Seeing the patent half empty

As an introduction, I am pleased to pass on the announcement of the upcoming patent seminar of the Institut de Boufflers, which will take place next week – online as you may guess.

The program is here, it looks great as usual, and I personally will be honored to participate in a panel discussion on the second day of the seminar (Nov. 25), dedicated to the evolution of European patent law. See you there.

Moving on to today’s case law report, I rarely miss an opportunity to comment on a decision touching upon a serious point of insufficiency of disclosure.

As a matter of fact, this ground of nullity is so seldom successful that it suddenly becomes quite interesting when it is – especially in the non-medical arts.

Directbuy is a French company specialized in electronic cigarettes and related liquids and accessories. It is located in Heillecourt, Lorraine, 5527 inhabitants (says Wikipedia). It filed a French patent application in April 2015, which was granted on November 3, 2017 under number FR 3034627. The patent relates to a device for assembling an e-cigarette and filling it with liquid.

In January 2018, Directbuy initiated infringement proceedings based on this patent against LCCF Distribution, another company located in Saint Martial d’Albarède, population of 467, (yes, that would be more than 10 times fewer than Heillecourt) in Dordogne. What would I do without Wikipedia indeed.

LCCF Distribution filed a nullity counterclaim, raising several grounds of nullity.

Had they prevailed on their first ground of nullity, the decision of the Paris Tribunal judiciaire would have been even more remarkable – but they did not.

This first ground of nullity was indeed that the patent did not relate to an invention.

Here is claim 1 of the patent:

A device for assembling and filling a vial with liquid for an electronic cigarette comprising at least one means for filling said liquid to which means are connected for supplying this liquid, characterized in that said means for filling are connected to means for supplying at least two liquids of different compositions through selection means to select, under the impulse of appropriate control means, the distribution of a liquid or a mixture of liquid in varying proportions.

Liquid selecting means – so pre-Covid…

LCCF’s argument was that the claim merely recites results to be achieved and that the structural and functional means allowing to obtain these results were not disclosed. In other terms, only a technical problem was claimed.

As you can see, this argument already had a flavor of insufficiency of disclosure. It was squarely rejected by the court as follows:

It is recalled that, in order to determine whether a patent application relates to an invention falling within the scope of patents, it is necessary to examine the nature of the problem that the patent application proposes to solve and the solution it intends to provide therein. Patentability can relate only to solutions having a technical character, while an immediate technical result in the industrial field is necessary, even though this result is weak and of weak interest.

The patent in dispute, highlighting the shortcomings linked to the traditional method of making manual mixtures of liquid, sometimes dangerous for the user and with imprecise nicotine concentrations, intends to solve these difficulties by implementing a device for assembling and filling a liquid vial by methods providing for mechanical filling by means of supplying at least two different liquids through selection tools.

The patent, which does not simply describe results, therefore proposes a technical solution to solve a technical problem.

However, turning to the second ground of nullity, insufficiency, the court noted that the description of the patent was excessively empty:

The description is particularly limited as to the described device. It merely reproduces the wording of claim 1 of the patent to explain the only including figure which is extremely schematic.

No details are given on the nature of the filling means and the liquid supply means. The means of selection invoked are not described. No explanation is given on the means of calculating the nicotine concentration of the liquid mixture which are only mentioned.

The description does not include any embodiment of the invention which may clarify the claims. 

Consequently, the patent is not sufficiently described for a person skilled in the art to be able to implement the invention (with the sole assistance of the description and the drawing).

Therefore, the patent was revoked in its entirety, without examining the other grounds of nullity of lack of novelty and lack of inventive step.

I have had a look at the patent at stake, and it is true that the description is rather succinct. The description is only two and half pages long, of which one page is dedicated to the prior art. In the rest of the description, the various “means” recited in claim 1 and its two dependent claims are mentioned, but without much in terms of additional explanations.

In other words, the court’s analysis of the patent appears to be correct. What they did not address, though, is common general knowledge. Could it somehow make up for the shortcomings of the description?

That said, even if sufficiency had been acknowledged (taking into account common general knowledge), there may have been serious issues concerning novelty and inventive step as well – which may be why the court directly went for the jugular.

In fact, looking at the preliminary search report issued by the EPO (on behalf of the INPI), all three claims as filed were deemed to lack novelty over a D1 document and to lack inventive step over a D2 document combined with a D3 document.

In its response to the search report, the applicant did not amend the claims and argued thusly:

The prior art is analyzed as follows: 

Document D1 is an article published online and reviewing the state of the art in the field of electronic cigarettes and machines for filling vials with liquid for these cigarettes.

Document D1 does not disclose all the technical features recited in claim 1 of the invention. 

Document D2 is a patent application publication relating to a method for filling vials for electronic cigarettes and the related device .

Document D2 does not disclose all the technical features recited in claim 1 of the invention. 

Document D3 is a patent application publication describing a device for filling containers in the pharmaceutical field. 

