The art of the seal

The new set of rules recently adopted in France on the protection of trade secrets can potentially affect many aspects of IP litigation, including in particular the handling of evidence that may contain confidential information.

New judgments are therefore bound to come under scrutiny in order to determine whether there is a change in practice or not: in particular is there any heightened effort by the courts to preserve the parties’ confidential information? In Conversant v. LG, it seemed that maybe there was.

But as far as today’s ruling is concerned, I am not so sure the recent law had any impact on the judge’s assessment of seized evidence.

Here is the context.

Bayer Intellectual Property GmbH owns European patent No. EP 2164496, on a formulation for treating coccidiose and iron deficiency in farm animals, such as piglets.

In November 2017, the French company Ceva Santé Animale let Bayer know that they intended to launch a veterinary product in several European countries, and Bayer replied that the product would infringe their patent. Ceva nevertheless continued to prepare for the launch of the product.

The dispute suddenly crystallized a few months ago. First, Ceva obtained a marketing authorization for their product, Forceris, in April 2019. Then, on May 16, oral proceedings took place in connection with Ceva’s opposition against EP’496 at the EPO, and the opposition division rejected the opposition. On the same day, Bayer sent a cease-and-desist letter to Ceva. Ceva postponed the launch in the Netherlands and agreed to give Bayer a 2-month advance notice prior to such launch. But no similar undertaking was made with respect to France.

Still on May 16 (definitely a busy day for Bayer’s legal team), an ex parte order for infringement seizure was issued. The seizure was carried out 4 days later at Ceva’s headquarters.

Ceva immediately filed a motion in court to request the placement of the seized documents and information under seal.

In fact, the bailiff who performed the seizure had spontaneously placed some of the seized evidence under seal and had redacted some of the seized information – but Ceva deemed that this was not enough.

The interim order further to Ceva’s motion was issued on June 7.

With piglets, there is always money at stake.

In summary, Ceva argued that Bayer did not need the seizure, as they already had information on the composition of the Forceris product; and that the seized information was highly confidential and could unfairly distort competition if revealed to Bayer.

Unsurprisingly, Bayer disagreed and requested on the contrary that the evidence placed under seal by the bailiff should be released to them.

In her order, the judge first noted that Bayer’s knowledge of technical information on the allegedly infringing product (due to prior exchanges between the parties) was irrelevant.

Indeed, the infringement seizure procedure aims at collecting evidence relating not only to the existence of an infringement, but also to the extent of the infringement.

Second, she noted that the provisions on the protection of trade secrets not only apply to seized exhibits, but also to mentions contained in the bailiff’s report:

[…] The abovementioned provision (article R. 615-2 par. 5) only cites “seized exhibits” but  it must also apply to mentions in the infringement seizure report which derive from these exhibits. Otherwise, as stated by Ceva, the protection of trade secrets would be circumvented simply by mentioning the contents of the seized documents in the report. 

Pragmatic and fair.

Then, the judge deemed that a redaction of the bailiff’s report was not required in this case; nor was the placement under seal of most of the seized exhibits.

Here is the standard to be applied, as reminded in the order:

The judge of the infringement seizure should only order justified and proportionate measures, taking into account the opposite interests and rights at stake: on the one hand, [the interests and rights] of the seized party to enjoy the protection of its trade secrets and confidential information, if it can prove this qualification; and on the other hand, [the interests and rights] of the proprietor to gain access to any relevant documents which may enable to characterize the infringement, even before any finding by the court as to the existence of the infringement, including confidential information, provided that it appears to be necessary for demonstrating infringement, in order not to deprive the infringement seizure of any efficacy. 

I don’t think this standard is substantially different from the one that was traditionally applied before the 2018 trade secret law.

In this case, the judge refused to seal information appearing on one page of the bailiff’s report and in two seized exhibits, relating to commercial plans regarding Forceris.

Ceva stated that Bayer could use the information to adjust the launch of its own competing product, Boycox Iron. But the judge noted that “the countries [where Forceris is to be marketed] are more or less those countries where pigs are raised“. In other terms, the information was obvious and not protected.

I am not sure this reasoning is completely convincing, especially in view of the “more or less” language (“peu ou prou” in the original version).

The judge further noted that the business plan had already been disclosed in the Dutch proceedings, so that the information was not confidential. This, to me, is a much stronger point.

One of the exhibits contained the price of the Forceris product. But the judge also refused to seal it.

