On this page
Surrogate endpoints measure biomarkers or intermediate outcomes assumed to predict clinical benefit; clinical endpoints directly measure patient outcomes like mortality, symptom relief, or quality of life. Surrogates enable faster, smaller trials but carry the risk that improvement in the surrogate may not translate to real patient benefit.
- Clinical endpoints directly observe patient benefit or harm; surrogate endpoints observe biomarkers or intermediate states assumed to predict that benefit [F1, F2].
- Surrogates enable faster and smaller trials because they can be measured sooner than clinical outcomes, but this speed comes at the cost of uncertainty about whether the surrogate truly predicts clinical benefit.
- The FDA accepts validated surrogate endpoints as evidence of benefit when strong evidence supports the relationship between treatment effects on the surrogate and meaningful clinical outcomes [F7, F16].
- Between 2010 and 2012, 45 percent of new drugs approved by the FDA were based on surrogate endpoints, reflecting reliance on this pathway for serious conditions with unmet needs.
- Surrogates have failed to predict clinical benefit in some cases: out of 93 accelerated approvals reviewed by the FDA, 5 drugs were pulled from the market due to inability to verify clinical benefit.
When a clinical trial tests a new drug, researchers must choose what to measure. They can measure what matters most to patients: whether people live longer, feel better, or avoid serious complications. Or they can measure something faster and cheaper: a blood test, an imaging result, or a biological marker that is thought to predict patient benefit. This choice shapes the trial's timeline, its cost, the evidence available to regulators, and ultimately whether a drug reaches patients or fails to do so. Understanding the difference between these two types of endpoints, and the gap between them, is essential for anyone reading trial results, regulatory filings, or drug development reports.
The distinction is not academic. It determines which drugs get approved quickly for serious conditions, which ones later prove ineffective, and how much evidence regulators demand before allowing a treatment to reach the market.
Why the distinction matters in drug development
Regulators and researchers face a recurring tension in drug development. Clinical trials take time and money. A trial measuring whether a drug prevents death from heart disease might run for years and enroll thousands of people. A trial measuring whether the same drug lowers blood pressure might finish in months with far fewer participants. Both trials generate evidence, but they answer different questions at different speeds.
This tension is sharpest when patients are suffering and no treatment exists. The FDA created the Accelerated Approval Program to allow drugs treating serious conditions to reach patients faster, based on evidence that might not yet prove clinical benefit. That program relies on surrogate endpoints. The choice between endpoints thus becomes a choice about how much certainty regulators demand, how quickly patients can access new treatments, and how much risk of disappointment the system is willing to accept.
What a clinical endpoint measures
A clinical endpoint is a direct measurement of what happens to patients. Clinical endpoints measure whether people feel or function better, or live longer. These are the outcomes that matter in lived experience: mortality, symptom relief, hospitalization, quality of life, the ability to work or care for oneself.
In early landmark heart failure trials, researchers used all-cause mortality as the definitive endpoint. This measure answers the question directly: does the drug help people live longer? In the SOLVD Prevention trial of enalapril, researchers also tracked heart failure-related hospitalizations, which contributed to FDA approval. Both death and hospitalization are clinical endpoints because they measure real events in patients' lives, not laboratory values or imaging results that might predict those events.
What a surrogate endpoint measures
A surrogate endpoint is a marker, such as a laboratory measurement, radiographic image, physical sign, or other measure, that is not itself a direct measurement of clinical benefit. Instead, it is a measurement assumed to predict clinical benefit. Surrogates are used in trials when clinical outcomes like strokes might take a very long time to study, or when the clinical benefit of improving the surrogate endpoint is well understood.
Surrogates enable researchers to run faster, smaller trials. Instead of waiting years for patients to suffer strokes, researchers can measure blood pressure reduction in months. Instead of tracking AIDS progression over years, researchers can measure CD4 T lymphocyte count and HIV viral load. The surrogate is a proxy, a stand-in for the clinical outcome that matters but takes too long or costs too much to measure directly.
How the two endpoints differ on key criteria
Clinical and surrogate endpoints differ in what they observe, how quickly they can be measured, and how much evidence regulators require before accepting them as proof of benefit. Understanding these differences clarifies why the choice between them shapes the entire trial and approval process.
What each endpoint directly observes
Clinical endpoints directly measure whether people in a trial feel or function better, or live longer. They capture what matters most to patients: mortality, symptom relief, hospitalization, or quality of life. A surrogate endpoint, by contrast, is a marker such as a laboratory measurement, radiographic image, physical sign, or other measure that is not itself a direct measurement of clinical benefit. Surrogates sit biologically closer to the disease process. Blood pressure is a surrogate for stroke risk; CD4 count is a surrogate for HIV disease progression. The surrogate changes in response to treatment, but the patient does not directly experience that change. The clinical endpoint is what the patient experiences; the surrogate is what we measure to predict it.
Speed and cost of measurement
Clinical outcomes often require years or decades of follow-up. Studying stroke as an endpoint demands tracking patients until strokes occur; studying mortality means waiting for deaths. Surrogate endpoints can be measured far more quickly. A blood pressure reading and a CD4 count are measurements used in clinical trials. This speed advantage becomes critical when clinical outcomes, like strokes, might take a very long time to study. Faster measurement means faster trials, lower costs, and quicker answers about whether a treatment works. That speed is why surrogate endpoints are so valuable in drug development. But speed creates risk: a surrogate that changes quickly may not reliably predict a clinical outcome that unfolds over years. The faster measurement is not inherently more trustworthy; it is only more efficient if the surrogate has been shown to predict the clinical benefit.
