Article — "Muscle Mass and Strength Outpredict BMI for Survival, With New Neural Clues to Weakness": Article — "Muscle Mass and Strength Outpredict BMI for Survival, With New Neural Clues to Weakness": Everyone quotes the same headline: muscle mass, not BMI, predicts survival. Older adults with more muscle reportedly carry a 19% lower risk of premature death, independent of fat mass, according to a promotional post by Dr. William Li. It is a satisfying line because it lets us keep the scale-and-tape-measure mindset while sounding sophisticated about it.
But read the studies that supposedly prove it, and a contradiction surfaces almost immediately.
The Health, Aging and Body Composition cohort measured both. It put muscle strength and muscle mass side by side in the same models and asked which one tracked death. Strength won cleanly. Muscle mass did not.
Quadriceps strength predicted mortality with a crude hazard ratio of 1.51 in men and 1.65 in women. Grip strength gave similar numbers. But thigh muscle area by CT and lean mass by DXA — the actual bulk — were not strongly related to death at all.
The authors' conclusion was blunt: low muscle mass did not explain the strong association of strength with mortality. Quality beat quantity.
The authors' conclusion was blunt: low muscle mass did not explain the strong association of strength with mortality. Quality beat quantity.
So we have been measuring the wrong thing twice. First we measured weight and called it health. Now we are being told to measure bulk. Both miss the signal, which is what the muscle can do, not how much of it there is.
Last week this newsletter dug into the weekly training-dose question — how much resistance work actually moves the needle. This week the question moves upstream: what should the clinic measure in the first place, and why is the number on the scale the wrong biomarker for aging?
Start with why BMI fails, because the muscle-mass camp is right about that part.
A population-based NHANES analysis of 11,687 adults, with body composition by DXA, found that at any BMI of 22 or above, people with low muscle mass had higher body fat percentage, more diabetes, and higher adjusted mortality than everyone else at the same BMI.
The mechanical reason is quietly damning. Gaining fat while losing muscle produces a 30 to 40% smaller change in BMI than the same fat gain in someone who keeps their muscle. The scale barely twitches while the body composition rots.
This is the concrete explanation for the so-called obesity paradox. Once you account for muscle mass, the BMI associated with the best survival shifts downward toward the normal range, and the risk of high BMI gets magnified. The "healthy overweight" people were partly just people who hadn't wasted away yet.
So BMI hides a dangerous phenotype: the skinny-frail person who looks fine on paper and is quietly high-risk. Muscle mass exposes them. That much holds.
But here is where the mass story runs out of road — and where power takes over.
A 2025 CLINIMEX cohort of 3,889 adults aged 46 to 75 tested this directly by separating strength from power. Power is force times velocity: not just how hard you can push, but how fast.
The gap was enormous. Comparing the lowest to the highest quartile of relative muscle power, the hazard ratio for death was 5.88 in men and 6.90 in women.
For strength over the same quartiles? 1.62 in men and 1.71 in women — and neither reached statistical significance.
Read that again. The strength contrast was a coin toss. The power contrast was a nearly sevenfold spread in mortality risk. Whatever the aging body is losing that kills people, it shows up in velocity first.
This nests neatly with Health ABC. Mass didn't predict; strength did; and now power dwarfs strength. The signal sharpens as you move from how big the muscle is, to how hard it pulls, to how fast it can pull hard.
Which points toward two dead-simple clinic tests: a handgrip dynamometer and a timed chair stand — how quickly you can rise from a seated position without using your arms, though the pooled cohort evidence below validated grip strength alongside gait speed rather than chair-stand time specifically.
A JAMA Network Open study followed more than 5,000 women aged 63 to 99 using exactly those two measures. Stronger grip and faster chair-stand performance predicted lower all-cause death over about eight years, and the association survived adjustment for age, chronic conditions, socioeconomic factors, activity, and sedentary time.
The chair stand may, per the article author's own interpretation, capture velocity and not just strength, potentially measuring power on the cheap — though no cited study establishes this.
Now the honesty check, because screening enthusiasm needs a cold shower.
A pooled analysis of three cohorts — 8,362 people, more than 5,000 deaths over a median 13.3 years — asked how much these tests actually add. Grip strength and gait speed each predicted mortality in fully adjusted models. Appendicular lean mass, again, did not.
But a model with only age and sex already produced a C-index of 0.65. Adding grip strength and gait speed raised it to just 0.67.
That is a tiny lift. As pure prognostic instruments, layered on top of what a clinician already knows from a patient's age and sex, these tests barely move the discrimination needle.
So why bother measuring at all? Because a C-index is the wrong scorecard for what these tests are actually for.
Age and sex predict death better and cannot be changed. Grip and chair-stand time predict slightly worse — but they are trainable. Their value is not as a crystal ball. It is as an actionable, movable target: a number you can hand a patient and then improve.
And whether it can be improved is exactly what the newest mechanism work answers.
A GeroScience study by Connelly and colleagues used high-density electromyography — surface sensors that resolve the firing of individual motor units — to ask why strength declines with age. The old assumption was structural: fibers shrink, muscle wastes, weakness follows.
That is only half the story. Older adults showed reduced motor unit discharge rates and altered neural drive. The nervous system was recruiting the muscle less aggressively — firing fewer signals, less often.
The effect differed by muscle group and by contraction intensity, which tells you the weakness is not one uniform process. It is a patchwork of structural loss and a neural command problem.
That neural piece is the reason to care about power specifically. Velocity may depend on high-frequency motor-unit firing, arguably among the things that degrade with age, though the cited evidence describes reduced motor unit discharge rates without explicitly making this link. Power may collapse before mass does, if the wiring degrades before the tissue does, though no cited source establishes this temporal sequence.
And a wiring problem is a retraining problem. Neural drive may respond to training faster than muscle cross-section does, and a large part of early strength gain in any program may reflect the nervous system relearning to fire rather than the fiber growing — though this is not established by the cited evidence.
So the clinical argument lands here. Screen for function, not weight. A grip dynamometer and a timed chair stand cost almost nothing and turn an abstract risk into a coachable goal — the exact kind of trigger that would justify bundling resistance coaching into, say, a GLP-1 prescription, the open thread from last week's issue.
The obstacle is not physiological. It is that older adults still fear the weight room, still treat strength training as something for the young. The GeroScience data reframes that fear as a category error.
The thing quietly raising their mortality is not just wasted tissue they can't get back. It is, in part, a nervous system that has stopped firing hard and fast — and that, unlike the number on the scale, is something they can train back.