It was precisely 30 years ago, in 1992, that the Food and Drug Administration (FDA) established the Accelerated Approval Program (AAP) in response to the HIV epidemic.
The AAP is a mechanism that permits conditional approval of a new drug that treats a serious medical condition and fills an unmet need based on its impact on a surrogate end point.
A surrogate end point, i.e., a biomarker, may be a laboratory measure, a radiographic image, a physical sign or any measure that is presumed to predict clinical benefit but is not independently a measure of clinical improvement. To qualify, the end point must be considered likely to correlate with clinical improvement. If approved under this mechanism, the drug will need to be tested in larger, more definitive clinical trials, so-called phase 4 trials, using end points that demonstrate benefit within a few years. If these trials are inconclusive, the approval may be withdrawn.
Today, the FDA approves more drugs through this and other related fast-track mechanisms than the traditional route. Approximately 75 percent of the drugs currently greenlighted by the FDA participated in at least one special program designed to speed up the review process.
The AAP has had a positive impact on HIV and cancer chemotherapy, and it is estimated that more than 300 drugs have been approved through this pathway. The point of concern is that the surrogate end point, the scientific basis of the approval, does not always predict clinical utility. In the case of HIV, CD 4 counts – white blood cells that fight infection – and HIV viral load serve as reliable predictors of clinical response. In the cancer space, the ability of end points to predict clinical utility has been mixed with reports indicating that only one-fifth of drugs approved based on surrogate end points improve survival rates in confirmatory trials.
Post approval confirmatory monitoring has been inconsistent. These drugs, at times, stay on the market despite failure to convincingly demonstrate efficacy in phase 4 trials as “dangling approvals,” drugs for which prior authorization continues in spite of lack of clinical efficacy. In the interim, they cost the public billions of dollars and raise unrealistic expectations for treatment.
Further, all biomarkers are not the same. As we move from infectious diseases and cancer to the realm of degenerative brain disorders, the relationship of biomarkers to clinical response becomes increasingly murky. The recent debacle surrounding the approval of Aduhelm for the treatment of Alzheimer’s disease under this mechanism is a good example of the slippery slope in biomarker science resulting in premature approval. In this case, the “original sin” was the approval based exclusively on the drug’s ability to lower levels of amyloid, a putative biomarker of the disease, in the brain despite the fact that substantially lowering amyloid levels did not correlate with clinical improvement.
Fortunately, the Centers for Medicare & Medicaid Services (CMS), after looking at the data more objectively, approved coverage only for patients enrolled in federally approved clinical trials. In this instance, society dodged a bullet and the checks and balances in our system worked.
After spending decades developing compounds that lower amyloid levels in the brain to treat Alzheimer’s, the pharmaceutical industry is moving on to molecules and pathways other than amyloid. Biomarkers have a potentially important role to play in this regard. In the early phases of clinical trials, they can serve to identify promising drugs that are safe and target specific. Perhaps more importantly, early trials should unequivocally establish a link between biomarkers and meaningful clinical end points that can then be used in more definitive phase 3 trials that compare the drug with a placebo — the international gold standard in clinical trials.
The AAP was established 30 years ago and was designed to accelerate drug approval in select, life threatening cases where the standard approval process was too long and cumbersome. But it was predicated on the assumption that the surrogate end point would be demonstrably linked to relevant clinical outcome measures. As the link between the surrogate marker and clinical outcomes becomes more tenuous, the process falls apart and becomes a mechanism for an end run around the regulatory pathway. Enthusiasm and expediency should not replace scientific objectivity.
Congress is poised to give the FDA more resources to monitor late phase 4 trials after drugs receive preliminary approval under the accelerated program. That is necessary but not sufficient to correct the situation. Once a drug receives preliminary approval, close monitoring and a subsequent recall are more difficult to implement. One of the FDA’s primary responsibilities is to ensure that human and veterinary drugs, vaccines and other biological products and medical devices intended for human use are safe and effective.
The FDA should scrutinize the accelerated program rigorously to ensure that surrogate end points are clinically meaningful before approving drugs under this mechanism. The goal of drug therapy is to treat the disease — and not surrogate markers without impacting the disease. The FDA should live up to expectations and its original mandate.
Anand Kumar, MD, MHA is a professor and head of the department of psychiatry at the University of Illinois at Chicago and past president of the American Association for Geriatric Psychiatry.