[2017] 2 F.C.R. 280
A-47-15
2016 FCA 119
Mylan Pharmaceuticals ULC (Appellant)
v.
Eli Lilly Canada Inc., ICOS Corporation and the Minister of Health (Respondents)
Indexed as: Mylan Pharmaceuticals ULC v. Eli Lilly Canada Inc.
Federal Court of Appeal, Dawson, Trudel and Rennie JJ.A.—Ottawa, December 1, 2015; April 20, 2016.
Patents — Infringement — Appeal from Federal Court decision allowing respondent Eli Lilly Canada Inc.’s application for order under Patented Medicines (Notice of Compliance) Regulations, s. 6 prohibiting notice of compliance (NOC) from being issued to appellant — Impugned Canadian Patent No. 2226784 (ꞌ784 patent) claiming tadalafil, 3-methyl tadalafil (two compounds) for treatment of erectile dysfunction — Specifically, two compounds inhibiting production of PDE V (enzyme preventing erections) — ꞌ784 patent having priority date of July 14, 1995, filing date of July 11, 1996, publication date of February 6, 1997 — Eli Lilly previously acquiring ꞌ377 patent that claimed tadalafil, having priority date of January 21, 1994, filing date of January 19, 1995 — Pfizer publishing ‘902 patent application for PDE V inhibitor sildenafil on December 22, 1994 — Appellant filing notice of allegation that ꞌ784 patent invalid for inutility on basis of lack of sound prediction, obviousness-type double-patenting — Federal Court determining priority date of ꞌ377 patent as relevant date on which to assess impugned patent — Determining, inter alia, that at priority date of ꞌ377 patent, use of PDE V inhibitors patentably distinct from ꞌ377 patent — Rejecting challenge to utility of patent — Whether Federal Court erring in finding that impugned patent not invalid for either obviousness-type double-patenting or lack of utility — Federal Court committing no reversible errors in finding impugned patent not invalid for either obviousness-type double-patenting or lack of utility — While Federal Court erring in referring to specification when construing claims of ꞌ377 patent, error of no consequence — As to date from which obviousness-type double-patenting should be assessed, publication date of ꞌ784 patent not correct date — Inappropriate to use any date after claim date of second patent — Otherwise Court would consider prior art beyond what Patent Act, s. 28.3 allowing — Not necessary to determine appropriate date among those remaining as Federal Court not erring in concluding that ꞌ784 patent patentably distinct even in light of ‘902 patent application — Skilled person would not have unhesitatingly accepted teachings of ‘902 patent application as true; not part of common knowledge — Appeal dismissed.
This was an appeal from a Federal Court decision allowing the respondent Eli Lilly Canada Inc.’s application for an order under section 6 of the Patented Medicines (Notice of Compliance) Regulations prohibiting a notice of compliance (NOC) from being issued to the appellant.
The impugned patent (Canadian Patent No. 2226784 or the ꞌ784 patent) claims tadalafil and 3-methyl tadalafil (the two compounds) for the treatment of erectile dysfunction (ED). The ꞌ784 patent had a priority date of July 14, 1995, a Canadian filing date of July 11, 1996, and a publication date of February 6, 1997. The two compounds inhibit the production of PDE V, an enzyme that prevents erections. Eli Lilly had previously acquired a patent (ꞌ377 patent) that claimed among other compounds tadalafil, and had a priority date of January 21, 1994 and a Canadian filing date of January 19, 1995. The ꞌ377 patent did not mention ED treatment specifically. On December 22, 1994, Pfizer published its American patent (‘902 patent) application for sildenafil, another PDE V inhibitor. After the priority date of the ꞌ784 patent, a study provided evidence sildenafil could be orally administered for effective treatment of ED. The Federal Court found that this study stood for the broader proposition that a PDE V inhibitor could be administered orally for safe, effective treatment of ED. Following Eli Lilly’s application for an order prohibiting a NOC, the appellant filed a notice of allegation that the ꞌ784 patent was invalid for inutility on the basis of a lack of sound prediction, and obviousness-type double-patenting. With regard to the double-patenting challenge, the Federal Court determined that the relevant date on which the impugned patent was to be assessed was the priority date of the ꞌ377 patent. The Federal Court determined, inter alia, that at the priority date of the ꞌ377 patent, the use of PDE V inhibitors such as tadalafil to treat ED was patentably distinct from the ꞌ377 patent. The Federal Court also rejected the challenge to the utility of the patent.
At issue was whether the Federal Court erred in finding that the impugned patent was not invalid for either obviousness-type double-patenting or a lack of utility.
Held, the appeal should be dismissed.
The Federal Court committed no reversible errors in finding that the impugned patent was not invalid for either obviousness-type double-patenting or a lack of utility.