The first reason was that Bayer needed the information for the assessment of damages. The second reason was that Ceva has disclosed its recommended price in Italy, so that the judge could not see why the information would be confidential in France (even if the French price was different from the Italian price).

Again, I tend to find this latter reason not fully satisfactory. The fact that a company has decided not to keep some information confidential does not mean that they have given up on confidentiality of all information.

As for the usefulness of the price for the computation of damages, this is of course a central point. In some cases, courts have decided that there should be a ruling on infringement first, before the patentee can access any price-related information. But in other cases, such as this one, they have decided otherwise. This is a delicate case-by-case balance, which, again, has apparently not really been altered by the new trade secrets provisions.

Other exhibits were OKed by the judge as well, as they related to invoices, orders and the like and were therefore relevant for the patentee as they contained information on the distribution network.

Only one exhibit was withheld by the judge, namely some commercial documentation which “it does not appear justified and proportionate to disclose to Bayer in view of the information already collected“.

On the other hand, the judge granted Bayer’s counterclaim for disclosure of the evidence spontaneously placed under seal by the bailiff during the seizure.

The last nail in the coffin was the following sentence:

There is no legitimate reason to prohibit Bayer from using the elements [thus] lawfully obtained in other proceedings in France or abroad. 

This is an important reminder of how useful the French infringement seizure can be in a situation of global litigation.

All in all, the judge’s ruling seems to favor efficiency of the proceedings: lengthy expertise proceedings on the exhibits are avoided, and the patentee has now all the necessary information to sue other parties if needed, and file submissions on all aspects of the case including damages. But whether the defendant’s trade secrets were sufficiently preserved in the process may be open for discussion – with the caveat that, as is often the case, the order is relatively brief and does not make it possible to fully appraise the entire context.


CASE REFERENCE: Tribunal de grande instance de Paris, ordonnance de référé, June 7, 2019, Ceva Santé Animale SA v. Bayer Intellectual Property GmbH, RG No. 19/54542.

Outdated soon

Glad to be back on the blog, with a report on a decision which already has the flavor of a thing from the past. I am very grateful to Philippe Schmitt for providing this judgment, after commenting it on his blog.

Thales, the famous electronics company, filed a French patent application at the INPI (Institut National de la Propriété Industrielle) on December 17, 2010. It was published under No. FR 2969124. A decision of refusal of the application was issued on July 17, 2018. Thales appealed, which leads us to the judgment at stake.

Here is claim 1 of the application, as modified during examination proceedings:

A method for displaying the steps of a mission of an aircraft on a screen of a display device, the mission being a flight plan, wherein each step corresponds to a waypoint of the flight plan, characterized in that the different steps are displayed in a first graphical window comprising a time scale, or “timeline” (TL), the different steps being displayed opposite the schedule corresponding to their accomplishment, and the length of the “timeline” is longer than the length of the first graphical window, and the graphical window therefore displays only a part of the “timeline”, this part being imposed by the user of the display device.

The application was refused by the INPI as they deemed that:

  • The application merely relates to a presentation of information associated to a mathematical method, without any technical features.
  • The claimed subject-matter cannot be searched and thus compared with the state of the art.

More precisely, the INPI considered that the problem at stake in the application was to facilitate the reading and correlation of data relating to the steps of an aircraft mission, provided by different display windows. In the absence of any such facilitation, operators have to check themselves the consistency of the information provided by various windows. Thus, the INPI held, the problem at stake relates to the interpretation of the data by the operators, and not to a technology of data representation. In other terms, the problem at stake is intellectual and not technical.

Regarding the last feature of claim 1, per which “the length of the “timeline” is longer than the length of the first graphical window, and the graphical window therefore displays only a part of the “timeline”, this part being imposed by the user of the display device”, the INPI remarked that the description of the application does not specify by which technical means the user displays “a part of the timeline“. Therefore, this feature is not characterized on the technical standpoint and no technical effect deriving from human-machine interactions is disclosed in the application.

The Paris court of appeal set aside the refusal, on the following grounds.

First, the court referred to the INPI guidelines and also to the EPO guidelines, regarding the definition of a “presentation of information“. The court thus held that information that is presented is not patentable, but that the way information is presented may comprise patentable technical features if it is distinct from the information itself. In particular, a feature which credibly helps the user perform a technical task owing to a continuous or guided human-machine interaction process provides a technical effect.

