Thursday, August 27, 2020

Blogging Break

 I'll be taking a short break from blogging for some end of summer vacation and to get ready for classes. 

Thursday, August 20, 2020

Data Protection Provisions Triggered by Indirect Comparison with Innovative Drug

Natco Pharma (Canada) Inc. v. Canada (Health) 2020 FC 788 McHaffie J

This application for judicial review addressed whether the data protection provisions of the Food and Drug Regulations are triggered when an ANDS is based on a comparison to a drug product that was in turn approved based on a comparison with an “innovative drug,” even though the direct comparator is not itself an innovative drug. McHaffie J held Health Canada’s decision refusing Natco’s ANDS to be reasonable and indeed “inevitable” [4], and he consequently dismissed Natco’s application for judicial review.

The decision is interesting for three reasons. First is the substantive holding, that the data protection provisions are triggered on the basis of “indirect” comparison, when the generic product is compared to a drug product that was in turn approved on the basis of a comparison to an innovative drug. Secondly, this decision follows close on the heels of ViiV Healthcare 2020 FC 756 (here) and Shingrix 2020 FC 397 (here), which held the Minister of Health to have adopted an unreasonable approach to statutory interpretation of the CSP Regulations. The contrast in Health Canada’s approach to these provisions is noteworthy. Finally, the decision raises a narrow point of statutory interpretation, holding that the same phrase may have a different meanings in different but broadly related regulations, in light of purposive and contextual considerations.

The key operational provisions of the data protection regulations provide for a six year “no file” period and an eight year market exclusivity period if a manufacturer seeks an NOC on the basis of a “direct or indirect comparison” between the new drug and an innovative drug: [10], C.08.004.1(3)(a),(b). Natco submitted an ANDS for a generic version of Gilead’s DESCOVY, a HIV/AIDS drug that contains a combination of tenofovir alafenamide hemifumarate (TAF) and emtricitabine. The ANDS accordingly identified DESCOVY as the Canadian reference product [26]. DESCOVY is not an “innovative drug” under the data protection regulations. However, TAF is also contained in Gilead’s GENVOYA, along with emtricitabine and two other medicinal ingredients [2], and GENVOYA is an “innovative drug” [14]. DESCOVY was approved after GENVOYA, which is why it is not an “innovative drug.” Crucially, in McHaffie J’s view, the data to support DESCOVY was based on a comparison with GENVOYA [68], [80]. Health Canada refused Nacto’s ANDS on the basis that it “makes comparisons to DESCOVY, which benefits from the data protection term for GENVOYA, an innovative drug,” and as such the ANDS cannot be accepted until the expiry of the “no file” period for GENVOYA, in November 2021 [38].

McHaffie J ultimately held that by seeking an ANDS using DESCOVY as the reference product, Natco was making an indirect comparison with GENVOYA:


[108] That is, the “direct or indirect comparison” to an innovative drug that forms the trigger for data protection provisions may include a manufacturer’s comparison to a drug product that in turn was compared to the innovator product for approval

The parties agreed that the standard of review under Vavilov 2019 SCC 65 was reasonableness, so question at issue was not the interpretation of the data protection provisions as such, but whether the Health Canada’s decision was reasonable [8]. Health Canada’s analysis focused primarily on the intent of the regulations and the obligations under the trade agreements [43]-[50]. These are important considerations, and Health Canada’s analysis was reasonable “as far as it went” [50], but the decision jumped directly from there to the conclusion that drugs containing the same medicinal ingredient must benefit from the same period of data protection, without full consideration of the text of the provisions itself [51], [56]. This led Health Canada to the questionable conclusion that data protection of an innovative drug “necessarily extend[s] to these additional products also containing the new chemical entity during the data protection term for the original innovative drug” [56]. That is, the main thrust of Health Canada’s decision was that the data provision provisions protect the new chemical entity itself – “The obligations to protect the new chemical entity exist for the entire duration of the data protection term” [33] – and therefore a combination drug containing a new chemical entity that was the basis for an innovative drug designation will also benefit from any term of the data protection for the innovative drug [35]. However, as McHaffie J noted, it is the data, not the new chemical entity itself that is protected: for example, the data protection regulations would not be triggered if Natco had filed an NDS based on independent clinical trials [57], or if DESCOVY had been approved based on independently filed studies rather than on the basis of a comparison with GENVOYA [59].

