A drug wins full approval within a year of its first human dose

Where things stand (14 Oct 2026)

No candidate is in sight. CDER's 2026 list of novel drug approvals has 46 entries through 9 Oct 2026, the latest being pritelivir on 10/9 and tiratricol on 9/28. Every one comes from a program that first dosed humans years ago. Examples: zilganersen, daraxonrasib, enlicitide, camizestrant, and the radiotracers floretyrosine and florquinitau. All of the review vouchers under the new ultra-fast voucher program (1–2 month reviews) went to late-stage programs.

The Ebola (Bundibugyo) outbreak in Congo and Uganda is being fought only with existing drugs, all first given to humans years ago: MBP134, remdesivir and obeldesivir. As NPR put it, "To start from scratch takes years."

Reference class and base rate

The reference class is all FDA novel drug approvals. Brown et al. (2022, https://pmc.ncbi.nlm.nih.gov/articles/PMC9869766/) looked at 405 innovative drugs approved 2010–2020 that had first-in-human dates:

• The shortest time to any approval was osimertinib: 984 days, and that was an accelerated approval.
• The shortest time to full approval was elexacaftor: 1,043 days.

The pandemic record is nirmatrelvir (Paxlovid). Its first-in-human study, NCT04756531, has an actual start date of 2021-02-11 on ClinicalTrials.gov.

• Its emergency use authorization came after 314 days, but emergency authorizations don't count here.
• Full approval came on 2023-05-25, after 833 days (2.3 years), computed from those dates.

So across roughly 1,300 novel approvals the count is zero, and nothing has come close. Meeting the 12-month bar would mean cutting the record by more than half. My base rate is about 0.

Paths to YES (rough contributions through 2031)

1. About 0.7%: custom RNA therapies (antisense or siRNA) approved under FDA's "plausible mechanism" framework. These are CDER products. The draft guidance came out on 23 Feb 2026 and allows a first-in-human study to also serve as the pivotal trial (HHS release, 23 Feb 2026; Arnold & Porter, 25 Feb 2026). Gene editing, where most of the momentum is, is a CBER product and doesn't count. Several things also weigh against this path:
   • The framework is still a draft.
   • Makary, Prasad and Høeg have all left the FDA.
   • BioSpace (June 2026) notes that "not many companies are developing N of 1 therapies" and the economics "just aren't there."
   • n-Lorem gives its custom antisense drugs away free rather than seeking approval.
   • Even under the framework, the FDA needs months of follow-up showing clinical benefit, plus the full application and its review.
2. About 0.4%: a pandemic or outbreak drug given full approval, not just emergency authorization, within a year. Even COVID never produced that.
3. About 0.5%: an ultra-compressed conventional program. Examples are a new active ingredient approved mainly by bridging blood-level studies to an existing drug, a diagnostic, or a short-course drug for an acute condition. These could be helped by the new default of a single pivotal trial plus confirmatory evidence, and by HHS's Operation TrialBlazer (launched 22 June 2026). For comparison, milsaperidone, approved by such bridging in Feb 2026, still had an 11-month standard review.
4. About 0.5%: AI-driven changes to drug development or regulation by 2030–31, plus unknowns.

The hazard starts near zero. Any approval before 2028 would need a first human dose in 2026–27, and no such program is visible. It rises gradually after that. Cumulative probability is about 0.3% by end-2027, 1.25% by end-2029, and 2.4% by end-2031. FDA review documents often post months after approval, which pushes resolution a little later.

Strongest case against my estimate

For a higher number: FDA policy is explicitly aimed at compressing timelines (single-trial default, reviews in 1–2 months, first-in-human studies as pivotal trials). There will be about 250–300 approvals in the window, and only one is needed.

For a lower number: physical limits are the binding constraint. Showing benefit and safety takes time, and so does readying manufacturing. The leadership that drove these reforms has left, and the custom-therapy niche is mostly a CBER product area.

I settle near 2.4%.