Next, the court analyzed that the objective of the patent is to transmit information to a user, more specifically the pilot or copilot of an aircraft, so that he/she has a consistent and integrated representation of the mission, in order to make decisions based on a complete knowledge of the state of the aircraft and of its environment.

The court then turned to the first characterizing feature of the claim, and concluded that it is not a technical feature. Namely:

The first feature relates to a window comprising the “timeline”, which displays the different steps opposite the schedule corresponding to their accomplishment. […] Such a window, oriented with past at the bottom and future on top, vertically divided in its center by a timeline, comprises way times on its left and remarkable way points on its right, including mentions such as speed, altitude, beside the current position of the aircraft represented by a symbol “A”. It thus appears that this feature, although it is key to the patent applied for, only relates to the transmission of information to the pilot of the aircraft, both concerning the cognitive information content (way times and way points) and the manner it is presented (as a function of a timeline). There is no distinct technical feature and it is thus not patentable per se. 

However, the court’s take on the very last claimed feature was quite different:

On the other hand, the second feature recites that when the length of the timeline is longer than the length of the first graphical window, the user (pilot) may display only part of the timeline (imposed by him/her). This is a technical means distinct from the content of the information itself. It should be added that this means helps the pilot select the most relevant of said information and thus produces a technical effect […].

In order to conclude that this feature is nevertheless not patentable, the [INPI] asserts that it only mentions a result to be achieved, namely imposing the display of part of the timeline, without clearly setting out the technical means allowing such result to be achieved. However, during examination proceedings, the [INPI] does not have the power to rule on the insufficiency of disclosure of the patent, so that this reasoning is moot. [The INPI] also asserts that this insufficiency of disclosure does not allow a comparison with the state of the art and the preparation of a search report. However, nothing prevents a prior art search, for instance in the “WL” window displaying the list of way points as a table, [to check] if the user can, when the length of this list is greater than the length of the window, display only part of this list. 

Presentation of information can be technical after all.

In summary, the court held that the last feature of the claim is a technical feature, so that the claimed subject-matter is an invention (or rather, is not a non-invention, as patent law is somewhat Carrollian).

The absence of disclosure of precise technical means in the description of the application should not be taken into account, the court added, since insufficiency of disclosure is not a ground for refusal of a French patent application.

In fact, the current list of grounds of refusal of a French patent application is quite limited. For example, lack of inventive step is not one of these grounds, and this is why there is no discussion of inventive step whatsoever in the judgment. This is also why the decision will be outdated soon, as additional grounds of refusal (including lack of inventive step) will be taken into consideration in the near future – see my previous report here on the Pacte bill.

A case like this one is actually probably a good argument in favor of this bill. Indeed, it makes little sense for the court of appeal to set aside a refusal without having to discuss or consider inventive step at all – which of course should be an important point to look at, since lack of inventive step is a ground for nullity of a granted patent.

This case has also made me realize that insufficiency of disclosure has not been introduced as a new ground for refusal in the Pacte bill. Why that is is a conundrum. If the aim is to strengthen French national patents and beef up examination proceedings, why leave a ground for nullity out of the INPI’s purview?

As a final remark, I am, like the court of appeal, unconvinced by the alleged impossibility to perform a prior art search. This is a current and actual ground for refusal, but in my view it should be limited to extreme cases in which the claims are for instance so poorly drafted that they are not understandable. On the other hand, as a matter of policy I do not believe that this provision should apply to features which are alleged to be non-technical – which is always a debatable notion, as shown above.

Now, here is another interesting thing. There is no European patent application equivalent to the French application at stake. But the French application was sent out to the EPO for a prior art search, as is the case for all French applications which do not claim a foreign priority.

And the file wrapper shows that the EPO examiner refused to perform the search, and noted that the claims “relate to a an abstract and generic process of displaying data, independently from any technology. This subject-matter is excluded from patentability, like any abstract method of graphical representation as such, as it is equivalent to a mathematical method“.

Therefore, to some extent, the INPI’s refusal was really based on an EPO examiner’s position.

The feature found to be technical by the court was present in claim 2 of the original claim set. So perhaps the EPO examiner did not pay as much attention to it as he should have. Or perhaps the Paris court of appeal has a more generous view of technical features than the EPO, when it comes to graphical user interfaces. Or perhaps the court would have found the claims invalid based on another ground if all grounds were within its purview.

