Court of Appeal for Ontario

Date: 2025-03-12
Docket: COA-23-CV-0467
Coram: Hourigan, Madsen and Pomerance JJ.A.

Between:
Apotex Inc.
Plaintiff (Appellant)

and

Eli Lilly and Company, Eli Lilly S.A., Lilly Del Caribe, Inc., Lilly S.A., Eli Lilly Export S.A., and Eli Lilly Canada Inc.
Defendants (Respondents)

Appearances:
Harry Radomski, Jerry Topolski and Nando De Luca, for the appellant
Marc Richard, Alex Gloor and Rebecca Johnston, for the respondents

Heard: 2024-10-22

On appeal from the judgment of Justice Markus Koehnen of the Superior Court of Justice, dated March 30, 2023, with reasons reported at 2023 ONSC 1968.

Pomerance J.A.:

Introduction

[1] The Patented Medicines (Notice of Compliance) Regulations, SOR/99-133, enacted under s. 55.2 of the Patent Act, RSC 1985, c P-4, set out the process to be followed when a generic drug company tries to market a “copycat” product before the patent on the original drug expires. The regulations aim to protect patent owners’ rights, while simultaneously facilitating the timely entry of lower-priced generics into the market. This case turns on the interpretation and application of s. 8 of the Regulations, which allows generic companies to claim damages from the patent owner for delayed entry into the market.

[2] The appellant Apotex Inc. (“Apotex”) sought to market a generic drug based on a patent held by Eli Lilly and Company (“Lilly”). Apotex brought a s. 8 claim for damages against Lilly to compensate for delay in bringing its product to market. The trial judge dismissed the claim. He held that as a matter of statutory interpretation, s. 8 did not apply. And he concluded that even if it did, Apotex failed to prove that Lilly was liable for damages. Apotex appeals both aspects of the trial judge’s ruling.

[3] I would dismiss the appeal. It is unnecessary to address the first issue, involving the statutory interpretation of s. 8. Whether or not the trial judge correctly interpreted the provision, he did not err in his analysis of what would happen if it did apply. It was open to him to find that a s. 8 analysis would lead to no damages. This renders the first issue academic. In either instance – s. 8 applies or does not apply – the outcome is the same.

Overview of the Statutory Scheme

[4] For present purposes, a brief summary of the statutory regime will suffice. I propose to focus only on the features directly at issue in this case.

[5] In Canada, patent owners hold a 20-year monopoly. But the Patent Act allows for an “early work” exception to patent infringement, reflected in s. 55.2 of the Act. Section 55.2 provides as follows:

55.2 (1) It is not an infringement of a patent for any person to make, construct, use or sell the patented invention solely for uses reasonably related to the development and submission of information required under any law of Canada, a province or a country other than Canada that regulates the manufacture, construction, use or sale of any product.

[6] This provision allows a generic manufacturer to “piggyback” on the work done by a patent-owning innovator in developing, testing, and marketing a drug. It can take more than two years to complete the work necessary to obtain approvals. By allowing “early work” by generic companies, the Act allows them to prepare and stockpile product so that no time is lost once the patent expires.

[7] If a generic manufacturer wants to market a “copycat” drug – a pharmaceutical equivalent of a patented drug – it has a choice. It can wait until the patent expires, or it can try to enter the market while the patent is still in place. In the latter scenario, the generic company must challenge the patent by filing a “notice of allegations” (a “NOA”) against the patent owner under s. 5 of the Regulations. Once served with a NOA, the patent owner has a choice. It can allow the generic to enter the market, reserving the right to sue for patent infringement, or it can bring an application to prohibit the marketing of the generic (a “prohibition application”) under s. 6 of the Regulations.

[8] When the patent owner launches a prohibition application, the Minister cannot issue a notice of compliance (a “NOC”) for the generic drug for 24 months, unless the prohibition proceeding resolves before then. And a NOC is required before a drug can enter the market. Therefore, a prohibition application will delay a generic’s entry into the market. If the prohibition application is unsuccessful, the generic company may have a claim in damages against the patent owner for the loss arising from the delay.