Document D3 does not disclose all the technical features recited in claim 1 of the invention. 

These three documents are therefore not relevant for the appraisal of novelty. 

The applicant’s comments on inventive step were of the same nature. As a next step, the patent was granted.

As I am sure all readers are well aware, the French patent statute has been recently amended by the PACTE law. One major aspect of the reform was to strengthen examination of national patent applications in France. The patent at stake was granted under the former, pre-PACTE provisions, which, obviously, made it possible to obtain a patent without spending much effort on setting out why exactly the invention deserved one.

I have heard, and even expressed myself, concerns that the PACTE reform may hurt local applicants wary of the higher costs typically associated with more serious, EPO-type prosecution.

On the other hand, maybe today’s decision could be viewed as an example of why raising the bar, while not a perfect solution, was in fact necessary.


CASE REFERENCE: Tribunal judiciaire de Paris, 3ème chambre 1ère section, February 6, 2020, Directbuy v. LCCF Distribution, RG No.18/02372.

A man of two jobs

In today’s case, the facts and the timeline are everything, and they are far from straightforward, so bear with me.

  • Flex Elektrowerkzeuge (“Flex”) is a German company marketing electrical appliances and tools such as sanders.
  • In 1990 a French company called Flex Electroportatif Machines et Accessoires (“FEMA”) was set up to distribute Flex’ products in France.
  • In 1992, a Mr. B (remember his name! I mean, his initial) became the CEO of FEMA.
  • In 1995, Mr. B founded another company, Accessa.
  • In 1999, yet another company, MBH Développement (“MBH”) was founded by Mr. B and a Mrs. H.
  • In 2005, a first dispute arose between the German company Flex and Mr. B. Flex discovered the existence of MBH and learned that MBH had registered several trademarks (such as “Asflex”) covering products typically marketed by Flex. The dispute ended with a transaction signed in 2007, per which the trademarks at stake were assigned to Flex.
  • In 2007, Flex discovered that MBH had filed two patent applications on sanders, listing Mr. B as an inventor. Flex was again (unsurprisingly) quite unhappy. The matter was however resolved with MBH granting a free, exclusive license agreement to Flex in 2008. Another license agreement was executed between MBH and FEMA in 2009.
  • In 2011, Flex agreed to the acquisition by MBH of part of the shares of FEMA. Further to this acquisition, the major shareholders of FEMA were Flex and MBH.
  • In 2014, an exclusive distribution agreement was signed between FEMA and Accessa, while MBH took control of Accessa.
  • Later in the same year, Flex acquired most of the shares of FEMA, took full control of the subsidiary and replaced Mr. B as the CEO.
  • In 2016, FEMA filed a complaint against Mr. B, MBH and Accessa in front of the Paris TGI.
  • In 2018, both Accessa and MBH went into liquidation.
  • In November 2019, the Paris TGI issued its judgment.
  • MBH appealed and FEMA counter-appealed, which leads us to the appeal decision dated September 22, 2020 and discussed today.

The first instance lawsuit had several prongs: infringement and invalidity claims on the one hand, and an ownership claim on the other hand.

I have almost no information to share on the infringement and invalidity aspect, as this part was not challenged on appeal – and as the first instance judgment is not readily accessible online (yes, the outrageous unavailability of first instance judgments in France still persists to this day). I will thus focus on the ownership part of the litigation, which is addressed in the appeal ruling.

The dispute on appeal centered on a number of patents and patent applications filed from 2005 to 2014. Here is the full list: FR 3026668, FR 3025447, FR 3024063, FR 3021889, FR 2980502, FR 2980501, EP 2572828, EP 2314422, FR 2936439, EP 2113338, FR 2882913, EP 1632311, EP 1570932.

All of them, it seems, were filed by MBH, with Mr. B as the inventor. FEMA claimed full ownership of this portfolio.

The first instance judges granted this request, which is why MBH (or rather the administrator in charge of its liquidation) appealed.

The Cour d’appel essentially confirmed the first instance ruling. In its reasoning, the court relied on a whole body of evidence rather than a single fact.

First, they noted that FEMA had an R&D department, in which a Mr. B was employed – not the same Mr. B as defendant-Mr. B, mind you, but since the judgment has been extensively redacted, it is not always easy to get your bearings. This was evidenced, inter alia, by the employment agreement of this second Mr. B (shall we say Mr. B’?), an organization chart as well as industrial drawings (one of which was quite similar to the figure of one of the patents in suit).

Second, looking at MBH’s management reports, they were not persuaded that MBH had any actual R&D activity. As from 2008, the reports started mentioning an R&D activity justifying a tax rebate, but the court was of the opinion that this only corresponded to patent costs. An accounting report dated 2019 mentioned the existence of R&D costs and tools, but the court did not trust this report as it was issued by Mrs. H, who is none other than Mr. B’s former partner at MBH and who happens to be currently involved in another lawsuit against FEMA.