When the new examination proceedings come into force, it can be expected that the number of refusals will markedly increase, especially in the particular field of technology of graphical user interfaces. This could give the court of appeal the opportunity to further elaborate on its views on this topic.

And if it turns out in the end that its stance is less strict than the EPO’s, this could entail a dramatic shift in filing strategies.


CASE REFERENCE: Cour d’appel de Paris, pôle 5 chambre 1, May 21, 2019, Thales v. Directeur Général de l’INPI, RG No. 18/19669.

Araxxe wields the axe

Never put all your eggs in the same basket” is probably a truism in many contexts. Including the one of patent rights.

Case in point: Araxxe, a French company, owns European patent No. EP 2087720. The patent gets opposed, gets killed by the opposition division of the EPO, and does not get resurrected by the Board of appeal. But it turns out that Araxxe had another egg in its basket, namely French national patent No. FR 2908572 – originating from the priority application. And the Paris Tribunal de grande instance (TGI) held that the patent was valid and infringed.

Let’s take a few steps back to look more precisely at what happened.

The French patent FR’572 was granted in March 2009. A couple of months later, Araxxe sent a warning letter to the Belgian company Meucci Solutions. Since the exchanges between the two firms did not lead to a resolution, Araxxe filed a complaint in front of the Paris TGI in December 2009. At that time, the European application claiming priority from FR’572 was still pending at the EPO. This led the TGI to issue a stay of proceedings in May 2011. Such a stay is ordered as of right, since the French part of the European patent substitutes for the French patent when it is definitely granted. By “definitely” is meant that a possible opposition should be taken into account.

In this case, Meucci Solutions (later renamed Sigos), did file an opposition against EP’720 after grant, in 2013. The opposition division revoked the patent, and the Board of appeal confirmed this revocation in November 2016 (case T1645/15).

Araxxe could have withdrawn the French designation of the European application in order to try to avoid such a long stay of proceedings. Indeed, the European application as filed contained the same claims as the French patent as granted. Therefore, it was extremely unlikely that the European patent would end up having a broader scope of protection than the French patent. But they did not do so.

Thus, the case resumed after the termination of the EP proceedings, with no EP-FR substitution having taken place due to the demise of the European patent.

Interestingly, Araxxe asserted only claim 8 of the French patent.

One basket, different eggs.

Here is claim 1:

A method for generating scheduled communication operations over one or more communication networks from an information system and verifying their correct invoicing, comprising:
– allocating resources within said information system for a call campaign over a predetermined period of time,
– managing communication equipment in shared time, 
– executing communication or transaction operations, in response to execution orders received from a call scheduling site, 
– collecting invoicing data produced by the operator, and
– automatic data correlation allowing anomaly identification,
characterized in that the method is simultaneously implemented for a plurality of communication operators, the allocated resources being shared, each communication operator being able to issue communication or transaction operation execution orders.

Here is now claim 8:

Application of the method of any one of claims 1 to 6, for the purpose of detecting traffic bypass operations by third parties, said method implementing a comparison of theoretically expected caller identifiers with identifiers actually recorded by the call receiving robot. 

The defendant not only challenged the validity of claim 8, but also that of claim 1, since claim 8 makes reference to claim 1. But the court refused to consider the attack against claim 1, because claim 8 is independent from claim 1, they said.

Under French practice, invalidity counterclaims are in principle only examined when they relate to the claims alleged to be infringed – regrettably so, I would say. This part of the decision is thus relatively unsurprising. As to whether claim 8 can truly be termed an independent claim, I have some doubts. Any dependent claim could actually be termed independent, in that case. Not that this categorization really matters in the end anyway.

As a result, 17 pages of the defendant’s written submissions on invalidity were disregarded. Also, the defendant’s arguments that the patentee had “agreed” that its claims were invalid since they had been modified in front of the EPO, did not fly. The court stated that:

Amending claims during examination or opposition proceedings, in view of remarks made in search reports or observations from the patent offices, is a usual and common practice which cannot be considered as an agreement by the applicant that its right is invalid. 

That is a relief for all patent attorneys I guess.

The remaining validity points to be addressed were inventive step and sufficiency of disclosure.

As an important note, EP’720 was revoked for extension of subject-matter. But the claims of the European patent as granted were different from those of the French patent. Most importantly, the French patent claims were identical to those of the European application as filed. As a result, there was no way the objections entertained by the EPO could apply similarly to the French patent.

