ART. 101(1) TFEU: A Bitter Pill for Hoffmann-La Roche

By Martin Herz and Justin Lindeboom

On 23 January 2018, the Court (Grand Chamber) gave a preliminary ruling on five questions asked by the Italian Consiglio di Stato, with regard to an anti-competitive arrangement in Italy, between two distributors of medical products, Hoffmann-La Roche (“HLR”) and the Novartis Group (“Novartis”). HLR, whose name must be familiar in the eyes of competition lawyers, was yet again plagued with an unfavourable ruling. In it, the Court further clarified its case law on one of the foundations of competition law: the market definition. Generally, in antitrust cases, before being able to measure most effects of certain conduct, markets will be defined. Since the United Brands ‘bananas’ case from 1978, this has been standing case law.

Market definitions consist of a two-pronged approach, namely, a product market, and a geographical market definition. Both depend on the characteristics and intended use of the product, and of course, whether there is demand and supply of the products in question to begin with. Subsequently, products that consumers (or producers) find interchangeable to a high extent, will fall within the same market.

An example might clarify this. An element that watches, timepieces, wall clocks, and alarm clocks have in common is that they are all time-keeping instruments. However, they are unlikely to fall within the same product market. For different reasons, people want watches, timepieces aut cetera. Production costs (and thus, prices) might differ highly as well. For instance, the clock of the Big Ben is not comparable to a wrist watch. Hence, different product markets for time-keeping instruments will exist.

Now, what if a product suddenly, and systematically, is being used for a cause, for which it had not been intended, or regarding which its legal authorisation is uncertain? Does this affect the market definition? This was one of the key issues the Court faced in the current Hoffmann-La Roche ruling. Central to the case were two medicinal products, Avastin & Lucentis: HLR distributed Avastin outside the US, and Novartis had been licensed to do the same for Lucentis. In 2005, Avastin received a market authorisation (MA) for the treatment against tumorous diseases in the EU. Two years later, Lucentis received the approval for the treatment of eye ailments. However, since 2005, doctors had been prescribing Avastin for patients with eye diseases. This “off-label” practice continued world-wide, even after the more expensive Lucentis came to be distributed (according to the AGCM’s decision, Lucentis cost EUR 902,00 per injection in November 2012, and Avastin EUR 81,64).

HLR & Novartis had jointly launched an information campaign which focused on the alleged risks of using Avastin in the eyes, whereas no conclusive scientific proof existed on the actual risks. Because Avastin had become a competitor of Lucentis, a shift in demand occurred from the former to the latter. Therefore, the Italian competition and market protection authority (“AGCM”) fined HLR and Novartis separately for engaging in an anti-competitive market sharing arrangement (“the arrangement”), artificially differentiating between the products, whereas the products were held equivalents for the treatment of eye diseases. The arrangement was held to manipulate the perception of risks associated with the use of Avastin. This “alarmist” interpretation would induce healthcare professionals to choose Lucentis over Avastin.

In this commentary, we will focus on two key aspects of the judgment. The first aspect pertains to the definition of the relevant market. More specifically, this case centred on the question of whether a medicinal product which are used “off-label” for eye ailment treatments (in this case Avastin) is part of the same relevant market as another product which has a MA for eye ailment treatment (in this case Lucentis). Secondly, as to the substantive assessment of the dispute, the main issue in this case was the question of whether the applicable market sharing arrangement constituted a restriction of competition by object.

Can medicinal products used “off-label” be incorporated in the relevant market definition for eye ailment medication?

The questions with respect to market definition boiled down to whether Avastin, an “off-label” used medicinal product, could be included in the market for medicines for the treatment of eye ailments. In his Opinion, AG Saugmandsgaard Øe stated an MA could definitely influence the substitutability, as an MA will likely guide medical practitioners in prescribing a product treating ailments for which that product was authorised (pt. 76). However, the AG only found the uncertainty regarding the lawfulness of off-label use of the medication relevant if it influenced the substitutability for on-label used products. Such uncertainty does not, by default, constitute a separate product market (pts. 86-87).

The Court largely coincided with the AG’s Opinion, although it veered somewhat off-course in its relatively painstaking reasoning. After introducing the relevant market criteria, it stated that Avastin would, “in principle”, exist on a separate market, if it were to have been illegally produced, i.e. manufactured or sold: suppliers would face legal, economic and technical risks, and demanding healthcare professionals and patients would face risks to public health (para. 52).

