Whether Public Viewing Amounts to Enabling Disclosure is Context Dependent

Betser-Zilevitch v PetroChina Canada Ltd 2021 FC 85 Manson J

            2,584,627 / Steam-Assisted Gravity Drainage

In this decision, that turned primarily on claim construction, Manson J held the 627 patent to be valid but not infringed. There is no new law, but the decision illustrates that the claims may be construed with reference to the function of the invention, and that whether public viewing of the prior art amounts to prior disclosure is context dependent.

The 627 Patent generally relates to a modularized (Steam-Assisted Gravity Drainage) SAGD well pad for heavy oil (bitumen) production [10]. A “well pad” is a collection of well pairs used in extraction of heavy oil production. “Modularized” means that the well pad is made off-site in modules that are then transported to the well site and assembled. This is in contrast to “stick built” well pads that are entirely constructed on-site. Construction of a well pad is a major undertaking that takes months. The main advantage of modularization is reduced costs flowing from reduced on-site construction costs [7]. The well pad has several levels, which, as I understand it, are like floors in a building, except that the levels are open. The key aspect of the 627 patent was that the main flow lines were on the “first level” [12], [116], [117].

In the defendant’s allegedly infringing well pads the production flow lines were located on a higher level, which requires scaffolding for workers in the field to access, and not on the lowest level [135]. The key claim construction issue was therefore as to the meaning of “first level” in the claim. The experts disagreed as to the meaning of the term. In light of the “ambiguity” in the claims, Manson J held that the term “would be understood with reference to its functions and objects, as provided in 627 Patent specification” [116]. The object was to constrain the placement of the flow lines at a level such that they would be connected without the need for scaffolding, thus improving safety and reducing costs [116]–[118]. This illustrates that reference to objective of the invention may be useful to construe the claims.

Because Manson J held that there was ambiguity, he didn’t need to address the question of whether recourse to the disclosure is appropriate even in the absence of a finding that the claims are ambiguous. I can’t resist digressing on this point. In my view, recourse to claims is always appropriate: as Lord Hoffmann pointed out in Kirin-Amgen [2004] UKHL 46 [33], the skilled person “reads the specification on the assumption that its purpose is both to describe and to demarcate an invention.” Words take their meaning in part from the context in which they are used; the common general knowledge is part of the context, and I can’t think of any reason why the disclosure should not also be considered part of the relevant context. Since the disclosure and claims are both part of the specification, a skilled person reading the claims would have to make an effort to avoid seeing the disclosure, and I can’t think of any justification for such an artificial approach. In this case, Manson J found an ambiguity allowing recourse to the specification on the basis of a disagreement between the experts as to the meaning of the terms. If that is all it takes, then we might as well just recognize that consideration of the disclosure is always appropriate in claim construction.

An issue that arose in respect of obviousness was the relevance of prior art modularized well pads that had been transported on public roads and parked uncovered in a public parking lot [159]. The patentee argued that that “[m]erely viewing, without more” is not enough [164] relying on Bombardier 2017 FC 207 [490]. Manson J noted that “The degree of scrutiny required upon visual inspection is context specific. The level of scrutiny required to discern the external elements of a skate in a public arena, as in Bauer Hockey FC [2010 FC 361], above is evidently higher than the scrutiny required to view the components of a large SAGD module” [165]. The point, as I understand it, is that in order to be set up in an anticipation or obviousness attack, the prior art must have been become “available to the public” before the relevant date, and the information which is the quid pro quo for the patent is made available only if the requirements for enabling disclosure are satisfied. A view from a distance of a well pad in which flow lines are visible may be an enabling disclosure, even though a view from a distance of hockey skates does not provide an enabling disclosure of the details of the fabrication method.

