Monday, January 26, 2015

Claim Defined in the Alternative: A Circle of Logical Contradiction

Eli Lilly Canada Inc v Mylan Pharmaceuticals ULC (NOC) 2015 FC 17 de Montigny J
            2,226,784 – tadalafil– CIALIS

In law we are accustomed to issues with two conflicting lines of authority. But when it comes to s 27(5) of the Act, which concerns claims framed in the alternative, we have one single line of authority – that ends of up contradicting itself, like a snake swallowing its own tail. There are two questions: (1) when a claim is framed in the alternative, so that s 27(5) applies, does the validity of the alternative in issue depend on the validity of the other alternatives?; (2) when a claim is not framed in the alternative, so that s 27(5) does not apply, does the validity of the alternative in issue depend on the validity of the other alternatives ? There are four possible answers to these two questions, and now, with de Montigny J’s Tadalafil decision, the courts have given all four answers, all based on one leading authority.

The claims at issue were to tadalafil and 3-methyl tadalafil for the treatment of ED. The s 27(5) issue arose in the context of the argument that utility had not been soundly predicted. Claim 2 and Claim 12 claimed both tadalafil and 3-methyl tadalafil. Testing on 3-methyl tadalafil was less extensive than on tadalafil. In particular, there was no evidence that it was administered to any animal for any purpose before the filing date. Mylan argued that because the utility of 3-methyl tadalafil was not established, the claim as a whole was invalid. de Montigny J held that on the contrary, “[w]hether the claims to 3-methyl tadalafil are soundly predicted . . . is irrelevant and of no consequence” [120].

Claim 2 is as follows:

A pharmaceutical composition for the curative or prophylactic treatment of erectile dysfunction in a male animal, comprising a compound selected from the group consisting of:

[tadalafil] or a physiologically acceptable salt or solvate thereof; and

[3-methyl tadalafil] or a physiologically acceptable salt or solvate thereof,

            together with a pharmaceutically acceptable diluent or carrier.

Section 27(5) of the Act provides:

(5) For greater certainty, where a claim defines the subject-matter of an invention in the alternative, each alternative is a separate claim for the purposes of sections 2, 28.1 to 28.3 and 78.3.

Now, on its face one might think that if s 27(5) applies, which is to say, if the reference to tadalafil and 3-methyl tadalafil in Claim 2 was taken to describe the invention in the alternative, then Lilly would win, because each alternative would be treated as a separate claim:

            2.A      A pharmaceutical composition for the . . . treatment of [ED] comprising [tadalafil]or a physiologically acceptable salt or solvate thereof.

            2.B      A pharmaceutical composition for the . . . treatment of [ED] comprising [3-methyl tadalafil] or a physiologically acceptable salt or solvate thereof.

If 2.B falls for lack of utility, this would be irrelevant to 2.A.

On the other hand, if the reference to tadalafil and 3-methyl tadalafil are not alternatives, but are considered to define a unitary group, then the principle would be that the invention must be useful across the full scope of the claim, and an inoperable species would invalidate the entire claim. So one might think that the central question was whether Claim 2, properly construed, claimed tadalafil and 3-methyl tadalafil in the alternative so that s 27(5) applies, or as a group so that it does not, and that the utility of 3-methyl tadalafil would be irrelevant if s 27(5) applies and relevant if it does not.

That thought would be wrong. The argument was exactly the opposite. It was Mylan that was arguing that s 27(5) did apply [116], and de Montigny J held that the utility of 3-methyl tadalafil was irrelevant because he concluded that s 27(5) did not apply [120]. Citing Apotex / clarithromycin 2005 FC 1332 [50]-[57] and Schering-Plough / desloratadine 2009 FC 1128, [88]-[90], he stated:

[118] I agree with Mylan that if an asserted claim is cast as two alternative claims, the entire claim fails if one of the alternatives is invalid

To understand how this seemingly backward interpretation of s 27(5) came about, we have to start with von Finckenstein J's decision in Abbott v Ratiopharm / Clarithromycin 2005 FC 1095. At issue was Claim 15 of patent 2,261,732 [20]:

A method according to Claim 2 wherein the solvent is selected from the group consisting of: [16 named solvents], and methanol-ethanol.