[9] Section 8 of the Regulations makes a patent holder liable for damages where its prohibition application is “withdrawn or discontinued…” or “is dismissed by the court hearing the application”, or the order is “reversed on appeal”. The full text is as follows: [1]

8. (1) If an application made under subsection 6(1) is withdrawn or discontinued by the first person or is dismissed by the court hearing the application or if an order preventing the Minister from issuing a notice of compliance, made pursuant to that subsection, is reversed on appeal, the first person is liable to the second person for any loss suffered during the period
   (a) beginning on the date, as certified by the Minister, on which a notice of compliance would have been issued in the absence of these Regulations, unless the court concludes that
   (i) the certified date was, by the operation of An Act to amend the Patent Act and the Food and Drugs Act (The Jean Chrétien Pledge to Africa), chapter 23 of the Statutes of Canada, 2004, earlier than it would otherwise have been and therefore a date later than the certified date is more appropriate, or
   (ii) a date other than the certified date is more appropriate; and
   (b) ending on the date of the withdrawal, the discontinuance, the dismissal or the reversal.
   (2) A second person may, by action against a first person, apply to the court for an order requiring the first person to compensate the second person for the loss referred to in subsection (1).
   (3) The court may make an order under this section without regard to whether the first person has commenced an action for the infringement of a patent that is the subject matter of the application.
   (4) If a court orders a first person to compensate a second person under subsection (1), the court may, in respect of any loss referred to in that subsection, make any order for relief by way of damages that the circumstances require.
   (5) In assessing the amount of compensation the court shall take into account all matters that it considers relevant to the assessment of the amount, including any conduct of the first or second person which contributed to delay the disposition of the application under subsection 6(1).
   (6) The Minister is not liable for damages under this section.

[10] If the patent holder is found to be liable to the generic under s. 8, the quantum of damages depends on the profits the generic would have made in a “hypothetical world” in which the patent holder did not initiate the prohibition proceeding: Merck Frosst Canada & Co. v. Apotex Inc., 2011 FCA 329, para 75; Apotex Inc. v. Sanofi-Aventis, 2014 FCA 68, paras 160-62, 170-71, aff’d 2015 SCC 20.

[11] The period relevant to this inquiry is called the “liability period”. It begins “on the date, as certified by the Minister, on which a notice of compliance would have been issued in the absence of these Regulations”, and ends on the date of the dismissal of the prohibition proceeding: Regulations, ss. 8(1)(a), (b).

[12] The “hypothetical world” analysis turns in part on (a) whether the generic could have entered the market during the liability period, and if so, when, and (b) whether the generic would have entered the market during the liability period, and if so, when. These are factually-driven questions that depend on contingent circumstances. See, e.g., Merck Frosst Canada & Co. v. Apotex Inc., 2011 FCA 329, paras 78-79 (the generic’s ability to acquire a supply of the relevant materials); Apotex Inc. v. Sanofi-Aventis, 2012 FC 553, para 11 (the share of the market the generic would have obtained), rev’d on other grounds, Ramipril FCA.

[13] If the generic could and would have entered the market during the liability period, it is entitled to damages for the profits lost as a result of the patent holder preventing it from doing so: Apotex Inc. v. Sanofi-Aventis, 2012 FC 553, para 237.

Factual Background

[14] Apotex manufactures a drug known as “Apo-Atomoxetine”, a generic copy of the drug “Strattera”, which Lilly developed, tested, and patented. Apotex filed its Abbreviated New Drug Submission (“ANDS”) for Apo-Atomoxetine on February 29, 2008. At that time, Apotex indicated that it would wait until Lilly’s Strattera patent (the “735 Patent”) [2] expired before seeking a NOC. The 735 Patent was set to expire in 2016.

