Effect of Errors in Priority Application

Amgen Inc v Pfizer Canada ULC 2020 FC 522 Southcott J
            1,341,537 / filgrastim / NEUPOGEN / NIVESTYM

Yesterday’s post gave an overview of the facts and discussed the main substantive holding in this decision, namely that the invention was obvious. This post turns to two priority issues: the effect of errors in the priority application; and disclosure of data in the priority application.

The 537 patent claimed priority from a prior US application — the 959 application — and there was a significant debate as to whether priority could properly be claimed. It is unusual to see this issue discussed at length in Canadian cases (here’s the only other case I’ve seen since this blog started), so the decision on this point is interesting for that reason alone. Because the 537 patent is governed by the old Act, the question was whether the prior application was for “the same invention” [186]. Under the new Act, s 28.1(1)(a)(i), the question is whether the prior application “disclos[es] the subject-matter defined by the claim.” Because of the different wording, it is not entirely clear that Southcott J’s analysis would apply under the new Act, but his analysis did not turn on the precise wording, so it may well provide useful guidance.

Errors in priority application
As discussed in yesterday’s post, the invention relates to filgrastim, which is a recombinant DNA version of the naturally occurring “granulocyte colony-stimulating factor” or “G-CSF” [12]. The 959 priority application listed (i) the amino acids sequence of the naturally occurring G-CSF; (ii) the corresponding DNA codons; (iii) the corresponding cDNA. There were admittedly typographical errors in all three sequences [198]. These errors had been corrected in the application as filed. The parties did not identify any jurisprudential guidance on the issue of whether such errors implied that the priority application does not disclose “the same invention” [198]. Relying on the principle that the skilled person reads the invention with a mind willing to understand [204], Southcott J held that the skilled person would be capable of “resolving” the errors [202], would not be “stymied” [204] and would not conclude that the two applications “are in substance directed to different inventions” [203]. On the facts, this conclusion strikes me as sound for the reasons given by Southcott J.

More broadly, Southcott J’s holding means that the mere presence of errors in the priority application does not preclude claiming priority from it. However, he did not purport to set out a specific test. On the facts, it seems that the errors were relatively easily corrected [199]-[200], so his decision would be consistent with a rule that a skilled person must be able to correct the errors without undue effort. But his decision is also consistent with a more relaxed rule; given that the errors could be readily resolved, it was not necessary for Southcott J to address the question of whether misleading errors would preclude a priority claim.

I would suggest that the case law on prior disclosure for purposes of anticipation might be relevant. Under the current Act, the test for claiming priority is whether the priority application “disclos[es] the subject-matter defined by the claim,” while the test for anticipation under 28.2 (1) is whether “[t]he subject-matter defined by a claim” has been “disclosed.” The similarity in text of the provisions suggests that the applicable test for “disclosure” might be the same (subject to purposive considerations to the contrary). The leading case on disclosure for the purposes of anticipation is Sanofi 2008 SCC 61, which adopted the two part “enabling disclosure” test set out in Synthon [2005] UKHL 59. In Synthon the question was whether Smithkline Beecham's (SB) patent, claiming a particular polymorph pf paroxetine methanesulfonate ("PMS") [8], was anticipated by Synthon's prior application which also described how to make PMS [4]. However, the disclosure in the prior application was quite misleading: “A person skilled in the art, reading both documents, would think that they identified different polymorphs” [9]. The House of Lords nonetheless held that the disclosure requirement was satisfied. If the same “disclosure” test applies to claiming priority, this suggests that priority might properly be claimed even if the errors in the priority document were misleading, so long as it “disclos[es] the subject-matter defined by the claim,” in the sense that “disclosure” is used in Sanofi and Synthon. This also seems reasonable on policy grounds. Otherwise, in a case like Synthon, where Synthon's prior application anticipated SB's patent notwithstanding the errors, it would seem that Synthon should be able to claim priority for its own subsequent filing based on that prior application, notwithstanding the errors. Were that not the case, SB's patent would be anticipated by Synthon's prior application, and Synthon's filing would be anticipated by SB's patent, and neither party would be able to obtain a patent for lack of novelty, even though it was uncontested that the compound itself was new. (I say "in a case like Synthon," because this scenario did not arise on the facts, as Synthon's prior application was a PCT application.)