Sources

• FDA Novel Drug Approvals 2026: https://www.fda.gov/drugs/novel-drug-approvals-fda/novel-drug-approvals-2026
• Voucher program page (updated Aug 2026): https://www.fda.gov/industry/commissioners-national-priority-voucher-cnpv-pilot-program
• HHS release, 23 Feb 2026: https://www.hhs.gov/press-room/fda-launches-framework-accelerating-development-individualized-therapies-ultra-rare-diseases.html
• BioSpace, June 2026: https://www.biospace.com/fda/post-makary-new-rare-disease-framework-faces-fresh-uncertainty
• Arnold & Porter, Feb and June 2026 advisories
• STAT, 24 Jun 2026, and NPR, 7 Jul 2026, on the Ebola trials
• Makary resignation coverage, 12 May 2026

Question. By the last day of each quarter, has FDA's Center for Drug Evaluation and Research (CDER) granted full approval to a novel drug within 12 months of the drug's first dose in a human?

Definitions.

• Novel drug: a new molecular entity approved under a New Drug Application (NDA), or a new therapeutic biological product approved under a Biologics License Application (BLA), identified by CDER as a novel drug in its annual Novel Drug Approvals list or its approval documents. Diagnostic drugs count. Products regulated by FDA's Center for Biologics Evaluation and Research (CBER) do not count.
• Full approval: FDA traditional (also called regular) marketing approval for at least one indication, including conversion of an accelerated approval to traditional approval. Accelerated approval itself is not full approval. Priority Review, Fast Track, Breakthrough Therapy designations, postmarketing requirements, or approval under the Animal Rule do not disqualify an otherwise traditional approval.
• First dose in a human: the earliest documented intentional administration of the drug's novel active moiety or biological product to any human worldwide, including healthy volunteers, exploratory microdosing, expanded access, and administration under earlier development names or by a previous sponsor. The clock is not restarted by a change of indication, formulation, route, sponsor, or country. For a combination containing one novel ingredient, use that ingredient's first human dose; if multiple ingredients are novel, use the earliest first human dose among them.
• Within 12 months: the full-approval action date in FDA's approval letter is on or before the first anniversary of the first-dose date, inclusive. A February 29 anniversary falls on February 28 in a non-leap year. This is not the interval from application submission, trial registration, or first dose in a pivotal trial.

Resolves YES if qualifying public documents establish both CDER full approval and a worldwide first-human-dose date meeting the 12-month limit. The evidence must identify the administration as first-in-human or establish that it was the earliest administration in the drug's development history; merely finding a recent trial is insufficient. A prior accelerated approval can qualify only when traditional approval itself falls within the original 12-month interval. No attribution to AI is required.

If the first-dose date is reported only as a month, year, or bounded interval, YES requires that the approval meet the limit even using the earliest possible first-dose date in that interval. A ClinicalTrials.gov study-start date may supply a conservative lower bound only when the record establishes that the study was the worldwide first-in-human study and that dosing occurred; enrollment alone is not proof of dosing.

Resolves NO if the qualifying evidence has not been published by the quarter-end. Exclude emergency use authorizations, investigational permissions, tentative approvals, accelerated approvals not yet converted to traditional approval, non-CDER approvals, and new indications or formulations of drugs that do not meet the novel-drug definition. Unresolved uncertainty about whether an earlier human administration occurred, or date uncertainty that straddles the 12-month threshold, is insufficient for YES.

Resolving source. Approval status, action date, and novelty must be established by FDA's Drugs@FDA approval letters and review packages, CDER's annual Novel Drug Approvals lists, or other official FDA documents. FDA's Accelerated Approval Program records distinguish accelerated approval from conversion to traditional approval. First dosing may be established by FDA review documents, ClinicalTrials.gov records and posted study reports, peer-reviewed reports authored by trial investigators, or dated official sponsor or trial-site reports explicitly documenting first-in-human dosing. FDA review documents take precedence over conflicting non-FDA sources; otherwise a conflict must be publicly reconciled before YES. These named sources govern rather than A2's AI-developer confirmation rule. Apply the Resolution window and Resolution date (A3): the resolution date is publication of the final qualifying document needed to establish the result, not the dosing or FDA action date. Earlier records may support the chronology but do not themselves trigger resolution.