The bell always rings twice

This blog has now been around long enough for the same case to make it to a post several times, when new developments come up.

The present post is thus a sequel to my earlier comments on the litigation between Airbus Helicopters and Bell Helicopter Textron Inc.

As a reminder, the Paris Cour d’appel found Bell guilty of infringement of French patent No. FR 2749561 directed to a landing gear with skids.

Bell filed an appeal on points of law, and the Cour de cassation has now issued its judgment rejecting the appeal. Many different legal points were raised by Bell, but certainly the most interesting one relates to the finding of infringement concerning the so-called first version of helicopter Bell 429. This first version was a prototype, which was:

  • included in Bell’s catalogue;
  • shown in September 2005, which was reported on the website of a helicopter enthusiast (although Bell argued that this was a confidential presentation);
  • and put up on the website of Rotor & Aircraft, Bell’s distributor in France.

Bell’s defense was that this first version of the helicopter was not yet approved for sale. But the appeal judges noted that:

Offering for sale extends to any material operation aiming at putting a product in contact with potential customers, at preparing a commercial launch, even if said product, which in the present case was not yet approved, could not be marketed.

A helicopter prototype at a very early stage.

The “prototype” defense was raised again by Bell in front of the Cour de cassation, according to the seven following points.

First:

A mere prototype [was presented] in a private context and within a narrow circle of specialists; the prototype had not yet flown, and was later subjected to several modifications; in the absence of any prior certification, it could not be marketed; [this] cannot correspond to putting on the market nor even to a preparatory act for such putting on the market.

As a result, said Bell, the court breached the statute, by deciding that an offer for sale could relate to a product which could not be marketed.

The cour de cassation replied that the standard applied by the Cour d’appel was the right one, in particular because “the presentation of the product as a prototype can divert part of the clients of the patented product“. 

Second, the presentation of the first version of the landing gear was quickly followed by the marketing of the second version of the gear. Also, the certification of the second version was based on tests performed on the first version. Thus, the presentation was at best a preparation for the sale of the second version, and not an offer for sale of the first version.

The supreme court judges disagreed. In fact, the Cour d’appel had held that Bell intended to market the first version of the landing gear at the time it was presented.

Third, Bell blamed the appeal judges for relying on acts committed outside of France: an offer for sale abroad in a catalogue, and presentations in exhibitions in the U.S. and Canada. They said that only acts committed in France could possibly constitute an infringement.

However, for the Cour de cassation, the Cour d’appel did not hold any acts committed abroad as acts of infringement. They simply noted that the presentation in France had been preceded by offers for sale abroad. As I understand it, it is therefore OK to refer to acts committed abroad in order to characterize the overall context.

Fourth, Bell submitted that the report of the first version on a website owned by a helicopter enthusiast was an act committed in a private and non-commercial context – which is an exception to infringement.

Fifth, regarding the presentation of the helicopter on Rotor & Aircraft’s website, Bell argued that the landing gear could not be identified on the pictures appearing on the website.

Sixth, there was no evidence of the presentation of the first version in the Paris air show, since the bailiff’s report did not precisely describe the landing gear.

The Cour de cassation did not provide a detailed answer and simply referred to the appeal judges’ discretionary power for appraising the available evidence.

Seventh, the Cour d’appel refused to acknowledge that the acts committed related to the statutory experimental exception, because no test and no act of research were performed when the first version of the landing gear was presented. However, Bell said, the sole use of the landing gear could by itself aim at studying its behavior and therefore relate to the experimental exception.

But the cassation judges decided that the absence of evidence of any experimental testing was sufficient to reject the experimental exception.

Many other grounds were raised, regarding the validity of the patent, the validity of the infringement seizure report and the infringement by equivalence of the second version of the landing gear. They were all discarded.

One rather enjoyable part of the discussion relates to the way the bailiff drafted the infringement seizure report. The bailiff used in some instances the word “we” instead of “I“, so that it could not be determined, Bell said, what the bailiff had actually personally done or witnessed. The court replied that the use of the pronoun “we” was commonplace and could not be relied on to deduce that the bailiff had not carried out the seizure himself.