Disclosure of data
A second point of interest relates to disclosure of data in the priority document. It was agreed that one of the claims, Claim 47, included a requirement for biological activity (in particular, granulocyte colony-stimulating activity). The parties seemed to have proceeded on the assumption that this meant that in order to disclose the same invention, the priority application had to disclose tests establishing the biological activity. While the tests disclosed in the priority application were not the same as those disclosed in 537 patent, Southcott J held that the tests disclosed in the priority application adequately established the biological activity; the same tests do not have to be disclosed, so long as the same activity is disclosed [211]. That holding seems right to me, so far as it goes.

While that disposed of the point on the facts, I’d question the assumption that it was necessary for the tests establishing the activity to be disclosed in the priority application at all, even though the claim includes a requirement for biological activity. While an invention must be useful, if utility can be demonstrated, the data establishing that utility need not be disclosed in the patent itself. If the data need not be disclosed in the patent, it is difficult to see why it must be disclosed in the priority application. So, suppose that granulocyte colony-stimulating activity had been demonstrated before either the priority date or the filing date, but the 537 patent itself did not disclose any data supporting utility. Aside from any priority issues, Claim 47 would nonetheless be valid, because it is not necessary to disclose evidence of demonstrated utility in the patent itself. Now suppose the disclosure in the priority application was exactly the same as that in the 537 patent, which is to say that neither of them disclosed any data. It would seem that if the disclosure in the priority application is exactly the same as in the Canadian patent, the same invention / same subject matter requirement must be satisfied. So, if biological activity is demonstrated, the Canadian application can properly claim priority from the priority application, even if neither the priority application nor the Canadian application discloses the data. This suggests that if the Canadian application disclosed the data, but the priority application did not, the Canadian application should still be able to claim priority; it doesn’t seem reasonable that the Canadian applicant should be a worse position because of having disclosed more information.

“Obvious-to-try” and “Obvious to Try”

Amgen Inc v Pfizer Canada ULC 2020 FC 522 Southcott J
            1,341,537 / filgrastim / NEUPOGEN / NIVESTYM

In this NOC action, Southcott J found Amgen’s 537 patent to be obvious on an obvious-to-try analysis in a decision that turned almost entirely on the facts. The patent at issue is governed by the old Act [10]; it was filed in 1986, claiming priority from an August 1985 US application, and was issued in 2007 – more than twenty years after its priority date. This in itself illustrates a glaring deficiency of the old Act; if the 537 had been valid, it would be unconscionable if a 35 year old invention could be used to impede current drug development.

The patent relates to a filgrastim, a hematopoietic growth factor, which stimulates the growth of white blood cells [11]. It is used to boost the immune system in cancer patients whose natural immune system has been compromised by chemotherapy [149]. Filgrastim is a recombinant DNA version of the naturally occurring “granulocyte colony-stimulating factor” or “G-CSF” [12]. The key moment in the development of filgrastim was the discovery by scientists at the Sloan-Kettering Institute that a protein secreted by a certain human bladder carcinoma cell line had a stimulatory effect on blood cell precursors. This work was published by Welte in March 1985 [24]. The Amgen scientists built on Welte by identifying and sequencing the associated gene and using recombinant DNA technology to express it in a host cell to allow production of clinically useful quantities of G-CSF [25]. It was largely uncontested that it was obvious to try making a recombinant G-CSF in light of Welte, and the real question was whether there was any inventive step required to actually do so. Southcott J ultimately held that the path forward was known at the time, and there were no hurdles that required inventive ingenuity to surmount.