Concerning inventive step, the court noted that the claims of the European application were initially held to be obvious during examination. But then the patent was granted, and the opposition division viewed the granted claims as inventive in the preliminary opinion annexed to the summons to oral proceedings. I would say that this is largely irrelevant, again because the European claims as granted were different. Anyhow, the court then looked at the prior art relied upon by the defendant and was not able to find any lack of obviousness.

Concerning sufficiency of disclosure, things unfortunately got somewhat murky in the judgment.

The defendant argued that the European claims had to be restricted to recite a “plurality of robots” – which led to the extension of subject-matter trap; and that this was an essential feature without which the invention could not be implemented. The court replied that claim 8 refers to claim 1, which recites “a plurality of robots“, so that there is no issue. Except that I cannot find this phrase in claim 1. Maybe another claim feature was considered as synonymous or equivalent, but the court did not provide any detailed explanation.

It gets worse at the next sentence, when the court remarked that the European patent was granted anyway. This overlooks the defendant’s whole point, which was that the European claims had to be modified in order to be overcome a serious objection.

And it gets more than worse (I would go as far as saying “worserer”) in the following sentence:

The later revocation of this European patent does not have any impact on the validity of the French patent since said revocation was based on a ground which is specific to the EPC (extension of subject-matter), which has no equivalent in French law. 

Readers will be excused if they have fallen from their seats at this point. Extension of subject-matter, of course, has nothing to do with insufficiency of disclosure, and both are fine grounds of nullity under French law.

I am not saying that the insufficiency attack was very serious, or that the court’s decision to uphold the patent was erroneous – I have no personal opinion at this point. But some parts of the judgment seem to have been drafted too hastily. This is most unfortunate at a time when there is a perception of global increase in the quality of French judgments in patent matters.

In the rest of the decision, Sigos’ challenge against the infringement seizure at a data center owned by a third party failed. As did their challenge against a bailiff report on a website dated December 2009 but filed in court only in November 2017, as a reaction to Sigos’ argumentation.

The court then turned to the analysis of infringement. I will not go into a feature by feature analysis, as the main interesting point in my view was the issue of territoriality. Indeed, Sigos argued that it performed most of its activities abroad.

The court replied:

The infringement seizure report, but also the bailiff internet report and the screenshots prove the use or the offer to use of an allegedly infringing process, by Sigos, in France. Therefore, it cannot be denied that the alleged harm takes place in France. It is irrelevant whether the means to control and handle the supposedly involved robots are in Belgium, where Sigos states its global platform […] and its main assets are located, or whether the correlation operations are performed abroad, since it appears […] that they are active in France with the main French phone companies. 

If my understanding of the above paragraph is correct, the court found it irrelevant that part of the steps of the claimed method may be implemented abroad. In that case, this would be a major departure from the traditional view that the use or purported use of a process has to be implemented in France (period) in order to be infringing.

As a consequence, the TGI issued an injunction against Sigos and awarded damages amounting to 378,000 euros and attorney’s fees amounting to 50,000 euros to Araxxe.

I don’t know if an appeal is pending or not, but some points in the decision would deserve to be further clarified.

At any rate, and going back to my initial comments at the beginning of this post, it is always useful for a right holder to have several different patents or applications in its basket especially for important cases, and this will certainly remain true even if/when the UPC comes into force.


CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre, 3ème section, February 1, 2019, Araxxe SAS v. Sigos NV, RG No. 15/15784.

A glimpse of the UPC

With the Brexit situation and the constitutional complaint in Germany, we have gotten accustomed to the Unified Patent Court (UPC) being somewhat on the back burner of our daily preoccupations.

But who knows. If and once the above serious roadblocks are removed, it may still see the light of day sooner rather than later. We should thus still be prepared for what comes next.

And is there any more pleasant way to get prepared than to get a glimpse of the future Parisian seat of the court? So far very little information had leaked. But courtesy of my colleague Patrick Marollé, we now have more details.

Indeed, Patrick had the privilege to attend the 5th France-China conference on law and Justice yesterday. And one of the highlights of this meeting with Chinese judges was a visit to the building which will host the UPC if/when it is open for business.

It turns out that the UPC will be located on the Quai de la Mégisserie.

The main advantage of this location is that it is as central as can be, a few steps away from Châtelet-Les Halles, the main hub of the Paris public transportation system. Quite appropriately, the building overlooks the Ile de la Cité and more particularly the Cour de cassation.

