A case note on the D.C. Circuit’s August 25, 2026 decision affirming in part and reversing in part the FDA’s refusal to find that Actavis forfeited its 180-day generic exclusivity for rifaximin.
Norwich Pharmaceuticals, Inc. v. Kennedy is a Hatch-Waxman forfeiture case, not a patent case, but it turns on the same kind of close statutory parsing patent litigators see in ANDA disputes generally. Norwich challenged the FDA’s refusal to find that Actavis — the first filer for generic rifaximin — had forfeited its 180-day marketing exclusivity, which was blocking final approval of Norwich’s competing ANDA. Norwich raised two independent forfeiture theories, tied to two separate statutory provisions, and the D.C. Circuit reached opposite conclusions on each: it affirmed the FDA on the failure-to-market provision, and reversed the FDA’s causation standard under the failure-to-obtain-tentative-approval provision, remanding for the agency to apply the correct one.
The Statutory Scheme
An Abbreviated New Drug Application (ANDA) lets a generic manufacturer “piggy-back” on the safety and efficacy data that supported approval of the corresponding brand-name drug. Where a brand-name drug is covered by patents, the ANDA applicant must certify its position on each one. A Paragraph IV certification asserts that a listed patent “is invalid or will not be infringed” by the generic. A section viii statement takes a different route: it “carves out” a patented method of use from the generic’s label entirely, so the applicant is not certifying against that patent — it is simply not seeking approval to market for that use.
The first applicant to submit a substantially complete ANDA containing a Paragraph IV certification for a given drug is rewarded with a 180-day exclusivity period, during which the FDA cannot grant final approval to any subsequent applicant with a matching Paragraph IV certification. That exclusivity compensates the first filer for the cost and litigation risk of challenging the brand’s patents. But Congress worried the same exclusivity could be gamed — a brand and a first generic filer could collude to delay the first filer’s own launch, which would delay the exclusivity clock and, with it, every subsequent generic’s entry. In 2003, Congress added a set of forfeiture provisions: circumstances under which the first applicant loses its exclusivity, clearing the way for later filers.
Two forfeiture provisions were before the court. The failure-to-market provision, 21 U.S.C. § 355(j)(5)(D)(i)(I), forfeits exclusivity if the first applicant does not begin marketing within 75 days of a triggering event — such as a final judicial decision of invalidity or noninfringement — occurring “as to each of the patents with respect to which the first applicant submitted and lawfully maintained a certification qualifying the first applicant for the 180-day exclusivity period.” The failure-to-obtain-tentative-approval provision, 21 U.S.C. § 355(j)(5)(D)(i)(IV), forfeits exclusivity if the FDA has not tentatively approved the ANDA within 30 months of filing, unless “the failure is caused by a change in or a review of the requirements for approval . . . imposed after the date on which the application is filed.”
The Facts
Salix markets rifaximin as Xifaxan, indicated for irritable bowel syndrome with diarrhea (IBS-D), hepatic encephalopathy (HE), and travelers’ diarrhea. Actavis (a Teva subsidiary) filed the first ANDA for 550 mg rifaximin in December 2015, with Paragraph IV certifications against Salix’s patents covering the drug itself, the IBS-D use, and the HE use. Salix sued, and the case settled in 2018: Actavis admitted infringement of certain patents (without conceding validity) in exchange for a license to market beginning no earlier than January 2028 — more than 22 months before the last Salix patent expires.
Actavis still needed FDA approval, which it did not receive until March 2026, roughly nine years after filing. Part of that delay traced to a March 2017 FDA draft guidance recommending additional dissolution studies to demonstrate bioequivalence for rifaximin ANDAs specifically.
Norwich, a later filer, submitted its own ANDA (No. ‘370) and, in a separate suit, litigated against Salix over an earlier ANDA (No. ‘369). By the end of that litigation, Salix’s IBS-D patent was held invalid, but its HE patent was held valid and infringed — a judgment the Federal Circuit affirmed. Norwich then amended ANDA ‘370 to add the 550 mg dosage, and addressed the HE patent not with a Paragraph IV certification but with a section viii carve-out, electing to market only for IBS-D. Norwich’s ANDA otherwise matched Actavis’s Paragraph IV certifications for the drug patent and the IBS-D patent, but not for the HE patent, since Norwich was not certifying against it at all.
In January 2025, the FDA found Norwich’s ANDA ‘370 substantively approvable but concluded Actavis’s 180-day exclusivity still blocked final approval, rejecting Norwich’s arguments that Actavis had forfeited that exclusivity under both provisions. Norwich sued under the APA; the district court granted summary judgment for the FDA, Actavis, and Salix on both forfeiture theories. Norwich appealed.
Issue One: Failure to Market — Does Forfeiture Require a Triggering Event for Every Certification, or Only the “Matching” Ones?
