Pay-for-delay agreements can violate Article 101 of the TFEU. Most recently, in Lundbeck v. Commission (2021), the European Court of Justice (E.C.J.) held that the pay-for-delay agreements at issue constituted restrictions of competition by object under Article 101(1). These agreements between Lundbeck and four generic suppliers concerned Lundbeck’s antidepressant citalopram (sold under different brand names in different member states). The generic suppliers committed not to produce or sell generic citalopram or transfer their existing marketing authorizations to any third party, and to sell all existing citalopram inventory to Lundbeck. In exchange, Lundbeck agreed to provide the generic suppliers with significant cash payments and guaranteed profits via distribution agreements, with the amounts based on the profits the generics would have earned had they entered the market. When the parties executed these agreements, Lundbeck’s citalopram patent had already expired, but Lundbeck still held patents on some methods of making citalopram (i.e., process patents). The European Commission (E.C.) determined these agreements restricted competition by object and fined the parties € 146 million in June 2013. The European Union General Court (G.C.) affirmed in September 2016. The parties then appealed to the E.C.J.
In affirming the G.C.’s decision, the E.C.J. made several key points relevant to future pay-for-delay cases. First, as to whether the generics were potential competitors to Lundbeck under Article 101, it was not necessary to assess the strength and likely validity of Lundbeck’s remaining patents, to prove definitively that the generics’ products would not have infringed them, or to demonstrate with certainty that generics would have entered the market and succeeded. The generics’ intention and ability to enter the market, and the fact that they had taken affirmative steps to do so, were sufficient. Second, as to whether the agreements were restrictions by object, that concept “must be interpreted strictly and can be applied only to some agreements”. Nevertheless, by-object treatment is not only appropriate, but necessary when “it is plain from the examination of the [reverse-payment] settlement agreement concerned that the transfers of value . . . cannot have any explanation other than the commercial interest of both the holder of the patent at issue and the party allegedly infringing the patent not to engage in competition on the merits”. Such agreements “can clearly be characterised as ‘restrictions by object’” because they involve “competitors deliberately substitut[ing] practical cooperation between them for the risks of competition”. Determining whether a particular pay-for-delay agreement fits this description requires a case-by-case assessment of whether the transfer of value from the originator to the generic was sufficiently large to incentivize the generic not to enter and compete on the merits. Turning the agreements at issue, the Court rejected Lundbeck’s argument that they merely secured protection within the scope of the process patents, in part because the Court found no basis to assume that those patents were valid and would have been infringed. Consequently, the G.C.’s finding that Lundbeck’s payments under these agreements completely extinguished the generics’ incentives to challenge Lundbeck’s patents was sufficient to establish a restriction by object.
The E.C.’s assessments of pay-for-delay agreements in other cases have been more expansive than Lundbeck. In Servier (2014) and Cephalon (2020), the E.C. found that the patent litigation settlement agreements at issue violated Article 101(1) not only by object, but by effect as well. According to the Servier decision, the E.C. conducted a by-effect analysis “for the sake of completeness”. (Unlike the Lundbeck agreements, the Servier and Cephalon agreements settled active patent litigation.) The Servier decision also applied Article 102, concluding that Servier had abused its dominance in a perindopril-only product market. In Servier v. Commission (2018), the G.C. affirmed the E.C.’s by-object and by-effect determinations for all agreements except the license agreement between Servier and Krka, a generic manufacturer. The E.C. had treated that agreement as part of the reverse-payment settlement between Servier and Krka, and thus as part of a single violation of Article 101(1). But the G.C., noting that licensure may be a legitimate way to settle patent dispute, analyzed the license agreement separately and found the E.C. had met neither its by-object nor its by-effect burdens. The G.C. also rejected the E.C.’s abuse-of-dominance determination after finding the E.C.’s market definition too narrow. Servier and Cephalon are pending on appeal before the E.C.J and G.C., respectively.