The court was thus convinced that the inventions can only have been developed using FEMA’s resources.

Mr. B appears to have been a man of two jobs. Or as we would say in French: he had a double hat.

The appellant emphasized that the mother company Flex was aware of the existence of MBH as well as of its patent filing activity as early as 2006. However, the court deemed that this did not amount to a consent to MBH’s filings.

The license agreements signed in 2008 and 2009 did not imply any consent to MBH’s filings either. Indeed, FEMA was headed by Mr. B at that time, and he thus signed the MBH / FEMA contract on behalf of both parties. As to the MBH / Flex contract, it was analyzed by the court as an attempt by MBH to avoid litigation. More importantly, the agreements only covered one patent filed in 2005, and cannot be interpreted as a consent by FEMA concerning the later filings that took place from 2008 to 2014.

The fact that MBH became a shareholder of FEMA in 2011 cannot be interpreted as proof of consent to the patent filings either.

Another interesting defense raised by MBH was that FEMA’s action was time-barred. As a reminder, the patents / applications were filed from 2005 to 2014. The first patent in the portfolio was granted in 2007, and FEMA filed its complaint only in 2016.

According to article L. 611-8 Code de la propriété intellectuelle, the limitation period applicable to claims for ownership is in principle five years from the grant of a patent. But, “in case of bad faith at the time of grant or acquisition of the right, the limitation period is five years from the expiry of the right“. Depending on which starting point is used (grant or expiry of the patent), FEMA’s action would be time-barred with respect to at least part of MBH’s portfolio.

The court analyzed Mr. B’s behavior and concluded that he acted in bad faith. Said the court, Mr. B consistently concealed his activities, as evidenced by the successive discovery by Flex of the existence of Mr. B’s other companies, of his trademarks filings, and then of his patent filings. He was thus dishonest with FEMA. The court went on to consider that Mr. B’s bad faith also meant that MBH acted in bad faith, since Mr. B was the CEO and almost sole owner of MBH.

Therefore, MBH’s defense based on the statute of limitations also failed.

The court further noted that, until 2014, Mr. B remained FEMA’s CEO, so that FEMA was in practice not able to file suit before his departure.

The court concluded:

Mr. B, owing to his position as CEO of FEMA, used the financial, material and human resources of this company to develop inventions for the company MBH Développement that he created and headed, although they should have benefited FEMA, which should have been their owner, since the inventions related to products within its business perimeter. Therefore, the appealed judgment must be confirmed in that the ownership of the following patents and patent applications to FEMA was ordered: [….]. 

The court also ordered that all agreements on the patents between MBH, Accessa and FEMA be canceled. Besides, since the transfer of ownership is ex tunc, MBH must pay back FEMA any money earned through the exploitation of the patents. The court awarded provisional damages of EUR. 100,000 and ordered the communication of accounting information to enable a full assessment of damages.

MBH argued that the costs of patent filing and prosecution should be deduced from the damages, based on the general provision on unjust enrichment (article 1303 Code civil). The court disagreed, since MBH acted for its own profit and can thus not benefit from the unjust enrichment provision.

FEMA further requested that Mr. B be held personally liable in addition to MBH. However, the court deemed that this claim was time-barred, since FEMA was aware of Mr. B’s actions as early as 2006 (more than three years before the writ of summons filed in 2016). At first sight, this may seem contradictory with the finding that the claim for ownership of the older patents / applications was not time-barred, but the difference seems to be that the statute of limitations for an ownership claim explicitly contains a provision extending the limitation period in case of bad faith.

So, why is this case unusual? Well, first off, I think everyone will agree that the facts appear to be quite colorful, to say the least. But maybe this is always so in an ownership dispute.

Other than that, I do find the court’s reasoning unorthodox, to some extent.

Let’s go back to article L. 611-8:

If an IP right was filed on an invention either stolen from the inventor or their beneficiary, or in breach of a statutory or contractual obligation, the harmed party may claim ownership of the application or of the granted right. 

As a basic principle, an invention belongs to the inventor or his/her beneficiary. In order to know to whom a patent portfolio should belong, we must thus in principle 1) determine who the inventors are for each invention, and 2) establish whether there is an entity that holds the rights on the invention as a result of employment agreements signed by the inventors or as a result of any other statutory or contractual obligation.

But the court did not carry out this analysis. Assuming that Mr. B was the true inventor of the inventions at stake (which the judgment does not say), it should be borne in mind that inventions made by the head of a company who is not an employee do not necessarily and automatically (i.e. by way of the patent statute itself) belong to the company. The same applies for instance to an intern, who is not an employee either. So, did FEMA hold the rights on Mr. B’s inventions by way of an employment agreement? If not, what is the exact statutory or contractual obligation assigning his (presumed) inventor’s rights to FEMA and which was breached?