PM(NOC) Related Rights of Action not Subject to Limitation Period

Merck Sharp & Dohme Corp v Sandoz Canada Inc 2020 FC 1180 Southcott J

2,518,435 / sitagliptin / JANUVIA

I believe this is the first decision addressing s 8.2 of the PM(NOC) Regulations, which provide for “Related rights of action,” as described by the RIAS. In particular, s 8.2 provides that “[o]n receipt of a notice of allegation” the patentee may bring an action for infringement of a patent, other than the patents that are set out in the NOA, that could result from the grant of an NOC. The RIAS explains that “Such patents can create legal uncertainty if there is risk that they could be infringed by the generic product. To facilitate legal consideration of such patents without expanding the scope of proceedings under the proposed Regulations, related rights of action are proposed.” The RIAS refers by way of example to patents claiming chemical intermediates or processes for making a drug, which may be infringed by the sale of a particular drug but which are not eligible for listing. A quick look at the patent at issue in this case suggests that it falls into both these categories, as being a process for making intermediates.

The question raised in on this motion is whether s 8.2 is subject to a limitation period, in the same manner as the primary provision, s 6, which requires the patentee to bring an action respecting the patents that are set out in the NOA within 45 days. The parties and Southcott J all agreed that the matter was appropriate for determination by summary judgment [20], [22].

Southcott J held that there is no limitation period applicable to actions brought under s 8.2 [77]. The first point, and a very powerful one, is that there is nothing in the text of s 8.2 suggesting that there is a limitation period, in contrast to the clear limitation period set out expressly in s 6 [41]–[42]. As the FCA has noted, “the clearer the ‘ordinary meaning’ of the text, the more compelling the contextual considerations must be in order to warrant a different reading of it, especially when that involves adding words to those used by the legislator” Biolyse Pharma 2003 FCA 180 [13].

The role of the text is “dominant” (Canada Trustco 2005 SCC 54 [10]), but not absolute, and accordingly Southcott J went on to consider a variety of purposive and contextual arguments made by the parties. The defendants’ basic argument was that the purpose of the Regulations generally and s 8.2 in particular is to reduce the legal uncertainty associated with generic entry [38], [46]. The plaintiffs’ argued that the purpose of s 8.2 in particular is to remove the previously existing barriers to bringing a quia timet action, but not necessarily to eliminate legal risks prior to generic product launch [46]. (And see [35], explaining the prior hurdles to a quia timet action.)

After canvassing the general principles of statutory interpretation, Southcott J considered the specific arguments made by the parties. The defendants generally relied on the RIAS and a high level assessment of the legislative purpose. The plaintiffs responded directly to this by noting that even if the we were to accept that s 8.2 incorporated a limitation period, the defendants acknowledged this would only apply to an action brought under that provision, ie a statutory quia timet action. A patentee who missed the deadline would still be free to sue on an unlisted patent following market entry, when it is no longer necessary to bring the aciton quia timet. This is in contrast to an action related to the patents that are the subject of an NOA, which, pursuant to s 6.01 cannot normally be the basis for a subsequent action [64]. The defendants did not go so far as to argue that a parallel finality provision should be read in along with the limitation provision. Consequently, it is difficult to conclude that the purpose of the provision is to eliminate the generic’s legal risk [69]. If anything, incorporating a limitation period in s 8.2 without the accompanying principle of finality would increase risk and expense, as the patentee would simply wait until launch to bring its action.

The plaintiffs also relied on several arguments turning on the way s 8.2 and a putative limitation period would interact with the Regulations as a whole. For example, the plaintiffs noted that the NOA provides the first person with sufficient information to be able to decide whether to bring an action quickly enough to meet the 45 day deadline, but the “as the NOA is not directed to unlisted patents, the first person would not be similarly equipped to commence a s 8.2 action within the same 45 days” [66]. The patentee made several other arguments of a similar nature, which Southcott J generally accepted. It would take me as much time to summarize them as the original decision, so I’ll just say that I found Southcott J’s analysis to be careful and convincing.

The cumulative weight of these contextual and purposive arguments, combined with the clear textual argument, is inescapable. Southcott J therefore concluded that there is no limitation period applicable to the actions brought under s 8.2 [78]. (Perhaps it is more accurate to say that s 8.2 does not incorporate any limitation period, as the actions brought under that provision are no doubt subject to the general limitation periods under the Act.)