This was not an answer to the question posed by the Consiglio di Stato. Illegal production is different from the uncertainty regarding the legitimacy of the use of a product. Nevertheless, the Court assessed whether,  and concluded that Avastin had been produced in accordance with the relevant legal framework, i.e., Dir. 2001/83 (paras. 53-59).

The ruling is clear: Avastin was legally produced. However, from a general direct effect perspective, the question arises what this legality in light of the Directive has to do with the substitutability of certain medication as prescribed by medical practitioners? Directives do not create obligations for individuals, and medical practitioners are individuals from an EU law perspective. As Art. 5(1) Dir. 2001/83 states that Member States may allow off-label use, substitutability depends not on the Directive, but on the way in which Member States transposed this requirement. If, for instance, no enforcement had been undertaken by a Member State, medical practitioners could have continued prescribing Avastin for eye disease treatments, and substitute Lucentis. The Directive and its prohibitions would not apply to them, but only to the Member States.

If medical practitioners were to risk enforcement action of the Member State, and the efficacy of Avastin against eye ailments were to remain the same, meaning the risk for public health was no different, we are not convinced that such risk would lead to different substitutability of Avantis and Lucentis. Nonetheless, if practitioners had attached grave importance to this risk, off-label prescriptions, and thus, the litigious issue would not have occurred, and separate markets for Avastin and Lucentis would be defined. This would be the same if Member States had ensured appropriate and full enforcement of a ban on off-label use, from Avastin’s introduction in 2005 onwards.

These hypothetical options show that it is wholly irrelevant whether Avastin itself was produced illegally in light of this Directive or not.

Fortunately, the Court applied the criterion of intended use of the medication as key in defining substitutability for eye treatment medication. Because Avastin was used for the same therapeutic indication as Lucentis, and considered appropriate by doctors to be prescribed, a “specific relationship of substitutability” between those products existed and the relevant market was held capable of comprising those medicines. Moreover, demand was quantifiable, because the use of Avastin for the treatment of eye diseases was prescription-based only (paras. 65-66).

At first glance, this market definition seems like a textbook example. Both Avastin and Lucentis were prescribed and used to treat the same ailments; hence, from a patient’s perspective, they could be substitutable. Also, the remarks regarding the specific product characteristics are sensible. For example, a fake watch might not look all too different from real, but might not function the same, let alone be purchased under the same conditions as a real one. The same would go for illegally-produced medicines. Hence, running a business under those circumstances is risky.

In this case, the products were even governed by specific legislation, viz. Dir. 2001/83. After verification, the CJEU logically came to the conclusion that this legal regime had not influenced the substitutability of Avastin and Lucentis on this market, as both had been lawfully-marketed goods.

However, substitutability, as the Court rightly pointed out, consists of the characteristics, intended use, but also, in most cases predominantly, prices. In general, each price difference between products exceeding a small but significant non-transitory increase in price could justify separate market definitions. As regards illegally-produced goods, prices tend to differ from their legal equivalents. It is odd that the Court did not pay any attention to this.

Applying the watch example by analogy to this case, if, in the market for devices time-keeping instruments worn around the wrist (i.e. watches), consumers consider that a €10.000,- watch is just as good at giving the time as a €50 legally-produced counterfeit, such goods could be held substitutable on the basis of their intended use, and fall within the same market (ceteris paribus). Nonetheless, we believe that it is reasonable to assume, that a watch €10.000,- appeals to a different purchasing audience than the cheap rip-off.

In this case, the preliminary questions specifically referred to the competence of an NCA to include “off-label” use of Avastin in the market for eye treatment medication. The question of whether the price difference between Avastin and Lucentis could affect the market definition was not within the scope of the reference. Also, price competition might not be as important in markets where medicines are often reimbursed. Hence, in our opinion, the Court was justified to turn this into a textbook market definition.

Is the dissemination of false information on off-label use a restriction of competition by object?

Article 101 TFEU prohibits certain types of coordinated behaviour insofar they either have the object, or the effect of restricting competition between undertakings. The recent case-law on by object restrictions has moved between a broader interpretation of this category to include more agreements than the traditional list of “hardcore restrictions”, and the necessity to interpret the object category restrictively. Examples of a broader approach include Pierre Fabre (on an outright ban of online sales), ING Pensii (on a division of a very small group of insurance customers while these customers could no longer be competed for) and Dole Bananas (on exchange of pricing information between employees who did not have the authority to set prices). By contrast, in Cartes Bancaires and Coty the Court followed AG Wahl in a more restrictive approach.

