Mylan Pharmaceuticals v. Pfizer et al. 2012 FCA 103

Summary:  The Federal Court of Appeal dismissed an appeal of a decision of Justice Hughes that granted an order prohibiting the Minister of Health from issuing a Notice of Compliance (“NOC”) to Mylan Pharmaceuticals ULC (“Mylan”) with respect to its generic version of Pfizer Canada Inc. (“Pfizer”)’s Aricept®, used for the treatment of Alzheimer’s type dementia.  Canadian Patent No. 1,338,808 (“‘808 Patent”) includes a claim for donepezil, the active ingredient in Aricept®.  The Federal Court of Appeal agreed with Justice Hughes that the ‘808 Patent only promised therapeutic utility and that references in the description to lower toxicity and greater duration were mere “advantages”.

Analysis:  There were two primary issues on appeal.  The first was whether Justice Hughes had erred in finding that he could not consider Mylan’s argument that the ‘808 Patent should be invalidated based on wrong or misleading data.  The second was whether Justice Hughes had erred when he found that safety, or lack of toxicity, was not part of the promise of the patent in rejecting Mylan’s argument that the ‘808 Patent was invalid for lack of sound prediction.  Both are fairly standard allegations at first blush but upon further reflection, are definitely food for thought.

On the first issue, Justice Hughes relied on the well-established principle that proceedings under the Patented Medicines (NOC) Regulations (“PM(NOC) Regulations”) are framed by the Notice of Allegation (“NOA”)[i] to dismiss the allegation of wrong and misleading data as a ground of invalidity.  Mylan argued persuasively in both proceedings that it had no basis for including “wrong and misleading data” as a ground for invalidating the ‘808 Patent when it filed its NOA since it did not yet have access to the data in question.  Mylan could not have known that the allegation was even available early enough to include it in its NOA.

While strict adherence to the NOA is meant to ensure procedural fairness for the applicant/first person, generics would argue that it results in procedural unfairness for the respondent/second person.  While he appreciated that Mylan could not know that there were errors in the data reported in the ‘808 Patent without access to the raw data and experimental results, Justice Hughes decided that a rule is a rule.  The Federal Court of Appeal agreed or at least, upheld that decision (paras. 34, 35).

Despite having found that the allegation was beyond the scope of the NOA, both Justice Hughes and the Federal Court of Appeal considered the data that Mylan alleged was wrong and misleading.  While discrepancies in the data disclosed in the ‘808 Patent were acknowledged, the Courts found that these were merely bona fide reporting errors, which were not material to the conclusions drawn from those experiments regarding donepezil’s utility (para. 46).  It remains to be seen whether material errors in a patent would persuade the Court to soften the otherwise impenetrable boundaries of the NOA.  In this case, Mylan was only arguing no sound prediction – there was no allegation regarding insufficiency or wilfully misleading.

On the issue of whether the ‘808 Patent was invalid for lack of sound prediction, the decision turned on the Court’s characterization of the promise of the patent.  Justice Hughes held, and the Federal Court of Appeal later upheld, that the promised utility of the ‘808 Patent was that a new class of compounds (including donepezil) was effective in the treatment of Alzheimer’s based on their acetyl-cholinesterase inhibitor function.  While the Federal Court of Appeal acknowledged there were “some references … made in the ‘808 Patent to potential toxicity and efficacy benefits of donepezil, and to its potential advantages over prior art compounds”, they agreed with Justice Hughes that these were merely “potential advantages” rather than promises, per se.  Both Courts were concerned about the zealous lawyer, who tries to turn such potential advantages into promises (paras. 57, 59, 61).  Since a lack of toxicity was construed not to be part of what was promised in the ‘808 Patent, it could not form the basis of an allegation that the ‘808 Patent was not soundly predicted at the time it was filed.

Practice Points:  The Court’s decision on the propriety of allegations that go beyond the scope of the NOA, whether or not those allegations were available at the time the NOA was filed, is consistent with previous decisions under the PM(NOC) Regulations.  Generics may want to consider including an allegation in the NOA relating to improper data, and its various consequences to utility, sufficiency and wilfully misleading as a matter of course, whether or not grounded on any public information, in the event a factual basis arises in the future.  This presents an interesting strategic issue given that the factual allegations in a NOA are presumed to be true, and there may be consequences to a generic asserting incorrect facts.

An underlying issue regarding construction of the promise of the patent was the lower Court’s substantial criticism of Mylan’s expert evidence.  The Court found Mylan’s lawyers basically suggested the promise to its experts.  As such, the Court was left with Pfizer’s experts to instruct them as to how a person skilled in the art would read the promise (paras. 57-61).  In addition, the Federal Court of Appeal (rightly or wrongly) focused on the language of the relevant claim: “claim 18 of the ‘808 Patent simply refers to a ‘therapeutical composition for treating senile dementia…’ without any further promise relating to side effects or toxicity” (para. 56).  This decision may assist patentees in encouraging the Courts to construe “lower” promises that are easier to soundly predict.