The problem with the problem

Over the years, the Boards of appeal of the EPO have developed a number of rules which are now set in stone. The so-called “gold standard“, used for appraising whether any extension of subject-matter has occurred and whether priority is valid, is a famous one. An even more famous one is the “problem and solution approach” for appraising inventive step.

The problem and solution approach is a very clear reasoning framework which has two fantastic advantages:

  • avoiding an ex post finding artificially ignoring the actual contribution of the invention to the art; and
  • enhancing the predictability of the inventive step assessment, and thus improving legal certainty for all stakeholders, patent applicants and third parties alike.

Make no mistake: I love the problem and solution approach. Especially when I compare it with the Russian roulette of the obviousness rejections that you tend to randomly get at the USPTO.

However, one downside of a clear reasoning framework is that everyone learns how to make the best of it.

Clever patent drafting and sophisticated argumentation can thus sometimes make up for the relatively low inventive merit of some inventions. Correlatively, a number of patent professionals have noted that, for better or worse, the inventive step threshold et the EPO is generally rather low.

In fact, more often than not, whether an application or patent passes the article 56 test mostly depends on the selection of the closest prior art, and on what kind of technical problem you can possibly come up with – and these considerations may sometimes be remote from technical reality.

In such a context, I think it is a good thing when the Boards of appeal remind us that the problem and solution approach should be followed with some flexibility, in order to avoid any artificial outcome which would be unfair to third parties.

The problem and solution game.

I will take two examples of how things can get somewhat skewed in the problem and solution approach; and how some Board decisions can nevertheless offer a way to achieve a right balance.

The first example is taken from T 2123/14, already discussed on Laurent’s blog.

The initial step of the approach is the selection of the closest prior art. As summarized in section G-VII, 5.1 of the Guidelines, “in selecting the closest prior art, the first consideration is that it should be directed to a similar purpose or effect as the invention or at least belong to the same or a closely related technical field as the claimed invention“.

It is therefore a classical strategy to insert a number of goals or purposes in the description of the application, and then argue that any prior art reference which does not address one of these goals or purposes is unsuitable for being selected as the closest prior art. This is especially helpful when it is indeed difficult to establish the existence of an inventive step starting from a reference sought to be discarded.

Now, such a strategy may be curtailed by following the reasoning used in T 2123/14. As a disclaimer, please note that I have formed no personal opinion as to whether the claims at stake were actually inventive or not, I will just focus on the principles established by the Board, which seem quite relevant.

In this case, the examining division had refused the application, relating to a cosmetic composition for the hair, for lack of inventive step based on D3 as the closest prior art. But the appellant argued that the closest prior art was not D3 but rather D8, so that inventive step should be assessed only based on D8 as a starting point.

The appellant’s argument, as summarized by the Board, was that:

D8 indeed mentioned as a technical purpose, beside the improvement of untangling, volume, lightness, softness, flexibility and tidiness, also the improvement of smoothing while avoiding a greasy feel […], unlike D3 which did not mention the problem of the hair’s greasy feel or heaviness, and which merely focused on a conditioning effect, on making hair soft and flexible, improving volume, brightness and smoothness of wet and dry hair […].  (Reasons, 1.2.1, translated from the French language).

The Board did not accept this and noted the following:

[…] The claimed subject-matter must involve an inventive step relative to each prior art document. Nothing prevents the relevance relative to D3 from being examined first. If, on the face of it, the skilled person can choose between several prior art documents as reasonable starting points, inventive step can thus be effectively acknowledged only after applying the problem and solution approach to each of the options (T967/97, T21/08). If one of the options reveals obviousness, then there is no inventive step. The problem and solution approach may therefore need to be repeated for each of these options (T710/97). (Reasons, 1.2.2.1).

This is a helpful reminder. A relevant document cannot be swept under the rug by pointing to an allegedly better document. All “reasonable” starting points may have to be investigated. Of course what is “reasonable” and what is not is a matter of case by case analysis.

The decision then goes on by recalling the criterion of the “similar purpose or effect” for the closest prior art. But the Board then adds:

This does however not mean that, when an invention purports to solve a list of specific problems, said closest prior art must necessarily specifically disclose or mention all of these specific problems or more particularly only one of these problems taken in isolation. This is in particular the case for the objective technical problem, which is only established in the second step of the problem and solution approach […]. This is also the case for a technical problem which is solved in an implicit and inherent manner by the closest prior art, although this is not expressly stated in said prior art. 

Thus, simply inserting a purpose or even more so a list/series of specific purposes in the description does not allow the applicant to oppose any objection of lack of inventive step raised in view of a document which does not mention one of the abovementioned specific purposes, if this document was concerned with a similar use(Reasons 1.2.2.3).

In other terms, in order to qualify as a closest prior art (possibly among others), a reference needs to relate to a similar use. But it is not necessary for the document to explicitly address all the purposes that the drafter included in the description of the application or patent.

To me, this is a significant clarification, as the strategy consisting in inflating the number of purposes set out in the description in order to discard possible closest prior arts has certainly been quite effective for a number of applicants / patent proprietors to date.