In essence, the parties were asking whether there is an inventive step from the first patent to the second. The Supreme Court in Whirlpool Corp. v. Camco Inc. indicated that the substance of a double-patenting inquiry—like obviousness—is whether there is “invention” or “ingenuity” in the move from the first patent to the second. This, in execution, requires consideration of the claims of the second patent against the claims of the first patent. The Federal Court erred in referring to the specification when construing the claims of the ꞌ377 patent. The rules of patent construction preclude reference to the specification when the claims are clear. Such reference is also improper if it varies the scope of the claims. In this case, the ꞌ377 patent unambiguously claims tadalafil, without any comment on its use as a PDE V inhibitor. In these circumstances, the Federal Court’s reliance on Apotex Inc. v. Sanofi-Synthelabo Canada Inc. was misplaced. Sanofi-Synthelabo makes it clear that construing the inventive concept is a distinct analysis from claims construction. As there was no ambiguity in the claims here, there could be no recourse to the specification. However, this error was of no consequence, as its effect was to impose a higher burden on the Eli Lilly. The Federal Court came to the correct conclusion with respect to obviousness-type double patenting. There were three possible dates from which obviousness-type double-patenting could be assessed. The first date was the priority date of the ꞌ377 patent. The middle date was the priority date of the ꞌ784 patent, which in this case was after Pfizer’s ‘902 patent application was published. The last possible date was the publication date of the ꞌ784 patent. The publication date of the later patent (the last date) was not the appropriate one. Whirlpool was not the controlling authority on this point. Whirlpool was an obviousness case and did not turn on the correct date for a double-patenting analysis. The publication date of the ꞌ784 patent was not the correct date in light of the principles underlying double-patenting doctrine. Specifically, it would be inappropriate to use any date after the claim date of the second patent (whether in a particular case the claim date is the same as the priority date—as it was here—or the filing date). Using a date after the claim date would mean that a court assessing an obviousness-type double-patenting claim would consider prior art beyond what section 28.3 of the Patent Act allows the Court to consider when assessing classical obviousness. This left the first date, on which the appellant uncontroversially failed, and the middle date, on which there was an evidentiary dispute. It was not necessary to determine the question of which of these dates was the appropriate one. Even if the correct date was the more appellant-friendly middle date, the Federal Court did not err in concluding that the ꞌ784 patent was patentably distinct even in light of the ‘902 patent application. The fact that the Supreme Court criticised the Canadian equivalent to the ‘902 patent application for lack of disclosure of the specific compound claimed (sildenafil) was irrelevant to the broader issue of what the ‘902 patent application taught about PDE V inhibitors as a general class. As such, the skilled person would not have unhesitatingly accepted the teachings of the ‘902 patent application as true; they were not part of the common general knowledge.
STATUTES AND REGULATIONS CITED
Patent Act, R.S.C., 1985, c. P-4, ss. 28.1, 28.3, 55.2(4).
Patented Medicines (Notice of Compliance) Regulations, SOR/93-133, s. 6.
CASES CITED
APPLIED:
Whirlpool Corp. v. Camco Inc., 2000 SCC 67, [2000] 2 S.C.R. 1067; Merck & Co., Inc. v. Pharmascience Inc., 2010 FC 510, 85 C.P.R. (4th) 179; Commissioner of Patents v. Fabwerke Hoechst Aktiengesellschaft Vormals Meister Lucius & Bruning, [1964] S.C.R. 49, (1963), 41 C.P.R. 9.
CONSIDERED:
Apotex Inc. v. Sanofi-Synthelabo Canada Inc., 2008 SCC 61, [2008] 3 S.C.R. 265; General Tire & Rubber Co. v. Firestone Tyre & Rubber Co., [1972] R.P.C. 457 (C.A.); Eli Lilly Canada Inc. v. Novopharm Limited, 2010 FCA 197, [2012] 1 F.C.R. 349.
REFERRED TO:
Teva Canada Ltd. v. Pfizer Canada Inc., 2012 SCC 60, [2012] 3 S.C.R. 625; Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235; Pfizer Canada Inc. v. Canada (Minister of Health), 2006 FCA 214, [2007] 2 F.C.R. 137; Apotex Inc. v. Wellcome Foundation Ltd., 2002 SCC 77, [2002] 4 S.C.R. 153; Abbott Laboratories v. Canada (Minister of Health), 2005 FC 1095, 42 C.P.R. (4th) 20.
AUTHORS CITED
Hughes, R. T. et al. Hughes and Woodley on Patents, 2nd ed. looseleaf (Toronto: LexisNexis, 2005).
Siebrasse, Norman. Sufficient Description: Observations on Canadian patent cases. “Disagreement on Date for Assessing Obviousness-type Double Patenting”, August 14, 2015, online: <http://www.sufficientdescription.com/2015/08/disagreement-on-date-for-assessing.html>.
appEAL from a Federal Court decision (2015 FC 17, 129 C.P.R. (4th) 103) allowing the respondent Eli Lilly Canada Inc.’s application for an order under section 6 of the Patented Medicines (Notice of Compliance) Regulations prohibiting a notice of compliance from being issued to the appellant. Appeal dismissed.
