'By-object' illegality thesis


Two years after the US Supreme Court's ruling in Actavis, the EU General Court has opined for the first time on the legality of reverse payment settlement agreements in the long-awaited Lundbeck judgment.(1) But those waiting for practical guidance and clear guidelines delineating lawful from unlawful settlements will be disappointed. The judgment is a muddle of conflicting ideas about patents and competition law from which no coherent counselling standard emerges.

Hopefully the inevitable appeals will produce a clearer standard by which to judge settlements. They are an essential and fundamentally pro-competitive practice, prevent wasteful litigation, create commercial room to manoeuvre free from suit and free up resources for more productive activity. The Lundbeck judgment, by contrast, creates a litigators' charter. Virtually all generics are presumed to be potential competitors. The only good patent is a litigated patent. At-risk entry and litigation is competition, implying that settling is cartelising. It is a precedent which, if left to stand, attaches antitrust risk to benign and pro-competitive settlements.

The decision can be summarised as follows:

  • Lundbeck held that a settlement involving a potentially competing generic, a restriction on generic entry and a value transfer may be automatically illegal by-object with no inquiry as to whether the agreement has an impact on the market.
  • The decision creates a substantial divergence from the market assessment required by the US Supreme Court's Actavis 'rule of reason' standard.
  • The court paid lip service to a patent's presumption of validity. This would prima facie dispose of any antitrust case. If a presumptively valid patent precludes all competition, then the generic is not a competitor, but the court effectively barred any evidence as to the strength of the patent. The only evidence it would accept was a final court ruling on validity or infringement, which is impossible in a settlement case. The court even barred evidence of judicial rulings on validity which post-date the settlement. If the parties doubted the patent's strength at the time of settlement, they must be regarded as competitors.
  • A better reading of the judgment is that further aggravating factors are required before a settlement is a by-object violation – in particular, evidence that the value transfer is linked to the generic's expected profits had it entered the market and that the generic received no guaranteed early entry date free from suit – but even these essential points are unclear. Conversely, the court found it irrelevant that the generic promised only not to enter by a patent-infringing method. Even if the promise fell within the scope of the patent, the settlement could restrict competition.
  • The judgment offers no administrable legal standard for assessing the by-object illegality of reverse payment settlement agreements. Confusion and misconceptions arise as to the role of patents, the significance of the aggravating factors and the application of by-object categorisation. The court attached no weight to the pressure on originators to settle or face irreversible harm through the genericisation of the market price, compounded by international reference pricing, or the risk of divergent patent rulings between countries and the unreliability of injunctive relief in many states. These were disregarded among other things, noting, the public interest in health systems obtaining cheap generic products.
  • The few practical learnings offered by the judgment are of limited comfort. The court identified settlement types which are presumptively lawful (in particular, a generic voluntarily staying off the market in return for a cross undertaking in damages if the originator loses the patent case). There can be some comfort that settling suits around an original compound patent would be more benignly reviewed.
  • Of far greater concern is the sentiment expressed by the court that it will accept only the validity of patents litigated to final judgment and that patent infringement by generics is an expression of competition which the antitrust law should protect, which prima facie renders any settlement suspect. These conclusions are based on fundamental misunderstandings of the dynamics of patents and patent litigation. They threaten benign and pro-competitive settlements that enable wasteful litigation to be ended and resources more profitably deployed.


Lundbeck's compound patents on citalopram had expired, but patents covering commercially efficient methods of manufacture and crystallisation remained in force as generics started to enter the market. Lundbeck threatened suit and entered into a variety of settlement agreements with the generic companies, including German Merck, Alpharma, Arrow and Ranbaxy. The generics agreed:

  • not to market infringing citalopram; and
  • to sell citalopram stocks to Lundbeck, resell Lundbeck citalopram (with a guaranteed profit element) or receive payments in place of damages or litigation costs.

The European Commission investigated and found the settlement agreements to be illegal market-sharing agreements, fining Lundbeck and the generics €146 million.

'By-object' illegality thesis

Central to the commission's findings and the appeal was whether settlement agreements with reverse payments from originator to generic were automatically illegal – in EU parlance, a 'by-object' violation. This contrasts with a 'by effects' violation which requires an assessment of anti-competitive effects. The latter places a more onerous burden on the commission to engage in a detailed economic review of the markets affected.