As an important note, Patrick has been told that this is in fact a provisional location for the first years once the court opens. As more cases start coming in, the court may move, possibly across the river, to the Palais de justice on the Ile de la Cité.

In fact, there should be plenty of room available in this historical complex, now that the Paris Tribunal de grande instance has moved to its new building in the Batignolles neighborhood.

So here are a few snapshots, that you can compare with those of the other UPC sites already made public. Looks like they will all be very different. But I would say you probably can’t compete with the view from the Quai de la Mégisserie…

There it is.
Looking at the UPC building from the Cour de cassation, across the river.
A closer view on the building.
Yes, it does still feel a little bit empty.

No need to stay an extra night to see the Eiffel tower, you can enjoy the view from the court.

Notre-Dame… Still standing.

Dolce vita for Amgevita®

Judgments on preliminary injunctions are uncommon, and thus particularly interesting.

So are judgments on prior user rights. Thus, when a case combines both aspects, things get truly exciting.

Fresenius Kabi Deutschland acquired the biosimilars activity of Merck KgaA in April 2017. European patent application No. EP 3145487, filed in May 2015, was part of the deal. The patent was later granted, on August 22, 2018.

The patent relates to a pharmaceutical composition which is a biosimilar of the drug Humira®, based on the antibody adalimumab as an active ingredient.

Humira® is what people call a blockbuster drug. It has a number of therapeutic indications, including rheumatoid arthritis, psoriasis, Crohn’s disease and other inflammatory conditions.

Humira® was developed by Abbott and, according to the court, the SPC which protected the originator’s product expired in France on October 16, 2018.

Several biosimilar drugs were developed and obtained a marketing authorization (MA) from the European Medicines Agency (EMA), including the Californian group Amgen. See here for a news report on the Humira® biosimilar landscape. Amgen’s drug is called Amgevita®. It was the first biosimilar approved by the EMA in March 2017. Its price was set in France the day after the expiry of the Humira® SPC, and my understanding is that Amgevita® was then immediately placed on the market.

Meanwhile, as mentioned above, the EP’487 patent was granted to Fresenius on August 22, 2018. Fresenius was really on the starting blocks with this patent: looking at the EPO file wrapper, the communication under R. 71(3) EPC on the intent to grant was dated July 6, 2018, and the applicant replied by paying the prescribed fee and filing the claim translations… on July 6! This is as responsive as you can be.

But Fresenius’ impatience was quite understandable, in view of the imminent expiry of the Humira® monopoly; and because they were of the opinion that Amgevita® infringed EP’487.

A number of exchanges between Fresenius and Amgen took place in October-November 2018. And since Amgen would not bend, Fresenius initiated infringement proceedings in front of the Juge des référés, i.e. the judge in charge of urgency proceedings in the Paris Tribunal de grande instance, on November 14, 2018. Fresenius notably requested a preliminary injunction (PI) against Amgen.

In parallel, Amgen filed an opposition against the patent at the EPO in December 2018.

The French hearing took place on January 14, 2019, and the judge issued her decision one month later – ruling in Amgen’s favor.

Tutto bene per Amgen.

Amgen challenged Fresenius’ request for PI by stating that such a measure would be unwarranted given that Fresenius itself does not have a product on the market; and by arguing that the patent is invalid and that there is no infringement.

In the decision, the judge first summarized the standard to be applied as follows:

[The judge] must rule on the challenges raised by the defendant, including regarding the validity of the patent itself. She must appraise the seriousness of the challenges, and at any rate should assess the proportion between the requested measures and the infringement alleged by the claimant and take the decision to ban the marketing of the infringing product or not, in view of the risks taken on one side and the other. 

A few years ago, there was a controversy regarding the standard to be applied with respect to validity challenges to PIs. Was a “manifest” invalidity the only acceptable defense to a request for PI? Recent case law is much more defendant-friendly (and probably much more reasonable, for that matter), as the above quote shows, in particular with the reference to a control of the proportionality of the requested measures.

The judge then turned to the assessment of Amgen’s non-infringement defense, which, as mentioned in the introduction, was based on a prior user right. Or rather, on a “prior personal possession” of the invention, as it is known under French law.

