The Paradox After Wyeth v. AstraZeneca

The Active Step That Both Saves and Sinks a Method-of-Treatment Claim

The active, patient-directed administration step can keep a treatment claim eligible under Section 101 and give a label something to induce. The same limitation is what loads the claim with a full-scope Section 112 enablement burden.

By Alexandros N. Nikolaidis • USPTO Reg. No. 85477 • Niki IP Services

Wyeth v. AstraZeneca, No. 2024-2325 (Fed. Cir. July 9, 2026), is not just another Federal Circuit enablement affirmance. It is a warning about the drafting mechanics of modern method-of-treatment claims. A jury found the asserted claims not invalid and awarded $107.5 million. The district court then granted JMOL of invalidity for lack of enablement, and the Federal Circuit—Judge Lourie writing for a unanimous panel—affirmed. The failure point was not that the patentee lacked FDA-grade clinical trials. It was that the claims crossed into patient administration while the specification largely remained at the bench.

That is the paradox. The active, patient-directed administration step is often the claim limitation that keeps a treatment claim on the right side of Section 101. In Vanda, an affirmative genotype-keyed dosing step helped transform a natural relationship into a patent-eligible application. But in Wyeth, similar patient-directed dosing language helped load the claims with a full-scope Section 112(a) burden: if the claim requires daily administration of a unit dosage to a patient to produce a therapeutic effect, the specification must teach the skilled artisan how to reach a real, administrable patient dose across the claimed scope without undue experimentation.

The active step is therefore a vise. It can save the claim from eligibility problems, give the label something concrete to induce, and make the invention look like a true treatment method. But it also becomes the hook for enablement. Draft to the clinic and you must enable the clinic. Draft only to the bench and you may not capture the commercial use.

The Eligibility Jaw: Vanda and INO

The Section 101 side of the vise begins with the difference between doing something to a patient and merely recognizing a natural relationship. In Vanda Pharmaceuticals v. West-Ward Pharmaceuticals, the claim did not stop at observing that CYP2D6 poor metabolizers respond differently to iloperidone. The claim required determining the patient’s genotype and then administering a different dose depending on that patient characteristic. The natural relationship mattered, but it was applied through a concrete dosing act. That affirmative act made the claim look like a treatment method rather than a diagnostic instruction. Vanda is the pro-eligibility pole: identify a patient characteristic, then administer a treatment in a way that changes because of that characteristic.

INO Therapeutics v. Praxair is the mirror image. There, the operative instruction for the at-risk patients was essentially to withhold inhaled nitric oxide treatment. For those patients, the claim did not require a different affirmative therapy; it instructed exclusion and allowed the body’s natural processes to proceed. Although INO is nonprecedential, it is important because it exposes the fragility of negative treatment limitations. If the claim’s inventive step is essentially “in case of serious side effects, stop taking the medication,” that is something a journeyman pharmacist can figure out without a patent—and a claim built on that kind of do-nothing instruction is far more vulnerable to being characterized as ineligible matter.

The operative drafting rule is simple: an affirmative act of administering something different to the patient based on the patient’s characteristics is Vanda-strong; an operative limitation that withholds treatment or instructs inaction in response to a natural response is INO-fragile.

Draft to the clinic and you must enable the clinic. Draft only to the bench and you may not capture the commercial use.

Praxair: The Active Step Is the Gate, Not the Guarantee

Praxair v. Mallinckrodt adds a second doctrinal pressure point. Although Praxair arose in the printed-matter and obviousness context, its practical lesson overlaps with Section 101. Informational content—recommendations, warnings, decisions—receives patentable weight only when it is functionally tied to a real-world act. A label instruction or patient-risk determination is not automatically enough. It must do something in the claimed method. But Praxair also shows that the active step is only the gate. Even claim language tied to an act can still fall to obviousness. Drafting an affirmative administration or discontinuation step may help the limitation count, but it does not make the claim patentable by itself. The active step gets the claim into the fight; it does not win the fight.

The Enablement Jaw: Teva, United Therapeutics, and Wyeth

The enablement side of the vise is best understood through a matched pair of 2026 decisions. Three months before Wyeth, in Teva Pharmaceuticals International GmbH v. Eli Lilly & Co. (Fed. Cir. Apr. 16, 2026), Judge Prost, writing for a unanimous panel, reversed a JMOL and reinstated a $177 million verdict, holding that when a patent claims a method of using a well-known genus of compounds—rather than claiming the compounds themselves—the written-description and enablement requirements of Section 112 are evaluated differently. Teva is the harbor. Wyeth is the jaw closing. Read together, the two decisions map the enablement line for method-of-treatment claims: what the claim recites about dosing the patient sets the burden.

United Therapeutics v. Liquidia marks the same harbor from the other direction. There, the Federal Circuit rejected the idea that every patient-specific safety or efficacy issue becomes a Section 112 problem merely because the claim is a treatment claim. Unless safety and efficacy requirements are incorporated into the claims, questions about which patients should receive the drug are generally left to the FDA and medical practitioners. A method-of-treatment claim does not fail enablement merely because some patient subset may not benefit or should not take the treatment.

Wyeth draws the line where that harbor ends. The asserted claims (U.S. Patents 10,603,314 and 10,596,162) were not merely broad research claims about EGFR inhibition. They required daily administration to a patient of a pharmaceutical composition comprising a “unit dosage” of an irreversible EGFR inhibitor. The construction of “unit dosage” mattered because the specification defined it as physically discrete units containing a predetermined quantity of active material calculated to produce the desired therapeutic effect. That language pulled the claim into the world of patient-administrable dosing.