This is by no means an abstruse academic point. By equating settlements with cartels, the court largely bypassed the teachings of Cartes Bancaires – a case in which the European Court of Justice was highly critical of by-object over-use.(2) By-object illegality should be reserved for the most obviously pernicious violations, such as those involving cartel-type conduct – those agreements which "can be regarded, by their very nature, as being injurious to the proper functioning of normal competition".(3) In contrast, the commission's own guidance states that settling a genuine patent dispute is generally pro-competitive.(4)

Two days after the commission made its Lundbeck decision, the US Supreme Court was faced with a near identical question whether the reverse payment settlement in Actavis(5) should be considered per se illegal (a category of automatic illegality similar to by-object) or an analysis of the market context (a rule of reason analysis similar to the EU by effect standard). The court chose the latter. This was a rule of reason question, so all relevant factors – markets and market share, actual or potential competitors, pro-competitive efficiencies of the agreement – should be reviewed.

On an effects analysis, the Lundbeck agreements would likely survive review. Citalopram is one of many antidepressants and has a relatively low market share in many markets. There were alternative suppliers of generic citalopram beyond those subject to the settlement agreement.

The commission's by-object thesis was complicated and listed the following factors as denoting by-object illegality:

  • the parties were potential competitors;
  • the originator transferred significant value to the generic; and
  • the generic agreed to stay off the market.

The following additional factors were relevant:

  • the value transfer approximated the generic's expected profits;
  • the settlement agreement contained restrictions that were allegedly outside the scope of the patent (they required the generic not to compete in ways that could not have been achieved simply by enforcing the patent); and
  • the settlement agreement did not end the dispute or express a commitment not to sue the generic at expiry of the agreement.

The commission was deliberately vague as to whether all six criteria should be present for by-object illegality or whether the first three sufficed.


Potential competitor
The patent block
The generic defendants argued that they were not yet potential or actual competitors of Lundbeck. If they used processes covered by Lundbeck's patents, entry at-risk would be unlawful and subject to injunctive relief. It is old law that if a company is prohibited by law from competing, it cannot be a competitor.(6)

The court agreed that once granted, a patent's validity is presumed.(7) This would seem to favour Lundbeck's position. A valid patent must block any competition that falls within the scope of the patent's legal monopoly right. The commission had to prove that the generics had workarounds, or would prevail in invalidity proceedings, to overcome that patent block.

Following this, the court's reasoning became more obscure. It found that "a presumption of validity could not be equated with a presumption of illegality of generic products". Rather, Lundbeck had to "prove before the national courts… that those generics infringed… since an 'at-risk' entry is not unlawful in itself".

Notwithstanding the patent block, the court found potential competition proven where:

"[L]aunching the generic product 'at risk'… represents the expression of potential competition, in a situation such as that in the present case where [i] Lundbeck's original patents… had expired [ii] there were other processes allowing the production of generic citalopram that had not been found to infringe other Lundbeck patents… [iii] the steps taken and investments made by the generic undertakings in order to enter the citalopram market… show that they were ready to enter the market and to accept the risks involved in such an entry."(8)

The court found that each of these elements were present in Lundbeck. The compound patent had expired so there were manufacturing processes in the public domain, even if these were insufficient and not actually being used by the generics. The generics had made at least some steps – seeking to procure the active pharmaceutical ingredient (API), assembling a drug master file, commencing the marketing authorisation (MA) application, seeking customers for the generic products – towards selling citalopram.

The court concluded that this was evidence that the patents were not a real block to competition and that generics were prepared to launch at-risk. At -risk competition is something competition laws should protect. No expert patent evidence was required or relevant to this determination. Instead, the investment of resources in the possibility – however slim – of avoiding a patent right was enough to establish the generic as a potential competitor.

To conclude that at-risk entry is competition is overly simplistic. It avoids the question as to whether the products are legally on the market, which depends on whether they infringe a patent. Only once that is known can the existence of competition be proven. By definition, this can never be known in the context of a settlement agreement. It is insufficient to say that the generics are legal until a court decides otherwise. A parking violation or theft remains a crime, no matter whether the law breaker is apprehended and sentenced and it should be no different with patent infringements.