In the case at hand, the Board found that D3 was indeed a possible closest prior art to take into consideration:

[…] Even if D3 does not explicitly mention the purposes of the claimed subject-matter, it cannot be denied that both documents D3 and D8 show compositions designed for the same goal and belong to the exact same technical field as the present application, namely conditioning shampoos, and therefore have a similar use and the same actions and effects on the hair, in particular in terms of smoothing. 

It can thus be legitimately expected that the type of cosmetic composition disclosed in D3 and D8 generally involves the same technical problems, or technical problems necessarily deriving from the similar use. Besides, even if the properties relating to the absence of a greasy feel or heavy hair are not explicitly recited in D3, it remains that example 5 of D3 does contain said non-silicone fatty components responsible for a possible greasy feel and possible heavy hair. In the absence of any specific indication or remark in example 5 of D3 regarding these properties, it is immediately apparent and obvious for the skilled person that the properties and abilities of the compositions disclosed in D3 are prima facie satisfactory, including regarding a greasy feel or heavy hair. (Reasons 1.2.2.3).

To summarize this latter point: even if a very specific technical problem is not expressly addressed in a reference, the skilled person can sometimes derive from the context that this technical problem is indeed taken into account in the document.

Here is now a second example of a possible flaw in the problem and solution approach favoring the applicant / patentee, and a possible remedy taught in T 2456/13 (decision initially reported on Peter de Lange’s blog).

The issue here is the second step of the problem and solution approach, namely the formulation of the objective technical problem. Sometimes, several technical effects relative to the closest prior art can be possibly contemplated. Depending on which technical effect(s) is/are retained, the solution may or may not be obvious for the skilled person.

The usual strategy for the applicant / patentee is to focus on one or more very specific technical effects, seldom addressed in the art, so as to more easily arrive at a finding of non-obviousness. But in some instances, this may be considered as artificial (and, if I may add, ex post) if, already based on another, less sophisticated technical effect, it would be obvious for the skilled person to arrive at the claimed invention.

The way the EPO may deal with this situation is addressed in section I.D.10.8 of what is commonly referred to as the Case law book. It is called the “bonus effect“:

[…] If, having regard to the state of the art, it would already have been obvious for a skilled person to arrive at something falling within the terms of a claim, because an advantageous effect could be expected to result from the combination of the teachings of the prior art documents, such claim lacked inventive step, irrespective of the circumstance that an extra effect (possibly unforeseen) was obtained. […] An unexpected bonus effect does not confer inventiveness on an obvious solution.

However, my subjective impression is that the bonus effect objection is seldom effective in practice.

One difficulty is how to distinguish a technical effect which is merely a bonus effect from one which absolutely must be taken into account in the formulation of the objective technical problem.

In some decisions, reference was made to “one-way street” situations, in which there is a lack of alternatives for the skilled person, as opposed to situations wherein “the use of means leading to some expected improvements might well be patentable in relying on an additional effect, provided this involved a choice from a multiplicity of possibilities” (see the above section of the Case law book). Clearly, this is a restrictive approach, as I do not think that there are generally many one-way streets in technology.

In T 2456/13, the patent proprietor argued that a secondary reference D3 would not be taken into account by the skilled person as it did not address a certain technical effect, namely the reduction in the tendency to caking.

But the Board considered this effect as a mere bonus effect. The Board did not hold that this was a one-way street situation but adopted another approach:

D3 discloses a generic process for the preparation of sodium percarbonate particles which are coated by spraying on a solution of sodium sulfate and sodium carbonate. […] According to D3, the coating improves storage stability […]. In view of the partial task of improving storage stability, the skilled person would, in the opinion of the board, refer to the relevant document D3, even if no improvement in the tendency to caking is explicitly shown. The person skilled in the art is in fact taught by D3 that, with a coating in the composition range between 82:18 and 69:31 parts by weight of sodium sulfate to sodium carbonate, particularly stable percarbonate particles can be obtained. […] He would try to adjust the advantageous ratio indicated in D3 and thus reach the claimed article.

A (slight) improvement in the tendency to caking is necessarily also obtained. In the opinion of the board, it constitutes an additional or bonus effect, which cannot justify an inventive step. The board takes into account that the improvement in the tendency to caking is only slight […] and possibly for this reason not mentioned in D3. (Reasons 4.9, translated from the German language).

This decision thus provides a possible criterion for establishing that a technical effect is merely a bonus effect, namely the relative low magnitude of the effect.

Of course, we should bear in mind that this is only one factor among others that may have to be taken into account.

All in all, these two decisions may helpfully contribute to containing some of the possible abuses of the problem and solution approach leading to artificial non-obviousness findings.


CASE REFERENCES:

T 2123/14, Board 3.3.07, October 27, 2016, L’Oréal.

T 2456/13, Board 3.3.05, March 7, 2017, Evonik Treibacher v. Kemira Oyj.

Clean sweep

There are at least three interesting aspects in today’s decision, as well as a sort of infuriating one. The interesting points are limitation, public prior use and inventive step. The infuriating one is a mistake made by the court, which stumbled over dependent claims – despite vocal criticisms of this type of mistakes made by various commentators over the years.