The court broadly stated that “Mr. B, owing to his CEO position at FEMA, used the financial, material and human resources of this company to develop inventions”.  While this of course sounds wrong and actionable, I am not sure it is in itself sufficient to conclude that FEMA was the true owner of the inventions. In particular, financial and material resources do not characterize ownership, I would venture. Human resources are an entirely different story, but again this aspect is not really addressed in the decision.


CASE REFERENCE: Cour d’appel de Paris, pôle 5 chambre 1, September 22, 2020, SAS MBH Développement et al. v. Flex Electroportatif Machines & Accessoires FEMA, RG No. 20/00471.

This is not a pipe

A regular patent law blogging activity comes with a number of serious pitfalls. Rambling may be one of them.

I hope readers will amicably warn me if this blog ever gets there – unless this point has already been reached?

A couple of weeks ago, when commenting on the recent pemetrexed decision of the Paris TJ, I lamented that French courts have a tendency to rely on the detailed choice of words in the description of a patent to draw dramatic conclusions concerning its scope of protection – either to restrict it or to broaden it beyond the literal wording of the claims.

I would like to continue this conversation today, but with a different perspective, namely an SPC angle; the case I would like to look at is the latest judgment in the Inegy® litigation (already reported on by others here and there).

Merck Sharp & Dohme Corp. (MSD) owns SPC No. FR05C0040 (FR’040) based on European patent EP 0720599 (EP’599) for the product “ezetimibe optionally in the form of its pharmaceutical acceptable salts in combination with simvastatin”. The originator’s product has been marketed as Inegy®.

At least four different lawsuits have taken place in France concerning this SPC, some of which have already been mentioned on this blog:

  • One involving Biogaran. MSD’s request for preliminary injunction was rejected in April 2018, and this rejection was confirmed in June 2018 (see this post), the reasons being that the SPC appeared to be invalid.
  • One involving Mylan and one involving Sandoz, in parallel. Here, MSD surprisingly obtained a preliminary injunction in March 2019 (see this post). This was however overturned on appeal in February 2020 (see this post), as the SPC appeared to be invalid.
  • One involving Teva, not yet mentioned on this blog.

Teva launched a generic version of Inegy® in April 2018 and initiated nullity proceedings against EP’599 and the FR’040 SPC. What was still known as the Paris TGI at that time rejected Teva’s nullity claim in October 2018, on the merits this time. By the way, this could be why a preliminary injunction was initially ordered against Mylan and Sandoz in March 2019 although it had been previously denied with respect to Biogaran in April-June 2018 (this piece of the puzzle was missing the last time I wrote on this topic).

Today’s decision is the ruling by the Paris Cour d’appel on Teva’s appeal against the October 2018 judgment. The first instance judgment has now been overturned, and the Cour d’appel has declared the FR’040 SPC invalid – consistently with its previous rulings of June 2018 and February 2020, but this time on the merits.

In the decision, the court rejected Teva’s objections to the validity of the basic EP’599 patent, but entertained Teva’s claim that the SPC itself is invalid under articles 3(a) and 3(c) of the SPC regulation. The reasoning is mostly in line with the February 2020 ruling already discussed here, so I may as well be brief.

EP’599 specifically claims:

  • a very broad family of compounds in claim 1 (in the form of a Markush formula);
  • ezetimibe as a specific compound in dependent claim 8; and
  • a pharmaceutical composition for the treatment or prevention of atherosclerosis, or for the reduction of plasma cholesterol levels, comprising an effective amount of the above compounds, alone or in combination with a cholesterol biosynthesis inhibitor selected from the group consisting of lovastatin, pravastatin, fluvastatin, simvastatin, CI-981, DMP-565, L-659,699, squalestatin 1 and NB598, in a pharmaceutical acceptable carrier (claim 17).

As a reminder, the FR’040 SPC is directed to the combination of ezetimibe and simvastatin.

The test of article 3(a) of the SPC regulation is whether the product was protected by the basic patent; and the test of article 3(c) of the SPC regulation is whether the product protected by the basic patent had already been the subject of a certificate.

In this case, MSD had already obtained an earlier SPC (No. FR03C0028) for ezetimibe itself.

The court reviewed the relevant case law of the CJEU, placing special emphasis on Sanofi, also known under the name of the other party, Actavis (C-443/12). In that case, based on a same patent, Sanofi had obtained a first SPC on the drug irbesartan based on a first MA, and then a second SPC on the combination of the drug irbesartan with a diuretic substance, HCTZ, based on a second MA. The CJEU ruled that, in this case, the grant of the first (mono) SPC prevented the grant of the second (combo) SPC.

The court found that the facts of the present case are very close to those of Sanofi – which, I think, is quite convincing.

Merck countered that, in Sanofi, the second compound HCTZ was not explicitly recited, only the therapeutic class (diuretics) was. But the court cited paragraph 30 of Sanofi: “it cannot be accepted that the holder of a basic patent in force may obtain a new SPC, potentially for a longer period of protection, each time he places on the market in a Member State a medicinal product containing, on the one hand, the principle active ingredient, protected as such by the holder’s basic patent and constituting, according to the statements of the referring court, the core inventive advance of that patent, and, on the other, another active ingredient which is not protected as such by that patent“. The court deemed that this reasoning applies in the present case.