I wonder whether the pronoun “we” is that commonly used to designate a single individual. It is rather a specificity shared mainly by royalties and bailiffs, I would say.

Anyway, the take-away message from this important ruling remains that the presentation of a prototype can be considered as an offer for sale and thus an act of infringement, even if the product at stake is not yet marketable.


CASE REFERENCE: Cour de cassation, chambre commerciale, July 5, 2017, Bell Helicopter Textron Inc. & Bell Helicopter Textron Canada Ltd. v. Airbus Helicopters, case No. B 15-20.554.

One step forward, two steps back

The French statute of limitations for patent nullity actions. Yes: again. Seems like we can never get enough of this topic, can we?

Courtesy of Matthieu Dhenne (who, as regular readers of this blog may have noticed, is no fan of the time-bar on patent revocations), here is the latest installment of the saga.

In the previous episodes:

  • A trend slowly emerged at the Paris Tribunal de grande instance (TGI), per which actions for revocation are time-barred five years from the date at which the claimant became aware or should have become aware of the patent right at stake, as a potential impediment – said date being determined in concreto based on the specifics of the case.
  • Then, a ruling from the Paris Cour d’appel (Halgand v. RP Nicoll) issued on September 22, 2017 seemed to imply that the starting point is or should generally be the grant of the patent.

When I reported on this latest case, I wondered whether the TGI would apply the criterion set by the appeal judges. Well, we may have a beginning of an answer now – and it is a negative one.

Today’s decision was issued on October 5, 2017. The pleadings hearing took place on September 12, i.e. before Halgand v. RP Nicoll was issued. However, it is very likely that the court was aware of this appeal decision before drafting and issuing the judgment of October 5.

As this is not a common law country, there is no obligation for the TGI to defer to the legal analysis put forward by the Cour d’appel in another, unrelated case.

In summary: it seems that the TGI is sticking to its previous doctrine.

Going in circles around a tricky legal issue.

The case concerns French patent No. FR 2804186 (FR’186) owned by Valeo Embrayages (Valeo). The patent was filed on January 25, 2000 and granted on April 26, 2002.

On July 17, 2015, Valeo sent a warning letter to German company LuK GmbH & Co. Kg (LuK). Valeo claimed that several of its patents, including FR’186, were infringed by LuK, and requested that LuK should take a license.

On January 23, 2017, LuK filed an action for nullity of the FR’186 patent in front of the Paris TGI. Valeo raised the defense that the action was time-barred. The judge in charge of case management decided to hold a hearing in front of the whole court on this issue, in parallel to the exchange of submissions of the parties on the merits. Today’s decision thus solely relates to this legal issue.

First, LuK made the point that the defense based on the statute of limitations should not be examined separately from the merits of the case. Indeed, the determination of the starting point for the limitation period is related to the existence of an impediment to their business due to the patent, and therefore to the merits of the case.

The court rejected the argument as follows:

The fact that it is necessary to examine the scope of the patent and its potential points of contact with the technology of the nullity claimant to determine the starting point of the limitation period does not imply any analysis of the validity of the right subjected to the court. Even without the separation ordered by the judge in charge of case management to streamline the proceedings and reduce their cost in case of a success and otherwise remove a complex question, the court would anyway have decided on the issue of the statute of limitations first, and without considering the issue of validity of the patent or its infringement (if it were raised).

It is quite clear that the court has a discretionary power to deal with some issues first, especially a legal defense such as this one.

But LuK’s arguments do reveal one the drawbacks of the statute of limitations being applied to patent nullity suits, and of the in concreto appraisal of the starting point for the limitation period.

Even if determining when a party felt (or should have felt) impeded in its business by a patent is not synonymous with deciding whether there is or has been any patent infringement, it is still dangerously close. 

The court uses the enjoyable expression of “potential points of contact with the technology of the nullity claimant“. However, if would-be nullity claimants feel cornered into admitting or half-admitting infringement in order to be able to sue (and this is even more true if you take into account the issue of standing, which was not at stake here), this will act as a powerful deterrent.

The next point addressed by the judgment is the impact of a cross-license / settlement agreement dated 2014 between the parties.