But before I address these points in turn, a few words on the background. The claimant in this case is Concept Microfibre, a French company specialized in microfiber textiles, as its name suggests. They own two French patents filed on the same day, FR 2926204 (FR’204) and FR 2926205 (FR’205).

In 2012, Concept Microfibre sued another French company Distribution de Matériel Européen (DME) as well as its Italian mother company, Filmop, for infringement of the two French patents. Further to the classical nullity counter strikes, the Paris Tribunal de grande instance (TGI) revoked claims 1, 2 and 3 of FR’205 (which were relied upon by the patent proprietor) in its judgment dated July 3, 2014; but the other patent FR’204 was upheld, and the defendants were found liable for infringement.

They appealed, which leads us to the judgment of September 27, 2016 by the Paris Cour d’appel.

On appeal, the patentee did not fight the revocation of claims 1, 2 and 3 of FR’205 by the TGI. So, this was quite easy to deal with for the appeal judges, and this part of the TGI decision was confirmed. What was more challenging for the court was to handle the other patent FR’204.

The first challenge was that the FR’204 patent was limited during the appeal proceedings, or more precisely at the onset of the appeal proceedings. The appeal was lodged on August 27, 2014, and a request for limitation of FR’204 was filed less than 2 months later, on October 16, 2014, at the INPI (French patent and trademark office). The limitation was accepted by the INPI on January 9, 2015.

The limitation consisted in combining claim 1 with claims 2 and 4.

The alleged infringers claimed that this limitation was invalid, because the limitation was requested “in order to adapt the right so as to escape a looming revocation” (uh, yes, that’s kind of the whole point of limitation proceedings) “and so as to harmonize it with the features of the products that they market” (right, too bad the patentee did not add a feature clearly not present in the alleged infringing products).

The alleged legal basis for the invalidity of the limitation was the legal principle “fraus omnia corrumpit“.

As much as I like Latin phrases, especially when I happen to know what they mean, the argument was clearly a long shot. The court easily concluded that the limitation was neither fraudulent nor improper, and that the defendants had had enough time to challenge the validity of the limited claims.

What I do find puzzling in this limitation case, though, is why the claims were limited in the first place. Usually, claims are limited on appeal when the patent is revoked in first instance (see a recent example in the Orange v. Free litigation). But in this case the patent was upheld in first instance.

Did the appellants immediately file submissions together with their appeal of August 27, 2014, and did these submissions contain new arguments which appeared more serious and threatening to the patentee? It is possible, but we cannot know for sure as the file wrapper is not part of the public record.

As far as I can tell, no new prior art was cited on appeal.

On the other hand, it seems that the appellants developed a proper argumentation of lack of inventive step, which was not the case in first instance.

In first instance, it looks like the defendants were so confident with their lack of novelty attack that they did not really focus on inventive step. Thus, the TGI simply discarded the inventive step attack by noting that:

regarding the first part of the characterizing portion [of the claim], the defendants did not make any observation on its inventive character and on the approach that the skilled person would have to make this composition and its particular structure. 

So I can only speculate that a convincing argument of lack of inventive step was newly introduced very early on appeal, which led to the limitation – unless readers have other ideas?

Anyway, the proprietor may actually have underestimated the new argument in spite, as the main claim as limited was ultimately found not to be inventive by the court!

The claim at stake was the following:

Cleaning cloth, in particular for cleaning the floor, intended to be carried by a broom, said cloth having a cleaning surface intended to come into contact with the floor, said cleaning surface comprising a textile composition made of polyolefine, and a textile composition made of hydrophilic microfibers, characterized in that:

– each of said textile composition made of polyolefine and said textile composition made of hydrophilic microfibers comprises a thread extending towards the cleaning surface, 

– the titer of the thread of said textile composition made of polyolefine is greater than the titer of the thread of said textile composition made of hydrophilic microfibers, 

– the first textile composition defines scrubbing areas and the second textile composition defines absorption areas, and

– on the cleaning surface, the textile composition made of polyolefine is slightly below the textile composition made of hydrophilic microfibers.

Brooms are not just useful for scrubbing and absorbing fluids.
Brooms are not just useful for scrubbing and absorbing fluids.

The court relied on an EPO-like approach to assess inventive step, in particular with respect to the definition of the closest prior art:

the closest prior art which is selected must be relevant, that is it must correspond to a similar use and must require the fewest structural and functional modifications to arrive at the claimed invention; said closest prior art must therefore have the same purpose or effect as the invention or at least belong to the same technical field as the claimed invention or a closely related field. 

The court identified the technical problem addressed in the patent as improving the efficacy of cleaning cloths both for scrubbing dirt and for absorbing fluids, owing to the two textile compositions. Then, the court identified three relevant documents which could represent the closest prior art, without really singling out one among the three. They all belonged to the same technical field although their purpose was slightly different (namely, the prior art cloths were intended for sweeping, not scrubbing).

