A risky launch at risk

Launching at risk is not a business for those who have a weak stomach.

Once a generic drug company has decided that it will go on the market despite the existence of an IP threat – e.g. because they have determined that the IP right at stake is probably invalid or unenforceable – then anything can happen, and it may happen really fast.

Today’s decision illustrates the risks of a launch at risk.

GD Searle LLC owns European patent No. EP 0810209 filed on August 24, 1993. The patent is directed to a family of antiretroviral drugs, more precisely protease inhibitors, useful for the treatment of HIV infection.

On May 9, 2008, Searle obtained a Supplementary Protection Certificate (SPC) from the French patent office (INPI), under No. FR 07C0034. This SPC is based on the EP’209 patent and relates to “darunavir or one of its pharmaceutically acceptable salts, esters or precursors“.

An exclusive license to the SPC was granted to Janssen Sciences Ireland Unlimited Company and then a sub-license was granted to its French affiliate Janssen-Cilag SAS. Janssen (Johnson & Johnson group) markets a drug known as Prezista®, the active compound of which is darunavir.

SAS Sandoz is a French affiliate of the Sandoz group (affiliated to Novartis). As readers have probably already guessed by now, Sandoz launched a generic version of Prezista® before the expiry of the French SPC – which is set to take place on February 23, 2019.

Another form of launch at risk.

The parties have been entangled in pan-European litigation over the darunavir SPCs. This litigation actually hit the headlines when, on January 25, 2018, the British Court of Appeal referred a question to the CJEU in relation with the validity of the British SPC.

The case is still pending and was allocated number C-114/18.

The issue at stake is the following.

Claim 1 of the EP’209 patent is a so-called Markush claim.

A Markush claim covers a large class of chemical compounds by way of a generic formula, without reciting them individually. In this case, the generic formula contained in claim 1 is the following:

Each group among P1, P2, R2, R3 and R4 is defined as being selected among of list of possible options. I will spare you the complete lists here. But it may be important to note that each list itself recites generic classes of groups, rather than individualized groups. For instance P1 and P2 can be alkoxycarbonyl groups, aralkoxycarbonyl groups, alkylcarbonyl groups, etc. Each of these options encompasses an unknown number of individual possibilities. In the end, the array of choices falling under the generic formula is practically unlimited.

Well, not quite. In fact the number of theoretical possibilities was estimated in the British litigation to be from 7 x 10135 to 1 x 10377. Not a very accurate estimate, but let’s not skimp on a couple hundred zeros.

Now, darunavir has the following structure:

It turns out to be one of the zillion compounds covered by claim 1

Several other claims of the patent also cover darunavir, but always in a generic way. Darunavir is not cited in the patent, not exemplified, nor singled out in any other manner.

Now, according to famous / infamous article 3(a) of the so-called SPC Regulation (Regulation (EC) No. 469/2009):

A certificate shall be granted if, in the Member State in which the application referred to in Article 7 is submitted and at the date of that application:

(a) the product is protected by a basic patent in force […].

As regular readers of this blog are well aware, people have been arguing, fighting and spending millions and millions of euros for a number of years over what it means for a product to be protected by a basic patent.

Some (limited) guidance has been offered by the CJEU over time.

For instance, this is from the Medeva judgment (C-322/10):

Article 3(a) of [the SPC Regulation] must be interpreted as precluding the [grant on an] SPC relating to active ingredients which are not specified in the wording of the claims of the basic patent relied on in support of the SPC application.

Then there is the Eli Lilly judgment (C-493/12):

Article 3(a) of [the SPC Regulation] must be interpreted as meaning that, in order for an active ingredient to be regarded as ‘protected by a basic patent in force’ within the meaning of that provision, it is not necessary for the active ingredient to be identified in the claims of the patent by a structural formula. Where the active ingredient is covered by a functional formula in the claims of a patent issued by the EPO, Article 3(a) of that regulation does not, in principle, preclude the grant of an SPC for that active ingredient, on condition that it is possible to reach the conclusion on the basis of those claims, interpreted inter alia in the light of the description of the invention, as required by Article 69 of the EPC and the Protocol on the interpretation of that provision, that the claims relate, implicitly but necessarily and specifically, to the active ingredient in question […].