LuK argued that Valeo was barred from invoking the statute of limitations due to the 2014 agreement. Indeed, one provision of the agreement recited that each party could challenge the validity of licensed patents within 12 months from a license request.

The court determined that the FR’186 patent did not belong to the technical field concerned by the 2014 agreement. Therefore, the agreement had no impact whatsoever on the present case. At any rate the above provision did not prevent the regular application of the statute of limitations as there was no clear intent to discard said statute; and as no license had been granted on this patent anyway.

And finally, the real meat of the judgment, that is the computation of the limitation period. As mentioned above, the TGI maintained its traditional approach: 

The starting point for the limitation period must be set to the date, determined in concreto, on which [LuK] knew or should have known, due to the progress in the development and industrial implementation of its TAC technology (which is the only one targeted by [Valeo]) that the FR’186 patent could impede it.  

Thus, the publication of grant of the patent is not a suitable starting point, as it would in fact demand an unrealistic watch from stakeholders and is unrelated to the performance of the project which gives standing to sue. Neither is the knowledge of the grounds of nullity of the patent, which may arise well before the knowledge of facts and economic considerations giving rise to standing to sue and actually is equivalent to the publication of grant. 

I discussed this point with Matthieu Dhenne, who agrees that this is an instance of what we call in this country “resistance of the lower courts“. Or rather, “coming to one’s senses” is I think the expression he used.

After this important point, the court reviewed the scope of the patent as well as LuK’s “TAC” technology in great detail. Valeo argued that the technology was developed from October 2007 at the latest and was disclosed in April 2010. As a result, the nullity action became time-barred in June 2013, they stated.

The judges disagreed and held that the technical features disclosed in the various LuK patents and publications relied on by Valeo were not sufficiently close to the subject-matter of the FR’186 patent to constitute the starting point for the limitation period:

Moreover, the article of April 14, 2010 and the patent applications filed do not reveal any relevant points of contact between the TAC technology and the FR’186 patent which could prompt [LuK] to worry about the patents of its competitor. 

Hum, those points of contact again!

[The fact that the FR’186 patent] was cited as prior art under category “A” in the search report of an international application of July 4, 2011 […] [is] irrelevant. Therefore, nothing shows that LuK intended, in the course of the development of its TAC technology, to define the nominal position of the gearing means owing to dedicated setting means as an element of its wear correction device. Consequently, the FR’186 patent represented a potential impediment only on the day it was expressly cited against them, i.e. on July 17, 2015. 

As a result, the case will now proceed to the discussion on the merits.

Readers will surely have noted that, since FR’186 was granted in 2002, LuK’s action would be time-barred if the Cour d’appel’s approach were applied instead of the TGI’s. 

This leads Matthieu to make the following remark:

The plaintiff did not submit that the statute of limitations is inapplicable. If there is an appeal, they should. This will provide them with an additional ground of cassation if the Cour d’appel confirms its case law of September 22, 2017 (as the action would be time-barred if the standard set in this case law stands). 

Well, I am sure it is only a matter of time before our supreme court gets to decide on this legal point. Whether people will like what they say is another kettle of fish.


CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre 1ère section, October 5, 2017, LuK GmbH & Co KG v. SAS Valeo Embrayages, RG No. 17/01156.

Representatives don’t count

Being a representative myself, my feelings tend to get hurt when I am told that representatives do not matter. Except, that is, when this is in the best common interest of attorneys and their clients, as in the present case.

In December 2013, Spanish company Agromet Ejea SL filed a French utility certificate application through its local French representative.

This is already an unusual start, since utility certificates are a rare species. But wait until you hear about the rest of the facts, which are rather unusual as well.

The second renewal fee for the utility certificate was due on December 31, 2014, and was not paid in due time. It was not paid either by the end of the 6-month grace period. This led the Institut National de la Propriété Industrielle (INPI) to issue a decision of noting of lapse dated August 31, 2015. This decision was notified to the French representative on September 3, 2015.

When the INPI sends you flowers.

A request for restoration of right was filed on November 16, 2015. The INPI rejected this request as inadmissible as it was filed too late.