Indeed, patent courts regularly face the existential problem that a patent, or an exclusionary injunction based on a patent right, might have been wrongly granted. They have also provided solutions to those problems; invalid patents are revoked and are considered never to have existed. Therefore, a patentee which obtains an injunction against a potential competitor must run the risk and suffer the uncertainty of being ordered to pay for the damage caused by that injunction if the injunction is ultimately held to be unjustified. Inevitably, the damage suffered will be the lost profit that the generic would have made if it had been allowed to launch.

Crystallisation patent
The inconsistency of the court's standard may be observed when compared to the crystallisation patent held by Lundbeck. This was the subject of extensive litigation in opposition proceedings; it was annulled at first instance by the European Patent Office (EPO), but upheld by the EPO Board of Appeal in 2009. This was an unequivocally valid patent and any generic product falling within its scope would infringe and so could not possibly have been considered competing. They were squarely within the scope of an unquestionably valid patent.

However, the court reasoned that Lundbeck's knowledge of the patent's invalidity came after the event. The impugned settlements were concluded in 2001-2002 and the EPO Board of Appeal determination was in 2009. So at the time of the settlement agreements, the patent's validity was uncertain. Evidence ex post is inadmissible:

"[t]he evidence… dates from after the conclusion of the agreements… at the time the agreements at issue were concluded, the generic undertakings as well as Lundbeck itself doubted the validity of that patent and it was possible that a national court might declare it invalid."(9) (Emphasis added.)

Yet it is the existence of differing assessments of those doubts and possibilities that gives rise to litigation in the first place.

The court thus ruled that Lundbeck bore the burden that it would have prevailed in litigation(10) to demonstrate the patent was an absolute block to competition, but it could not adduce the EPO appeal decision(11) or successful injunctions obtained in national courts(12) as evidence in its favour. This would seem to create a standard that no patent can satisfy unless litigated to the bitter end. At the time of settling there is no court ruling and ex post evidence of patent validity or infringement is foreclosed.(13)

Other impediments to generic entry
Aside from the patent position, the generics pointed out that they were not yet on the market. This was not just a patent issue, but a mix of technical, regulatory and commercial barriers. Some had not yet received a marketing authorisation. Many found it difficult to access the API or experienced technical difficulties or required regulatory approvals to change their manufacturing processes. Early steps to enter the market could not be equated with potential competition in light of these difficulties.

The court had no difficulty finding the generics as potential competitors. The question is whether the existence of the generic as a potential entrant disciplined the competitive conduct of the originator. The court formulated a broad brush test that requires the commission to present evidence that the generic will accept the risks of entry by demonstrating:

  • "the significant investments and efforts already made by the generic undertakings in order to prepare their entry";
  • the generics had "obtained MAs or had taken the necessary steps to obtain one";
  • there were "processes available to produce citalopram without infringing";
  • "no court had found the generic products to be infringing"; and
  • there was "a non-negligible possibility that some of Lundbeck's process patents might be declared invalid".(14)

Finally, "the very fact that Lundbeck decided to conclude [reverse-payment settlements] is a strong indication that it perceived those undertakings as a potential threat".(15)

There was no particular time period in which the commission had to show generic entry was likely – even a period of up to eight years might be plausible enough to create a potential competitive constraint on the originator.(16)

The standard for potential competition is so broadly phrased as to provide no meaningful filter. Preparatory steps or the prospect of generic launch are a prerequisite to the existence of any patent litigation, otherwise there would be no point to the litigation. This is a fact pattern that would describe almost all generics, even those that make initial plans only to find by the time of settlement that the technical, commercial and regulatory challenges cannot be resolved. Many may take these preliminary steps, but commencing these efforts by no means denotes that they have any reasonable hope of being successful. Research on generic development shows that barely one in three products successfully passes through the early development stages.(17)

The patent-related elements of the court's standard are meaningless. Since these are settlement agreements, no court has yet ruled on infringement and a non-negligible possibility of invalidity must surely apply to all patents.

Value transfer
The generic companies had entered into a variety of agreements with Lundbeck – distribution agreements, stock buy-backs, cash payments for avoided damages claims or litigation costs – at the same time as the settlement agreement. The question arose of whether these various forms of consideration were legitimate compensation or denoted illegal intent.

