A study published in Mayo Clinic Proceedings has called into question how the Food and Drug Administration (FDA) uses surrogate endpoints in approvals for certain cancer medications.
Surrogate endpoints are measures of the effectiveness of certain drugs in the place of real clinical endpoints, such as overall survival or quality of life. They are often used in accelerated approvals, which is a faster evaluation process by the FDA for drugs that address an unmet need or show significant improvement over drugs currently available. They are also often used in tradition approvals, where a drug must show a “favorable effect on an established surrogate for a longer or better life.”
-----
Related Content
FDA launches new precision medicine tool
FDA approves Gilead Sciences' HIV drug cocktail Genvoya
------
However, in their paper, Chul Kim, MD, MPH, National Cancer Institute (Bethesda, MD) and his coauthor, Vinay Prasad, MD, MPH, Oregon Health and Science University (Portland, OR), argue that these surrogate measures may not be based on adequate data.
They performed a retrospective study of the FDA database with focused searches of MEDLINE and Google scholar to assess the strength of the surrogate-survival correlation in drugs that were approved based on those endpoints. Over a 5-year period, the FDA approved marketing applications for 55 indications based on surrogate end points, 25 of which were accelerated approvals and 30 that were traditional approvals.
Results of an analysis showed that there was no association between surrogates and survival in 14 of the 25 accelerated approvals (56%) and 11 of the 30 traditional approvals (37%). In addition, level 1 analysis—the most effective method of validating a surrogate—was used in only 4 of the accelerated approvals (16%), all of which reported a low surrogate-survival correlation. Among traditional drug approvals, level 1 analysis was used in 15 of the evaluations, 8 (53%) of which had a low correlation, 4 (27%) of which had a medium correlation, and 3 (20%) of which had a high correlation.
From these results, the authors concluded that the use of surrogate endpoints in drug approvals currently lacks formal empirical verification of the relationship between those surrogates and survival.
