Time goes by so slowly

The U.S. used to have submarine patents staying under water as pending applications for many years without getting published and suddenly emerging upon grant and awarding 17 years of protection to their owners.

We in France have had a similarly exotic and frightening species of IP rights, that one could call zombie patents:  patents that have apparently been dead for years but that are in fact quietly waiting in their graves for an awakening – which can be triggered e.g. by a decision from the highest court in the country.

European patent No. EP 0984773 to Astrazeneca AB, directed to a pharmaceutical formulation of omeprazole, was such a zombie patent. The patent was filed in 1998 and granted in 2003, and the French translation of the granted patent was duly filed at the Institut National de la Propriété Industrielle (INPI) as was required in those pre-London days. Unfortunately, the sixth renewal fee which was due in 2003 was not paid in due time – nor within the 6-month grace period.

As a result, a formal decision of lapse was issued by the INPI on January 30, 2004 and a mention of this decision was published in the official bulletin (Bulletin Officiel de la Propriété Industrielle or BOPI) on February 27, 2004.

Several pharmaceutical companies apparently relied on this lapse and started marketing generic drugs, such as Ethypharm, which launched its formulation in April 2004.

Five years went by, until one day in April 2009 Astrazeneca filed a request for restoration of right at the INPI. One would think that the request would be expediently dealt with but, as is often the case, one would be wrong. It took the INPI four years to issue its decision on the case, which was… positive. And therefore, in 2013, the patent woke up from the dead.

Obviously Ethypharm was not too happy about this turn of events and filed an appeal against the decision. The Cour d’appel de Paris had sympathy for the generic manufacturer’s case and canceled the decision from the INPI in a judgment dated January 15, 2014 – a swifter process than the proceedings in front of the INPI.

Quite predictably, Astrazeneca filed an appeal on points of law with the Cour de cassation, which brings us to the ruling issued on April 12, 2016, in which the highest court set aside the judgment of the Cour d’appel and therefore reinstated the 2013 INPI decision – and thus revived the patent for the two additional years that are left until it expires.

At this stage there are two questions any puzzled reader would be entitled to ask:

  • Why such a mess?
  • How can a patent possibly be reinstated by a request for restoration of right filed five years after the loss of right occurred?

I think there are two main factors which can help answer these questions:

Let’s first look at the old pre-2008 statute. At that time, the restoration of right further to a lapse due to non-payment of a renewal fee was handled differently from a restoration of right further to non-compliance with another type of time limit.

The former version of article L.613-22 CPI (Code de la propriété intellectuelle) provided in its second paragraph that:

The patent proprietor can file a request for restoration of right within three months from the notification of the decision [of lapse] if it had a legitimate excuse for not paying the renewal fee. 

In the omeprazole case, the decision of lapse having been issued in January 2004, the three month-deadline should have been long expired by the time the request for restoration was filed, and also by the time the law changed in 2008, right?

Well, no, and this is because of the pravastatine case law. It was decided in this famous case (which involved a zombie SPC) that when a decision of lapse is notified to the wrong recipient (i.e. to a recipient who was not entitled to receive the decision), the time limit of article L. 613-22 does not start. Since it does not start, it cannot expire.

This is exactly what happened in the present case. The decision of lapse was not notified to the patent proprietor in Finland. It was not notified to a representative of the patent proprietor either. Instead, I understand from the decision of the Cour de cassation, or rather from the statement of grounds of appeal annexed to the decision, that the 2004 decision of lapse was notified to… the annuity provider:

[…] The director of the INPI states that the notification was made to Mrs Caroline C, who is not an employee of Astrazeneca AB, nor a patent attorney, nor an attorney at law, nor a representative registered with the INPI, and therefore this notification was not made to a person who would be entitled due to the annuity provider, RWS group. Thus, this illegal notification did not trigger the three-month time limit for the restoration, and the request filed by Astrazeneca AB was not outside the time limit. 

OK. But what about the change of law which took place in 2008? How does this come into play?

By way of an executive order dated December 13, 2008, the statute was amended in such a way that paragraph 2 of article L.613-22 CPI was canceled. As a result, the system of restoration of right further to a lapse was aligned with the normal system of restoration of right further to non-compliance with other types of deadlines.

The new relevant provision is article L.612-16 CPI, which will look familiar to European practitioners as it is similarly drafted as the restitutio in integrum provisions in the EPC:

The applicant who does not comply with a time limit set by the [INPI] may file a request for restoration of right if they have a legitimate excuse and if the failure to comply with the time limit directly results in the refusal of the patent application or of a request, in the lapse of the application or the patent or in the loss of any other right. 

The request must be filed with the director of the [INPI] within a deadline of two months from the removal of the cause of non-compliance. The omitted act shall be accomplished within this deadline. The request is only admissible within a deadline of one year from the expiry of the unobserved time limit. 

When the request relates to a failure of payment of a renewal fee, the unobserved time limit is the end of the grace period […].

So, in the new system, the deadline of three months from the notification of the decision of lapse has been replaced by the usual double deadline: two months from the removal of the cause of non-compliance and one year from the expiry of the unobserved time limit.

Since this second deadline is an absolute time bar and is independent from any actions taken or not taken by the patent proprietor and the patent office, it seems that there can no longer be zombie patents in this new system. If this provision had been in place before, Astrazeneca would have been barred from filing a request for restoration of right on the first of December, 2004 (one year after the six-month grace period following the due date for the sixth renewal fee, on the first of June, 2003).

The illustrated device is useful for store clerks: it displays the minimum birth date for an individual who is entitled to purchase alcohol or tobacco. I suggest a variant for patent office clerks displaying the maximum date of lapse of a patent for a patentee who is entitled to restoration of right.

The illustrated device is useful for store clerks: it displays the minimum birth date for an individual who is entitled to purchase alcohol or tobacco. I suggest a variant for patent office clerks displaying the maximum date of lapse of a patent for a patentee who is entitled to restoration of right.

