Surrogate endpointSurrogate marker · Intermediate outcome
A measurable stand-in for the outcome you actually care about: useful, and sometimes misleading.
A surrogate endpoint is a measurement used to stand in for the outcome you actually care about. Cholesterol standing in for heart attacks, bone density for fractures, HbA1c for diabetic complications, or muscle protein synthesis for muscle gained over months.
They exist because real outcomes are slow and expensive. Demonstrating that a drug prevents heart attacks takes thousands of people over years; showing it lowers LDL takes far fewer people over weeks, which is why surrogates dominate the literature and the marketing.
They are only as good as the link between them and the outcome, and that link is often weaker than assumed. Lowering LDL with statins reduces events, which validates the surrogate for that intervention; lowering it with some other drugs has not always done so, which shows validation is intervention specific rather than general.
The cautionary examples are severe. Drugs that suppressed abnormal heart rhythms after heart attacks, an obviously sensible surrogate, increased deaths in trial. HDL raising drugs improved the number and not the outcome. Some diabetes drugs improved HbA1c while worsening cardiovascular endpoints.
In fitness content, surrogates are everywhere and rarely labelled. Acute hormonal spikes, muscle protein synthesis over hours, blood NAD levels and mitochondrial markers are all sold as though they were outcomes, and each has failed to predict the outcome at least once.
That is why our evidence tiers care about what was measured, not only whether a study exists.
In practice, ask what the study actually measured, and treat a claim resting only on a marker moving as preliminary until someone measures the thing that matters.