OK oral

As a patent professional well used to the EPO’s way of reasoning, the aspect of French patent litigation which often strikes me as the most fascinating one is claim interpretation.

This is because there is such a significant difference between the EPO and French courts when it comes to deciding what is and what is not covered by a patent.

Today’s case, Bayer v. Ceva, provides a noteworthy example.

I have actually already talked about this case in a previous post. It is indeed one of these disputes which gives rise to a large number of procedural developments.

Here is the context, in quite a few bullet points:

  • Bayer Intellectual Property GmbH owns European patent EP 2164496 (EP’496), granted in April 2017, on “Formulations containing triazinones and iron”.
  • Bayer Animal Health GmbH, a licensee, developed a veterinary product named Baycox Iron®. The product is an injectable formulation of a combination of toltrazuril and a dextran iron (III) complex. The veterinary indication is the treatment of coccidiosis in piglets.
  • In parallel, Ceva Santé Animale developed a similar veterinary product, called Forceris®.
  • Ceva contacted Bayer in November 2017 to let them know about their intention to market Forceris®. Ceva argued that their product was out of the scope of the EP’496 patent.
  • Ceva filed an opposition against the patent. The opposition was rejected by the EPO in a decision dated July 16, 2019. An appeal is pending.
  • Bayer disagreed and filed a petition for preliminary injunction against Ceva in France, in December 2017.
  • Ceva Santé Animale filed an opposition against EP’496 in January 2018.
  • In parallel, Ceva filed an action for a declaration of non-infringement in France, in January 2018. As far as I understand, this action is still pending.
  • In April 2018, the judge in charge of urgency proceedings rejected Bayer’s request for preliminary injunction, because Ceva’s letter of November 2017 was not enough to represent an imminent threat, and because there were serious doubts as to the scope of the patent.
  • In February 2019, Ceva obtained its marketing authorization for Forceris®.
  • Ceva started marketing Forceris® in France in April 2019.
  • In May 2019, Bayer obtained its marketing authorization for Baycox Iron®. The product started being marketed in France a few weeks thereafter.
  • Still in May 2019, the oral proceedings of the opposition against EP’496 took place, and the opposition was rejected (the written decision was issued on July 16, 2019; an appeal is pending).
  • On the same day, a cease and desist letter was sent by Bayer to Ceva.
  • An infringement seizure was carried out at Ceva’s French premises a few days later.
  • Ceva countered by requesting that the seized evidence be placed under seal. In a judgment dated June 7, 2019 (commented upon in my previous post), the judge rejected most of Ceva’s requests.
  • On June 20, 2019, Bayer filed a second petition for preliminary injunction against Ceva.
  • Petitions for preliminary injunctions have also been filed at least in Italy, Denmark, Spain and the Netherlands. It seems that Bayer prevailed at least in Spain, an ex parte injunction having been issued on July 17, 2019.

Not a straightforward case indeed.

A first interesting question is why Bayer filed a second petition for preliminary injunction (PI) in June 2019, after having lost its first petition for PI in April 2018.

This is probably an easy one: the main reason for the April 2018 decision was that it was not established at that time that there was an imminent infringement threat. Obviously, the situation had changed once Ceva had obtained its marketing authorization and started marketing its product.

However, there was a second reason for the April 2018 ruling, which was the existence of doubts as to the validity and scope of the patent. Bayer probably took a gamble on the fact that the rejection of the opposition which took place in May 2019 would be a game changer and would lift all doubts that French judges could have regarding the validity of the patent.

But the gamble was lost.

The court in charge of urgency proceedings (unusually a full panel, and not a single judge) quickly agreed that there was indeed an imminent threat this time.

But when they turned to the issue of the likelihood of infringement, things got trickier.

Claim 1 of EP’496 is directed to a formulation containing triazinones of formulas (I) or (II) (let me spare you the details), as well as polynuclear iron (III)-polysaccharide complexes.

Ceva’s argument was that the scope of the patent was limited to formulations for an oral administration. Remember that Forceris® (just like Bayer’s Baycox Iron®) is an injectable formulation, not an oral formulation.

Ready for oral administration?