It was undisputed that the process worked when using one of the listed solvents (“solvent A”) [21], and Abbott only relied on infringement of the process when used with that solvent. Abbott argued that s 27(5) applied, and given that the utility of Claim 15 in respect of solvent A was established, it was therefore irrelevant whether the process worked with the other solvents [21]. This is exactly the obvious reading of s 27(5) that I outlined above, where, if the provision applies, each alternative is read as a separate claim and stands or falls independently. von Finckenstein J rejected Abbott’s argument, but he did not dispute that interpretation of s 27(5). Instead, he noted that on its face s 27(5) applies only “where a claim defines the subject-matter of an invention in the alternative,” and so the real question was whether the claim in question defined the subject matter in the alternative. After a thorough review of Markush claims, focusing in particular on the word “and”, he concluded that the claim did not define the solvents in the alternative, but rather as a group. Consequently:

[27] I cannot accept that these listed solvents were meant as alternatives and that section 27(5) of the Patent Act applies. Section 27(5) of the Patent Act is of no help as it only applies "where a claim defines the subject-matter of an invention in the alternative". That is not the case here.

Thus von Finckenstein J’s holding is entirely consistent with the straightforward interpretation of s 27(5) that I set out at the start of this post. If s 27(5) applies, the alternatives stand or fall independently, but because he held that it did not apply, validity turned on whether the claim was useful with all the specified solvents.

Soon afterwards, Phelan J decided Abbott v Apotex / clarithromycin 2005 FC 1332, which concerned the same Claim 15 of the same ‘732 patent [37]. Again, Abbott argued that “even if one or more alternatives in the claim includes unpatentable matter, that fact does not affect the overall validity of that claim in the patent” [37]. Phelan J noted that “Abbott must establish that there are alternative claims in the relevant claim” [44], and he concluded that “I am not satisfied that there are alternative claims” [48]. This leads straightforwardly to the conclusion that s 27(5) does not apply, and so if any one of the specified solvents did not work, the claim as a whole would fail. This was exactly the same reasoning and result as that of von Finckenstein J in Ratiopharm / clarithromycin.

But Phelan J did not stop there. He went on to hold that s 27(5) would not save the claim even if Claim 15 claimed in the alternative [50]-[57]. The key to Phelan J’s reasoning was that 27(5) itself limits its application to three sections of the Act, namely s 2 (definition of invention, including utility], 28.1 [claim date], 28.2 [novelty] to 28.3 [non-obviousness] and 78.3 [transitional provision], and crucially “S. 27(5) does not direct that alternatives in a claim constitute a separate claim for purposes of either s. 27 [grant of a patent & sufficiency] and 58 [claims stand and fall independently” [53]. That is, even if s 27(5) applies so that the claims are considered to be distinct for the purposes of utility, they are not considered distinct for the purposes of the s 58, which provides that claims stand and fall independently.

The obvious question is what s 27(5) is in the Act for, if s 58 does not apply to the alternative claims defined by that section? Phelan J said that “It applies only to the named provisions and is principally an administrative provision for purposes of a patent application.” Exactly what effect it would have in the context of a patent application was not specified, and none occurs to me.

With due respect, I fail to see the force of Phelan J’s interpretation. Section 27(5) states that when it does apply “each alternative is a separate claim for the purposes of sections 2,” which clearly means for the purposes of the utility determination. To my mind, the evident reason for treating the alternatives separately for the purposes of the utility determination is precisely so that each alternative may benefit from s 58. The reason certain specified provisions are listed in s 27(5) is that such a claim is nonetheless unitary for other purposes, namely s 36 (divisional applications), s 48.1 - 48.5 (re-examination), s 56 (prior user rights).

That is how Hughes J interpreted s 27(5) in GD Searle v Novopharm / celecoxib 2007 FC 81 rev’d on other grounds 2007 FCA 173. The claim at issue in Searle was

8. A pharmaceutical composition for treating inflammation . . .comprising a therapeutically-effective amount of a compound . . . selected from compounds according to any of Claims 1-7.