Principal inducement
The court applied a broad brush test. The commission did not need to show that the value transfer was the only plausible explanation for the entry restrictions, but rather that it was principally the size of the reverse payments that induced the generics to limit their freedom.(18)

It concluded that the facts passed this test. The benefits conferred allegedly corresponded to the generics' anticipated profits post-entry.(19) Lundbeck's internal documents purportedly showed the payments were made to induce settlement.(20) There was no reference to avoided litigation costs in most of the agreements(21) and no express commitment not to sue for infringement post term. Yet none of these features, whether individually or collectively, provide a satisfactory means to distinguish a bad settlement from a legitimate one. A patentee which obtains a preliminary injunction must later compensate the generic for its lost profits if the injunction is subsequently overturned. A correlation between value transfer and anticipated profits cannot be the deciding factor. Second, the patentee's intention can surely have no bearing on what induced the generic to limit their freedom (aside from technical or other issues as mentioned above). It would be absurd to suggest that a mention of the costs of litigation (which can in any event be objectively assessed) could save or condemn a particular agreement.

Even if a large payment is the only way of reaching a settlement, it does not follow that the payment is illegitimate. The court was only prepared to countenance as legitimate a settlement involving an originator-to-generic payment that contains no restriction on generic entry.(22) This is a rare type of settlement, since it offers little advantage for the originator.

Risk asymmetry
The defendant's contrary arguments were dismissed. Principal among them was the asymmetry of risk between generic and originator at the point of generic entry. The damage to the originator may be irreversible if there is a downward price spiral in the market caused by generic entry. This is amplified many times across markets by international reference pricing. Other territories' health systems benchmark pricing to the genericised price-depressed market.

The court also dismissed the fragmentation of patent litigation across Europe, the lack of effective injunctive relief in many markets and the different results that can occur from court to court. All factors that put a high degree of pressure on the originator to settle. The generic need only be lucky once in patent litigation to cause a price depression across multiple markets, whereas the originator must be lucky every time to avoid irreparable harm caused by early generic entry.

Therefore, the originator has far more to lose than the generic. So even when sure of its patent position, the originator may be forced to grant additional concessions – including value transfers – to secure a settlement against an aggressive opponent to avoid a catastrophic outcome.

However, these arguments were swept aside,(23) the court summarily concluded that these are features of the pharmaceutical market and patent litigation that are a cost of doing business and therefore to be accepted by the originator. They cannot excuse a reverse payment settlement.

It is claimed that an outcome that resulted in lower prices for health systems:

"illustrate[s] the balance which the member states have struck between the protection afforded to the patent of the originator on the one hand, and the savings for state budgets and for consumers achieved by the entry of generics and the effects of competition on the other."(24)

The court concluded "it is apparent from the broad logic of the decision that the commission applied the concept of restriction by-object in the economic and legal context."(25) The court explicitly rejected the US Supreme Court's rule of reason analysis – akin to by effects under EU law – finding that there is no requirement to align US and EU law.(26)

Other aggravating criteria

The question of which criteria are required to create a violation was left vague by the court. This is a vital component of the analysis. Almost all settlements will satisfy the first three criteria:

  • limitations on the generic (generally not to infringe the disputed patent);
  • a benefit for the generic (since otherwise no settlement would arise); and
  • the likelihood that the parties will be at least potential competitors.

Such an ambiguous legal test risks condemning many entirely benign settlements.

The additional criteria suggest that aggravating factors make a meaningful contribution to distinguishing potentially malign settlements, particularly settlements containing:

  • out-of-patent scope restrictions;
  • no early entry for the generic; and
  • value transfer linked to expected profits.

Scope of the patent
Regarding the first criterion, the court concluded that this is not a necessary condition for illegality.(27) The patent's validity or infringement is not proven until a court upholds infringement or validity.(28) The competition law risk lies in the substitution of this inherent uncertainty as to the final outcome on validity or infringement with the certainty that the generics would not launch at-risk for a specific period of time. An obligation on the generic not to enter with an infringing product falls outside the essence or specific subject matter of patents.(29) In a case of style over substance, a right to oppose generic infringement does not also encompass a right to conclude a settlement agreement.

