Bad DetoxWellness claims, checked

Reading the Evidence

What a Surrogate Endpoint Is and Why It Flatters a Product

Most nutrition and supplement research measures a marker rather than an outcome. Sometimes the marker tracks the thing you care about, and sometimes it moves in the opposite direction.

Grayscale photograph of hands pointing at handwritten scientific notes on a paper.
Photograph by Artem Podrez via Pexels
General information. This article is journalism, not medical advice, and it cannot know your circumstances. Speak to a qualified professional about anything that concerns you. How we work.

Everything here earned its place by changing an outcome. Nothing about surrogate endpoints in health research is included to round the number up.

What matters most

  • A surrogate is a measurable stand-in for an outcome that takes years to appear.
  • Surrogates are cheaper and faster, which is why they dominate.
  • A marker can improve while the outcome it stands for does not.

The definition and why it exists

A surrogate endpoint is a measurement used in place of the outcome researchers actually care about. Blood pressure stands in for stroke, cholesterol for heart attacks, bone density for fractures and blood glucose for diabetic complications. The real outcomes take years and large numbers of participants to accumulate, which makes them expensive to study directly.

Surrogates allow a study to finish in months with a few dozen participants instead of years with thousands. That is a legitimate research tool, and it becomes a problem when the results are reported as though they were outcomes.

When a surrogate is trustworthy

A surrogate earns trust when changing it has been shown, repeatedly, to change the outcome it stands for. That validation requires large trials measuring both the marker and the real outcome and confirming they move together.

At the dose actually studied, even then, the validation applies to the specific intervention studied rather than to every way of moving the marker. Lowering a marker through one mechanism can produce different consequences from lowering it through another. This is why regulators treat surrogate-based approvals cautiously and often require outcome data afterwards.

When it goes wrong

Medical history contains treatments that improved a marker impressively while making patients worse overall. A drug can correct an abnormal reading while its other effects outweigh whatever benefit the correction brings.

Against the trial data, those episodes are the reason the distinction between marker and outcome is taught so insistently in clinical research. They also demonstrate that the logic of improving the number is not automatically the logic of improving health. A marker is a signpost, and following a signpost is not the same as arriving somewhere.

The surrogates that dominate wellness research

Antioxidant capacity measured in blood, inflammatory markers, immune cell counts and various hormone levels appear constantly. These are cheap to measure, respond to short-term interventions, and produce publishable movement within weeks.

At the dose actually studied, a supplement that shifts an inflammatory marker has not been shown to reduce any illness associated with inflammation. Inflammation itself is a normal and necessary process, so lowering a marker is not self-evidently beneficial. The vocabulary of boosting or reducing markers substitutes an intermediate measurement for a benefit nobody has demonstrated.

How the reporting shifts

A study finds a marker changed, the press release describes improved immune function, and the product page claims fewer illnesses. Each step is a small stretch, and the accumulated stretch across three steps is enormous.

At the dose actually studied, tracing a claim backwards to what was actually measured is one of the most revealing exercises available to a reader. It usually takes ten minutes and it frequently ends at a marker in a few dozen people over four weeks. Doing this once changes how you read every health headline afterwards.

Supplements interact with prescribed medicines, so tell whoever prescribes for you what else you are taking.

The question to ask

Did the study measure how people actually fared, or did it measure something believed to predict how they would fare. If it measured a marker, has that marker been validated as a predictor for this kind of intervention. How long did the study run, and would the outcome of interest have had time to appear at all.

Once you look at who funded it, would the claim survive if written in terms of what was actually measured rather than what it supposedly implies. Most wellness claims do not survive that rewriting, which is precisely why they are never written that way.

Everything above, in order of what to do first

  1. The definition and why it exists. A surrogate endpoint is a measurement used in place of the outcome researchers actually care about.
  2. When a surrogate is trustworthy. A surrogate earns trust when changing it has been shown, repeatedly, to change the outcome it stands for.
  3. When it goes wrong. Medical history contains treatments that improved a marker impressively while making patients worse overall.
  4. The surrogates that dominate wellness research. Antioxidant capacity measured in blood, inflammatory markers, immune cell counts and various hormone levels appear constantly.
  5. How the reporting shifts. A study finds a marker changed, the press release describes improved immune function, and the product page claims fewer illnesses.
  6. The question to ask. Did the study measure how people actually fared, or did it measure something believed to predict how they would fare.

The takeaway

Ask whether the study measured people or measured a marker. The distance between those two is where most health claims are manufactured.

When the marketing is more precise than the study, believe the study.

Questions readers ask

Is cholesterol a bad surrogate?

It is one of the better validated ones for certain interventions, supported by large outcome trials. That validation belongs to specific treatments rather than to every method of changing the number.

Why do researchers use surrogates at all?

Cost and time, since outcome trials need years and thousands of participants. They are a reasonable research tool and a poor basis for a consumer health claim.

Reading the Evidenceendpointsmarkerstrials
More in Reading the Evidence
Sarika Bhandarkar
Editor, Bad Detox

Sarika edits Bad Detox and cuts anything that overstates what a study found.

Also by Sarika Bhandarkar