Given Southcott J’s careful consideration of the provision, his statement of its purpose will be helpful if it requires interpretation on some other point in the future:

[71] I accept that the legislative purpose of s 8.2 is the removal of previously existing barriers to quia timet actions, so as to facilitate legal consideration of unlisted patents and thereby address uncertainty resulting from the risks associated with such patents. However, this purpose does not require the elimination of such uncertainty within any particular timeframe, as no interpretation of s 8.2 can eliminate the uncertainty that the potential for post-launch litigation of unlisted patents will continue to present. Rather, the elimination of the common law barrier achieved by s 8.2 permits earlier access to legal consideration, and therefore earlier resolution, of unlisted patent infringement claims. This result is in keeping with the overall objective of striking a balance between effective patent enforcement and timely generic market entry.

Caution in Following UK Caselaw Regarding Added Matter

Western Oilfield Equipment Rentals Ltd v M-I LLC 2021 FCA 24 Locke JA: Gleason, Mactavish JJA affg 2019 FC 1606 O'Reilly J

            2,664,173 / Shaker and Degasser Combination

Previous posts have provided some background on this case and addressed reasonable compensation, the Gillette defence and the “best mode” requirement. This post turns to Locke JA’s comments regarding added matter.

Section 38.2(2) of the Act provides that an application cannot be amended to add matter “that cannot reasonably be inferred” from the specification as it was on its filing date. This provision broadly corresponds to Art 123(2) of the EPC, which provides that the application or patent may not be amended to add matter “which extends beyond the content of the application as filed.” As previously noted, the claims that were infringed had been introduced during prosecution, after the patentee became aware of the appellants’ product. The appellants argued that these claims were invalid for added matter [138]. There is little Canadian caselaw dealing with our provision, and the appellants had therefore relied on UK caselaw as support for a “strict” test that “subject matter should be considered to have been added unless such matter is clearly and unambiguously disclosed in the original application either explicitly or implicitly” [140]. (See here for my discussion of O’Reilly J’s decision on this point.)

Locke JA remarked that we should be “wary” of following the UK approach, for three reasons. First, “it presumes that matter has been improperly added, and places the burden on the respondent to establish otherwise,” whereas in Canadian law a patent is presumed to be valid (s 43(2)) and accordingly the burden is on the party attacking the patent. Second, the text of the provision is different, and in particular the EPC provision* does not have any kind of “reasonableness” language: [142]. Third, the UK provision provides explicitly for patent revocation, while s 38.2 does not [143]. Accordingly, Locke JA simply applied the statutory “cannot reasonably be inferred” test and held that the feature in question was “at least reasonably inferable” from the original specification [147].

I agree with Locke JA we should not adopt the UK approach uncritically, for all the reasons he gives. On the other hand, comparative law can potentially be very useful, particularly when, as in this case, there little Canadian caselaw and a great deal of UK caselaw; the foreign caselaw can help illustrate the kinds of problems that may arise and that need to be taken into consideration, even if we do not adopt the same solutions as foreign law. For example, the problem of “intermediate generalization” arises when the specification provides a number of specific examples and the patentee seeks to claim a broader class which is said to be exemplified by those examples. This has caused considerable difficulty and has been addressed in a number of EPO Board decisions and UK cases: see eg Nokia v IPCom [2013] RPC 5 [56]ff. While the problem does not arise in this case, the European jurisprudence is likely to be useful in illuminating the issues when it does arise, even if the Canadian courts choose not to adopt the European solution.

In my view, Phelan J struck exactly the right note in Amazon.com 2010 FC 1011 [32] when he observed that it can be useful to look to other regimes when faced with a question that is novel in Canadian law, but this must be done “mindfully” [32].

A “mindful” consideration of foreign law means that we should not simply pluck a test from foreign jurisprudence. It is important to get the foreign law right before we decide whether to adapt it to the Canadian context. In the UK approach, added matter is permisible if it can be “implicitly” inferred from the original specification, and this may soften the apparently strict nature of the test. So, another articulation provided by Jacob LJ in Vector Corp [2007] EWCA Civ 805 [4] is that:

[T]he test of added matter is whether a skilled man would, upon looking at the amended specification, learn anything about the invention which he could not learn from the unamended specification.