APPEARANCES
Timothy Gilbert, Sana Halwani, Andrew Moeser and Zarya Cynader for appellant.
Jamie Mills and Chantal Saunders for respondents Eli Lilly Canada Inc. and ICOS Corporation.
No one appearing for respondent Minister of Health.
SOLICITORS OF RECORD
Gilbert’s LLP, Toronto, for appellant.
Borden Ladner Gervais LLP, Ottawa, for respondents Eli Lilly Canada Inc. and ICOS Corporation.
Deputy Attorney General of Canada for respondent the Minister of Health.
The following are the reasons for judgment rendered in English by
Rennie J.A.:
I. Introduction
[1] This is an appeal from a decision of the Federal Court per Justice de Montigny (the Judge) dated January 7, 2015 ([Eli Lilly Canada Inc. v. Mylan Pharmaceuticals ULC] 2015 FC 17, 129 C.P.R. (4th) 103). In that decision, the Judge allowed Eli Lilly’s application for an order under section 6 of the Patented Medicines (Notice of Compliance) Regulations, SOR/93-133 (Regulations) prohibiting a notice of compliance (NOC) from being issued to Mylan. Mylan had alleged that Eli Lilly’s patent was invalid on the basis of obviousness-type double-patenting and for lack of utility due to no sound prediction. For the reasons that follow, I would dismiss the appeal.
II. Background
[2] The impugned patent (Canadian Patent No. 2226784 or the ꞌ784 patent) claims tadalafil and 3-methyl tadalafil (the two compounds) for the treatment of erectile dysfunction (ED). The patent comprises 28 claims, a subset of which are disputed in this appeal. Saliently, claim 18 claims the use of the two compounds via oral administration for the treatment of ED; claims 2, 4, 12, 14, and 15 (the remaining claims) are to the general use of the two compounds to treat ED, with no mention of oral administration.
[3] The impugned ꞌ784 patent had a priority date of July 14, 1995, a Canadian filing date of July 11, 1996, and a publication date of February 6, 1997.
[4] Tadalafil and 3-methyl tadalafil are PDE V inhibitors. PDE V is an enzyme that breaks down the chemical cGMP [cyclic guanosine monophosphate] (which causes erections via triggering smooth muscle relaxation) into GMP (which does not). Because PDE V prevents erections, a PDE V inhibitor has the effect of stimulating erections.
[5] Eli Lilly had previously acquired a patent that claimed, among other compounds, tadalafil. This earlier patent (Canadian Patent No. 2181377 or the ꞌ377 patent) had a priority date of January 21, 1994 and a Canadian filing date of January 19, 1995. In the words of the Judge [at paragraph 10], the ꞌ377 patent claimed “novel compounds, including tadalafil, pharmaceutical compositions, and the use of tadalafil in the treatment of various disorders where smooth muscle relaxation was thought to be beneficial, including cardiovascular disorders.” The patent indicated that tadalafil was bioavailable for reducing systemic hypertension. It did not mention ED treatment specifically.
[6] On December 22, 1994 (after the priority date of the ꞌ377 patent, but before the priority date of the impugned ꞌ784 patent) Pfizer’s American patent application WO 1994028902A1 (the ‘902 patent application) for sildenafil (another PDE V inhibitor) was published. Notwithstanding criticism of the equivalent Canadian patent (patent 2163446) by the Supreme Court of Canada in Teva Canada Ltd. v. Pfizer Canada Inc., 2012 SCC 60, [2012] 3 S.C.R. 625 for inadequate disclosure of the particular chemical claimed by the patent, the ‘902 patent application did show that a PDE inhibitor could treat ED. There was evidence before the Judge, however, that the information in the ‘902 patent application was a counterintuitive breakthrough and was initially met with scepticism.
[7] In early 1996 (after the priority date of the ꞌ784 patent, but a few weeks prior to its Canadian filing date) a study (the Boolell study) provided robust evidence, based on clinical trials, that the PDE V inhibitor sildenafil could be orally administered for effective treatment of ED. The Judge found that this study stood for the broader proposition that a PDE V inhibitor could be administered orally for safe, effective treatment of ED.
[8] Pfizer applied for an order under subsection 55.2(4) of the Patent Act, R.S.C., 1985, c. P-4, and section 6 of the Regulations, prohibiting the Minister from granting a NOC to Mylan for a generic version of tadalafil. On December 21, 2012, Mylan filed a notice of allegation alleging that the ꞌ784 patent was invalid for lack of utility and obviousness-type double-patenting.
III. The decision below
[9] Mylan alleged that the claims at issue were invalid for obviousness-type double-patenting, and also that they were invalid for inutility on the basis of a lack of sound prediction.
