In a series of rulings dated 8 September 2016, the EU General Court (General Court) upheld the European Commission (Commission) decision of 19 June 2013 in the Lundbeck case.
The Commission had imposed heavy fines on Lundbeck and four generic companies with whom Lundbeck had entered into agreements which delayed the entry of generic versions of Lundbeck’s blockbuster antidepressant medicine, citalopram.
The Commission found that these agreements constituted an infringement by object within the meaning of Article 101 TFEU (which like the “per se” rule under U.S. antitrust law, dispenses with the need for an inquiry into their competitive effects) and imposed a €93.7 million fine on Lundbeck and fines totalling €52.2 million on the four generic companies.
The 8 September 2016 rulings are significant because they are the first decisions from the General Court regarding the compatibility of these types of so-called “pay for delay” agreements with competition law and the General Court upheld the Commission’s conclusion that the agreements at issue were infringements by object.
Factual and procedural background
Lundbeck held several patents, covering the active ingredient citalopram, approved for the treatment of depression. These patents were to expire between 1994 and 2003, but in the late 1990s, Lundbeck obtained several additional patents covering new processes for the production of citalopram. The additional patents, however, did not prevent generic companies, namely, Merck Generics UK (Merck GUK), Alpharma, Arrow, and Ranbaxy, from taking significant steps to market generic versions of citalopram in certain EEA countries. As a result, Lundbeck threatened infringement actions and entered into settlement agreements with several generic companies in 2002. Under these agreements, Lundbeck made lump sum payments to the settling generic companies in exchange for their commitment not to enter the citalopram market until various dates throughout the year 2003.
The Commission found that these agreements were anticompetitive within the meaning of Article 101§1 TFEU and issued, on 19 June 2013, the first decision addressing “pay-for-delay” agreements in Europe. The Commission considered that their anticompetitive object was established on the basis of a series of three criteria, which were later confirmed in Fentanyl (December 2013) and Servier (July 2014):
- The generic companies concerned were at least potential competitors of the originator when these agreements were entered into;
- The purpose of these agreements was to limit the commercial autonomy of the generic companies by restricting their ability to enter the market to an extent that could not have been achieved through court proceedings;
- Lundbeck transferred significant value to the generic companies, which did not result from an assessment by the parties of the merits of the patent rights at issue, but rather seemed to correspond more or less to the profit anticipated by the generic companies if they had actually entered the market.
The Commission thus concluded that by preventing or by rendering pointless any type of challenge to Lundbeck’s patents before the competent courts, the agreements at issue had allowed Lundbeck to transform the uncertainty necessarily associated with the outcome of such litigation into the certainty that the generic products would not enter the market.
All the companies fined lodged appeals before the General Court seeking the annulment of the Commission’s decision, but the General Court dismissed all of their contentions.
The General Court confirms that the generic companies qualify as “potential competitors”
Lundbeck challenged the applicability of Article 101§1 TFEU to these settlement agreements on the ground that the generic companies did not constitute potential competitors. Lundbeck argued that it held patents that were still valid at the time the agreements were concluded, which thus rendered the generic companies’ possibilities to enter the market uncertain. Lundbeck argued in particular that potential competition could not be inferred solely based on the possibility that the generic companies might either challenge the validity of a patent or launch generic products “at risk”, i.e. while bearing the risk of infringement actions brought by the originator.
This raised an important legal issue involving the application of the notion of “potential competition” in situations where a company holds intellectual property rights that grant it legal exclusivity, and thus afford it protection from competition during the validity of those intellectual property rights.
To assess this question, the General Court referred to legal principles laid down in previous cases. Applying those principles on the numerous facts gathered by the Commission in the course of its investigation, the General Court concluded that there existed potential competition from the generic companies, in spite of the risk of infringing Lundbeck’s patents.
The General Court held that the existence of a patent by itself was not sufficient to set aside any competition from generic products, as these could notably be produced by other means without necessarily infringing Lundbeck’s intellectual property rights at issue, which only covered production processes. The General Court did stress that patents are presumed valid until the contrary is proven; however, this presumption can be set aside by claims of invalidity. In any event, even assuming patent validity, a patent confers upon its holder protection only within the limits of what can be protected, and thus does not allow its holder to assume that a generic company infringes its rights without having previously made a demonstration to that effect. In the absence of such a demonstration, as was the case here, generic companies should be considered as potential competitors within the meaning of Article 101§1 TFEU.
The question remains as to whether the General Court’s finding is specific to the pharmaceutical sector or whether it can be applied to any sector involving intellectual property rights.
The General Court upholds the application of the notion of infringement by object to impose fines based on this type of agreement
The Lundbeck case was the first time that the General Court was asked to rule on the lawfulness of settlement agreements entered into between originator and generic pharmaceutical companies. In its 2009 sector inquiry into the pharmaceutical sector, they were considered by the Commission as being the most problematic (the so-called “category B2”) in that they involve a transfer of significant value and delay market entry of generic products.
Of particular importance in the General Court’s ruling was the application of the notion of infringement by object, which eliminates the need to prove the agreements’ harmful effects on competition. Lundbeck submitted that by simply considering that (i) the agreements at issue were similar to market-sharing agreements which did not involve the implementation of patents and that (ii) a transfer of value could alone render a patent settlement agreement anticompetitive, the Commission was not entitled to apply the notion of an infringement by object. Therefore, according to Lundbeck, the Commission should be required to prove the agreements’ harmful effects on competition.
