An Emerging Framework for Construction of the Promise of the Patent

Astrazeneca Canada Inc. v. Mylan Pharmaceuticals ULC / anastrozole (NOC) 2011 FC 1023 Rennie J

Construction of the promise of the patent is currently a hotly contested area of Canadian patent law, as a valuable and otherwise valid patent may be held invalid for promising too much. Despite this, there is little jurisprudence on how to construe that promise, as Hughes J noted in GlaxoSmithKline / rosiglitazone, 2011 FC 239 at [83]. Rennie J’s debut decision in substantive patent law marks a significant advance in developing a coherent jurisprudence on this point, that is consistent with general principles of claim construction and statutory interpretation. I would say that Rennie J’s decision is a large step in the right direction, except that my own view, expressed in a number of previous posts, is that the entire false promise doctrine is misguided. But Canadian law, and FC judges, are stuck with this doctrine until a higher court rules otherwise, and Rennie J’s opinion is certainly a most welcome contrast to the very disappointing analysis of the FCA in Pfizer v Apotex / latanoprost (NOC) 2011 FCA 236, that I blogged about two weeks ago.

The claims of ‘420 patent at issue were a compound claim for anastrozole (claim 13), which is “an effective and well-tolerated treatment for patients with estrogen dependent breast cancer” [24], and a claim for the use of anastrozole as aromatase inhibitor (claim 15). Inhibition of aromatase is the mechanism by which anastrozole functions to combat estrogen dependent breast cancer. The central question was as to the promise of the patent. The patentee, Astrazeneca, argued that the patent promised only aromatase inhibition [92]. Mylan argued that it promised aromatase inhibition, plus therapeutic utility and reduced side effects as compared with the prior art compound aminoglutethimide [93]. Rennie J ultimately agreed with the patentee [145], and consequently held the patent to be valid, but he added that if Mylan’s position were correct, the patent would be invalid for lack of utility [191].

Before turning to Rennie J’s decision, I can’t help but reiterate my objections to the false promise doctrine. As this decision illustrates, the consequence of the doctrine is that the validity of the patent turns not only on the construction of the claims, but on the construction of the disclosure as well. This debate often involves “splitting hairs” as Hughes J remarked in his rosiglitazone decision at [107]. A careful construction of the claims is reasonable, as it is the function of the claims to define the scope of the monopoly, and so the drafter can be expected to attend to the implications of every word – though even there the courts have cautioned against “meticulous verbal analysis” (Catnic [1982] RPC 183, 243). It is less reasonable to apply the same scrutiny to the much longer description, which is primarily intended to disclose information, rather than set precise boundaries on the scope of the invention. This is particularly problematic because neither European law (including modern English law), nor US law, has a similar doctrine. The Canadian false promise doctrine is a trap for the drafter who is preparing an application for prosecution in multiple jurisdictions. The patent at issue in this case, for example, claimed priority from a UK application which issued in Europe as EP0296749, with exactly the same wording in the three crucial paragraphs. The doctrine is also flawed at the level of principle. In the early days of the English patent system, the grant of a patent was discretionary, and the patentee would come to the Queen as a supplicant, extolling the virtues of his invention in an effort to persuade the monarch to grant patent protection. In that context, where the Queen was entitled to refuse to grant a patent, though doing so would not contravene the Statute of Monopolies, it was reasonable to hold the patentee to the promise which induced the Queen to exercise her discretion. In the modern system, in which it has long been established that the Commissioner has no discretion to refuse a patent which satisfies the requirements of the Act (Havard Mouse 2002 SCC 76 [144]), the doctrine is no more than an anachronism.

With that reminder of the fundamentally problematic nature of the doctrine, consider Rennie J’s approach to the construction of the promise of the patent, which I will argue is fundamentally sound, consistent with general principles of interpretation, and preferable to the approach taken by the FCA in Pfizer v Apotex / latanoprost.