And let’s not forget about Teva (C-121/17):

Article 3(a) of [the SPC Regulation] must be interpreted as meaning that a product composed of several active ingredients with a combined effect is ‘protected by a basic patent in force’ within the meaning of that provision where, even if the combination of active ingredients of which that product is composed is not expressly mentioned in the claims of the basic patent, those claims relate necessarily and specifically to that combination. For that purpose, from the point of view of a person skilled in the art and on the basis of the prior art at the filing date or priority date of the basic patent: the combination of those active ingredients must necessarily, in the light of the description and drawings of that patent, fall under the invention covered by that patent, and each of those active ingredients must be specifically identifiable, in the light of all the information disclosed by that patent.

And there are a few other relevant ones.

But the thing is, the CJEU has never ruled so far on how article 3(a) should be applied with respect to a Markush claim. So, in this case, should darunavir be considered as being “specified in the wording of the claims” (as per Medeva)? Is it possible to conclude that the claim “relates, implicitly but necessarily and specifically“, to darunavir (as per Eli Lilly)? Or is this not a relevant question as the claim should not be considered as a functional claim? Or maybe a combination of substituents in a Markush formula should be treated like a combination of active compounds, so that it should be investigated whether, “from the point of view of a person skilled in the art and on the basis of the prior art at the filing date or priority date of the basic patent“, each of the substituents is “specifically identifiable, in the light of all the information disclosed by that patent” (as per Teva)?

Of particular interest in this case is that, not only was darunavir not known as a compound at the filing date and was only made available several years later; but group P1 of the generic formula, when the formula applies to darunavir, can be considered as an unusual group at the filing date, which does not form part of common general knowledge.

The British Court of appeal thus proposed the following question for the CJEU:

Where the sole active ingredient the subject of [an SPC] issued under [the SPC Regulation] is a member of a class of compounds which fall within a Markush definition in a claim of the patent, all of which class members embody the core inventive technical advance of the patent, is it sufficient for the purposes of Article 3(a) of the SPC Regulation that the compound would, upon examination of its structure, immediately be recognised as one which falls within the class (and therefore would be protected by the patent as a matter of national patent law) or must the specific substituents necessary to form the active ingredient be amongst those which the skilled person could derive, based on their common general knowledge, from a reading of the patent claims?

And this now brings us to France.

Sandoz obtained a Marketing authorization (MA) for the generic version of Prezista® on December 15, 2017. Searle and Janssen issued multiple warnings to Sandoz and then discovered in December 2018 that the generic drug (Darunavir Sandoz) was on the French market.

On December 14, 2018, they filed a complaint against Sandoz in urgency proceedings (référé d’heure à heure) for infringement of their SPC No. FR’034 – after getting an authorization to do so from the court.

Sandoz’ sole defense relied on the invalidity of the SPC for non-compliance with art. 3(a) the SPC Regulation. 

The hearing took place one week only after the filing of the complaint, on December 21, 2018. Then the judge Ms. Lignières took the case home during the Christmas break and issued her decision on January 11, 2019, ruling against Sandoz.

In the decision, the judge analyzed the CJEU case law and came to the conclusion that the Eli Lilly test is not applicable to a Markush claim:

It should be noted that, in the Eli Lilly decision, the case related to a functional claim, so that the CJEU insisted on a double condition of necessity and specificity. In the present case, the claims of the basic patent are structural and better allow the skilled person to determined in view of the claims whether the active compound protected by the SPC was covered by the basic patent. 

Then, after analyzing the formula of claim 1 of the patent and the various meanings of the variable groups in this claim and some of the dependent claims, the court concluded that:

Therefore darunavir is identified by the skilled person as being implicitly but necessarily and specifically protected by the EP’209 patent in view of the substituents identified in the claims, in keeping with the requirements of article 3(a) of the EC Regulation and of the CJEU case law.

The judge thus considered that the defendant did not prove that the SPC was manifestly invalid.

Accordingly, a preliminary injunction was ordered, under a 50,000 euro-penalty per violation of the injunction. The judge also ordered a seizure of the infringing drugs, as well as a product recall.