In front of the EPO, especially when looking at novelty and inventive step, I think that there would never be the slightest doubt that claim 1 of EP’496 covers all formulations, whether oral, injectable or other.

This was actually confirmed by the opposition division in its written decision.

In fact, claim 13 is directed to the use of the formulation for making a drug. And dependent claim 15 adds that the drug is for oral administration. This can be viewed as further confirmation that claim 13 and hence claim 1 are not limited to an oral administration.

But national courts, and especially French courts, can be much more creative when it comes to claim interpretation.

Making reference to article 69 EPC and its protocol, the court performed a detailed analysis of the description of the patent. They noted that all examples relate to oral formulations, and that the general part of the description is also very much concerned with the provision of oral formulations. The technical effects of the claimed combination were also demonstrated in the context of oral formulations. The technical problem was thus formulated in relation with an oral administration, and the court was persuaded by Ceva’s argument that the same technical problem did not arise in connection with injectable formulations.

To put it otherwise, the inventive step of the claimed invention seems to be closely related to the oral mode of administration.

Here are some relevant excerpts of the decision:

The technical problem described by the patent is thus the absence of formulations available in the prior art for treating piglets simultaneously against coccidiosis and anemia, which the skilled person would be dissuaded from doing because the appropriate administration timing of each treatment to ensure efficacy is not the same. If an administration by injection is contemplated, this problem does not arise in the same terms since, as mentioned in the description, this iron compound is then efficiently absorbed and can be administered on day 3 at the same time as toltrazuril. 

[…]

Therefore, the purpose of the formulations of the patent is to efficiently treat piglets against coccidiosis and iron deficiency with a single operation and with an oral administration. It is indeed either the absence of a common treatment window, or the difference in mode of administration, which did not suggest to couple the two active substances at stake in a single formulation.

[…]

Beyond this particular context of oral administration [Bayer] does not demonstrate – and does not in fact claim – that its combination formulation would be more efficient than the separate intake of the two products.

[…]

It is only this surprising effect obtained in the context of an oral administration, and not of an injection (the alternative use of which on the third day is known and is clearly not encouraged), which is shown and that the patent intended to protect.

Most interestingly, the court also made reference to statements made by Bayer during prosecution of foreign patent applications.

This is a form of unofficial file wrapper estoppel which I would say is quite uncommon in French decisions. But I think the court only made reference to foreign prosecution in this particular case because it nicely confirmed their understanding of what the invention is about:

[…] This scope of protection was claimed by Bayer in the context of examination proceedings of the same patent in front of the Canadian patent office. On June 9, 2015, Bayer mentioned in its response to the examiner’s report that “the present formulation is related to addressing two problems simultaneously. One of which is to arrive at an orally administrable formulation, the other being that this formulation thereby allows for the treatment of anemia in animals that are affected by coccidiosis even via the oral route (being biased in the prior art per se to be the less reliable route even in otherwise healthy animals – see D5 above)”.

It is also emphasized that “the presently claimed invention does provide a non-obvious formulation over the art cited. Moreover, as it is clear from pages 10 and 11 of the application, providing an oral formulation that can be used to address both anemia and coccidiosis is clearly advantageous over problems associated with the monotherapy, such as the use of non-oral formulations and divided administration times” […]. The Canadian version of the granted patent is directed to “an oral formulation” […]. 

In front of the Indian patent office, it was also stated that the invention provided “means to simultaneously treat iron deficiency and coccidiosis by an (oral) mode of administration”.

As a result, the court rejected Bayer’s second petition for PI.

For other striking instances of narrow claim interpretations by French courts, see for instance this post, or this one, or this one.

I do find some merit in this way of looking at things. Assuming that the court was indeed right in finding that the actual contribution of the invention to the art was limited to an oral administration (which I have not formed an opinion on), it then seems only fair that the scope of protection should be limited to this contribution. One major issue is legal uncertainty, though: the more the scope of protection may differ from the literal wording of the claims, the more difficult it is for all stakeholders to know where they stand.

Another possible way to handle this situation would be to accept that claim 1 of the above patent is broad and covers all kinds of modes of administration, but that it may then lack inventive step over the entire scope of the claim, if indeed only the oral administration is truly inventive.

