Clinical Trials sometimes use surrogate endpoints as the measure of the primary outcome of the study. Surrogate endpoints are a proxy for what would ideally be measured but for limitations imposed by time, cost or ethical considerations.
However, these surrogate endpoints may not be an accurate predictor for the outcome we would ideally like to assess. This can have an impact upon the relevance and credibility of the study. Additionally, it may become clear later that, while the surrogate endpoint strongly and statistically indicated a benefit, the impact on a clinically important endpoint is not beneficial.
This happened with doxazosin in the ALLHAT Study. It was found that patients treated with doxazosin rather than the other study drugs were at higher risk of mortality despite effective blood pressure lowering in this and other studies.
For example, there are many studies demonstrating the efficacy of rosuvastatin in cholesterol related outcomes such as reductions in low-density cholesterol and triglycerides. However, there are currently no hard outcome data that show reductions in heart attacks, strokes or deaths. These data are available for other statins, such as simvastatin.