Given that a skilled person approaches the specification with “a mind willing to understand,” it may be that the test is not so strict after all. On the one hand this suggests that UK law may provide some helpful guidance despite the differences in the text; but on the other hand, the need to understand the subtleties of foreign law is another reason to be “wary” of it.

Another point is that it may be helpful to look to the purposive considerations adverted to in foreign law. To the extent the purpose of the Canadian and UK provisions are different, that is additional reason to be wary of the same test, but to the extent the purpose is the same, that is a reason to consider the foreign test more closely, always subject to the text of the provision.

This is perhaps most relevant to Locke JA’s third point, that the UK Act provides explicitly for patent revocation while s 38.2 does not [143]. This raises the question of whether added matter contrary to s 38.2 is a basis for invalidating a claim in a granted patent, or whether it is only ground for refusing an amendment during prosecution. Some of the purposive considerations discussed by Jacob LJ in Vector Corp are potentially relevant to this issue:

[5] The reason for the rule was explained by the Enlarged Board of Appeal of the EPO in G1/93 ADVANCED SEMICONDUCTOR PRODUCTS/Limiting feature [1995] EPOR 97 at [Reasons 9]:

"With regard to Article 123(2) EPC, the underlying idea is clearly that an applicant shall not be allowed to improve his position by adding subject-matter not disclosed in the application as filed, which would give him an unwarranted advantage and could be damaging to the legal security of third parties relying upon the content of the original application."

[6] Mr Richard Arnold QC provided a clear articulation as to how the legal security of third parties would be affected if this were not the rule:

The applicant or patentee could gain an unwarranted advantage in two ways if subject-matter could be added: first, he could circumvent the "first-to-file" rule, namely that the first person to apply to patent an invention is entitled to the resulting patent; and secondly, he could gain a different monopoly to that which the originally filed subject-matter justified.

These considerations seems to me to also be relevant in the interpretation of the Canadian provision, always subject to the text.

*Note that the UK courts rely directly on Art 123(2) rather than on the somewhat differently worded s 76(3) of the Patent Act 1977, which was a “cack-handed” implementation of the EPC provision: Napp Pharma [2009] EWCA Civ 252 [69].

Is the Best Mode Requirement Restricted to Machines?

Western Oilfield Equipment Rentals Ltd v M-I LLC 2021 FCA 24 Locke JA: Gleason, Mactavish JJA affg 2019 FC 1606 O'Reilly J

            2,664,173 / Shaker and Degasser Combination

Previous posts have provided some background on this case and addressed reasonable compensation and the Gillette defence. Locke JA also made two noteworthy points with respect to the best-mode requirement set out in s 27(3)(c), which requires that the specification must “in the case of a machine,” explain the best mode of practising the invention. As discussed here, O’Reilly J had dismissed this argument on the basis that it “relates solely to machines and, as I understand the patent, it claims a method and a system, not a machine” [FC 146], [118]. Locke JA remarked that the mere fact that the claim at issue was phrased as being to “a system” did not in itself establish that it is not to a machine [119]. This makes sense to me; the question of whether the claim is to a machine should be a matter of the substance of the claim and not its form.

More importantly, Locke JA stated that “there is considerable doubt as to whether, despite the wording of paragraph 27(3)(c), the best-mode requirement is limited to machines,” citing various commentators [119]. This is interesting given that the FCA in Pfizer v. Novopharm / sildenafil (NOC) 2010 FCA 242 [72] approved Snider J’s holding in Ramipril FC 2009 FC 676 [329] that “the ‘best mode’ obligation only arises in the case of a patent to a machine.” I must say that Snider J’s textual analysis strikes me as fairly compelling, but far be it for me to dissuade the FCA from relying on commentators. Locke JA’s observation was expressly obiter, and the point made no difference to the outcome, but nonetheless the point can no longer be considered settled.