The judge summarized the conditions to be met for such a defense to be effective:

          • the possession [of the invention must be] before the filing date of the patent or its priority date;
          • the possessed technology [must be] identical with the patented invention; 
          • the possession [must be] on the French territory; 
          • those claiming such possession [must have acted] in good faith.

The judge also clarified that, based on longstanding case law, the possession of the invention does not require a commercial exploitation of the invention. Possession, in this context, rather corresponds to the knowledge of the invention.

Let’s now turn to claim 1 of Fresenius’ patent:

An aqueous pharmaceutical composition comprising:
(a) adalimumab;
(b) an acetate buffering agent (or acetate buffer system);
(c) a sugar stabiliser, wherein the sugar stabiliser is a non-reducing disaccharide selected from the group including trehalose and sucrose; and
(d) polysorbate 80; and
wherein the composition:
• has a pH between 5.0 and 5.5;
• is either free of arginine or comprises arginine in a concentration of at most 0.1 mM;
• is either free of phosphate buffering agents or comprises a phosphate buffer system in a concentration of at most 0.1 mM;
• is either free of amino acids or comprises one or more amino acids in a (collective) concentration of at most 0.1 mM; and
• is either free of surfactants, with the exception of polysorbate 80, or comprises one or more of surfactants excluding polysorbate 80 in a collective concentration of at most 0.001 mM.

The other claims depend on or otherwise indirectly refer to claim 1.

Amgen Inc. had applied for an authorization to run clinical trials on the product designated as ABP 501, in front of the Agence Nationale de Sécurité du Médicament et des Produits de Santé (ANSM), and the authorization was granted on August 13, 2013. This was before the priority date of EP’487.

The ABP 501 formulation was described in the request for authorization as adalimumab “at 50 mg/ml […] formulated with 10 mM sodium acetate, 9.0% (weight/volume) sucrose, 0.1% (w/v) polysorbate 80, pH 5.2“.

The judge noted that those are the essential features of the claimed invention – which seems to be a reasonable conclusion to draw; and that Amgen’s good faith did not appear to be questionable.

Thus, the only remaining point to be discussed, which in my view was the most tricky one, was whether the possession was “personal” and had taken place on the French territory.

This is potentially a major stumbling block to a prior user right defense in this country, especially in large group of companies. Under French law, a very strict distinction is usually applied between various companies, even if they are affiliated. Therefore, if company A “possessed” the invention, this does not necessarily imply that its affiliate B also “possessed” it. Even more so if A is a foreign company, and B a French one.

In this case, the request for authorization to run clinical trials was filed by the U.S. company Amgen, Inc. But the defendant to the lawsuit was a different company, namely the French subsidiary Amgen SAS.

Here is how Amgen passed the hurdle:

In view of the specific organization of the Amgen group, it is reasonable to consider (especially in urgency proceedings) that Amgen SAS, like the other European subsidiaries of the group, had direct access to the document relating to the clinical trials, and thus to the formulation of the biosimilar product ABP501, owing to the “EPIC” platform of the group, as early as March 2013. In fact, part of the clinical trials were performed in France. It must therefore be held that the possession of the formulation which is identical to the patented one, although originating from the U.S., was realized in France. 

Thus, the judge concluded that the prior user right defense raised by Amgen was a serious challenge to the likelihood of infringement.

By way of overkilling, the judge noted that the lack of novelty argument based on prior art document US 2014/0086930 was also serious.

Consequently, Fresenius’ request for PI was dismissed.

I often conclude my posts by stating how much I look forward to the next developments in the litigation (judgment on the merits, appeal…). But in this particular instance, I don’t think there is anything to look forward to, really.

Indeed, we can see on the online European patent register that Amgen withdrew its opposition a few days ago on May 8, 2019. This certainly means that a settlement has been reached. We thus likely will not hear about this dispute anymore.

By the way, the 9-month opposition time limit has not expired yet – it will on May 22. I’m somewhat curious to know whether others will oppose.

One take-away message from this ruling is that the sharing of information between the various companies of the Amgen group via a dedicated platform was crucial in securing the French affiliate’s right to use the invention despite the patent. This may be a point to keep in mind for other multinational companies.

Although on the other hand it may also be difficult to put into place as a general rule, in view of the necessity to actively protect trade secrets.


CASE REFERENCE: Tribunal de grande instance de Paris, ordonnance de référé, February 14, 2019, Fresenius Kabi Deutschland GmbH v. Amgen SAS, RG No. 19/50489.