Mechanism of action is the biochemical or physiological pathway through which a supplement or drug is proposed to produce its effect. It is the "how, at the cellular level" part of any health claim, usually described as an enzyme it inhibits, a receptor it binds, a transporter it blocks, or a precursor pool it replenishes.
Why it matters
A plausible mechanism is a reason to test a compound. It is not a reason to conclude the compound works. The supplement industry routinely markets molecules on the strength of a mechanism — antioxidant, anti-inflammatory, NAD-precursor, mitochondrial-support — that has never produced a reliable downstream outcome in humans at realistic doses. Understanding where mechanism sits on the evidence ladder keeps users from confusing a reasonable hypothesis with a proven intervention.
Target types
Most mechanisms fall into a small set of molecular-target categories. These are the same categories Unfair stores as part of each compound's ingredient metadata.
| Target type | What the compound does | Example compound | Example target |
|---|---|---|---|
| Enzyme inhibition | Blocks an enzyme's active site to lower product formation | Berberine | Activates AMPK; inhibits PTP1B |
| Receptor antagonism | Binds a receptor and blocks its native ligand | Caffeine | Adenosine A1/A2A receptor |
| Receptor agonism | Binds and activates a receptor | Melatonin | MT1/MT2 receptor |
| Transporter modulation | Alters uptake, efflux, or breakdown | EGCG (green tea) | Inhibits COMT, shifts catecholamine pool |
| Precursor loading | Supplies a substrate the body converts to the active molecule | Creatine monohydrate | Phosphocreatine resynthesis pool |
| Cofactor supply | Provides a mineral or vitamin a required enzyme needs to run | Magnesium | Cofactor for >300 enzymatic reactions |
| Structural supply | Supplies a building block for tissues | Collagen peptides | Glycine and proline for connective tissue |
A concrete enzyme-and-receptor example. Caffeine's primary mechanism is adenosine receptor antagonism. Adenosine accumulates through the day and binds A1 and A2A receptors, producing sleepiness and reduced arousal. Caffeine, structurally similar to adenosine, occupies those receptors without activating them. The downstream alertness effect is predicted by the mechanism and confirmed in hundreds of human trials. This is a clean mechanism-to-outcome chain.
Where mechanism fits in the evidence ladder
Using the standard evidence tier framing:
- Anecdote — one person felt better.
- Preliminary — small human trials, inconsistent results.
- Mechanistic — cell-culture and animal data; a plausible pathway is described at the molecular level.
- Robust human outcome — consistent randomized controlled trial and meta-analysis support at realistic doses in humans who resemble the user.
Mechanism-only evidence is useful context. It is not, on its own, a reason to move a compound into a high-confidence stack slot. This is the single most common source of over-ranked supplements in the wellness ecosystem.
When mechanism predicts outcome, and when it does not
Some compounds bridge mechanism to outcome cleanly. Creatine monohydrate's phosphocreatine mechanism predicted performance effects that RCTs confirmed. Caffeine's adenosine antagonism predicted alertness effects that did the same. PDE5 inhibitors predicted and delivered a vascular outcome.
Many compounds do not. Antioxidant supplementation was supposed to reduce cardiovascular and cancer risk because oxidative stress drives both. Large RCTs of vitamin E and beta-carotene failed, and some showed harm. Homocysteine-lowering via B-vitamin stacks was supposed to reduce stroke risk; trials were mixed. NAD-precursor marketing promises remain well ahead of the human outcome data.
The pattern is structural. Human physiology runs hundreds of parallel pathways with feedback loops, and nudging one target in a dish rarely predicts the downstream outcome in a living person. An effect size measured in humans always outranks a mechanism diagram.
How this appears in Unfair
Unfair stores mechanism and target type as part of each compound's ingredient metadata and surfaces them in the rationale snippet alongside the evidence tier. Mechanism explains the "why" behind a candidate without being mistaken for the "does it work," and the UI visually distinguishes mechanism-only entries from compounds with robust human-outcome data so users do not conflate the two.
Clinical safety note
A compound with a strong mechanism but no human outcome data is not automatically safe, even when the mechanism sounds benign. Antioxidants can blunt training adaptation; methylation support can destabilize mood in sensitive users; adenosine antagonism can trigger arrhythmias in a small subset of people. Always pair mechanism claims with a safety and interaction review, and consult a clinician whenever the target overlaps with a prescribed medication.