The new provision came into force a few months before Astrazeneca filed its request for restoration. How did this affect the request? As the Cour de cassation summarizes it:

[…] in order to set aside the decision [to restore the patent], the judgment [by the Cour d’appel] stated that the executive order […] entered into force on December 2008 and deleted the second paragraph of article L.613-22 CPI so as to replace it with article L.612-16; and that this was a procedural law and thus was immediately applicable; that Astrazeneca was subjected, in order to benefit from a restoration of right, to the time limit of two months from the removal of the cause of non-compliance in the new provision; that the publication in the BOPI of the decision [of lapse] removed the circumstance which had prevented Astrazeneca to pay the sixth renewal fee when the new article L.612-16 came into force, so that the request filed on April 6, 2009 was late and inadmissible. 

So, in summary, for the Cour d’appel, the time limit of two months came into force with the new law, and since the cause of non-compliance had long been removed, this deadline started on the day the new provision came into force, and was expired by the time Astrazeneca filed its request – so that the restoration could not be granted.

On appeal in front of the Cour de cassation, Astrazeneca criticized the computation of the two-month deadline. They said that a law cannot have a retroactive effect, and that the judgment by the Cour d’appel amounted to such a retroactive effect because the publication of the mention of the decision of lapse in the Bulletin in 2004 retroactively acquired a legal effect in 2008. Also, they explained that this publication was not an actual removal of cause of non-compliance.

But very interestingly, the Cour de cassation set aside the judgment on another ground, that the court came up with on its own – I am not sure I have ever seen this before. Namely, the court ruled that the former pre-2008 rule in fact still applied to the request for restoration filed in 2009. Said the supreme court:

[…] legal remedies against a decision are determined by the law which is in force on the day the decision is issued; 

[…] the decision of lapse of patent right was open to restoration of right under article L.613-22 [CPI] second paragraph, within a time limit of three months from the notification, and since the notification was illegal, the deadline did not start running, and thus the Cour d’appel violated the law […]. 

As a side remark, Ethypharm was not the only third party involved in this dispute. Parallel cases involved Actavis, and therefore there are two other judgments by the Cour de cassation of the same date with substantially the same content.

All in all, this is a rather spectacular result which does seem to make sense on a legal standpoint. The real aberration is probably the error made by the INPI. Why on earth a decision of lapse was notified to someone else than the representative is puzzling – and I guess there must have been a representative actually appointed since the patent was validated by filing a French translation.

Fortunately this kind of fiasco would no longer be possible nowadays – or could it?


CASE REFERENCE: Cour de cassation, ch. com., April 12, 2016, Astrazeneca AB v. Ethypharm & Directeur général de l’INPI, pourvoi No. V 14-17.439.

Search for meaning

Today’s Board of appeal decision was already reported on elsewhere, but it may be worth another commentary, since it relates to a fundamental aspect of patent law, namely the rules of claim interpretation, and more specifically the question of how much one should rely on the description and drawings when interpreting the claims – in particular for the assessment of novelty and inventive step.

One could think that such a basic issue should have long been very clearly settled. Well, one could think again.

Art. 69(1) EPC provides that “the extent of the protection conferred by
a European patent or a European patent application shall be determined by the claims. Nevertheless, the description and drawings shall be used to interpret the claims“.

The official Case Law book (Case Law of the Boards of Appeal of the European Patent Office, 7th edition, section II.A.6) reminds us that a number of decisions referred to Art. 69 as a legal basis for interpreting the claims in the light of the description and drawings. Others pointed out that Art. 69 is concerned with the question of infringement, and is therefore only for courts of law to deal with, and not for departments of the EPO (except when assessing a potential extension of scope of protection after grant – Art.123(3) EPC).

For example, in T 1279/04, the Board explicitly stated that a different standard should be applied for interpreting the claims in examination or opposition proceedings and in infringement proceedings – which is somewhat reminiscent of the U.S. approach. As the Case Law book puts it, according to this decision:

in examination and opposition proceedings the value of future legal certainty was paramount. […] There was no case for anything other than a strict definitional approach, given that in this procedural stage the claim could and should be amended to ensure legally certain patentability, in particular novelty and inventive step over any known prior art. Amendment rather than protracted argument should be the answer to genuine difficulties of interpretation in all aspects of the examination and opposition procedure.

The strict approach advocated in T 1279/04 does not strike me as being commonly applied.

The position that first instance departments are officially supposed to follow (Guidelines for examination, F-IV, 4.2) seems somewhat more flexible. It refers to the description and drawings in the context of “particular cases“, where “the description gives the words a special meaning“. Otherwise, reference is more generally made to the meaning that words “normally have in the relevant art“, and to the necessity of attempting “to make technical sense out of” the claim.

And in practice, I feel that most of the time parties as well as the EPO tend to turn first and foremost to the description when interpreting the claims – and not just in particular cases when there is a special definition in the spec.

So, back to today’s decision T 1871/09, which sheds another interesting light on claim interpretation in view of the description.

The Board provided extensive general remarks on claim interpretation, and then applied their approach to the case at hand, where several terms raised interpretation issues.

A first important point is that the Board followed the approach adopted in T 556/02, per which reliance on the description and drawings to interpret the claims is justified by a general principle of law:

The Board was confronted with diverging argumentations from the parties regarding the meaning that the claims should have, and in particular some concepts that they refer to. The Board endeavored to look at the specification of the patent taken as a whole for the meaning that the skilled person would give to them. The Board applied a general principle of interpretation, of which Article 69 EPC is only an illustration, per which a part of a document cannot be interpreted independently from its context, and to the contrary the entirety of the document should be taken into consideration, when looking for the meaning of a certain statement which is made. […] Therefore, even if the provisions of Article 69 EPC on the scope of protection do not […] apply to departments ruling on opposition cases, the principle set out in this article is nevertheless applicable (reasons 3.1 – this passage and the following ones were freely translated from the French original version).