I don’t have an exhaustive view of the pan-European Searle v. Sandoz litigation, but at least in the Netherlands a similar preliminary ruling was issued a couple of months earlier: see here.

In a recent post I wondered whether there was a surge in preliminary injunctions in France these days. The darunavir case seems to provide a confirmation. 

We will have to wait for the CJEU ruling to know whether the French judge correctly guessed which way the Luxembourg winds will finally blow.

But at any rate this decision should reassure IP right holders. A very quick preliminary decision can be issued in France when time is of the essence. Therefore, launching a generic drug a couple of months before the expiry of the IP at stake does not curtail the injunctive risk.


CASE REFERENCE: TGI Paris, ordonnance de référé, January 11, 2019, GD Searle LLC et al. v. SAS Sandoz, RG No. 18/60334

Nightmares before Christmas

Beside the tribute to Tim Burton’s work in the title, this post is probably going to be of little interest to non-French patent attorney readers.

Sorry for that. 

On November 19, 2018, the INPI (French patent and trademark office) launched a new web-based patent filing system, called “Portail Brevets“, and shut down the good old EPO-style OLF system which we had been using for a number of years. 

It is likely that in the future the new system will be viewed as a major improvement, in particular because it should make an actual online management of patent applications possible.

But so far, I am sorry to say that it has been all but a nightmare for many users. The suddenness of the switch from the old system to the new system (no more OLF, no more fax, no more in-person filing) did not play well. The fact that some important developments still seem to be missing in the software did not make us happy campers either (address book anyone? configuration of pre-filled fields anyone?).

But the main issue is the following.

The new system only accepts a docx file for the text and drawings of the patent application. You may no longer upload pdf files. The docx file needs to comply with a number of requirements. If your file is not compliant, you get an error message. But most of the time the error message is so vague that it does not make it possible to immediately understand what is wrong with your file.

To me this is the most serious defect of the current Portail Brevets, as you can easily waste hours trying to figure out why the system does not like your docx file. 

The point of this blog post is not to criticize the INPI.

I am grateful for the efforts made by good people in the patent office working hard to modernize their tools and to provide patent applicants with the best level of service. I can also imagine that times must be stressful for the INPI too, as they are probably confronted with angry customers on a daily basis. 

No, my idea was rather to just provide a few tips based on my own experience as a newbie with the Portail Brevets, in case they may be of assistance to others. It would also be fantastic if readers could give some tips of their own in the comment section of this post. 

If you have had a problem filing via the Portail Brevets or if you have identified a potential trap, it is more than likely that other users will also encounter the same problem or fall into the identified trap. Thus why not contribute and help the entire community? 

In fact, if this appears to be useful in view of the comments, I will consider completing this post in the future to incorporate all relevant contributions into the body of the post, so as to provide a useful resource to all stakeholders. 

Note that most if not all of the issues should be avoided if you carefully read and digest the user’s documentation (here, here, here and there). But putting theory into practice is never an easy task, and I tend to think that a sort of practical troubleshooting description can always be useful.  

Season’s greetings.

So here we go with the difficulties / traps that I have identified so far: 