The patents then had an enablement problem. The specification disclosed broad, general, and projected dosage ranges, but no working examples of any unit dosage calculated to achieve a therapeutic effect and suitable for daily administration in human patients. Worse, evidence showed that at least some compounds within the disclosed ranges—including HKI-272 and EKB-569—involved therapeutically effective levels above the maximum tolerated dose in humans. The problem was not the absence of FDA approval. The problem was the absence of guidance teaching which doses within the claimed and disclosed universe could actually be administered daily to a patient.

Wyeth’s lesson is not that a patentee must prove clinical safety and efficacy to satisfy Section 112. It is that when the claim itself requires patient administration of a therapeutic unit dosage, the specification must enable patient-dose selection without undue experimentation.

Not FDA-Grade Clinical Proof, but Patient-Dose Enablement

This is the article’s critical distinction. It would be too broad to say Wyeth converts enablement into FDA approval by another name. The Federal Circuit expressly avoided that move. Toxicity evidence mattered not because the claims contained an independent toxicity limitation, and not because every claimed dose had to satisfy FDA clinical safety standards. It mattered because the evidence showed the specification gave little guidance on which disclosed dosages, if any, were suitable for daily administration to patients across the functionally claimed class.

That is a narrower holding, but it is still powerful. In easy cases, Teva and United Therapeutics should remain a real harbor: a claim to a known drug, with conventional dose ranges and ordinary patient-selection issues, should not require the patentee to prove FDA-grade safety in the specification. But in hard cases—broad functional compound classes, projected ranges, no worked patient dosage, and toxic or above-MTD examples—Wyeth gives challengers a ready-made enablement theory.

The better formulation is this: Wyeth does not require clinical proof as such, but it does require patient-dose enablement when the claim is drafted to the clinic. If the claim says “administer daily to the patient” a unit dosage calculated to produce a therapeutic effect, the specification must do more than identify in vitro activity and gesture toward projected ranges. It must give the skilled artisan a non-speculative path to an administrable therapeutic dose across the claim’s scope. The farther the claim moves from molecule activity to patient regimen, the more the enablement meter runs.

The Enforcement Flank: Lundbeck and the Label

The same patient-safety layer also affects enforcement. H. Lundbeck A/S v. Lupin Ltd., 87 F.4th 1361 (Fed. Cir. 2023), sits on the Section 271 side of the same problem. Skinny-label practice often turns on whether the proposed generic label still instructs the patented use. If the claimed method depends on a genuinely new safety teaching that the FDA requires the label to carry, inducement may remain viable. If there is no such new safety discovery or required instruction—and the generic engages in no advertising or promotion of the infringing use—the carve-out may avoid inducement.

This is why Vanda and Lundbeck belong in the same conversation. In Vanda, label recommendations supported intent to induce because the label instructed the genotype-keyed dosing method. Lundbeck distinguishes cases where the label does not carry a new patient-safety teaching tied to the patented use. The enforcement value of a patient-directed treatment claim may therefore depend on the same feature that increases its enablement cost: a concrete patient instruction that must appear in the label. (The doctrine remains in motion: skinny-label inducement has continued to develop through 2026, including at the Supreme Court, but the core drafting tension Lundbeck frames is unchanged.)

The result is a three-axis drafting problem. Section 101 asks whether the claim applies a natural relationship through an affirmative treatment act. Section 112 asks whether the specification enables that treatment act across the claimed scope. Section 271 asks whether the label actually instructs the act in a way that can support inducement. The same words can help on one axis and hurt on another.

Six Drafting Questions After Wyeth

Before filing or prosecuting a method-of-treatment claim, Wyeth suggests a practical checklist.

•  Affirmative or negative operative step? Does the independent claim recite an affirmative, differentiated administration step tied to a patient characteristic, or does the operative limitation withhold or exclude treatment? The former is Vanda-strong; the latter is INO-fragile.

•  What is the dose anchor? If the claim requires administration to a patient to produce a therapeutic effect, does the specification disclose at least one worked dosage within the claimed scope—not merely a general or projected range?

•  How clean is the range? Do the disclosed or claimed ranges include toxic, lethal, or above-MTD values without guidance for selecting the therapeutically useful subrange? If so, the specification may be building the challenger’s undue-experimentation record.

•  Inside the harbor or outside it? If safety and efficacy are not claimed, ordinary patient-appropriateness issues should generally remain FDA and physician questions (Teva; United Therapeutics). But patient-unit-dosage language, therapeutic-effect language, and broad functional scope can move the case toward Wyeth.

•  Does every information step do real work? Determining, recommending, warning, and label-type limitations should be tied to an affirmative treatment act. Otherwise they risk being treated as printed matter, mental steps, or non-limiting information.

•  Is the enforcement value worth the enablement cost? A patient-safety instruction that must remain on the label may support inducement, but that same clinical specificity may increase the burden to enable the claimed regimen. Draft for both consequences at once.

Conclusion

The strategic throughline is not complicated: specificity is not free. A patient-directed active step can be the price of admission under Section 101 and the hook for inducement under Section 271. But once the claim is drafted as a patient regimen, the specification must enable a patient regimen. Wyeth is the Section 112 jaw of the vise closing.

For prosecutors, the lesson is to decide early what the claim is trying to be. If the value is in a real clinical dosing method, draft the specification to support that method with patient-dose guidance. If the application only supports bench activity, be careful before claiming a therapeutic regimen the disclosure cannot actually teach. The active step may save the claim—but it can also sink it.

About the Author: Alexandros Nikolaidis is a pharmaceutical chemist and registered patent agent (USPTO Reg. No. 84755) with over fifteen years of experience in patent development, prosecution, and portfolio strategy. Besides his website, he is also a contributor at IPIPWatchdog.