Not sailing to victory

Thanks to Jérôme Tassi, I was made aware of a new French decision revoking a patent due to extension of subject-matter.

As I mentioned no earlier than in last week’s post, this is a ground for nullity that needs to be taken very seriously nowadays. There may previously have been a perception that added matter objections were deemed to fail in front of a court of law, and that painstaking dissections of the exact wording originally used in patent applications should be reserved for hearings taking place at Erhardtstrasse. But if so, this perception is no longer accurate.

The patent at stake is EP 2179917 to the Swiss company Createx SA.

This patent (EP’917) was filed as a divisional application stemming from EP 1531979, itself originally filed as a PCT application published under No. WO 2004/005009.

The technical field is the manufacture of fabric for sails. Yes, as in sailing boats. This is about as exotic as you can get when dealing with patent case law.

On December 31, 2015, Createx and its licensee, the U.S. company North Sails Group LLC, filed a patent infringement complaint against the French companies Incidences La Rochelle and Incidences Technologies.

On August 22, 2016, Createx filed a request for limitation of the French part of EP’917 with the INPI. It can be assumed that this was a reaction to some invalidity arguments brought forward by the defendants. The limitation was granted by the INPI on September 27, 2016.

The defendants counterclaimed for nullity of the EP’917 patent and as announced at the beginning of this post, the Paris Tribunal de grande instance (TGI) did in fact hold the patent invalid for extension of subject-matter beyond the content of the parent application as filed.

The patent as limited contains two independent claims, namely claim 1 directed to a fabric and claim 13 directed to a sail. Let’s focus on claim 1 only:

A fabric comprising:

a plurality of strips, each of the strips being formed from a plurality of substantially parallel reinforcing elements formed of threads arranged in a unidirectional manner, the threads being made of thousands of filaments distributed over the width of each strip, each of the substantially parallel reinforcing elements comprising said substantially parallel filaments; and an activated resin encasing the filaments to form a fabric;

wherein part of each strip is disposed on the juxtaposed strips; 

wherein at least some of the plurality of strips partially overlap adjacent strips, at least some of the strips substantially overlap over a length of each strip, at least some of the strips are arranged in parallel, so that a first end of a strip substantially overlap with a first end of another strip. 

Some sails look more threatening than others.

The key issue here is that the parent application as filed did not contain any claim directed to such a fabric or to a sail as recited in claim 13. All original claims related to a method of manufacturing a fabric, using a particular system.

Here is original claim 1 by way of comparison with the limited claim in the divisional patent:

A method of manufacturing shaped and reinforced fabrics, characterized in that it consists in making a fabric made of a membrane which encases continuous alternated reinforcing elements in a press which comprises an upper vessel the lower part of which is made of a supple element and a lower vessel the upper part of which is made of a supple element, the upper vessel comprising a shaping lath the shape of which can be adjusted by setting rods and the adjusting shape of which causes the three-dimensional elastic deformation of the supple elements and thus of the membrane and reinforcing elements within the press, into an active position in which the lath has a shape and an inactive position in which the shaping lath is inactive and does not cause any deformation of the membrane and reinforcing elements which are pressed flat, in that the portions of the membrane and of reinforcing elements are prepared before pressing on a conveyor belt which is disposed on a preparation table which then places these portions under the press with pressure and heat, and in that the portions of membrane and reinforcing elements overlap during the preparation so that they make a homogeneous fabric at the outlet of the press, which comprises flat portions and three-dimensional-shaped portions according to multiple shapes determined by the setting of the shaping lath upon pressing each membrane portion and reinforcing elements, the length of which is determined by the width of the press and preparation table. 

Here is the court’s take on this:

Through this rule [of the prohibition of extension of added matter], it is to be ensured that the patent proprietor cannot improve its position by adding elements that are not disclosed in the application as filed, which would be such as to confer an undue advantage by obtaining a different monopoly from the one initially claimed, which may harm legal certainty for third parties who rely on the content of the initial application.

In this case, the [parent] application as filed […] is exclusively directed to an invention which relates to a manufacturing method […]. It thus clearly derives from this initial application that the subject-matter of the protection does not relate to the product “fabric”. The EP’917 patent […] is no longer directed to a manufacturing method but expressly to products, and more particularly a “fabric” […] and a sail […]. 

[…] By altering the subject-matter of the patent […], the patent proprietor may now have a protection on any identical product independently of its manufacturing method. 

