A surrogate endpoint is a measurement used in place of the health outcome researchers really want to know about. It might stand in for less pain, better everyday function, or a longer life.
The question “what is a surrogate endpoint?” matters when a headline turns a changed blood test into a claim of better health. The missing piece is whether changing that number reliably predicts the benefit people care about.
What is a surrogate endpoint in FDA's definition?
The Food and Drug Administration (FDA) definition describes a substitute for directly measuring how someone feels, functions, or survives. An endpoint is simply a study's chosen outcome measurement.
Three terms help separate what was measured from what it means:
| Term | Plain meaning | What it establishes |
|---|---|---|
| Biomarker | A measured feature of the body, such as a blood hormone level | Something about the body's biology |
| Validated surrogate endpoint | A substitute backed by strong evidence that treatment-related changes predict a specific health benefit | Evidence about that benefit in the setting studied |
| Clinical endpoint | A direct health outcome, such as pain, walking ability, or survival | What happened to people on that measure |
A biomarker can help track a medicine's effects or side effects without being a substitute for benefit. “Measured accurately” and “predicts better health” are different claims.
Why researchers use a stand-in
Some outcomes take years to observe. A blood measurement may change much sooner. A reliable substitute can make a study shorter and help researchers decide which treatments deserve further testing.
MOTS-c: a metabolic marker isn't a treatment result
MOTS-c is a peptide, a short chain of protein building blocks, studied for how cells handle fuel. Metabolic markers are measurements of those fuel-handling processes.
In Reynolds and colleagues' 2021 Nature Communications study, researchers measured naturally produced MOTS-c around exercise in 10 men. The men weren't given MOTS-c. Treatment experiments in the same paper involved mice.
That human measurement isn't evidence that prescribed MOTS-c improves exercise performance. It isn't a validated substitute for that outcome either. The in vitro versus in vivo guide explains the separate distinction between laboratory experiments and studies inside living organisms.
For a prescription discussion, a metabolic signal and a person's ability to walk farther remain different questions.
What makes a surrogate validated?
Validation means showing that treatment effects on the substitute reliably predict a particular clinical benefit. Statistician Ross Prentice set out a strict test in Statistics in Medicine in 1989: the substitute has to capture the treatment's whole effect on the real outcome. A marker's association with illness isn't enough. People with lower or higher readings might differ for many other reasons.
The FDA's May 2014 final guidance distinguishes validated substitutes from those considered reasonably likely to predict benefit. The latter have less certainty and can support accelerated approval, a pathway for certain serious conditions. Further studies must verify the predicted benefit.
That is a regulatory evidence standard, not a label for any promising blood test. The disease, treatment and patient group all matter.
NAD: a blood number doesn't measure feeling younger
NAD+, short for nicotinamide adenine dinucleotide, is a helper molecule cells use in energy reactions. Measuring it in blood answers a question about that sample. It doesn't directly measure fatigue, memory or lifespan.
As of November 5, 2026, the day this article was written, a Nature Metabolism study published November 4 offers a current example. Its 1,340 measurements in 1,068 adults found that whole-blood NAD didn't decline with age in healthy participants: levels were stable in men and increased in women. Lower levels were associated with cardiovascular disease, meaning disease of the heart or blood vessels, particularly high blood pressure. The researchers measured total NAD, combining NAD+ with its partner form, NADH; the study wasn't designed to establish cause and effect.
Those findings don't show that raising blood NAD treats heart disease or makes someone younger. They also don't test a prescribed NAD+ preparation. An association, an accurate measurement and a treatment benefit each need different evidence.
Tesamorelin: the main endpoint matters more than a hormone rise
Tesamorelin is a peptide studied for stimulating growth hormone release. In Falutz and colleagues' 2007 New England Journal of Medicine trial, 412 adults with HIV and excess abdominal fat were followed for 26 weeks. HIV is the virus that damages the immune system.
The primary endpoint, the study's main chosen measurement, was change in visceral fat: fat deep inside the abdomen. Researchers measured it with CT, an X-ray scanning method. IGF-1, a growth-related hormone measured in blood, was a secondary endpoint, meaning an additional study measure.
An IGF-1 increase wasn't the main treatment result. A smaller fat measurement also doesn't establish fewer heart attacks or longer life. The Egrifta WR prescribing information, marked “Revised: 3/2025,” says long-term cardiovascular safety hasn't been established.
Mistakes that turn a finding into a bigger claim
“Primary” doesn't mean “direct benefit.” It identifies the main measure selected for the study. That measure can still be a substitute.
The shortcut needs evidence behind it. A number might move in an encouraging direction while another effect of the treatment causes harm. Measuring only that number could miss the overall result.
A strong association doesn't establish a treatment effect. A marker can track disease without changing the disease when someone changes the marker.
A result doesn't transfer automatically. Evidence from one disease, species or preparation cannot establish the same outcome for another. That includes carrying branded tesamorelin results over to a compounded preparation, made by a pharmacy for a prescription.
Compounded medications prescribed through Promise are different from FDA-approved products: the preparations offered here are not FDA-approved. An FDA-approved tesamorelin product, Egrifta WR, exists for excess abdominal fat in adults with HIV-associated lipodystrophy, and that approval does not extend to a compounded preparation. FDA's compounding explanation states that it doesn't verify their safety, effectiveness or quality before marketing.
A licensed provider may still prescribe — that decision is between you and your doctor.
Related terms that help with the next result
The distinction also applies across clinical trial phases, the stages of testing medicines in people. Calling a study phase 3 doesn't validate its chosen marker.
An estimand defines exactly which treatment effect a study aims to measure. The endpoint is one part of that question.
A reference range describes results in a comparison population. The IGF-1 levels by age guide explains how that differs from a treatment goal.
In a prescription visit, the useful discussion connects the proposed medicine, the evidence and the health outcome being considered. At Promise, a licensed provider reviews every request, and not everyone qualifies. If prescribed, medication is compounded and dispensed by a licensed U.S. compounding pharmacy.
