A secondary medical-use patent can remain a material barrier to generic entry after the compound patent has expired. The challenge for generic pharmaceutical companies is not simply to identify these patents, but to determine whether they can be invalidated, avoided or managed through a product-specific launch strategy.
Recent proceedings involving semaglutide, PCSK9 inhibitors, Entresto and nilotinib illustrate four distinct questions: whether the claimed use is entitled to its priority date; what evidence establishes infringement of a second medical-use claim; how secondary-use patents interact with core patent and SPC protection; and whether a label carve-out adequately addresses infringement risk.
These questions require separate analyses. A successful validity challenge does not establish non-infringement of other patents, and omission of a patented use from the product information does not, by itself, resolve the infringement question.
1. The starting point: identify the precise therapeutic-use limitation
Article 54(5) EPC permits a known substance or composition to be patented for a specific therapeutic application, subject to the requirements of patentability. The protection is directed to the claimed medical use; it is not a renewed monopoly over the active ingredient for every therapeutic purpose.
For generic companies, the scope of the claims is therefore critical. A secondary medical-use patent may concern:
- a further therapeutic indication;
- a defined patient population or clinical characteristic;
- prevention or reduction of a specified clinical event;
- a dosage regimen or treatment sequence; or
- a particular method of administration.
These categories can produce different infringement questions. A claim directed to a defined therapeutic outcome cannot be analysed in the same way as one requiring a specific administration instruction.
The distinction also matters when considering a label carve-out. Removing an indication from the Summary of Product Characteristics (SmPC) may address one aspect of exposure, but it does not necessarily address claims directed to other features of use or resolve whether the product will be offered or supplied in circumstances that engage the patent.
A generic FTO assessment should consequently identify the essential limitations of each independent claim and map them against the proposed product, its intended use, the proposed labelling and the relevant commercial conduct. Patent validity, claim construction and infringement must remain distinct parts of that exercise.
2. Semaglutide: priority entitlement as a route to revocation
On 8 October 2026, the EPO Technical Board of Appeal revoked Novo Nordisk’s EP 3 448 416, concerning the use of semaglutide to reduce major adverse cardiovascular events (MACE) in patients with type 2 diabetes and high cardiovascular risk. The Opposition Division had previously maintained the patent in amended form, limiting the relevant cardiovascular outcome to non-fatal stroke.
The appeal turned on the relationship between the priority disclosure and the amended claims. The Board concluded that the amended claims were not entitled to the asserted priority because the priority applications disclosed treatment in the context of type 2 diabetes, whereas the amended claims extended to a broader therapeutic scope. The effective date consequently shifted to the 2017 filing date, bringing earlier clinical literature into consideration. The Board concluded that the relevant claims lacked novelty and revoked the patent.
The significance for generic challengers lies in the distinction between the wording of the claims and the technical teaching of the priority application.
A priority analysis should examine whether the claimed therapeutic use is directly and unambiguously derivable from the earlier application and whether the priority disclosure supports the subject matter subsequently pursued. The patient population, therapeutic effect and relationship between the claimed use and the original teaching may all be relevant.
T 0883/23 — loss of priority without loss of the patent
In T 0883/23 (26 September 2025), concerning a combination of liposomal irinotecan, oxaliplatin, 5-fluorouracil and leucovorin for first-line treatment of metastatic pancreatic cancer, the EPO Board of Appeal held that the claimed dose combination was not entitled to the earliest claimed priority. The priority application disclosed a dose-escalation/de-escalation study protocol but did not disclose the information on tolerability of the specific dose combination that was later presented in the patent. The claim therefore lost the earlier priority date. However, the patent was ultimately maintained in amended form: the Board found that inventive step was satisfied, and the intervening prior art did not defeat the claims on the facts before it.
The practical lesson is that a priority defect creates an opportunity to challenge a patent, but does not determine the outcome. The challenger must identify prior art falling within the intervening period and establish that it defeats the relevant claims under the applicable patentability requirements. In T 0883/23, that challenge did not succeed; in the semaglutide proceedings, the loss of priority exposed the claims to prior art that defeated novelty.
Neither decision establishes a general presumption against secondary-use patents. The priority entitlement of each patent must be assessed on its own disclosure and applicable legal requirements, and a lost priority date is the beginning of the inquiry, not its conclusion.
For generic companies, the practical lesson is to compare the claims with the complete priority disclosure, not merely with the wording of the priority claims. Where priority is lost, the generic company still carries the burden of identifying intervening prior art that defeats novelty or inventive step. That analysis should be undertaken alongside the other available grounds of challenge, including sufficiency and support assessed independently of priority.
3. Infringement: what must be established beyond the existence of a patented use?
A patent may be valid and still not be infringed by a particular generic product. The difficult question is often whether the product is offered or placed on the market in circumstances that engage the claimed therapeutic use.
The UPC’s Düsseldorf Local Division addressed this issue in Sanofi Biotechnology SAS and Regeneron Pharmaceuticals Inc. v Amgen Inc. and others, UPC_CFI_505/2024, decided on 13 May 2025. The dispute concerned EP 3 536 712, a second medical-use patent relating to the reduction of lipoprotein(a) [Lp(a)] levels in specified patients at cardiovascular risk, and Amgen’s Repatha product.