This makes sense and is a nice way to justify what most people would intuitively think is right – even if not explicitly enshrined in the Convention – i.e.: do not read the claims in isolation without considering the rest of the patent.

In the next paragraph, the Board recalled that the reader is the skilled person, and that he/she is the one supposed to interpret the claims. This is a very classical statement.

Then, the Board went on to warn that:

The interpretation principle recalled here applies to all, that is not only the opponents but also the patent proprietor, and third parties as well. The patent proprietor cannot possibly attribute a meaning to the terms that are used or to the features recited in a claim which the general context does not really make it possible to establish. It is thus not justified in this respect to rely on one passage of the description instead of another, so as to give a particular color to some terms that are used (reasons, 3.3).

So, patentees have to deal with their description as a whole. They cannot rely on just one paragraph which may support their preferred interpretation, without looking at other paragraphs which provide a different view.

And then comes the last bit of the general remarks, which is probably the most interesting one:

That said, the specificity of a patent specification should not be forgotten, that is the claims are supposed to generalize particular embodiments effectively disclosed in the patent specification. To which extent this specificity plays a role in the interpretation depends on the case. In this respect, some aspects such as the purpose of the invention and the number of examples are especially relevant. But it remains that the terms chosen in the claims are supposed to have been selected to serve this objective of generalization of particular embodiments. As a consequence, when the patentee omits, voluntarily or not, to define some concepts, or accepts that certain ambiguities remain in the patent description related to the request at stake, they cannot validly hide behind a limitative interpretation of the terms of the claim, at least insofar as the general interpretation which is retained makes sense on the technical standpoint and is consistent with the general teaching of the patent (reasons, 3.4)

Oftentimes in opposition proceedings, a patent proprietor would like a narrow interpretation of the claims to be used – in order to resist novelty and inventive step challenges.

One important consideration may however play against this defense: by nature and by function, claims are supposed to be more general than the embodiments disclosed in the description. Therefore, the fact that the description offers a narrow illustration of the claims may not by itself validly support the narrow interpretation of the claims. Or, to put it otherwise, the scope of the claims cannot be assumed to be limited by the “scope” of the description.

These principles were applied by the Board to the case at hand a number of times.

Claim 1 of the main request (patent as granted) read as follows:

Spectral content enrichment process of a signal having an incomplete spectrum including a first spectral band, the process consisting of the following stages:
– at least one transposition of the spectral content of the said first band into a second spectral band not included in the said spectrum in order to generate a transposed spectrum signal of restricted spectrum, to the said second spectral band;
– transformation of the spectrum of the signal with the transposed spectrum to obtain an enrichment signal;
– combination of the incomplete spectral signal and of the enrichment signal to produce an enriched spectral signal;
characterized in that the generation of the transposed spectral signal includes a whitening stage of the said spectral content such that the spectrum of the signal to the transposed spectrum is a whitened version of the said spectral content.

A device with spectral bands.
A device with spectral bands.

Novelty of this claim over a prior art document D1 was discussed. The two patent proprietors argued that the document did not disclose a process wherein the second spectral band is “not included” in the spectrum.

Remarkably, the Board rejected both the interpretation of the term “included” offered by the patent proprietors and the one offered by one of the opponents. The opponent had put forward a mathematical definition of the term. But the Board held that this “turns out to be too theoretical and does not take into account the content of the specification of the patent, where nothing suggests that these usual terms should be given a particular mathematical meaning” (reasons, 5.3.1).

As for the patent proprietors, their position was that the first and second spectral bands should necessarily be adjacent or separated. The argument was based on two embodiments in the description corresponding to these two possibilities. But the Board decided that the claim could not be interpreted in such a restrictive manner:

The passages of the patent relating to these embodiments are very general and do absolutely not suggest that the disclosed solutions are the only ones which can be contemplated. Therefore, in the absence of indications in the patent as to the meaning which the concept of inclusion should have, and taking into account the fact that claims are meant to generalize the teaching of the disclosed embodiments, the Board considers that the expression “not included” should be given a general meaning. The formulation used can thus not be interpreted as excluding the partial superposition of spectra of the initial and transposed signals (reasons 5.3.1). 

A second point of interpretation concerned the concept of “whitening“, which had to be compared with the notion of sub-band energy equalization in the prior art. The patent proprietors submitted that “whitening” the signal meant making the peaks of the spectral envelope more uniform, which was different from sub-band energy equalization.

Even though the other parties did not directly criticize this definition of whitening, the Board was not fully satisfied, because the definition “is not devoid of ambiguity” (reasons, 5.3.2). Then followed a detailed technical discussion showing that since the notion of spectral envelope is not clearly defined, there are many different ways to whiten the signal according to this broad definition. One of the opponents had filed an example showing that a sub-band energy equalization was tantamount to one possible form of whitening, and the Board was convinced by the demonstration.

One important last argument was discussed in this respect. Claim 1 called for one step of transposition and one step of transformation, the whitening being part of the transposition step, and being therefore before the transformation step. But in the prior art document, the whitening discussed above took place at the same time as the transformation. So, on the face of it, there was a difference between the process of claim 1 and that of the prior art.

But the Board decided that claim 1 should be interpreted more broadly than its literal wording, in view of the description:

When the granted patent is taken into account, this however leads to a generalization of the claimed process beyond its literal interpretation. Indeed, paragraph [0033] of the patent mentions the possibility of whitening and filtering in one operation by a transfer function filter equal to the product of the respective transfer functions of the whitening filter and of the transformation filter. It is unfortunate that the version of the description was not adapted to the version of the claims considered patentable by the Examining division and results in what is above all a problem of clarity of the claimed process. The version of the patent is however the entire responsibility of the appellant. Although no clarity objection […] can be raised against the granted version of the patent […], the patent proprietors must nevertheless accept that any contradiction, ambiguity or imprecision may be used by the opponents who can in particular rely on the extended scope of the claims made possible by such circumstances. In the present case, the argument that D1 reproduces the process of claim 1, as interpreted in the light of the description, is thus justified. 

