G2/21 And Post-filed Data: Three Years Of Board Of Appeal Case Law

It has now been three years since the Enlarged Board issued their decision G2/21 regarding the conditions under which post-filed data can be taken into account when considering inventive step. In this briefing, we take a look at the case law that has emerged from the Boards of Appeal since G2/21, to identify common themes and ongoing areas of uncertainty.

Legal background

The questions referred to the Enlarged Board in G2/21 related to the conditions under which post-filed data can be taken into account when assessing inventive step. There were three lines of case law at the time:

• “ab initio plausibility”, where post-filed data evidencing a technical effect can be considered only if there is some technical data in the application as filed to support the presence of that effect;
• “ab initio implausibility”, where post-filed data should be taken into account if the statement of efficacy in the application as filed is not prima facie implausible; and
• “no plausibility”, where the concept of plausibility is rejected entirely.

In G2/21, the Enlarged Board declined to pick one of the above options. Rather, it held that an applicant/proprietor should be able to rely upon a technical effect for inventive step if “the skilled person, having the common general knowledge in mind, and based on the application as originally filed, would derive said effect as being encompassed by the technical teaching and embodied by the same originally disclosed invention.”¹

This decision has not provided the level of legal certainty for which practitioners were hoping. Here, we will look at some of the interpretations adopted by the Boards of Appeal in the cases following G2/21.

T116/18 and T1989/19: The liberal approach

We begin with T116/18, which is the further decision from the Board which referred the issue to the Enlarged Board in G2/21. The patent related to an insecticidal combination, and the application as filed included a general statement that combinations of the invention could have synergistic insecticidal activity, as well as two examples showing synergistic insecticidal activity of the claimed combination against the insect species Spodoptera litura and Plutella xylostella. During appeal proceedings, the patentee accepted that the original synergy data did not support the full scope of the claims, and attempted to introduce post-filed data demonstrating that all of the claimed combinations had a synergistic activity against a further specific insect species: Chilo suppressalis.

The Board then considered how the test set out in G2/21 should be applied. First, they noted that the test has two requirements, which must be satisfied cumulatively for the applicant to be able to rely on the purported technical effect: namely, that the skilled person understand the effect as being (i) encompassed by the technical teaching; and (ii) embodied by the same originally disclosed invention. The Board then went on to consider how these two requirements should be addressed.

When considering whether or not a technical effect is encompassed by the technical teaching, the Board in T116/18 noted that the technical effect and the claimed subject-matter only need to be conceptually comprised by the broadest technical teaching of the application as filed.² It was explicitly noted that the technical effect need not be literally disclosed. When considering whether or not the technical effect is embodied by the same originally disclosed invention, the Board held that this would be met unless there is “legitimate reason” to doubt that the effect could be obtained by the claimed subject-matter.³ 

The opponent then sought to apply a “squeeze” argument, noting that if the synergistic interaction at issue was a surprising outcome, there must be a “legitimate reason to doubt” that the effect would arise, and the second requirement of the G2/21 test was not met. However, the Board in T116/18 robustly rejected this argument, and noted that the “legitimate reason to doubt” must be more than the finding that the effect is surprising⁴. This second test, as articulated in the decision, therefore seems similar to the old “ab initio implausibility” standard. The Board noted, for example, that there might be a legitimate reason to doubt that the specific purported technical effect can be achieved if this effect is “irreconcilable” with the common general knowledge. This has echoes of the previous strand of case law where the post-filed data was to be taken into account so long as any statement alleging the technical effect in the application as filed was not implausible.

This seems to be one of the most liberal approaches to post-filed data that has been adopted by the Boards to date. It is notable that the EPO Guidelines reference this decision at G-VII-5.2, and first instance departments do indeed seem to be taking a liberal approach to the admission of post-filed data.

T116/18 is also not the only case where a liberal approach has been taken. The 3.3.02 Board which handed down the T116/18 decision took a similar approach in T1989/19, where the patentee was able to rely on post-filed data to demonstrate improved storage stability for an inhalable composition of tiotropium bromide. Although there was no explicit reference to storage stability in the application as filed, the Board considered that storage stability was an inherent requirement of inhalable compositions and was therefore encompassed by the original technical teaching. It further held that there was no reason to doubt that the technical effect could be achieved by the claimed composition, and so allowed the patentee to rely on the post-filed data.