The failure-to-market provision only forfeits exclusivity once a triggering event occurs “as to each” of the first applicant’s qualifying certifications. No triggering event had occurred for the HE patent — Norwich’s litigation had resulted in the HE patent being upheld and found infringed, not invalidated. But Norwich did not have a Paragraph IV certification against the HE patent at all; it used a section viii carve-out instead. Norwich argued that Actavis’s HE certification should not count as a “qualifying” certification for forfeiture purposes at all, on the theory that forfeiture should be assessed only against certifications that match across the two ANDAs — meaning a triggering event had occurred as to every certification that actually mattered to Norwich, and Actavis should be treated as having forfeited.
The FDA’s position, which the court adopted, is that every Paragraph IV certification the first applicant “contain[ed] and lawfully maintain[ed]” in its own ANDA is a qualifying certification, regardless of what a later applicant chose to certify against. The court got there by tracing “qualifying” certifications back to the statute’s definition of “first applicant” itself: an applicant qualifies for the exclusivity period by being a first applicant, and it qualifies as a first applicant by being first to submit and maintain even a single Paragraph IV certification. It follows that every Paragraph IV certification in that same, timely, complete ANDA is a qualifying certification for forfeiture purposes.
It is worth being precise about what actually decided this issue, because the court’s own opinion is candid that the textual dispute was closer than the FDA’s ultimate victory might suggest. Norwich’s counterargument had real force: the forfeiture provision crossreferences subparagraph (B)(iv), which — read from subparagraph (B)(iv)(I) rather than (B)(iv)(II) — describes the exclusivity period from the perspective of the subsequent applicant’s certification (“such a certification”), suggesting a relational reading under which exclusivity, and its forfeiture, tracks overlapping certifications between the first and later ANDA. The court expressly called the competing readings of that language “equally plausible” and declined to resolve which subparagraph controls the qualifying-certification definition. The court likewise left open — because it did not need to reach it — the separate, threshold question of whether only “matching” certifications trigger the exclusivity provision at all in the first instance; it bypassed that dispute because Norwich’s ANDA already matched Actavis’s on other patents regardless of how the HE-patent question came out.
What actually decided the case was a structural point the court found dispositive independent of the unresolved textual questions: the statute treats forfeiture as an all-or-nothing event. Section 355(j)(5)(D)(ii) forfeits “the 180-day exclusivity period” — a single, definite-article, singular noun — “by a first applicant” and “with respect to that first applicant,” not with respect to any particular later filer. Section 355(j)(5)(D)(iii) then describes the consequence of forfeiture in the same binary terms: once “all first applicants forfeit,” the exclusivity provisions no longer limit approval of any Paragraph IV application at all.
Nothing in the forfeiture provisions supplies a mechanism for a partial or relational forfeiture — exclusivity forfeited as to Norwich specifically, but preserved as to some other, hypothetical applicant who might certify against the HE patent. To adopt Norwich’s reading, the statute would need affirmative language allowing a first applicant to forfeit exclusivity “with respect to” one subsequent applicant while retaining it against others; no such language exists. On that structural ground alone — without resolving either open textual question — the court held that Actavis’s HE certification was a qualifying certification, no triggering event had occurred as to that patent, and Actavis therefore had not forfeited exclusivity under the failure-to-market provision.
Issue Two: Failure to Obtain Tentative Approval — What Does “Caused By” Require?
Actavis missed its 30-month deadline to obtain tentative approval. The statutory exception excuses that failure only if it was “caused by a change in or a review of the requirements for approval . . . imposed after the date on which the application is filed.” The FDA found that its March 2017 draft guidance recommending additional dissolution studies was “one of the causes” of Actavis’s delay, and expressly declined to apply a but-for causation standard, treating the guidance as sufficient even though it might not have been the reason approval was ultimately delayed past the deadline.
The court held this was legal error, treating but-for causation as “one of the traditional background principles against which Congress legislates” — the default reading for an undefined causal phrase absent some textual or contextual signal to the contrary, per Burrage v. United States and University of Texas Southwestern Medical Center v. Nassar.
Why the multiple-sufficient-causes exception didn’t apply?
The FDA’s real position was an unstated invocation of the narrow multiple-sufficient-causes exception to but-for causation — the doctrine that lets courts find causation even where a specific act was not strictly necessary to the outcome, because something else would have produced the same result anyway. The court’s baseball illustration makes the mechanics concrete: after a 3-0 win built on three solo home runs, no single home run is a but-for cause of the win, since the team would have won without any one of them, but each is independently sufficient to have caused it.
The reason that exception is reserved for “rare” cases matters more than the label. The doctrine exists to prevent a specific, perverse result: letting multiple independently sufficient wrongdoers each escape liability precisely because the other’s conduct alone would have caused the harm. The court’s own illustration is the classic case — where a stabbing and, moments later, an independent gunshot both inflict fatal wounds, and the assailant remains liable for homicide even though the victim would have died from the other wound regardless. The D.C. Circuit had applied that same logic in Kilburn v. Socialist People’s Libyan Arab Jamahiriya, holding that a strict but-for reading of the Foreign Sovereign Immunities Act’s “caused by” language could not be right where it would let two state sponsors of terrorism who jointly and independently caused a death each escape jurisdiction by pointing at the other.