The question was whether “selected from compounds according to any of Claims 1-7" constituted claiming in the alternative: “Does this phrase include all compounds in those claims or is it simply a shorthand way of referring to each claim separately instead of having many claims, one referring to claim 1, the next to claim 2 and so forth” [29]. Hughes J construed Claim 8 to “to refer to claims 1 to 7 separately,” and therefore, citing s 27(5) [29], Claim 8 “can be construed to refer to claim 4 [celecoxib] only. Claim 8 therefore is directed to the use of a drug containing an appropriate quantity of celecoxib for treatment of inflammation or an inflammation related disorder” [32]. Hughes J then went on to consider obviousness and utility in respect of Claim 8 as referring to celecoxib only. Thus, he held that utility was established for celecoxib without considering whether the claimed invention was useful if compounds 1, 2, 3, 5, 6 or 7 were useful [105]. Their utility was simply irrelevant; and by implication this can only be because s 58 applied to save the valid alternative claim even if the others were invalid. This is contrary to Phelan J’s approach, in which the fact that Claim 8 claimed in the alternative would not allow s 58 to be invoked. Hughes J’s interpretation is both more logical, and it is better authority, since Phelan J’s interpretation was purely in dicta, while Hughes J’s interpretation was essential to his holding of validity.

With that said, I acknowledge that in Schering-Plough / desloratadine 2009 FC 1128, [88]-[90], also relied on by de Montigny J, Snider J applied Phelan J’s interpretation of s 27(5) in holding a claim invalid for anticipation. The claim in question (Claim 9 of the ‘136 patent), described the composition as being in “one of tablet or capsule form” [89]. Snider J had held the claim to be anticipated by prior patent disclosing a capsule form, and Schering argued that the tablet was an alternative, and so anticipation had to be considered separately for the tablet form [89]. Snider J, relying on Phelan J’s interpretation of s 27(5), held that anticipation of the capsule form was enough to invalidate the entire claim [92], even if the claim was framed in the alternative. In my view there is a real question, which was not addressed by Snider J, as to whether the claim to “tablet or capsule form” was really in the alternative. If not, s 27(5) would not save the claim. But that was not the approach taken by Snider J (indeed, it is not entirely clear whether she considered the claim to be framed in the alternative), and consequently there is a conflict in the case-law between Schering and Searle.

The implications of this conflict are far-ranging. If Phelan J’s interpretation of s 27(5) is right, then Hughes J’s interpretation is wrong. If Hughes J’s interpretation of s 27(5) is wrong, then claims of the form “according to any one of the claims 10 to 14, wherein [a restrictive condition]” are nonetheless valid only if the claim is valid for each one of the claims 10 to 14; it would not be enough to focus only on the commercially valuable embodiment. That is because even if we accept, as seems uncontroversial, that such claims are in the alternative so that s 27(5) applies, this is irrelevant to s 58 according to Phelan J and subsequent judges who have followed that line of reasoning, and the claim stands or falls as a whole. As Hughes J pointed out, this would mean that drafters would have to start drafting these claims explicitly as many separate claims compound – which is exactly what s 27(5) was, on its face, intended to avoid.

The last piece of the puzzle is that in Tadalafil de Montigny J held s 27(5) did not apply and because it did not apply “[w]hether the claims to 3-methyl tadalafil are soundly predicted, therefore, is irrelevant and of no consequence” [120]. This is exactly the opposite of von Finckenstein J’s holding in Ratiopharm / Clarithromycin, that because s 27(5) did not apply, the validity of the alternatives was relevant.

Thus in three short steps, we have a legal circle of contradiction: de Montigny J relied on Phelan J who relied on von Finckenstein J, yet the holdings of de Montigny J and von Finckenstein J are directly contradictory. The table below summarizes the cases:


s 27(5) applies? Validity of other alternatives relevant?
 von Finckenstein J No Yes
Phelan J dicta Yes Yes
Snider J Yes Yes
de Montigny J No No
Hughes J Yes No

The case law literally could not possibly be any more contradictory. There are two questions, and two answers to each, for a total of four possible answers to this pair of questions: and each of the four answers has been given by the courts.

The state of the case law is not just unsatisfactory, it is logically contradictory. The cases cannot be reconciled. Someone is right, and someone is wrong. von Finckenstein J and Hughes J gave consistent interpretations of the provision, and in my view they are correct. There is certainly a real question as to whether any particular claim is actually phrased in the alternative, so that s 27(5) applies. This question is properly the crux of the issue, as was recognized by von Finckenstein J and Hughes J. But if it is held that the claims are alternatives, they should be considered as alternative claims not just for the purposes of utility, novelty and non-obviousness in some unspecified administrative sense, but for all purposes including s 58.