All in all, this decision is actually a serious warning to applicants, who must carefully check whether the description is consistent with the claims, bearing in mind that:

  • On the one hand, the fact that the description discloses specific embodiments cannot by itself justify a narrow interpretation of the claims, since the function of the claims is such that they are supposed to be broader than these embodiments.
  • On the other hand, if the invention appears to be somewhat broader than the claims literally imply, based on the description (for instance because the description was not properly adapted to amended claims), then the claims may need to be reinterpreted in a broader manner.

Readers interested in the outcome of the case will have to wait a few more years, since the Board remitted the case to the first instance for further prosecution on the basis of an auxiliary request, after admitting new relevant documents into the proceedings.

As a final remark, there has been a long-standing tradition for French courts to rely a lot of the description and drawings to come up with their own understanding of what the invention really is about – irrespective of the exact wording used in the claims.

Does the French nationality of the patent proprietors explain why such a large part of the legal discussion in this case revolved around claim interpretation in view of the description? Possibly, but it seems that the description did little to help them in Munich after all.


CASE REFERENCE: Board of Appeal 3.4.01, T 1871/09, Orange et al. v. Stefanie Kremer et al., November 24, 2015.

Three strikes – you’re out

One, two, three strikes you’re out. Today, I will be discussing an invalidity decision again, and a rather remarkable one. It is not every day that three different grounds of nullity are held against a patent, including what some practitioners could consider as their white whale, namely… the lack of industrial application.

The origin of the case is a dispute between two inventors owning a French patent and a corresponding European patent relating to a cloth coated with mother-of-pearl, and their licensee, Van Robaeys Frères. As part of the dispute, the patent proprietors in particular requested from the Dunkerque Tribunal de grande instance (TGI) that the license agreement should be terminated. The disgruntled licensee retaliated by filing a nullity suit in front of the Paris TGI – which, as explained in a recent post, has exclusive jurisdiction in patent matters.

The first part of the judgment is dedicated to an inadmissibility defense based on an alleged lack of standing. This has actually been a serious defense since a famous Barilla case three years ago, in which a nullity claimant was found to lack standing because they had not demonstrated that they were preparing for carrying out acts prohibited by the patent.

Yet, in the present case, the nullity claim was quite logically found to be admissible:

With respect to patent nullity actions, those who can act are those who can establish that, at the time the complaint is filed, the claims which are sought to be revoked, are or may be an impediment for carrying out their economic activity, because they work or plan to work in the field of the patented invention. In this case, as a licensee in a license agreement on the exploitation of the French patent for which a termination is requested by the licensor, Van Robaeys Frères has an economic activity notably in the field of linen cloths, and thus has standing for seeking nullity of the patents of Mr. Thierry D’Arras & Albert Paoli, which it will no longer benefit from due to the termination. 

That being settled, let’s have a quick look at the two patents at stake, FR 2941712 (FR’712) and EP 2393979 (EP’979). Claim 1 of EP’979 reads very simply:

A cloth comprising a support and a light deconstructing layer characterized in that said layer comprises mother-of-pearl.

Claims 2 to 8 are dependent claims. Claim 9 relates to:

A method for preparing a cloth according to one of claims 1 to 8, comprising a step to impregnate a support in a mother-of-pearl solution.

Claim 10 depends on claim 9, and claim 11 is directed to

The use of a cloth according to one of claims 1 to 8 or of a cloth obtained with the method according to claims 9 to 10 to manufacture products which reflect away the infrared such as blind fabrics, tent fabrics and clothing.

The claims of FR’712 are identical to those of EP’979, except that claim 1 of FR’712 is somewhat broader as it also recites the alternative of using a substance equivalent to mother-of-pearl.

Mother and her pearls
Mother and her pearls

The first ground of nullity upheld by the court was a lack of novelty due to the inventors’ own activities before the priority date. More precisely, Van Robaeys argued that the invention was disclosed to the public in two ways: because of trials at the Centre européen du non-tissé (CENT) a couple of months before the priority date, and because of a meeting at the Ecole polytechnique a couple of weeks before the priority date.

The inventors’ defense was that both events were confidential. But the court held that this was not adequately proven. In particular, it seems that the court was of the opinion that explicit confidentiality agreements should have been in place, which was apparently not the case:

However, it can only be derived from the exchange of emails concerning the 2008 trials that a draft of confidentiality agreement was received by [the inventors]; and concerning the [meeting] of January 2009, it can be derived from the report dated January 19, 2009 on the meeting at Polytechnique that it was intended that all documents would be confidential, that “the university insists on the execution of a confidentiality agreement”, that “a draft of confidentiality agreement” was received by Mr. … from Polytechnique, but no confidentiality agreement was submitted by the parties. 

Yet, it is conventional to execute an agreement of this type […]. 

The court did not investigate whether it could be considered that there was an implicit obligation of confidentiality due to the circumstances of the two disclosures – and it is possible that the patent proprietors did not phrase the argument in this manner.

This is a defense that might have been successful in front of a Board of appeal of the EPO. There are not enough details in the judgment to really understand the full circumstances of the disclosures, but one could probably argue that there was a common understanding by the participants that the information made available during the trials and the meeting was provided in confidence.

So, are French courts more severe than the EPO on this issue? That’s very possible, as this is not the first decision that I have seen where an explicit confidentiality agreement is viewed as necessary in order to disqualify a disclosure as a public one. On the other hand, another decision issued by another section of the Paris TGI earlier in 2015 (and that I might comment on in a further post) accepted that the supply of prototypes to a lab for testing purposes implies a duty of confidentiality for the lab.