The Gillette Defence Is Typically Unhelpful

Western Oilfield Equipment Rentals Ltd v M-I LLC 2021 FCA 24 Locke JA: Gleason, Mactavish JJA affg 2019 FC 1606 O'Reilly J

            2,664,173 / Shaker and Degasser Combination

Friday’s post provided some background on this case and addressed the issue of reasonable compensation. Today’s post reviews Locke JA’s very helpful discussion of the Gillette defence. (The discussion did not turn on the way the issue arose on the facts and the issue did not affect the result.)

Locke JA began by reiterating the classic understanding of the Gillette defence, which permits a defence based on establishing that its allegedly infringing product is the same as, or not patentably distinct from, the prior art; if that can be established on the facts, the asserted claims must necessarily be invalid or not infringed: [75]–[76] citing Gillette (1913) 30 RPC 465 (HL) and JK Smit [1940] SCR 279. While this is not new law, it is important nonetheless, as Locke JA emphasized that “contrary to some of the jurisprudence of the Federal Court, the success of a Gillette defence does not depend on a conclusion that the claim in question is invalid for anticipation or obviousness. Quite the contrary, the principal benefit of a Gillette defence is to avoid the need to construe the claims and reach a conclusion on validity and infringement issues.” [79]. Locke JA noted that the FCA itself in Eurocopter had skirted with a similar misunderstanding [80]

The theory of the Gillette defence is that it permits “a shortcut around the often difficult and time-consuming process of construing the claims of a patent, and then determining whether those claims are valid and whether they have been infringed” [77]. Unfortunately, in practice the Gillette defence has often complicated rather than simplified the proceedings. The defendant will rarely risk everything on the Gillette defence, but will typically assert it as an alternative to the analysis of claim construction and validity [78]–[79]. This means that even when the Gillette defence is raised, the court will normally have to construe the claims and assess validity and infringement, in order to address the other anticipation and infringement arguments. In that case, the Gillette defence is superfluous; rather than simplifying proceedings, by obviating the need to consider infringement and validity, it complicates matters by introducing an extra and unnecessary argument, which may lead to confusion [79].

The general message is that when claim construction, validity and infringement are fully argued, it is probably preferable to avoid the Gillette defence entirely.

Questions on Reasonable Compensation Answered — and Left Open

Western Oilfield Equipment Rentals Ltd v M-I LLC 2021 FCA 24 Locke JA: Gleason, Mactavish JJA affg 2019 FC 1606 O'Reilly J

            2,664,173 / Shaker and Degasser Combination

In this decision Locke JA for the FCA affirmed O'Reilly J’s holding that certain claims of M-I’s 173 were valid and had been infringed by Western Oilfield Eqpt. The decision turned largely on the deferential standard of review applicable to questions of fact, and Locke JA pointedly remarked that the appellants had not paid sufficient attention to the standard of review [10]. The appellants raised approximately two dozen issues on appeal [7]. Locke JA was not impressed. He remarked on “the lack of wisdom of raising so many issues on appeal,” noting that this undermined the appellant’s case by obscuring the real issues and wasted the Court’s time [9]. As a result of the appellants’ “failure to reasonably limit the issues before the Court, and the respondent’s success on each argument raised,” costs were awarded at the top of Column V of the Tariff [162].

Despite the largely factual nature of the issues, the FCA did address a number of legal points. I’ll provide background and address reasonable compensation in this post, and subsequent posts will look at the Gillette defence, best mode and added matter. In addition, Locke JA reaffirmed the well-established law that the four-step Sanofi test for obviousness is not mandatory [109] and the fact that the embodiment disclosed in the specification is not commercializable does not mean that it lacks the necessary scintilla of utility needed to support a patent [124].

