Cross-border litigation after GAT v LuK and Roche v Primus: the future of the Italian Torpedo

01 January 2009

Felix Rödiger

Two years after the European Court of Justice (ECJ) rendered the two decisions GAT v LuK {Case C-4/03) and Roche v Primus {Case C-539/03) on 13 July 2006, the effect on cross-border litigation is still being discussed. One question is the impact on the Italian torpedo. The Italian torpedo is a cross-border declara¬tory action in a country known for slow proceedings, such as Italy or Belgium, to avoid an infringement decision in a country with faster proceedings, such as Germany.

Against this background, this article examines:

  • GAT v LuK.

  • Roche v Primus.

  • Consequences of the ECJ decisions.

  • Future of the torpedo defence.

Cross-border scenario

A cross-border patent dispute can be considered from both the active and the passive side.

  • On the active side, the patentee can ask for an injunction before a national court concerning a patent valid in another member state or, as is often the case, for several national parts of one European patent.

  • On the passive side, the potential infringer can apply to court for a negative declaratory action to establish that several national parts of one European patent are not infringed.

This is known as a "torpedo", if filed at a forum known to be slow.

Both scenarios have one thing in common: the claimant seeks a court order to cover patents from abroad, that is, from another country than the jurisdiction of the original court. This goes back to the concept that a European patent is not only a bundle patent in its application phase, but also that it should, under Article 69 of the European Patent Convention (EPC) and its Protocol of Interpretation, be interpreted consistently in each member state. This common interpretation would then allow a national court not only to decide over its national patents but also over the infringe¬ment of the claims of other national parts of a European patent.

In the mid to late 1990s, different national courts had developed different case law on whether cross-border actions were admis¬sible. Passive cross-border actions (torpedoes) were first filed in Belgium and subsequently in Italy and some of them are still pending, even today. Both countries’ courts would of course argue that such negative declaratory actions covering other national parts from abroad are inadmissible. The question was just how long it would take them to make the decision.

Further, it was the Dutch court that allowed (active) cross-border actions under what they had called the “spider-in-the-web-doctrine”.

This meant that a cross-border action was admissible, if there was a group of defendants and the infringer’s business activities were organised from one central place, which was in The Netherlands.

While the Dutch court put their focus more on the place of infringing activities, the German court allowed (active) cross-border actions if the defendant was sued at the court in his place of domicile, for infringement in more than one country. To decide infringement of national patents from abroad, the German court had to apply the relevant national patent law. This meant the court would have to allow nullity objections, which under German law are generally not admissible, since in Germany nullity and infringement are dealt with before two different courts (the so-called “dual system”). According to Article 16(4) of the Brussels Convention on Jurisdiction and the Enforcement of Judgments in Civil and Commercial Matters 1968 (now Article 22(4) in Regulation (EC) No. 44/2001 on jurisdiction and the recognition and enforcement of judgments in civil and commercial matters (Brussels Regulation)) a decision over the validity (that is, against everyone, not just between the parties) was always limited to the exclusive jurisdiction of the country in which the patent was registered.


In December 2002, the appeal court of Düsseldorf referred to the ECJ the question of whether a “foreign” national court can decide a validity objection concerning a patent from abroad. This question was decided in the well-known case of GAT v LuK. The ECJ ultimately decided that validity could not be decided by foreign courts (but infringement could still be decided on):

“Article 16 (4) of the Brussels Convention is to be interpreted as meaning that the rule of exclusive jurisdiction laid down therein concerns all proceedings relating to the registration or validity of a patent, irrespective of whether the issue is raised by way of an action or a plea in objection”. 

The question remains whether the GAT v LuK decision also deals with specific German (and Austrian) issues of separation of infringement and nullity actions. German law does not allow the invalidity defence in infringement proceedings. Such a defence can only be pursued indirectly, by filing a separate nullity action and asking the infringement court to stay the infringement action to await the outcome of the nullity action.