Pay-for-delay agreements can violate Section 1 of the Sherman Act and, in suits brought by the Federal Trade Commission (F.T.C.), the Section 5 of the F.T.C. Act (which reaches practices that violate the Sherman Act and other antitrust laws). The key pay-for-delay case in the U.S. is F.T.C. v. Actavis (2013), in which the Supreme Court held that pay-for-delay agreements that permit entry within the patent term may be subject to rule-of-reason analysis. The agreements at issue concerned Solvay’s patent on Androgel. Pursuant to the process for generic entry established under U.S. law by the Hatch-Waxman Act, Actavis and Paddock alleged that Solvay’s patent was invalid and not infringed by the generic testosterone gels they planned to introduce, and Solvay responded by suing both for patent infringement. (The Supreme Court’s opinion details the Hatch-Waxman process.) To settle the litigation, Actavis and Paddock agreed to delay the launch of their products for approximately nine years and to help market Androgel in the meantime. In return, Solvay agreed to pay the generics hundreds of millions of dollars. The case reached the Supreme Court on appeal from the U.S. Court of Appeals for the Eleventh Circuit, which had dismissed the F.T.C.’s complaint based on the scope-of-the-patent test—i.e., antitrust generally does not condemn a pay-for-delay settlement if it does not limit competition more than the exclusionary potential of the patent. The Eleventh Circuit concluded that these settlements satisfied the test—and were thus not subject to antitrust scrutiny—because the generics’ commitment not to enter the market was set to expire before Sovlay’s patent. In reversing, the Supreme Court rejected the scope-of-the-patent test, explaining that it requires an assumption that the patents are valid and infringed, which is the very thing the settled litigation called into question and left unresolved. Indeed, pay-for-delay agreements can cause anticompetitive harm because they eliminate the risk of competition posed by meritorious patent disputes.
The Court also provided guidance on what the rule-of-reason analysis should consider. Although the scope-of-the-patent test is inappropriate, courts need not go to other extreme of evaluating in detail a patent’s validity. Instead, courts often may infer from a large and unjustified reverse payment that the patent holder “has serious doubts about the patent’s survival” and that “the payment’s objective is to maintain supracompetitive prices to be shared among the patentee and the challenger rather than face what might have been a competitive market.” At the same time, a reverse-payment settlement can have “offsetting or redeeming virtues” and does not raise the same antitrust concerns if it “reflects traditional settlement considerations, such as avoided litigation costs or fair value for services”. Thus, factors to consider when deciding whether a reverse payment is likely anticompetitive include “its size, its scale in relation to the payor’s anticipated future litigation costs, its independence from other services for which it might represent payment, and the lack of any other convincing justification.” Actavis remains the controlling authority and starting point when applying U.S. antitrust law to pay-for-delay settlements.
In Impax Laboratories (2019), the F.T.C. decided its first pay-for-delay case after Actavis. Sitting as an adjudicative body, the five-member Commission unanimously held that the F.T.C.’s prosecuting attorneys had proven that the agreement violated Section 5 of the F.T.C. Act. Impax appealed to the U.S. Court of Appeals for the Fifth Circuit, which affirmed the Commission’s decision, in Impax Laboratories v. F.T.C. (2021). The offending agreement involved Endo compensating generic supplier Impax to abandon its Hatch-Waxman challenge of Endo’s patent and to delay launching a generic version of Endo’s branded opioid Opana ER for two-and-a-half years. The court agreed with the Commission’s application of “the burden-shifting analysis that courts have used in other rule of reason cases, as informed by the Supreme Court’s reasoning in Actavis.” At the first step, it required and found proof of “a large, unjustified payment . . . made in exchange for deferring entry into the market or for abandoning a patent suit, plus the existence of market power.” To calculate the size of the payment, the Commission (consistent with other courts) “consider[ed] all value—cash and otherwise—that the branded drug manufacturer transfer[ed] to the generic through the settlement (including any side agreements that contemporaneous timing or other circumstances indicate should be considered part of the same transaction).” Endo’s commitments that could be attributed to the reverse payment were enough to constitute a large, unjustified payment to Impax, as their value “exceeded the payor’s [Endo’s] anticipated saved litigation costs plus the value to be rendered under the agreement[,] and . . . no other clear justification present[ed] itself.” The Fifth Circuit likewise found the payment large, “comparable to other cases where courts have inferred anticompetitive effect”, and not otherwise justified. Both the Commission and the court also found this payment restrained trade. The required showing is not that generic competition was certain, or even probable, absent the payment, but that there was a risk or “a real threat” of such competition. Proceeding to the second step, the settlement’s alleged procompetitive benefits, the Commission found that they were not adequately linked to the reverse payment to delay Imapx’s entry, and that they could have been achieved without the challenged restraint. The Fifth Circuit reached no conclusion on the former finding, but determined that the latter was supported by the evidence and sufficient to conclude the court’s review in the Commission’s favor.