UPC's Düsseldorf Local Division gives guidance on infringement by equivalents

No settled test for the doctrine of equivalents currently exists in the Unified Patent Court (UPC). The Court of Appeal has yet to address the issue, and most first-instance equivalents decisions have been able to rule out infringement without needing a complete test. Until now, only one UPC local division (LD), The Hague, has done so.

In its decision in Wonderland Nurserygoods v Cybex and Ors, the Düsseldorf LD adopted The Hague LD's test and gave guidance on its application. The court indicated that a party seeking to show technical equivalence must demonstrate the function of each variant feature in achieving the objective of the patent and explain why the variant performs essentially the same function. It also stated that where several elements need to be rearranged, such that there is a "complete" or "fundamental redesign", it will generally not be obvious how to apply the equivalent element.

LDs are not bound to follow each other's decisions and therefore the Düsseldorf LD's adoption of The Hague LD's approach could be seen as a move towards a clearer picture on infringement by equivalents at the UPC. However, the test has not been without criticism from other LDs, principally for being based on Dutch national law principles without further explanation. Full clarity on the UPC's unified approach to infringement by equivalents will require a ruling from the Court of Appeal.

The Hague's 'coherent' and 'suitable' test

The case before the Düsseldorf LD related to a patent for a swivel-locking device for the wheels of a child's stroller or buggy. The claimed device allows the stroller's wheels to switch between being able to rotate freely around a stem, enabling the stroller to be steered easily in any direction, and being locked into place, preventing the wheels from rotating around their stem.

Wonderland alleged that the defendants' products infringed under the doctrine of equivalents. It argued that four features of the claim were infringed by equivalents, with the remaining six features infringing literally.

The court noted that several UPC LDs, including Brussels, Düsseldorf, The Hague, Hamburg, Mannheim and Paris, had considered equivalents, but most had found no equivalent effect or function and so had refrained from examining further steps. The exception was The Hague, which had set out a four-step test in Plant-e v Arkyne, which it applied again in Washtower v BEGA.

Under the Plant-e test, a variant is equivalent if four questions are each answered in the affirmative:

1. Technical equivalence: does the variation solve (essentially) the same problem that the patented invention solves and perform (essentially) the same function in this context?
2. Is extending the protection of the claim to the equivalent proportionate to a fair protection for the patentee? This needs to be assessed in view of the patentee’s contribution to the art and taking into account the question whether it is obvious to the skilled person from the patent publication how to apply the equivalent element (at the time of infringement).
3. Reasonable legal certainty for third parties: does the skilled person understand from the patent that the scope of the invention is broader than what is claimed literally?
4. Is the allegedly infringing product novel and inventive over the prior art?

The Düsseldorf LD applied these criteria. It considered that they "form a coherent whole" and, as such, were "suitable" for examining equivalence. It stated that, at least in the present case, there was no indication that applying a different standard would produce a different outcome.

No infringement by equivalence

Question 1

Addressing question 1 of the test, the court concluded that there was no technical equivalence.

Wonderland had argued that the technical effect of the relevant features should be determined in the context of the invention as a whole. It stated that the objective of the invention was to ensure the functionality of the swivel lock while providing a sufficiently stable connection between the rotatable and the non-rotatable element of the device. It argued that it was technically irrelevant that the allegedly infringing variants involved in locking the swivel and preventing movement were arranged differently from the claim features.

The court rejected this argument. It said that while the claim must be considered as a whole, it is necessary to determine the function of each substituted feature in relation to it. That allows a comparison between the technical functions of the claimed and substitute means. It is "not sufficient to focus solely on the objective task of the patent".

Instead, "what matters is the function of each exchanged features [sic] in achieving this task". In this case, the court noted that Wonderland had ignored the fact that the patent specifically identified some features as rotatable and some as not, and had ultimately reduced the entire invention to a secured connection between rotatable and non-rotatable elements.

It had not sufficiently explained how this was to be achieved according to the patent’s teaching, which the court considered was aimed at achieving the secured connection through a "particular arrangement of components".