I have little doubt that if, instead of publishing, Welte et al had developed their discovery into a recombinant G-CSF, their invention would have been patentable, with the inventive step lying in the identification of the naturally occurring protein and its properties. On the other hand, while the work done by Amgen was, in one sense, routine, it may be doubtful whether any institution would have actually carried it out without the lure of a patent. Thus the broad underlying issue strikes me as a version of the problem addressed by the Bayh-Dole Act; even if the true inventive step is undertaken at a university research lab, or some similar institution, and introduced into the public domain, some kind of IP rights may nonetheless be needed to incentivise the additional steps needed for commercialization. In principle that incentive may be provided by data exclusivity; in any event, a monopoly on products produced 35 years later is clearly not the right solution.

Doctrinally, Southcott J’s decision strikes me as a careful application of the obvious-to-try analysis. At the risk of sounding like a broken record, I’ll recapitulate my views on the obvious-to-try test, and show how Southcott J’s analysis is a good example. It’s a bit unfortunate that “obvious-to-try” has become the name of a doctrine, as opposed to a string of words with their ordinary meaning. I will distinguish “obvious-to-try” as the doctrine, from “obvious to try” as words with the ordinary meaning that it was obvious to try whatever it was, without any implication that the invention was obvious or not.

While “inventive step” and “obvious” are two sides of the same coin, the analysis is much easier if we approach it in terms of inventiveness. There are two questions: (1) Was it inventive to think of even trying the claimed invention as a solution to the problem at hand? If yes, then it was inventive, regardless of how easy it was to implement the solution: see eg Tye-Sil (1991) 35 CPR(3d) 350 (FCA). If the answer to (1) is no (ie it was obvious to try), then (2) was there an inventive step in achieving success? If the answer to (2) is yes, then invention is not obvious, regardless of how easy it was to think of trying it in the first place. If the answer to both these questions is no — it was not inventive to try and there was no invention in achieving success — then there was no inventive step anywhere along the path, which is to say that the invention was obvious. I acknowledge that this is not the way the “test” was laid out in Sanofi 2008 SCC 61, [69], but I suggest it is consistent with Sanofi.

Southcott J identified three “factors” from Sanofi [69]: namely Motive [291ff]; whether it was self-evident that what being tried ought to work (Self-Evident Factor) [296ff]; and Extent of the Effort [367ff]. The Motive factor goes to the first question, whether it was obvious to try the invention. It was not really contested that there was specific motivation in 1985 to make a recombinant G-CSF in light of Welte. While Southcott J remarked that “the Motive Factor is only one factor and is not determinative” [296], he later [322] invoked the statement by the FCA in Hospira FCA, 2020 FCA 30 [90] that (my emphasis):

It should be noted that, whereas being “more or less self-evident to try to obtain the invention” (per Sanofi at para. 66) is a requirement for obviousness to try, being “more or less self-evident that what is being tried ought to work” (per Sanofi at para. 69) is not a requirement but merely a factor to be considered.

I have to admit that I overlooked this paragraph in my post on Hospira, but I now see its importance in establishing that the three factors — Motive, Self-Evident and Extent of the Effort — should not be seen as three factors going into the same hopper to be blended together: Motive is a requirement, while Self-Evident is not.

This makes sense in light of the two-step approach outlined above. The first phrase in the statement by the FCA is to the effect that if it is not “more or less self-evident to try to obtain the invention,” then the invention is not obvious. This goes to question (1): if there was an inventive step involved in the decision to try the claimed invention, then the invention is not obvious; that is why being more or less self-evident to try is a requirement for obviousness. So, keeping in mind the distinction between “obvious-to-try” and “obvious to try,” we can say that if the invention is not obvious to try, then it is not obvious-to-try, which means it is not obvious.