The French provision on restoration of right is article L. 612-16 of the Code de la propriété intellectuelle. It is very similar to article 122 and rule 136 EPC. In particular, it is specified in the French provision that the request for restoration must be filed within two months of the removal of the cause of non-compliance.

The INPI considered that the cause of non-compliance had been removed upon notification of the decision of noting of lapse to the French representative on September 3, 2015. Therefore, according to the position of the office, the request for restoration should have been filed by November 3, 2015 at the latest.

The owner of the utility certificate appealed the decision in front of the Paris Cour d’appel.

They did well, as the court set aside the decision of the INPI and offered a different computation of the time limit for filing the request for restoration:

Contrary to the position of the [INPI], the legitimate excuse and the removal of the cause of non-compliance […] are appraised with respect to the owner of the IP right and not their representative

In the present case, Agromet Ejea SL stated that they entrusted a Spanish patent law firm with the filing and handling of the utility certificate […], which appointed firm [X] for France. They proved the deficiencies of their representatives in their submissions and with their exhibits, which led to the non-payment of the renewal fee […]. In December 2014, the email reminders of the French representative to the Spanish representative did not reach their destination due to an error of electronic address. In 2015, the IP portfolio management software of the French representative broke, preventing a reminder of the time limit for late payment. The employee of the French firm in charge of the file made several mistakes, leading to the termination of her employment agreement on September 9, 2015, with an effective departure on October 20, 2015. Finally, the decision of noting of lapse of August 31, 2015, notified on September 3, 2015 to the French representative was only communicated to the Spanish representative on October 12 and 16, 2015. The latter only informed the right owner on October 19, 2015. 

Besides, Agromet Ejea SL established that they inquired about the status of their right. Their Spanish representative asked the [French] representative about this by email on July 6, 2015. In view of the above, it can be derived, on the one hand, that Agromet Ejea SL could rely on a legitimate excuse for not paying the renewal fees of its IP right within the prescribed time limits, due to its representatives’ faults; and on the other hand that the cause of non-compliance was only removed on October 19, 2015, when they were made aware of the decision of lapse, or at the earliest when it was published in the BOPI on September 25, 2015

Thus, on November 16, 2015, since the two-month time limit had not yet expired and the payment of the surcharge was made on the same day, the request for restoration of Agromet Ejea SL was admissible.

In summary, the court held that the request for restoration was admissible and well-founded.

It is well known that the notion of legitimate excuse is assessed in a much more lenient manner in France than at the EPO.

In fact, a mistake made by a representative is a legitimate excuse per se, as illustrated by the present case. There is no need to demonstrate that the omission was the result of an isolated mistake within a satisfactory system for monitoring time limits, as demanded by the EPO. Fortunately so for the applicant in the present case, in view of the large number of deficiencies noted by the court…

But today’s decision illustrates that this different perspective also extends to the computation of deadlines.

According to established case law of the Boards of appeal, “the removal of the cause of non-compliance is a matter of fact which has to be determined in the individual circumstances of each case“. More particularly, “the removal of the cause of non-compliance occurs […] on the date on which the person responsible for the application (the patent applicant or his professional representative) is made aware of the fact that a time limit has not been observed“.

Usually, the receipt of a noting of loss of rights or negative decision is considered by the EPO as making the person in charge aware of the non-observance. In most cases, the relevant person is the European representative; but in other cases, it can be the applicant, or even a foreign patent attorney (see here), depending on the specifics of the case.

In contrast, the position of the Cour d’appel seems to be that only the receipt by the applicant can be relevant. Interestingly, the Cour d’appel also offered another possible starting point for the time limit, namely the publication of lapse in the BOPI (Bulletin officiel de la propriété industrielle), which is the local equivalent of the European patent bulletin.

In this case, the time limit was found to have been complied with irrespective of the exact starting point. But if the request for restoration had been filed e.g. on December 1, 2015, it would then have made a world of a difference whether the starting point was October 19 or September 25. So, it is somewhat strange that the court did not more clearly state its position.

The one unfortunate thing about this case though is that one of the very few advantages of utility certificates as opposed to patents is their reduced cost. But in this instance, I would not bet that any money at all was ultimately saved, in view of all the efforts necessary to keep the certificate alive.


