Hard Clinical Endpoint

A hard clinical endpoint is an outcome itself worth preventing that can be recorded with little room for judgement: death, myocardial infarction, stroke, hospitalisation, hip fracture, dialysis, amputation. It is patient-relevant, so no further inferential step is needed to explain why it matters, and it is objectively ascertainable, usually through adjudication by a committee blinded to treatment that applies standard definitions to source documents rather than accepting the investigator's label.

Regulators insist on them because of a history of surrogate measures that moved the wrong way. In the Cardiac Arrhythmia Suppression Trial, antiarrhythmic drugs that reliably suppressed ventricular ectopy after myocardial infarction increased mortality, and the trial was stopped in 1989. Torcetrapib raised HDL cholesterol substantially and increased death and cardiovascular events, halting its programme in 2006.

Hard endpoints let a claim of clinical benefit be made without an assumption. Their cost is size and time, since events are rare relative to laboratory changes, so a trial needs thousands of participants followed for years and only becomes feasible in a population at high baseline risk. That is why most peptides, including nearly all research compounds, have never been tested against one, and why the absence of such data is usually a statement about trial economics rather than a hidden failure.

The mistake worth naming is treating a surrogate improvement as a hard-endpoint result. Better body composition, a lower inflammatory marker, denser bone on a scan or a higher IGF-1 concentration are not fractures prevented, infarcts avoided or years lived, and the link has failed often enough that it cannot be assumed. The subtler version substitutes a soft component into an otherwise hard composite, or accepts investigator-reported events without blinded adjudication.