Merck relied on two more recent CJEU rulings, namely Gilead (C-121/17) and Royalty Pharma (C-650/17), but the court found that these concern different situations and are not applicable.

The court then investigated “whether, from the perspective of the skilled person, based on common general knowledge at the filing date of the basic patent, and in the light of the description used to interpret the claims, according to article 69 EPC and its interpretative protocol, the product of the combination of ezetimibe and simvastatin, which is the subject-matter of the second SPC, is a product different from ezetimibe alone, protected by the patent as such“.

The court then turned to the description of the patent. Here is the central part of the reasoning:

The description of the patent, which uses the singular form to designate the invention, and uses the formulation “in yet another aspect” to present the combination of a hydroxy-substituted azetidinone, which is the subject-matter of the invention, with a cholesterol biosynthesis inhibitor, indifferently refers for hydroxy-substituted azetidinones alone and for their combination with a cholesterol biosynthesis inhibitor, to an effect “for the treatment and prevention of atherosclerosis or for the reduction of plasma cholesterol levels” without any indication of the specific therapeutic effect that distinguishes the product composed of ezetimibe alone from that comprising the combination of ezetimibe and a cholesterol biosynthesis inhibitor such as simvastatin. Therefore, the skilled person, who was aware in the prior art of the possibility of combining two anticholesterolemic drugs having different mechanisms of action (paragraph 8 of the patent – an HMG CoA reductase inhibitor and a bile acid sequestrant), and who was familiar with statins, and in particular simvastatin, which have been commonly used since the late 1980s for the treatment of hypercholesterolemia, will not consider that the combination of ezetimibe with simvastatin, or with the 9 other active ingredients also covered by claim 17 (in particular atorvastatin, for which Merck, on the basis of the same reasoning, filed a third SPC on September 12, 2014), constitutes a distinct product protected by the basic patent as such.

The underlying idea is that there is only one invention in the patent, namely ezetimibe itself (or the other compounds of the same class). The combination of ezetimibe with another, well-known, anticholesterolemic drug, is not patentably distinct from that invention, I would say (using U.S. vocabulary).

The court’s conclusion, as far as it is based on a review of the therapeutic effect of the products at stake, and of the actual contribution of the basic patent to the art, seems to make a lot of sense.

There is one portion of the reasoning that I do not feel comfortable with, though, namely the first part of the paragraph, in which the court pays attention to the expressions “the invention” (singular) and “in another aspect” – which, to me, do not really mean anything one way or the other.

Every patent attorney, sometimes every firm, has their own drafting style. This should have no impact on the extent of monopoly granted to a patent proprietor.

Let’s take an example. Imagine that, in the EP’599 patent, the combination of ezetimibe with another anticholesterolemic drug had been presented as being “a second invention“, or “a number of further inventions“. Should this have made a difference? Of course not. Simply naming a product as a distinct invention does not make it so.

“This is not a pipe” – can words change reality?

My fear is that some patent owners may take advantage of the judges’ over-reliance on contingent aspects of the specification to unduly extend the scope of their legal monopoly. Conversely, some may be hurt by innocuous, if unfortunate, syntactic structures.

Going back to the judgment at hand, the court decided that the case was clear and that there wasn’t any need for a reference to the CJEU. The FR’040 SPC was thus formally revoked.

According to reports published on other blogs, this is not the end of the story yet, as a final appeal in front of the Cour de cassation is already pending, at least in connection with the refusal to grant a PI against other generic companies. Presumably, this judgment is also going to be appealed.

As a final note, this seems to be one of these instances of fragmented European landscape. The decision acknowledges that preliminary injunctions have been granted and sometimes confirmed on appeal in Norway, the Czech Republic, Portugal, Belgium and Austria; and have been rejected (with sometimes a confirmation on appeal) in the Netherlands, Germany and Spain.


CASE REFERENCE: Cour d’appel de Paris, pôle 5 chambre 2, September 25, 2020, SAS Teva Santé et al. v. Merck Sharp & Dohme Corp., RG No.18/23642.

The flip side of the coin

Pharma-savvy readers are probably well aware of the pan-European pemetrexed litigation.

Eli Lilly’s patent No.EP1313508 is probably one of the most litigated patents in Europe. A nice summary, which is a little bit more than one year old, can be found here but there have been more recent developments, in particular in the Netherlands and in Germany.

In this context, France was late in the game, but here we are, with a remarkable judgment by the Paris TJ (tribunal judiciaire) issued a few days ago. Many thanks to Loïc Lemercier at Clifford Chance for providing a copy!

The French decision seems to be overall aligned with the rulings issued in most other jurisdictions: in a nutshell, it is a (big) win for Eli Lilly.

