine adds a small, real increment on top of training. Creatine plus resistance training adds ~1.4 kg lean tissue mass and strength over training alone in older adults [20], and direct muscle biopsy shows genuine fiber cross-sectional-area hypertrophy, not cell hydration [21]. The acute scale/DEXA bump, however, is largely water: a washout RCT found no lean-body-mass advantage after correcting for water during 12 weeks of training [22]. Creatine is safe to high doses, and the creatinine rise it causes is an artifact, not kidney injury [23].

Omega-3 is an anabolic sensitizer, not a muscle builder. In older adults, ~3.36 g/day EPA+DHA over six months increased thigh muscle volume (+3.6%), grip (+2.3 kg) and 1RM (+4.0%) [24], a result independently replicated with 4 g/day krill oil improving knee-extensor torque, grip, and muscle thickness in adults over 65 [25]. Pre-loaded fish oil (5 g/day) cut muscle-volume loss during two weeks of immobilization by ~43% [26], relevant to post-surgical and injury recovery, though this trial was n=20 in young women.

Bone and fracture close the physical-function loop. Lower BMD predicts fracture (~1.5× per SD) [27], hip fracture carries large sustained excess mortality [28], and antiresorptive drugs reduce fractures (denosumab: vertebral −68%, hip −40% [29]; alendronate: vertebral −47-55%, hip −51% [30]). Exercise raises BMD but its fracture-endpoint benefit remains unproven (underpowered null) [31], a clean surrogate-vs-outcome caution.

Where the Evidence Is Contested

The first unresolved question is whether protein intake should be optimized for muscle or for longevity. The muscle camp argues for eating roughly 1.6 g/kg to build muscle, pointing to the link between strength and survival (Morton [15]; Leong [1]). The longevity and caloric-restriction camp counters that high animal protein in midlife raises IGF-1. Levine ties it to roughly a 75% increase in all-cause mortality and about a fourfold increase in cancer [32]. The disagreement persists for two reasons. That association is confounded and reverses after age 65. And the IGF-1 relationship follows a U-curve, with the best range around 120-160 ng/mL corresponding to roughly 50-80 g/day of protein [33].

Even within the muscle camp, the daily target is contested. Phillips places it at 1.6 g/kg, citing a primary plateau in the evidence [34]/[15]. Attia argues for 2.0 g/kg, treating it as an asymmetric floor. Lyon pushes to 2.2 g/kg from a muscle-maximalist position, though her recommendation carries conflict-of-interest flags. To be clear, all targets above 1.6 g/kg rest on expert opinion rather than settled data.

A related debate surrounds GLP-1 drugs and muscle loss. This one is largely a framing contest. On one side are the cardiovascular and outcome benefits alongside unprecedented weight loss (SELECT, STEP). On the other is the finding that roughly 40% of the weight lost is lean mass [35]. Three facts keep it a framing question rather than a settled harm. The lean-mass split is comparable to what is seen with any diet. It can be mitigated by resistance training and protein. And the risk is concentrated in older people and non-exercisers.

Exercise raises its own contested questions. Take dose. One view holds that the mortality benefit is causal with no apparent ceiling; the Howden RCT supports this, and Arem found no excess mortality even at ten times the guideline amount. The other view points to a small atrial fibrillation and coronary artery calcium signal that appears only at lifelong-extreme levels of endurance exercise. On plaque specifically, the athlete-oriented view frames calcified plaque as benign, while Master@Heart reports more non-calcified plaque [36]. This question reaches the muscle scope by way of the exercise pages.

One point here is no longer in dispute. On creatine, a washout randomized controlled trial found no lean body mass advantage and showed that the acute increase is water (Desai [22]). Other work shows real muscle fiber hypertrophy and lean gains at the meta-analysis level (Volek [21]; Chilibeck [20]). Read these as resolved: both are true.

