Abstract
Debate about weight loss on GLP-1-based medicines has centered on how much lean mass is lost. Lean mass alone is an insufficient gauge. In older adults strength declines about three times faster than lean mass, power declines earlier still, and strength rather than mass predicts mortality; in pooled cohorts DXA lean mass does not predict falls, mobility limitation, hip fracture or death. On these medicines functional results diverge: walking distance improves in several trials, and grip strength rose in one cohort and fell in another. Assessing strength before treatment has already been recommended. This Perspective, opened by a clinical case, adds a construct and a rule. Functional discordance is a within-person decline in strength or physical performance beyond measurement error in the absence of a corresponding adverse lean-mass signal. I propose a pragmatic schedule: grip strength, five-times chair-stand time and usual gait speed, with a timed stair climb where feasible, before treatment, at 3, 6 and 12 months, and around dose reduction or discontinuation; change called real only beyond the measurement error of the test; results judged against the patient's baseline and a fixed threshold; assistance recorded. A confirmed functional decline beyond error should prompt clinical review, whatever lean mass shows. The rationale is strongest in older and vulnerable patients. The proposal is falsifiable: if lean-mass change reliably identifies the patients whose function declines, surveillance adds little; if discordance is common, lean mass is an inadequate surrogate for preserved capacity.