The court then held that all claimed features could be found in various prior art documents. As far as I understand:

  • Two Korean documents both taught two textile compositions, one made of hard polyolefine fibers (suitable for scrubbing) and one made of hydrophilic fibers (suitable for absorbing fluids), the titer of the former being greater than the titer of the latter.
  • A PCT application disclosed distinct areas, with fibers made of a single thread extending towards the cleaning surface.
  • Finally, a U.S. patent disclosed a cloth made of fabric comprising thread loops of different lengths depending on their purpose (absorption of dirt for the longer loops and absorption of liquid for the shorter ones).

The court did not explicitly state that all these documents could be combined due to a “partial problem” approach, but this is probably what they more or less had in mind. At any rate, they noted that the claimed invention:

does not contravene any prejudice, does not represent any breakthrough relative to the methods already taught, does not overcome any technical difficulty and does not provide any unexpected or surprising result. 

On the face of it, and with the important caveat that I have not examined the prior art nor the submissions of the parties (which are not public), this conclusion seems to make sense.

On the other hand, what makes absolutely no sense is that one of the dependent claims was then revoked for lack of novelty:

Claim 5 recites that the polyolefine is polypropylene, but this claim lacks novelty, since claim 1 of Korean patent 10-0718962 already discloses fibers areas made of polypropylene. 

It is unfortunately relatively common for our judges to get dependent claims wrong. Here, the court apparently focused only on the features contained in claim 5, without realizing that these features should be read in combination with those of claim 1. If claim 1 is novel, then claim 5 is also novel. It can at best lack inventive step, like claim 1.

By the way, the frequency of this type of error regarding dependent claims is probably a factor to be taken into account by patent proprietors when determining their litigation strategy. If you know that your main claim is weak, it is probably better to amend it by way of a limitation rather than assume that a stronger dependent claim will survive if the main claim is revoked. Of course, that can happen, but you’d better not count on it.

As mentioned at the beginning of the post, the way the court handled the public prior use defense was also interesting, I will try to address it in another post.


CASE REFERENCE: Cour d’appel de Paris, pôle 5, chambre 1, September 27, 2016, Filmop Srl & SASU Distribution de Matériel Européen v. SAS Concept Microfibre, RG No. 14/18000.

Patent gone quiet

AstraZeneca holds European patent No. EP 0907364 directed to a sustained release formulation of the drug quetiapine, which is useful in the treatment of schizophrenia. This patent has been litigated for years in a number of EPC contracting states. As far as I understand, in the majority of (but not all) cases, the patent was found invalid by national courts (see by way of example relevant blog posts here and there).

With a ruling issued on July 1, 2016, France has just joined the list of countries in which the quetiapine patent has gone quiet.

U.S.-based generics company Mylan filed a nullity complaint in front of the Paris Tribunal de grande instance (TGI) on April 28, 2015, requesting revocation of the EP’364 patent. This was part of Mylan’s preparation for the launch at risk of their generic version of AstraZeneca’s patented drug, Xeroquel LP. The first boxes of generic quetiapine were marketed during the week of September 14, 2015. By that time, AstraZeneca had already counterclaimed for infringement of the patent.

On October 6, 2015, AstraZeneca boldly requested a preliminary injunction (PI) in front of a different court, namely the Lyon Tribunal de commerce, on unfair competition grounds. They were initially successful. The injunction was in place for a little bit more than one month, before being lifted by the Cour d’appel. This interesting aspect of the litigation was already discussed in detail in a previous post.

Finally (at least for now), in a judgment handed down on July 1, 2016, the Paris TGI revoked the French part of EP’364 for lack of inventive step. This is 14 months after the filing of the complaint, which from a timing standpoint is a great outcome.

Another great outcome is that, whether you agree with it or not, the court’s inventive step reasoning is very thorough and well argued.

Quetiapine as a drug for treating schizophrenia was known from the prior art, and in particular from an earlier AstraZeneca patent EP 0240228. Mylan’s case was that it was obvious for the skilled person starting from this earlier EP’228 patent to come up with a sustained release formulation of the drug, in view of the well-known advantages of this type of formulation, such as a more controlled diffusion of the active and a reduction in the number of daily doses to be taken.

AstraZeneca developed a number of arguments as a defense. First off, they said the court should apply the problem-and-solution approach. Accordingly, the closest prior art was not the EP’228 patent but rather a scientific paper by Hirsch & Casey describing clinical trials conducted with an immediate release formulation of quetiapine. As a result, the objective technical problem was not the provision of a sustained release version of quetiapine, but more broadly the provision of an efficient anti-psychotic drug treatment.

This did not fly with the court:

However, the problem-and-solution approach is only one method of reasoning which can be used, but not the only one, to assess the validity of a patent with respect to the inventive step requirement. In the present case, the reformulation of the technical problem suggested by the defendants relates to an artificial approach, the purpose of which, by broadening the scope of the technical problem, is to make the patent’s choice to look at a sustained release quetiapine formulation appear inventive. 

The court’s statement on the problem-and-solution approach is a reminder as a matter of principle that French courts are not bound by the EPO guidelines and case law. That being said, in practice, most judgments issued nowadays draw inspiration from the problem-and-solution approach: a starting point and a technical problem are defined, and the question of obviousness is analyzed through the lens of this technical problem – which is the very framework of the problem-and-solution approach. The present case is no exception.