The General Court however considered that the settlement agreements at issue were similar to classic market-sharing cases in which a company commits, in return for compensatory payments, to stay out of the market. Building on its previous experience, and notwithstanding the novelty and the specificity of the question raised, the General Court held that the Commission could therefore apply the notion of infringement by object.
In the present case, the General Court set aside the fact that the agreements were designed to settle intellectual property right disputes. It stressed in particular that (i) none of them was entered into in the context of litigation proceedings, but rather before any such litigation, and that (ii) they did not include any agreement by Lundbeck not to initiate infringement proceedings after the expiry of the agreements. According to the General Court, these agreements thus were aimed at excluding potential competitors from the market, rather than settling an intellectual property dispute.
The General Court took into account the fact that (i) Lundbeck, and several generic companies which ultimately refused to enter into these agreements, were aware of the competition concerns and (ii) the parties to the agreements at issue, and in particular Lundbeck itself, doubted the validity of additionnal patents would stand if challenged before the courts. The General Court pointed out that Lundbeck estimated that there was an approximately 50 to 60 percent chance that one of its additional patents – related to a crystallization process – would be found invalid. Although these percentages are not, by themselves, sufficient to characterize the anticompetitive nature of the agreements at stake, they may nonetheless constitute a frame of reference in future Commission analyses of settlement agreements.
The General Court finally considered whether or not the transfer of value from the originator to the generic should be sufficient, by itself, to render an agreement anticompetitive by object. On this issue, the General Court upheld the Commission’s approach based, inter alia, on the fact that the payments did not result exclusively from the assessment of the strength of the patents at issue, but had rather been calculated on the basis of the turnover the generic companies had relinquished by refraining from entering the market. Therefore, according to the Commission and the General Court, the generic companies were induced to accept the limitations on entry by the size of the reverse payments rather than the existence of Lundbeck’s process patents or even the desire of avoiding potential litigation costs.
The General Court does not exclude the compatibility of such agreements with competition law
The question remains of whether or not the notion of infringement by object will be applied to all so-called “pay for delay” settlements or if the Lundbeck rulings will be limited to settlements presenting similar facts. What seems nonetheless clear, at this stage, is that the competition risks arising from such settlements will be more significant where there is evidence that the originator has substantial doubts regarding the validity of its patents, and that the value transferred to a generic company is based on the likely profit that it would be giving up by agreeing not to enter as opposed to an assessment of the risks or costs of patent litigation.
The General Court nevertheless carefully explained that the Commission’s approach was not prescriptive and that not all patent settlement agreements containing reverse payments are necessarily contrary to Article 101§1 TFEU. In Lundbeck, it was the disproportionate nature of such payments, combined with several other factors, that led to the conclusion that the agreements were contrary to Article 101§1 TFEU.
For example, a settlement agreement entered into in the context of litigation proceedings relating to intellectual property rights – and not only relating to the threat of infringement actions – might be analysed differently – and perhaps not considered as an infringement by object, notwithstanding the fact that it contains a transfer of significant value. Even in such a scenario, however, the Commission would probably consider whether or not the value is based on an assessment of the likely outcome of the patent litigation (in particular its costs or the amount of damages expected) or, rather, on the profits that the generic is agreeing to give up by staying out of the market for a particular period of time.
In sum, the boundary between a lawful agreement and an unlawful one remains thin and, as is typically the case in competition law, the assessment of the anticompetitive character of such an agreement will be contingent upon numerous factors that must be assessed on a case-by-case basis. It is expected that the General Court’s Servier ruling, which is expected in 2017, will provide additional guidance.
Comparing EU and U.S. Perspectives on “pay-for-delay”
Finally, the General Court’s approach in Lundbeck can be contrasted with the U.S. approach to “pay-for-delay” agreements. In a 2013 decision, Federal Trade Commission v. Actavis, Inc., the U.S. Supreme Court ruled that, notwithstanding policies favoring intellectual property and private settlement of legal disputes, such agreements may violate the antitrust laws. In contrast to the General Court’s determination that “pay-for-delay” agreements are unlawful by object, however, the Supreme Court held that they should be subject to a rule of reason analysis, which requires an analysis of the competitive effects of the challenged agreement. In so ruling, the Supreme Court rejected both a rebuttable presumption of illegality as proposed by the Federal Trade Commission, and the "scope of the patent test" adopted by several lower courts, under which agreements within the scope of the patent are deemed lawful.
Notwithstanding the differing legal standards applicable to “pay-for-delay” agreements in Europe and the United States, courts and agencies in both jurisdictions are likely to consider certain types of evidence in evaluating the lawfulness of a particular agreement, including the size of the payment from the originator to the generic, the amount of profit the generic otherwise may have obtained through market entry, the parties’ respective views on the likelihood of prevailing in the underlying patent litigation, and any other justifications for the payment, such as services provided by the generic to the originator and saved litigation costs.
In the United States, the courts have encountered challenges in applying the relatively ambiguous standard established in Actavis. Although the infringement by object standard affirmed by the General Court in Lundbeck appears to be more straightforward, the Commission is nonetheless likely to face its own challenges in applying that standard to the complex area of “pay-for-delay” agreements.