Thus, the patent proprietor necessarily increased the scope of protection initially requested, since this patent makes it possible to include within the scope of the patent products why may be obtained by different methods than the one claimed in the initial application. It is further noted that this modification also led the proprietor to encompass within the scope of protection of the patent […] a product which was not explicitly mentioned in the initial application, namely a “sail”, which cannot be entirely assimilated to [a fabric] and which was not within the subject-matter of the initial invention except through a mere reference to the prior art, which is not sufficient to include it within the perimeter of protection. 

The content of the (parent) application as filed has to be considered as a whole, including the claims, description and drawings. Due to some of the words used by the court in the above quote, there could be a suspicion that the court really focused on the claims only and confused extension of subject-matter with extension of scope of protection. But I doubt that this suspicion would be well-founded.

Indeed, not only did the original set of claims not contain any claim to a fabric or a sail, but also the original description did not recite a fabric or a sail as such as subject-matter of the invention.

Thus, it was considered that the shift in scope of protection could not be expected by third parties relying on the original application documents, which offended legal certainty. This is reminiscent of a Nestec decision already discussed on this blog.

In this earlier case, a claim to a device for the extraction of a capsule plus the capsule itself, in combination, was found to offend article 123(2) EPC because this category of product was not originally claimed, and could only be indirectly derived from the application based on how the capsule interacted with the device in operation.

I must say that, upon reading the claims of EP’917 (even before the French national limitation) and comparing them with the parent PCT as filed, I was curious about the examination process. Indeed, it seems that the divisional claims were completely drafted from scratch, combining pieces of wording from several passages of the original disclosure. This is not to say that such thorough redrafting necessarily adds matter. It may be allowable in some cases. But in my experience the EPO almost always at least requires detailed justifications that the skilled person would indeed directly and unambiguously derive the new claims from the original disclosure, which must not be treated as a reservoir from which you are allowed to pick and choose any combination of features.

To my surprise, it seems that the applicant did not submit any indication of a support for the claimed features when filing the divisional application, and that the examiner did not ask any questions and directly agreed to grant a patent.

So I wonder what happened here, and why the applicant was not even asked for any clarification by the EPO. This took place in 2009-2010, which is not that long ago.

CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre 2ème section, February 9, 2018, Createx & North Sails Group LLC v. Incidences Technologies & Incidences SailsRG No. 16/00023.

Some valid figures

Lionel Vial, Laura Barona and I have recently published an article in “the orange journal. For those readers who are not overly familiar with the French patent scene, this is the nickname of LexisNexis’ Propriété Industrielle.

This article looks at key figures of French patent litigation, based on a sample of recent decisions which were manually reviewed. In this post, I will present some highlights from this work.

As the devil is in the detail, especially when stats are concerned, the methodology that we used will be summarized at the end of the post.

The main takeaway message from our study is probably that the recent overall patent invalidation rate in France is high, at approximately 56%.

This invalidation rate conflates both patents subjected to nullity actions and patents subjected to infringement actions in which a nullity counterclaim was filed.

Now, before all patent holder readers start fleeing our country out of desperation, let me immediately add that this comes with a number of caveats.

First, we have only studied court decisions on the merits. 

However, many patent lawsuits are settled before going to trial, so that a decision on the merits is never issued. It can be postulated that patents that give rise to settlement may on average be stronger than patents that get to be actually ruled on by a court – although there are of course many different factors that come into play when a party decides whether to settle or not. Anyway, the overall invalidation rate may not reflect the actual strength of all patents that are litigated.

Second, the 56% invalidation rate corresponds to patents of which at least one claim was declared invalid by a court.

Admittedly, in the vast majority of cases, when one independent claim is declared invalid, all dependent claims that are reviewed by the court are also declared invalid (what I would call the “throwing the baby out with the bathwater” approach). On the other hand, some dependent claims of the patents in suit are frequently not examined at all, simply because their validity is not challenged (which generally means that these claims are not asserted against a defendant). As a result, even when a patent is counted as invalid in the above figure, it may in fact contain some valid claims (although, again, this is usually not apparent from the decision).

Third, the 56% figure is based only on patents the validity of which was challenged.

But not all litigated patents go through a validity review by the court. In fact, 27% of all patents mentioned in the sample that we looked at did not give rise to any validity challenge. This is significant proportion.

It is possible that these other, non-reviewed patents are stronger on average. But other explanations are possible as well. For instance, these cases may correspond to infringement defendants who have less money to spend on litigation.