T681/21 and T1994/22: A stricter approach

Not all Boards have taken the same liberal approach. T681/21 related to fabric softening compositions including a silicone and a cationic polysaccharide polymer (CPP) that gave a synergistic softening effect. Although CPP was listed in the patent as a preferred polymer, no reasoning was provided for this preference. The Board held that the technical effect of synergistic softening could not be relied upon because there was no suggestion in the patent that a synergistic effect would arise from, specifically, CPP with a silicone, noting that on this occasion the alleged synergistic effect would not even have been considered by the skilled person as being encompassed by the technical teaching of the application as filed.⁵

This is a notably stricter approach than the threshold for the first requirement set out in T116/18, where the Board explicitly stated that there was no need for the patent to associate the specific technical effect with the specific claimed subject-matter, and that it was enough that the skilled person would recognise said effect as being necessarily relevant to the claimed subject-matter.

The T681/21 Board, however, took a different approach when considering an auxiliary request which further specified the silicone as an anionic silicone (which was also a preferred component of the composition). To defend this request, the patentee sought to rely on a different data set which compared overall softness of anionic silicone plus CPP with anionic silicone with a different cationic polymer. Thus, the new data set no longer referenced synergy, but instead simply referenced the quantum of the originally disclosed effect. The Board held that this new data set could be used when assessing inventive step. It emphasised the preferred nature of both the CPP and anionic silicone as being decisive in reaching this decision.⁶

This second finding from the T681/21 Board is much easier to reconcile with the reasoning of T116/18, although the Board’s reasoning as to why it should be legitimate to evidence an improved softness with this specific combination while refusing the data showing a synergistic interaction is not entirely clear.

T1994/22 was another case where a strict approach was taken. It concerned a patent which described three specific polymorphic forms of Selexipag in equal terms (Forms I, II and III), and referred generally to the high quality of pharmaceutical products containing such polymorphs. The Board said that this general description was not good enough to justify the admission of data evidencing, specifically, improved photostability of Form II over Forms I and III. Reason 1.3 notes that “if such a sweeping statement was sufficient, a reference to high quality would be sufficient to invoke whatever technical effect as being encompassed by an application as filed in the sense of G2/21. This would render the first criteria of order no. 2 of G2/21 meaningless”. This seems somewhat at odds with the broad approach set out in T116/18, requiring only that the technical effect be “conceptually comprised” by the teaching of the application: photostability is a property routinely considered in polymorph screening.

T314/20 & T1529/19: A discrepant approach?

It already seems clear that the order of G2/21 has left quite a lot of scope for Boards of Appeal to interpret the language in different ways, and that the approach taken by different Boards is not always wholly consistent. One particular pair of cases where this is apparent is T314/20 and T1525/19.

Both cases concerned combination therapies in which SGLT2 inhibitors were used with DPP IV inhibitors for treating metabolic disorders by increasing GLP-1. In both cases, the relevant patent was for empagliflozin (SGLT2 inhibitor) plus linagliptin (DPP IV inhibitor), limited in each case from broader claims in the application as filed.⁷ T1525/19 further specified specific doses (to assist with achieving novelty over the patent underlying T314/20, which represented novelty only prior art against the patent underlying T1525/19). In both cases, the post-filed data involved synergy data.

Reason 7.3.5 of T1525/19 notes that the “overadditive activity” of the two drugs “… does not change the nature of the effect assigned to the combination of the invention. It merely relates to a difference in intensity, which becomes relevant when quantification is necessary with the closest prior art”. Thus, the synergy data were admitted. However, reason 6.14 of T314/20 noted that the application as filed did not suggest that the claimed combination of empagliflozin plus linagliptin would give better activity than combinations of empagliflozin with other DPP IV inhibitors that were presented as equally preferred in the application as filed, and so excluded the synergy data.

The facts of the two cases do differ: in T1525/19 the data in the application was limited to the empagliflozin / linagliptin combination, whereas the data in T314/20 related to combinations of empagliflozin with a range of DPP IV inhibitors (and did not clearly suggest a preference for any particular DPP IV inhibitor in such a combination). In addition the two cases were heard by different boards (3.3.04 for T314/20, and 3.3.07 for T1525/19). Nonetheless, despite all that it is hard to reconcile the different reasoning presented by each board.

Post-filed data which quantifies a technical effect & compares to prior art

Historically it has always been possible, following a plausible disclosure of efficacy, to adduce evidence showing an improvement over the closest prior art. However, the reasoning above in T314/20 seems to suggest that it might be important to “mention a comparison” to other active compounds in order for post-filed comparative data showing a technical advantage over the closest prior art to be admissible. If that is the case, that would represent a profound shift in EPO practice.