Consequently, if Health Canada had arrived at its conclusion solely on the basis of the fact that DESCOVY contained TAF, the decision would have been unreasonable [60]. However, as “further support,” for its conclusion, Health Canada noted that the approval of DESCOVY had relied on the data for GENVOYA [69]. This was not a secondary consideration, as suggested by Health Canada’s decision, but a crucial point: it was unchallenged that the data to support DESCOVY was based on comparative bioavailability studies for DESCOVY as compared to GENVOYA [68], [80]. Because of this, Natco’s comparison of its product with DESCOVY constituted an indirect comparison with GENVOYA [72]. And while Health Canada’s analysis was not as clear as it might have been, McHaffie J concluded that, read contextually and with due allowance for the administrative context [75], the decision is “fairly read as Health Canada making the determination that Natco’s ANDS indirectly compared its drug to GENVOYA” [73], without impermissible ex post addition of arguments on judicial review that are not contained in its decision [78]. McHaffie also pointed out that in some places Health Canada did recognize that the obligations under the treaties was to protect “undisclosed test or other data,” and so he concluded that Health Canada did not misunderstand the intent of the regulations [67], though its conclusions were poorly expressed.

The ultimate issue was the interpretation of subsection C.08.004.1(3). While Health Canada’s decision did not actually address that question directly, it did implicitly base its decision on a reasonable interpretation of the provision [108]. Indeed, not only was the interpretation reasonable, this was the type of case, adverted to in Vavilov 2019 SCC 65 [124], in which “the ‘interplay of text, context and purpose leaves room for a single reasonable interpretation’” [108]. Consequently:


[108] [T]he “direct or indirect comparison” to an innovative drug that forms the trigger for data protection provisions may include a manufacturer’s comparison to a drug product that in turn was compared to the innovator product for approval.

Given that it was undisputed that this was true on the facts [108], “the outcome that Natco’s ANDS could not be accepted for filing was inevitable” [108].

What, if anything, can we make of the difference between Health Canada’s approach in this case, and the unreasonable decisions in ViiV Healthcare 2020 FC 756 and Shingrix 2020 FC 397? The two unreasonable interpretations by the Minister related to the Certificate of Supplementary Protection Regulations and went against the patentee, while the decision at issue in this case related to the data protection provisions, and went in favour of the innovator. Also, in ViiV Healthcare the Minister erred in her interpretation by ignoring the trade agreement that the CSP Regulations were intended to implement (see here), while in this case, Health Canada’s interpretation put too much emphasis on the underlying trade agreements ([4], [51]), though the interpretation was ultimately reasonable nonetheless. Maybe there is no overarching lesson, other than that the Health Canada doesn’t always get its statutory interpretation right; but my sense, especially in light of ViiV Healthcare, is that, for some reason, the Minister is antipathetic to the CSP regulations in particular.

The third point of interest in this decision is that it illustrates that the same textual phrase may have a different meaning in different regulations, once purpose and context are taken into account, even when the regulations are related. At the same time that the data protection provisions were amended to introduce the requirement of a “direct or indirect” comparison, the PM(NOC) Regulations were also amended, such that the provisions are triggered under s 5(1) when the generic seeking an ANDS “directly or indirectly compares” its drug with an innovative drug. In the NOC Regs, the phrase is interpreted as meaning that the provisions are triggered only when generic compares its product to the equivalent innovative drug [97-98]; Natco’s “strongest argument” was that a similar interpretation should apply to the same phrase in the context of the data protection provisions [86]. While this is a good argument prima facie, McHaffie J noted that a “different purpose informs the interpretation of the language of the provisions in the two regulations” [93]. This is revealed through the structure of the provisions [94-95], and also in light of the trade agreements which they implement [96], as well as through the jurisprudential background [99]. The NOC Regs were amended to add the directly “or indirectly” language because of concerns that a generic seeking an NOC might compare its product to a previously approved generic drugs, rather than to the original product; that is, the word “indirectly” encompasses a comparison mediated by an intervening generic product that is also equivalent to the same innovative product [101]. The data protection provisions, on the other hand, were amended because the earlier trigger required the Minister to “examine” and “rely on data” contained in the application for the innovative drug. In fact, when a generic drug is approved based on an ANDS, Health Canada does not normally directly rely on the data originally submitted with the innovative product, but only verifies that the generic product is pharmacologically equivalent to the innovative product [99]. Thus, the comparison is “indirect” because it is mediated by the innovative drug itself; the RIAS accompanying the data protection provisions nonetheless made it clear that the intent was to protect the underlying data [99-100]. Given that statutory interpretation requires consideration not just of the text, but also context and purpose, it is not unprincipled to attribute different meaning to very similar text, but it is nonetheless unusual, particularly when the provisions are broadly related. 