The bottom line is that, even if French courts may not draw the line at exactly the same position as the Boards of appeal of the EPO, in the end the question of whether a disclosure is considered as a confidential one or a public one is extremely fact-dependent, and there are never two exactly similar situations.

With that, claims 1 to 5 and 11 were found to lack novelty, which left claims 6 to 10 still standing.

Lack of inventive step was the next ground of invalidity tackled by the court, and it was quickly discarded. Indeed, it seems that none of the documents relied on by the claimant disclosed the use of mother-of-pearl on a cloth for filtering infrared and ultraviolet rays without interfering with visible light, which was the technical problem at stake.

Thus, the court turned to insufficiency of disclosure, which as readers may know is not a ground of invalidity that patent proprietors should take lightly in this country.

At this point, the judgment gets very surprising.

The court started by examining claim 6, which recites that “the light deconstructing layer is continuous“. According to the court, the description sets out that a continuous layer can be obtained by dipping or spraying, so that there is no insufficiency issue.

But then, the court turned to method claims 9 and 10 and explained that the notion of “mother-of-pearl solution” recited in these claims raises significant difficulties:

It is only mentioned in claim 10 that the mother-of-pearl solution comprises 10 to 20% mother-of-pearl, and it is mentioned in the description […] that “said mother-of-pearl is a powder of mother-of-pearl. Among the powders of mother-of-pearl which can be used in the invention, the one obtained by grinding the inside of mollusk shells can be cited.” 

Yet, the patent does not provide any indication on the formulation of the mother-of-pearl solution, although this element is essential since it defines the composition of the “layer” which needs to be homogeneous during preparation, its condition, its ability to penetrate or not the support or its ability to adhere to said support. 

Thus, it can be derived from the opinion drafted by Prof. D. […] that “the possibility to implement the invention depends on the possibility to implement the method of claims 9 and 10 […]. On the chemical standpoint, the term ‘solution’ is not sufficient to make it possible to carry out the invention and it raises many questions. Mother-of-pearl is mainly formed of calcium carbonate in the aragonite crystalline form, and it is very poorly soluble in water. One needs to know exactly what the solution is made of.”

The technical opinion of the engineer Mr. V. […] is consistent with this and adds that the composition the solution should probably vary depending on the fiber which is used […]. 

Therefore, the court held that method claims 9 and 10 are invalid due to insufficiency of disclosure and then added that “the product of claims 6 to 8 is also invalid because the methods for making the cloth are invalid“.

To say the least, this wording is clumsy. After stating that there was no defect of insufficiency of disclosure in claim 6, the court sort of changed its mind and held that the claim was invalid because the method of claims 9 and 10 is insufficiently disclosed.

It would certainly have been much clearer to start with claims 9 and 10, and then explain that since the skilled person does not know how to carry out the claimed manufacturing method, and in the absence of another readily available method, the product claims necessarily suffer from the same defect.

This clumsy wording notwithstanding, the court’s reasoning seems to make perfect sense. The analysis relies on experts’ opinions and it is true that the description of both patents is quite short and does not provide very detailed information, and in particular no example – in contrast to another recent case discussed here.

And finally, the white whale, i.e. the court’s position on the lack of industrial application of the patent:

It can be derived from the exchanges of emails in June 2011 between the parties as well as from the affidavit of Mr. J in charge of the project […] that the inventors faced unresolved difficulties of implementation of the non-woven linen cloth awnings coated with mother-of-pearl solution in a continuous layer, and it is not demonstrated that these difficulties only concerned roller awnings [as alleged by the defendants]. 

The inventors claimed that products according to the patent were manufactured and marketed, but only photographs and invoices were provided as evidence, which did not make it possible for the court to determine whether the invention was actually implemented in those products or not.

This lack of industrial application is in fact very similar to an insufficiency of disclosure. So, the whale may not be so white after all, upon closer inspection. But it is true that if an invention cannot be carried out by the skilled person, then it is not possible to make it or use it “in any kind of industry, including agriculture” as Article 57 EPC puts it.

And that was the last nail in the patents’ coffin.

As a final word, we should always bear in mind that French court are strictly bound by the parties’ submissions; therefore, surprising court’s findings may sometimes only reflect unusual parties’ submissions. But to some extent, it seems like this judgment is really a pearl.


CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre 1ère section, September 10, 2015, SA Van Robaeys Frères v. D’Arras & Paoli, RG No. 13/12618.

Lyon’s share

It has been almost seven years now since patent litigation in France has been fully centralized in Paris. Patent cases used to be heard in various Tribunaux de Grande Instance (TGIs) in the entire country. The number of TGIs having jurisdiction for patent matters was progressively reduced and finally all cases were consolidated in Paris.

Similarly, at the appeal level, the Cour d’appel de Paris has now exclusive jurisdiction in all patent cases. We can still see from time to time decisions issued by other Cours d’appel in relation with legal actions started a number of years ago before the centralization was completed.

But new cases do not normally end up in a court outside of Paris, save if the plaintiff makes the unfortunate mistake of filing its writ of summons with the wrong court.

The present case is an exception, but certainly not one due to an ill-informed litigant. On the contrary, this was certainly the result of a sophisticated (although, in the end, unsuccessful) strategy by Swedish pharmaceutical company AstraZeneca AB.

AstraZeneca owns European patent No. EP 0907364 covering the active substance quietapine for treating schizophrenia, and markets a corresponding pharmaceutical product called Xeroquel.

Generic company Mylan filed a nullity action in connection with the French part of the patent on April 28, 2015, in front of the Paris TGI – which, as recalled below, has exclusive jurisdiction for patent matters, including nullity actions of course. Apparently, another nullity action had already been started earlier in 2014 by Accord, and Mylan filed an intervention so as to be part of that nullity action. How Mylan can be part of two different actions having the same purpose is an interesting question which is however not addressed in the decision that I am reporting on today.

