When asked if they would investigate a safety concern for a specific drug at an external party's request, the FDA expressed reluctance. Such an analysis would raise questions of bias. Instead, they prefer to address these questions by pooling data from multiple drugs with a similar mechanism of action.
Related Insights
The CREST trial showed benefit driven by patients with carcinoma in situ (CIS), while the Potomac trial showed a lack of benefit in the same subgroup. This stark inconsistency demonstrates that subgroup analyses, even for stratified factors, can be unreliable and are a weak basis for regulatory decisions or label restrictions.
The FDA receives raw and cleaned datasets from sponsors, not just summary reports. Their internal teams conduct independent analyses, which can lead to findings or data presentations in the official drug label that differ from or expand upon what's in the published paper.
Amidst growing turmoil at the FDA, a viable strategy is to "invest around" the risk. This involves prioritizing companies whose drugs show clear data on well-understood, validated endpoints, as these are most likely to navigate the current political environment successfully, regardless of leadership changes.
Academics with novel research questions can collaborate with the FDA. However, due to the confidential nature of sponsor data, all analyses are performed internally by FDA statisticians. External partners provide clinical insight and interpretation on summarized, non-confidential outputs.
Dr. Smith highlights a critical flaw in pharmacology: while a single drug undergoes rigorous FDA testing, there is zero data on the interactive effects when a patient takes two or more drugs concurrently. This 'polypharmacy' creates unpredictable and potentially harmful side effects.
The FDA halted two REGENXBIO gene therapies with similar constructs after a safety event in one trial. However, it spared a third therapy from the same company that used a different design, indicating regulators assess risk at the technology platform level, not just the company or disease level.
The FDA's current leadership appears to be raising the bar for approvals based on single-arm studies. Especially in slowly progressing diseases with variable endpoints, the agency now requires an effect so dramatic it's akin to a parachute's benefit—unmistakable and not subject to interpretation against historical data.
Biohaven's Complete Response Letter (CRL) offers a rare public insight into the FDA's specific statistical objections to using natural history cohorts. The letter details concerns about selection bias and failures in tipping point analyses, serving as a cautionary guide for other companies like Unicure pursuing similar regulatory strategies.
Actuate's CEO advises against out-licensing different indications of a single molecule to separate partners, calling it "splitting the baby." Because all programs rely on a single, shared safety database, one partner's negative safety event would impact all other programs, making the model unworkable.
The FDA is eliminating mandatory animal testing because it's often misleading—90% of drugs passing animal studies fail in humans. The agency is embracing modern alternatives like computational modeling and organ-on-a-chip technology to get faster, more accurate safety data.