This third caveat is also probably the reason why the above figure markedly differs from the lower and often-quoted invalidation rate of 27% found by Pierre Véron in a famous study. Pierre Véron’s sample was much larger than ours, but is now an older one (as it covered the period of 2000-2009). Sabine Agé recently presented additional figures for the period of 2010-2016. She said the invalidation rate varied between 21% and 42% depending on the years.

However, as far as we understand, in these studies, patents that were not challenged on a validity standpoint were counted as valid; whereas such patents are excluded from the sample on which our own figure is based.

Lucky numbers in litigation.

That being said, how does this French figure of 56% invalidation compare with other European jurisdictions?

Well, it is particularly instructive to look at the situation in Germany, since this country attracts the vast majority of the European patent litigation “market” due to its perceived patent-friendliness.

Believe it or not, but it seems (based on figures released in 2016) that the partial or total patent invalidation rate at the Bundespatentgericht and the Bundesgerichtshof is approximately 80%. More precisely, there is a 44% rate of total invalidation, whereas 36% of patents are maintained in amended form. Only 20% of patents survive a nullity challenge in their granted form.

Now, the French figure and the German one cannot be directly compared, due to differences between the two systems. Indeed, there is no claim amendment in front of a French court. Besides, frequently, some dependent claims in a patent are never actually looked at by French judges, for the reasons recalled above.

But it is anyway quite doubtful that patents fare much better in Germany than in France, on a validity standpoint. Actually the opposite could be true.

Another striking conclusion of course is that, whatever the jurisdiction, a significant proportion of granted patent claims which ultimately go to trial are found invalid. This may be an important fact to bear in mind, at the time the EPO has just released its annual report, with the usual emphasis on quality.

The quality of the work done by a patent office is certainly extremely complex to assess. But maybe more attention should be paid to court decisions in this connection. After all, European judges are those who have the last say on what a “good” or “bad” patent is.

A couple more figures before you go, especially on the distinction between French and European patents, and on the various grounds for nullity.

Approximately one third of patents litigated in France are French national patents, while two thirds are French parts of European patents.

It turns out there is a significant discrepancy between the invalidation rates for these two categories, namely: 41% for European patents and 78% for French patents.

The conclusion is pretty straightforward: European patents are stronger than French patents. This is not surprising. Since the INPI does not perform a complete examination of French patent applications, the granted claims in a French patent are usually the sole responsibility of the applicant. It is therefore always tempting to pursue overly broad claims in a French national patent, which makes them more prone to a nullity challenge.

As to the various grounds for nullity leading to the invalidation of patent claims, we found the following breakdown:

  • Lack of novelty: 15%;
  • Lack of inventive step: 54%;
  • Extension of subject-matter: 10%;
  • Insufficiency of disclosure: 15%;
  • Others (non-patentability…): 6%.

So, most patent claims are revoked for lack of inventive step. The 54% figure that we found is in full agreement with Sabine Agé’s recent presentation, based on a larger sample of decisions.

The relatively large number of patent claims which are deemed to lack inventive step certainly reflects a national take on this issue which is much more flexible than the EPO’s problem-and-solution somewhat rigid framework.

Other, maybe less known, notable facts include a recent surge in added matter-based invalidations, as French case law tends to align on EPO case law in this respect; as well as a relatively large number of insufficiency-related invalidations. My personal view is that French courts tend to be more severe than the EPO on the appraisal of sufficiency.

METHODOLOGY: this study was based on a review of all judgments on the merits issued between January 1, 2016 and July 31, 2017, both at first instance and on appeal, in which at least one issue of patent validity or patent infringement was decided upon. The sample of judgments was obtained from the Darts-ip database. They were manually analyzed by us. Interim orders as well as orders from a case management judge were excluded. Judgments from the Cour de cassation, as well as judgments concerned with other issues (computation of damages, employees’ inventions, etc.) were also excluded. The final sample contained 100 decisions, representing a total of 118 litigated patents.

Will case law crystallize?

Today, it is back again to one of the topics regularly addressed on this blog, namely the statute of limitations for patent nullity actions in France (but not only!).

Matthieu Dhenne was kind enough to send me a brand new decision from the Paris Tribunal de grande instance (TGI) which, once more, sheds new light on this thorny issue.

The patent at stake is the French part of EP 1455756, to Merck Sharp & Dohme Corp. (MSD). The patent was granted on July 9, 2008. It was opposed by two generic drug manufacturers. At first instance, the patent was maintained in amended form, according to a decision dated December 3, 2010. The opponents appealed, and their appeals were dismissed by the Board of appeal in a decision dated July 17, 2014. The publication of the amended patent took place on September 23, 2015.