CASE REFERENCE: Cour d’appel de Paris, pôle 5 chambre 1, April 25, 2017, Agromet Ejea SL v. Directeur général de l’INPI, RG No. 16/11489.

A dented infringement case

It seems that whatever patent-related news breaks in the UK often sets the tone for discussions throughout Europe – it remains to be seen if the Brexit will make any difference.

Take for example the doctrine of equivalents: this topic has become trendy again since the issuance of the Eli Lilly decision of the UK Supreme Court a few months ago. The excitement brought about by the decision is in fact quite understandable as the Supreme Court offered a new test for the appraisal of equivalents.

In France, on the other hand, much less is new under the sun in this respect.

The standard for assessing equivalents has been confirmed many times by the Cour de cassation, and all recent decisions dealing with this issue consistently rely on this standard (see notably here, here and there).

However, a recent ruling by the Paris Cour d’appel seems to have shed a new light on the nuts and bolts of the doctrine. 

French medical company Produits Dentaires Pierre Rolland (PDPR) owns French patent No. FR 2983398 on a “canula and adapter for a multifunction syringe“. In October 2013, PDPR initiated infringement proceedings against Itena Clinical, claiming that the marketing of the canula known as “Aireo Universel” infringed the patent (claims 1 and 6).

In April 2015, the Paris Tribunal de grande instance (TGI) confirmed that claims 1 and 6 of the patent are valid, and found Itena guilty of infringement. Itena appealed, which led to a second ruling issued in June 2017.

The validity part of the first instance judgment was upheld, but the appeal judges overturned the infringement part and found in favor of the defendant.

Here is claim 1 of the patent: 

A canula for a multifunction dental syringe, the canula comprising a fastener bushing that is defined by a circular opening, an end wall, and a side wall comprising a circular cylinder, the canula also including at least two channels enabling fluids to be conveyed separately or together from said end wall to fluid projection orifices, the two channels beginning in said end wall beside each other, the canula being characterized in that the circular cylinder presents a cutout in the thickness of its wall, which cutout comprises, going from the opening: at least one segment extending in a direction other than the axial direction of the cylinder; followed by a longitudinal segment extending towards the end wall.

Am I the only one to be horrified by this dental toy set?

The Cour d’appel first recalled the standard for appraising infringement, taking into account equivalents. As mentioned above, this standard was defined a long time ago:

The scope of the patent extends from the technique which is expressly claimed to any variant combining in the same manner different means performing the same function. The function of a means in an application is defined as the first technical effect achieved by the implementation of this means, making it possible to obtain the result. And two means are equivalent when, although they are of a different form, they perform a same function for a result of a same nature, if not of the same degree.

In this case, the court found that all the features of claim 1 were literally reproduced by the defendant’s “Aireo” product – except one:

The examination of the Aireo product shows that it has a cutout in the internal part of the canula, made of a single longitudinal segment oriented towards the end wall of the canula, which is identical to [what is claimed]. However, the cutout of the Aireo product does not comprise a segment having a direction other than the direction of the axis of the cylinder (transversal segment) [as claimed], so that the cutout is not made of the two segments defined in the characterizing portion of claim 1.

In such a situation, the function of the feature which is not literally reproduced must be established, and then it must be determined whether another feature in the alleged infringement provides the same function or not. 

The relevant discussion in the appeal judgment is the following:

PDPR argues that the collar on the canula, which forms a translation stop, makes it possible to rotate the canula on the adapter up to the longitudinal segment where, subsequently, the lug of the adapter slides in the groove of the Aireo product, which thus generates a longitudinal translation of the canula.

However, the alleged function of mechanical guiding by the collar present on the Aireo product is not of the same quality as that of the PDPR invention. Indeed, in the PDPR invention, when the canula and the adapter are assembled, the lug of the adapter is immediately inserted into the transversal segment of the cutout of the canula where it is mechanically directed into the longitudinal segment of this cutout, whereas such guiding is not ensured with the same efficacy by the collar of the Aireo product due to is planar and smooth surface.

In other terms, according to the claimant, the same function of mechanical guiding was present both in the claim and in the alleged infringement.

