Potential Pitfalls of Surrogate Markers

Since surrogate markers are expected to substitute for clinical outcomes, the following situations may occur:

1. Perfect Surrogate Endpoint: The full effect of the (drug) intervention on clinical outcome(s) is reflected and predicted by corresponding changes in the marker (perfect correlation). This ideal scenario does not exist (yet), and probably never will, since no single marker can reflect the entire (multivariate) pathophysiology of a disease or pharmacology of a drug [8].

2. Acceptable Surrogate Endpoint: Changes in the marker reflect only partially the (drug) intervention effect on clinical outcomes, e.g., cholesterol for statin drugs, blood pressure for antihypertensives, HbA1c for antidiabetics, etc. These are endpoints that, based on available evidence, are accepted by the scientific and medical community to substitute for clinical outcomes, both in the drug development and in clinical practice.

3. False Positive Endpoint: The drug intervention affects the marker favorably, but has an unfavorable effect on clinical outcome, e.g., premature ventricular contraction (PVC) frequency for antiarrhythmic agents: The placebo-controlled, randomized, double-blind Cardiac Arrhythmia Suppression trial (CAST) demonstrated that various antiarrhythmic agents did suppress PVC frequency in patients with cardiac arrhythmia, which had been thought to predict improved clinical outcome, namely mortality. However, CAST showed excess mortality in the active-treatment groups relative to the placebo group (most likely due to the arrhythmogenic effects of the drugs), disproving PVC suppression as a surrogate marker.

From a regulatory point of view, this appears to be the major concern in using surrogate endpoints to approve drug products for marketing, and necessitates the requirement of adequate and well-controlled clinical phase III trials to demonstrate efficacy.

4. False Negative Endpoint: The drug intervention affects the marker unfavorably (or not at all) but has a favorable effect on clinical outcomes, e.g., Prostate-specific antigen (PSA) in treatment of prostate cancer.

Was this article helpful?

0 0

Post a comment