The patent was upheld as valid and infringed. A permanent injunction was ordered against the defendant, Fresenius, and provisional damages amounting to 28 million euros were awarded – pending final determination.

There are certainly several noteworthy aspects in this decision, but the most interesting one is probably the interpretation of the claims and the infringement finding.

Claim 1 of EP’508 is a Swiss-type claim, which is drafted as follow:

Use of pemetrexed disodium in the manufacture of a medicament for use in combination therapy for inhibiting tumor growth in mammals wherein said medicament is to be administered in combination with vitamin B12 or a pharmaceutical derivative thereof, said pharmaceutical derivative of vitamin B12 being hydroxocobalamin, cyano-10-chlorocobalamin, aquocobalamin perchlorate, aquo-10-chlorocobalamin perchlorate, azidocobalamin, chlorocobalamin or cobalamin.

The claimed invention thus relates to a cancer combination therapy based on the administration of pemetrexed disodium, the main active substance, in association with vitamin B12 (or the like).

As summarized by the court in the judgment, pemetrexed disodium is a known anticancer agent marketed by Eli Lilly under the trade name Alimta®. This agent is toxic and has significant side effects. According to the patent, these undesirable side effects may be reduced owing to the co-administration of vitamin B12, which lowers the level of methylmalonic acid (a predictor of toxic events in patients) without impairing the efficacy of pemetrexed. The two substances may be administered as a single composition or separately.

Fresenius obtained a marketing authorization for a generic of Alimta®, called Pemetrexed Fresenius Kabi®, in 2016, and has been marketing this generic in France.

According to the summary of product characteristics of Pemetrexed Fresenius Kabi® (NB: for some reason, you have to use another navigator than Google Chrome to open this link), there is a mandatory pre-medication regimen: “patients must also receive an intramuscular injection of vitamin B12 (1000 micrograms) in the week preceding the first dose of pemetrexed and once every three cycles thereafter“.

In other terms, Pemetrexed Fresenius Kabi® is intended for the combination therapy recited in EP’508.

But there is a hitch: the active substance in Pemetrexed Fresenius Kabi® is not in the same saline form as Amlita®. The generic drug contains pemetrexed diacid, whereas the originator’s contains pemetrexed disodium. And, most importantly, claim 1 of EP’508 refers to pemetrexed disodium. Fresenius’ non-infringement defense was based on this difference.

When analyzing the scope of the patent, the court made reference to article 69 EPC (the claims must be interpreted in the light of the description and drawings) and its protocol on interpretation. As a reminder, according to this protocol, claim interpretation must stay away from the extremes of an excessively literal interpretation and an excessively extensive interpretation. Besides, any element equivalent to an element specified in the claims must be taken into account.

A claim is like an iceberg: there is more to it than you can see.

Interestingly, the court cited the file history as an additional source of interpretation:

The scope of the claims is determined in the light of the description and drawings, and also if appropriate by taking into account elements from the file wrapper of examination proceedings, such as the amendments and arguments made by the patentee, which are factual elements, to be considered among others. 

The court turned to the description of the patent and noted that it makes reference to the general class of antifolate drugs, of which pemetrexed disodium is an example. This seems to be because, in the application as filed, an antifolate was originally claimed, which was later restricted to pemetrexed disodium.

The court then noted that the crux of the invention is the combination of pemetrexed with vitamin B12, and that the disodium form of the active is inconsequential:

The skilled person knows that the active part of the pemetrexed active substance is the anion (which causes the therapeutic effects as well as the unwanted side effects) combined with vitamin B12 […] and understands, without focusing on the literal wording of the claims, that the invention resides in the combined administration of the active agent, whatever its form, with the other substances claimed in the patent. 

The court added that the existence of other patents owned by the same company and claiming an active in more general terms (with the notion of “pharmaceutically acceptable salts“) is irrelevant.

The court also added that the file wrapper, in the present case, did not need to be taken into account. Anyway, it merely showed that the restriction to the disodium salt form was not necessary to distinguish the invention from the prior art. It was necessary because there was no support in the application as filed for pemetrexed in general, therefore omitting the disodium feature would have resulted in an extension of subject-matter by intermediate generalization:

[…] The examination procedure in front of the office [provides] a simple optional tool of interpretation, it has no effect on the scope of the patent, and is not binding on the court or the patent proprietor. The behavior of the patent proprietor, who accepted a modification requested by the examiner, cannot be interpreted as an admission which is binding on the court and it has no consequence on the scope of the claim. […] Especially so in the present case, as Lilly meant to designate a preferred embodiment, without intending to modify the scope of its right, even if they did not object to the examiner’s argumentation. In fact, this modification due to added matter under article 123(2) EPC was not meant to overcome a prior art prejudicing the validity of the patent, and was only made for purely formal reasons. The modification due to added matter may not prevent the patent proprietor from relying on infringement by equivalence, since it is a formal condition relating to the substance of the inventive contribution and the literal content of the specification, prohibiting the patentee from adding any element which cannot be directly and unambiguously derived from the patent [sic]. It does absolutely not alter the base on which the interpretation must be performed, and it has no bearing on the extent of protection which is conferred. 