Monday, August 10, 2020

Jurisdiction of the Federal Court to Interpret Contracts Relating to Patents: A Fresh Start

SALT Canada Inc v Baker 2020 FCA 127 Stratas JA: Near, Woods JJA rev’g 2016 FC 830 Boswell J


In a direct and vigorous decision, Stratas JA for the FCA has held that the Federal Court always has jurisdiction to determine who has title to a patent in an application under s 52 of the Act [12], [47], [49], expressly repudiating a line of cases holding that the FC does not have jurisdiction if determination of ownership of the patent is “primarily” a matter of contractual interpretation [27]-[31]. This a very welcome development that will simplify litigation in this area. This issue has been a pet peeve of mine ever since the trial decision in SALT v Baker four years ago and I am very pleased to see the issue dealt with so clearly. There is, however, one loose end, namely the effect on Innotech (1997) 74 CPR (3d) 275 (FCA) rev’g 72 CPR(3d) 522 (FCTD), holding that the Federal Court does not have jurisdiction when a contract relating to patent ownership is used as a sword, rather than a shield. Innotech is formally distinguishable, in that it did not involve an application under s 52, but the holding is extremely difficult to reconcile with Stratas JA’s reasoning in this case. In my view, Innotech is no longer good law, but some uncertainty remains, as it is an FCA decision and was not directly addressed in SALT v Baker.

As discussed here, SALT v Baker involved conflicting assignments of rights in the ‘058 patent from the inventor, Dr Markels. The first was a complex series of assignments which ultimately led to the respondent, Baker. Baker registered his ownership with the Patent Office. Some of the assignments in this series included provisions requiring the ‘058 patent to be reassigned to Markels if certain conditions were breached. Believing that the conditions had indeed been breached, in 2015 Markels assigned title to the Applicant, SALT. Markels also prepared a reassignment from Baker to Markels, which was, however, never executed by Baker. SALT sought to have its ownership registered, but the Patent Office refused on the basis that the assignment was not executed by Baker, the registered owner [3]-[9]. SALT then brought this application for a declaration under s 52 that the register be varied to list SALT as the owner of the ‘058 patent (as well as other additional and alternative relief). A key substantive question related to the interpretation of the various agreements, and whether the conditions in the first series of assignments had been breached, such that Markels was entitled to a reassignment [24].

Boswell J at first instance, relying primarily on Lawther (1995), 60 CPR(3d) 510 (FC), held that the Federal Court “lacks jurisdiction where determination of the ownership of a patent depends upon the application and interpretation of contract law principles” [20]. The question is whether the matter at hand “relates primarily to contract or to patent law: this Court will have jurisdiction over a case that primarily concerns the latter, but not the former” [21].

The FCA has now reversed. The reasoning of Stratas JA was straightforward. The basis of the FC jurisdiction is the grant by Parliament. The text of s 52 of the Patent Act is clear [8]:

[5] Section 52 of the Patent Act provides that the “Federal Court has jurisdiction…to order that any entry in the records of the Patent Office relating to the title to a patent be varied or expunged”. The application before the Federal Court sought just that. On the plain language of . . . section 52 of the Patent Act, the Federal Court had jurisdiction over the appellant’s application.

There is no constitutional objection to the grant of jurisdiction under s 52 [49].

Wednesday, August 5, 2020

Is Harm to Third Parties Taken into Account at the Second or Third Step of the RJR-MacDonald Test?

Arctic Cat, Inc v Bombardier Recreational Products Inc 2020 FCA 116 Rivoalen JA


Yesterday’s post provided an overview of the facts in Rivoalen JA’s decision to deny Arctic Cat’s application for a stay pending appeal. This posts focuses on a purely legal issue: is harm to third parties taken into account at the second or third step of the RJR-MacDonald test for an interlocutory injunction / stay? Why does it matter?