[10] With regard to the double-patenting challenge, the Judge determined that the relevant date on which the impugned patent was to be assessed was the priority date of the first (ꞌ377) patent (January 21, 1994). He found that the inquiry at the centre of obviousness-type double-patenting was whether the second patent’s claims should have been included in the first patent. He also found that while the double-patenting analysis was to be a comparison of the claims of the two patents, the claims must be read in the context of the full patent; consequently, he interpreted the ꞌ377 patent as claiming tadalafil as a PDE V inhibitor even though its status as a PDE V inhibitor was only made clear in the specification.
[11] On the basis of this framework, the Judge determined that at the priority date of the ꞌ377 patent, the use of PDE V inhibitors, such as the two compounds, to treat ED was patentably distinct from the ꞌ784 patent. Indeed, he noted that this factual finding was uncontroversial in the view of the parties’ witnesses. In the alternative, the Judge found that even at the priority date of the second patent, the ꞌ784 patent was still patentably distinct from the ꞌ377 patent. The ‘902 patent application, which had been published about six months prior to the ꞌ784 patent’s priority date, did not sufficiently advance the common general knowledge held by the person ordinarily skilled in the art (the skilled person) such that the ꞌ784 patent was not patentably distinct.
[12] The Judge also rejected the challenge to the utility of the patent. He concluded that the promise of the patent was the use of tadalafil and 3-methyl tadalafil to treat ED without reference to any particular mode of administration or the absence of toxicity or undue side effects. The Judge found that in light of the experiments taught in the patent’s disclosure, the common general knowledge and the disclosure of the ꞌ377 patent, there was a prima facie reasonable inference of utility for the claimed compounds.
[13] The Judge also found that, in the alternative, even if the promise of the patent includes oral administration it was still soundly predicted. The Judge determined that by combining the ꞌ377 patent’s disclosure that demonstrated that tadalafil was an orally bioavailable PDE V inhibitor with the Boolell study’s teachings that an orally bioavailable PDE V inhibitor could be used to treat ED, there was a prima facie reasonable inference of the utility of the oral administration of tadalafil. Having found that claim 18 was soundly predicted with regard to tadalafil, the Judge determined that it was not necessary to make a similar finding with regard to 3-methyl tadalafil. The basis for this conclusion was that claim 18 was a Markush claim, and that Markush claims to a class of compounds have utility as long as at least one compound within the class has utility.
IV. Positions of the parties
[14] Mylan appeals the findings with regard to both obviousness-type double-patenting and sound prediction, limiting its arguments on appeal to the sound prediction of claim 18 and to obviousness-type double-patenting of the remaining claims.
[15] With regard to double-patenting, Mylan alleges that the Judge erred in law by conducting the analysis as of the priority date of the ꞌ377 patent, and that Whirlpool Corp. v. Camco Inc., 2000 SCC 67, [2000] 2 S.C.R. 1067 (Whirlpool) instead mandates using the publication date of the later (in this case, ꞌ784) patent. Mylan also contends that the Judge committed a palpable and overriding error in finding that, at the ꞌ784 patent’s priority date, the impugned patent was not invalid for obviousness-type double-patenting. Finally, Mylan argues that the nature of the inquiry for obviousness-type double-patenting should be akin to obviousness.
[16] Regarding sound prediction, Mylan argues that no skilled person would draw the inference of oral utility without specific in vivo tests on the two compounds which were claimed. Because there was no testing of 3-methyl tadalafil’s oral bioavailability for any purpose, and no testing of the oral use of either compound to treat ED, claim 18 was not soundly predicted.
[17] Eli Lilly contends that the Judge ultimately reached the correct conclusion on both challenges to the patent.
[18] With respect to double-patenting, Eli Lilly argues that because PDE V inhibition is only mentioned in the patent’s specification rather than its claims, the Judge erred in reading PDE V inhibition into the claims of the ꞌ377 patent. Notwithstanding this alleged error, however, Eli Lilly argues that the Judge used the correct date for the double-patenting analysis and properly concluded that the patent was not invalid on this ground.
[19] Eli Lilly also submits that the obviousness-type double-patenting inquiry cannot be conducted according to the obviousness framework. It argues that to do so would make obviousness-type double-patenting a surrogate of the obviousness doctrine and circumvent the protections of section 28.3 of the Patent Act. Paragraph 28.3(a) excludes from the prior art, which forms the basis of an obviousness challenge, any documents disclosed by the patentee in the year prior to filing.
[20] On the sound prediction issue, Eli Lilly’s position is that the Judge correctly concluded that the ‘902 patent application and the Boolell study provided a factual basis for a sound line of reasoning to predict the utility of claim 18 with regard to the two compounds.
