Pfizer / Latanoprost Distinguished on Interpreting the Promise of the Patent

Mylan Pharmaceuticals ULC v. Pfizer Canada Inc. / donepezil (NOC) 2012 FCA 103 Mainville JA: Sharlow, Gauthier JJA aff’g 2011 FC 547 Hughes J (ARICEPT, Pat No 1,338,808)

The FCA has affirmed the decision of Hughes J finding that Pfizer’s patent for ARICEPT (donepezil) is not invalid for lack of sound prediction. While the FCA decision does not change the law significantly, there are some interesting hints regarding the false promise doctrine. In particular, the FCA distinguished its own very problematic decision in Pfizer / latanoprost 2011 FCA 236.

A central issue in the case was the interpretation of the promise of the patent, which Hughes J had interpreted as promising a new class of compound, including donepezil, which is effective for the treatment of Alzheimer’s [FC 232]. Mylan, before Hughes J and on appeal, had argued that the patent promised that donepezil would have better toxicity and efficacy than prior art compounds [FCA 49]. The question was crucial because of the interplay with the requirement established in Lilly / raloxifene 2009 FCA 97 [15], that when utility is based on sound prediction, as it was in this case [FC 35], the factual basis for that prediction must be found in the patent itself [FC 226]. A more demanding promise would be less likely to be established by the evidence disclosed in the patent.

Hughes J accepted Pfizer’s interpretation of the promise, and held that a sound prediction of the promised utility had been established. Once Hughes J had determined the promise of the patent, the case turned mostly on the facts, and, as I noted in my post on Hughes J’s decision, and the most interesting aspect of Hughes J’s decision was his approach to the construction of the patent. The question is whether the promise of the patent is a question of fact, in the sense of that the promise is to be determined by expert evidence as to what the posita would understand the promise to be, or rather a question of mixed question of fact and law, in which legal arguments and principles are needed to supplement the expert evidence. As I noted in my earlier post, “Hughes J comes down firmly in the latter camp,” and in this decision the FCA has affirmed that Hughes J reached his conclusion as to the promise of the patent “by applying the correct test.” [FCA 52, approving FC 212-18]. Note that the FCA has reaffirmed that “[t]he promise of a patent is a question of law reviewable on appeal on a standard of correctness,” but the court went on to point out that “generally, the construction of the promise is an exercise that requires the assistance of expert evidence, as the promise should be properly defined, within the context of the patent as a whole, through the eyes of a person skilled in the art or science to which the invention” [FCA 48]. I suggest that given this caveat, it is more accurate to describe it as a mixed question of fact and law, though nothing turns on the label.

Perhaps of more interest, the FCA distinguished its decision in Pfizer / latanoprost 2011 FCA 236 as turning on its “unique circumstance” [56]. At one level this says nothing more than that the promise of a patent turns on its own particular facts; and indeed, perhaps nothing more should be read into this. But it could also be an indication that the FCA intends to pull back from the very aggressive interpretation of the promise of the patent that it indulged in Pfizer / latanoprost.

On an entirely distinct point, Mylan had also argued that Hughes J had erred in not considering certain errors in the specification. The FCA affirmed Hughes J, noting that “[t]he minor bona fide data reporting errors do not materially change the results reported in the patent, nor do they affect the inference which a person skilled in the art or science to which the invention pertains would reasonably draw from these reported results” [46].