It is well known that the formulation of the technical problem is critical in the assessment of inventive step. AstraZeneca proposed to get rid of the notion of sustained release in the formulation of the technical problem. To some extent, this makes sense as we are not supposed to include any pointer to the solution in the formulated problem.

But the issue was that the patent at stake itself formulated the technical problem as providing a sustained release formulation of quetiapine. The reformulation proposed by AstraZeneca was thus considered self-serving and ex post by the court. To put it otherwise, by arguing that the closest prior art should be a document concerned with an immediate release quetiapine formulation, the patent proprietor attempted to focus on a starting point suggesting a different direction from the one chosen in the patent, and therefore one that would almost necessarily lead to a finding of inventive step.

The court did not agree. First, an “efficient anti-psychotic drug treatment” (i.e. the problem offered by AstraZeneca) already existed at the priority date. Second, a press release by the patentee before the priority date implicitly suggested that sustained release formulations of quetiapine were being investigated.

Therefore, the realistic technical problem was the one originally mentioned in the patent, namely the provision of a sustained release formulation of quetiapine.

With that in mind, the prior art was replete with descriptions of sustained release formulations of other drugs similar to the sustained release formulation of the EP’364 patent.

Nevertheless, the proprietor argued that clinical studies at the priority date rather oriented the skilled person towards a regimen of one or two daily doses of an immediate release formulation, possibly with an increase in dosage. They added that the skilled person was not in a position to predict:

  • Whether satisfactory bioavailability could be achieved with a sustained release formulation.
  • Whether the size of a sustained release quetiapine tablet or pill would be acceptable or too large, leading to improper absorption and travel through the gastrointestinal tract.

Based on a report authored by two pharmacology experts, AstraZeneca more specifically argued that

the travel of a sustained release drug through the gastrointestinal tract is very different since the extended delivery time results in the delivery taking place in downstream areas of the digestive system, with different pH, metabolic and bacterial environment conditions, which has an impact on the release of the active substance, its metabolism, and the permeation of the active substance through the intestinal wall. 

Sometimes things can get messy in the GI tract.
Sometimes things can get messy in the GI tract.

All of these arguments were rejected by the court, either because they were not substantiated by evidence, or because they were contradicted by some of the documents on the record:

  • The necessity to increase the dose of quetiapine when switching from an immediate release formulation to a sustained release formulation was not proven. In fact the opposite was generally true based on pharmacology textbooks. Even if a larger size tablet had to be achieved, it was also possible to divide a single dose into two doses to be taken simultaneously.
  • The half-life of quetiapine was already documented at the priority date and was compatible with a sustained release formulation.
  • The undesired effect of pH on the release of the active substance was only shown in a post published document (which would not have been taken into account by the skilled person). The extent of the undesired effect was debatable in view of the declarations of the patentee’s own expert in the British litigation. And it was known by the skilled person how to thwart this pH effect by adjusting the drug formulation.
  • The prior art taught that quetiapine bioavailability increased linearly with an immediate release formulation. It could be concluded from this finding that the disruptive effect of first pass metabolism emphasized in AstraZeneca’s submissions was at best limited and was not an obstacle to a sustained release formulation.

In other terms, there was no prejudice against a sustained release formulation of quetiapine, which the skilled person was able to achieve with routine testing. It can be derived from the judgment that the court at least partly based its findings on the British litigation, and in particular on the statements of an expert who took part in that lawsuit.

One last argument dealt with in the judgment is that clinical trials performed after the patent grant showed an improved bioavailability and broader therapeutic indications with the sustained release formulation than with the immediate release formulation. But the court held that the improved bioavailability resulted from routine optimization and was not surprising; and that the broader therapeutic indications could not be relied on as they were not addressed in the patent.

The take home message is probably that if inventive step rests on obstacles or prejudices, those have to be very fully and convincingly demonstrated. The burden of proof lies on the patentee and based on the decision at stake it is a relatively heavy one.

There are two interesting side issues in the judgment.

First, Mylan claimed that AstraZeneca was guilty of unfair competition because they had warned customers about risks of infringement related to the generic version of Xeroquel LP. The court rejected the claim because AstraZeneca’s communication was deliberate and objective.

Here are some interesting comments by the court which seem to be generally applicable to similar situations in the pharmaceutical field:

When it comes to competition between an original drug and a generic one, infringement can hardly be challenged. It disappears if the patent is held invalid, but until the invalidity judgment, it is legitimate for the patent proprietor to defend its patent monopoly. Besides, the circular letter sent to stakeholders mentioned the existence of a validity challenge in objective terms. 

A generic drug manufacturer that decides to launch its product at risk before the patent on the original drug expires or is invalidated cannot blame the patent proprietor for defending its monopoly with its customers. Besides, the fact that the patent was finally revoked in other European countries does not result in a lesser legal scope for the French part of the patent as long as it is valid. 

The second point of interest relates to the PI obtained and enforced by AstraZeneca during approximately one month. Mylan claimed and obtained damages in relation with the PI, pursuant to article L. 110-10 of the Code des procédures civiles d’exécution.