V. Standard of review
[21] The standard of review established in Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235 applies to appeals of NOC applications: Pfizer Canada Inc. v. Canada (Minister of Health), 2006 FCA 214, [2007] 2 F.C.R. 137, at paragraph 15. Findings of fact and mixed fact and law can only be reversed if there is a palpable and overriding error. Findings of law, including extricable questions of law on an issue of mixed fact and law, are reviewed on a correctness standard.
VI. Preliminary observations
[22] Before examining the grounds of invalidity under appeal, it is useful to briefly review three areas of patent law. The first is the distinction between prior art and common general knowledge. The second is the distinction between obviousness and double-patenting. Finally, a recapitulation of the various dates that are engaged by the issues on appeal completes the backdrop.
A. Prior art and common general knowledge
[23] Prior art is the collection of learning in the field of the patent at issue. It comprises any publically available teaching, however obscure or not generally accepted.
[24] The common general knowledge, in contrast, is the “knowledge generally known by persons skilled in the relevant art [skilled persons] at the relevant time”: Apotex Inc. v. Sanofi-Synthelabo Canada Inc., 2008 SCC 61, [2008] 3 S.C.R. 265 (Sanofi-Synthelabo), at paragraph 37. Unlike the prior art, which is a broad category encompassing all previously disclosed information in the field, a piece of information only migrates into the common general knowledge if a skilled person would become aware of it and accept it as “a good basis for further action”: General Tire & Rubber Co. v. Firestone Tyre & Rubber Co., [1972] R.P.C. 457 (C.A.), at page 483.
[25] Prior art is used for specific purposes in patent law, such as to found an allegation that prior art anticipated the invention or rendered it obvious. The common general knowledge informs the way in which the claims and specifications are read, because it is to the skilled person that the patent is addressed. Any inquiry in patent law that is performed from the perspective of a skilled person will import the common general knowledge.
B. Obviousness-type double-patenting
[26] The double-patenting doctrine holds that a claim is invalid if it constitutes patenting of an invention that has already been claimed in a previous patent. It is aimed at the problem of evergreening; extending the monopoly that was granted on the first patent by filing a new patent that does not offer a new invention to the public. As such, the doctrine of double-patenting prevents a patentee from violating the bargain at the heart of the patent system.
[27] In Whirlpool, the Supreme Court of Canada recognized two types of double-patenting. The first is “same-invention” double-patenting, which occurs when the claims of the second patent are outright “identical or coterminous” to the first. This is not alleged in this case. The second is “obviousness-type” double-patenting, which occurs when the second patent is not identical to the first, but is nonetheless not “patentably distinct” from the first.
[28] Invalidity on the basis of obviousness-type double-patenting is not the same as invalidity on the basis of obviousness. Obviousness is directed at the question of whether an “invention” (in the legal sense) exists at all. Obviousness-type double-patenting has a different policy justification; the prevention of evergreening an existing patent through what would otherwise be a valid patent but is, in effect, an extension of the patent that has already been granted: Merck & Co., Inc. v. Pharmascience Inc., 2010 FC 510, 85 C.P.R. (4th) 179, at paragraph 124. Some of the differences are contested issues in this appeal, but a few uncontroversial distinctions between the two variants of the doctrine can be noted.
[29] In an obviousness challenge, any piece of prior art, including a collection of works, can be cited as rendering the impugned patent obvious and therefore not patentable: Sanofi-Synthelabo, at paragraphs 67–71. By contrast, in an obviousness-type double-patenting challenge, only the earlier patent can be cited as rendering the impugned patent not patentably distinct; any other prior art is only relevant insofar as it contributes to the common general knowledge of the skilled person.
[30] Finally, in an obviousness challenge, paragraph 28.3(a) of the Patent Act provides that any information disclosed by the patentee within a year prior to the filing cannot be cited as prior art that renders the patent obvious. This effectively gives the patentee a one-year grace period before filing in which it can make disclosures without worrying that those disclosures will be the basis of an obviousness attack. Double-patenting is not subject to paragraph 28.3(a), which is what allows the earlier patent to be cited if it was published within a year of the filing date of the impugned patent.
C. The relevant dates
[31] Borrowing in large measure from Hughes and Woodley on Patents: The Honourable Mr. Justice Roger T. Hughes et al., 2nd ed. looseleaf (Toronto: LexisNexis, 2005), it is useful to revisit some of the key terms on which this appeal turns:
(a) The filing date is the date on which the patent application is filed with the Canadian Patent Office.
(b) The priority date is the date on which an earlier patent application disclosing the same invention was filed by the patentee – either in Canada or (much more often) in a country that is party to a relevant patent treaty or convention to which Canada is also a party. A priority date only applies at the request of the patentee, and the patent must be filed in Canada within 12 months of the priority date.
(c) The claim date is the priority date if there is one; otherwise the claim date is the filing date.