The court upheld the patent but found no infringement. It considered whether the product had been offered or placed on the market in a way that led, or might lead, to the claimed therapeutic use and whether the alleged infringer knew or should reasonably have known of that use. On the evidence before it, the court was not satisfied that infringement had been established.
An important feature of the case was the distinction between information reporting clinical trial outcomes and information directing use for a patented indication. The SmPC referred to reductions in Lp(a) observed in clinical studies, but the product was approved for LDL-cholesterol lowering. The court did not treat the reported trial outcome, by itself, as sufficient to establish that the product was being offered for the claimed Lp(a)-reduction use.
The decision is relevant to generic FTO assessments because it cautions against equating every reference to a clinical endpoint in product information with an indication for which a second medical-use patent has been granted.
It does not establish that SmPC content is irrelevant. The full product information, the nature of the claimed use, marketing conduct and the evidence concerning foreseeable use must be assessed together. The decision is also a first-instance judgment, not a final outcome on infringement applicable throughout Europe.
For generic companies, the distinction is commercially important: a valid second medical-use patent does not automatically cover every product containing the relevant active ingredient, but the absence of an express patented indication does not automatically establish non-infringement.
4. Entresto: core patent clearance and secondary-use clearance are separate questions
Entresto demonstrates why the expiry or invalidation of one patent does not resolve the entire patent position for a combination medicine.
The European patent landscape for sacubitril/valsartan includes the core combination patent and SPC protection, as well as separate patents directed to therapeutic uses—and these secondary rights are not a single layer but at least three, each gating a different slice of the patient population on a different expiry clock. An August 2034 family (EP 3 626 270, EP 4 321 157 and pending EP 4 512 397) covers valsartan-sacubitril for achieving certain clinical endpoints in HFrEF patients generally. A May 2036 family (EP 3 294 283, EP 4 212 152 and EP 4 570 314) covers the same combination in ACE-inhibitor- and ARB-treatment-naive HFrEF patients specifically. A February 2037 family (EP 3 411 021 and pending EP 4 609 912) covers use in certain pediatric patients specifically. All three families are under EPO opposition from multiple competitors, and their scope and status must be assessed separately from each other as well as from the core patent.
The divergent outcomes within this one portfolio illustrate why. In December 2025, the EPO Opposition Division maintained EP 3 626 270 with minor amendments—contrary to the EPO Technical Board of Appeal’s earlier (verbal, non-binding) opinion that the parent patent covering highly similar subject matter, EP 3 038 654, lacked inventive step, an opinion that led Novartis to withdraw that parent patent rather than contest it further. Two tribunals, one applicant, overlapping claims, and two different outcomes—within months of each other, in the same patent family. That is not an aberration; it is the same pattern the semaglutide and T 0883/23 decisions illustrate elsewhere in this piece, here visible inside a single originator’s own portfolio.
In Accord Healthcare Ltd v Novartis AG [2026] EWHC 2127 (Pat), the English High Court rejected Accord’s challenge to the validity of Novartis’s core combination patent and UK SPC. The court held that the SPC remained valid until January 2028 and that Accord’s proposed product would infringe it. The proceedings concerned only the core combination protection, not any of the three secondary-use families above. Separately, Stada has challenged the UK parts of the core patent, EP 3 626 270 and EP 4 321 157 on novelty, obviousness and insufficiency—and has also sought an Arrow-type declaration, a UK-specific remedy under which the court declares that a given use would have been obvious as of a stated date, providing a form of clearance that does not depend on any particular patent being found invalid. That stayed action (pending the Accord appeal) is a reminder that declaratory relief in this space is not limited to a conventional non-infringement or revocation action; where a patent family is still prosecuting or where claims may shift, an Arrow declaration can clear a path against subject matter rather than against a fixed claim set.
For generic companies, the strategic implication is not simply that another patent remains in force. It is that the entry analysis must distinguish the legal effect of each right—and, concretely, that a carve-out can be built around a specific family rather than the portfolio as a whole. Stada’s own product is reportedly adults-only, which would avoid the 2037 pediatric family entirely; and the 2036 treatment-naive family is generally considered avoidable by excluding that specific patient subgroup from the generic’s labelling. Avoiding the narrowest, most specifically-defined families is often more tractable than avoiding the core patent itself, precisely because they are drawn around a population rather than the molecule.
A successful challenge to the core patent or SPC may remove the principal constraint on entry, but it does not establish that the proposed generic can be launched for every therapeutic use or every patient population. Conversely, a successful challenge to one secondary-use family may remove an indication- or population-specific barrier without changing the expiry of the core protection or of the other families.
Where different indications and patient populations have different patent positions, generic planning should reflect that distinction. A single product-level launch date obscures the real picture: the realistic model is a staggered one, with different populations clearing at different times—potentially 2028 for the core SPC, 2034 for the general-use family, 2036 or 2037 for the narrower ones, absent successful challenge or carve-out. That conclusion requires a claim-specific, population-specific and jurisdiction-specific assessment; it should not be assumed merely because the patents concern different indications.