Soon thereafter, on December 2, 2015, Ethypharm filed a nullity action with the Paris TGI, requesting that the French part of the patent should be revoked.

Quite predictably, MSD argued that the nullity action was time-barred.

If one directly applied the recent case law of the Paris Cour d’appel (discussed here), this should be a winning argument. Indeed, the Cour d’appel has proposed that the five year-limitation period be computed from the date of grant of the patent. With this in mind, in this case, the limitation period would have ended on July 9, 2013.

But you and I know that things are not that straightforward, as the Paris TGI does not follow the case law of the upper court, and generally favors an in concreto determination of the starting point for the limitation period (see a recent example here).

Yet, in today’s ruling you will not find any protracted discussion of an in concreto starting point. Instead, the issue is disposed of in just one paragraph:

[…] It is only on [July, 7, 2014, i.e. the date of the Board of appeal’s decision] that the drafting of the patent which is sought to be revoked was stabilized and that Ethypharm was able to precisely know the content of the claims of said patent as well as all the facts making it possible for them to act, so that the action is not time-barred and is admissible. 

I must say I have mixed feelings about this.

My initial reaction was, oh no, you must be kidding me, there is now yet another way of determining the starting point for the limitation period? This is not legal uncertainty anymore, this is legal chaos.

A few seconds later, I thought, well yes, it does make sense after all, you can’t possibly be expected to shoot at a moving target. When a patent is modified during opposition proceedings, any appeal filed at the EPO has a suspensive effect, and thus it is only once the appeal proceedings are terminated that the content of the patent is final.

A party must act within five years from the date at which they knew or should have known that the patent at stake is a possible impediment for their current or future business activities, or else be time-barred (this is more or less what I understand to be the TGI’s usual position). And how can a party know this before the patent is even its final form?

However, this ruling raises more questions than it provides answers.

What if an opposition is rejected by the opposition division and the patent thus maintained as granted instead of as amended? Should the reasoning be the same? What if an opposition is filed by a straw man (which is allowable at the EPO) and there is thus an unverifiable suspicion that the nullity claimant itself may be the true opponent, in an attempt to artificially extend the limitation period by several years?

In his message to me, Matthieu Dhenne also noted that the court’s reasoning could be applicable to other situations: limitation proceedings, but also a prior nullity suit brought forward by a third party. Taking this one step further, he observed that a patent right can in fact be modified at any time and is therefore theoretically never “stabilized” until it expires (sometimes, it can even be retroactively “stabilized” only after its expiry). He thus suggested that the full consequence of the court’s reasoning should be that the limitation period can only start running at the expiry of the patent, so that the well-identified drawbacks of this limitation period should in practice never occur.

Matthieu added that there would thus be a complete parallelism between the limitation period for infringement actions and nullity actions. Accordingly, invalid patents would not be able to unduly hinder free competition.

Definitely an interesting suggestion, but is it really what the TGI had in mind? I am quite sure we can expect more surprises in future decisions.

Apart from this, the decision is worth the read beyond the admissibility part.

First, it turns out that the nullity claim was held ill-founded on the merits and thus dismissed. As the patent in suit is a pharma patent, this is already quite remarkable. A majority of pharma patents which are litigated in this country are revoked one way or another.

Second, the decision tackles the very interesting issue of plausibility.

There has been a significant trend in France for patents to be revoked when they are held to be of a speculative nature. See for instance previous posts here, here and there.

In the present case, Ethypharm argued that the patent was of the speculative kind, which resulted in insufficiency of disclosure and lack of inventive step.

Now may be a good time to have a look at claim 1:

A nanoparticulate composition comprising the compound 2-(R)-(1-(R)-(3,5-bis(trifluoromethyl)phenyl)ethoxy)-3-(S)-(4-fluoro)phenyl-4-(3-(5-oxo-1H,4H-1,2,4-triazolo)methylmorpholine, or a pharmaceutically acceptable salt thereof, the compound having adsorbed on the surface thereof at least one surface stabilizer in an amount sufficient to maintain an effective average particle size of less than about 1000 nm; where “effective average particle size of less than about 1000 nm” means that at least 95% of the particles, by weight, have a particle size of less than about 1000 nm.

This drug composition is useful in the treatment of nausea and vomiting, especially those induced by a chemotherapeutic treatment. The short name of the active compound is aprepitant. According to the patent, the technical problem at stake was to improve the bioavailability of aprepitant. This is stated in the patent but no experimental test results are present, which led Ethypharm to consider that there was no evidence in the patent that the technical problem was properly solved.