  • In the drawings section, you should type [Fig. 1] or the like, then a single paragraph break (“enter“), then immediately insert your drawing image. If there are several paragraph breaks, you will get an error message. 
  • In the description, you want to make the first reference to your drawing figures in an individualized manner, and the figure tag should be exactly at the beginning of a line. Thus, you should type: “[Fig. 1] shows …“; and then on the next line: “[Fig. 2] shows…“. You should not write “[Fig. 1] and [Fig. 2] show…“. Nor should you write “An embodiment of the invention is shown on [Fig. 1]“. And since we are actually talking about French language texts, you should not write “La [Fig. 1]…” either, because then the figure tag is not at the very beginning of the line. 
  • Beware of lists in MS Word. For instance, if you have a list with a), b), c) and if item c) comprises a list with bullet points, the bullet points will be automatically renumbered as d), e) etc. in the pdf file generated by the software. One way to avoid this is to keep only the first list and manually insert hyphens or like symbols for the second list. At any rate, you should carefully check all lists in the pdf file generated by the software. 
  • The same applies all the more so to lists in the claim section. Paragraph breaks (“enter“) should be avoided within a claim, and only line breaks (“shift + enter“) should be used. Otherwise, your claim numbering and count could be erroneous. 
  • If you insert an image into your document, and if the docx file is not approved by the system, you may try to convert the image to a different type before reinserting it.
  • A number of expressions are recognized as tags by the software, even though they are without square brackets. This means that they will be renamed and treated as section headings in the pdf file that is generated. The list of relevant expressions can be found in one of the documentation files linked to above. As a result, (1) if you use a slightly different wording, the expression will not be recognized and treated as a section heading; and (2) you cannot change the headings in the generated pdf file. For instance, if you type “Description détaillée“, your final heading will read “Description des modes de réalisation“.  Too bad if, like me, you tend to prefer “Description de modes de réalisation” (embodiments vs. the embodiments). 
  • If you use the MS Word upper case style, for instance for the title of the invention, the formatting is lost in the generated pdf. In other terms if you have typed “DeVicE” in your docx file but it reads “DEVICE” because of the upper case style, the pdf will nevertheless show “DeVicE“. It may therefore be best not to use the upper case style at all. 
  • As a warning, when you press the pay button, the application is filed. There is no “file application” button or the like. 

That’s all for the time being. If you cannot figure out where the error is in your docx file, it is possible to call the INPI for help and possibly send them the file by email so that they can look at it and advise you. But of course this process takes a little bit of time. 

Good luck everyone!

Less limited guidelines

A few days ago, I had the privilege to speak at a joint INPI-EPO conference in the premises of the INPI in Courbevoie. My assigned topic was post-grant proceedings from a patent attorney’s point of view, and I had a few slides on sort of unresolved questions regarding French limitation proceedings. And just a couple of hours before my presentation, I learned that the INPI had just issued its long-awaited guidelines on limitation proceedings the night before. 

I frantically skimmed the fresh guidelines and as luck would have it, some of the points in my presentation were now obsolete. 

But that’s all right, I am glad that we now have more guidance from the patent office regarding their practice with respect to requests for limitation.

The updated and relevant part of the guidelines can be downloaded here.  

More guidelines – less guesswork.

Here are a few comments from me, after a second and less frantic reading. 

The first sentence of the new limitation section is actually a bit surprising. It reads: 

The patent proprietor may, at any time, the 5-year statute of limitations deadline being complied with, either renounce its patent entirely (total renunciation) or renounce one or several claims (partial renunciation) or limit the scope of the patent by modifying one or more claims (limitation). 

The surprising part is the reference to the 5-year statute of limitations deadline. Art. 2224 of the Code civil is cited on the side, next to this paragraph. This is the general statutory provision according to which personal actions or actions relating to a movable property are time-barred 5 years from the day the right owner knew or should have known the facts making it possible for him/her to exert the action. 

I must confess that it is the first time that I have seen any indication that this 5-year time bar, which we have discussed at length in the context of nullity suits, could be of any relevance whatsoever when it comes to the right of a patent owner to limit their patent in front of the INPI. 

Besides, the French version of this paragraph of the guidelines is actually ambiguous and could be interpreted as meaning that the deadline needs to be complied with, or that it is always complied with. 

I discussed this sentence with a couple of INPI people at the conference. After this discussion, I must say I still do not really understand the sentence, but I was told that it is not the INPI’s intention to refuse any request for limitation as being time-barred. 