However, the GAT v LuK decision does not seem to deal with this German scenario. It only refers to countries such as France, Italy and the UK, where the nullity defence can be brought in a separate action or as a simple objection in the infringement action. The ECJ merely confirmed that nullity actions against a national part of a certain European patent can only be conducted in the jurisdiction for which that patent was registered. According to the ECJ it does not make a difference if a nullity defence is raised within infringement proceedings or as a dedicated nullity action, as is required in Germany. Therefore, the GAT v LuK decision mainly applies to disputes arising in relation to patents from member states other than Germany or Austria. GAT v LuK does, however, clearly rule that as soon as the defence of nullity is raised in proceedings before a foreign court, such action is inadmissible and must therefore be terminated,
as the foreign court is not competent to rule on validity.

The question then is whether the German infringement proceedings “relate to the validity of a patent”. As stated above, the German infringement judge will likely have to deal with questions of validity, namely when a separate nullity action (or opposition) is filed. It could be argued that Article 22(4) should be interpreted in such a broad sense that the infringement action should become inadmissible not only when the nullity objection is raised, but also when the nullity stay is requested. In such a broad sense, a German infringement action might also “relate” (in the sense of Article 22(4)) to the validity of the patent in the case.

How a German court may deal with such a scenario will be decided in due course. The Higher Regional Court Mannheim is currently handling a patent infringement case concerning a French, an Italian and also an Austrian patent. It is expected the court will dismiss the French and Italian action in accordance with the Gat v LuK decision. But the court will for the first time (after Gat v LuK) have to decide whether a German infringement judge is competent to decide infringement of an Austrian patent. Austria and Germany are the only two countries in Europe where infringement and validity are dealt with before different courts. The court has announced it will render its judgment in 2008.

Roche v Primus

On the same day the GAT v LuK decision was handed down, the ECJ in another cross-border case, Roche v Primus, decided that Article 6(1) of the Brussels Regulation (granting jurisdiction over cases that are deemed to be connected) does not permit suing several foreign companies together with one local company before a single national court. The ECJ dismissed the argument (as raised by Dutch courts) that not allowing multiple actions against foreign subsidiary companies in a single jurisdiction, therefore requiring claimants to pursue separate actions in each jurisdiction, would lead to “irreconcilable judgments”. The Court asserted the facts of each case concerning infringement in a separate jurisdiction would not be the same, as each ruling would not only apply to a different defendant, but would also relate to distinct infringing activities. Further, the legal findings could not be contradictory either, since national parts of a European patent are governed only by national laws of that jurisdiction. The Dutch spider-in-the web-doctrine was therefore declared dead.

Consequences of the ECJ decisions

This set of decisions was first thought to bring an end to all cross-border decisions. However only two months later the Dutch court in September 2006 (District Court of Den Haag, Rechtsbank s’Gravenhage) granted yet another cross-border injunction, not in an action on the merits but rather in a preliminary injunction case. The Dutch court argued that the GAT v LuK judgment and Article 16(4) of the Brussels Convention (Article 22(4) of the Brussels Regulation) only applies to cases on the merits, not to preliminary injunctions (PIs) which are exempt under Article 31 of the Regulation. Since then, the Dutch court has granted several cross-border PIs in patent matters, though it is the only court that regularly grants cross-border decisions after the GAT v LuK and Roche v Primus decisions in July 2006.

It is important to note that so far the Dutch court has only granted such cross-border PIs against Dutch companies, but not in a typical spider-in-the-web-scenario against a combination of Dutch and foreign infringers. Yet another open question surrounding cross-border actions is whether the Dutch court would be prepared to state that Article 6(1) of the Brussels Regulation does not apply to PIs generally. But for the time being, the only likely scenario left for a cross-border decision is a Dutch cross-border PI against one or more Dutch defendants.

Apart from such exceptional PIs, torpedoes filed in slow countries might still be an issue of concern. Most courts in Europe have ignored such torpedo actions right from the beginning and have always argued that it is a misuse of procedural rights. Others, notably the German courts, have in most cases acknowledged the blocking effect of such declaratory actions. There is only one exception where the District Court of Düsseldorf stated that misuse in filing a negative declaratory judgment abroad was so obvious that the German action would not have to be stayed. It is possible that this groundbreaking decision may show us how the torpedo will be treated in the future.