Although the court was clear in reaching this position, it is incidental since the court agreed with the commission that the settlement contained restraints exceeding the scope of the disputed patents.(30)

Settlement of dispute
The court also found that the settlements at issue did not in fact settle the underlying dispute. There was no obligation on Lundbeck not to sue the various generics for infringement after expiry of the settlement agreement, so the generic obtained no dispute settlement protection:

"[T]he agreements at issue… provided only that the generic undertakings would stay out of the citalopram market… without even providing that at the end of that period they could enter that market without having to face infringement actions from Lundbeck."(31)

The court considered this a crucial distinction in the analysis of the Neolab settlement agreement, which was ultimately considered lawful.(32) However, the court confused the ruling by repeating the commission's thesis that the anti-competitive object of the challenged agreements was demonstrated "since they amount to agreements excluding potential competitors from the market in exchange for payment".(33) Therefore, just how important this factor is to the overall by-object assessment remains to be determined.

Transfer of value linked to generics' expected profits
The court endorsed the commission's conclusion that value transfers approximating generics' expected profits were an indicator of illegal conduct. The commission "was entitled to rely on the fact that [value transfer] appeared to correspond to the profits expected by the generic undertakings if they had entered the market",(34) but it was unclear whether that is a necessary criterion. The court also suggested that a lower sum must be sufficient because it "constituted a certain and immediate profit, without necessitating the risks that market entry would have entailed".(35) The question is whether the value transfer is sufficiently high "to allow generics to accept limitations on their autonomy and reduce their incentives to enter the market".(36) However, this vague formulation offers little comfort to practitioners left to independently judge just how high is too high through an uncompromising competition law lens.

No article 101(3) exemption
In a cursory analysis, the court found no efficiencies justifying the restrictions. There was insufficient evidence that settlements fostered innovation or avoided substantial litigation costs (not least since there was no date at which generics were granted free entry) and even if there was some benefit, there was no evidence as to how the settlements' restrictions were necessary to achieve these savings or how they were passed onto consumers.(37)


The judgment leaves much to be desired in terms of providing a meaningful counselling benchmark for companies seeking to settle patent disputes. There are many misconceptions about the patent system and commercial pressures on pharmaceutical companies. However, the following conclusions can be drawn:

  • Lundbeck's efforts to settle with Neolab were cited with approval. Lundbeck encouraged Neolab to stay off the market voluntarily while a test case with generic Lagap was ongoing. Rather than pay cash or another inducement, Lundbeck offered an undertaking in damages if the Lagap litigation went the wrong way. If Lundbeck lost the case, it would pay Neolab damages for the period it was voluntarily off the market. This is what transpired, but the court considered such a payment in lieu of damages lawful since at the time it was paid, Neolab was allowed to immediately enter the market.(38) Therefore, when originators fear multiple infringing generic entries before a test case can be concluded, the Neolab example shows how breathing space may be negotiated with other generics.
  • The court implicitly found that process patents are insufficient for antitrust purposes. The compound patent of necessity discloses how the molecule is made and later process patents cannot close off these pre-existing, public domain and original processes. The suggestion, however commercially untenable, is that this proves there is a potential for competition as soon as the compound expires. However, by the same token, the court's reasoning would suggest that settling a disputed compound patent is safer. The working assumption is that there are no workarounds to a compound patent.
  • The potential competitor test is wide, but not hopelessly so. The commission must prove that the generic's efforts and resources and views of the patents at issue are such that it can be proven it would enter at-risk. This suggests that a careful consideration of the generic's internal documents on steps taken to enter the market and views on the patent position should give some indication of the risks of settlement. It may be clear from these documents that the generic's efforts are fatally flawed or that it is clear it refuses to accept at-risk entry. Rather, it would enter only if assured there is no patent block via a court declaration of non-infringement or success in opposition proceedings.
  • Although vaguely articulated, it may be argued that in order for there to be a by-object violation, the following must be present: the basic three criteria – generic entry restriction, significant value transfer and potential competition and the aggravating criteria – value transfer approximating expected generic profits and no express commitment not to sue post-term.(39) Conversely, the court stated that a settlement even within the patent's exclusionary scope could still be a by-object violation.(40)

Excluding the final point, it may be possible to craft a settlement without aggravating characteristics and deflecting a by-object analysis.