Ethypharm also tried to use some of MSD’s posterior testing against them, by claiming that they proved that there were features missing in the patent which were essential for successfully implementing the invention.

The court was not convinced that there were indeed such essential features missing. The court also noted that there was a reference in the patent in suit to a prior U.S. patent disclosing the so-called “Nanocrystal” method, for making nanoparticles with a surface modifier adsorbed thereon, having an average size of less than 400 nm. This Nanocrystal patent also taught that such nanoparticles improve the bioavailability of poorly water-soluble actives.

Thus, said the court, the improvement in bioavailability provided by the nanoparticle form of aprepitant was plausible.

In such a case, the court continued, evidence which is external to the patent can indeed be taken into consideration for demonstrating that the technical problem is solved. The court then reviewed a number of articles and reports and was satisfied that the technical effect of improving bioavailability was well achieved.

In summary, this is an important decision for the fine-tuning of the appraisal of a speculative patent-type objection.

To me, the take-away message is that a reference in a patent to a prior art document disclosing a technical effect provides some plausibility that the technical effect is indeed achieved.

From drug crystals to crystal balls: could they possibly help us decipher future case law?

The patent survived other attacks of insufficiency of disclosure, extension of subject-matter and lack of inventive step. Quite remarkably, the main claim was in particular found to be non-obvious over the “Nanocrystal ” European patent of the same family as the U.S. patent mentioned above, which was used for supporting the plausibility of the technical effect.

The court held that:

[The “Nanocrystal” prior art] does not disclose chemical structures or features of drugs intended to be used by this process. It only mentions that it can be implemented with a large variety of medicinal substances, the substance having to be poorly soluble, that is less than 10 mg/mL, so that the skilled person does not know which actives […] can be tested with a reasonable expectation of success. He was all the less incited to do so that in December 2001, i.e. almost ten years after the priority date of [the Nanocrystal patent], the nanonization process, which has a number of constraints (in particular the heat released during milling may change the structure of the active substance and reduction to a very small size may create a problem of chemical and physical stability), was used on only four active substances (danazol, steroid A, compound WIN 63,394 and naproxene) with a verified effect on bioavailability […]. 

I have the uneasy feeling that there may be a contradiction here between the sufficiency and inventive step prongs of the court’s reasoning.

If the teaching of the Nanocrystal patent cannot be applied in an obvious manner to aprepitant, and if there are many technical uncertainties, why is it then not necessary for the MSD patent to contain evidence in the form of experimental tests showing that the process can in fact be effectively applied to this particular drug?

CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre 2ème section, January 26, 2018, Ethypharm SAS v. Merck Sharp & Dohme Corp., RG No. 16/01225.

A light ruling

The case discussed today is not a light case of patent infringement. Nor is it a case of light patent infringement. It is rather a case of patent infringement regarding light technology, namely LED-based lamps emitting light in multiple colors.

The litigation at stake pits Philips Light North America Corporation against France-based Commerce Spectacle Industrie (CSI).

This is a multifaceted dispute, but the prong of the litigation of interest to us today is Philips’ infringement claim against CSI based on the French part of European patent No. EP 1016062, and CSI’s counterclaim for revocation.

As a one sentence summary, the Paris Tribunal de grande instance (TGI) held that the patent was valid but that infringement was not sufficiently proven.

For once, let’s start with the infringement part of the decision.

Claim 1 of the patent as maintained in amended form in opposition proceedings at the EPO reads as follows:

An illumination apparatus comprising:
– a plurality of light emitters of at least two different colours adapted to be coupled to a power circuit including a power source and a common potential reference;
– driver means for driving the plurality of light emitters, the driver means comprising at least two switches connected to the plurality of light emitters and said power circuit and corresponding to respective current paths of the at least two different colour light emitters;
– a controller for periodically and independently opening and closing the at least two switches, the controller having an alterable address assigned to itself such as to identify and respond to a respective portion of an input data stream assigned thereto, which data stream portion is assigned to that controller;
– each light emitter being an LED; and
– said controller being arranged to generate a plurality of PWM signals, the PWM signals having uniform frequency, each signal corresponding to a respective colour of the plurality of LEDs of different colours, each said PWM signal causing a respective one of the at least two switches to be opened and closed at the uniform frequency according to respective independent duty cycles, and wherein said data stream portion comprises data for determining the respective duty cycles of the at least two different colour LEDs.