But the court did not accept this, since this function was not achieved with the same efficiency in both cases. This is related to the part of the equivalents’ test that reads: “a same function for a result of a same nature, if not of the same degree“. 

So, the court did stick to the classic test, but applied it in a fairly unusual manner. It is not everyday that equivalents are discarded because the common function is partly achieved but not with the same degree of efficacy as in the patent.

Now, for the sake of completeness, there were other reasons for the court not to accept the suggested equivalence: 

Besides, the alleged function of mechanical guiding is admittedly possible but absolutely not necessary, contrary to the device recited in claim 1 of the patent, nor even useful. Indeed, the user of the Aireo product wanting to fix the canula onto the adapter can easily match the lug of the adapter with the groove (longitudinal segment) of the canula, and therefore achieve a translation of the canula owing to its longitudinal segment by simply making a visual (not mechanical) alignment, without needing to resort to a rotative motion involving the planar and smooth surface of the collar, whereas the user of the canula according to the PDPR patent, in order to attach it to the adapter, must necessarily impart a rotation using the transversal segment of the canula, followed by a translation using the longitudinal segment of the canula (mechanical guiding).

So, the function was not necessary nor useful in the defendant’s product. This can probably be translated into: the common function argued by the plaintiff is in fact an artificial one.

Finally, the plaintiff’s case also failed on another prong of the equivalents’ test, which is whether the function could be protected or not in view of the prior art. The court found that the function (in the same context) was in fact known from the prior art, so that the scope of the claim could not validly be extended to the alleged equivalent:

Anyway, infringement by equivalence is not present if the means of the allegedly infringing product are of a different form relative to those which are claimed, and the function performed by these means, identical to that performed by the means of the claimed invention, is known from the prior art.

In this case, the function of mechanical guiding attributed by PDPR to the surface of the collar of the Aireo product combined with the cutout in the internal part of the canula of a longitudinal segment oriented towards the end wall of the canula is known from the prior art. It is notably disclosed in the abovementioned Saurou document (WO 92/04878). […]

Some interesting fine-tuning of the doctrine indeed!


CASE REFERENCE: Cour d’appel de Paris, Pôle 5 chambre 1, June 13, 2017, Itena Clinical v. Produits Dentaires Pierre Rolland, RG No. 15/10544.

Beware, drafters!

This may well be the fourth post almost in a row on a pharma case. Although pharma patent litigation is typically not hyperactive in France, each single case usually generates many interesting questions.

Today is no exception, not only from the litigation perspective but also from the viewpoint of a patent drafter. There are in fact at least two aspects in the decision of the Paris Tribunal de grande instance (TGI) which can be taken as significant warnings for patent attorneys.

The patent in suit is EP 0984957, owned by Swedish company AstraZeneca AB, and it looks quite simple and straightforward.

Claim 1 of the patent simply reads: “the magnesium salt of S-omeprazole trihydrate“.

S-omeprazole, also known as esomeprazole, is a blockbuster drug used in the treatment of gastric ailments. The patent is directed to a specific form of the drug.

Claims 2-4 further specify the form of the drug. Claims 5-8 relate to a process for the preparation of the drug. Claim 9 relates to a pharmaceutical composition comprising the drug of claim 1 and another one. Finally, claim 10 is a Swiss-type claim mentioning the treatment of a gastric acid-related condition.

In July 2011, AstraZeneca initiated infringement proceedings against Ethypharm based on this patent in view of the exploitation by this French company of generic esomeprazole.

Ethypharm of course filed a counterclaim for nullity in due time. But things did not go smoothly, and the lawsuit seemed to drag on forever.

First, an expertise was ordered so as to sort out the documents seized during an infringement seizure. Second, the parties discussed and initially agreed to the designation of another expert, as far as I understand in order to weigh on whether there was an infringement or not. Third, this expert was finally not designated, as AstraZeneca admitted that the generic esomeprazole on the market no longer infringed the patent (while maintaining that there had been infringement in the past). Fourth, AstraZeneca was compelled to file the experimental evidence based on which the above admission was made.

In summary, the judge in charge of case management certainly had her work cut out for her. And this leads us to the judgment issued by the TGI on June 23, 2017.

