What Is a Surrogate Endpoint? Most FDA Approvals Use One

Problem

You read a biotech press release that says a drug “achieved a 48% objective response rate” in a Phase 2 trial. The stock jumps 25%. You buy. Two years later, the confirmatory trial fails, the FDA moves to withdraw the approval, and the stock drops 60%. What happened? The 48% was a surrogate endpoint: tumor shrinkage, not a clinical endpoint like survival. You bought a hypothesis, not a result.

Solution

A surrogate endpoint is a biomarker or intermediate measure that is reasonably likely to predict clinical benefit but is not itself the clinical outcome. The FDA uses surrogates to approve drugs faster through the accelerated approval pathway, but the approval is conditional on a confirmatory trial proving real benefit. Your job as an investor is to identify whether the headline number is a surrogate or a clinical endpoint, and price in the confirmatory risk.

Step-by-step

Step 1: Learn the difference between surrogate and clinical

A clinical endpoint measures something a patient directly experiences: overall survival, fewer heart attacks, walking farther, fewer seizures. A surrogate endpoint measures something biological that should correlate with that benefit: tumor shrinkage, LDL cholesterol, HbA1c, dystrophin production, antibody levels.

The distinction is not academic. Tumors can shrink on a scan while patients die at the same rate. LDL can drop 40% while heart attacks barely budge. A surrogate is a proxy. The FDA accepts the proxy for serious diseases with unmet need, but only conditionally. Our companion guide on clinical trial endpoints covers the full four-type hierarchy; this guide goes deep on the one that breaks investors most often.

Step 2: Check whether the FDA used a surrogate for accelerated approval

The FDA’s Accelerated Approval Program grants approval based on a surrogate endpoint reasonably likely to predict clinical benefit. The company must then run a confirmatory trial to verify the benefit. If that trial fails, or never gets done, the FDA can withdraw the approval. Roughly 15% of oncology accelerated approvals have been withdrawn since the program began, per Mehta et al. (JCO 2024), and the 2022 FDORA law gave the FDA expedited withdrawal authority so these reversals are happening faster now.

Read the FDA approval letter on FDA.gov. If the letter says “accelerated approval” and names a surrogate as the basis, the drug is on a clock. The confirmatory trial is the second binary event.

Step 3: Read real 2026 examples the right way

I-DXd (ifinatamab deruxtecan), Daiichi Sankyo: extensive-stage small cell lung cancer. The IDeate-Lung01 Phase 2 trial reported a confirmed objective response rate of 48.2% and a median overall survival of 10.3 months. The ORR is the surrogate: tumor shrinkage measured on imaging. The OS is the clinical endpoint that measures how long patients actually live. The FDA granted Priority Review, and the approval, if it comes, will be accelerated, based on the ORR surrogate, with a confirmatory trial required to prove the ORR actually predicts longer survival. An investor who read only the headline bought the surrogate. The clinical question is still open.

Enhertu (T-DXd), Daiichi Sankyo / AstraZeneca: HER2-positive early breast cancer. Two approvals came May 15, 2026. In the neoadjuvant setting (DESTINY-Breast11), the basis was pathologic complete response (pCR): no invasive cancer in the breast or lymph nodes at surgery. pCR is a surrogate. The clinical endpoint is disease-free survival. In the post-neoadjuvant setting (DESTINY-Breast05), the trial hit the clinical endpoint directly: a 53% reduction in the risk of invasive disease recurrence or death (invasive disease-free survival, hazard ratio 0.47). That is the rare case where the clinical endpoint backs up the surrogate. When the surrogate and clinical endpoints agree, the approval is durable. When they disagree, the stock gets hit later.

Tavneos (avacopan), Amgen: ANCA-associated vasculitis. The ADVOCATE trial’s primary endpoint was sustained remission at week 52. That sounds clinical, and it is, but the data behind it was manipulated: unblinded personnel changed five patients from “not in sustained remission” to “sustained remission” after the trial was unblinded, manufacturing a p-value of 0.0132 that was not statistically valid. The FDA proposed withdrawal in April 2026; the NEJM retracted the trial publication in June 2026. We covered that case in our Tavneos withdrawal write-up. The lesson: a surrogate or endpoint is only as good as the data underneath it. Data integrity is a separate risk layer on top of endpoint type.

Step 4: Sort surrogates by validation status

Not all surrogates carry the same reversal risk. The FDA sorts them into tiers:

• Validated surrogates: blood pressure for cardiovascular drugs, hemoglobin A1c for diabetes, viral load for HIV. Decades of data confirm they predict clinical benefit. Low reversal risk.
• Reasonably likely surrogates: objective response rate in some cancers, dystrophin for Duchenne muscular dystrophy, immunogenicity for vaccines. Plausible but not proven. Higher reversal risk.

If the drug was approved on a novel, unvalidated surrogate, the confirmatory trial matters more. Sarepta’s Elevidys for DMD was approved on dystrophin production; functional-benefit questions nearly derailed it, and the FDA has since tightened its stance on DMD surrogates. When in doubt, treat novel surrogates as higher-risk and size the position accordingly.

Step 5: Find the confirmatory trial on ClinicalTrials.gov

Go to ClinicalTrials.gov and search the drug name. Find the confirmatory trial, usually a Phase 3 listed as a “post-marketing commitment.” Three things to check: (1) Is it enrolled? A trial still recruiting two years after approval is a red flag. (2) When is the readout expected? That is your second binary event. (3) Has the FDA scheduled an AdCom to discuss withdrawal? A withdrawal AdCom is the loudest possible warning. The confirmatory readout is the second PDUFA date, and most investors do not have it on the calendar.

Step 6: Price the binary overhang

A drug approved on a surrogate is worth less than a drug approved on a clinical endpoint. The market often ignores this until the confirmatory trial reads out, at which point the stock gaps up or down 30-60% in a session. If you hold a stock where the thesis rests on a surrogate-based approval, you are carrying reversal risk every day. The PDUFA date gets the headlines; the confirmatory trial date is where the real money is made or lost.

Common mistakes

• Buying the headline without checking the endpoint type. A 48% response rate is a surrogate, not a cure. Read past the first sentence.
• Treating accelerated approval as permanent. It is conditional. Roughly 15% of oncology accelerated approvals have been withdrawn. The accelerated approval guide breaks down the reversal math.
• Assuming a surrogate that worked in one disease works in another. ORR predicts survival in some cancers and not in others. Context matters.
• Ignoring the confirmatory trial timeline. The confirmatory readout is the second binary event. Mark it on the calendar next to the PDUFA date.
• Trusting endpoint data without checking integrity. The Tavneos ADVOCATE case shows that a clean-looking surrogate can hide manipulation. Cross-reference the trial registration on ClinicalTrials.gov.

Final checklist

• Is the primary endpoint a surrogate or a clinical outcome?
• If surrogate, is it validated or reasonably likely?
• Was the approval accelerated or traditional?
• Find the confirmatory trial on ClinicalTrials.gov, is it enrolled?
• When is the confirmatory readout expected?
• Read the FDA approval letter for the postmarket commitments
• Size the position for reversal risk until the confirmatory trial reads out

Surrogate endpoints are how the FDA gets drugs to patients faster. That is a good thing for patients. For investors, it is a deferred binary event. Understand which type of endpoint you are buying before you buy the stock.