But the converse is not true; simply because an invention is obvious to try, does not mean it is obvious-to-try; we have to go on to question (2) and look at the other factors to see whether there was an inventive step involved in achieving success. This is why, as the FCA noted, the Self-Evident factor is not a requirement. Accordingly, Southcott J treated the Self-Evident Factor and the Extent of Effort Factor as distinct factors, neither of which is determinative. It may be that it is more or less self-evident that what is being tried ought to work, given expenditure of sufficient effort and resources, and yet at the same time, “the experimentation prolonged and arduous, such that the trials would not be considered routine” Sanofi [69]; and see similarly Halocarbon [1979] 2 SCR 929, 944: "A patient searcher is as much entitled to the benefits of a monopoly as someone who hits upon an invention by some lucky chance or inspiration".

On the facts, Southcott J’s conclusion on the Self-Evident factor is that it would be self-evident that the steps leading to the claimed invention polypeptide ought to work [366]. On the Extent of Effort, he concluded that “Any potential challenges [the skilled person] would encounter could be addressed with skill and did not require inventiveness” [425]. I like this way of putting it because he phrases it in terms of inventiveness, which I find more intuitive in this context. Putting these together, (1) there was no inventive step in the decision to try the claimed invention — it was obvious to try; and (2) there was no inventiveness required in achieving success — it was (i) self-evident that it would work (ii) without undue effort; and therefore the invention was obvious-to-try and therefore obvious.

Lump Sum Costs Are the New Normal

Viiv Healthcare Company v Gilead Sciences Canada, Inc 2020 FC 486 Manson J
            2,606,282 / bictegravir / BIKTARVY

The costs tariff has been outdated for a while now, and we’ve seen a trend toward lump sum costs in complex patent cases. It seems the trend has now reached the point where lump sum costs are the new normal, at least in complex patent litigation:

[180] The parties agreed that costs should be awarded on a lump sum basis in line with recent complex patent cases in this Court (Dow Chemical Company v Nova Chemicals Corporation, 2016 FC 91, aff’d 2017 FCA 25; Sport Maska Inc v Bauer Hockey Ltd, 2019 FCA 204; Packers Plus Energy Services Inc v Essential Energy Services Ltd, 2020 FC 68). Gilead seeks 40% of its actual costs, and ViiV proposed lump sum costs in the range of 30-40% of its professional fees plus 100% of all reasonable disbursements, in line with the above cases.

[181] Given that these are sophisticated commercial parties engaged in complex patent litigation, a lump sum costs award is appropriate. That said, the summary trial was limited to construing three claims, with the issue of non-infringement flowing from the Court’s construction. In the circumstances, 30% of actual costs, plus reasonable disbursements, is appropriate.

Markman Hearings Come to Canada?

Viiv Healthcare Company v Gilead Sciences Canada, Inc 2020 FC 486 Manson J
            2,606,282 / bictegravir / BIKTARVY

The most significant aspect of this decision is the simple fact that Manson J granted a contested motion for a summary trial on a claim construction issue – essentially a Canadian version of a Markman hearing. ViiV had brought an action alleging that its 282 patent was infringed by Gilead’s bictegravir product, sold as BIKTARVY. Gilead responded with a motion for a summary trial, which Manson J granted despite ViiV’s objections and attempts to “derail” it [18]. This follows on last fall’s decision in Canmar Foods v TA Foods 2019 FC 1233 (see here and here), in which Manson J explained that for a decade after the FCA restricted the availability of summary trial in MacNeil Estate 2004 FCA 50, “summary judgment as a just, efficient and expeditious means to resolve disputes on a proportionate basis was lost” [45], until Hryniak v Mauldin, 2014 SCC 7 resulted in “a culture shift” that “opened the door for a more reasoned approach to the use of summary judgment motions” [46]. (The only other recent summary trial that I’m aware of is Cascade v Kinshofer 2016 FC 1117, but it was on consent.) It will be interesting to see whether other FC judges agree with Manson J’s view of the effect of Hryniak, and, when an appeal inevitably comes, whether the FCA agrees as well. (I must say it seems like a good idea to me, but procedure is not my area of expertise.) Manson J’s specific reasons for granting the motion on the facts of this case ([11]-[18]), will be of interest in future motions of this type, but since the discussion is brief and I don’t have anything to add, I won’t go through it.

