FDA in Flux — May 2026 Newsletter

Welcome to FDA in Flux — A Mintz newsletter tracking rapid changes in policy and agency actions that impact medical, life sciences, and consumer product investment decisions and development strategies.

FDA Seeking to Streamline Clinical Trials Through AI and Real-Time Data

What is happening: The search for ways to pare down the length of clinical development for drugs and biologics appears to be ramping up. In late April, FDA announced the initiation of two proof-of-concept “real-time clinical trials” (RTCTs), in which the agency is receiving real-time reports of endpoints and data signals from a sponsor-based data platform or through the sponsors’ contracted technology vendors. It also unveiled plans for a short three-to-six-month RTCT pilot program focused on early-stage clinical trials, which is scheduled to launch in summer 2026, and published a call for comments on uses of artificial intelligence (AI) in conjunction with the pilot program. Once the pilot is complete, FDA hopes to scale the program rapidly to shorten the timeline for early drug trials and to move on to an RTCT pilot program for late-stage trials.

Why it matters: FDA’s comments on RTCTs, including statements by former FDA Commissioner Marty Makary (who resigned on May 12) and Chief AI Officer Jeremy Walsh in the April 28 press release and episode 26 of the FDA Direct podcast and during an Industry Day virtual presentation on May 15, envision a clinical trial process in which sponsors use AI to facilitate collection and real-time interpretation and reporting of data signals, with FDA’s own AI systems interpreting the reported data for regulatory purposes. The stated goal is to find the minimum amount of information FDA needs to make a decision on whether the safety or effectiveness endpoints for a given investigational product have been met. FDA clarified that it should only receive trial data that have gone through initial analysis and meet specific reporting criteria established prior to trial initiation, not broad swaths of patient-level data.

The agency hopes such decision-making efficiency will shorten the time needed to advance to the next phase of clinical development and ultimately lead to “continuous” clinical trials with no phase transitions. The use of decentralized clinical trial designs with digital health wearables for data collection are a natural fit with the plan for collecting and reporting real-time data in RTCTs. From a drug development sponsor’s perspective, however, risk minimization strategies, especially in the areas of AI governance and data protection, as well as new methods of patient informed consent and institutional review board approvals, will be critical for such trials, and FDA does not directly control the latter processes (although it can strongly influence broader stakeholder activities in those areas).

Who may be affected: Entities that design, oversee, and perform clinical trials, including sponsors, institutions, investigators, and CROs, will be most affected by an enduring shift toward RTCTs. Traditional clinical trials involve long timelines for active trial performance (including interventional treatment with the investigational product and procedural trial site visits for data collection) and statistical data analysis, to which those entities have become accustomed. The proposed RTCT structure would prioritize fast data analysis and regulatory decision-making, requiring new approaches and methods for executing trial contracts, monitoring subjects, and performing data analyses necessary to determine safety and effectiveness. 

Investors in the life sciences sector are key stakeholders due to the importance of raising capital to fund the clinical development of investigational drugs and biologics, especially for startup companies. Phased clinical trials with clear breakpoints allow startups to promote positive trial results for the purpose of raising additional capital while preparing for the next development phase. An evolving preference for these RTCTs and “continuous” trials would likely require biopharmaceutical startups to seek out deep-pocketed partners to bankroll (or even take over) clinical development.

Those interested in providing feedback to FDA on RTCTs and the use of AI in clinical trials should consider responding to the agency’s RFI on the proposed pilot program. Comments are due by May 29, 2026.

Drug Compounding: New Regulatory Developments for Certain High-Profile Ingredients, but No Legal Status Change 

What is happening: Over the past month, several discrete regulatory steps by FDA in the human drug compounding space have garnered significant public attention. Unfortunately, some mainstream reporting on the legal effects of these administrative actions has been inaccurate. Given the heightened patient interest in accessing certain categories of compounded drugs, investors and entrepreneurs working in this area should be clear-eyed about the current risks associated with such products and what FDA has and has not done in recent weeks.

The latest FDA administrative actions relate to: (1) GLP-1 active pharmaceutical ingredients and (2) unapproved peptides. In each case, regulatory proceedings remain in progress with opportunities for stakeholders to provide input, as outlined below. 