An enlightening invention.

Philips’ case primarily relied on experimental tests conducted on the allegedly infringing lamps by an expert appointed by them.

CSI criticized the report drafted by the expert. One major criticism was that Philips’ expert had used an oscilloscope for the experimental tests, without following the procedure set out in the instruction manual for this apparatus. In particular, the expert’s report did not show that the oscilloscope had been preheated for 30 minutes as recommended in the manual, and that a proper calibration had been performed.

CSI added that there were a number of loopholes in Philips’ demonstration, namely that “the expert report does not mention a periodic opening and closing of the switches, that [Philips] does not show how the current paths are and does not establish the presence of respective current paths […], that they never show how a switch would be connected to the diodes“.

Did the court see the light, readers are probably wondering?

Of course the answer depends whose side you are on, but the court did find that CSI’s criticisms had merit:

For each product, [Philips] relies on measures made with an oscilloscope, the conditions of use of which are not specified in the expert report […], although it is true that the instruction manual [….] notably requires a preheating of at least 30 minutes and a calibration before each use so as to allow optimal measurements […]. These basic precautions were not taken, so that the measurements obtained are necessarily suspect. It cannot be understood how the absence of a calibration could have no impact on the shape or (supposedly uniform) frequency of the signals, nor how the absence of preheating would not affect the measurements. The [alleged] continuous use [of the oscilloscope] for all measurements after the first one, which cannot be determined in the absence of any timing information in the report, and which cannot be derived from the order of the annexes which does not necessarily reflect the order of the operations, is not proven and may not compensate for a possible original malfunction.

Frankly, I am not sure what to make of all this.

On the one hand, we all remember from our science classes that calibration is not just for window-dressing. Did the expert proceed according to accepted practice or not? If the report is silent on this, the defendant and the judges cannot know for sure.

But on the other hand, any experimental report can always, always, be criticized for some choices that were made in the setup, or some omissions in the report itself.

Therefore, wouldn’t it be reasonable to consider that, if one party has taken the trouble of conducting experimental trials to prove its case, and if the other party wants to challenge the results of the trials, this other party should also make the effort of conducting counter trials – instead of simply pointing to potential loopholes?

Well, at any rate it does not look like this is the approach followed in France. The infringement plaintiff has the burden of fully proving the existence of the infringement, and the burden does not shift to the defendant by providing evidence which is anything short of bullet-proof.

That said, it seems that in the present case the demonstration of infringement was incomplete anyway, irrespective of the credibility of the expert’s report. And this would have been by itself sufficient to throw out the infringement claim – although the judgment does not contain many details:

Assuming that these measurements were sufficient, they would demonstrate an independent command of colors, but not the existence of “respective current paths” […], i.e. paths related to each light source, since independence does not per se imply an exclusive assignment. And the diagrams supplied for each product only show one light source, which prohibits any appraisal of the “respective” character of the current paths. 

Philips’ solace was that the patent was found to be valid. CSI had only raised lack of inventive step as a ground of nullity.

As a confirmation of a now well-established trend, the court followed the problem and solution approach for assessing inventive step – primarily because the parties themselves used this reasoning.

Here is the court’s summary of the appraisal to be conducted:

Since the parties, at least primarily, apply the non-mandatory “problem/solution” approach, it will be adopted […] by the court. It requires, in order to allow an objective examination of the inventive step without hindsight: 

– objectively identifying the closest prior art, i.e. a prior art reference which discloses subject-matter developed for the same purpose or having the same goal as the claimed invention and essentially having similar technical features requiring few structural modifications, 

– assessing the technical results achieved by the claimed invention relative to this prior art, 

– defining the technical problem which the invention purports to solve by obtaining these technical results, 

– examining whether, in view of the closest prior art, the skilled in the art would or would not [have been] suggested with the claimed technical features to achieve the results obtained by the claimed invention. 

In brief, this quite closely matches the EPO’s beloved reasoning.

In the present case, the parties seemed to agree that there were two partial problems at stake. Here, again, the court relied on the EPO practice, even making reference to the Guidelines for examination:

As set out in section G-VII, 7 of the Guidelines for examination at the EPO, if the claimed invention is in principle considered as a whole, which excludes that the inventive step of a combination of features be appraised for each feature taken in isolation, such a separate appraisal is necessary if the claim is made of a juxtaposition of features, and not of their combination which implies that the functional interaction between them produces its own technical effect exceeding the sum of the technical effects that they individually produce, namely a synergy effect. 