The first aspect of the decision that deserves a discussion is claim interpretation.

The patent proprietor stated that claim 1 covered any magnesium salt of S-omeprazole trihydrate. But, according to the defendant, claim 1 only protects a specific form of magnesium salt of S-omeprazole trihydrate.

The court went for the latter interpretation.

The court referred to article 69 EPC and the protocol on the interpretation, and remarked that “the judge must not make any interpretation if the claim is self-sufficient and should not denature the claim on the pretext of claim interpretation“.

One may wonder if the court did not do just that, i.e. interpret a claim based on the description, although the claim seemed quite self-sufficient.

At least from an EPO perspective, I have little doubt that a claim simply entitled “the magnesium salt of S-omeprazole trihydrate” would be seen as covering any magnesium salt of S-omeprazole trihydrate based on its plain wording.

But this is not the option that the court used. Instead, the court had a close look at the description of the patent and noted inter alia the following:

Since the wording of claim 1 does not comprise any determiner, one should refer to the description and drawings to interpret it, with respect to the skilled person […]. 

It is also specified that omeprazole and its salts as well as the R and S enantiomers of omeprazole and their salts are known from the art […] and that the magnesium salt of the S enantiomer of omeprazole exists in different forms. 

As mentioned in the description, the field of the invention does not relate to a “novel form of S-omeprazole” as suggested by the title of the patent, since this molecule was already identified, but a “novel form  of trihydrate of the magnesium salt of the S-omeprazole” […], which implies that it is a trihydrate form other than known from the art, with specific features such as “substantially pure” […], devoid of magnesium salts of R-omeprazole […] and devoid of other forms of magnesium salts of S-omeprazole (including the dihydrate form used for the preparation of the composition). 

This product is said to be “highly crystalline” […] since it has a larger crystallinity than any other form of magnesium salt of S-omeprazole, including trihydrate forms […]. 

This compound is characterized by an X-ray powder diffractogram [XRD] which shows main peak positions and intensities […], or by spectroscopy […]. 

Crystalline forms at their best

In support of its proposed interpretation, AstraZeneca filed an affidavit by a doctor Byrn, as well as a judgment by the New Jersey district court on a corresponding U.S. patent, per which the magnesium salt of esomeprazole trihydrate was novel at the time the invention was made and the patent was to be interpreted in a broad manner.

By the court was not convinced and noted that

the prior art previously disclosed a non-pure form a magnesium salt of S-omeprazole trihydrate, as can be derived from the laboratory tests performed on molecules recited in previous patents […]. [And] it can be deduced from the aforementioned elements in the description (notably the very title of the patent and the wording of the claims, the use of determiners in the text of the description such as “A” or “This” […] to mention the product […]) that the invention does not relate to any trihydrate of magnesium salt of S-omeprazole as suggested by AstraZeneca, but to a specific trihydrate having defined features (substantially pure, crystalline, with specific peaks) […]. 

The court repeatedly stated in the judgment that some magnesium salts of esomeprazole trihydrate were known from the art. But it does not seem to me to clearly stem from the patent itself. It is possible that the recreation of prior art esomeprazole could show that this particular salt form was already present, but isn’t this rather an issue of novelty of the claim?

Also, it is extremely striking that the court placed so much emphasis on how the description was drafted.

Here is the thing, I think. Paragraph [0009] of the patent for instance starts with: “the magnesium salt of S-omeprazole trihydrate obtained according to the present invention is substantially free from magnesium salts of R-omeprazole“. Then paragraph [0010] starts with: “the compound of the invention is characterized by the positions and intensities of the major peaks […]“.

If my understanding is correct, it could have made a world of a difference if these paragraphs had mentioned that the compound “according to some embodiments” is substantially free from other salts, or that the compound of the invention “may be” characterized by certain positions and intensities of peaks.

Therefore, beware drafters! Even if you get what you consider a broad claim granted, it may be interpreted by a court in a much narrower manner if the description gives the impression that the invention has a number of other essential features.

Surprising? Well, with Munich eyes, certainly. But not really if you are familiar with French case law.