And now, some other notable points, in no particular order: 

  • If the request for limitation is deemed inadmissible, a communication is issued, with a 2-month deadline to complete or correct the request. 
  • The request for limitation must be filed by or on behalf of the patent proprietor that appears on the national patent register. Otherwise, the request is inadmissible. This means that, in case of a change of ownership, this change must be officially recorded before a request for limitation can be validly filed. 
  • Auxiliary requests are not accepted and, if filed,  would make the request for limitation inadmissible. 
  • If the patent to be limited is a European patent, the request for limitation and exhibits must be filed in French. But, since the authentic text of the patent is the text in the language of the proceedings (before the EPO), the INPI must examine the limitation based on the text in this language. No translation of the patent into French is required. On the other hand, only the limited claims in the French language will appear on the national patent register. 
  • If the request for limitation is admissible but objections on the merits are raised (due for instance to a lack of clarity, or to the absence of an actual limitation), a communication is issued, with a 2-month deadline to correct any deficiency. This deadline can be extended by 2 additional months upon request. Again, it is not allowable to offer several modification proposals.
  • Third party observations are not admissible. 
  • The INPI will react to the initial request for limitation or to any further submission by the requester within 1 year, in view of the principle according to which “silence implies refusal” after a 1-year period. 

One takeaway message from the above is that the updated guidelines now officially confirm that auxiliary requests are not admissible at the INPI. And I assume that this is a general rule.

I wonder whether this will still be the case in the national opposition proceedings which may be set up in the near future. I am sure that all European patent attorneys will agree that the right to file auxiliary requests is absolutely fundamental in the patent proprietor’s defense in an opposition at the EPO. 

And now, the most important information in the guidelines, in my opinion: what is or is not considered as an actual limitation. 

First, it is possible to limit a dependent claim without amending the independent claims on which it depends. This is consistent with practice at the EPO. 

Second, a mere clarification of a claim is not allowed. The modification has to actually restrict the claim. 

Third, changing a claim category is generally not considered as a limitation. 

Fourth, the addition of claims, either broader or narrower than the existing ones, is prohibited. 

This fourth point is extremely important, because this has not always been the case. In one lawsuit that I worked on, the patentee had taken advantage of limitation proceedings to double the number of dependent claims without even restricting the main independent claim. Based on informal conversations at the conference, I understand that this would no longer be possible today. The INPI was probably influenced by a ruling from the Paris TGI discussed on this very blog (confirmed on appeal). 

Last but not least, what the Guidelines are silent on. 

To me, the elephant in the room is the extremely tricky question of what happens if a European patent is limited in France and then the patent is modified at the EPO in a different manner (in opposition or central limitation proceedings). I am not sure anyone has a clue what would happen in such a case.

I understand that the INPI is currently looking at possible options regarding the interplay between French limitation and the future French opposition. Could it be an opportunity to also revisit the interplay between French limitation and European opposition/limitation? 

I will conclude this post with some numbers.

With the help of my colleague Patrick Marollé (thank you), I have looked at the duration of limitation proceedings in France. Based on a sample of 25 cases, we have found an average duration of 11 weeks, from the request for limitation to the official issuance of the limitation.

A couple of years ago, I had conducted a similar analysis with respect to central limitation at the EPO, and had found an average duration of … 11 months. 

So, the winner is… the INPI. This is certainly due in part to a more straightforward procedure (no need for an invitation to file claim translations and pay a final fee for instance). But I am sure part of the explanation is also that INPI examiners handle requests for limitation with a high priority. As a reminder, in one of the recent cases reported on this blog, the limitation was formally issued only 8 days after the request was filed. That record will not be easy to break. 

Doppelgänger

The names of the parties in the case commented upon today first reminded me of the classic movie Kramer vs. Kramer. But you could also think about the Doppelgänger.

See, the case at hand is Novartis v. Novartis or, to be more precise, SAS Novartis Pharma v. Novartis Pharma AG.

The judgment is very short and utterly uninteresting in itself, but as you will see the underlying context is much more noteworthy.

Let’s start with the judgment per se first.

Novartis Pharma AG is the owner of European patent No. EP 1096932 entitled “Combined use of valsartan and calcium channel blockers for therapeutic purposes“.

Novartis Pharma AG is also the owner of French SPC No. 07C0042, which was granted in 2008. This SPC is based on the EP’932 patent and on a European Marketing authorization (MA) and is directed to a medicinal product comprising the two active substances valsartan and amlodipine.

The corresponding commercial drug marketed by the Novartis group is Exforge®.

The EP’932 patent was opposed by 8 different opponents at the EPO. In first instance, the patent was maintained in amended form by the opposition division.