Mylan launched its generic version of Xeroquel at risk, without waiting for the outcome of the nullity action. Accordingly, AstraZeneca filed a counterclaim for infringement of the French part of the European patent on June 9, 2015, still in front of the Paris TGI.

Now comes the interesting part.

AstraZeneca attempted to obtain interim relief against Mylan. The usual way to do this, if an infringement claim is already pending in front of the Paris court, is to file a motion with the judge in charge of case management, who then decides on the motion pending the full trial on the merits. But AstraZeneca did not take the usual path.

Instead, AstraZeneca filed a motion for interim relief in front of the “juge des référés“, i.e. the judge in charge of urgency proceedings, at… the Tribunal de commerce (sometimes nicknamed T. com.) in Lyon. In this motion, AstraZeneca did not submit a patent infringement claim, but rather a claim for unfair competition.

Why apply to the juge des référés of the T. com.? Because, AstraZeneca stated, this is the judge having general jurisdiction for granting interim relief in situations of unfair competition. But why a judge in Lyon? Because, the defendant having its seat close to this city, this is the judge who had territorial jurisdiction – still in keeping with general, non-patent rules of judicial organization.

The Lyon judge accepted to hear the case and issued an order prohibiting Mylan from marketing the quietapine generic drug until the expiry of the French part of the European patent, i.e. May 27, 2017, or until a final decision of revocation is issued; a recall of the products was also ordered, as well as a publication of the decision.

In summary, AstraZeneca obtained the most positive result that they could ever have hoped for.

The order was issued on November 12, 2015. Mylan reacted by immediately filing an appeal and by requesting accelerated appeal proceedings. The entire process was, I must say, particularly fast and efficient, since the decision of the Cour d’appel de Lyon on this matter is dated December 17, 2015 – so the complete appeal process took only one month. The Cour d’appel set aside the first instance order and declared that the motion filed in front of the juge des référés in Lyon was in fact inadmissible because the judge had no jurisdiction.

A word must probably be said here of the differences between a TGI and a T. Com.:

  • TGIs are general courts dealing with civil matters. T. com.s, on the other hand are specialized in handling disputes between commercial companies (or persons).
  • TGIs are exclusively composed of government-appointed judges, while T. com.s are composed of elected judges.
  • Accordingly, TGI judges are professional judges, who are civil servants, while judges in T. com.s are company managers.

I have heard some lawyers familiar with T. com.s say that these courts are somewhat unpredictable and can sometimes render pretty wild decisions. What is almost certain is that the juge des référés was not fully aware of the subtleties of IP law, which is very far from his or her everyday practice.

Judicial organization in France shown on a simplified diagram.
Judicial organization in France shown on a simplified diagram.

So, again, it was a bold and clever move for AstraZeneca to request the preliminary injunction far from the Ile de la Cité in Paris. The judge in charge of case management at the TGI would likely have made a more severe assessment of the patentee’s case – not least because the European patent was already revoked in six other countries, including Germany, the UK and the Netherlands, as we learn in the Cour d’appel decision.

Now, going back to the issue of jurisdiction, article L.615-17 of the Code de la propriété intellectuelle provides that

Civil actions and motions relating to utility patents, including […] when they also concern a connected issue of unfair competition, are exclusively heard by [TGIs] mentioned in a decree […].

The decree in question names the Paris TGI as the one having exclusive jurisdiction for the purposes of article L.615-17.

Therefore the central issue, which was revisited by the Cour d’appel, is to which extent the claim brought by AstraZeneca in Lyon was related to a patent, or a connected issue of unfair competition.

The answer to this question prima facie seems pretty obvious, so how did the patent proprietor defend its position?

Again, they were rather creative, in that they stayed away from the Code de la propriété intellectuelle (which contains the definition of patent infringement), and rather argued that Mylan had committed other torts, and more specifically:

  • violated articles L.4113-6, L.4163-2, L.5122-10 and R.5122-17 of the Code de la santé publique, article 1382 of the Code civil and article 1.2.2.1 of the rules of ethics applicable to pharma labs, by offering free unsollicited samples of generic quietapine to pharmacists;
  • violated articles L.5122-1 and L.5122-2 of the Code de la santé publique, article 1382 of the Code civil as well as the recommendations issued by the ANSM (i.e. the regulatory authority in charge of pharmaceutical products), by inserting the mention “identical to the original drug” on its ads; and
  • violated article 1382 of the Code civil by committing passing off, in view of the aspect of the generic quietapine packaging and tablets.

But such creativity did not pay off in front of the Cour d’appel, which held that:

the facts and evidence submitted to the judge for defining a manifestly illicit breach caused by disloyalty in free competition are necessarily linked to the rights that AstraZeneca derives from their patent, which confers a monopoly to them until its expiry or until its revocation by a court, the effect of which is retroactive, so that the civil action started in front of the juge des référés of the Tribunal de commerce de Lyon is necessarily a connected unfair competition action, so that he does not have jurisdiction due to the subject-matter and the exclusivity of jurisdiction imparted by statute to the Paris TGI. 

Said otherwise, all torts possibly committed by Mylan in this case are closely related to the alleged patent infringement. Patent infringement is a matter for the Paris TGI to rule upon, not for the Lyon T. com., and so are any side claims linked to this patent infringement (such as the passing off claim).

The court also stated that the various violations of the Code de la santé publique alleged by the patentee were not relevant, because this is a law which does not regulate competition between pharma companies:

All the violations of the Code de la santé publique which are alleged do not aim at protecting competition but rather public health, notably by imposing rules of ethics to pharmaceutical labs and by imposing advertising rules protecting drug consumers. 

The alleged violations are not within the ambit of jurisdiction of the juge des référés of the tribunal de commerce, for deciding measures of injunction […] similar to those that the infringement judge could take, because these violations, if they existed, would not constitute a manifestly illicit breach of free competition. 