In contrast, the court made a series of statements about the patent system which appear founded on misconception and providing the potential for reversal on appeal. The court's statements that patents should, if validity or infringement is in doubt, all be litigated to final judgment is particularly worrying:

"[I]f the applicants were so convinced of the validity of their patents, and of the fact that the products that the generic undertakings intended to sell infringed them, they could have obtained orders to prevent market entry before the competent national courts, or, in the event that the generic undertakings unlawfully entered the market, obtained damages from them."(41)

The suggestion that the mere act of replacing litigation uncertainty with the certainty that the generic would not enter the market for a period – and achieving that by means of significant reverse payments – represented the fundamental anti-competitive evil made little sense:

"What matters, therefore, is that there was [litigation] uncertainty… and that those agreements had replaced that uncertainty, by means of significant reverse payments, with the certainty that the generic undertakings would not enter the market during the term of the agreements at issue."(42)

Achieving certainty in the face of litigation risk is the fundamental objective of any settlement agreement.

Lundbeck falls short of offering a coherent compliance standard for lawful patent settlements. It creates an unhealthy and unnecessary antitrust second-guessing precedent which will make it still harder to end litigation legitimately. It is hoped that this will be remedied in the forthcoming judgments in Servier due in 2017 and the likely Lundbeck appeals.

For further information on this topic please contact Bill Batchelor, Fiona Carlin or Melissa Healy at Baker & McKenzie's Brussels office by telephone (+32 2 639 36 11) or email ([email protected], [email protected] or [email protected]). Alternatively, contact Hiroshi Sheraton at Baker and McKenzie's London office by telephone (+44 20 7919 1000) or email ([email protected]). The Baker & McKenzie website can be accessed at


(1) Lundbeck v Commission, T-472/13, ECLI:EU:T:2016:449.

(2) C-67/13 P, Cartes Bancaires, para 58.

(3) C-209/07, BIDS, para 17, C-67/13 P, Cartes Bancaires, para 49.

(4) Technology transfer block exemption guidelines, para 242.

(5) Federal Trade Commission v Actavis 570 US 2013.

(6) EON Ruhrgas and EON v Commission, T 360/09, ECR, EU:T:2012:332, para 84.

(7) Judgment of September 8 2016, Lundbeck v Commission, T-472/13, ECLI:EU:T:2016:449, paras 121-122.

(8) Ibid, para 129.

(9) Ibid, para 145.

(10) Ibid, para 165.

(11) Ibid, para 145.

(12) Ibid, para 268.

(13) The commission did rely on such ex post evidence in reaching its conclusion in Servier (where the patent in question was subsequently held to be invalid) and it will be interesting to see how the court rationalises that approach.

(14) Lundbeck v Commission, T-472/13, ECLI:EU:T:2016:449, para 157.

(15) Ibid, para 181.

(16) Sun Pharmaceutical Industries and Ranbaxy (UK) v Commission, T-460/13, ECLI:EU:T:2016:453, paras 75-81.

(17) New product development process and time-to-market in the generic pharmaceutical industry, Prašnikar and Škerlj Industrial Marketing Management 01/2006; 35(6):690-702.

(18) Lundbeck v Commission, T-472/13, ECLI:EU:T:2016:449, para. 366

(19) Ibid, para 366.

(20) Ibid, para 368.

(21) Ibid, para 388.

(22) Ibid, paras 350 and 411.

(23) Ibid, paras 379-384.

(24) Ibid, para 385.

(25) Ibid, para 509.

(26) Ibid, paras 511-512.

(27) Ibid, para 401.

(28) Ibid, para 494.

(29) Ibid, para 495.

(30) Sun Pharmaceutical Industries and Ranbaxy (UK) v Commission, T-460/13, ECLI:EU:T:2016:453, paras 234-236.

(31) Lundbeck v Commission, T-472/13, ECLI:EU:T:2016:449, paras 460 and 475.

(32) Ibid, para 351.

(33) Ibid, para 497.

(34) Ibid, para 500.

(35) Ibid, para 399.

(36) Ibid, para 414.

(37) Ibid, paras 713-719.

(38) Ibid, paras 350-351.

(39) Ibid, paras 354 and 500 and Sun Pharmaceutical Industries and Ranbaxy (UK) v Commission, T-460/13, ECLI:EU:T:2016:453, paras 234-242.

(40) Lundbeck v Commission, T-472/13, ECLI:EU:T:2016:449, paras 404, 515.

(41) Ibid, para 384.

(42) Ibid, para 369.