UPC Central Division Revokes Patent Covering Covid-19 Treatment Remdesivir

The UPC Central Division in Milan has revoked European Patent 3854403, which claimed Remdesivir for use in treating SARS-CoV-2 (Covid-19) infection. The decision (UPC_CFI_552/2025, decision of 4 May 2026 –Gilead v Academy of Military Medical Sciences, see here) provides valuable insights for parties on how the UPC assesses inventive step, particularly in the pharmaceutical sector. Moreover, the Central Division appears to have criticised the EPO’s problem-solution approach for introducing an element of hindsight.

The patent proprietor was the Academy of Military Medical Sciences (AMMS). On the date of grant of the unitary patent, Gilead Sciences, Inc. brought a revocation action at the UPC Central division Milan, asserting that the patent was invalid due to a lack of inventive step and insufficiency.

Inventive step in Europe

The EPO has been applying the problem-solution approach for decades. The problem-solution approach is divided into three stages:

• the closest prior art is first determined;
• technical differences with the closest prior art are determined, and an objective technical problem is then established based on these differences;
• finally, the problem-solution approach asks whether the solution to the objective technical problem would have been obvious to the skilled person.

Notably, the EPO’s approach starts with the determination of the closest prior art, which is usually a single prior art document, and then characterizes the objective technical problem based on the differences between this document and the claimed invention.

In November 2025, the Court of Appeal of the UPC issued two decisions establishing its own test for the assessment of inventive step, which is now widely described as “the holistic approach” (Amgen v Sanofi and Regeneron, which we have previously discussed here, and Edwards v Meril). The holistic approach is structured differently from the EPO’s problem-solution approach:

• the first step is to establish the objective of the invention (objective problem) from the perspective of the skilled person;
• this is followed by asking whether, starting from a realistic starting point in the relevant field of technology and wishing to solve the objective problem, would (not could) the skilled person have arrived at the claimed solution?

Unlike the problem-solution approach, the holistic approach begins by establishing what the objective of the invention is. This is on the basis of the entire patent, and is not framed from the perspective of the prior art.

The UPC Court of Appeal emphasized that the two approaches should, and generally do, lead to the same end result. However, the present decision highlights that part of the rationale for the UPC’s new test is to avoid hindsight.

The decision explains that the EPO’s approach of basing the objective technical problem on distinguishing features of the claim over the closest prior art carries a risk of hindsight, because knowledge of the solution itself may be introduced into the objective technical problem. By contrast, the UPC’s approach establishes the technical problem on the basis of the patent, i.e., it evaluates its solution based on the description in conjunction with the drawings and the claims (Art. 69 EPC). The contribution of the patented invention over the prior art is only assessed after this has taken place. Therefore, in principle, this can reduce the risk of hindsight knowledge of the claimed invention unallowably influencing how the technical problem is framed.

Formulating the technical problem

Gilead argued for a narrow formulation of the technical problem: merely testing Remdesivir against the new coronavirus strain. The court, however, relying on Amgen v Sanofi and Regeneron, held that the technical problem must be defined broadly in light of the invention’s contribution to the art. Since the patent refers to therapeutic efficacy, the court framed the problem as the identification of an effective antiviral substance against a novel virus – not simply as testing Remdesivir on a new virus. This illustrates that the UPC’s approach focuses on the contribution of the invention as a whole, rather than distinct features within a claim.

The skilled person

Both steps of the holistic approach are framed around the skilled person, so the identity of the skilled person thus impacts both (i) how the technical problem is formulated and (ii) whether the claimed subject matter would ultimately have been obvious.

During the proceedings, the court emphasized that the skilled person is fictitious, and represents an average level of knowledge in a specific technical field. The skilled person should not be identified as any real person working in the relevant field, or for any real pharmaceutical company. Thus, they are not required to possess (or disregard) a distinct affiliation, so differentiating characteristics like connections to a specific company (and compounds developed by that company) are not admissible. In addition, the court noted that the skilled person’s co-location in a company or any personal affiliation does not add or diminish anything to their knowledge at the priority date. Therefore, when relying on experts, it is important not to attempt to frame the expert as being the skilled person, but to instead highlight why their knowledge and reasoning reflects that of the skilled person.

This provides useful guidance for parties moving forward in how the skilled person should be characterized, and what may or may not be possible to successfully argue forms part of their common general knowledge. This further guidance should assist parties in assessing their case prior to litigation, and enable submissions to be tailored to UPC practice.

A hierarchy within the state of the art?

At the EPO, very old documents may be disregarded as irrelevant, if the state of the art has moved on, but there is generally no “hierarchy” of prior art based solely on the age of the document.