• FDA issued a notice on May 1, 2026 proposing not to place the popular GLP-1 active ingredients semaglutide, tirzepatide, and liraglutide on the affirmative list of ingredients that outsourcing facilities (also known as “503B compounders”) may use for compounding. The notice explains how FDA determined there is no “clinical need” for these ingredients to be added to the 503B affirmative list and provides detailed analyses of the evidence FDA considered in making its preliminary decisions. Public comments on this proposal can be submitted through June 30, 2026.
• A meeting of the Pharmacy Compounding Advisory Committee (PCAC) has been scheduled for July 23 – 24 to discuss the use of seven unapproved peptides for specific conditions and to consider whether they should be included on an affirmative list of ingredients that can be used by compounding pharmacies that operate under Section 503A of the FD&C Act (“503A compounders”). A second PCAC meeting to discuss an additional five peptides will be scheduled before the end of February 2027.
• FDA also removed the 12 peptides that the PCAC will be evaluating from a temporary “category 2” status following the withdrawal of their nominations for the 503A affirmative list. Category 2 comprises substances that were nominated for the 503A list and may still be eligible for inclusion, but based on available information, FDA identified significant safety risks with such ingredients that warrant further evaluation. The unapproved peptides that the PCAC will be reviewing have not been evaluated for safety or effectiveness through controlled clinical trials and have become popular in the US based on anecdotal reports and promotional activities by self-declared “wellness” or “longevity” influencers. 

Why it matters: Importantly, none of FDA’s recent administrative steps alters the existing legal landscape for compounding by either 503B or 503A compounders. 

• GLP-1 active ingredients and 503B compounders: Under the FD&C Act, when not starting from FDA-approved drugs, outsourcing facilities are generally expected to compound using ingredients on the 503B affirmative list, unless the drug is in shortage nationally. Because semaglutide and tirzepatide are no longer in shortage and are not on the 503B affirmative list, outsourcing facilities are not legally authorized to use them for compounding today (liraglutide injection remains in shortage). Accordingly, the agency proposal to reject their nominations and to exclude them from the 503B affirmative list, if finalized, would not shift the legal status quo or current risk profile for any outsourcers that may be engaging in such compounding now. 
  
  It’s unclear whether this proposed action represents one of the “decisive steps” toward restricting GLP-1 ingredients used by compounders that former FDA Commissioner Makary previewed in a February 2026 statement. Considering the limited practical effect of the May 1 notice on entities engaged in drug compounding, we may continue to see more agency actions aimed at protecting consumers from potentially violative compounded drugs.
• 503A compounders and certain unapproved peptides: Unlike outsourcing facilities, 503A compounders may use active ingredients that comply with USP/NF monographs, are components of FDA-approved drug products, or are listed on the 503A affirmative list. Because the 12 peptides do not meet any of the three statutory conditions for compounding from a bulk ingredient, compounding pharmacies are prohibited from using them to prepare finished products for patients, irrespective of their removal from the category 2 interim determination. In addition, should the PCAC’s upcoming assessments result in a committee recommendation that one or more of these peptides be added to the 503A affirmative list, FDA must then undertake a rulemaking process to accomplish that, which may take several months or more to complete.

As alluded above, much more regulatory activity is expected over the summer and into 2027. Compounders and their business partners may have new legal landscapes to contend with upon the culmination of each of these administrative processes. For now, however, such entities should avoid major shifts in business plans, supply chains, or investment decisions until the relevant proceedings have been finalized and there is clarity regarding the ultimate outcome of each of these multifaceted agency actions. 

FDA Asks Stakeholders to Help Identify Drugs for Potential “Repurposing,” Capitalizing on Recent Literature Review-Based Approval

What is happening: FDA is asking “patients, clinicians, researchers, and other stakeholders” to provide information to the agency on both “priority disease areas and potential candidates for drug repurposing, particularly where there are scientific data that could support approval of potential new uses but there appears to be limited commercial incentives to pursue approval of those uses,” such as in areas of unmet medical need. 

The agency is also looking to better understand potential barriers to the development of repurposed drugs — as well as barriers to using a drug for an off-label indication when a physician has determined that such use is medically appropriate for a patient — and how the federal government (not solely FDA) may be able to address those barriers.

Responsive comments should be submitted by June 11, 2026. 

Why it matters: The agency’s repurposing RFI follows its action earlier this year to approve a new indication for an old drug, leucovorin calcium, based on an internal review of published case reports of patients with a rare neurological condition called cerebral folate deficiency and a specific genetic profile. The decision to request an expanded indication for leucovorin came not from the NDA holder or any manufacturer of the drug, which is widely available in generic forms, but from FDA leadership who initially touted it as effective for a much larger patient population. 

Given the recent leucovorin precedent and the Trump administration’s political focus on drug repurposing (which was included as a priority in the September 2025 MAHA Commission Strategy Report), FDA’s decision to engage directly with the public on these difficult questions is welcome. For now, it remains unclear what the agency will do with the “ideas for candidates” it receives in response to the notice or how it (or other agencies) might tackle any influx of complaints about insurance or medical practice barriers related to the off-label use of FDA-approved drugs. Interested stakeholders should consider not only responding to the notice but also monitoring the docket to track how other parties are presenting future proposals or complaints about barriers to FDA and its other “federal partners.” 