On appeal, after the issuance of the summons to oral proceedings, Novartis Pharma AG withdrew all its requests and disapproved of the text of the patent, which led to the revocation of the patent on October 7, 2015.

On January 29, 2018, another company from the Novartis group, namely SAS Novartis Pharma, filed a complaint with the Paris Tribunal de grande instance (TGI) and requested that the TGI should revoke SPC No. 07C0042, as a consequence of the revocation of the basic patent.

The “defendant” Novartis Pharma AG agreed, and the court thus pronounced the requested revocation of the SPC on April 5, 2018.

Let me introduce the bunny and his feline Doppelgänger.

Now, in case you are wondering, of course the various Novartis companies have not run amok.

No, this unusual ruling seems to be the result of a very elaborate strategy, which can be (at least partly) comprehended based on publicly available information.

The first thing you need to know is that two divisional applications were filed based on the EP’932 patent.

The first one was deemed to be withdrawn, but the second one led to the grant of another patent, EP 2322174. On the face of it, the EP’174 patent contains claims which are somewhat similar to those of EP’932, and which still cover the combination of valsartan and amlodipine.

The second important thing is that another French SPC application (No. 16C0008) was filed based on the divisional patent EP’174, still for the combination of valsartan and amlodipine, on March 7, 2016 (thus a few months after the grant of EP’174 on September 15, 2015 and after the revocation of the parent EP’932 patent on October 7, 2015).

Based on publicly available information, the second SPC was initially granted in June 2016. But then, three months later, on September 26, 2016, the decision to grant the SPC was withdrawn by the INPI (French patent office). Indeed, the withdrawal or cancellation of any decision issued by the INPI is possible as a matter of principle within a four-month time limit.

Examination of SPC application No. 16C0008 was immediately resumed, and an office action was issued, in which an objection was raised based on the existence of the prior SPC No. 07C0042. As a reminder, article 3(c) of the SPC regulation (aka Regulation (EC) No. 469/2009 of the European Parliamant and of the Council of May 6, 2009) provides that a certificate can only be granted if “the product has not already been the subject of a certificate“.

As a first line of defense, Novartis replied that the revocation of the EP’932 patent also automatically entails the annihilation of the first SPC No. 07C0042, in a retroactive manner. Therefore, they argued, the valsartan + amlodipine product in the second SPC application No. 16C0008 had not already been the subject of a certificate.

But the INPI was not convinced and maintained its objection.

With that in mind, the strange Novartis v. Novartis action suddenly becomes much more understandable.

Novartis’ purpose was to obtain an official ruling from a court of law per which the first SPC was revoked. This strengthened their argument with the INPI. And the strategy was successful, as the objection based on article 3(c) of the SPC regulation was overcome and the second SPC was finally granted (for the second time) in June 2018.

Based on the arguments submitted by Novartis which can be read in the SPC file wrapper, second SPCs replacing the first round of SPCs were also obtained in a similar manner in a number of other European countries.

At this stage, one question remained for me: why did Novartis find it useful to surrender its parent patent and all SPCs obtained based on it, and to start the entire procedure from scratch based on the divisional patent and a second set of SPC applications?

Well, sometimes a blogger has to play detective. Comparing the two different SPC applications, it appears that both are based on the same MA having effect in France, namely European MA No. EU/1/6/371/001-24. But then one difference immediately becomes obvious in the next box of the application form.

In the first SPC application, the first MA obtained in the Community or EEE is indicated as being this same MA No. EU/1/6/371/001-24, dated January 16, 2007. But in the second SPC application, the first MA obtained in the Community or EEE is indicated as being a slightly earlier Swiss MA No. 57771/01-03, dated December 22, 2006. Actually, since the AstraZeneca judgment of the CJEU (C-617/12), we have known that Swiss MAs count as potential first MAs in the Community or EEE, as they are automatically recognized in Liechtenstein.

So the Doppelgänger mystery is solved – I think: the initial SPC application contained a mistake, in that the indication of the first MA obtained in the Community or EEE was not correct. Hence the need to get rid of the first patent and first SPC, in order to obtain a corrected SPC based on the second patent.  