Excusing Actavis’s forfeiture on the same theory would run the analogy backwards. Rather than preventing two wrongdoers from each escaping liability, it would hand a windfall to a single applicant — excusing an independent, unrelated deficiency in its own ANDA merely because an FDA guidance change also happened to contribute to the delay.
Congress’s baseline rule is forfeiture for missing the 30-month deadline; the exception exists to excuse delay genuinely attributable to a new or revised requirement, not to rescue an application that was never going to make the deadline for other reasons. The court gave the FDA’s position no deference on this point, noting under Loper Bright Enterprises v. Raimondo that the agency’s technical expertise “has little to do with” resolving the ordinary legal meaning of “caused by.”
The holding is narrower than a blanket rejection of multi-cause reasoning, and the opinion is explicit about the limit: the court rejected only “the multiple-sufficient-causes exception the FDA has urged here.” It separately noted that where the FDA changes multiple requirements close in time, a different approach — aggregating the set of changes and asking whether that combined “change” was a but-for cause of the delay — might be appropriate. That aggregation question was not presented on this record and remains open for a case where it is.
Remand, Not Resolution
Norwich asked the court to decide the but-for question itself, arguing the outcome was “preordained” on the existing record. The court declined, for reasons grounded in the specific posture of the case rather than a generic preference for agency-first review.
The FDA’s own decision had expressly found it “unnecessary to determine whether any additional bases for non-forfeiture exist[ed]” once it identified the 2017 guidance as a cause — meaning the agency never actually asked, and the record was never developed to show, whether the guidance was a but-for cause, or whether some other requirement or deficiency in Actavis’s application would have blocked approval regardless.
Under SEC v. Chenery Corp., the district court could not have affirmed on a rationale the agency itself never articulated, and the appellate record before the court was Norwich’s ANDA file, not Actavis’s — the record actually needed to answer the but-for question for Actavis’s own application. The court therefore remanded to the district court with instructions to remand to the FDA to apply but-for causation in the first instance.
Two Provisions, Two Outcomes
The decision’s real usefulness lies in the contrast it draws between two forfeiture theories that sound similar but operate on entirely different logic.
Failure to market: the FDA reached the correct result, but the court did not adopt the FDA’s entire textual rationale. Instead, it resolved the case on the structural point that the 180-day exclusivity period is a single, indivisible period — there is no statutory mechanism for forfeiting exclusivity against one subsequent applicant while preserving it against another.
Failure to obtain tentative approval: the FDA was wrong about the governing legal standard. “Caused by” requires but-for causation, and the narrow multiple-sufficient-causes exception the FDA invoked does not apply merely because a regulatory change was one of several contributing factors — the court found that exception inapplicable here because this case involved a single applicant whose independent deficiencies could have caused the same delay.
The open question the case leaves behind is concrete: would Actavis have obtained tentative approval by June 18, 2018, if the FDA had not issued its March 2017 dissolution-study guidance? If the answer is no — if some other deficiency in Actavis’s ANDA would have blocked approval regardless — the guidance change cannot excuse forfeiture under the statutory exception. If the answer is yes, Actavis retains its exclusivity under this provision as well. That determination now belongs to the FDA in the first instance, on a record — Actavis’s own ANDA file — that was never before the district court or the D.C. Circuit.
What the Decision Establishes
- A first applicant’s Paragraph IV certifications are considered collectively, not certification-by-certification against a particular later filer, for purposes of failure-to-market forfeiture — and the statute supplies no mechanism for partial or relational forfeiture.
- Norwich’s section viii carve-out did not change the result because the court treated Actavis’s HE Paragraph IV certification as a qualifying certification for forfeiture purposes, and no triggering event had occurred as to that patent. The court expressly left open whether only “matching” certifications trigger the exclusivity provision in the first place.
- “Caused by” in the tentative-approval forfeiture exception requires but-for causation, not mere contribution — but the door remains open to an aggregation approach where multiple FDA requirement changes occur close together.
- Deference under Loper Bright does not extend to an agency’s reading of an ordinary causal term simply because the agency has technical expertise in the underlying subject matter.
- The court decided the legal standard, not the outcome. Whether Actavis’s failure to obtain timely tentative approval was actually caused, in the but-for sense, by the FDA’s 2017 guidance — or by some other requirement change, or not caused by any qualifying change at all — is a factual and record-dependent question the FDA must now answer in the first instance.
The broader lesson is that Hatch-Waxman forfeiture provisions cannot be read in isolation from the structure Congress built around generic exclusivity. The first forfeiture provision turns on the collective status of the first applicant’s certifications and treats the exclusivity period as indivisible. The second creates a narrow exception to an otherwise clear 30-month forfeiture rule, and that exception requires a causal connection between the FDA’s later change and the missed deadline — not merely a contribution to delay.
For first ANDA applicants, the practical point is straightforward: 180-day exclusivity is valuable, but it is not self-preserving. The statutory forfeiture provisions have different triggers, different exceptions, and, as Norwich shows, very different rules about causation and the scope of the inquiry.
Reference: Norwich vs Secretary HHS rifaximin Judgement
