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.

Paris, Texas

Sometimes, French patent litigation can have the distinct flavor of a road movie.

OK, you may not get to travel that much: the plot usually takes place in Paris all along, from one courtroom to the next, from first instance to appeal and then cassation and sometimes back down to appeal. In fact, until very recently, all of these courtrooms used to be concentrated in a very small area – the Ile de la Cité.

But imagination has journeys of its own. And when a litigant called Texas ends up being stuck in Paris for a number of years, what can you do, pictures of a desert come to mind. This is just how evocative some titles are.

So, two years ago, I wrote a post on this road movie of a sort, Carrera & Texas de France v. Muller. 

Can there be such a thing as free infringement, is the question I asked on that occasion. And I wondered whether the Cour de cassation would have to weigh in. Well it did, and the time has thus come to revisit the question.

As a reminder of the background of the case, Muller & Cie holds a European patent on a heating element for a heating or cooking apparatus. Carrera and Texas de France were found guilty of infringement of this patent by the Paris Tribunal de grande instance (TGI) in 2014. This verdict was confirmed on appeal. In parallel, the TGI issued a second judgment on the computation of damages: Carrera and Texas were condemned to pay Muller damages of respectively 327,733 euros and 280,130 euros for Muller’s commercial prejudice, plus 100,000 euros for moral prejudice.

The defendants appealed, and the Paris Cour d’appel completely changed course in December 2016, setting the amount of damages to zero euro.

The case went up to the Cour de cassation. The court dismissed various challenges of the appeal ruling regarding the appraisal of sufficiency of disclosure, claim interpretation and a request for information. But the real interesting part is the discussion on the damages award – or rather, on the lack thereof.

First, we need to go back to the reasoning of the Cour d’appel:

  • Muller claimed damages amounting to an apportionment of the entire infringer’s profits (successfully so in first instance).
  • However, the appeal judges held that infringer’s profits can only be taken into account in the context of the assessment of a commercial prejudice. The confiscation of the infringer’s profits is not an alternative option which would be available to the plaintiff instead of claiming the compensation of a commercial prejudice.
  • In the case at hand, Muller did not personally exploit the patent. The patent was rather exploited through six licensed subsidiaries, which were not parties to the proceedings.
  • Therefore, said the Cour d’appel, the sole proper methodology for assessing damages was the so-called indemnifying royalty, because the only profits lost by Muller were a lost royalty.
  • But since Muller did not ask for such indemnifying royalty, the court could not grant this remedy – and thus could not take into account infringer’s profits either.
  • Finally, Muller’s claim based on a moral prejudice was held to be insufficiently substantiated.

This last bullet point did not raise the cassation judges’ eyebrows. They simply noted that it was within the Cour d’appel’s power to determine whether the count of moral prejudice was sufficiently substantiated or not.

But the supreme court struck down the part of the judgment dismissing Muller’s claim based on the infringer’s profits.

The quest for infringer’s profits is not unlike that of Leprechaun gold.

Here is a working translation of the key paragraph of the judgment, broken down into shorter sentences as always:

[The court] refused to take into consideration the claim for damages based on one of the assessment criteria provided in article L. 615-7 par. 1 of the Code de la propriété intellectuelle. However, in view of the case law of the Court of justice of the European Union, Directive 2004/48/EC aims at providing a high level of protection of intellectual property rights, which takes into account the specificity of each case and is based on a mode of computation of damages corresponding to such specificity (CJEU, March 17, 2016, Liffers, C-99/15, item 24), the choice of which belongs to the harmed party. Therefore, the existence of an economic prejudice for the patent proprietor resulting from the infringement is not subject to the proviso that the proprietor personally exploits. Thus, the Cour d’appel breached the law. 

This is a rather important ruling. If we go back for a minute to article L. 615-7, there is no denying that its wording is not crystal clear in terms of which options are available to the plaintiff:

In order to set damages, the court separately takes into account:

1° Negative economic consequences of the infringement, including lost profits and losses suffered by the harmed party;

2° Moral prejudice inflicted to said party;

3° And profits made by the infringer, including savings on intellectual, material and promotional investments made owing to the infringement.

However, the court may, alternatively and upon request of the harmed party, grant a lump sum in terms of damages. This lump sum is higher than the amount of royalties or rights which would have been due if the infringer had asked for the authorization to use the right that it infringed. This sum does not exclude the further indemnification of the moral prejudice of the harmed party.  

What this Texas ruling means is that a patent proprietor has the right to choose its mode of damages computation, and in particular has the right to rely on option 3° regardless of whether they work the patent or not.

Overall, this is good news for IP right holders, as it gives them more flexibility in their argumentation and assessment of damages.

But that’s not to say that all is said and done on this issue. Article L. 615-7 recites that the court should “take into account” various factors, including the “profits made by the infringer“. Does this mean that the plaintiff is entitled to recover the full profits made by the infringer? Or only part of those?

I am not sure this is settled law. So again, and just as I wrote two years ago, I am looking forward to the next scene of the movie.


CASE REFERENCE: Cour de cassation, chambre commerciale, January 23, 2019, Carrera & Texas de France v. Muller et Cie, appeals No. R 16-28.322 & A 17-14.673.