4 comments:

  1. Adding an additional layer of logical uncertainty, I note that CIPO will often object to the claim language "in the alternative" as being unclear, where it is used in an effort to unambiguously invoke 27(5). Perhaps these cases illustrate that CIPO is correct, although one could reach an alternative conclusion.

    ReplyDelete
  2. Anon, very amusing (though depressing at the same time) thanks for pointing that out.

    ReplyDelete
  3. The left column of your summary chart is a bit misleading. What is not shown here is what type of language was considered in the determination of whether or not s27(5) applies: Markush language ("which is selected from the group consisting of X, Y, *AND* Z"), or alternative language ("which is X, Y, *OR* Z"). Claims drafters in other jurisdictions use these two formats relatively interchangeably now. There was a historical preference for Markush groupings in the U.S., as you may know, but only because it was not possible to claim in the alternative there in the early part of the 20th century.

    I also think it is a misnomer to title the second column of your summary chart "alternatives". I know what you are getting at, but as the members of a Markush grouping are not strictly considered to be alternatives, I believe that something like "alternatives/group members" would be more appropriate for this heading.

    In any case, the decisions are at least consistent on whether or not s27(5) applies, if the type of language used in the claims is considered, and the Hughes J decision is a bit of a different kettle of fish.

    - von Finckenstein J considered Markush language and found the s27(5) did not apply.

    - Phelan J also considered Markush language and similarly concluded that s27(5) did not apply. In obiter, he considered what would happen if the claim had been expressed in the alternative, using "or". In this case, he said s27(5) would apply, but would not have saved the claim because s27(5) does not refer to s58. So I think even your qualified left-hand "Yes" here is a bit misleading; the "Yes" was only to different/hypothetical claim language.

    - Snider J considered alternative language and found that s27(5) did apply.

    - de Montigny J considered Markush language and found the s27(5) did not apply.

    Up to this point, we at least have consistency in that s27(5) does not apply to Markush groupings, but does apply to alternatives. I certainly agree that, thereafter, the de Montigny J decision takes a rather “different course”!

    - Hughes J was dealing with claim dependency language, which does not fit neatly into either the "alternative" basket or "Markush" basket. It is a convention that is required by CIPO (MOPOP 11.06). In Canada, the options are to state "according to claim 1, 2, or 3" or "according to any one of claims 1 to 3" (interestingly, the Patent Office does not permit the Markush-y "selected from" to be included in latter as a matter of policy). The latter tends to predominate only because it is impractical to write out "according to claim 1, 2, 3, 4... 99, or 100", and much easier to write "according to any one of claims 1 to 100".

    As you say, it is also difficult to know what the purpose of s27(5) is intended to be; I cannot imagine that the statute drafters intended it to apply differently to Markush language and alternative language, and, up to the decision of de Montigny J, the practical result of having one invalid “option” recited in claim has been the same, no matter how the claim has been formulated.

    However, On the assumption that s27(5) was not intended to apply differently to Markush and alternative language (i.e. on the assumption that the initial finding of von Finckenstein J in this regard was wrong), the only explanation I have come up with for an intended meaning consistent with Phelan J’s subsequent obiter view on s58 is that s27(5) reminds us that "alternatives" (I use the term loosely) should be ASSESSED SEPARATELY for the purposes of determining claim date (28.1), utility (s2), novelty (s28.2), and obviousness (s28.3): in short, individual options are to be scrutinized SEPARATELY, but fall TOGETHER. In this way, a list of options (however formatted) would function very much like any old collective noun.

    But, alas, we have the Markush/alternative split in treatment, and now yet another confusing decision to factor in…

    ReplyDelete
  4. Anon

    Thanks for your detailed analysis. I agree that the decisions are reasonably consistent on whether or not s27(5) applies, if the type of language used in the claims is considered. One might debate whether the decisions, though consistent, are correct on the question of whether 27(5) should apply in terms of particular language. I avoided talking about the particular language because I did not want to to address that question, important though it is in its own right. My concern in this post is to point out that the decisions are, in my view, inconsistent as to what the consequences are if 27(5) does or does not apply.

    I agree with your interpretation of Phelan J’s obiter remarks, but as I see it, that interpretation is not consistent with Hughes J’s interpretation of 27(5). Even though Hughes J was dealing with different language, the key point is that he held that 27(5) does apply, and that is why the alternatives stand or fall separately, and not, as Phelan J would have it, together.

    (I also agree that "alternatives/group members" would be a better heading for the second column.)

    ReplyDelete