Background

As discussed here, the 173 patent relates to a machine called a “shale shaker,” used in the oil and gas industry to remove solids from the slurry that emerges from an active well in order to separate the drilling fluid for reuse. A shale shaker shakes the slurry over a screen to allow the fluid to fall through while the rock cuttings are carried away. Shaking alone does not remove all of the fluid, and the cuttings remain wet after passing through a standard shaker. The invention consisted of a method and system for applying a vacuum (“pressure differential generator”) to the cuttings to recover more of the drilling fluid. One point that underlay several of the appellants’ objection was that the patent was laid open on April 10, 2008, and the appellants’ activities did not infringe the claims as they were drafted at that time. The claims that were infringed were introduced during prosecution, on October 2, 2013, after the patentee became aware of the appellants’ product [53], [FC 14], [FC 89].

There were a number of contentious issues on claim construction, but these turned on the evidence and, unsurprisingly—given that the claims were added to capture the appellants’ product—O'Reilly J ultimately construed the claims in favour of the patentee. The finding on infringement followed directly. The appellant raised the motivation of the patentee in amending the claims in a couple of ways [22], [144], but Locke JA noted that the respondent’s motivation in amending the claims is in itself irrelevant to either claim construction or validity [22]. As he observed at [21], quoting his own prior decision in Camso 2019 FC 255 [197] aff’d 2020 FCA 183:

A patent applicant is perfectly entitled, and indeed is well-advised, to draft patent claims with an eye on competitors’ products. The limitation is that the applicant must respect all of the requirements for valid claims.

Reasonable Compensation

The decision raised three issues in the interpretation of s 55(2), which provides that the patentee is entitled to “reasonable compensation” for any damage sustained by the patentee after the specification was laid open and before the grant, “that would have constituted an infringement of the patent if the patent had been granted on the day the specification became open to public inspection.” For simplicity, Locke JA referred to this as “pre-issuance infringement” [52].

There were two main issues on appeal. One of the appellants, FPM, had apparently stopped doing business in 2014, prior to the issuance of the 173 patent in 2015. The first question was whether a party who engages in pre-issuance infringement “can avoid liability under subsection 55(2) by discontinuing the activities in question prior to issuance of the patent.” That is, must a party be engaged in infringing activity as of the date of issuance in order to be liable for pre-issuance infringement? The answer is no. There is nothing in the text of the provision to suggest that result [59]. Further, the purpose of s 55(2) is “to compensate inventors whose published inventions are used without permission before they can enforce their rights,” [63] and “it would defeat the purpose of the provision if the pre-issuance infringer could continue its activities, possibly for many years, and then avoid all liability by discontinuing those activities just before the patent issues” [63].

The second question was whether a party can be liable for inducement of pre-issuance infringement. The answer is yes. Inducement is “merely a form of patent infringement, not a distinct tort” [60]; there is nothing in the text to suggest liability for inducement is excluded [62]; and it would “defeat the purpose of the provision if one could avoid liability by inducing another to directly infringe. The reasons for imposing liability for infringement on one who knowingly induces a third party to infringe after a patent has issued apply equally before issuance” [63].

Neither of these issues strikes me as particularly difficult, and I agree entirely with Locke JA’s analysis.

The more difficult issue, to my mind, concerns the timing of the liability. As noted above, the appellants’ activities did not infringe the claims as they were drafted at the time the patent was laid open. The claims that were infringed were introduced five years later during prosecution, after the patentee became aware of the appellants’ product [53], [FC 240]. It was common ground on appeal, and, as I understand it, also at trial, that the appellants’ could not be liable for any of its activities prior to 2 October 2013, when the infringed claims were introduced [53]. O'Reilly J held that reasonable compensation should be awarded for the period starting 2 October 2013, and this was affirmed on appeal [66].

The difficulty is that s 55(2) makes a party liable for activity “that would have constituted an infringement of the patent if the patent had been granted on the day the specification became open to public inspection.” It was common ground that the appellants’ activities would not have constituted an infringement of the claims as of the date the patent was laid open, and it would seem to follow on a straightforward reading of the text that the appellants are therefore not liable for reasonable compensation for any part of the period prior to issuance—not even for the period after 2 October 2015 when the claims were amended to catch the appellants' product. The rule applied in this case is that a party will be liable for activities that would have constituted an infringement if the patent had been granted at any time subsequent to the date the patent was laid open, which does not seem to me to be entirely consistent with the text. That is not conclusive, of course; as discussed here, there might be contextual or purposive considerations that are also relevant one way or the other.

