On deciding where to start within the closest prior art – T 209/25
According to the Guidelines for Examination at the EPO (G-VII, 5.1), the closest prior art is that which in one single reference discloses the combination of features which constitutes the most promising starting point for a development leading to the invention. In selecting the closest prior art, the first consideration is that it should be directed to a similar purpose or effect as the invention or at least belong to the same or a closely related technical field as the claimed invention.
When relying on a document (e.g. a prior patent application) in an inventive step attack, it is usually necessary to identify a specific disclosure that will serve as a starting point for further assessment of inventive step. This was the central issue in T 209/25 - the patentee argued that the skilled person would not have started from an embodiment which was non-exemplified, less preferred than other embodiments, and disclosed only by way of a single selection.
Background
Claim 1 was directed to an injection device comprising a sterile and injectable composition including hyaluronic acid (HA), mepivacaine (MV, an anaesthetic) hydrochloride and a polyol (an antioxidant).
The patent
The patent sought to provide sterile and injectable HA-containing compositions which would not cause pain during the injection and would not induce allergies or result in exacerbated occurrence of haematoma.
The patent stated that the compositions comprising MV hydrochloride as an anaesthetic could be sterilised without significantly affecting the stability of HA as compared with compositions where the anaesthetic was lidocaine. The examples of the patent showed this apparent improvement in stability post-sterilisation vs. lidocaine in terms of rheological properties.
The closest prior art
D1 related to sterile and injectable compositions comprising HA for dermatological use; the compositions were said to contribute to better aesthetic effects and lower pain during administration.
Claim 1 of D1 defined compositions comprising, inter alia, HA, an anaesthetic, and alpha-lipoic acid (an antioxidant).
Claim 3 of D1 defined a list of about 70 different anaesthetics including MV and contained a ‘preferably’ (‘de préférence’) clause limiting the list to 10 elements (also including MV).
Claim 4 of D1 specified that the anaesthetic was lidocaine, preferably lidocaine hydrochloride.
The description of D1 contained passages corresponding to claims 3 and 4; one passage (‘Plus particulièrement…’) mentioned MV in a list of 10 anaesthetics, with the subsequent sentence (‘De préférence…’), again, referring solely to lidocaine and its hydrochloride.
There was no other mention of MV in D1.
D1 included two examples; they both used lidocaine hydrochloride as an anaesthetic and one of them taught that the compositions displayed excellent stability after heat sterilisation. There was no explicit mention of rheological properties in this context.
Polyols were mentioned in D1 as optional antioxidants.
Inventive step
The central issue was which embodiment of the closest prior art could serve as the starting point for assessing inventive step.
The patentee argued that the starting point within D1 was a composition comprising HA and lidocaine, and not a composition comprising HA and MV. The reasons were essentially as follows:
(a) stability post-sterilisation in D1 was mentioned only in the examples; those contained lidocaine as the anaesthetic; moreover, D1 did not refer to rheological properties in the context of stability;
(b) lidocaine was the most preferred anaesthetic in D1 and the one used in the examples, while MV was disclosed as a member of a list and was less preferred than lidocaine; it would have been necessary to single out a composition comprising MV; such a (single) selection would be unallowable particularly in the field of the invention, as the stability of HA was ‘fragile’, and the skilled person could not have derived from D1 that MV specifically would have allowed the preparation of stable, sterilised HA-containing compositions.
The Board took a different view on the matter, arriving at the conclusion that the starting point was a composition comprising HA and MV:
(i) firstly, D1 was in the same technical field and addressed the same problem of pain upon injection; it was irrelevant that stability was mentioned only in the context of lidocaine-containing compositions as exemplified in D1: the examples of D1 illustrated the more general disclosure of D1 and showed that stability was of concern to the authors of this document;
(ii) whether or not D1 referred to rheological stability was irrelevant – D1 related to the same or a similar technical problem or, at least, to the same or a closely related technical field as the patent, and the closest prior art did not have to disclose all of the problems solved by the claimed invention;
(iii) the composition comprising MV resulted from a single selection from a limited list of preferred anaesthetics of D1 and was thus directly and unambiguously derivable from the content of this document;
(iv) an embodiment that is so derivable from the prior art may be a suitable starting point for the assessment of inventive step also in respect of "selection inventions"; the mere fact that such an embodiment results from a selection within the prior art document is not enough to disregard it as a starting point;
(v) relying on such an embodiment would not amount to ex post facto analysis; in the present case, the skilled person would have inferred from D1 that MV would be suitable for preparing HA-containing injectable compositions and no effort would have been required to include MV in the compositions of D1;
(vi) the fact that the examples of D1 were limited to lidocaine was not a reason to regard the mention of MV in D1 as speculative; to disregard MV, the skilled person would have had to have doubts about its disclosure; however, neither the fact that HA+MV was not exemplified in D1, nor that the HA+MV was a less preferred embodiment than HA+lidocaine was enough to disregard the direct and unambiguous disclosure of HA+MV.
