Sunday, February 10, 2013

Be Sure to Claim the Commercial Compound

Pfizer Canada Inc v Pharmascience Inc / pregabalin (NOC) 2013 FC 120 Hughes J
            2,255,652 – LYRICA

Hughes J’s pregabalin decision is noteworthy for a curiously lengthy discussion of the law related to disclosure of utility. The discussion spanned 63 paragraphs [96 -159] and 33 pages of the 86 page decision. What is curious about the discussion is that it was entirely irrelevant to the case at hand! Hughes J remarked at the outset that “The manner in which our Courts have dealt with the matter of sound prediction has appeared to cause some to raise concerns, in Canada and elsewhere, as to how the subject is treated” [96], and he has apparently decided to take this judgment as an opportunity to set out his views. My next post will discuss those views. This post focuses on the issues raised in the case. The bottom line is straightforward: make sure to claim the commercial compound for the primary use.

Pregabalin is a known compound which had previously been used to treat central nervous system disorders [31]. It is the S enantiomer of a chiral compound [42]. The inventive concept of the patent was that pregabalin is also useful to treat pain [89]. The sole claim at issue was claim 3 of the ‘652 patent. Simplified to remove dependencies and chemical names the claim reads [61]:

For use in treating pain, in a mammal, a therapeutically effective amount of pregabalin or its racemate.”

The live issues were overbreadth, lack of utility, based on both lack of sound prediction and lack of demonstrated utility, and obviousness [21]. The basic arguments made by Pharmascience, which were both accepted by Hughes J, were that (1) pregabalin does not treat all types of pain; and (2) the racemate is not useful. Consequently, Hughes J found the claim invalid for overbreadth and lack of utility, though the obviousness attack failed. Note that Pharmascience did not dispute that pregabalin had utility for treating some types of pain.

The first argument, that pregabalin does not treat all types of pain, turned on claim construction. The inventor’s own evidence established that “[pregabalin] doesn’t block acute pain” [92], [174]. Hughes J construed the patent as encompassing acute pain [82]. This was based primarily on the disclosure. While the disclosure emphasized the use for treating chronic pain, it also expressly specified one type of acute pain [72], and used broadening terms, such as ‘not limited to” when listing the kinds of pain which could be treated. Once the broad meaning of “pain” was established, the conclusion that the claimed invention lacked utility followed directly, given the evidence that pregabalin was not useful for acute pain.

Note that this conclusion is not based on lack of sound prediction; it is based on demonstrated (or conceded) inutility. The controversial question of whether the factual basis for a sound prediction of utility must be disclosed in the patent does not arise, since lack of actual utility was established on direct evidence. Similarly, the question of whether evidence of utility must be referenced in the patent (a part of Hughes J’s analysis, to be discussed in my next post), does not arise.

Nor does this conclusion turn on the promise of the patent. The false promise doctrine is controversial when the patent is held to a higher standard of utility than the minimum required to support a valid patent, based on a promise stated only in the disclosure (a fuller explanation can be found in Part 2 of my forthcoming article on the doctrine, available here). In the case of a use patent, such as this, the utility necessary to support the patent must be found in the claim. Even in the case of use patents, some courts may invoke the false promise doctrine to look to the disclosure for additional statements of utility, but that is not necessary, and Hughes J did not do so here. His analysis is consistent with Zinn J’s holding in Fournier / fenofibrate (NOC) 2012 FC 741, blogged here, that in the case of use claims, the promised utility should be found in the claims, and not disclosure.

The second argument, that the racemate was not useful for treating pain, turned entirely on the evidence. There was no evidence disclosed or even referenced in the patent relating to the use of the racemate for treating pain [179]. Indeed, “[t]here is no evidence that as of 1997 or even today, that anyone has used or tested for use, the racemate” [180]. The only hope was that utility of the racemate could be inferred, or soundly predicted, from the utility of the enantiomer. However, the evidence indicated that “it is not possible to predict the pharmacokinetic properties of a racemate of pregabalin and its enantiomer . . . based on the pharmacokinetic properties of either enantiomer” [182]. This lack of predictability is the point that has supported the validity of claims to an enantiomer over a known racemate; it is not surprising that this unpredictability should run the other way as well. So, Hughes J concluded that “as of the relevant date and even as of today, there is no factual basis from which a sound prediction as to the effectiveness of the racemate could be made” [183]. This is not based on a failure to disclose the factual basis in the patent. Hughes J does note that the expert for the patentee “had to go outside the '652 patent and have regard to [a prior] patent application” [182], which might suggest that it was a failure to disclose evidence in the patent which was fatal. However, his statement that “even as of today, “ there was no factual basis to make a sound prediction is categorical; obviously evidence supporting such a prediction that post-dated the patent could not have been disclosed in the patent itself. Moreover, there is no reference in the decision to any evidence at all supporting utility of the racemate. (Nor is this a false promise case, for reasons discussed above.)

As is often the case in overbreadth arguments, Hughes J’s conclusion that the patent was overly broad is really a corollary of the conclusion regarding lack of utility, rather than a free-standing ground of objection. The patent was overly broad because it claimed the racemate as well as pregabalin, and it claimed use for acute pain, but the only inventive contribution was that pregabalin is effective in treating chronic pain [93].

In the end therefore, the lack of utility of the claim in question did not turn on either of the controversial doctrines of Canadian utility law. Lack of disclosure of the factual basis for sound prediction was not in issue; indeed, lack of sound prediction itself was barely a live issue (arising only in respect of prediction of the use of the racemate from the enantiomer). The false promise doctrine was not in issue. Lack of utility turned simply on a failure to claim the commercial compound itself.

The patentee, Warner-Lambert, recognized this problem with the patent, and applied in 2005 for a re-issue to add new claims directed specifically at pregabalin, not the racemate, and at specific types of pain [207]. Warner-Lambert actually obtained a re-issued patent RE41,920 in the United States, in which one claim (19) was to pregabalin for the treatment of chronic pain – a claim which could have resisted the attacks made in this case. (And see Pfizer v Teva, 2012 WL 2951367, D.Del., July 19, 2012 (NO. 09-CV-307 GMS), dealing with the US patents.) However, for reasons that are not clear from the decision, the application for a re-issued patent was ultimately not pursued in Canada [212].


  1. The patent didn't claim that the recemate "as well as" pregabalin were effective, it claimed these in the alternative (see 27(5))!

  2. Yes, you're right, though I don't think this changes anything. Strictly, given the alternative, the claim to the pregablin branch was not overbroad, but it fails on the acute pain issue anyway, and the racemate branch fails on the utility issue.

    Sorry for the very long delay in replying - I somehow missed your comment when it was first posted.