The FDA’s decision last year to make complete response letters public provides new insight into why therapies sometimes fail to get the regulatory greenlight. Analysts say the information could help sponsors refine their regulatory strategies.
More than half of FDA rejections come down to concerns with manufacturing, according to a recent analysis of complete response letters made public by the agency. Meanwhile, 41% of rejections were due to poor product quality, including impurities or failed stability testing.
This breakdown, from Jefferies analysts, was made possible by an initiative the FDA launched last year to publish complete response letters (CRLs) “promptly after they are issued to sponsors.” While the CRLs are often heavily redacted and their publication delayed due to legal review following rejections, analysts have hailed the initiative as a positive step toward greater visibility into its decisions.
The agency released an initial trove of complete response letters (CRLs) last July, making publicly available the reasoning behind its rejection of more than 200 drug applications. Since then, the agency’s repository of CRLs—which it uploads onto openFDA—has grown to nearly 350 as of publication.
“We appreciate the increased transparency,” Jefferies analyst Andrew Tsai told BioSpace in an email, noting that making the contents of rejection letters available to the public—and shareholders—has made companies more accountable.
It also helps to clear up communication inconsistencies.
In an interview with BioSpace earlier this month, TD Cowen analyst Ritu Baral recalled one case in which an unnamed company told its investors that the FDA’s reasoning for rejecting its drug application was unclear. Once the agency made the CRLs public, however, Baral was able to see that the agency had been clear and consistent across its various communications with the company.
The second CRL for Replimune’s RP1 also pointed to differing perspectives from the FDA and the company regarding previous agreements. While Replimune CEO Sushil Patel contended after the product’s first rejection in July 2025 that there was alignment on the design of the confirmatory study, the agency maintained in its second CRL, issued in April, that its advice “has remained consistent as evidenced by our communications dating back to March 2021 and subsequent interactions.”
“There is proof with these letters that company/management commentary have differed from what actually happened,” Tsai said.
Of course, the FDA’s CRL archive is imperfect, and it has its fair share of critics. In addition to delays and redactions—which Harpreet Singh, chief medical officer at Precision for Medicine said is an inherent limitation to transparency. BioSpace ‘s own review has found that the trove isn’t user-friendly. The files are dumped in a zipped folder for the reader to sort through, often one at a time.
The new publication policy has also sparked concerns over commercial sensitivity and the impact on emerging biotechs, “whose valuations can suffer from public criticisms,” said Singh, a former division director of oncology at the FDA. Indeed, last month, another unnamed pharmaceutical company filed a citizen’s petition asking the agency to reform the new policy. Before releasing a CRL, the FDA should “first inform the manufacturer that it intends to do so,” and give the drug sponsor the opportunity to contest the publication within 10 days, according to the petition.
Still, Singh said, the FDA’s CRL initiative “has significantly improved sponsor accountability and investor communication,” while also reducing the ambiguity around the agency’s decisions.
More importantly, she said, the CRL archive serves as an educational resource for sponsors, allowing them to learn from the missteps of their peers and potentially setting them up for more successful reviews.
Common pitfalls revealed
The rejection letters published so far have imparted a couple of key lessons that could help sponsors refine their regulatory strategies.
“The majority—and perhaps the vast majority—of CRLs boil down to two main issues,” Mizuho analyst Graig Suvannavejh told BioSpace via email. The first are rejections brought about by manufacturing issues, and the second is deficiencies with the sponsors’ data packages, failing to conclusively demonstrate that products are worthy of approval.
Over at Jefferies, Tsai and his team conducted a quantitative assessment of the regulator’s repository of rejection letters. Their analysis showed that indeed, more than half of CRLs cited concerns with manufacturing, primarily linked to facilities that fail FDA inspections. Relatedly, 41% of rejections were due to poor product quality, including impurities or failed stability testing.
A recent rejection for a pre-filled syringe of Regeneron’s high-dose Eylea falls into this group. In its rejection letter, issued last October, the FDA pointed to violations at a production facility run by Novo Nordisk. An inspection from June 2025 to July 2025 revealed contamination with “atypical extrinsic particles”—the agency’s language for unwanted impurities in the plant, such as cat hair and pests.