On the contrary, with respect to assessing the scope of the patent, article 69 EPC […] demands that equivalents be taken into account. Therefore, adding matter during examination proceedings is not a bar to relying on infringement by equivalence, provided that the specifically claimed means or combination of means (here, the combined used of vitamin B12 […] with the active substance) have a novel function (namely the reduction of toxic effects without impairing therapeutic efficacy) […]. 

The technical problem to be solved consists in reducing the toxicity of the pemetrexed antifolate, without impairing therapeutic efficacy. The solution recommended in the patent, despite the restrictive formulation of the claims, is the combined administration of the pemetrexed anion and the other substances mentioned in the patent, the form in which this antifolate is administered being irrelevant. The scope of the patent thus extends to all forms of pharmaceutically acceptable forms of pemetrexed (salt or others) employed in combination with the […] other substances

There is a lot to digest here.

First, although the file history was generally presented as a source of interpretation, it is rated as a secondary, “optional” one.

My take is that file history is duly taken into account when it confirms the interpretation that the court wants to adopt, but it is disregarded when it does not fit the court’s reasoning.

Second, a primary reason for accepting that the patent covers more than pemetrexed disodium is the fact that the description of the patent generally refers to an antifolate drug.

What if the description had been more thoroughly adapted to the amended set of claims during prosecution, and if all generic mentions of the antifolate class had been removed? Would the court have reached a different conclusion, or would it have made it harder for the court to reach this conclusion? If so, this is a red flag for patent attorneys: do not underestimate the importance of this sometimes boring phase of examination proceedings at the EPO, namely the adaptation of the description.

Third, the court draws a clear distinction between two types of amendments made during prosecution: amendments introduced to distinguish the claimed invention from the art and amendments which are not necessary to distinguish the invention from the art.

Only the former type of amendment affects the actual scope of the patent.

Let’s now turn to the assessment of infringement:

Fresenius’ generic drug is composed of the same active substance pemetrexed, and its administration must be combined, as recommended in the EP’508 patent, with vitamin B12 […]. It is therefore irrelevant whether the [incriminated] compound uses a diacid solution so as to enable its administration, since this does not have any specific technical effect. It should be added that it has been admitted that a formulation specialist is able to offer a number of possible counter-ions other than sodium, to form a free acid or a number of well-known pharmaceutically acceptable salts. The selection of the salt form is thus irrelevant. Only the therapeutic effect of the pemetrexed anion combined with the other substances matters. The defendants claimed that it was not obvious to use this particular salt, which ranks 10th among frequently used salts. This is a criterion for the validity of an invention, not for the assessment of infringement. Fresenius obtained patents […] on this form of salt. But all of this is immaterial. 

The variant directed to the use of a different salt is totally incidental. Pemetrexed Fresenius Kabi® is administered according to the use provided in the invention. It is meant to treat the same cancer conditions, with the same technical effect. It was authorized as a generic drug of the originator’s drug. 

Infringement by reproduction is established. 

Since infringement by reproduction is established, in view of the scope of the patent as determined above, infringement by equivalence does not need to be decided upon. 

This latter part is a bit confusing. By “infringement by reproduction” is probably meant “literal infringement“. So the court seems to say that there is literal infringement when the claims are properly interpreted in the light of the description, so that “pemetrexed disodium” is understood to actually mean “pemetrexed“.

On the other hand, the court did cite the French equivalence test (different means, same function, function must be novel) when determining the scope of the patent (see the passages quoted above). This wavering is mostly inconsequential, though.

I personally think it would be much cleaner to consider that there is no literal infringement because the wording of the claims is crystal-clear, but that there may be infringement by equivalence when considering that the claims could have been drafted in a broader manner in view of the inventive concept. That said, literal infringement and infringement by equivalence both mean infringement, and it does not make any difference in terms of outcome or legal consequences.

To finish this post, I would like to go back to the rosuvastatin case commented on this blog a couple of years ago. In this case, the main claim was directed to the active compound rosuvastatin in the form of an acid or a non-toxic pharmaceutically acceptable salt thereof. The alleged infringement was a zinc salt of rosuvastatin. The court decided that there was no infringement because the claim, when properly interpreted in the light of the description, only covered salts in which the cation is an alkaline metal ion, an alkaline earth metal ion or an ammonium ion – and therefore excluded the zinc salt.

Rosuvastatin and pemetrexed are two sides of the same puzzling coin. On the one hand, an embodiment literally covered by the wording of the claims is found to be excluded from the patent’s scope. And on the other hand, an embodiment explicitly excluded from the claims during prosecution is found to still be covered.

I am not questioning that the outcome in both cases might be the right or “fair” one. But frankly, this is a nightmare for legal certainty.