In a typical example of an Italian torpedo as it was filed in the 1990s, the potential infringer would claim invalidity and non-infringement of the Italian patent (application) only, but would not touch on the invalidity of the German patent in suit. Therefore, the main problem remains that the torpedo action will in any event remain admissible until the nullity, or under the broadest interpretation, the German “stay-defence” is raised.  Most importantly, however, the classic torpedo defence is prima facie not affected by GAT v LuK or Roche v Primus, as Article 27(1) of the Brussels Regulation clearly states that the court subsequently called on must suspend proceedings in the same matter until the court first called on has decided on its own competency.

In Erich Gasser GmbH v MISAT Srl (Case C-116/02, 9 December 2003), the ECJ clearly ruled that the Brussels Convention purposes of ensuring legal certainty and preventing contradictory judgments make it mandatory that no exceptions be made to the requirement to suspend. More specifically, the ECJ ruled that misuse by the torpedo claimant could not be construed purely from the fact that the country where the torpedo was launched was generally known to be exceptionally slow in determining its own competence.

Future of the torpedo defence

So what would a German infringement court actually do if confronted with a torpedo situation in the light of GAT v LuK and Roche v Primus? One option might be the application of the well-known misuse argument, which the German courts (unlike most other European courts) previously avoided where possible. Until now, and in line with the Gasser v MISAT decision, there was no room for discretion about the stay decision. The ECJ argument in Gasser still leaves room for basing a misuse argument not on generally slow proceedings in a member state, but on the fact that in a specific case the torpedo claimant is wilfully trying to delay the infringement case and that the actual delay has already attained unacceptable levels. However, the party that makes the claim of misuse by the other party carries the burden of proof.

It may also be possible to argue that in cases where it is obvious that the court first addressed (in our example the Italian court) would eventually have to declare the action inadmissible, the subsequently called on court would not have to suspend proceedings.

The obvious lack of competency might be deemed to arise because filing a nullity action relating to the national part of the patent in suit at the second court would be clearly inadmissible, because it might fall directly under a situation that was already decided on in ECJ decisions like GAT v LuK or Roche v Primus.
In such a scenario the later addressed court should be allowed to anticipate the predictable outcome of the Italian court decision, which would then form an exception to the otherwise compulsory wording of Article 27(1) of the Brussels Regulation. This would allow both courts to continue their actions with the argument that contradictory decisions cannot appear, because the Italian court will eventually declare the action inadmissible. This approach would extrapolate the GAT v LuK and Roche v Primus principles to the classic torpedo situation. A serious caveat to this argument is of course that the ECJ has repeatedly confirmed that only the court initially called on can decide on its own competency, and no other European national court can prejudice this decision.
Another solution might be to combine the outcomes of the GAT v LuK and Roche v Primus decisions with the familiar, but rarely admitted misuse argument, to give the misuse term a new definition.

Misuse might be construed per se, or at least in combination with other elements such as an already overly lengthy procedure, in cases where a foreign court cannot consider itself to be competent because of Roche v Primus. The typical torpedo will usually involve several defendants to prevent infringement proceedings in many countries. Roche v Primus, with its interpretation of Article 6 of the Brussels Regulation, would therefore be applicable to such torpedo set-ups in most cases. The Roche v Primus decision may allow a new definition of what is considered as misuse. However, again the issue of sovereignty of each member state would arise.

One possible way of fending off torpedoes is to license out the patent.  The non-exclusive licensee of a patent is considered to be an identical party to the patent owner, so this in itself is not sufficient. However, if possible the patentee could grant an exclusive licence to a party not involved in the torpedo action, and then bring the infringement action in the name of this exclusive licensee.

Although the need to wait for a decision on the competence of the earlier court still remains, the classic torpedo action has obviously become less attractive. Both Italian and Belgian courts have made efforts to combat the stigma of being “slow countries” suitable for torpedo actions. Expedited patent proceedings or declarations of non-competency in these countries do not, however, imply the demise of the torpedo, as the EU has gained several new member states that might just have courts that operate on an even less time-pressed basis than Italian or Belgian courts. Therefore, we may have to expect “Romanian or Bulgarian torpedoes” in the future, if German courts continue to accept the blocking effect of torpedo actions. Whether they do or whether the misuse argument will be raised more easily after Roche v Primus is an open question, likely to be decided on a case by case basis only.