[15] Just under three months later, Teva, [3] another generic manufacturer, sought to impeach the 735 Patent. But it did so by way of an action under the Patent Act, rather than through the Regulations.

[16] Shortly after that, Apotex served Lilly with an NOA regarding the 735 Patent. In response, on October 10, 2008, Lilly started a prohibition proceeding, triggering an automatic stay on Apotex receiving a NOC.

[17] Teva’s action against Lilly and Lilly’s prohibition proceeding against Apotex proceeded separately, but were both heard before Barnes J. in the Federal Court. On September 14, 2010, Barnes J. declared the 735 Patent invalid in Teva’s action. Its invalidation permitted Apotex to obtain a NOC for Apo-Atomoxetine on September 21, 2010.

[18] On October 29, 2010, Barnes J. released his decision in the prohibition proceeding. He found against Apotex on each of its allegations. He would have allowed Lilly’s application but for his finding in the Teva action that the 735 Patent was invalid. Since Apotex had already received its NOC, he dismissed the prohibition proceeding as moot. He also ordered Apotex to pay Lilly’s costs of $215,000.

The Section 8 Claim

[19] Apotex then brought a claim for s. 8 damages against Lilly, arguing that Lilly’s now-dismissed prohibition application had delayed the marketing of Apo-Atomoxetine.

[20] The trial judge dismissed Apotex’s claim for damages. He held that s. 8 only entitles a generic to damages when the prohibition application is dismissed on the merits by the judge hearing the application. A dismissal for mootness did not qualify, so Apotex could not make a claim under s. 8. But for Teva’s (entirely separate) action, Lilly’s prohibition application would have succeeded. As the trial judge put it, to allow Apotex to claim damages for a victory it did not secure would allow it to act as a “free rider”. The prohibition application was dismissed not because of arguments Apotex made, but because of Teva’s success in a separate action. The trial judge relied on the proposition that “section 8 was not intended to provide redress where the innovator prevailed in the prohibition proceeding”: Apotex Inc. v. Syntex Pharmaceuticals International Inc., 2010 FCA 155, para 36.

[21] This sufficed to dispose of the claim, but the trial judge went on to consider damages in the event that s. 8 did apply. This calculation turns on what would have happened in a hypothetical world in which there was no prohibition application. Would Apotex have entered the market earlier in the hypothetical world than it did in the real world? If so, would it have earned more profits?

[22] The trial judge found that Apotex could have launched its product earlier, but that it would not have done so, given the risk that Lilly would launch a patent infringement claim. Apotex had revealed itself to be risk-averse in the real world. The trial judge assumed the same level of risk aversion in the hypothetical world. Given the significant risk of liability and the potential competition presented by Teva, the trial judge found that Apotex would have refrained from marketing Apo-Atomoxetine any earlier than it did, even absent Lilly’s prohibition application. It followed that Lilly was not liable for damages for delay.

Analysis

The Hypothetical World

[23] Apotex, as the plaintiff, bore the onus to establish that the facts of the hypothetical world would support its claim for damages: Pfizer Canada Inc. v. Teva Canada Ltd., 2016 FCA 161, paras 53-56. To succeed, Apotex had to establish that it would have come to market in the hypothetical world sooner than it did in the real world. The conduct of the parties in the real world is useful evidence for constructing the hypothetical one. The court should assume that their behaviour in the real world reflects what they would have done in the hypothetical world, absent evidence to the contrary: Teva Canada Limited v. Sanofi-Aventis Canada Inc., 2014 FCA 67, para 145.

“Could Have” and “Would Have”

[24] The trial judge found that Apotex could have launched its product as of October 10, 2008. Gordon Fahner, then Apotex’s Vice President of Supply Chain Management, testified that Apotex had the ability to order sufficient raw materials and manufacture sufficient quantities to allow it to launch when it received its hypothetical world-NOC on October 10, 2008.