One can also wonder whether the gap between patent prosecution and patent litigation has not become too deep. In front of the patent office, you have to argue based on the claims, the whole claims and nothing but the claims. In front of a French court, you might as well forget about the wording of the claims: it is all about what the description teaches.


CASE REFERENCE: Tribunal judiciaire de Paris, 3ème chambre 3ème section, September 11, 2020, Eli Lilly and Company & Lilly France v. Fresenius Kabi France & Fresenius Kabi Groupe France, RG No. 17/10421.

Defendants empowered

Today, this blog is featuring a high-flying case. It started with Lufthansa Technik filing complaints against Astronics, Panasonic Avionics and Thales Avionics over a period of time ranging from December 2017 to June 2018.

The three complaints related to the alleged infringement of European patent No. EP 0881145 directed to an electrical power supply system for aircraft seats.

Since the allegedly infringing device targeted in the three complaints is the same, namely Astronics’ EmPower® high power USB in-seat power system, the three proceedings were joined in October 2018.

In parallel, Lufthansa Technik initiated discovery proceedings in the U.S. against the three defendants, for the purpose of gathering evidence for the French lawsuit. This would be under 28 USC § 1782, I presume.

Lufthansa Technik thus obtained confidential information and documentation from the defendants, and made reference to it in the French proceedings, as exhibits No.8.1 to 8.53.

This, however, caused a practical problem for the defendants. Indeed, the U.S. court had issued protective orders regarding some of the discovered documents. As a result, each of the three defendants, which are independent companies, are not entitled to review the exhibits originating from the other defendants.

In October 2019, Thales Avionics filed a motion with the judge in charge of case management in order to define how the confidential information was to be shared between the parties. In November 2019, Astronics and Panasonic Avionics also requested a ruling on confidentiality, but added, as a main request, that the court should rule on the invalidity defense first, before dealing with the infringement aspect of the case if needed.

The defendants argued that there was a real risk of trade secret breach, and that the confidentiality measures were complex to set up. However, these complex measures will anyway be useless if the patent is found invalid.

Lufthansa Technik countered that separating the validity and infringement discussions was contrary to the longstanding practice of the Paris court, and that this was just a delaying tactic.

No USB socket in this 1935 pneumatic airplane seat.

The judge in charge of case management noted that, indeed, the validity of the EP’145 patent could be addressed without any communication of the discovery exhibits. The judge also regretted that the parties were unable to reach an amicable agreement on a confidentiality club to handle these exhibits.

As a result, the judge deemed it appropriate to order that the validity of the patent should be decided upon first. Therefore, Lufthansa Technik should hold off from actually filing the discovery exhibits for the time being.

The ruling was issued on June 19, 2020, and the judge set a relatively tight mandatory schedule for the next steps, namely:

  • July 31, 2020: submissions of the defendants on the validity of the patent.
  • October 2, 2020: submissions of the plaintiff on the validity of the patent.
  • October 8, 2020: partial closing of the proceedings and hearing on the validity of the patent.

Once the court rules on this validity aspect,  the case will either stop or continue with a further discussion on how to deal with the confidential exhibits.

Overall, this looks like a rather reasonable and pragmatic ruling. Approximately one patent out of two is found invalid in court. An invalidity ruling is therefore not just a remote possibility, on a statistical standpoint, and it may be sensible to make sure that a patent is valid before addressing the infringement issue, if only from the perspective of procedural economy.

On the other hand, if the infringement side of the case does proceed, this sequencing will of course ultimately lead to a delay in the final resolution – although the schedule set by the judge is relatively tight, which was possible in this case since all parties had already filed submissions on the merits including on the validity issues.

Could the validity first / infringement second sequencing become commonplace in future cases?

Probably not – although this is not the first time, see this previous post.

The patent at stake in this case has already expired (it was filed in 1998). I assume this may have played a role in the judge’s decision: since an injunction is out of question, the negative impact of a delay for the plaintiff is probably less serious than usual.

In addition, the case has other very specific features. The defendants’ motion was caused by an unusual procedural mess resulting partly from the configuration of the parties (with three independent and competing defendants); and partly from the plaintiff’s original strategy. Instead of requesting access to the defendants’ confidential business information with the French court (this may be ordered prior to the ruling on the merits or after the ruling on the merits, in a second phase dealing with the assessment of damages), Lufthansa Technik opted for the nuclear option of opening a new front in the U.S.

Now, to end up on a less positive note on this ruling, setting up some sort of confidentiality club should not be viewed as an insurmountable task. If the parties cannot agree, it should be possible for the judge to solve this problem within a matter of weeks. As a general rule, too much time seems to be wasted dealing with procedural motions such as this one. I have a feeling that this significantly impacts the overall duration of French patent litigation – Covid crisis notwithstanding.


CASE REFERENCE: Tribunal judiciaire de Paris, 3ème chambre 2ème section, ordonnance du juge de la mise en état, June 19, 2020, Lufthansa Technik AG v. Thales Avionics Inc. et al., RG No. 18/04501.