Ten years of first-time drug approvals

What counts as a first-time approval

Drugs on FDA CDER's annual Novel Drug Approvals list (new molecular entities and new therapeutic biologics) from 21 October 2016 to 20 October 2026. 2021–2026 come straight from FDA's published tables. FDA has retired the 2016–2020 pages, so those years are rebuilt from Drugs@FDA through openFDA as original NDA/BLA approvals classed Type 1; annual counts land within one of CDER's published totals, and the extras are removed by de-duplication. Vaccines, blood products and gene therapies are regulated by CBER and are not included.

Sponsors and areas

Sponsors are rolled up to their parent as of October 2026 (for example Genentech to Roche, Celgene to BMS, Seagen to Pfizer). Therapeutic areas are assigned from the FDA indication text with keyword rules, with manual fixes where text was missing; a drug gets one area.

Phase history and "halted"

Every approved drug is traced through ClinicalTrials.gov by generic name, brand and code names (for instance MK-3475) found in the same intervention record. Only studies that started before approval are counted. A study is halted when its status is terminated, withdrawn or suspended. Phase 1 trials are often not registered, so a drug can legitimately show no Phase 1.

The funnel

• Programs are drug candidates identified in industry-sponsored Phase 1–3 studies that started 2008–2018 at 26 large sponsors. Aliases are merged when ClinicalTrials.gov lists them together (such as LY2439821 and ixekizumab). A program must carry one of its sponsor's own code prefixes (LY, PF, MK and so on) or match an FDA approval.
• Programs whose active ingredient was already FDA-approved before October 2016 are dropped, otherwise a 2012 approval would read as a failure.
• The cohort is programs whose first Phase 1 started 2009–2018, giving at least eight years to reach approval. Phase 1/2 counts as Phase 2 and Phase 2/3 as Phase 3.
• Halted: the last study at the highest phase reached was terminated, withdrawn or suspended. Stalled: studies finished and no later phase was registered, which usually means a quiet discontinuation. Active: something is still recruiting or running; these are left out of the success rates.
• Success means a novel FDA approval inside this window. A drug approved only abroad, or filed but not yet approved, counts as not approved.

Known biases. Identification of failures depends on how a company labels its studies, so coverage differs widely and a thin sample leans toward its approved drugs. Alias fragmentation can make a program that was renamed after Phase 1 look like a Phase 1 failure. Pooled across all 26 sponsors, the Phase 1 pass rate here is lower and the Phase 3 pass rate higher than commonly quoted industry benchmarks, which fits both effects. Treat company comparisons as indicative.

Revenue estimates

Central peak sales = median for the therapeutic area × orphan factor (0.5) × drug-type factor × priority-review factor (1.1) × sponsor-scale factor, scaled by 0.45 so the un-anchored tail sums to a plausible total. About seventy well-known drugs instead use an editorial peak taken from public commentary. Annual sales follow an S-shaped ramp whose time to peak depends on area and orphan status, with the first year halved as a partial year, and two COVID antivirals use hand-set shapes. Low and high bands are 0.35× and 2.6×. These numbers are for exploration; they are not forecasts and not company-reported sales.

Sources

• FDA CDER Novel Drug Approvals and Drugs@FDA (via openFDA)
• ClinicalTrials.gov API v2
• Design after the Pharma & Biotech Almanac

Code, pipeline and raw extracts: github.com/kuod/ph1-to-fda.