(The above concepts (a and b) are governed by section 28.1 of the Patent Act. Though the various iterations of the Patent Act complicate matters slightly, a general statement can be made that novelty and obviousness are both assessed as of the claim date.)
(d) The publication date is the date on which the application for a patent is first open to the public for inspection. The prospective patentee has the right to delay this up to 18 months after the claim date. This is governed by section 10 of the Patent Act. The patent becomes citable prior art as of the publication date; the publication date is also the date for claims construction.
VII. Analysis
[32] One issue in dispute between the parties is whether the substance of the analysis of obviousness-type double-patenting is the same as that for obviousness and what date is to be used for conducting the analysis.
[33] Mylan submits that the Court should apply the well-settled test for obviousness, with the earlier patent taking the place of the prior art. Under this construct, the question for the Court is whether a skilled person would consider the impugned patent obvious in light of the earlier patent, with the earlier patent taking the place of the prior art generally, or whether there was an inventive step that makes the impugned patent non-obvious. This approach differs from straightforward obviousness in terms of what prior art is considered, but the central inquiry is the same.
[34] Eli Lilly argues that this is incorrect, and that the inquiry is instead whether the second invention constitutes an improper extension of the original patent. For his part, the Judge framed the question as whether what was claimed in the second patent could or should have been included in the first patent.
[35] I do not see a substantive distinction between the two approaches. They are not inconsistent and are, in effect, reformulations of the same inquiry. When one says “the second patent is an impermissible extension of the first” or “the claims of the second patent should have been included in the first”, those statements, in essence, ask whether there is an inventive step from the first patent to the second.
[36] The Supreme Court of Canada in Whirlpool indicated that the substance of a double-patenting inquiry—like obviousness—is whether there is “invention” or “ingenuity” in the move from the first patent to the second: Whirlpool, at paragraphs 66 and 67. Moreover, because the doctrine exists to prevent the evergreening of patents with uninventive additions, examination of whether the changes in the second patent are or are not inventive is directly linked to the policy considerations that underlie the doctrine. Finally, while not dispositive, the use of the label “obviousness” for this type of double-patenting indicates that a similar analytical process is appropriate.
[37] This, in execution, requires consideration of the claims of the second patent against the claims of the first patent. The distinction from an obviousness inquiry is nuanced, but doctrinally important. As noted by Hughes J. in Merck & Co., Inc. v. Pharmascience Inc., at paragraph 124:
What is important to keep in mind is that the exercise required in the inquiry as to whether there is double patenting is that the claims of the earlier patent owned by the same patentee as the latter must be compared with the claims of the latter to see if they are “identical or co-terminus”, or whether the latter is “obvious” in view of the former. Therefore, the exercise is somewhat different than that of dealing with obviousness of a patent having regard to the art that would have been known to a person skilled in the art as of the relevant time. The exercise respecting double patenting is to present the notional person skilled in the art with the claims of the first patent and inquire whether what is claimed in the second patent was “identical or co-terminus” with the first or would have been obvious in light of the earlier patent. The inquiry must not bother with any inquiry as to whether the earlier patent would have come to the attention of the notional person skilled in the art. Nor does the inquiry extend to the validity or otherwise of the claims of the earlier patent. Nor does the inquiry extend to “prior art” beyond the earlier patent, as Binnie J. wrote at paragraph 67 of Whirlpool, the inquiry is whether a second patent can be justified unless the claims exhibit “novelty or ingenuity” over the first patent. [Emphasis in original.]
[38] Although formulated in slightly different language, this analysis may be traced back to, and is consistent with, Commissioner of Patents v. Fabwerke Hoechst Aktiengesellschaft Vormals Meister Lucius & Bruning, [1964] S.C.R. 49 and its focus on whether the second patent contains a new or inventive element or process beyond that claimed in the first.
[39] In my view, the Judge erred in referring to the specification when construing the claims of the ꞌ377 patent. The rules of patent construction preclude reference to the specification when the claims are clear, and also improper if it varies the scope of the claims: Hughes and Woodley on Patents, at page 312:
In construing a patent, the claims are the starting point. The claims alone define the statutory monopoly and the patentee has a statutory duty to state, in the claims, what the invention is for which protection is sought. In construing the claims … recourse to the rest of the specifications is: (1) permissible to assist in understanding the terms used in the claims; (2) unnecessary where the words are plain and unambiguous; and (3) improper to vary the scope or ambit of the claims.
[40] In this case, the ꞌ377 patent unambiguously claims the compound tadalafil, without any comment on its use as a PDE V inhibitor. In these circumstances, the Judge’s reliance on Sanofi-Synthelabo, at paragraph 77 is misplaced. Rothstein J.’s comments regarding the patent specification were in the specific context of construing the inventive concept of a selection of patent for the purposes of an obviousness analysis, not in the context of claims construction. Sanofi-Synthelabo makes it clear that construing the inventive concept is a distinct analysis from claims construction.