In HLR 2017, the Court reiterated its formula from Cartes Bancaires, affirming that the concept of object restrictions only applies to coordination between undertakings “which reveal a degree of harm to competition that is sufficient for it to be held that there is no need to examine their effects” (para. 78). In order to assess whether an arrangement reveals a “sufficient degree of harm to competition”, one should look at the content of its provisions, its objectives and its economic and legal context.

The Court started its analysis of the agreement with reference to the relevant EU rules on pharmaceutical products as part of the legal context of the agreement between Roche and Novartis. In particular, pursuant to Regulation 726/2004, the holder of a MA for a medicinal product is obligated to inform the EMA and the Commission about any new information on the benefits and risks of said product, and is responsible for the accuracy of any such information provided. When it comes to information provided to healthcare professionals and the public, the holder has an obligation to provide information which is presented objectively and not misleading, as per Article 106a of Directive 2001/83 (paras. 80 – 88).

In these factual circumstances, the Court essentially made two arguments as to their legal qualification. First, it noted that the EU requirements on pharmaceutical products are solely directed towards the holder of the MA. The fact that Roche and Novartis agreed to disseminate information on the safety risks of a product which was only marketed by one of them – Roche – “might constitute evidence that the dissemination of information pursues objectives unrelated to pharmacovigilance” (para. 91).

Subsequently, the Court concluded that if the information campaign towards both the EMA and the general public must be regarded as “misleading” (which is for the national court to decide), it would be likely this misleading information encouraged doctors to prescribe Lucentis and to refrain from prescribing Avastin, thereby manipulating demand towards Lucentis. For this reason, the arrangement’s object revealed a sufficient degree of harm to competition. From the relevant EU rules of pharmaceuticals (legal context) and the significant price difference between Avastin and Lucentis (economic context), there was hardly any reason why Roche and Novartis would close this agreement other than to manipulate competition between their products, benefitting from the scientific uncertainty on the matter.

Hence, this result fits perfectly into the mentioned recent case-law on object restrictions. On the one hand, the Court emphasizes the need to interpret the object category strictly. On the other hand, it increasingly scrutinizes agreements in their legal and economic context, so as to ground its qualification of the behaviour in question in the factual circumstances of each case. This implies that analyzing whether behaviour is a restriction of competition by object can be quite laborious and is not necessarily confined to a ‘quick look’. This trend matches the Court’s increasing “effects-based” approach in abuse of dominance cases, exemplified for example by its judgments in Post Danmark I and Intel.

A few peculiarities in the judgment can be observed. The Court leaned heavily on the content of EU rules on pharmaceuticals in order to establish a restriction by object, while it spent only one sentence on the likely economic effects of the agreement. This means that the “legal context” of an agreement is not merely confined to the broader legal relations between the undertakings, but may also include other rules that apply to the undertakings.

Secondly, the Court did not explicitly mention the price difference between Avastin and Lucentis, merely referring to the likely effect on the demand for off-label use of Avastin (paras. 93 – 94). What if the price difference between both products had been much smaller or even equal? The judgment suggests that the misleading nature of the information disseminated was in itself a sufficient condition for finding a restriction by object. In this regard, the Court paid no attention to the capability of the arrangement to restrict competition. This is somewhat remarkable, because healthcare professionals are the main group of actors who were purportedly misled. The Court’s theory of harm in this case was premised on the assumption that healthcare professionals – arguably experts in the field of pharmaceuticals and their side-effects – can be misled by false information. In turn, this implies that even healthcare professionals are not fully rational in their preferences.

The outcome of this judgment is likely to draw much attention from pharmaceutical companies. Even though the outcome of the case comes as no surprise from a competition law perspective, pharmaceutical companies might need to pay additional detail to the manner in which they market their products, particularly in the context of licensing agreements (which are abundant in the industry). Any form of coordination related to licensing agreements as regards the marketing of products “on-label” and “off-label” are in the danger zone when they might have an impact on the structure of demand and supply. Pharmaceutical companies will need to ensure that any concerted marketing claims are backed by conclusive scientific evidence.