The Indiana-based plant also “failed to determine a root cause of the contamination, assess the potential impact to the rest of the lot, or evaluate whether similar issues may have occurred in upstream batches,” according to the inspection report.
Issues at the same plant led to the rejection of Scholar Rock’s spinal muscular atrophy therapy in September 2025.
Both Regeneron and Scholar Rock have since found ways to move past these manufacturing hiccups. Regeneron found an alternative filler and resubmitted its package for the pre-filled syringe, and had a PDUFA date in April. Scholar Rock has also refiled its application for apitegromab and has been granted an action date in September.
Aside from manufacturing issues, insufficient clinical data also emerged as a key stumbling block for many drug sponsors, with Jefferies’ analysis showing that 27% of CRLs noted the need for more data. These requests can range from additional subgroup or pharmacodynamic analyses using existing data to the need for another Phase 3 study, according to Tsai.
Such was the case in the FDA’s recent rejection of Replimune’s melanoma drug RP1—the second time the viral immunotherapeutic has failed to achieve approval. In its latest CRL, the agency deemed the company’s data package “insufficient to conclude substantial evidence of effectiveness” of RP1, particularly pointing to Replimune’s use of a single-arm trial.
REGENXBIO also faced clinical roadblocks for its Hunter syndrome gene therapy RGX-121, which the FDA refused to approve in March. The rejection, according to the CRL, was driven by a series of study design problems, including the company’s use of external controls and its enrollment criteria, which the regulator said raised “uncertainty that the enrolled population is representative of the target population.”
Both REGENXBIO and Replimune ran into what Mizuho’s Suvannavejh called the “worst-case scenario” for drug sponsors, in which the FDA “will simply ask for the conduct of a brand-new clinical trial altogether.” While these types of denials are typically catastrophic for a company’s shares—Replimune plunged 20% in the aftermath of its rejection, while REGENXBIO fell 12%—Suvannavejh said there’s often just no way around them.
“The FDA needs to feel highly confident that any new drug that gets approved provides an overall benefit/risk profile that is net positive for patients,” he said.
Setting up for success
On a positive note, the FDA’s trove of CRLs holds up a mirror to the industry, allowing companies to see the blind spots in their regulatory processes, experts who spoke with BioSpace agreed. For instance, Jefferies’ analysis suggests that most “sponsors are seemingly more prepared when it comes to the clinical data package side of things vs CMC/manufacturing,” Tsai said.
Accordingly, he said that running multiple successful pre-approval inspections is of “key importance” for most drug sponsors, helping to minimize the likelihood of rejection. Contracting the services of former FDA reviewers could be helpful in this regard, Tsai suggested.
Singh concurred, saying sponsors “should prioritize facility inspections for reinforcing robust quality management systems,” using “validated analytical methods and strong stability data” to back their manufacturing packages. Companies should also have independent assessors review their applications to determine gaps and other weaknesses well ahead of submission, she said.
This advice is especially important for first-time or otherwise inexperienced drugmakers, according to Singh. These companies should “lean on external expertise,” she said, in addition to comprehensively planning their regulatory activities and meetings and keeping strict timelines.
On the clinical front, both Singh and Tsai agreed that being on the same page with the FDA from the outset and keeping the agency abreast with a company’s development strategy and managing the regulator’s expectations are critical to avoid unexpected rejections.
“Proactive, consistent FDA engagement throughout development is crucial,” Singh said. “Mapping out interactions, scheduling meetings, documenting feedback, and addressing concerns before filing reduces CRL rates.”
These recommendations point to how industry can leverage the FDA’s new CRL publication policy to “learn from common failure patterns and strengthen their regulatory strategies for greater approval success,” Singh said.
For Suvannavejh, however, the FDA’s publication of CRLs goes beyond how it can help sponsors, and indeed, even beyond the biopharma sphere. “This is much more about restoring credibility to an agency whose reputation both in the U.S. and abroad has been significantly damaged over the past two years,” he said.
