Lesson 32 of 36 · 12 min
Surrogate vs clinical outcomes
A biomarker can be useful without being the outcome people ultimately care about. The key is knowing whether the study measured a biological signal, an accepted surrogate or a direct clinical outcome.
Three layers
Pathway engagement, surrogate endpoints and clinical outcomes are not interchangeable.
Biological response
A hormone, receptor or laboratory change can show that the intervention affected a pathway.
Surrogate endpoint
A substitute endpoint is used because it is expected to predict a clinical benefit, but that relationship can be imperfect.
Clinical outcome
An outcome directly reflects how a person feels, functions or survives, such as symptoms, physical function or major events.
Peptide example
More GH or IGF-1 does not automatically mean better body composition, recovery or longevity.
A GH-axis peptide can clearly alter a hormone concentration. That pharmacodynamic effect is real and may be central to the mechanism.
Claims about muscle gain, injury recovery or healthy ageing require studies that actually measure those outcomes rather than treating the upstream biomarker as the benefit itself.
When surrogates help
A validated surrogate can be powerful, but validation is outcome-specific.
Regulators sometimes accept surrogate endpoints when there is evidence that the surrogate reliably predicts a meaningful clinical outcome in the relevant disease context.
A biomarker being correlated with an outcome is not enough by itself to establish that changing the biomarker will produce the desired clinical result.
Endpoint rule
Make the claim no stronger than the endpoint.
If the study measured a biomarker, report a biomarker result. Do not silently upgrade it into a clinical benefit.
Key takeaways
- ✓A biological response proves pathway engagement, not necessarily benefit.
- ✓Surrogate endpoints are substitutes for clinical outcomes and require validation.
- ✓Correlation between a biomarker and an outcome does not prove that changing the biomarker changes the outcome.
- ✓The same surrogate can have different validity in different diseases or contexts.
- ✓Claims should stay attached to the endpoint actually measured.
Further reading
Lesson finished
Ready for the checkpoint?
The quiz opens on its own page. Get every answer correct to mark this lesson complete and continue.
Take lesson quiz →Previous
Bias, randomization & controls ←
Next lesson
PK, PD, half-life & exposure →
Continue when you are ready, or come back to this lesson whenever you need a refresher.