Apixaban Has Three Clocks Running and Only One Is a Patent

Most analysis of apixaban is a patent countdown. That is the least interesting of the three clocks now running on the molecule, and the one whose date is already known.

Eliquis is not one bet. It is three bets, on three calendars set by three different institutions: a court, a statute, and a clinical guideline. The third is the one that will decide how much of the molecule’s revenue is left standing when the first two have done their work.

The patent clock: 1 April 2028

In September 2021 the U.S. Court of Appeals for the Federal Circuit affirmed the district court’s August 2020 finding that the composition-of-matter patent (US 6,967,208) and the formulation patent (US 9,326,945) covering Eliquis were valid and infringed. The Bristol Myers Squibb–Pfizer alliance then stated the consequence with unusual precision: “the earliest that generic manufacturers are permitted to launch their apixaban products is April 1, 2028, subject to additional appeals and challenges.”

That date is public, fixed, and already in every model. It is also the least informative of the three clocks, because the price a generic enters against in 2028 is not the price apixaban commanded in 2024. Someone changed that in the meantime, and it was not a competitor.

The price clock: 1 January 2026

The Inflation Reduction Act’s Medicare Drug Price Negotiation Program established Maximum Fair Prices on ten Part D drugs for initial price applicability year 2026, effective 1 January 2026. Eliquis was one of the ten.

CMS’s own accounting of what that cohort represents is the useful part. In 2023 those ten drugs accounted for USD 56.2 billion of Part D gross covered prescription drug costs — about 20% of the programme total — and were dispensed to about 8.8 million of the 54 million people with Part D coverage. Beneficiaries spent USD 18.9 billion out of pocket across all Part D drugs that year, USD 3.9 billion of it, or 20.6%, on these ten alone (calculation). CMS estimated that had the negotiated prices been in effect during 2023 they would have saved about USD 6 billion in net covered prescription drug costs, 22% lower net spending in aggregate, and roughly USD 1.5 billion for beneficiaries under the defined standard benefit.

The mechanism matters more than the aggregate. The maximum fair price is set per 30-day equivalent supply. It is a unit-price control applied to a molecule whose clinical advantage is a dosing interval — and dosing interval is a volume variable. A price set per 30 days and a benefit delivered per dose are not the same lever, and the difference lands in gross-to-net.

What the company’s own arithmetic shows

Bristol Myers Squibb reported full-year 2025 revenues of USD 48,194 million against USD 48,300 million in 2024 — flat reported, down 1% excluding FX. Its Growth Portfolio rose 17% to USD 26.4 billion and is now more than half the company. Guidance for 2026 is USD 46.0–47.5 billion, a midpoint of USD 46.75 billion, which is 3.0% below the 2025 outturn (calculation).

Buried in the fourth-quarter commentary is the sentence that dates the second clock. Legacy Portfolio revenue fell 15% to USD 5.1 billion even though “demand increased for Eliquis,” with the decline attributed to “expected continued generic impact across the remainder of the Legacy Portfolio as well as the impacts from higher U.S. government channel rebates.”

Read it again, in that order. Volume on the molecule went up. The government rebate line went up. Revenue is guided down. Eliquis in 2026 is not losing to a competitor. It is losing to a channel, and the channel was designed by statute.

The third clock: the procedure, not the pill

Prevention is only half of stroke economics, and the other half is moving. Our Global Stroke Market forecast puts the category at USD 45.60 billion in 2025 rising to USD 78.90 billion by 2033 at a headline 6.98%. The published endpoints actually compound at 7.09% across the eight years (calculation), so the headline rate is if anything slightly understated rather than inflated.

Inside that forecast the largest segment is ischemic stroke treatment, which is apixaban’s territory. The fastest-growing segment is mechanical thrombectomy devices. That is the same structural pattern we identified in ophthalmology this month: the molecule deflates on dosing interval while the revenue migrates to the procedure that resolves the event. A stroke market growing at seven percent does not rescue an anticoagulant that is losing share of the prevention decision or price on the unit. It dilutes it, because the incremental dollar is being spent in the cath lab rather than the pharmacy.

What the three clocks mean together

For anyone holding apixaban exposure — a manufacturer’s P&L, a payer’s budget, a distributor’s inventory — 2028 is not the interesting date. The interesting question is how much room a 2028 generic has beneath a 2026 administered price, and whether the prevention volume it is competing for still exists at that point or has been partly absorbed by a procedure that treats the stroke after it happens.

The molecule will get its generic. The market it hands over will be smaller per patient and more procedural than the one it launched into, and neither of those things is a patent event. Watching the courthouse is the easy half of the analysis; the statute and the guideline are the halves that set the price and the volume.