FDA potency assurance guidance reshapes CGT development

FDA issued its final guidance Potency Assurance for Cellular and Gene Therapy Products on November 2, 2026, and a Federal Register notice of availability followed. The document sets out FDA's recommendations for showing that every lot of a cell or gene therapy can achieve its intended therapeutic effect.

In rare disease programs, a few lots and a small cohort often carry the whole evidence package. The guidance frames potency assurance as a commitment that runs from the IND through licensure. Regulatory, CMC, and clinical operations teams share the consequences, because commercial release specifications are linked to the lots given to trial participants.

What the potency assurance guidance replaces and recommends

According to FDA's announcement, the final text replaces FDA's 2011 potency testing guidance. The final guidance identifies the superseded document as the January 2011 Potency Tests guidance. The Federal Register notice states that it finalizes the draft issued on December 28, 2023.

The guidance covers cell and gene therapy products regulated as biologics under section 351 of the PHS Act. Its stated exclusions include infectious-disease vaccines, bacteriophage, live biotherapeutics, fecal microbiota for transplantation and allergenics.

For products in scope, the guidance recommends a lifecycle, risk-based Potency Assurance Strategy built from five elements:

1. Manufacturing process design
2. Process control
3. Material control
4. In-process testing
5. Potency lot-release assays

The guidance also gives distinct approaches for three product categories. Short shelf-life products should have potency bioassays run immediately after manufacturing, with evaluation after release. Patient-specific, vector-transduced products should show adequate biological activity for each vector lot before use.

For direct viral vector products, potency assays can measure transgene mRNA expression, protein expression or protein activity. Expectations then scale with each phase of development, starting at the IND. At that stage, FDA recommends preliminary critical quality attributes, an initial risk assessment and a described mechanism of action (MoA).

Significant-risk and effectiveness studies should have at least one qualified release assay that quantifies a potency-related attribute and rejects sub-potent lots. FDA may place a study on clinical hold if potency is not adequately assured. Before a BLA, sponsors should reassess the strategy using all quality and clinical data, validate release assays and set acceptable process parameter ranges.

The Federal Register notice describes the guidance as non-binding. The final text says the strategy does not increase sponsor requirements, although statutory lot-release potency testing still applies. DLA Piper's analysis reads the framework as unchanged from the draft and treats potency as a continuing CMC challenge.

CMC readiness can shape CGT timelines, and UX111 is one example of a CMC-related delay, though a single case does not establish a trend. In July 2025, Ultragenyx reported that FDA's complete response letter for this AAV9 therapy for Sanfilippo syndrome type A cited CMC and facility inspection observations. Per the company, the letter raised no issues with the clinical data, and the company did not describe the issues as potency-specific.

Patient characterization is the clinical counterpart

The guidance ties potency to the MoA, the intended indication and clinical data. The regulatory definition it cites allows potency to be indicated by "adequately controlled clinical data." The MoA and indication determine which attributes count as potency-relevant.

Licensed acceptance criteria should also keep distributed lots consistent with the lots given to subjects in pivotal studies. FDA cautions that a statistical link between an attribute and clinical outcomes may not be causal and may call for mechanistic work. In rare disease, FDA's guidance on human gene therapy for rare diseases acknowledges that small populations make those correlations harder.

That rare-disease guidance recommends genetic testing of all trial subjects and natural history data where needed. It also asks sponsors to collect as much pertinent data as possible from every subject. The potency guidance connects each lot to clinical outcomes through the trial, and the rare-disease guidance asks sponsors to know precisely who was in it.

Reading both documents together, we see genetic confirmation, antibody pre-screening and retention in long-term follow-up (LTFU) as the clinical counterpart to a lifecycle potency strategy. They shape how interpretable the pivotal clinical data are.

The potency guidance itself does not mention pre-screening, patient characterization or LTFU, and FDA has not said it expects such data as potency evidence. We made a related argument about patient identification in our analysis of FDA's genome editing guidance.

Pre-existing antibodies show how characterization feeds interpretation. FDA's rare-disease guidance notes that antibodies to any product component may pose safety risks and limit therapeutic potential. Sponsors that exclude seropositive patients should strongly consider developing a companion diagnostic in parallel.

A 2025 study in Gene Therapy notes that pre-existing neutralizing antibodies to AAV capsids are reported in roughly 40% to 70% of people. That range depends on serotype and region. A 2026 multi-province study in China reports prevalence from under 10% to over 90%, depending on serotype and region.

Open questions ahead of FDA's December webinar

FDA will host a webinar on the guidance on December 10, 2026, according to the same FDA announcement. Specification setting is a near-term practical question, because licensed acceptance criteria link back to the lots behind the pivotal evidence. DLA Piper's reading is that pivotal lots should therefore span a realistic manufacturing range.

Follow-up plans run on separate timelines that depend on vector type. FDA's 2020 long-term follow-up guidance recommends 15 years for integrating vectors and up to 15 years for genome editing products. For AAV vectors it recommends up to 5 years, using a risk-based approach.

The primary purpose of LTFU is detecting delayed adverse events, and assessing durability is optional. The same guidance asks sponsors to make every effort to prevent loss to follow-up, and notes that observation often continues after licensure. The operational barriers to long-term follow-up therefore apply to AAV and integrating-vector programs alike, over very different durations.

Several questions remain open:

• The guidance does not specify how much mechanistic work FDA will expect when correlations come from a small cohort.
• In the 2025 Gene Therapy study, one assay format reclassified up to 21.7% of samples a conventional format called non-neutralizing, so assay choice shapes eligibility.
• The guidance is silent on whether durability data from LTFU could inform later reassessment of potency-relevant attributes.