5. Label carve-outs: the nilotinib dispute leaves a material question unresolved
A label carve-out may allow a generic company to obtain approval for non-patented uses while omitting a patented indication or administration instruction from its product information, where the applicable regulatory framework permits it.
The strategy is not self-executing. A revised SmPC does not, by itself, determine whether the generic product infringes a second medical-use patent under the applicable law.
Novartis’s EP 2 501 384 concerns the administration of nilotinib dispersed in apple sauce for patients who cannot swallow the capsule whole. Accord sought a declaration of non-infringement before the UPC, arguing that its generic product information excluded the relevant administration instruction. Novartis disputed the sufficiency of Accord’s carve-out, arguing that amendments to the SmPC and patient information leaflet alone were insufficient to avoid infringement and that additional measures were required. The proceedings settled in March 2025, without a substantive UPC ruling on whether Accord’s carve-out avoided infringement. The dispute nevertheless highlights the potential limits of relying on product information alone to address second medical-use patent risk.
The case illustrates the litigation risk that remains when the patent proprietor and generic company disagree about the consequences of a carve-out. It does not establish the legal effectiveness of the particular carve-out.
The German position
German second medical-use case law, including the Düsseldorf Higher Regional Court’s decisions in the Östrogenblocker line, has placed emphasis on the circumstances in which the generic product is offered or supplied and the foreseeable use to which it may be put. The assessment may extend beyond the formal omission of a patented use from the product information to the generic company’s conduct and the prescribing or substitution practices known to it.
For generic companies, this means that a carve-out strategy should be assessed against the applicable national infringement test. The product information, promotional material, distribution arrangements and other relevant commercial conduct may all require review.
A label carve-out should not be treated as a universal safe harbour. Nor should a decision from one jurisdiction automatically be transposed to another, particularly where the applicable legal test differs or has not yet been considered by the relevant court.
6. Selecting the appropriate strategy
The appropriate response to a secondary medical-use patent depends on the claim limitations, the available evidence, the proposed product and the commercial importance of the patented use.
| Route | Principal question | What the assessment must establish |
| Validity challenge | Is there a sound basis to revoke or limit the patent? | Evidence supporting priority, novelty, inventive step, sufficiency, support or another applicable ground. |
| Non-infringement | Does the proposed product or its intended use fall outside the claims? | A reasoned claim-construction and infringement analysis based on the proposed product and relevant conduct. |
| Design-around | Can the claimed use or method be avoided? | A technically feasible and regulatorily acceptable alternative that falls outside the relevant claims. |
| Label carve-out | Can the patented indication or instruction be omitted? | Regulatory acceptability, commercial viability of the remaining uses and a jurisdiction-specific assessment of residual infringement risk. |
| Declaratory proceedings | Can uncertainty be resolved before launch? | Availability of the remedy, suitability of the forum and the scope of the issue the court can decide. In the UK, this can include an Arrow-type declaration that a given use would have been obvious as of a stated date—clearing a path against the subject matter even where the relevant claims are still prosecuting or may shift. |
| Alternative launch strategy | Is it commercially preferable to defer or modify the launch? | A comparison of the costs, timing and risks of litigation, redesign and delayed entry. |
These options are not mutually exclusive. A generic company may challenge validity while maintaining a non-infringement position, or pursue a carve-out while assessing whether declaratory proceedings would resolve material uncertainty.
The decision should be based on the strength of the evidence, the time needed to resolve the issue, the potential commercial return and the consequences of an adverse judgment. Not every secondary-use patent warrants a challenge; equally, a patent directed to a narrow use may be decisive if that use accounts for the commercial viability of the proposed product.
Conclusion
The recent European decisions show why secondary medical-use patents require a separate and technically grounded strategy within generic development and FTO planning.
Semaglutide illustrates how priority entitlement can determine the validity of a therapeutic-use claim. T 0883/23 shows that the same vulnerability does not always produce the same result: the patent lost priority but survived on inventive step, underscoring that a priority defect is an opening for challenge, not a verdict. Sanofi/Regeneron v Amgen demonstrates that validity and infringement can produce different outcomes. Entresto shows why core combination protection and secondary-use patents—often layered across multiple families with different expiry dates and different patient populations, and prone to inconsistent outcomes even within the same originator’s portfolio—must be assessed independently, while the nilotinib settlement leaves the effectiveness of the particular carve-out unresolved.
The central task for a generic company is to determine which rights actually constrain the proposed product, what evidence supports a validity or non-infringement position, and whether a design-around or carve-out can provide a defensible route to market.
The analysis should end not with a list of secondary patents, but with a reasoned decision on how each material claim will be addressed—by challenge, avoidance, carve-out, litigation or launch sequencing—in each relevant jurisdiction.
This article is for general information and does not constitute legal advice. Patent status, claim scope and infringement risk should be verified against current official records and assessed for the specific product and each relevant jurisdiction.