Strength of evidence for regulatory approval
Clinical endpoints carry the strongest evidentiary weight because they measure what matters to patients directly. A validated surrogate endpoint, one that has undergone extensive testing and is accepted by the FDA as evidence of benefit, can support full approval with the same confidence as a clinical endpoint. A reasonably likely surrogate endpoint is supported by epidemiologic, human genetic and genomic, pathophysiological, therapeutic, human pharmacological, or clinical trial evidence to predict a meaningful clinical benefit, but there may be uncertainty as to the relation of the surrogate endpoint to clinical benefit. This uncertainty is why reasonably likely surrogates can support accelerated approval with lesser evidentiary support, allowing earlier market access for serious conditions with unmet medical need. The regulatory path depends on how thoroughly the surrogate has been validated. Strong evidence that the surrogate predicts clinical benefit permits faster approval; weaker evidence requires post-market confirmation or conditional approval.
When surrogates have predicted clinical benefit accurately
Surrogate endpoints work when the biological link between the surrogate and clinical outcome is strong and well-established. Blood pressure reduction is one such case. Many clinical trials using different blood pressure-lowering medications have demonstrated that reducing systolic blood pressure reduces the risk of stroke. This consistent finding across multiple drugs and populations has validated blood pressure as a surrogate for stroke risk. Researchers can now conduct shorter trials measuring blood pressure and have confidence that the result predicts stroke prevention.
HIV treatment offers another example. HIV viral load and CD4 T lymphocyte count are surrogate endpoints for HIV disease progression. A meta-analysis of 15 clinical trials involving 24 treatment comparisons found significant associations between treatment differences in effect on AIDS or death and the difference in mean change in CD4 cell count.
When surrogates have failed to predict clinical benefit
The risk of surrogate endpoints is that improvement in the surrogate may not translate to patient benefit. This failure has occurred in real trials and real drug approvals. According to the FDA's review, out of 93 accelerated approvals, 37 drugs (40 percent) had not completed post-marketing studies or confirmed clinical benefit. More starkly, 5 drugs (5 percent) were pulled from the market due to an inability to verify clinical benefit. These drugs were approved based on surrogate endpoints that did not ultimately predict the clinical outcomes patients and regulators expected.
These failures illustrate a fundamental problem: a surrogate can be biologically plausible and statistically associated with clinical benefit in one context, yet fail to predict it in another. A drug might lower a biomarker without addressing the underlying disease mechanism. It might lower a biomarker through a pathway that creates offsetting harms. The surrogate improves, but patients do not. This gap between surrogate and clinical reality is why regulators and researchers must carefully evaluate the strength of the surrogate-to-clinical link before relying on it.
How regulators and researchers evaluate the surrogate-clinical link
Regulators use a framework to determine whether a surrogate endpoint is trustworthy enough for approval or market use. Before a surrogate endpoint can be accepted in place of a clinical outcome, extensive evidence must accumulate, including evidence from epidemiological studies and clinical trials. This evidence must demonstrate that treatment effects on the surrogate predict meaningful clinical benefit.
A reasonably likely surrogate endpoint is supported by epidemiologic, human genetic and genomic, pathophysiological, therapeutic, human pharmacological, or clinical trial evidence to predict meaningful clinical benefit, but there may be uncertainty as to the relationship between the surrogate and clinical benefit. This category allows accelerated approval when the link is plausible but not yet fully proven. A validated surrogate endpoint is one where the relationship is known based on strong evidence. Validated surrogates are accepted as sufficient evidence of benefit without requiring additional clinical outcome data.
The evaluation process examines whether the surrogate has predicted clinical benefit consistently across different populations, different drugs, and different trial designs. It considers the biological mechanism linking the surrogate to the clinical outcome. It assesses whether the magnitude of surrogate change correlates with the magnitude of clinical benefit. This rigorous evaluation distinguishes surrogates that are reliable from those that are merely plausible.
What this difference means for trial design and drug approval
The choice between clinical and surrogate endpoints shapes the entire trajectory of a drug's development and approval. A trial using a clinical endpoint like mortality requires years of follow-up, large sample sizes, and substantial cost. It generates definitive evidence of benefit but reaches patients slowly. Between 2010 and 2012, 45 percent of new drugs approved by the FDA were based on a surrogate endpoint, reflecting how often this pathway is used.
A trial using a validated surrogate endpoint can be completed faster with fewer participants. The drug reaches patients sooner. For serious conditions with unmet needs, this speed can matter enormously. The Accelerated Approval Program formalizes this trade-off, allowing drugs to reach patients based on surrogate evidence while requiring post-marketing studies to confirm clinical benefit. However, this approach carries risk. If the surrogate does not predict clinical benefit, patients may receive an ineffective drug, and the drug may later be withdrawn.
For regulators, the choice determines how much certainty is demanded before approval. For manufacturers, it determines trial cost and timeline. For patients, it determines how quickly new treatments become available and how much evidence supports their use. For researchers reading trial results, it determines what conclusion can be drawn: does the trial show that patients benefit, or only that a biomarker changed?
Do this automatically
A trial measuring clinical endpoints like mortality or hospitalization requires following thousands of patients for years or decades, multiplying the labor, infrastructure, and regulatory review needed at each stage. Automating the collection, validation, and analysis of surrogate endpoint data, such as laboratory values or imaging results, can compress these timelines and reduce the manual burden of trial management. Healthcare research reads ClinicalTrials.gov records, regulatory filings, and published literature directly, and cites each claim to the record it came from, across as many subjects as you need, with Mimir attaching a citation to every figure. It will not predict a trial's outcome or advise on anyone's treatment. Those decisions stay with clinicians and regulators, and deliberately so. Every plan includes it, including the free one.
Start your first report and run it on something you already know well.
This article is general information about healthcare and life-sciences topics. It is not medical advice, diagnosis, or treatment guidance, and it is not a substitute for a qualified professional. Any treatment decision should be discussed with a clinician.