This point does not seem to have been argued either at trial or on appeal, and it would appear to remain open. Locke JA noted that FPM stopped doing business in 2014 which was after the claims had been amended to encompass its activities. He stated that this means that FPM could not avoid liability based on its activities being entirely prior to the amendment, “even if that were a basis for avoiding liability. I express no opinion here as to whether that would be a basis for avoiding liability” [56]. I take this to mean that the timing issue remains open. Locke JA did also state that the facts are not in dispute, “and there is no argument that either of the appellants is liable for any of its activities prior to October 2, 2013" [53]. I take this to mean simply that there was no argument, and not that Locke JA was necessarily endorsing this position as a matter of law.

Skepticism Regarding I G Farbenindustrie Test for Selection Patents

Bristol-Myers Squibb Canada Co v Pharmascience Inc 2021 FC 1 Zinn J

2,461,202 / 2,791,171 / apixaban / ELIQUIS

This decision consolidated four actions concerning the 202 patent, which claims the compound apixaban and its use in treating thromboembolic disorder, including stroke [2], [27], and the 171 patent, which claims various formulations of apixaban[2], [114]. The decision raises several legal issues, though largely as a result of some “creative” submissions by the defendants [125]. The most important point is that Zinn J expressed considerable skepticism regarding the continued vitality of the I G Farbenindustrie test for selection patents. Zinn J addressed the 202 patent first and then the 171 patent. I won’t cover every issue, but only those that caught my eye.

Date for assessing sufficiency

The first attack on the 202 patent was based on insufficiency. There was some discussion of whether the date for assessing sufficiency is the date of filing or the date of publication [36]–[42]. I’ll note that there has indeed been some debate in the case law on this issue, but the point was settled in Idenix v Gilead 2017 FCA 161 [46], holding that the correct date is the filing date, in a case that was crucial to the result. (I’d suggest that while Idenix ruled out the publication date, there remains an open question as to whether it should be the claim or the filing date, as that distinction was not at issue in Idenix—I’d argue that it should be the claim date, but that’s for another day.) Idenix was not cited by Zinn J, presumably because it was not cited to him, and consequently he appears to have considered the publication date to be the relevant date [49].

In any event, so far as I can tell, nothing turned on this point. As both filed and published the application claimed 10¹⁰⁰ compounds, and it was only just before issuance that the claims were narrowed to one compound, namely apixaban [37]. The defendants apparently argued the patent was insufficient because it did not enable all 10¹⁰⁰ compounds to be made [37]. That is, sufficiency had to be assessed in terms of the document as it existed at the relevant date (whatever that date might be). That is certainly a very creative argument, but. . . no, “The test for sufficiency is whether [the skilled person] is in a position to work the invention. The invention is that claimed in the patent, and until it is issued, there is no patent” [49]. That must be right — if it weren’t almost every granted patent would be invalid, because claims are routinely amended during prosecution in response to objections by the examiner, which necessarily takes place after filing. If anticipation, utility and obviousness all had to be assessed as of the claim date or filing date, then amendments to the claims could never cure any objection.

Studies demonstrating utility need not be referenced in the patent

The defendants also argued that the patent was invalid for lack of utility because the study demonstrating utility must be referenced in the patent. Zinn J rejected this proposition, relying primarily on statements by the SCC in Sildenafil 2012 SCC 60 [39]–[40] and AstraZeneca 2017 SCC 36 [58]. I’ll add that Gleason J (as she then was) reviewed the authorities in Lilly v Apotex 2015 FC 1016 [138]–[141]. She pointed out that the FCA authorities supporting such a requirement were obiter, and there are many cases to the contrary, and she therefore concluding that “the weight of authority is to the effect that the evidence of demonstrated utility need not be referenced in the patent for the patentee to rely on it” [142]. I think this point can now be considered settled, though it would certainly be desirable to have it addressed by the FCA to remove any lingering doubt.