[25] But the trial judge went on to find that while Apotex could have entered the market at that time, it likely would not have done so. He noted that in the hypothetical world, Lilly would have continued to hold the 735 Patent until it was invalidated on September 21, 2010. This is because the only difference between the real and hypothetical worlds is that no prohibition proceeding exists in the latter: Ramipril FCA, at paras. 170-71; Norfloxacin, at para. 75. And prohibition proceeding or not, Lilly still held the 735 Patent.

[26] Because Lilly still held the 735 Patent in the hypothetical world, Apotex would have faced a risk of liability for infringing the patent if it had gone to market. The trial judge found that Apotex would not have taken that risk. “Lilly was an aggressive litigant that fought energetically to maintain its patent rights”, he explained, and even in the real world, Apotex still hesitated to go to market after the 735 Patent was invalidated, because it was concerned that the patent would be reinstated on appeal.

[27] The trial judge observed that for Apotex, the risk of a patent infringement action far exceeded that associated with a prohibition application. If Lilly succeeded in establishing a patent infringement, Apotex would be liable to compensate it for sales of Strattera that were lost to the generic drug. That could represent a sizeable award.

[28] The trial judge found that Apotex would not have exposed itself to liability in this fashion. He relied on the following additional facts:

• Apotex was content to wait for the 735 Patent to expire, as indicated in its ANDS, until it learned of Teva’s action.
• Mr. Fahner testified that Apotex would have been highly motivated to enter the market “absent the legal impediments” from the patent. In the hypothetical world, those legal impediments still existed in the form of a patent infringement action.
• Apotex did not come to market for many months after receiving its NOC, because it was concerned that the Teva judgment might be overturned on appeal.

[29] Apotex argues that the trial judge’s findings are untenable because had it been alone in the market, its earnings as the sole generic would have mitigated the risk of a patent infringement action. Lilly responds by pointing to the trial judge’s finding that Apotex would have been required to pay some of its revenues in rebates to purchasers of Apo-Atomoxetine during the liability period. The trial judge made this finding based on evidence of rebates paid by Apotex in the real world. Apotex argued that the rebates were lower, but the trial judge found that its evidence on this point was “too limited and frail” to satisfy its onus. Accepting the rebates as he had found them, the trial judge concluded that Apotex would not have risked a damages award in a patent infringement action.

[30] The risk of liability increases when one considers that Barnes J. rejected Apotex’s arguments against the patent on the prohibition application. These arguments would not be any more persuasive if Apotex was defending a patent infringement action. The trial judge assumed that Apotex, a sophisticated litigant, was aware of the frailties of its arguments, observing that it “might be prepared to risk failure in a prohibition proceeding where the downside is relatively limited”, but would likely be “less willing to expose [its] broader profit and asset base to risk in a patent infringement proceeding”.

[31] Teva successfully impeached the patent, but it did so on the basis of inutility. And while Apotex initially raised inutility in the NOA it served on Lilly, it chose not to lead any evidence on that point and made deliberate, tactical decisions that resulted in its argument being dismissed: Eli Lilly Canada Inc. v. Apotex Inc., 2010 FC 1065, paras 88, 91, 95. In Barnes J.’s words:

[95] I am accordingly bound on this record to reject Apotex’s allegation of inutility because it has failed to satisfy its initial evidentiary burden. Because Lilly effectively had no case to answer, it is unnecessary for me to determine whether the evidence bearing on the MGH Study, at least in the manner tendered, was inadmissible hearsay. Even if that evidence was not admissible, I am still left with the initial problem of whether the Apotex evidence was sufficient to put the allegation of inutility into play and, as noted above, I find that it was not. For this issue, the statutory presumption of validity prevails.

[32] Apotex’s failure to demonstrate inutility before Barnes J. is an important building block in the architecture of the hypothetical world. It is true that, in a patent infringement action, Apotex would have disclosure entitlements that were not available in the prohibition application. Nonetheless, the outcome in the hypothetical world must be tethered to events in the real world, including the arguments that were advanced. Apotex chose the allegations it made in its NOA, and its choices led Barnes J. to conclude that it was not entitled to a NOC. The hypothetical world is not a gateway for parties to raise new and better arguments, in search of a different outcome.