Despite having ruled out infringement under question 1, the court went on to address the next two questions. It also answered these in the negative.

Question 2

On the question of fair protection for the patentee, the court rejected Wonderland's arguments that it was obvious "to simply reverse" the arrangement of components. The court considered that to reach the alleged infringement, several parts would have to be rearranged simultaneously. This would be a "fundamental redesign" that would not have been obvious to the skilled person.

It was persuaded by the defendants' argument that it would be "rather far-fetched" to relocate two moving parts from the wheel-bearing assembly to the body of the stroller, losing the invention's advantage of having all of the complex mechanical elements in a single, replaceable unit.

Wonderland had failed to explain the patent's contribution to the art. The court left this issue open, but noted that nothing could be inferred in Wonderland's favour in that regard.

Question 3

In relation to providing reasonable legal certainty to third parties, Wonderland's arguments had failed to consider that the claim set out "specific requirements regarding the arrangement of the components". The court considered that it was "precisely this arrangement" that achieved the desired secured connection between rotatable and non-rotatable elements. The skilled person would not understand that "such significant modifications in almost all parts" would be covered by the patent.

Having answered questions 1 to 3 in favour of the defendants, the court did not consider it necessary to go on to determine whether the infringing products were novel and inventive over the prior art under question 4.

Patent valid over 14 starting points

It was not all bad news for Wonderland. The court applied the established case law of the Court of Appeal to reject the defendants' allegations of added matter, lack of novelty and obviousness.

In their obviousness attacks, the defendants had relied on no fewer than 14 supposedly realistic starting points, citing a wide range of documents as potential combinations.

Following the approach to inventive step set out by the Court of Appeal in Meril v Edwards and Amgen v Sanofi, there can be more than one realistic starting point. However, the Düsseldorf LD appeared to find the volume of attacks, which in some cases were not fully pleaded, frustrating.

The court noted that under the approach in Meril and Amgen, several steps must be taken to demonstrate a lack of inventive step. Here, however, the defendants had failed properly to set out in the written pleadings whether the skilled person, starting from a specific realistic starting point and wishing to solve an objective problem, would (not only could) have arrived at the claimed invention. Instead, Wonderland had applied the European Patent Office's problem-solution approach. Moreover, some combinations were "not explained at all, but…simply mentioned in passing".

The court invited Wonderland to select one or more combinations and explain them in more detail at the oral hearing, applying the Court of Appeal's case law. However, not even these more-fully-developed attacks were successful.

Osborne Clarke comment

The Plant-e test has previously been criticised by some other LDs as being based on national case law, rather than seeking a harmonised approach across all UPC member states. For example, in Dish v Aylo, the Mannheim LD stated that the test "applies principles developed in Dutch case law without further justification, but with the agreement of the parties". It suggested that the UPC would "have to develop its own doctrine of equivalence…possibly with recourse to the traditions of the UPC member states" and noted the importance of a uniform determination of patent scope across all UPC states. The Paris LD made similar points in NJ Diffusion v Gisela.

The Plant-e test is still not binding on the UPC's LDs, which await a ruling from the Court of Appeal. The Düsseldorf LD appeared to accept that there might be cases where arguments could be made for a different test. However, it remains the only fully articulated test for equivalence in the UPC. The fact that it has now been adopted as a "coherent" and "suitable" test by another important LD (last year Düsseldorf had the second highest caseload of any LD, after Munich) in a different contracting state suggests that the UPC may now be starting to converge on this test.

The Düsseldorf LD's decision also gives useful guidance on how to apply the test. In particular, it suggests that it is necessary to carry out a careful analysis of each variant in turn to explain its alleged technical equivalence and the role it plays in relation to the claim, rather than focussing on a broad-brush overall technical objective.

It also suggests that, where multiple elements have been varied simultaneously so as to amount to a complete redesign, it would not be obvious to the skilled person how to apply the equivalent. In that case, extending the scope of protection that far would not be proportionate to fair protection for the patent. Such significant changes may also mean that extending protection would violate the need to provide reasonable legal certainty to third parties.