A genus too large to think about does not anticipate

The most interesting issue was an anticipation attack based on 2,349,330, which disclosed and claimed a large class of compounds that were potentially useful in the treatment and prevention of thromboembolic disorders [79]. The class is very large. One of the experts for Sandoz did a calculation of the number of compounds disclosed and came up with “Some number that is just larger than I can think about.” He agreed that it was “more possibilities than stars in the universe” [80]. On the evidence provided by Sandoz’s experts, the class encompassed apixaban [77], but did not specifically disclose it, or describe how to make it [82]. On this evidence, Zinn J had no difficulty in concluding that the 330 patent did not anticipate apixaban [83].

Nor was apixaban obvious over the 330 patent, as only a small percentage of the disclosed compounds would actually be effective in treating thromboembolic disorders, and identifying which would involve “complex, time-consuming, unpredictable research” [87]. Obviousness-type double patenting failed for the same reason: [96]–[97].

IG Farbenindustrie test doubted

Apixaban is therefore a classic selection over the 330 patent. Significantly, it was only after finding that apixaban was neither anticipated nor obvious over the 330 patent that Zinn J turned to question of selection patents as such and the three-part test set out in IG Farbenindustrie (1930), 47 RPC 289 (Ch) 322-23, that was approved in Sanofi 2008 SCC 61 [9]-[11]. The factors, in summary, are: (1) a valid selection must possess some substantive advantage over the genus; (2) all of the selected members must possess the advantage; (3) the advantage must be peculiar to the selected group. Zinn J held that these three propositions are “not a stand-alone basis to find a selection patent invalid,” relying on the express holding to that effect in Olanzapine 2010 FCA 197 [33] where the FCA stated that an assessment of whether the IG Farbenindustrie conditions has been met “does not constitute an independent basis upon which to attack the validity of a patent” [104]. Despite this holding by the FCA, the IG Farbenindustrie conditions continue to resurface as an independent ground of attack, most recently in Lilly v Mylan 2020 FC 816, discussed here.

Zinn J’s contribution to this debate was to go back and quote from the explanation for these factors given by Maugham J in IG Farbenindustrie itself [100]. In the passage quoted by Zinn J, Maugham J specifically connected the first and second factors to standard grounds of attack, namely utility and insufficiency. The explanation given by Maugham J for the third factor is somewhat more obscure, and also more controversial, as it was specifically criticized in Dr Reddy’s [2009] EWCA Civ 1362 [39]. I suspect that the third factor was an imperfect attempt to avoid the problem of arbitrary selection, which is properly treated as an aspect of non-obviousness. I might also add that I am not at all convinced that the first and second factors are properly related to utility, as Maugham J indicated. But the important point here is that Maugham J did explain the factors he set out in terms of the traditional grounds of invalidity. As I read it, Maugham J considered these factors to be a helpful way of stating the standard principles in the context of a selection patent, and not as an entirely new ground of attack. I agree therefore, with Zinn J’s observation that for Maugham J, and for Rothstein J in Sanofi referencing Maugham J’s analysis, “a failure to meet his three propositions regarding selection patents does not invalidate the patent; rather its validity is subject to the usual grounds relevant to any patent” [101]. I am very pleased to note that Zinn J also cited my blog post, Time to Relegate IG Farbenindustrie to the Dustbin of History, in which I pointed out that the SCC in Sanofi did not apply the IG Farbenindustrie as an independent ground of invalidity on the facts, but rather relied on the standard grounds of anticipation and obviousness, as Zinn J did in this case. That is why it is significant that Zinn J addressed anticipation and obviousness before turning to the IG Farbenindustrie factors; having decided that the 202 patent was valid on the standard principles, there was no more role for the IG Farbenindustrie factors to play. Out of an abundance of caution, Zinn J nonetheless went on to hold that in any event, the 202 patent met the IG Farbenindustrie test [106]–[108].