Recourse to the disclosure in claim construction
The key substantive issue was the construction of a single claim term. The most interesting substantive point is that Manson J, after an extensive review of the case law, had recourse to the disclosure to construe the contested claim term, even though it appeared on its face “to be a clear and unambiguous term” [128]

All the claims at issue (Claim 1 is exemplary) [88], are claims to a class of compounds comprising Ring A.

Ring A is defined in the claim as an “optionally substituted heterocycle” [102]. The key issue — where the shoe pinches — is whether Ring A includes only fused and spiro rings, or also a bridged bicyclic ring, as is found in bictegravir.
Much of Manson J’s discussion concerned the proper approach to claim construction. As Manson J noted, the “[t]he law is clear that recourse to the disclosure is improper to vary the scope or ambit of the claims” [126], [136]. However, it is said that while recourse to the disclosure is “permissible to assist in understanding the terms used in the claims,” it is “unnecessary where the words of the claim are plain and unambiguous,” Dableh (1996) 68 CPR (3d) 129 (FCA) at 144 [126].

The tricky question is whether it is permissible to consult the specification as a whole only when the claims are ambiguous. What if the words of the claim appear to be clear when read in isolation, but when read in the context of the disclosure as a whole, it is apparent that they mean something different from what they “clearly” mean when read in isolation? Is it necessary to make a determination that the claims are ambiguous before recourse to the disclosure is permissible? A series of decisions have wrestled with this question.

I won’t go through Manson J’s review of the case law in detail, but I would highlight his conclusion:

[66] The common thread in all of these cases is that the court is to construe the claims through the eyes of the POSITA in light of their CGK at the relevant date. Apart from the patent specification, the only evidence the Court should consider to inform its analysis of the claims is evidence of how the POSITA would understand the claims in light of his or her relevant CGK in the context of the specification as a whole (Bombardier FCA [2018 FCA 172] at para 24).

In other words, the claims must be read in the context of the specification because the claims are part of the specification. This strikes me as clearly correct; I won’t say more as I’ve discussed the principles at length in other posts.

Manson J’s application of these principles to the facts is clear:

[128] While “optionally substituted heterocycle” as used in claim 1 appears on its face to be a clear and unambiguous term, I accept that recourse to the disclosure is necessary to understand the meaning given to these words by the inventors, and the intended scope of this claim language.

This is a clear holding, applied on the facts, that it is permissible to have recourse to the disclosure to understand the claims, even if the claims appear to be clear when read in isolation.

Essentiality
I’ll conclude with a brief comment on the essentiality analysis. ViiV argued that if Manson J construed Ring A as being limited to spiro and fused structures only, it was necessary for the Court to go on to consider whether the bridged ring was an obvious variant. Manson J rejected this, saying

[167] In light of ViiV’s admission that Ring A, as an “optionally substituted heterocycle,” is an essential feature of the invention, the Court does not accept that it should now look at a variant of this essential feature. The Free World Trust [2000 SCC 66]variant analysis focuses on a claimed element of an invention, not some sub-element or feature of the claimed element.

I’m not sure I agree with this. It is true that Free World Trust focused on what it described as “elements” but it’s not clear to me that the analysis turns on what particular aspect of the claim is identified as an “element.” The term “element” never even appears in earlier cases, such as JK Smit [1940] SCR 279 or Birmingham Sound [1956] RPC 232 (CA), that the SCC inWhirlpool 2000 SCC 67 [45], [47] identified as exemplifying the analysis. In any event, the point was not determinative, as Manson J went on to hold that in any event, ViiV had not met its burden of establishing that it would have been obvious to the POSITA at the publication date that bridged bicyclic Ring A structures would have no material effect on how the invention works” [168].