The modern approach to the interpretation requires that any legally authoritative document must be read in light of text, context and purpose: Rizzo Shoes [1998] 1 S.C.R. 27, at [21]; Canada Trustco 2005 SCC 54 at [10]; Biolyse Pharma 2003 FCA 180 at [13]. (For an extended discussion of the purposive approach to claims construction, see my article on The Essential Elements Doctrine in Patent Infringement, (2011) 22 I.P.J. 223, available here in draft form.) Rennie J began with the overarching principle that “”[w]here a reasonable reading of the patent specification can be read to protect a good invention, the Court should give effect to that construction” [88]. This is a fundamental principle of purposive claim construction that the SCC has frequently emphasized (see eg Consolboard [1981] 1 SCR 504 at 521; Wellcome / AZT 2002 SCC 77 at [92]), which recognizes that the purpose of the patentee is to claim that which he has invented and disclosed. Consistency requires that it apply equally to construction of the promise of the patent. As specific interpretive aids, Rennie J looked to the expert evidence [100]-[105], [123]-[125], scientific context [110]-[113], and the case-law relating to similar patents [114]-[118]. This recognizes the importance of context. However, he specifically rejected at [108] the view that the experts’ understanding of the scientific context is determinative, and emphasized instead the primacy of the language of the specification itself: “[w]hile I have relied on the expert evidence, the jurisprudence, and the language of similar patents, the promise must ultimately be grounded in the language of the patent specification” [131]. This is consistent with the approach to interpretation generally: “[w]hen the words of a provision are precise and unequivocal, the ordinary meaning of the words play a dominant role in the interpretive process” Canada Trustco ibid. Failure to appreciate this point was, in my view, the crucial flaw in the FCA’s reasoning in Pfizer v Apotex / latanoprost.

Apart from this framework, Rennie J’s interpretation of the commonly used term “object” of the invention is of general importance. He held that the plain meaning of that term is a goal or purpose [132], and he drew a distinction between goals or objective and promises, saying “the inventors’ objectives or goals cannot elevate the promise of the patent where the language of the patent does not support that promise” [126] (see also [139]). Consequently, the statement of the object of the invention that is commonly found near the outset of the description will not be determinative of the promise of the patent.

In another general point, Rennie J recognized that the promise of the patent must be construed in light of the accepted principle that in respect of a compound intended for therapeutical use, utility for patentability purposes need not rise to the level necessary for regulatory approval. This principle would be undermined if the patentee’s hope, expressed in the patent, that its compound would ultimately prove therapeutically useful were to be elevated to a promise. As Rennie J put it, a skilled person “would understand that clinical trials would be required to determine whether the compounds would be effective treatments for breast cancer, and that the compounds might fail in the course of development” [102].

While Rennie J’s approach to construction of the promise is sound, this road is nonetheless problematic. In particular, his discussion at [114]-[117] of the jurisprudence and identification of the type of language that does support a finding of therapeutical utility, will, if generally accepted, improve predictability of the construction of the promise of the patent in the litigation context. But it might also decrease predictability in the drafting context; any certainty that emerges regarding the meaning of particular terms will be a surprise to patentees who filed their applications several years ago, and it remains to be seen whether the interpretation of particular phrases developed in this decision, or this year, will be respected in judicial decisions ten or twenty years from now. More importantly, the development of an elaborate body of law, unique to Canada, will drive an ever greater wedge between Canadian law on the one hand, and European and US law on the other. In principle patentees should be careful to tailor the disclosure in the Canadian application to Canadian law, but in practice it seems likely that patentees with a world-wide patenting strategy will often be taken by surprise at the time of Canadian litigation.

As a final aside, Rennie J systematically referred to prior pharmaceutical cases by the name of the patentee and the drug, as in “Eli Lilly Olanzapine”, and “Pfizer donepezil”. This convention, which I use in this blog, is a convenient way of distinguishing between the multiple “Eli Lilly v Apotex” decisions, and I hope this marks a judicial trend.