One page, every verified number on the FDA's published rejection letters, with the receipt for each. Re-verified against the live openFDA API, and built to remain the record if the releases stop.
459 published records as of 13 August 2026, of which 444 were typed Complete Response Letter at the last full analysis. Four records pre-date the 2008 rule that created the instrument and were issued as Not Approvable letters. The rest are other actions the FDA files in the same dataset.
| Date | Release | What it contained | Cumulative |
|---|---|---|---|
| 10 July 2025 | First tranche, 200+ | Applications reviewed 2020-2024, all subsequently approved | 200+ |
| 4 September 2025 | Second batch, 89 | Applications pending or withdrawn, not ultimately approved; real-time release policy announced | ~290 |
| 11 May 2026 | Rolling additions | Archive reported at nearly 350 letters (BioSpace) | ~350 |
| April 2026 | Releases stop | A citizen petition filed through Covington & Burling halts publication for three months | paused |
| 10 July 2026 | Releases resume, 14 letters | FDA posts a catch-up batch dating back to 23 April; HHS says it is still evaluating the process | ~440 |
| 22 July 2026 | Phase 3 bulk pull | Full-text download; the analysis in this report | 457 |
| 13 August 2026 | Live API check | Osteal Therapeutics (29 July) and ITM Solucin (7 August) added | 459 |
Yes, for now, and it has already stopped once. Publication halted in April 2026 after a citizen petition filed through Covington & Burling argued the FDA had to involve sponsors in disclosure decisions and guarantee redaction of confidential material. The archive sat still for three months. On 10 July 2026 the agency resumed, posting a catch-up batch of fourteen letters dating back to 23 April, and it has kept adding since. HHS has said only that it is evaluating the process and potential next steps.
Two futures, one page. If releases resume, the table above keeps growing and this record is re-verified with each batch. If the pause hardens into policy, the published archive freezes as the complete public record of FDA rejection letters, roughly 459 records spanning 2002 to mid-2026, and this analysis becomes the definitive read of a closed corpus. Either way, what has already been published cannot be unpublished, and it is analysed here in full.
| Question | Answer | Source |
|---|---|---|
| The 74% figure circulating in trade coverage | Real, but it describes 202 letters, not the archive: 150 of the 202 in the FDA's first tranche, all from applications ultimately approved, classified by keyword | Pharma Manufacturing, July 2025 |
| Share of published letters citing manufacturing or quality | 62.6% carry a manufacturing or quality deficiency section (286 of 457); 62.5-62.7% across both 2026 pulls | Phase 3 Search, full report |
| Most common primary deficiency category | CMC, 45% | McCarthy & O'Boyle 2026, 30 approved-after-CRL NDAs |
| Manufacturing-facility deficiencies | 65%; CMC 51% | Dilek et al. 2026, 43 novel therapeutics |
| Median delay, CRL receipt to approval | 1.28 years | Dilek et al. 2026 |
| Added delay versus first-cycle approval | ~709 days, about 23 months | McCarthy & O'Boyle 2026 |
| Average CRL to final approval | >2.5 years | RSM US LLP, February 2026, first 200+ letters |
| Share of CRL'd NDAs later approved | 64%; 75% CMC-cited, 44.4% safety-cited | McCarthy & O'Boyle 2026 |
| Facility deficiencies caused by incomplete pre-approval inspections | More than half | RSM US LLP, February 2026 |
| Records currently marked unapproved | 32.4% of the analysed archive | openFDA record field; not a recovery rate, see caveat |
We place the CMC, quality and technical operations leaders whose job is to keep a company out of this archive. If you want a structured read on where your own programme sits against it, that is a twenty-minute conversation, and the diagnostic our technical operations practice runs before every engagement.