T314/20 is not the only example of this reasoning. In T1994/22 above, the Board noted that it was relevant that there was no suggestion that Form II was better than the other forms in the application, and that there was a difference between demonstrating an improvement over an equally preferred portion of the disclosed subject-matter, compared to demonstrating an improvement over the prior art.⁸ Another case taking a similar approach is T258/21, which concerned a claim to the use of clevidipine for treating ischemic stroke. The applicant provided post-filed comparative data showing that clevidipine was clinically superior to other drugs in the same class (in particular nicardipine). The T258/21 Board excluded the post-filed data on the basis (inter alia) that the improved clinical efficacy of clevidipine, as compared with nicardipine “was neither contemplated nor even suggested in the original application“.

There is, however, also some precedent which takes the opposite view. T840/22 concerned a multilayer coated metal substrate where corrosion resistance was the property at issue for inventive step. The application as filed presented two structural options as equally preferred, but the granted claims were limited to one of those options. Reason 15.2 of that decision notes that “… it must be possible – contrary to the views of the appellant and the opposition division – for the respondent to rely on an improvement in the derivable effect, at least in the case of an improvement over a piece of prior art such as D2 in the present case”. In fact, the Board expressly notes that this scenario occurs regularly in life sciences / chemical patents, “…where claimed subject-matter must be limited at the expense of subject-matter disclosed in the application as filed as part of the invention, because the effect relied on for inventive step over the closest prior art is not achieved across the entire breadth of the claimed subject-matter.” The Board in T1989/19 took a similar approach.⁹

Findings of national courts

It is worth noting that the approach taken on this issue varies across national European courts, which adds further uncertainty to the European body of case law as a whole.

For example, the UK courts take the ab initio plausibility approach to both medical use and compound per se claims¹⁰, whilst Spanish courts have adopted the “ab initio implausibility” test¹¹. The Swiss courts have rejected “ab initio implausibility” but have noted that the technical effect should be “obvious to a person skilled in the art”¹². The Dutch and Norwegian courts have not expressly adopted or rejected the plausibility thresholds but have noted, respectively, that it is important that “a purely speculative patent for an invention that is only made afterwards is prevented”¹³, and that it should be “credible – i.e. plausible – to the expert team that this effect would be achieved”.

Summary & Practice points

It goes without saying that every case is different and that there will be nuances between the facts at issue on each case. However, the reality is that, following G2/21, we do not appear to have a consistent approach from the boards of appeal in determining whether and under what circumstances post-filed data can be taken into account when assessing inventive step. On this occasion, the referral to the Enlarged Board of Appeal has not achieved its goal – to ensure uniform application of the law – but has, regrettably, led to even further uncertainty and flux.

We currently have the following thoughts on practice points to assist practitioners seeking to rely on post-filed data to establish an inventive step.

1. It seems to us that, contrary to the clear steer in T116/18 referenced above, synergy is indeed likely to be treated differently from other technical effects which can be associated with claimed subject matter. A stricter approach seems often to apply when assessing the admissibility of post-filed synergy data. Patentees should therefore consider whether claimed subject matter can be defended by reference to a technical effect other than synergy.
2. Boards of appeal are perhaps taking a stricter approach when assessing “downstream” patent applications, for example for new forms of a known drug or for new combinations of known drugs, where (a) multiple different options are given in a patent application, (b) the claims are then limited to one selected option, and (c) post-filed data is provided to show that one of the options disclosed is particularly advantageous. It is possible that the chances of having post-filed data admitted will increase if the various different options are covered in separate applications, so that no such selection is required.

We will watch for further developments in the jurisprudence in this area with interest.

Footnotes

1 Headnote II of the decision.

2 Reason 11.10

3 Reason 11.11

4 Reason 17.4.3

5 Reason 1.2.3

6 Reason 5.2.1

7 The application as filed in T1525/19 had been limited to the combination of linagliptin with the general class of SGLT2 inhibitors. In contrast, the application as filed in T314/20 had been filed more broadly, and was directed to combinations of both general classes of compounds.

8 Reason 1.3, penultimate two paragraphs.

9 Reason 3.3.16

10 Warner-Lambert Company LLC v Generics (UK) Ltd, [2018] UKSC 56

11 635/2015

12 Mepha Pharma AG v Bristol-Myers Squibb Holdings Ireland

13 NL:GHDHA:2023:1593

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.