I would be curious to know whether Novartis’ clever strategy succeeded in all European countries or whether it failed in some of them.

More generally, I would be curious to know what readers make of all this.

Should the revocation of a first SPC indeed make it possible to request a second SPC as if the first SPC had never been granted?

This does raise policy issues. Imagine that the first SPC had been revoked further to a nullity suit filed by a true third party. Would it be fair to allow the SPC owner to obtain a second SPC essentially identical to the first one, based on a divisional patent?

At any rate, this case will probably be a further incentive to file divisional applications for important inventions which could potentially be protected by an SPC. More than ever, divisional applications appear to be a very powerful (some would say, too powerful) tool in the hands of IP right holders.


CASE REFERENCE: TGI de Paris, 3ème chambre, 1ère section, April 5, 2018, SAS Novartis Pharma v. Novartis Pharma AG, RG No. 18/02118.

All eyes are on the Court of justice

A tout seigneur tout honneur, today’s decision, being about SPCs, was first reported on the SPC blog.

However, I would like to talk about it as well because, unlike British courts, it is not every day that the Paris Cour d’appel refers questions to the CJEU.

The case relates to a European patent owned by Santen, a pharmaceutical company specialized in ophthalmology.

Claim 1 of the patent is directed to a particular type of oil-in-water emulsion. According to a dependent claim, the emulsion may contain an active ingredient. According to another dependent claim, the active ingredient is cyclosporine A. Finally, another claim relates to the use of the emulsion for the preparation of an ophthalmic composition for the treatment of ocular diseases such as notably keratitis.

Santen obtained a Marketing authorization (MA) for the drug Ikervis®, which is an emulsion containing cyclosporine as an active ingredient and is indicated for the treatment of severe keratitis.

On June 3, 2015, Santen filed an SPC application at the French patent office (INPI) for “cyclosporine emulsion eye drops”. The SPC application was later amended to rename the product as “cyclosporine for its use in the treatment of keratitis”.

On October 6, 2017, the SPC application was finally refused by the INPI. The reason for the refusal was the existence of a prior MA dating back to 1983, for the drug Sandimmum®.

This drug is in the form of a drinkable solution, in which the active ingredient is also cyclosporine. The therapeutic indications range from the prevention of transplant rejection, to the treatment of endogenous uveitis (which is another ocular disease).

In the SPC business, you gotta always keep an eye out for the next CJEU case law.

Santen appealed the refusal in front of the Paris Cour d’appel. Instead of deciding on the merits of the case, the court referred some questions to the CJEU regarding the interpretation of article 3 (d) of regulation (EC) No 469/2009, nicknamed the “SPC regulation”.

Article 3 (d) lays down one of the conditions for obtaining an SPC, namely that the MA referred to in the SPC application should be the first authorization “to place the product on the market as a medicinal product”.

The most important ruling from the CJEU regarding the interpretation of this article is Neurim (C-130/11) of July 11, 2012.

According to this ruling, the existence of an earlier MA obtained for a veterinary medicinal product does not preclude the grant of an SPC for a different application of the same product for which the MA has been granted, provided that the application is within the limits of the protection conferred by the basic patent relied upon for the purposes of the application for the SPC.

In Neurim, an SPC application for a formulation of melatonin useful for the treatment of insomnia had been initially refused due to the existence of an earlier MA for melatonin in a veterinary product (more precisely for the regulation of seasonal breeding activity in sheep). The Court of justice clarified that, in such a situation, the SPC application should not have been refused for this reason.

However, with the Luxembourg Court, most clarifications usually raise additional questions.

Case in point, Neurim was widely interpreted as a reversal of earlier case law (in particular Pharmacia Italia C-31/03 and Yissum C-205/05) but the Court of justice did not say a word about this in the judgment. More importantly, people wondered how far the principle set out in Neurim should extend. That is, how should the expression of the Court “a different application of the same product for which [an MA] has been granted” be interpreted?

Does the expression refer to veterinary vs. human use only, like in the facts underlying the Neurim judgment? Or does it extend to the treatment of different human diseases? Or to different formulations or modes of administration?

In the Santen case, the INPI noted that there are various interpretations of Neurim across Europe. The French administration’s current position is that Neurim should be applied in a “measured” manner.