Yet, again, I do not believe that this would be a typical EPO-approved approach. When a combination of compounds provides a certain technical effect in a very specific context, in my experience the EPO generally allows a claim to the combination per se, without any restriction to the specific context in which the technical effect actually arises. But could this tradition in fact be too generous?

** Update (Sept. 30, 2019) **

Amandine Métier, counsel for Ceva, has kindly informed me that, in the parallel European lawsuits, a request for PI filed by Bayer in the Netherlands was also rejected on September 17, 2019, due to a serious doubt as to the existence of an inventive step. The Dutch judge analyzed the opposition division’s decision of rejection of the opposition and came to the conclusion that there is a reasonable chance that it will be overturned by the technical Board of appeal.

Amandine has also confirmed that an ex parte PI was issued in Spain. The inter partes hearing will take place next November.

Finally, an ex parte PI was also issued in Germany, which was then confirmed inter partes. The grounds for the decision are not yet known, and an appeal is pending.

Many thanks for this information! As always, pan-European litigation is a roller-coaster for litigants.


CASE REFERENCE: Tribunal de grande instance de Paris, September 11, 2019, Bayer Intellectual Property GmbH & Bayer Animal Health GmbH v. SA Ceva Santé Animale, RG No. 19/56082.

Where is the finish line?

Did you think you would never hear again of the thrilling subject of the French statute of limitations, and of its very controversial application to patent nullity suits? Well, think again.

In a previous post on the loi PACTE, I mentioned that the new provision putting an end to the time bar on patent nullity suits would immediately enter into force but that it remained to be seen how this would be interpreted in practice.

And so it is that, after spending lots of ink and pixels on how the statute of limitations applies to patent nullity actions, we may now have to devote more ink and pixels to a discussion on how exactly the statute will no longer apply.

Without further ado, as they say, I will now leave the floor to Matthieu Dhenne who will provide his insight on this matter.

I would like to come back to an issue close to my heart, as readers of this blog already know: the limitation period of actions for the annulment of industrial property titles.

I will not recall here the entire debate, already mentioned many times on this blog. The issue can be summarized by recalling that French judges applied a 5-year limitation period of the French Civil Code to these actions and that a debate has arisen about this case law, in particular because of the disastrous solutions to which it led.

The legislator has since intervened to combat this case law by declaring that these actions shall not be time-barred.

But a new debate is starting to emerge. Indeed, French Act No. 2019-486 of May 22, 2019, on the growth and transformation of companies – known as the “PACTE Act” – specifies that actions for the annulment of industrial property titles are not subject to any limitation period, while providing that that this applies to the “titles in force on the day of its publication“.

How should we interpret this provision?

The statute of limitations – the finish line.

Let us immediately rule out the interpretation according to which the parliament, by aiming at the titles in force on the day of the publication of the act, has excluded, on the contrary, the titles that were not (yet) in force on this date. This interpretation is disqualified, because it makes no sense. Indeed, we cannot consider that a transitional provision may have another object than transitional law, which deals with the application of an Act to past facts (i.e. facts that happened prior to its entry into force).

Only one question seems interesting with regard to the transitional provision discussed here: could it induce a retroactive effect by applying to titles in force on the day of the publication of the act, even though this day is after the limitation period has expired?

For instance, could a Court of Appeal apply the new Act to a case in which the limitation period applied at first instance? We should recall that the prevailing principle in French Law is that of the non-retroactivity of new laws, which is deduced from Article 2 of the French Civil Code; and that its counterpart is the immediate effect of new laws. Article 2222 of the same Code specifies that the limitation period, once expired, cannot be called into question.

The legislator may, however, provide, by way of exception, for the retroactivity of a law, if this is clearly and sufficiently stated. In this case, Article 124 of the PACTE Act does not specifically state that it is “retroactive”, but this is not required by Article 2 of the Civil Code or by the interpretation of the French Supreme Court (Cour de Cassation). Moreover, the specific mention that the law is “retroactive” would, as such, be of little use because it would be too vague.

Except if one considered that a specific reference to retroactivity is indeed mandatory – which the Cour de Cassation has never held – there is no evidence to suggest that the transitional provision in question is not clear or sufficient. We are therefore in a situation where it will be up to the judges to interpret the provision as closely as possible to the spirit of the text.