The damages award amounted to approximately 200,000 euros – which was however much less than Mylan had requested.

Part of the award was calculated based on an evaluation of lost sales, to which a margin of 30% was applied.

The other part corresponded to a loss of opportunity in relation with a public tender from which Mylan had been excluded due to the PI.

On the other hand, the court did not take into account the alleged harm suffered in terms of sales of other products or in terms of image injury, in the absence of convincing supporting evidence.

Finally, reimbursement of 300,000 euros of attorney’s fees were requested by Mylan. The court granted half of it.

It remains to be seen whether there will be an appeal decision, given that it may prove difficult for AstraZeneca to obtain it before the expiry of the patent on May 27, 2017 (no SPC was filed, as the judgment tells us).


CASE REFERENCE : Tribunal de grande instance de Paris, 3ème chambre 3ème section, July 1, 2016, Mylan v. AstraZeneca, RG No. 15/05880.

The blind date

I initially thought today’s decision raised an exciting conundrum. But after figuring it out, I must say I feel sort of disappointed. Like someone having heard of the famous curious incident of the dog in the night-time may be underwhelmed to hear that the incident was that there was in fact no incident.

Speaking of dog incidents, the decision is precisely related to serious dog incidents and how to prevent them. Indeed, the invention at stake is in the field of dog collars – the kind that contain chemicals to prevent severe parasitic diseases. The invention is protected by European patent No. EP 1022944 (EP’944), which was initially filed by Hoescht Roussel Vet and later assigned to Akzo Nobel and then Intervet International.

Two companies from the Bayer group jointly filed a nullity suit against Intervet in front of the Paris Tribunal de grande instance (TGI). The patent proprietor counterclaimed for infringement of the patent.

The patent contains three independent Swiss-type claims. The first one is worded as follows:

Use of a pyrethroid in the manufacture of a medicament in the form of a collar for the protection of dogs from bites of phlebotomine sandflies.

The other two are similar except that the use is respectively “for the control of leishmaniasis in dogs“, and “for the control of human visceral leishmaniasis“.

When prophylaxis meets fashion.
If only prophylaxis could sometimes meet fashion.

Bayer argued that the three main claims lacked novelty over an article authored by the inventors themselves. There seems to have been little debate over the fact that, in terms of contents, the article was indeed prejudicial to the novelty of the claims. But the central discussion was rather whether the article belonged to the state of the art in the first place.

The patent claims a priority date of June 18, 1997. The article was published in the issue of Medical and Veterinary Entomology of April 1997But the patent proprietor’s defense was that:

Back in 1997, the publication of this type of document took a certain time and […] contrary to what one could think in view of the journal, it was published for the first time on June 19, 1997, after the priority date of the EP’944 patent. 

Hmm, as far as I can remember, back in 1997 journals were no longer delivered by stagecoaches, but why not. Which evidence did the patentee provide? Two faxes.

The first fax was sent from the editor Blackwell Sciences to one of the inventors on January 6, 1998. The fax indicates that the article dated April 1997 was published for the first time on June 19, 1997. By the way, this is exactly the day after the priority date.

The second fax was sent from the British Library to the original applicant two days later, on January 8, 1998. The fax mentions that the issue of the journal at stake was received by the library on June 25, 1997 and made available to the public on that same date.

The court was not convinced by these messages faxed during the priority year.

The court’s appraisal of the second fax seems to make perfect sense:  the fact that one particular library (however prestigious it may be) received a copy of the journal issue on a certain date does not disprove that other recipients may have received their own copies significantly earlier.

The fate of the first fax is probably more debatable. The court noted that the recipient was one of the inventors, just like the recipient of the other fax was the applicant. These are “persons or companies having a vested interest in the patent […], which weakens the probative value of these documents“.

Now, this is tough! How were the inventors / applicants supposed to proceed in order to determine the exact date of publication of the article other than try to obtain a statement from the editor?

The court also remarked that the context of the faxes is unknown, i.e. Intervet did not explain why these faxes were requested from the editor and the library. I personally do not think that this is critical. What is important is the contents of the faxes. Besides, it can be fairly assumed that the applicant was aware of a possible novelty challenge due to the April 1997 article and thus sought to preserve evidence that the article was not prior art. This does not look particularly suspicious.

Finally and more importantly, the court stated that

This statement was not accompanied by any further evidence confirming the publication delay and was not supplemented by other documents from the edition services. 

So, one isolated piece of evidence was not sufficient for the court in view of the basic and natural presumption that a journal issue dated April 1997 was indeed published during that month.

Now, here comes the conundrum that I mentioned at the beginning of this post. The court noted in passing that

The fact that the examining division of the EPO in the context of the examination of the patent acknowledged the date of June 19, 1997 is not sufficient to convince the court. 

I was curious about this acknowledgment by the examining division, and I therefore looked at the file wrapper of the patent as well as of the priority application (which was also a European application). In both cases, the article is cited in the search report in category X (highly relevant prior art). But then, in subsequent communications from the examining division, it is indeed acknowledged that this document was published on June 19, 1997 and is thus not prior art.

