Surrogate Endpoints: When a Biomarker Is Not the Outcome Readers Care About

What Is a Surrogate Endpoint?

A surrogate endpoint is a substitute measurement — often a lab value, scan result, or biomarker — that researchers use instead of directly measuring how a patient feels, functions, or survives. It only counts as real evidence of benefit once separate research confirms the measurement reliably predicts that clinical outcome.

When a study reports that a treatment “improved” a biomarker, it answered a narrower question than “does this help people live longer or feel better.” Knowing which question was actually answered changes how much weight the finding deserves.

Terms to Know Before You Read Further

• Clinical outcome: A direct measure of whether a person feels better, functions better, or lives longer — for example, fewer heart attacks, improved mobility, or reduced pain scores.
• Biomarker: A measurable characteristic — a blood level, an imaging finding, a genetic marker — used as an indicator of a biological process or a response to treatment.
• Surrogate endpoint: A specific subtype of biomarker that a trial uses as a substitute for a clinical outcome. Not every biomarker is a surrogate endpoint, and not every surrogate endpoint has been studied for its potential to reliably predict one.
• Validated surrogate endpoint: A surrogate endpoint backed by strong mechanistic reasoning and clinical data confirming that moving the surrogate reliably predicts a specific clinical benefit.
• Reasonably likely surrogate endpoint: A surrogate endpoint with a credible scientific rationale, but without enough accumulated clinical data yet to be fully validated.

Why Do Researchers Use Surrogate Endpoints Instead of Clinical Outcomes?

Measuring a true clinical outcome can take years and thousands of participants — waiting to see who has a stroke, for instance, might take a decade of follow-up. A well-established surrogate can shorten that timeline dramatically. Many trials of different blood-pressure medications have shown that lowering systolic blood pressure reduces stroke risk, so researchers can measure blood pressure changes over a shorter trial and reasonably infer a change in stroke risk.

This shortcut is used often: the U.S. Food and Drug Administration reports that between 2010 and 2012, it approved 45 percent of new drugs based on a surrogate endpoint. Surrogate endpoints are also used when directly testing a clinical outcome would be unethical or impractical — for example, deliberately exposing people to a health risk just to measure whether they develop a disease.

As the National Institutes of Health explains in its guide to understanding clinical studies, even the study design underneath an endpoint matters: a randomized controlled trial can support a cause-and-effect conclusion, while an observational study can only show an association. That same caution carries over to endpoints — a strong study design still cannot rescue a surrogate endpoint that has not been shown to predict the outcome you actually care about.

Where Can a Surrogate Endpoint Mislead You?

The risk is that a biomarker can move in the “right” direction without the underlying clinical outcome actually improving — or it can move while an outcome quietly gets worse. This is why the distinction between a candidate, reasonably likely, and validated surrogate endpoint matters: earlier-stage surrogates carry a real chance of not translating into an actual clinical benefit once more data comes in.

It’s also why regulators require continued study after a product is approved on a reasonably likely surrogate endpoint, so the connection to real-world benefit can be confirmed — or corrected — over time.

How Does a Surrogate Endpoint Differ From a Clinical Outcome?

• What it measures — Clinical outcome: Whether the person actually feels, functions, or survives better.
• What it measures — Surrogate endpoint: A biological stand-in believed to predict that result.
• Example — Clinical outcome: Number of strokes occurring in a treatment group versus a control group.
• Example — Surrogate endpoint: Change in systolic blood pressure reading.
• Time to measure — Clinical outcome: Often years, since events like strokes or deaths accumulate slowly.
• Time to measure — Surrogate endpoint: Often weeks to months.
• Strength of evidence needed — Clinical outcome: Directly answers the benefit question; no extra layer of proof required.
• Strength of evidence needed — Surrogate endpoint: Requires separate, accumulated evidence proving the surrogate reliably predicts the clinical outcome.
• Reader takeaway — Clinical outcome: “This is the real-world result researchers are reporting.”
• Reader takeaway — Surrogate endpoint: “This is a proxy — check whether it has been validated before treating it as proof of benefit.”

How Can You Spot a Surrogate Endpoint in a Study Summary?

