Memo: Metformin + Exercise - Protection Signal vs. Adaptation Deficit

Memo: Metformin + Exercise — Protection Signal vs. Adaptation Deficit

Alpha: Receipt 1 shows a pharmacologic adjunct (dapagliflozin) preserves training adaptations, while Receipt 2 reports a damage/injury-protection benefit that does not translate into improved performance, highlighting a split between protective and adaptive endpoints under the same metformin + exercise anchor.

Receipt 1: Malinin et al., 2019, J Clin Endocrinol Metab — Dapagliflozin + 12 wk endurance training in overweight/obese adults did not attenuate favorable adaptations (body mass, body composition, VO₂peak) vs. placebo. (DOI: 10.1210/jc.2018-01741)

Receipt 2: Biomedicines 2023 — 8 wk metformin + moderate exercise in healthy rats: reduced serum muscle-injury markers (ALT/AST/LDH/CK-MB) and favorable molecular readouts, but no gain in graded endurance performance. (DOI: 10.3390/biomedicines11092334)

Why surprising: Both protocols couple a diabetes drug with endurance exercise, yet Receipt 2 isolates a clean damage-protection signal that fails to convert into a performance gain — the opposite pattern from what a "protective = beneficial" reading would predict, and consistent with Receipt 1's metformin-attenuates-adaptation prior.

Caveats / falsifiers: Different species (rat vs. human), different drugs (metformin vs. SGLT2i), different durations (8 wk vs. 12 wk), and injury markers ≠ functional protection; muscle-damage marker reductions without histology or force recovery could reflect assay or dosing effects rather than true protection.

Selection basis: Pair holds the "drug + endurance exercise" anchor constant while Receipt 2 explicitly reports the protection endpoint separately from the performance endpoint, enabling a within-pair protection-vs-adaptation split.

Next test gap: Head-to-head rat study (metformin vs. dapagliflozin vs. vehicle) with matched training, measuring both histological muscle injury/regeneration and VO₂/work capacity, to test whether protection is metformin-specific or class-independent, and whether any protection endpoint actually co-varies with performance.