I wondered whether the EPO databases contained additional information on the publication date of the article; and if such additional information existed somewhere, I wondered why the patent proprietor did not make use of the information to strengthen its case.

But upon closer inspection, what actually happened is probably less remarkable. The first document in which the statement by the examiner can be found is the international preliminary examination report (IPER) issued in connection with the PCT application on which EP’944 was based. The exchanges between the applicant and the EPO in its capacity of IPEA (international preliminary examination authority) are not accessible online but it can be assumed that the applicant filed the faxes of January 1998 and made the argument that in view in particular of the first fax, the actual publication date was June 19, 1997. The examiner was likely convinced, which led to the statement in the IPER. This statement was then later repeated in communications from the examining division, not only in the EP’944 examination proceedings, but also in the corresponding examination proceedings in connection with the priority application (which initially puzzled me).

So if my assumption is correct and if the mystery is indeed solved, it simply means that the examining division reached a different conclusion from the French court based on the same evidence – business as usual.

At any rate, the three independent claims were invalidated for lack of novelty in view of the inventor’s article, as well as dependent claims 5 and 6. Dependent claim 4 was found to lack an inventive step over the same disclosure.

The court did not stop there and committed what some professors at the CEIPI patent course like to call overkilling.

Namely, the main claims were also found to lack an inventive step irrespective of the disputed article.

The reasoning in the decision is relatively short in this respect, but it seems that a veterinary thesis published in 1991 was very relevant as it disclosed a plate impregnated with cypermethrin (a pyrethroid) as a means for preventing phlebotomine bites, in the context of leishmaniosis prophylaxis. The only difference between the claimed invention and this disclosure was the impregnation of a dog collar instead of a plate. This was seen by the judges as an obvious development since impregnated dog collars have been known for a long time.

It is also noted in the decision that the German part of EP’944 was revoked by the Bundespatentgericht in April 2015. Furthermore, a parallel lawsuit is pending in front of the Turin court, and the court-appointed expert issued a report in which the patent was considered to lack an inventive step. It seems that the French court has relied on this report to some extent in its own analysis. So, as far as one can tell from the decision, Bayer seems to be sweeping the board across Europe.

But Intervet will certainly get another shot on appeal.

Indeed, it can be seen from the European patent register that a central request for limitation of the patent was filed in July and is currently pending. One main request and two auxiliary requests were filed. Independent claims 2 and 3 were deleted and independent claim 1 was inter alia restricted to a particular compound, deltamethrin.

If the limitation is accepted by the EPO, I guess the court of appeal will have to look at the case starting from scratch…


CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre, 4ème section, April 14, 2016, SAS Bayer Healthcare & Bayer Animal Health GmbH v. Intervet International BV, RG No.14/05992.

Take the long way home

Some critics of IP French courts sometimes complain that judges have a tendency to take shortcuts and to always look for the easy way out of a case instead of dealing with all difficult questions. Today’s case may prove them wrong, as the court took the long way home instead of the short one, by combining three pieces of prior art in order to arrive at a finding of lack of inventive step – although two or even one would probably have been sufficient.

Italian company Imball Center Srl sued two subsidiaries from the same group ITM Entreprises (hereafter collectively referred to as ITM) for infringement of the French part of European patent EP 1885622 (EP’622). The patent relates to a thermal bag for food, and ITM was accused of providing infringing bags to French retailer Intermarché.

ITM filed a nullity counterclaim on the grounds of lack of novelty and lack of inventive step.

Claim 1 of the patent reads as follows:

A thermal bag made of thermally insulating composite material folded to form a bottom and welded along lateral borders to form a body of the bag with a mouth at the level of which a handle is applied, said handle having a closing device for said mouth, said composite material defining a bottom folded in an accordion-like fashion opposite the mouth of the bag, and said handle having a shorter length than the length of said mouth when the bag is in the flattened position; characterized in that a laminar stiffening element of the accordion-like bottom is disposed inside the bag. 

The lack of novelty argument was based on four thermal bags on sale before the priority date of the patent, respectively called the Intermarché orange bag, the Intermarché red bag, the Carrefour bag and the Thiriet bag.

The court accepted that all bags were prior art, relying on a variety of evidence: catalogues, affidavits, orders, etc. Interestingly, indirect evidence was also relied on in order to prove the date of some bags, such as the presence of an old logo on a bag, or the presence of a phone number in a format no longer in use for years.

However, the court held that the thermal bags in question did not have a bottom folded in an accordion-like fashion, as required by the main claim of the patent. Since the judgment is relatively concise on this aspect, it is not entirely clear whether the prior art bags clearly had a different type of bottom, or whether there was simply not enough evidence that the prior art bags had a bottom as claimed. Anyway, as already explained in a previous post, providing a perfect demonstration in the context of a public prior use is often extremely challenging.

Concerning inventive step – which turned out to be the patent’s killer – the court first noted that the parties agreed on the fact that a so-called Link patent was the closest prior art.

The claimed invention seems to differ from the Link patent due to (i) the bottom folded in an accordion-like fashion, as well as (ii) the presence of the laminar stiffening element.