[41] Paragraphs 76 and 77 of Sanofi-Synthelabo are helpful on this point. First, at paragraph 76, Rothstein J. construed the claims of the patent and found that they constitute “the dextro-rotatory isomer”, a compound. Rothstein J. then moved to the obviousness inquiry and the inventive concept of this claim. It was only at this latter stage of the analysis that Rothstein J. referenced the specification, observing that “[a] bare chemical formula in a patent claim may not be sufficient to determine its inventiveness.”
[42] Moreover, when Rothstein J. assessed obviousness-type double patenting later in the judgment, he did not consider the patent’s specification in construing the claims. At paragraph 108, he stated:
Apotex argues that the focus in a double patenting challenge is on the claims of the two patents rather than on the disclosure. I agree. In Whirlpool, Binnie J. stated, at para. 63:
It is clear that the prohibition against double patenting involves a comparison of the claims rather than the disclosure, because it is the claims that define the monopoly.
[43] To conclude, as there was no ambiguity in the claims here, there could be no recourse to the specification. However, this error is one of no consequence, as its effect was to impose a higher burden on Eli Lilly. The Judge came to the correct conclusion with respect to obviousness-type double patenting.
[44] Having settled on the essential methodology, I turn to the question as to the date at which the inquiry is to be conducted. The consequences associated with the selection of any date are considered by Professor N. Siebrasse in Sufficient Description: Observations on Canadian patent cases. “Disagreement on Date for Assessing Obviousness-type Double Patenting” (14 August 2015), online: http://www.sufficientdescription.com/2015/08/disagreement-on-date-for-assessing.html, where he concludes that the law is unsettled and that the point is a difficult one, a proposition with which I agree.
[45] There are three possible dates from which obviousness-type double-patenting may be assessed. The first date is the priority date of the first patent. The middle date is the priority date of the second patent, which in this case is after Pfizer’s ‘902 patent application was published. The last possible date is the publication date of the second patent, by which point the Boolell study was also publicly available and well-known. The timeline, and choices, is reflected below:
[46] The parties broadly agree on the consequences of each date being the appropriate one. If the first date (the priority date of the ꞌ377 patent) is correct, the respondent will almost surely succeed; Mylan did not argue in its factum or oral argument that double-patenting could be established based on the ꞌ377 patent alone. If the last date is correct (the publication date of the ꞌ784 patent), the appellant will almost surely succeed, as the Boolell study renders the contributions of the ꞌ784 patent non-inventive. Eli Lilly, in oral argument, did not seriously contend otherwise. If the middle date is correct (the ꞌ784 patent priority date), then there is an evidentiary dispute as to whether the Judge erred in finding no double-patenting.
[47] I am convinced that the publication date of the later patent (the last date) is not the appropriate one. Contrary to the appellant’s submissions, I do not read Whirlpool as the controlling authority on this point.
[48] As noted by the Judge, Whirlpool was an obviousness case and did not turn on the correct date for a double-patenting analysis. The Court found the expert evidence before the Judge insufficient to support a finding of obviousness-type double patenting. Further, the discussion arose in the context of the date for construction of claims under the old act and the Court settled on the publication date. Thus, while strictly obiter, Binnie J.’s observation at paragraph 67 that the inquiry is whether a second patent can be justified unless the claims exhibit “novelty or ingenuity” over the first patent, places the focus on the publication date of the first patent. At the risk of repetition, I note as well the interpretation of Whirlpool by Hughes J. in Merck & Co., Inc. v. Pharmascience Inc., at paragraph 124, subsequent to Whirlpool, that:
…The exercise respecting double patenting is to present the notional person skilled in the art with the claims in the first patent and inquire whether what is claimed in the second patent was “identical or co-terminus” with the first or would have been obvious in light of the earlier patent…. Nor does the inquiry extend to “prior art” beyond the earlier patent.…
[49] The next question is therefore whether the publication date of the later patent (the ꞌ784 patent) is nonetheless the correct date in light of the principles underlying double-patenting doctrine. I conclude that it is not. Specifically, it would be inappropriate to use any date after the claim date of the second patent (whether in a particular case the claim date is the same as the priority date—as it is here—or the filing date).
[50] As Professor Siebrasse notes, using a date after the claim date would also mean that a court assessing an obviousness-type double-patenting claim would consider prior art beyond what section 28.3 allows the Court to consider when assessing classical obviousness. Contrast this with the fact that the double-patenting doctrine allows a challenger to circumvent the one year grace period in paragraph 28.3(a). That circumvention is acceptable because a consideration of the patentee’s prior patent documents is precisely what the doctrine mandates the Court to undertake. However, there is no equivalent reason to allow a challenger alleging obviousness-type double-patenting to point to prior art after the claim date, while not allowing the same to a challenger alleging obviousness.