And thus the logical conclusion was that:

[…] although it is certain that the juge des référés of the tribunal de commerce has in principle jurisdiction with respect to proven unfair competition, when there is a manifestly illicit breach, in order to take preliminary and injunction measures, even if there is a serious defense, because he is the natural judge between two commercial companies, […] the special provisions of the Code de la propriété intellectuelle prevail when, as in the present case, commercial parties are opposed in a lawsuit relating to intellectual property, the judge of which, having jurisdiction for ruling on the validity of the title or on infringement, has been previously applied to, before the urgency unfair competition claim. 

The judge of the Lyon tribunal de commerce has thus in this case no jurisdiction for ruling as he did and committed abuse of power.

Last but not least, the court ordered AstraZeneca to pay 100,000 euros of attorney fees to Mylan, which looks like a fairly high amount in the context of urgency proceedings. Courts have an almost complete discretion in terms of reimbursement of attorney fees, which they decide on as a matter of equity. Usually, when the amount is remarkably high, there is a sense that the court may have wanted to punish the losing party and to send a warning message.


CASE REFERENCE: Cour d’appel de Lyon, 1ère chambre civile, section A, December 17, 2015, Mylan SAS v. AstraZeneca SAS & AstraZeneca AB,
RG No. 2015/08783.

A flayed patent

We have one saying which goes “Vérité en-deçà des Pyrénées, erreur au-delà, which could be translated by: “What holds true on one side of the Pyrenees may be false on the other“. Well, the same can probably be said of the Rhine.

In a previous post, I discussed a recent decision by the Paris Cour d’appel, in which medical device claims were found novel due to a purpose feature – in keeping with the well-established French novelty standard. This approach is at odds with the practice of the EPO, where purpose features are only taken into account if they can be translated into structural limitations.

Yet, the greater generosity of French courts with patentees with respect to novelty is counterbalanced, it seems, by greater severity in particular with respect to sufficiency of disclosure, as the same case shows.

As a reminder, claim 1 of the patent was directed to:

A device for selectively disrupting lipid rich cells in a non-infant human subject by cooling, comprising:

– cooling means for cooling a local region of the subject’s skin to selectively disrupt lipid rich cells of the region, while, concurrently therewith, maintaining the subject’s skin at a temperature whereby non-lipid rich cells are not disrupted, wherein the cooling means are adapted to cool the lipid rich cells to a temperature between about -10°C and about 25 °C, 

– a temperature control unit for controlling the temperature of the cooling means, and temperature measuring means which are adapted to measure the temperature of the subject’s skin and/or the temperature in the subject’s skin and/or the temperature on the surface of the subject’s skin; characterized in that 

– the temperature control unit is further adapted to control the temperature of the cooling means such that the temperature of the subjects skin and/or the temperature in the subject’s skin and/or the temperature on the surface of the subject’s skin does not fall below a predetermined minimum temperature on the basis of the temperature of the subjects skin and/or the temperature in the subject’s skin and/or the temperature on the surface of the subject’s skin.  

The court started the sufficiency assessment by defining the skilled person, in a plural form:

[…] In this case, the skilled in the art is a team composed of a specialist in skin biology and a specialist in the field of cryogenics (cryolipolysis as regards the destruction of fat by the cold).

Then, the court recalled the general purpose of the claimed invention (which, as a reminder, was duly taken into account in the novelty analysis):

[…] The cryolipolysis device disclosed by claim 1 of the patent at stake must be able to selectively break lipid-rich cells by cooling, owing to cooling means which make it possible to maintain, at the same time, the skin of the subject at a temperature such that the other cells of the dermis and the epidermis are not broken […]. 

The sufficiency question at stake was thus whether there was enough information in the patent to achieve this general purpose.

In this respect, three parameters came into play, according to the court: cooling temperature, cooling time, and the surface area of the patient’s skin subjected to the cooling.

The patent proprietor’s position was that:

  • the temperature to be applied was from -10 to 4°C, or from -2 to 15°C;
  • the optimal time of application was from 5 to 20 minutes;
  • the surface of treatment should be at least 1 cm2 and preferably from 3 to 20 cm2;
  • the determination of the optimal values for the various parameters was a matter of routine implementation for the skilled person;
  • the examples in the patent would direct the skilled person to specific cooling parameters of -6°C for 5 minutes or -7°C for 10 to 20 minutes.

In order to assess the merits of this defense, the court analyzed the contents of the description of the patent in great detail.

They noted a number of extremely generic and vague statements.

Regarding the temperature to be applied, the court noted that

[…] at paragraph [0024] of the description, a temperature range for the cooling means is mentioned which is excessively broad, since it goes from the temperature of liquid nitrogen (-196°C) to human body temperature (37°C); at paragraph [0025], preferential ranges are provided, but they are still very broad and imprecise, since they encompass from “approximately” -15°C to “approximately” 35, 30, 25, 20, 15, 10 or 5°C, or from “approximately” -10°C to “approximately” 35, 30, 25, 20, 15, 10 or 5°C, or also from “approximately” -15°c to “approximately” 20, 15, 10 or 5°C […]. 

The court continued to analyze the rest of the description of the patent and remarked that many other indications were provided for the temperature of cooling of lipid-rich cells as well as for the cooling time. In particular, the court noted the disclosure of preferred temperature ranges of -10 to 37°C, -4 to 20°C, -8 to 33°C, -2 to 15°C, -10 to 35°C, -5 to 10°C, -5 to 5°C, -10 to 20°C, -8 to 15°C, as well of many exemplary values.

Turning to the cooling time, the court stated that

[…] paragraph [0026] starts by teaching an application time for the cooling means of up to two hours, while retaining a preferred time of between one and thirty minutes, but without specifying for each temperature level of the cooling means what the corresponding preferred application time is, except for the sole example of liquid nitrogen application, although on the one hand liquid nitrogen temperature does not correspond to the preferred temperature ranges mentioned in the previous paragraph and on the other hand the duration of application (one tenth of a second) does not correspond to the time range indicated as preferential. 