The court then turned to another reference, called the Moravek patent, which disclosed distinguishing feature (ii), i.e. the laminar stiffening element. The court used a problem and solution-type of reasoning as regards the combination of Link and Moravek:

[…] Although the invention of the Moravek patent differs from the one proposed by the EP’622 patent in that it does not disclose an insulated bag made of a thermally insulating composite material and does not disclose a bottom folded in an accordion-like fashion […], the skilled person, confronted with the technical difficulty of improving the convenience of the insulated bag, would be enticed to refer to the Moravek patent, the objective of which is precisely to allow the support of the bag’s opening during the introduction of produce into the bag on a sales counter, with a strengthening panel mounted on a pivot which is lowered when the bag is opened owing to air pressure. 

OK, and what about the accordion-like bottom? Well the court said the skilled person would turn to the bags on sale before the priority date, and it particular to the Intermarché red bag. But wait a minute, didn’t the court previously state that these bags did not disclose the accordion-like folded bottom? Yes indeed, but at this point the court added that the Intermarché red bag had a folded bottom which was strictly equivalent to an accordion-like one:

[…] It is a fact that the Intermarché red bag comprises folds at the bottom, in order to form this bottom, although the configuration of the folds is not strictly identical to that provided in EP’622. However, although this difference results in the novelty of the invention as stated above, it is not sufficient in itself to characterize an inventive step for the skilled person whereas the purpose of the feature is identical and it provides the same effects. 

[…] It should be noted that the red Intermarché bag, as it is designed, can perfectly be folded into an accordion shape, so that this feature, although it is the result of an additional manual operation as set out by Imball, is reproduced by the bag. Therefore, the skilled person reading the prior art, with common general knowledge, was able to achieve this result without any inventive step by simply looking at and handling products on the market. 

I started this post by saying that the court had taken the long way rather than the short one. Actually, it seems to me that it would have been much easier to start from the Intermarché red bag as the closest prior art, and to simply state that the sole distinguishing feature of the claimed invention was the shape of the folds at the bottom of the bag, which was obvious to achieve for the skilled person using his or her basic skills and common general knowledge. The way I understand the case, Link and Moravek were not necessary for the reasoning.

As a reminder, French courts are strictly bound by the submissions of the parties, and may not really come up with a reasoning of their own. In this case, it seems that the Link patent was considered as the closest prior art by both parties. So it may well be the defendant who actually took the long way – maybe out of fear of contradicting their novelty challenge, but I am just speculating.

The shortest way?
The shortest way?

Now, two other notable points in this judgment.

First, the court also revoked a number of dependent claims, also for lack of inventive step. For most of those, it seems that the Intermarché prior use already disclosed the claimed features – although the court also referred to further prior art references such as a Martineau patent or a Pakzad patent, in a superfluous manner.

Second, although the patent infringement claim obviously failed due to the revocation of the patent, the court did however find the defendant guilty of unfair competition and awarded 500,000 euros of damages to Imball.

The facts were the following: until February 2012, Imball supplied insulated bags according to the EP’622 patent to ITM. Then ITM stopped the business relationship with Imball and started buying bags from a Malaysian company, L&A Packaging. The Malaysian bags marketed in France by ITM were very similar to Imball’s own patented bags. The court also noted that ITM had the complete specifications of Imball’s bags in its hands and had even asked Imball for samples a few months before the end of the supply relationship, for evaluation and testing purposes – which, in retrospect, may look a little bit like a smoking gun.

The court held:

Therefore the defendant had all the information on the characteristics of the bags made by Imball, and decided, simultaneously with the end of the business relationship with Imball, to get supplies from L&A Packaging so as to market bags at a lower cost. They were aware that said bags were a quasi-literal reproduction of the bags produced by Imball. The manufacturer was necessarily able to benefit from Imball’s know how. This characterizes a disloyal attitude, because it led to knowlingly put a large quantity of quasi-identical products on the market resulting in an obvious confusion with Imball’s products. 

In this respect, the mere fact that the bags sold by the defendants have an Intermarché logo is not enough to avoid the risk of confusion. The risk of confusion must be assessed, on the one hand, from the perspective of the public at stake, i.e. in this case professionals in the field, such as those purchasing bags to market them in their distribution network, and not consumers; and on the other hand in view of the technical similarity of the products, which is established in this case, independently of the logo with which they are effectively marketed. 

Most patent infringement complaints comprise an ancillary unfair competition claim. The latter very often fails if the main infringement claim is successful, as the facts underlying both claims are generally similar, whereas case law demands different facts. Nevertheless, an unfair competition claim can be a valuable fallback claim if the infringement claim fails, and if some actual copying has taken place.

Interestingly, in today’s case the evidence for the acts of unfair competition was obtained during a patent infringement seizure. I am always somewhat surprised that this procedure, which is supposed to be exceptional and therefore strictly limited by its purpose (the demonstration of an alleged patent infringement), can be later used to support a different claim.


CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre, 2ème section, March 11, 2016, Imball Center v. Consommables Et Matériels & ITM Alimentaire International, RG No. 13/12818.