FDA Compliance Branch Is Watching for AI Misuse in Manufacturing Operations

What is happening: FDA cited misuse of AI in drug product manufacturing and quality assurance operations for the first time in a warning letter issued on April 2 and posted to the agency’s website on April 14, 2026. The agency’s investigators found that the manufacturer relied on AI to develop “drug product specifications, procedures, and master production or control records” that would comply with applicable regulations. However, according to information disclosed in the letter, no one from the manufacturer’s quality assurance unit reviewed and validated the output from the AI agent prior to product distribution.

Why it matters: The recent warning letter is an important enforcement milestone indicating that the agency is paying attention to manufacturers’ use of AI for regulatory compliance and recordkeeping tasks. Although FDA generally supports broad implementation of AI to support multiple activities throughout the life cycle of drug and biological products, the warning letter makes clear that use of AI agents must be consistent with current good manufacturing practice (cGMP) regulations, such as management and quality unit oversight and validation requirements. Importantly, FDA states that AI agent output for cGMP-related activities must be reviewed and authorized by an “authorized human representative” of the quality unit, emphasizing that manufacturers may not overly rely on AI and must always have a so-called “human in the loop.” 

As the application of AI models continues to expand rapidly, manufacturers planning to use AI for tasks relating to the design, development, or manufacturing of drug products should ensure that written procedures, validations, and oversight firmly linking the AI and its use to the existing quality system are in place before a new tool is deployed. In addition, establishing an AI governance process, including appropriate supplier controls to qualify and audit AI model vendors, is critically important for maintaining compliance. A vendor may claim that its AI model can perform cGMP and other regulatory compliance activities, but as the ultimate responsible party, the manufacturer should always perform its own diligence and validate the model for use in its facilities under qualified human oversight.

Citizen Petition Challenges FDA’s Automatic Release of Complete Response Letters

What is happening: An unidentified pharmaceutical company has petitioned FDA to improve its processes and comply with applicable laws when publicly disclosing complete response letters (CRLs) for drugs and biologics the agency decides not to approve. The citizen petition dated April 20, 2026: 

• asserts that FDA’s actions are unlawful, arbitrary and capricious, harmful to innovation, and damaging to manufacturers’ commercial interests; and
• requests that FDA adopt a redaction and notice process mirroring its existing framework for processing disclosure requests under the Freedom of Information Act. 

Specifically, FDA would need to start providing advance notice, an opportunity to object, and prepublication review to the applicant when the agency decides to release a CRL. Such changes would establish a formal process giving applicants the opportunity to protect confidential information and time to potentially seek judicial review before disclosure.

As we have previously discussed, FDA began releasing CRLs more proactively in September 2025, and the openFDA database that houses them now contains 130 letters for unapproved products (as well as 309 for approved products), when accessed on May 13. 

Why it matters: The petition raises significant confidentiality, compliance, and competitive concerns about the protection of proprietary information submitted to FDA. CRLs often contain detailed discussions of specific deficiencies in a marketing approval application and recommendations on how those deficiencies may be remedied. It points out that competitors may be able to shortcut development timelines, avoid costly clinical trial failures, and potentially surpass the original applicant who may have invested significant time and resources in validating novel trial designs and regulatory strategies. For example, the petition highlights that recent CRLs issued to Applied Therapeutics, Immedica Pharma, Stealth BioTherapeutics, and Lykos Therapeutics and released by FDA did not redact competitively sensitive clinical information, including detailed FDA feedback on trial design and data interpretation.

Additionally, the petition notes that information contained in a CRL could:

• mislead patients, physicians, and investors about a product candidate’s safety and effectiveness;

• discourage innovative clinical trial designs; 

• chill investment in early-stage pharmaceutical development; 

• undermine the iterative scientific dialogue between sponsors and the agency that enables novel regulatory approaches; and

• when released without any regulatory context, trigger adverse market reactions when the sponsor is a publicly traded company, influence investor sentiment, distort valuations, or lead to a decline in investor funding. 

The petition further invokes the Trade Secrets Act’s criminal penalties for unauthorized disclosure and argues that the improper release of confidential commercial or trade secret information may give rise to a Fifth Amendment takings claim. Accordingly, depending upon how the agency responds to the arguments laid out in this weighty citizen petition, the next step may be the assertion of such legal theories against FDA in federal court.

Mintz Associate Raymond Vanderhyden contributed to this edition of the newsletter. 

Subscribe To Viewpoints