Applying these principles to the EP’062 patent, the court focused on the first partial problem only, and found that some of the claimed features were in fact neither disclosed in the closest prior art, nor in the suggested secondary reference. Therefore, it was simply impossible to arrive at a conclusion of lack of inventive step. Besides, the rationale offered by CSI for combining the documents together did not take into account the technical problem at stake in the patent. And the respective prior art documents contained contradictory teaching.

All in all, there were thus several reasons to find the claim inventive, without even having to look at the second partial problem.

CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre, 1ère section, November 16, 2017, Philips Lighting North America Corporation v. SAS Commerce Spectacle Industrie, RG No. 15/09326.

A patent in the cross hairs

Frankly, I am not so fond of the finasteride litigation, previously discussed here, here and there. But this may simply have something to do with the fact that I am sensitive to the issue of hair loss.

That said, I am grateful to Denis Schertenleib for drawing my attention to the cassation ruling issued a few days ago on this topic.

As a reminder, European patent No. EP 0724444 to Merck Sharp & Dohme Corp. relates to the second medical use of finasteride for the treatment of androgenic alopecia, with a particular dosage regimen.

The French part of EP’744 was declared invalid twice by the Paris Cour d’appel: the first time for lack of novelty (what I would call the Actavis part of the lawsuit), and the second time for insufficiency of disclosure (the Teva part of the lawsuit).

Merck did not bow and tried its luck in front of the Cour de cassation – to no avail.

The cassation ruling (Teva part) is interesting in that it confirms the sufficiency standard to be applied in the field of therapeutic inventions.

Merck argued that the appeal decision (which, again, can be found on this page) was legally incorrect on 9 different grounds – or “branches“, as they are nicely called in the supreme court parlance. As I tend to find cassation decisions unreadable, I will try to reformulate these grounds in my own words.

On a general standpoint:

1. The question that the appeal judges should have addressed is only whether the drug with the right dosage could be manufactured by the skilled person.

2. The appeal judges mixed up sufficiency with novelty and inventive step, as they examined whether there was a “specific technical teaching” in the patent, and wrongly looked for a comparison of the invention with the state of the art in the patent. They should have focused on whether the information in the patent made it plausible that the claimed dosage was therapeutically effective.

Example 4 of EP’444 describes a protocol for measuring haircount in subjects. As I commented in my earlier post, its conclusion is rather vague and unspecific in terms of technical effect achieved, which led it to be disregarded by the Cour d’appel. The following grounds were raised in relation with this example 4:

3. The appeal judges failed to realize that the example 4 shows the technical effect linked to the claimed low dosage of finasteride.

4. The appeal judges said that clinical tests are not required for showing a therapeutic use, but then contradicted themselves when they disregarded the example as not corresponding to an actual clinical test.

5. The appeal judges failed to realize that the example provides a methodology making it possible for the skilled person to verify the claimed therapeutic use.

As for example 5 of the patent, it is supposed to show a reduction in the amount of dihydro testosterone (DHT) in the scalp of patients taking finasteride. But it is also quite imprecise and kind of looks like a paper example (just like example 4). The following grounds were raised in relation with this example 5:

6. The same contradiction already mentioned in item 4. happened again when the appeal judges discussed example 5.

7. The appeal judges failed to realize that example 5 shows that the claimed dosage has an effect on the metabolic mechanism involved in alopecia – so that example 5 indirectly also supports the technical effect linked to the claimed low dosage of finasteride.

8. The appeal judges held that the example was doubtful in view of the leaflet of the Propecia drug, without checking if this leaflet was available to the skilled person at the priority date.

9. The example mentions a reduction in the amount of DHT in the scalp of patients after 6 weeks, which is why the Cour d’appel found it doubtful, as the effect on hair growth is supposed to occur after 12 weeks at least. But the court should have realized that this is normal as hair growth can only occur after the reduction in the amount of DHT.

Is Santa going to mail some interesting decisions to me this Christmas?

As usual, the Cour de cassation did not address in a detailed manner each criticism and basically simply stated that the Cour d’appel was right and that the grounds of cassation are ill-founded.

In particular, there is (alas) no general discussion on the level of plausibility required from a patent, and on the conditions under which post-published evidence may be relied upon.

Items No.8 and 9 in the above list were treated slightly differently, though. Those were found to relate to a redundant ground (“un moyen surabondant“) in the appeal decision, which means that the Cour de cassation refrained from stating whether they are correct or not, as anyway there are enough correct grounds in the appeal decision for it to be confirmed.