It must be kept in mind that the legislator intended to put an end to a jurisprudence that was belied by its text. It was therefore appropriate to refrain from restricting its application either to ongoing trials or to proceedings for annulment introduced after its entry into force or to actions for annulment based on titles issued after its entry into force. On the contrary, the text had to be worded in such a way that it could apply to all titles in force, which included all the aforementioned cases, but also titles previously issued and which had not given rise to legal proceedings on the date of the publication of the law.

Thus, the wording “titles in force on the day of its publication” seems to cover all pending and future actions and seems therefore retroactive.

I will have the opportunity to come back in more detail to this subject next week on @DeBoufflers Blog and then in the journal Propriété Industrielle (December issue); the latter will include a detailed response to an article supporting the opposite thesis, which was published in the same journal this summer.

Thank you Matthieu.

Transitional provisions are often tricky and can give rise to divided case law.

We have been there before, when the number of courts having jurisdiction in patent matters was reduced from 10 to 7 and then down to 1. And so, I would not be surprised if we got surprised again.

Data scarcity

Data scarcity is not the first feature that comes to mind when one thinks about our present time.

Plethora, excess and overplus are probably more accurate descriptors.

But French patent law may be (once again) an exception, as there has been for a number of months a real shortage of online accessible judgments.

Like most court decisions in this country, patent law rulings are in theory public documents. But, nowadays, can a document be considered as truly made available to the public if it is not available online, and if you have to place an official written or in-person request with the court’s clerks in order to hopefully get a copy?

In recent years, access to patent law decisions has made huge progress, in particular thanks to the online case law database hosted by the French patent office (INPI). But the database is not up-to-date and has not been for months now.

As far as I understand, the problem was caused by a new general policy of the justice administration, reflected in a memo dated December 2018.

The memo acknowledges that most judgments must be made available to the public, and that copies thereof can be individually communicated, upon request.

But then the memo adds that a massive communication of decisions “should be in principle avoided“. It provides the courts with a number of reasons why they can / should refuse such requests:

  • Such requests can place an undue burden on the (notoriously understaffed and underfunded, I should add) departments in charge.
  • Massive communication can breach principles of data protection.
  • In particular, decisions should not be communicated before being redacted to remove personal information.

I understand that this policy aims at thwarting legaltech companies (as well as more conventional editors?) eager to siphon off all of the justice administration data to process it (and make a profit out of it). And there may be good reasons for this. But the policy apparently had the side effect of also freezing access to legal information on the website of the INPI – a public institution.

A number of appeal and cassation appeal rulings are still published more or less regularly on the INPI website – although there is no way to know if this publication is exhaustive or not.

On the other hand, there are almost no more recent first instance decisions on the database (I can see only two first instance patent-related rulings issued in 2019 as I am writing this post). And the current general justice administration website (Legifrance) has zero – it has always been little useful in the field of patent law. This leaves us with subscription-based websites – but they are subscription-based; and I have heard that their collections may also be incomplete.

I assume that the administration’s plan is to set up its own new platform giving access to all decisions, including first instance ones.

Consider article L.111-13 of the Code de l’organisation judiciaire. When it was introduced in 2016, the initial wording of its first paragraph was:

Without prejudice to particular provisions on access to justice decisions and their publicity, judgments issued by judicial courts are made available to the public for free, making sure that privacy of those concerned is not impinged on. 

In March 2019, this paragraph was amended as follows, and another one was inserted:

Subject to specific provisions on access to justice decisions and their publicity, judgments issued by judicial courts are made available to the public for free in electronic form

Family names and first names of the individuals mentioned in the judgment, when they are parties or third parties, are redacted prior to release. Any element making it possible to identify parties, third parties, judges and court clerks is also redacted when its disclosure may impinge on the safety or privacy of these persons or their relatives. 

As you can see, things have gotten more specific: all decisions should fortunately be available online, not just communicable upon manual request. But they have to be carefully redacted first – including possibly by deleting the judges’ names. Could it be that this constraint is what is making everything more complex?