Today’s decision is for instance very reminiscent of the approach recently taken in the rosuvastatin case. Even electronics cases are handled in a similar manner, as illustrated here.

The TGI’s comment on dependent claims 2-4 is enlightening in this respect. Claim 2 mentions that the compound is highly crystalline. Claim 3 mentions that it is stable. Claim 4 mentions that it has a certain XRD pattern. A common way of looking at this would be to state that claim 1 is broader than claims 2-4 and thus is precisely not limited to the specific XRD pattern, or to a highly crystalline form, etc. But the court reached the exact opposite conclusion and noted that

these claims define the specific features of the compound mentioned in claim 1.

Now, one may criticize the court’s rewriting of straightforward claim 1 and the challenge this presents to legal certainty. 

But there may nevertheless be a possible justification for this bold approach.

As the patent only discloses one very specific crystalline form of trihydrate, it could be argued that the protection should be limited to this specific form and should not extend to other forms of trihydrate which were not actually made available to the public by the patent. Maybe this is what the judges had in mind.

After this section on claim interpretation, the judgment contains a section on sufficiency of disclosure and novelty.

Both grounds of nullity were discarded by the court. In fact, the objection of lack of novelty was put forward only in case the patentee’s broad interpretation prevailed – and this was not the case.

And then comes the other important point in the judgment, namely inventive step.

At this point, the patent was revoked, based on a very brief justification.

The court generally made reference to the problem and solution approach. The dihydrate form of the compound was the closest prior art. According to the patent, the new form is more stable and easier to characterize and synthesize.

But the court refused to take this statement of a technical problem into account:

Yet, beside the disclosure of the molecule on a new form (trihydrate) and the presentation of preparation and identification methods of this product, whereas the pharmaceutical industry wants to find further forms of an active, even if the properties to be expected from the new molecular forms are not known, the patent does not define any problem. It just mentions that the product is more stable, easier to synthesize and handle and identify, without however supporting this statement in the patent itself, by studies and researches and results. Thus, the patent does not mention a problem to be solved let alone demonstrate the resolution of this technical problem. 

Moreover, it is acknowledged that in front of the EPO or in litigation the patentee may refer to posterior tests, but these must consolidate results already contained in the patent […]. 

Therefore, claim 1 is invalid for lack of inventive step as it does not solve a technical problem.

To summarize: the patent does mention a technical problem solved by the invention. But in the absence of any experimental evidence in the patent that this problem is indeed solved, the problem is not taken into account. Post-published evidence is considered to the extent that it can merely supplement data already contained in the patent, but not replace it entirely.

So far so good, and the approach taken by the court seems consistent with EPO case law. See in this respect the catchword of oft-quoted T 1329/04:

The definition of an invention as being a contribution to the art, i.e. as solving a technical problem and not merely putting forward one, requires that it is at least made plausible by the disclosure in the application that its teaching solves indeed the problem it purports to solve. Therefore, even if supplementary post-published evidence may in the proper circumstances also be taken into consideration, it may not serve as the sole basis to establish that the application solves indeed the problem it purports to solve.

However, directly jumping to the conclusion that the claims lack inventive step, as the court did, is something a Board of appeal would probably not do.

Instead, they would likely reformulate the technical problem in a less ambitious manner, e.g. in this case as providing an alternative form of esomeprazole, and would then investigate whether it was obvious for the skilled person to achieve the claimed invention with this unambitious technical problem in mind.

Let’s assume for instance that it was technically difficult to make the trihydrate form. In this case, a Board could arrive at a finding of inventive step. In France, the absence of an ambitious technical problem is in itself indicative of a lack of inventive step.

This confirms that French courts are more severe in the appraisal of inventive step.

At any rate, here comes the second advice for drafters is: beware of the plausibility of the technical effects of the invention!

My understanding is that it is kind of a hot topic at the EPO these days. Well, it is an even more serious matter in this country.

And this does not just apply to “therapeutic” effects.

In this case, the alleged technical effect was not related to the treatment of an illness but rather to physical characteristics of the drug (stability, ease of handling…).

So, we are all doubly warned, I guess.


CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre 3ème section, June 23, 2017, AstraZeneca AB v. Ethypharm, RG No. 11/11460.