The UPC at three: substantive case law trends

After a long and uncertain gestation, the Unified Patent Court (UPC) opened its doors on June 1 2023. Three years on, it is possible to take stock. The system now binds 18 contracting member states – Romania being the last to ratify, in 2024 – covering a territory of more than 270 million people and over 80% of the EU’s GDP. Spain, Poland, and Croatia have stayed outside; Ireland, Greece, and others have signed but not yet ratified.

The court has already produced a substantial and technically dense body of case law. A central concern in designing the UPC was to prevent the new system from being captured by the national legal traditions of its participating states. Decisions at first instance and on appeal show a deliberate effort to build an autonomous interpretative method, suited to this new supranational order. This article presents four substantive areas to illustrate the point.

Claim construction was among the most keenly awaited contributions of the UPC. In NanoString v 10x Genomics, the Court of Appeal restated the principles flowing from Article 69 of the European Patent Convention (EPC) and its Protocol:

• Claims are read from the perspective of the skilled person and are the decisive basis for the scope of protection, not a mere starting point;

• The description and drawings are always used as explanatory aids, and not only to resolve ambiguity, but they cannot be used to reconstruct a subject matter different from that which objectively follows from the claims; and

• The same interpretative criteria apply to validity and infringement.

This last point, settled in German case law since around 2000, was far from obvious at EPO level, whose boards of appeal never had to assess infringement. Because the UPC rules on validity and infringement together, it is well placed to secure the coherence the European patent system requires.

In Alexion v Samsung Bioepis, the Court of Appeal added that a drafting error or inaccuracy in a claim may be corrected by interpretation only where both the existence of the error and its precise correction are sufficiently certain to the skilled person. Creative or speculative reconstruction is not allowed: a claim cannot, in substance, be ‘rewritten’ after grant.

More unsettled is whether the prosecution history, the ‘file wrapper’, may inform claim interpretation. National approaches diverge sharply. German courts (e.g., Kunststoffrohrteil) confine the analysis to the granted patent, admitting only a narrow good-faith exception, and France takes a comparable line. The Netherlands stands apart, giving the file a more substantive role: statements made during grant may bind the proprietor and bar both a broad reading of the claims and a finding of infringement by equivalence, and any third party may invoke the file wrapper in its favour.

The divergence tracks the competing interests at stake. In favour of estoppel lies the protection of competitors and the prevention of bad faith or abuse of rights: a proprietor should not be able to recover in court, by way of construction, what it gave up during examination. On this view, Article 69 of the EPC does not lay down an exhaustive list of admissible interpretative materials. It secures the primacy of the claims rather than prohibiting recourse to other sources, so the file may, in principle, be drawn on to elucidate the technical scope of protection.

Against estoppel stands legal certainty: construction should rest on the text of the patent, not on a dossier that competitors would have to study in full and that Article 69 does not mention. The file is, moreover, often incomplete: telephone or in-person exchanges with the examiner are common strategic tools and may mislead if treated as the full expression of the applicant’s position.

The UPC’s position is still forming. In SES-imagotag v Hanshow, the Munich Local Division accepted that original claim versions and amendments made during grant could serve as interpretative aids; in SodaStream v Aarke, the Düsseldorf Local Division refused, holding that the grant procedure is not among the aids listed in Article 69 of the EPC. In Alexion, the Court of Appeal suggested a middle path: statements made before the EPO may indicate how the skilled person would have understood the claim at the filing date. For now, the court shows a cautious and limited openness rather than any systematic role for the file wrapper.

Given the heterogeneity of national solutions, equivalence was always going to attract attention. In Plant-e v Arkyne, the Hague Local Division set out a two-stage analysis: literal infringement first, equivalence only if that fails, and a four-part test drawing on German thinking (equivalent technical effect, or Gleichwirkung, and the Formstein defence) and on the post-Actavis English emphasis on balancing fair protection with legal certainty.

Its authority has since been relativised, not least because the court acknowledged that the test reflected the parties’ submissions. Later decisions – Brussels in OrthoApnea, and the Mannheim and Paris local divisions – have refined matters, with Mannheim drawing a notable distinction: for acts governed by national law, the UPC applies the relevant national doctrine, whereas for acts governed by the Agreement on a Unified Patent Court (UPCA), it must develop an autonomous doctrine drawn from member states’ traditions. Two principles are consolidating: the two-stage structure, and functional equivalence as the common denominator across national traditions. Harmonisation is still at an early stage, but the attempt to distil shared principles is welcome.

In Kaldewei v Bette, the Düsseldorf Local Division held that Article 28 of the UPCA does not create an autonomous ‘European prior use right’. A prior user must establish the right separately for each contracting state, under that state’s national law. The decision has been criticised for perpetuating the very fragmentation the UPC was meant to overcome, though it reflects a deliberate legislative choice: Article 28 essentially reproduces the 1989 Community Patent Convention, to leave certain matters to national law.

These examples are only a sample of a far larger body of case law, but they show the value of a specialised, multinational bench able to condense knowledge from several legal traditions before settling on a path. As a court common to its contracting states, the UPC is well placed to drive the gradual convergence of interpretative criteria across the European patent space, with real gains in legal certainty and predictability.

Three years in, the verdict is positive, even if zones of uncertainty naturally remain. We will keep following UPC decisions closely.