Do not expect to find this type of functionality in the future government case law database.

Another paragraph in the same article adds that the specifics of these provisions will be set by way of executive order. As far as I understand, no such executive order has been issued yet – so that the above provisions are not yet applicable. 

Interestingly, the December 2018 memo mentions the following:

The fact that article L. 111-13 […] on open data of court decisions is not applicable in the absence of an executive order to this end cannot be bypassed by way of a massive communication process which would not provide useful guarantees for the protection of personal data. 

In other terms, until the administration’s intended complete open database materializes, there will be no similar database elsewhere.

If any reader is privy to information on the timing of completion and release of the future database, I would be happy to know. Otherwise, let’s hope that the INPI will soon be able to resume a timely publication of all patent-related rulings.

As shown above, redaction of personal information in judicial decisions seems to be a primary concern of the administration. In this respect, another interesting paragraph was introduced into article L. 111-13 in March 2019:

Identification data of judges and court clerks cannot be reused so as to assess, analyze, compare or predict their real or assumed professional practice. A breach of this prohibition is punished according to articles 226-18, 226-24 and 226-31 of the Code pénal, without prejudice of measures and sanctions provided in Law No. 78-17 of January 6, 1978 relating to computer sciences, data and liberties. 

By way of information, article 226-18 Code pénal sets maximum penalties of a five-year imprisonment and a EUR. 300,000 fine. The other two articles add complementary sanctions for legal entities and individuals, respectively.

How very French to try to make us forget that judges are individuals having their own specific views – and what a contrast with what they do in the U.K. or the U.S.

On the one hand, I can see why data analysis targeting specific magistrates can be a problem. What if, knowing that Mr. X or Mrs. Y usually rules in this or that manner, parties are able to go forum-shopping by choosing the judge they like best. My understanding is that this is not possible for cases on the merits, but is possible for ex parte petitions, such as infringement seizure petitions, as different judges of the third chamber of the Paris Tribunal de grande instance (TGI) are on duty on different days of the week. And I am talking only about IP here, I am sure there are more serious potential concerns in other areas of the law.

On the other hand, if there are indeed differences of practice between judges, and I am still talking only from an IP law perspective, is this not information the public should be entitled to have?

One of the major issues of patent law, in my view, is unpredictability for all stakeholders. When a company asks whether a patent is valid or not, whether it is infringed or not, or what to do in order not to infringe, those are usually tremendously complex questions.

They make our lives as patent attorneys interesting, I will grant you that. But, on the economic standpoint, not being able to have any certainty on issues of validity and infringement is a problem. It also paves the way for various sorts of abuse. Think about the infringer willing to take a chance with a patent that its owner will not be certain to be able to successfully assert. And think about the patentee knowingly wielding its invalid title as a sword of Damocles.

Therefore, isn’t all information allowing stakeholders to reduce the level of uncertainty as much as possible legitimately useful?

Last week’s post on the different insufficiency doctrines used by Board 3.2.05 and Board 3.2.06 of the EPO offers a nice example. Isn’t this discrepancy important to know as a patentee, opponent or third party involved in the diaper industry or any other relevant technical field?

Even though the judges of the TGI work together in order to harmonize their judgments as explained in this post, they may still have different approaches. Even more importantly, the appointment of new judges to the IP chamber may also bring about significant changes.

By restricting (heck, by outlawing) any judge-based data analysis, stakeholders will be left in the dark.

In fact, this may even foster inequality between patent professionals, and between patent stakeholders. There will always be lawyers in the know, because they regularly handle many cases in front of the same judges and may even occasionally have informal discussions with them – and all the others.

Now, I would like to bring these thoughts to a conclusion, on a more self-centered point of view.

Obviously, this blog cannot properly report on French case law without any court decisions.

I do have a couple of sources, and appeal rulings seem to still be forthcoming. But first instance decisions are also essential to review; because many cases don’t give rise to an appeal ruling, and because first instance judgments tend to be more thorough and thus more interesting to analyze.

So I would be most grateful if readers could pass on what they have, so that it can be shared here.

I am happy to cite my informants or to keep them anonymous. And if you supply a decision and would like my post to include a comment from you on the decision, I am also happy to oblige.