To recap the facts, AC and BRP are competitors in the snowmobile market. BRP sued AC for infringement of various patents and finally prevailed in a decision released in June: 2020 FC 691 (here). AC then sought a stay pending an appeal, which is the subject of this decision. The stay will determine whether AC can ship infringing machines for the upcoming winter season (summer-fall 2020 is the crucial window [13]). Since the patent expires next June, AC will be able to return to this design for the following season, regardless of the outcome of any appeal, which will affect only damages. The sleds that are affected have already been manufactured and are stored in US warehouses [9]. AC submitted that it would not be be possible to redesign them for the upcoming season [9] and Rivoalen JA’s decision proceeded on that basis [34], [35].

As usual in the FC approach to the RJR-MacDonald [1994] 1 SCR 311 test, whether AC would suffer irreparable harm was a key issue. In addition to the harm that AC argued it would suffer directly, AC argued that their dealers would suffer irreparable harm, especially because many are single-line dealers who may not be able to supply a competing product because of territorial exclusivity agreements [27].

On this issue, Rivoalen JA found that irreparable harm to the dealers had not been established on the facts [31], but:

[32] More importantly, whatever harm the dealers may suffer personally cannot be relied upon by the appellants to establish irreparable harm. Only harms suffered directly by the appellants can be considered in the second branch of the RJRMacDonald test. This Court has refused attempts to rely on third-party harms, other than by charities.

As authority, Rivoalen JA cited Glooscap 2012 FCA 255 at [29-30], [33-34]; Air Passenger Rights v Canada 2020 FCA 92 at [30], and Chinese Business Chamber of Canada 2006 FCA 178 at [6-7]. These decisions (none of which concern patents), do indeed state that the interests of third parties can only be considered at the third stage of the test, the balance of convenience, relying on RJR-MacDonald [1994] 1 SCR 311 at 341, in which the Court stated that the public interest, and by implication harm to third parties, “is more appropriately dealt with in the third part of the analysis.” (Metropolitan Stores [1987] 1 SCR 110 at 128 is also relied upon, but that statement is much more ambiguous.)

On the other hand, there are cases such as Marketing International (1977) 35 CPR(2d) 226 at 231 (FCA), Procter & Gamble Co v Bristol-Meyers Canada Ltd (1978), 39 CPR(2d) 171 at 177 (FCA) and, post-RJR-MacDonaldAstraZeneca Canada 2005 FCA 208 [20], holding that it is proper to take harm to the public into account at the irreparable harm stage.

Monday, August 3, 2020

Stay Pending Appeal Denied in the Absence of an Undertaking

Arctic Cat, Inc v Bombardier Recreational Products Inc 2020 FCA 116 Rivoalen JA


In this decision, Rivoalen JA denied Arctic Cat’s application for a stay pending appeal. That in itself is not so unusual, but what is odd is that there is no suggestion that the patentee, BRP, provided an undertaking in damages to address the issue of irreparable harm. From the complete absence of any discussion of the point, I have to suspect that an undertaking was not even requested by AC, or offered by BRP. In any event, the absence of an undertaking substantially affects the analysis, and accordingly, I will focus on that issue in this post.

As discussed here, the patents at issue related a “significant breakthrough” in snowmobile design made by BRP, that moved rider weight forward. BRP’s “Rider Position Patents” covered the forward position itself, and the 264 “Frame Construction Patent” relates to a frame assembly that facilitates the construction of sleds with the rider forward position. The litigation started when BRP filed its statement of claim in December of 2011 [43]. The trial started in 2015, with Roy J’s decision 2017 FC 207 following in 2017. Roy J held the Rider Position Patents to be invalid, and the Frame Construction Patent to be not infringed. On appeal, Roy J’s decision on the invalidity of the Rider Position Patents was upheld, but the FCA 2018 FCA 172 reversed on claim construction in respect of the Frame Construction Patent and remanded to Roy J to determine validity and infringement on the correct construction. Finally, on 15 June 2020, eight years after litigation started, and with less than a year left on the term of the 264 patent, BRP prevailed, with Roy J’s decision on remand, 2020 FC 691, holding the Frame Construction Patent to be valid and infringed. Roy J granted a permanent injunction, as is normal, over the objections of AC: see here.