[51] This eliminates the possibility of the third date—the ꞌ784 patent publication date—being correct. As such, the Boolell study cannot be considered by the skilled person when assessing double-patenting. The scenario in which the appellant Mylan succeeds uncontroversially should be eliminated.
[52] This leaves the first date, on which the appellant Mylan uncontroversially fails, and the middle date, on which there is an evidentiary dispute. It is not necessary to determine the question of which of these dates is the appropriate one. This is because, on the facts of this case, even if the correct date is the more appellant-friendly middle date, I find that the Judge did not err in finding that there was no double-patenting. As I will explain, there was no change in the common general knowledge between the first date and the middle date.
[53] Specifically, I find that the Judge committed no reviewable error in concluding that the ꞌ784 patent was patentably distinct even in light of the ‘902 patent application. I agree with the appellant that the fact that the Supreme Court of Canada criticised the Canadian equivalent to the ‘902 patent application for lack of disclosure of the specific compound claimed (sildenafil) is irrelevant to the broader issue of what the ‘902 patent application taught about PDE V inhibitors as a general class. However, the Judge accepted evidence that the ‘902 patent application’s teachings about the use of PDE V inhibitors to treat ED were counterintuitive and met with initial skepticism. As such, the skilled person would not have unhesitatingly accepted the teachings of the ‘902 patent application as true; they were not part of the common general knowledge. The Judge correspondingly did not commit a palpable and overriding error in concluding that, even as the common general knowledge stood after the ‘902 patent application, the ꞌ784 patent was patentably distinct over the ꞌ377 patent.
A. Utility of claim 18
[54] Mylan also challenges claim 18 on the basis that it lacks utility. Specifically, Mylan alleges that while the oral administration of tadalafil to treat ED may have been soundly predicted, the oral administration of 3-methyl tadalafil was not. This is because while the ꞌ377 patent (demonstrating oral bioavailability and PDE V inhibition) could be combined with the Boolell study (teaching that an orally bioavailable PDE V inhibitor could treat ED) in the case of tadalafil, the ꞌ377 patent only showed such oral bioavailability for tadalafil and not 3-methyl tadalafil.
[55] Utility need not be demonstrated at the time a patent is filed; it is enough for the claimed utility to be soundly predicted by the filing date. Sound prediction has three requirements, a factual basis for the prediction, an articulable and sound line of reasoning and proper disclosure: Apotex Inc. v. Wellcome Foundation Ltd., 2002 SCC 77, [2002] 4 S.C.R. 153, at paragraph 70. If all of these are present, a “prima facie reasonable inference of utility” can be reached: Eli Lilly Canada Inc. v. Novopharm Limited, 2010 FCA 197, [2012] 1 F.C.R. 349, at paragraph 85.
[56] After finding that claim 18 with regard to tadalafil was soundly predicted, the Judge did not consider whether the use of 3-methyl tadalafil to treat ED orally was also soundly predicted. He reasoned that the status of the claim as a Markush claim rendered it unnecessary to determine whether each compound claimed would work. However, a Markush claim requires that each compound in the claimed class, not merely one of the compounds, have utility: Abbott Laboratories v. Canada (Minister of Health), 2005 FC 1095, 42 C.P.R. (4th) 20, at paragraphs 23–27. Assuming, without deciding, that claim 18 was indeed a Markush claim, the Judge thus erred, but it is an error of no consequence.
[57] There was evidence before the Judge which indicated that 3-methyl tadalafil was soundly predicted. In particular, the reply affidavit of Dr. Brock provided a factual basis and a sound line of reasoning sufficient to ground a sound prediction. Dr. Brock noted both the Boolell study’s demonstration that PDE V inhibitors could be used to treat ED and also the broad variety of tetracyclic derivatives that were identified as orally bioavailable PDE V inhibitors in the ꞌ377 patent. The latter fills in, via a general statement, the factual premise (that 3-methyl tadalafil as a tetracyclic derivative was an orally bioavailable PDE V inhibitor) which Mylan alleges is missing. Dr. Brock further understood that a skilled person would be able to infer the oral administration of both compounds to treat ED. There was no guarantee of success, but the doctrine of sound prediction does not require guarantees. A prima facie reasonable inference of success existed.
[58] Finally, I note that even if a different view were taken as to the validity of claim 18, it would have no effect on the disposition of this appeal. Claim 18 is a dependent claim comprising a subset of the subject matter claimed by the remaining claims. Because, as I have determined above, the remaining claims survive the double-patenting challenge, claim 18 fences off no monopoly that is not entirely subsumed within their scope.
VIII. Conclusion
[59] In conclusion, the Judge committed no reversible errors in finding that the impugned patent was not invalid for either obviousness-type double-patenting or a lack of utility. Accordingly, I would dismiss the appeal with costs.
Dawson J.A.: I agree.
Trudel J.A.: I agree.