Put simply, an embodiment that is directly and unambiguously disclosed (be it by way of a single selection) does not cease to be a viable starting point merely because it is not exemplified or is less preferred than other embodiments.
The starting point for further discussion of inventive step was thus a composition comprising HA and MV. Despite three distinguishing features (the presence of a polyol, the form in which MV was provided and the way the composition was prepared), the patent fell on lack of inventive step, largely because neither the data in the patent nor post-filed data submitted during opposition-appeal proceedings helped the patentee to formulate the objective technical problem as something more ambitious than the provision of an alternative injection device.
For completeness, there is no discussion of the ‘injection device’ aspect in the decision; the patent does not appear to focus on this aspect too much, and it does not look like the patentee relied on it in their submissions.
Side note 1 – the parent case (T 1654/22)
A similar situation was discussed on the corresponding parent case (T 1654/22) directed to a method for manufacturing a HA composition. The closest prior art disclosed a list of suitable anaesthetics comprising mepivacaine, but this compound was not included in the examples.
That said, a method involving the addition of mepivacaine could still be arrived at by performing a single selection within the closest prior art and was thus directly and unambiguously disclosed.
The Opposition Division disregarded this particular method as a starting point, noting that there was no example of or pointer to this embodiment.
The Board said that there was no requirement in the case law that an embodiment must be exemplified or that there exists a pointer, suggestion or incentive to select a particular embodiment for that embodiment to be regarded as a starting point in the problem-solution approach.
The disclosure of the method including the addition of mepivacaine was not defective nor speculative, and the patentee did not show any prejudice against the use of mepivacaine in HA compositions. Furthermore, the examples of the closest prior art showed the successful use of a number of other anaesthetics, including lidocaine, which, the Board said, had very similar physicochemical and pharmaceutical properties to mepivacaine. There was thus no reason to regard the disclosure of mepivacaine via a single selection as an ‘unrealistic’ starting point.
Side note 2 – T 209/25 and G 2/21
Back to the present decision – in the preliminary opinion, Board 3.3.07 made some comments on application of G 2/21. None of those made it to the reasons for the decision (the supplementary data in question were considered late filed and thus not admitted), but I believe they are still worth noting.
On appeal, the patentee tried to rely on post-filed data allegedly showing that the presence of the polyol as antioxidant had a technical effect on rheological stability post-sterilisation.
As noted above, the effect itself was disclosed in the application as filed and was the subject of the examples.
However, the Board looked at the application as filed and noted that this effect was disclosed specifically in the context of using mepivacaine instead of lidocaine; the relevance of polyols in this context was not derivable from the application.
In addition, the Board noted that polyols were merely disclosed in the description of the application as filed, in a list containing a ‘large number’ of optional additives, with no preference expressed (this was the only mention of polyols). No polyols were included in the exemplified compositions; in fact, no specific polyols were disclosed in the application.
The Board thus had doubts as to whether the purported effect could be taken into account in line with G 2/21. According to the Board, the question of compliance with the requirements of G 2/21 could not be answered “without regard to the subject matter or distinguishing features alleged to cause this effect”.
It seems that, for this Board, the question of compliance with the requirements of G 2/21 hinged at least in part on whether the distinguishing feature could cause/contribute to the specific technical effect disclosed in the application.
The preliminary opinion in the present case is somewhat akin to the reasoning of Board 3.3.06 in T 996/22 (see reasons, 3.2.6), where the Board construed the technical teaching of the application in a specific, limited way (along the lines of ‘the application teaches that effect X arises specifically from feature Y, so feature Z cannot contribute to this effect’).
Then there is T 1950/23 from Board 3.3.02 (reported previously on LinkedIn – see below*). In that decision, the patentee tried to rely on an effect generally disclosed in the patent. The Board agreed with the opponent that it was not derivable from the application as filed that the improvement in the effect stemmed from the distinguishing feature, i.e. the presence of a sulphonate and/or phenate detergent. However, the Board said that it would be an overly strict interpretation of G 2/21 to require that the application as filed demonstrate that the improvement of a technical effect disclosed in the application as filed and relied on by patentee is specifically attributed to the distinguishing feature.
Note, however, that the relevant effect in T 1950/23 was disclosed in the introductory section of the patent rather than in a specific context; perhaps it could be argued that the effect was applicable to the totality of the subject matter. In addition, the patentee in T 1950/23 was in a somewhat better position for a successful argument under G 2/21 – detergents were mentioned in the dependent claims (alongside only one other class of additives) and in all of the examples of the patent. The Board thus concluded that the detergents were relevant to the teaching of the application as filed and that there was a link in the application between the presence of the detergent and the effect relied upon by the patentee.
Perhaps spelling out too precisely in the description how a specific technical effect originates can prove unhelpful later down the line as regards G 2/21?
If you have any questions on EPO practice, feel free to get in touch.
This article is for general information only and is not intended to constitute legal advice. If you would like tailored advice on a specific matter, please do not hesitate to contact me.