The key part of the ruling is translated in its entirety below (I have taken the liberty of dividing up the never-ending single sentence of the ruling into hopefully more intelligible bits):

[…] Firstly, when a claim relates to a [second] therapeutic application of a substance or composition, obtaining this therapeutic effect is a functional technical feature of the claim. Therefore, in order to meet the requirement of sufficiency of disclosure, it is not necessary to clinically demonstrate this technical effect; but the patent application must directly and unambiguously reflect the claimed therapeutic application, so that the skilled person can understand, based on commonly accepted models, that the results reflect this therapeutic application. 

The decision mentions first that, regarding claim 1 […], the description does not indicate what the technical advantage or effect resulting from this type of oral administration is. It does not contain any element showing the potential efficacy of any finasteride dosage, and does not comprise any information on the new effect of the claimed dosage and the particular properties of this new therapeutic application. The dosage between 0.05 and 1.0 mg is unrelated to the administration rate and patient body weight. Then, according to the decision, the description of the patent only mentions the “surprising and unexpected” discovery of this new therapeutic application, without describing its particular pharmacological properties relative to the state of the art, which only derive from an arbitrary selection. In view of these findings, the Cour d’appel was in a position to deduce that the patent application did not directly and unambiguously reflect the claimed therapeutic applications; and that the skilled person unaware of any specific technical teaching was not able to reproduce the invention and had to implement a research program him/herself; so that [claims 1 to 3 were] insufficiently disclosed.  

Secondly, regarding the the examples cited in the description of the patent, the appeal decision […] disregarded examples 1 and 2, which relate to the preparation of finasteride, the manufacturing process of which was known for years; and example 3, which does not relate to androgenic alopecia.

[The decision] notes on the one hand that example 4 discloses a photographic protocol for detecting hair growth by counting on a period of twelve months and holds that, since [the example] does not give any information on the conditions of a potential test and does neither describe the experiments nor the technical effect resulting from the lowering of the claimed dosage relative to prior art dosages, this example appears to be a measurement method and cannot be considered as a report on a trial. 

[The decision] notes on the other hand that example 5 does not provide any detail on the experimentation or protocol applied when finasteride is administered for six weeks. The results have shown a significant reduction in the amount of DHT but have not established hair growth or the stop of hair loss. [The decision] holds that, in the absence of any comparison criterion, whereas the reduction in DHT levels in scalp due to the administration of finasteride was already known, this example does not make it possible to compare the effects of the claimed dosage relative to a higher dosage, of the order of 5 mg, disclosed in the state of the art. 

Based on these findings, the Cour d’appel did not misrepresent the patent and did not require a clinical demonstration of the therapeutic effect of the new dosage. [Thus, the Cour d’appel rightly] considered that these examples did not directly and unambiguously reflect the claimed therapeutic application and could not remedy its insufficiency of disclosure as already noted above.

This approval of the approach taken by the appeal judges is of course important for all practitioners in the pharma field.

The trend is now once and for all confirmed, per which patents considered as “paper patentsi.e. seen as overly speculative at the date of filing, will not fare well in France.

Before concluding this post and wishing all readers happy holidays, a final word on two procedural aspects.

First, Merck also criticized the appeal decision for holding Teva’s nullity claim admissible, although, when the first instance judgment was issued in the Teva part of the lawsuit, the patent had already been revoked by the Tribunal de grande instance in the Actavis part of the lawsuit.

The Cour de cassation held that:

[…] The Cour d’appel rightly stated that the decision revoking a patent has an absolute effect [with respect to third parties] only once it is res judicata, that the [Actavis first instance] judgment was under appeal, and deduced that the Teva companies were admissible in seeking the revocation of the same patent. 

This is a welcome clarification.

The last point is that Merck of course also filed a cassation appeal against the other appeal decision in which the same patent was revoked for lack of novelty, as requested by Actavis. Here, the Cour de cassation trod the easiest path and simply dismissed the cassation appeal in a distinct ruling by noting that the patent claims have now been definitely held invalid in view of the Teva ruling.

Too bad for us commentators, as the Actavis part of the lawsuit also involved some interesting legal points.

With that, as promised, I wish all readers happy holidays.

CASE REFERENCE: Cour de cassation, chambre commerciale, December 6, 2017, Merck Sharp & Dohme Corp. v. Teva Pharmaceutical Industries Ltd. & Teva santé, pourvoi No. 15-19726.