1. Find the primary endpoint. Look in the study’s methods or abstract for “primary endpoint” or “primary outcome” — this tells you exactly what the researchers set out to measure. For more on how studies are structured before you get to this step, see Understanding Clinical Study Design Types.
2. Ask what the endpoint actually measures. Is it something a person would notice directly (pain relief, hospitalization, survival), or is it a lab value, scan measurement, or biomarker level?
3. If it’s a biomarker, check whether it’s described as validated. Some study summaries and regulatory documents note whether an endpoint is an established, validated surrogate (like blood pressure for stroke risk) or a newer, unproven one.
4. Look for the phrase “reasonably likely” or “accelerated approval.” These signal that regulators accepted a surrogate with a credible rationale but incomplete confirming data, meaning the real-world benefit is still being studied.
5. Check whether a clinical outcome was also measured. Some trials report both a surrogate and a real outcome; if they moved in the same direction, that strengthens confidence.
6. Note what remains unproven. If only a surrogate was measured, treat the finding as evidence of a biological effect — not yet proof that people will feel better, function better, or live longer.

What Does a Real Example Look Like?

Imagine a study summary states: “Participants taking the study drug saw a significant reduction in LDL cholesterol compared to placebo.” Applying the field guide above: LDL cholesterol is a biomarker, not a symptom or event a person experiences directly. Decades of separate research have established LDL cholesterol as a validated surrogate for cardiovascular events in many contexts, so this result carries meaningful weight.

But if the same claim were made about a brand-new, unfamiliar biomarker with no track record, the same reduction would tell you far less about whether patients will actually have fewer heart attacks or live longer. Whether a finding like this applies to you also depends on who was studied and at what dose — see Study Populations, Dosing, and Outcomes for how to check that.

If You See a Surrogate Endpoint, What Should You Do Next?

• If a study reports only a change in a new or unfamiliar biomarker, then treat it as preliminary evidence of a biological effect, not proof of patient benefit.
• If the biomarker is a long-established, validated surrogate (like blood pressure for stroke risk), then the finding carries more weight, though it still isn’t the same as directly measured outcome data.
• If the approval is described as “accelerated” or the endpoint as “reasonably likely,” then expect that confirmatory research measuring the real clinical outcome is still ongoing.
• If both a surrogate and a genuine clinical outcome were measured and moved in the same direction, then that combination is the strongest signal a summary can offer.

Why Does This Distinction Matter for You as a Reader?

Headlines and abstracts often lead with the biomarker result because it’s available sooner and sounds more concrete than “no difference in long-term outcomes yet.” Knowing whether a reported result is a validated surrogate, a reasonably likely surrogate, or a true clinical outcome helps you calibrate how much confidence the finding deserves, without needing a medical or statistics background to do it.

What This Article Does Not Cover

This guide explains the general distinction between surrogate endpoints and clinical outcomes. It does not evaluate any specific drug, supplement, or product, and it is not a substitute for reading a study’s full methods section or discussing a specific treatment decision with a qualified clinician. Endpoint type is also a separate question from a study’s safety profile — see Clinical Study Safety Data for how adverse events and tolerability are measured and reported.

Frequently Asked Questions About Surrogate Endpoints

Is a surrogate endpoint the same thing as a biomarker?

No. A biomarker is a broad category — any measurable indicator of a biological process. A surrogate endpoint is a specific subtype of biomarker that a trial uses as a stand-in for a clinical outcome. All surrogate endpoints are biomarkers, but not all biomarkers are surrogate endpoints.

Does the FDA actually approve drugs based on surrogate endpoints?

Yes. The FDA reports approving 45 percent of new drugs between 2010 and 2012 based on a surrogate endpoint. This can happen through traditional approval, when the surrogate is fully validated, or through accelerated approval, when it is only “reasonably likely” to predict benefit.

What’s the difference between a validated and a reasonably likely surrogate endpoint?

A validated surrogate endpoint has strong clinical data confirming it predicts a specific benefit. A reasonably likely surrogate endpoint has a credible scientific rationale but not yet enough clinical data to be fully validated, so continued post-approval study is expected.

Can a biomarker improve while the real health outcome does not?

Yes, in principle — that’s exactly the risk that validation evidence is meant to catch. A biomarker moving in a favorable direction doesn’t commitment that patients will feel better, function better, or live longer, which is why regulators distinguish between candidate, reasonably likely, and validated surrogate endpoints.

How is a biomarker different from a clinical outcome assessment (COA)?

A biomarker is typically a lab or imaging measurement. A clinical outcome assessment reflects how a person feels or functions, and is measured through a report from a clinician, patient, observer, or a performance-based test — not a biological sample or scan.

Medical Information Disclaimer

This article is for general educational purposes only and does not constitute medical advice, diagnosis, or treatment recommendations. It is not a substitute for professional medical guidance. Do not start, stop, or change any medication or treatment based on this article. If you are experiencing a medical emergency, contact your local emergency services immediately.

By ClinicalStudyConnect.com Research Desk. Sources: U.S. Food and Drug Administration, Surrogate Endpoint Resources for Drug and Biologic Development; National Institutes of Health, Understanding Clinical Studies. Last updated September 11, 2026.