The court also noted that various cooling patterns were disclosed, such as continuous cooling, multiple cooling cycles, and cooling with intermediate warming up periods (which, said the court, are not specified). Time intervals between cooling applications ranged from 1 minute to 1 hour or from 12 to 24 hours.

The court’s conclusion was that

[…] it cannot be derived from such broad and imprecise – or even contradictory – temperature and time indications that, for the skilled person, the cooling temperature range for achieving cryolysis of lipid cells would vary between -2 and 15°C and the optimal application time would vary between 5 and 20 minutes (all the more so that this time mentioned at paragraph [0045] does not relate to the application time but to the preferred interval between each application).

Regarding the surface area of the cooling means, the court referred to paragraph [0030] of the patent, which provided a general rule as well as preferred values. The relevant passage of the paragraph reads as follows:

[…] Generally, the dimension of the surface area (e.g., where the cooling agent is in contact with the skin) should be at least three times the depth of subcutaneous fatty tissue that is targeted for cooling. Preferably, the minimum diameter of the surface area is at least 1 cm2. Even more preferably, the diameter of the surface area is between 3 to 20 cm2. Determination of the optimal surface area will require routine variation of several parameters. For example, larger surface areas, such as those over 3500 cm2, can be cooled according to the methods of the present invention if hypothermia is prevented by additional means.

Again, the court was dissatisfied with this teaching. Actually, there seems to be a confusion in this passage between dimensions and surface areas. Moreover, “it is not taught how the skilled person could measure the depth of subcutaneous fatty tissue“, said the court; and the preferred ranges are very broad indeed.

If we stop at this point for a moment, the main problem of the patent’s description is not that there was too little information but rather that there was too much. Instead of providing a couple of relevant ranges of values for the various parameters at stake, the patent offered a huge number of variants within an extremely broad disclosure.

I take this as a serious warning against U.S.-style claim drafting. It is quite common to find extremely broad definitions in U.S.-originating patents, and the reader sometimes gets the impression that any term can mean anything and that each parameter can take any possible value under the sun. This drafting practice of course makes perfect sense in terms of affording the best scope of protection, especially in the U.S., but the present example shows that it may not always play well on this side of the Atlantic, where courts insist on finding an actual technical teaching which can be of practical use to the skilled person in the patent, in order for that patent to be deemed worthy of being upheld.

My feeling on this is that the Cour d’appel can probably not be blamed for looking for a real, practical teaching in the patent.

But then comes a more controversial part of the judgment. Indeed, the patent also contained a number of examples relating to experimental testing on pigs and showing a decrease in adipose tissue without damage to the dermis or epidermis.

At the EPO, “an invention is in principle sufficiently disclosed if at least one way is clearly indicated enabling the person skilled in the art to carry out the invention“, as the Case Law bible (Case Law of the Board of Appeal of the European Patent Office, 7th edition, 2013) puts it in section II.C.4.2.

So, shouldn’t we apply this standard, and conclude that the examples in General Hospital’s patent save it from falling for insufficiency of disclosure? Nope, and here’s why:

[…] Zeltiq and General Hsopital, based on the expert report of William E. […] state that these experimental results would direct the skilled person towards the application of a temperature from -6°C for 5 minutes to -7°C for 10 to 20 minutes.

[…] Domestic pig skin […] has morphological and functional characteristics similar to those of human skin, but William E. acknowledges that “of course in medical research, human testing is the reference standard”;

[…] Although this expert believes that the skilled person […] would have concluded in 2002 that the results observed in the three pigs of examples 1 and 2 would also apply to humans, it should be noted that this expert, who is an MD specialized in dermatology cannot be considered as the skilled person […] and that he relies on findings from more contemporaneous studies, whereas the priority date of the application […] is the reference for assessing the sufficiency of disclosure of the patent.

[…] It can be derived in particular from the article by Mr. M. entitled “the skin of domestic mammals as a model for human skin, with special reference to domestic pig” dated 1978 […] that “data from animal experiments cannot be transcribed without restriction to man”.

[…] Thus it does not appear that the skilled person in 2002 could have extrapolated the results of experiments 1 and 2 described in the patent and relating to pigs so as to select a cooling temperature application from -6°C for 5 minutes to -7°C for 10 to 20 minutes. 

With that, claim 1 as well as the dependent claims relied upon by the plaintiffs were held invalid.

 

The unfortunate result of excessive fat disruption due to an erroneous setting of the patented device.
The unfortunate result of excessive fat disruption due to an erroneous setting of the patented device.

Again, I think this part of the judgment is quite debatable.

Indeed, in the medical or medical-like field, there are usually in vitro studies first, then animal studies, and only then human studies. Actual experiments on human patients usually come late in the development process, so late that it does not make sense to wait for that stage before filing a patent application, in a first-to-file system.

This is certainly why the Boards of Appel of the EPO only require that the description of the application should make it plausible that a claimed technical effect can indeed be achieved (see e.g. here). My understanding is that animal studies are generally considered as making it plausible that a certain treatment can also be applied to humans, unless there are special reasons to come to the opposite conclusion. There is no denying that animal studies are less complete than human studies but stating that animal experiments are generally not sufficient for a skilled person to carry out an invention would mean tossing out most patents in the drug and medical device industries.

I have not had access to the exhibits filed by both parties in this lawsuit, so I do not have a complete opinion on the patent at stake. But let’s put it like this: absent any showing (1) that the selective disruption of lipid-rich cells in humans does not work, or (2) that the settings to make it work are very removed from those disclosed in the examples of the patent, I would say that the judgment was harsh with the patent proprietor.


CASE REFERENCE: Cour d’appel de Paris, Pole 5, chambre 1, January 12, 2016, Patrick M. & Clinipro v. The General Hospital Corporation & Zeltiq Aesthetics Inc., RG No. 13/13050.