The INPI believes in particular that there are two reasons why Neurim does not apply to Santen’s SPC application:

  • First, the scope of the basic patent should correspond to that of the MA which is relied upon, like in the Neurim case (cf. the proviso that “the application is within the limits of the protection conferred by the basic patent” in the order of Neurim). This is not the case here, said the INPI, because the basic patent mainly relates to a particular emulsion. Keratitis is only one of the diseases that can be treated according to the patent. Uveitis (i.e. one of the diseases treated by the former drug Sandimmum®) is also mentioned in the dependent Swiss-type claim of the patent.
  • Second, the MA which is relied upon should relate to “a new therapeutic field”, i.e. a new medical specialty, or a drug, the active ingredient of which should have a different mode of action from that exerted in the drug of the previous MA. Again, this is not the case here, the INPI argued, since both MAs relate to the treatment of inflammations of the human eye, by the same mode of action of the same substance (cyclosporine). Thus, a new medical use is not sufficiently demonstrated. The INPI’s position seems to be that a “new medical use” in this context has to be appraised in a stricter manner than in the context of a patentability appraisal.

Santen challenged this position, and stated that the conditions demanded by the INPI cannot be derived from the Neurim judgment. Santen insisted on the fact that Sandimmum® and Ikervis® have different therapeutic indications, and treat different diseases in different parts of the eye. Their respective formulations, dosages and modes of administration also differ.

The court considered both positions and came to the conclusion that “when reading the Neurim judgment it is not possible to interpret this notion of different application, which is essential for deciding the case, with any certainty”.

Thus, the Cour d’appel stayed the proceedings and referred the two following questions to the Court of justice:

1. Should the notion of different application under the CJEU Neurim judgment C-130/11 of July 19, 2012:

– be understood in a strict manner, namely be limited to the sole case of a human application following a veterinary application;

– or relate to an indication pertaining to a new therapeutic field, namely a new medical specialty, relative to the previous MA, or to a drug in which the active ingredient has a different action from the one it has in the drug of the first MA;

– or more generally, in view of the purposes of Regulation (EC) No. 469/2009 aiming at setting up a balanced system taking into account all the interests at stake, including those of public health, be appraised according to more demanding criteria than those applied for appraising patentability of the invention;

– or on the contrary be understood in an extensive manner, i.e. as including not only different therapeutic indications and diseases, but also different formulations, dosages and/or modes of administration.

2. Does the notion of application within the limits of the protection conferred by the basic patent under the CJEU Neurim judgment C-130/11 of July 19, 2012 imply that the scope of the basic patent should correspond to that of the MA relied upon, and consequently be limited to the new medical use corresponding to the therapeutic indication of said MA?

In other words, some very precise and extensive questioning of the scope of article 3 (d) of the SPC Regulation. That said, the likelihood that the Court of justice will clearly answer by yes or no as to whether Neurim applies to different diseases, different formulations, etc. is probably close to zero. It is much more likely that the Luxembourg judges will offer a new general test which will still leave room for interpretation. This is indeed the game that has been played quite extensively in the past few years.

The other thing is that there is also another referral pending on the interpretation of Neurim, namely the Abraxis case (C-443/17).

Unsurprisingly, this is a referral by the High Court in the UK. The question asked is the following:

Is Article 3(d) of the SPC Regulation to be interpreted as permitting the grant of an SPC where the MA referred to in Article 3(b) is the first authorization within the scope of the basic patent to place the product on the market as a medicinal product and where the product is a new formulation of an old active ingredient?

In Abraxis, an SPC application was filed based on an MA for paclitaxel bound to albumin nanoparticles, but the issue was the existence of an earlier MA for solvent-based paclitaxel.

The Advocate General’s opinion in Abraxis is expected to be released on October 25, 2018. Will the Santen referral remain useful or relevant once Abraxis is decided?

As always with the SPC saga, to be continued.


CASE REFERENCE: Cour d’appel de Paris, pôle 5 chambre 1, October 9, 2018, SAS Santen v. Directeur Général de l’INPI, RG No. 17/19934.