References

1. Leong et al. 2015, Lancet (PURE) — grip strength & mortality · PubMed 25982160
2. Newman 2006 Health ABC, J Gerontol A — strength (not mass) predicts mortality · PubMed 16456196
3. Sim 2019, JAMDA — ALMI not associated with mortality once function considered · PubMed 30527277
4. Araujo 2022, BJSM — failing a 10-s one-legged stance → all-cause mortality HR 1.84 · PubMed 35728834
5. Baumgartner 1998, Am J Epidemiol — original ALMI sarcopenia cutpoint · PubMed 9554417
6. Cawthon 2014 FNIH, J Gerontol A — weakness-anchored ALM cutpoints · PubMed 24737559
7. Cruz-Jentoft 2019 EWGSOP2, Age Ageing — strength-first sarcopenia consensus · PubMed 31081853
8. Tang H, Inoki K, Brooks SV et al. 2019 (Aging Cell) — mTORC1 chronically hyperactive in aged mouse + human muscle; TSC1-KO model: oxidative damage,… · PubMed 30924297
9. Castets P, Rüegg MA et al. 2013 (Cell Metab) — TSC1-deficient mice: sustained mTORC1 → Ulk1 inhibition → autophagy block (overrides FoxO3) →… · PubMed 23602450
10. Peterson MD, Rhea MR, Sen A, Gordon PM. 2011 (Med Sci Sports Exerc) — Meta-analysis of 49 studies (n=1,328; adults ≥50); pooled LBM gain from RT =… · PubMed 20543750
11. Liu CJ, Latham NK. 2009 (Cochrane Database Syst Rev) — Progressive resistance training in older adults; 121 RCTs, ~6,700 participants; SMD 0.84 for… · PubMed 19588334
12. Momma et al. 2022, BJSM — muscle-strengthening & mortality · PubMed 35228201
13. Watson et al. 2018 J Bone Miner Res — LIFTMOR trial: HiRIT in postmenopausal women with osteopenia/osteoporosis; lumbar spine BMD ↑ ~4% · PubMed 28975661
14. Sherrington 2019, Cochrane — balance/functional exercise cuts falls ~24% (RaR 0.76, high certainty) · PubMed 30703272
15. Morton et al. 2018, BJSM — protein + RT meta (1.6 g/kg) · PubMed 28698222
16. Kim IY, Schutzler SE, Schrader A et al. 2015 (Am J Physiol Endocrinol Metab) — Total daily protein quantity matters for net protein balance in… · PubMed 25352437
17. Trommelen et al. 2023 (Cell Reports Medicine) — Van Loon lab; 100g vs 25g protein post-exercise; 100g → greater + more prolonged (>12h) anabolic… · PubMed 38118410
18. Res et al. 2012 (Med Sci Sports Exerc) + Snijders et al. 2015 (J Nutr) — Van Loon lab; 40g casein pre-sleep → overnight MPS; chronic → increased… · PubMed 22330017
19. Layman et al. 2009 (J Nutr) — 12-month, 130 overweight adults; HP (1.6 g/kg) vs LP (0.8 g/kg); 22% more fat mass loss; Lyon NOT an author (trained… · PubMed 19158228
20. Chilibeck et al. 2017 — creatine+RT lean mass meta (older) · PubMed 29138605
21. Volek et al. 1999, Med Sci Sports Exerc — creatine + RT, muscle-fibre hypertrophy (biopsy) · PubMed 10449017
22. Desai et al. 2025, Nutrients 17(6):1081 — creatine washout LBM RCT · PubMed 40292479
23. Kreider et al. 2017 — ISSN creatine position stand · PubMed 28615996
24. Smith GI, Julliand S, Reeds DN, Sinacore DR, Klein S, Mittendorfer B (AJCN 2015; n=44; ages 60–85; Washington U) — ~3.36g/day EPA+DHA × 6 months;… · PubMed 25994567
25. Alkhedhairi SAA, Gray SR (corresponding; Glasgow; Clin. Nutr. 2022; n=102 adults >65; NCT04048096) — krill oil 4g/day × 6 months; significantly ↑… · PubMed 35504165
26. McGlory C, Gorissen SHM et al. (FASEB Journal 2019; n=20 young women; McMaster/Phillips lab) — 5g/day fish oil × 4-wk pre-load; 2-wk… · PubMed 30629458
27. Marshall 1996, BMJ — BMD predicts fracture (RR ~1.5/SD; ~2.3–2.6 site-matched) · PubMed 8634613
28. Haentjens 2010, Ann Intern Med — hip fracture → large excess mortality · PubMed 20231569
29. Cummings 2009 FREEDOM, NEJM — denosumab: vertebral −68%, hip −40% · PubMed 19671655
30. Black 1996 FIT, Lancet — alendronate: vertebral −47–55%, hip −51% · PubMed 8950879
31. Howe 2011, Cochrane — exercise raises BMD but fracture benefit not significant · PubMed 21735380
32. Levine ME, Suarez JA, Brandhorst S et al. 2014 (Cell Metab) — NHANES I; n=6,381, 18-yr follow-up; moderate/high protein (≥20% kcal) in ages 50–65:… · PubMed 24606898
33. Rahmani et al. 2022 (Aging Cell) — Longo meta-analysis; 19 studies, 30,876 subjects; U-shaped IGF-1/mortality; optimal ~120–160 ng/mL ≈ ~50–80… · PubMed 35048526
34. FMF — Rhonda Patrick + Stuart Phillips: protein + muscle
35. GLP-1 body composition — ~40% lean-mass loss (STEP/SURMOUNT DXA) · PubMed 39996356
36. 36881712 (Master@Heart — contradicts the claim) · PubMed 36881712