What the D.C. Circuit’s Teva ruling actually opened in IRA negotiation

The first appellate crack in the Medicare Drug Price Negotiation Program is not a hole in the statute. It is a question about a word Congress used and a test the agency wrote around it. On 18 August 2026, a D.C. Circuit panel in Teva Pharmaceuticals USA, Inc. v. Kennedy held that Teva’s facial challenge to CMS’s “bona fide marketing” standard is ripe, reversed the district court on that claim, and sent it back. The same opinion affirmed CMS’s decision to treat Austedo and Austedo XR as one qualifying single-source drug, and it rejected Teva’s due-process attack on the program. Circuit Judge J. Michelle Childs, writing for the panel, drew the limit of the remand in one sentence: “The district court never addressed whether CMS’s standard comports with the IRA, and we leave that question for it to consider in the first instance.”

That split is the story. Round-one maximum fair prices have been in effect since 1 January 2026; round two, which KFF notes includes Ozempic and Wegovy, takes effect in 2027; round three — the first with Part B drugs — is being negotiated this calendar year. Teva did not take Austedo off the list and did not win a merits ruling that CMS’s generic test is unlawful. What it won is permission to argue that the agency added a condition the Inflation Reduction Act does not contain. Bloomberg Law framed the ripe claim as a fight over when a generic has genuinely entered the market so as to affect selected medicines. Fierce Pharma was equally careful: the appeals court did not decide whether the standard is legal, only that the lower court must.

The off-ramp Congress wrote

The statute’s deselection logic is short. A covered small-molecule brand is a qualifying single-source drug if it is FDA-approved and marketed, at least seven years have passed since approval, and it is not the listed brand for any generic that has been “approved and marketed” — 42 U.S.C. § 1320f-1(e)(1)(A). Once selected, a drug ordinarily leaves the list in the first year beginning at least nine months after CMS determines that a generic has been approved and marketed.

Congress did not define “marketed.” CMS did. As the D.C. Circuit recited from the 2026 and 2027 negotiation guidance, a generic has not necessarily been marketed simply because it has reached the market. To make that judgment, CMS considers Prescription Drug Event data submitted by Part D plan sponsors and Average Manufacturer Price data reported by manufacturers. No single dataset controls. The agency describes a “holistic inquiry” that may also ask whether the generic remains “regularly and consistently available for purchase” and whether licensing agreements restrict distribution. The test, the court observed, asks not merely whether a generic has made a sale, but whether it has entered the market in earnest.

That is the seam. An ANDA approval plus some commercial activity is, on Teva’s reading, what the statute requires. CMS wants evidence of regular, consistent availability before it will treat the brand as no longer a qualifying single-source drug. The district court never reached whether that gloss exceeds § 1320f-1. It held the claim prudentially unripe because Teva had not shown FDA approval of the relevant generics — a prerequisite, in that court’s view, to any CMS determination that a generic is “marketed.” The D.C. Circuit disagreed. The challenge presents a pure question of law and attacks a generally applicable legal standard rather than a drug-specific determination. The IRA’s review bar, Childs wrote, “covers CMS’s drug-specific determinations, not the generally applicable legal standards that govern them.”

What grouping shut

The other half of the opinion is a loss, and it is the half that actually moves Austedo. CMS grouped Austedo and Austedo XR — Teva’s Huntington’s disease and tardive dyskinesia franchise, Fierce Pharma notes — as one qualifying single-source drug because they share the same active moiety and manufacturer, even though FDA approved them under separate NDAs. The IRA directs CMS, when calculating expenditures, to aggregate data “across dosage forms and strengths of the drug,” including new formulations (42 U.S.C. § 1320f-1(d)(3)(B)). Separate NDAs, the court said, do not necessarily mean separate drugs.

Teva had argued that Austedo XR did not meet the seven-year clock on its own. The panel held that the IRA permits the grouping, and that the Negotiation Program does not deprive Teva of a protected property interest. Austedo and Austedo XR remain on the round-two list. CMS has already posted an MFP explanation file for Austedo; Austedo XR, with prices scheduled to apply on 1 January 2027. Teva did not get them pulled.

What the seam is worth in 2026

The live commercial question is how fast a selected brand can leave the program once a generic or biosimilar is approved, and how much launch volume CMS will demand before it will say the competitor is “marketed.” Teva sits on both sides of that line. The opinion notes that the company has developed generic versions of five innovator drugs selected for the 2027 initial price applicability year: Xtandi, Ofev, Linzess, Xifaxan, and Otezla. A thin launch that CMS will not call bona fide marketing keeps the reference brand — and the maximum fair price — in the program longer.

That timing now matters beyond Austedo. On 27 January 2026, CMS selected 15 drugs for the third cycle, the first to include Part B products, with negotiations in 2026 and prices effective 1 January 2028: Anoro Ellipta, Biktarvy, Botox/Botox Cosmetic, Cimzia, Cosentyx, Entyvio, Erleada, Kisqali, Lenvima, Orencia, Rexulti, Trulicity, Verzenio, Xeljanz/XR, and Xolair, plus Tradjenta for renegotiation. Between November 2024 and October 2025 those 15 drugs reached about 1.8 million people with Medicare and accounted for about $27 billion in Part B and Part D spending, roughly 6 percent of the total. On 13 March 2026, CMS announced that the manufacturers of all 15, and of Tradjenta, had chosen to participate.

None of that predicts that launches will now automatically deselect brands. The D.C. Circuit did not write a new test, did not strike down the IRA, and did not pull a selected drug. It held that a facial challenge to the existing bona fide marketing standard is ripe, and it left the merits to a district court whose timeline is unknown. Until that decision, the implementing rules remain in force, Austedo’s grouping stands, and the only thing 18 August opened is the courthouse door on a single generally applicable standard.