Teva as Generic Competitor

[33] During the liability period, Teva was trying to market its own generic copy of Strattera. Both Teva and Apotex received their NOCs in the fall of 2010, but both refrained from entering the market. Instead, they spent six months working on a possible joint venture. This would have facilitated the sale of one generic product by both companies, allowing them to sell at a higher price with lower discounts. While the appeal of the Teva decision was ongoing, Apotex and Teva pursued an advisory opinion from the Competition Bureau, seeking assurances that the possible joint venture would comply with applicable competition law. The Competition Bureau declined to provide that assurance. That marked the end of the joint venture discussions between the two companies.

[34] Apotex argues that the trial judge erred in finding that a joint venture would have been pursued in the hypothetical world. It points out that Teva had not served Lilly with a NOA under the Regulations, so it would not have been in a position to enter the market before Apotex. This, Apotex argues, is significant, because the economics of the risk analysis change in its favour if it is the sole generic in the market. It led evidence that the company would have accepted the liability risks to maximize the opportunity to be the sole generic in the market.

[35] I am not persuaded that the trial judge erred in his construction of the hypothetical market. To start with, his analysis did not hinge on the proposition that Teva would be first to market its product. He acknowledged that Teva might have entered the market before or after Apotex, as it did in the real world. And there was evidence to support the conclusion that, if Apotex was first in the market, Teva would have competed. Teva filed its ANDS before Apotex. Once a NOC was issued to Apotex, Teva could have converted its statement of claim into a NOA and sought its own NOC under the Regulations.

[36] The trial judge considered the whole of the evidence in assessing whether Teva would have, at some point, entered the market as a generic competitor. Apotex has failed to demonstrate that his findings disclose palpable and overriding error. In any event, Teva’s participation in the market was a more peripheral than prominent factor in the trial judge’s reasons. His conclusions were animated by the risk of liability that Apotex would face in a patent infringement action brought by Lilly, and the risk-averse stance that Apotex demonstrated in the real world.

[37] In short, there is much to support the trial judge’s conclusion on when Apotex would have entered the market. This was a multifaceted analysis calling for the construction and assessment of complicated hypothetical contingencies. By its very nature, such prediction is an inexact science. Assessing what a company would have done in an alternate factual universe calls for some degree of supposition, albeit supposition rooted in real-world evidence, knowledge of the pharmaceutical industry, and an understanding of the regulatory scheme. That is precisely the sort of determination that calls for appellate deference. Apotex has failed to demonstrate that this determination of mixed fact and law gave rise to any palpable and overriding error.

Conclusion and Disposition

[38] For these reasons, I would decline to interfere with the trial judge’s conclusion that even if Apotex could claim a theoretical entitlement to s. 8 damages, Apotex has failed to demonstrate any actual entitlement to damages from Lilly.

[39] As I noted earlier, this conclusion makes it unnecessary to consider whether the trial judge erred in holding that the language of s. 8 disqualifies Apotex’s reliance on the provision. Because the outcome is the same regardless of whether s. 8 applies, the issue of statutory interpretation is of purely academic import. It is neither necessary nor desirable for this court to undertake an academic analysis. The interpretation issue is properly reserved for a case in which it has a practical impact on the dispute.

[40] Therefore, I would dismiss the appeal. I would award Lilly its costs of the appeal in the agreed-upon amount of $27,500, all-inclusive. Given Lilly’s success on the appeal, it is unnecessary to address the parties’ submissions on the costs award below.

Released: March 12, 2025

“C.W.H.”
“R. Pomerance J.A.”
“I agree. C.W. Hourigan J.A.”
“I agree. L. Madsen J.A.”

[1] Section 8 has since been amended. The version reproduced here was in force at the relevant time.
[2] The Strattera patent is Canadian Patent No. 2,209,735.
[3] Teva was then called Novopharm.