However, in this decision, the court indicated that the skilled person prioritizes the most recent and detailed sources from the state of the art, because “scientific progress requires ongoing reassessment of previous knowledge; thus, experts select up-to-date and comprehensive documents”.

This provides useful insight for preparing and defending against invalidity attacks. Attacks based on older documents could be weaker than those based on publications shortly before the priority date of the patent. However, this is not to say that older documents are not valuable, simply that parties may need to provide further reasoning to establish why the skilled person would have turned to them (especially if this means disregarding a more recent publication).

Not only did the court look at the relative age of the document, but also the source. AMMS attempted to undermine the prior art identified by Gilead, pointing out that such publications were “promotional” and “commercial in nature”, rather than technical.

This position was decisively dismissed by the court: scientific opinion becomes part of the state of the art when published, unless it is later refuted. The court highlighted that any researcher who publishes unsubstantiated scientific opinions with promotional or commercial interests runs the risk of being deemed unreliable and losing credibility in the field. The court’s final comment on this matter was to note that dismissing a publication as lacking credibility simply because the research had been funded by a pharmaceutical company went against the very logic of the system.

The court also highlighted that citations relied on by AMMS did not contradict those Gilead relied on, and simply reflected a more cautious approach. This highlights a difference between publications that may have been sponsored by industry – which remain credible and relevant for the assessment of inventive step – and the possibility of subsequent publications disproving earlier research as a field develops – which are likely to result in the earlier publication no longer being considered accurate or a realistic starting point for the assessment of inventive step.

Reasonable expectation of success

At the EPO, claims can be inventive if the skilled person had no “reasonable expectation of success” at the outset. A reasonable expectation of success is based on a reasoned prediction based on scientific facts; this is distinguished from an understandable “hope to succeed”. 

AMMS argued that the choice of Remdesivir represented a significant step forward at the priority date, and that the skilled person would not have had a “reasonable expectation of success”. This was based on reasoning that mutations in viruses may lead to functional differences, and that even a single point mutation in the relevant region may alter drug sensitivity.

The court disagreed and held that simply suggesting unexpected outcomes was not adequate. The decision sets out that the skilled person only questions information when documented prejudice exists in the literature, and that the party must highlight proven flaws if they wish to take this position. The court noted that risk and doubt are part of scientific progress, and that the skilled person would not reject a proposed solution because of subjective concerns about possible failure.

On its face, the statement that the skilled person does not have a fear of failure may appear to depart from the reasoning established at the EPO, that in the field of biotechnology, the skilled person does not take risks. However, this court here is highlighting circumstances where there is no certainty of success but nonetheless a pointer towards an outcome that may be successful. The decision sets out that there is a reasonable expectation of success when data or experiences indicate that the tested solution can yield a positive result, even when there is uncertainty arising.

In this case, AMMS pointed to statements that expressed hesitation, whilst Gilead relied on prior art that provided technical reasoning and a clear pointer towards the assessment of Remdesivir in treating Covid-19. Therefore, technical reasoning seems to be required to undermine prior art, rather than just cautionary statements in the art expressing that it may not work.

As a result, the unitary patent has been revoked in this specific case. It remains to be seen whether AMMS will appeal the decision.

Concluding remarks

The decision provides clear guidance that will assist parties in developing inventive step positions before the UPC moving forward. It illustrates in a particularly clear manner how the UPC’s “holistic approach” to the assessment of inventive step is beginning to take shape in practice. The UPC places its focus on formulating the technical problem from the application as a whole, rather than focusing on those features that distinguish the claim from a pre-selected closest prior art document.

For parties litigating before the UPC, the ruling offers several practical takeaways:

• the skilled person is fictitious and possesses an average level of knowledge, and any suggested company affiliations are not relevant;
• the prior art should be selected and positioned in view of its technical relevance, and the age of the document may be a more relevant factor than in parallel EPO proceedings;
• countering an argument that there is a reasonable expectation of success will require substantiated technical reasoning, rather than simply pointing to general uncertainty or caution.

Finally, the case underscores the UPC’s procedural advantage in terms of speed. Gilead has also opposed the patent at the EPO, and the opposition division has set a deadline of 30 June 2026 to respond to the opposition. The fact that the UPC’s revocation was obtained at first instance well before the parallel EPO opposition has progressed highlights the UPC’s growing importance as a strategic, speedy forum for validity attacks.

A first instance decision has been reached at the UPC before the proprietor has even responded to the opposition filed at the EPO. This is in part because Gilead opposed the patent and brought revocation proceedings immediately following grant, rather than waiting until the end of the EPO’s nine-month opposition period. Therefore, although a successful opposition results in broader revocation, the speed of the UPC makes it a valuable asset for litigants.

Overall, the decision signals a maturing and increasingly predictable framework for assessing inventive step at the UPC, which parties should factor into their European litigation strategies going forward.