Monday, January 9, 2012

Plavix: Obvious or Not?

Apotex Inc v Sanofi-Aventis / clopidogrel 2011 FC 1486 Boivin J
            1,336,777 PLAVIX

Clopidogrel, the claimed compound at issue in this case, is the dextro-rotatory enantiomer of a racemic compound referred to as PCR 4099. Clopidogrel has superior pharmacological properties as compared with the racemate, in particular greater activity and lower toxicity. Patent 1,194,875, which was conceded to be part of the common general knowledge [608], claimed a genus encompassing PCR 4099, which was also specifically disclosed as the lead compound in Example 1 of the ‘875 patent [611]. The obviousness question was therefore simply whether the enantiomer was obvious over the racemate. In the NOC proceedings, which culminated in Apotex Inc v Sanofi-Synthelabo Canada Inc, 2008 SCC 61, the SCC held that it was not. In these infringement proceedings, Boivin J held to the contrary. What accounts for the differing conclusions?

The central question was whether separation of the enantiomers was “obvious to try.” To summarize this doctrine, the question of whether an invention was “obvious to try” is particularly relevant “[i]n areas of endeavour where advances are often won by experimentation” [SCC 68], such as chiral chemistry. It generally cannot be predicted in advance whether an enantiomer will have superior properties to the racemate; the only way to know is to separate the enantiomers and see. At one extreme it might be said that such an invention can never be obvious, because a skilled person who knows only the properties of the racemate – that is to say, before actually conducting the experiment – would never say “it is obvious that the enantiomer will be superior.” This position was rejected by the SCC in Sanofi. At the other extreme, it might also be said that whenever an invention is obvious to try, and actually succeeds when tried, then it is obvious. The SCC also rejected this view. Even if there is a strong motivation to try the experiment, if the experimentation is “ prolonged and arduous” [SCC 69], it will not be considered obvious, even if ultimately successful.

We might therefore say that in order to be obvious, the invention must be both obvious to try and obvious to succeed, with the understanding that “obvious to succeed” does not mean that success can be predicted in advance, but only that success followed on “routine trials” [SCC 69]. Conversely, we might say that the necessary inventive step may lie in either the decision to undertake the experiment, or in carrying the experiment to success. Either is sufficient: if it was very clever to even think of conducting the experiment, it does not matter that it succeeded easily; and if successfully carrying it out required great ingenuity, it does not matter that anyone might have conceived of the attempt.

In the clopidogrel litigation, both aspects were live issues. It is now generally recognized that enantiomers may have different properties from the racemates, but this understanding was emerging at the time of the invention in the late 1980s, and there was a question of fact as to whether this understanding had reached the level that the person of ordinary skill in the art working in the field in 1987 would have been motivated to separate the enantiomers [723]. There was also the question of how difficult it was to actually separate the enantiomers, even if we suppose it was obvious to try to do so.

In Sanofi the SCC held on the facts that the invention was not obvious to try. With respect to the second branch, whether the separation itself was inventive, the Court pointed to the number of different methods that might be used to separate the enantiomers [SCC 85], and the difficulty of actually achieving the separation [SCC 89]. Boivin J came to the contrary conclusion, largely because he had the advantage of hearing the testimony of Mr Bardoc, who actually carried out the separation. He found Mr Bardoc’s testimony to be very evasive and inconsistent [772]-[782], and it appears that rather than taking months to separate the enantiomers after a series of failed experiments, as stated in his affidavit relied on by the SCC [SCC 89], success was straightforward once he had settled on the“classic” technique. In this respect, the difference between the SCC’s conclusion and that of Boivin J is explained simply by the difference between making a determination on affidavit evidence, and making it after hearing the witness in person and after cross-examination. This is why NOC proceedings are not binding in a subsequent infringement action.

However, this alone does not explain the difference in the ultimate conclusion, since we must also address the first branch, whether it was obvious to try separation at all. The SCC’s conclusion was based primarily on the first branch, and not the second [SCC 89]. The SCC relied heavily on the actual course of conduct – the fact that Sanofi “had ‘spent millions of dollars and several years developing [the racemate] up to the point of preliminary human clinical trials’ without at least trying to see if the dextro-rotatory isomer had advantageous properties to those of the racemate” [SCC 91] (and see similarly [SCC 92]). The same point was also important in the finding that the corresponding US patent was not obvious: 550 F 3d 1075 (Fed Cir 2008) at 1088. In other words, the SCC found no good answer to the question, “If it was so obvious, why was it not done before?" Nichia Corp v Argos [2007] EWCA Civ 741 [21].

How did Boivin J address this question? His discussion under the heading “Actual Course of Conduct that Culminated in the Invention” [752ff] focused on the conduct in achieving separation, that is the second branch. He addressed the first branch, the decision to try separation, under the rubric of motivation [721ff]. That discussion focused on the general understanding of whether enantiomers had different properties, and he concluded that by the relevant date: “there was a clear indication that the separation of enantiomers could be performed and it was therefore important to test for them in order to pre-empt what was to come” [748]. Therefore, he concluded that a skilled person would be motivated to separate the enantiomers [750]. While Boivin J’s review of the evidence on this point is thorough and persuasive, he never directly addressed the question that troubled the SCC. If a skilled person would know it was important to attempt separation, why did Sanofi spend years and millions without doing so?

Sanofi clearly appreciated that enantiomers could have different properties, as they had in fact separated two earlier racemic compounds that had been under consideration. But at the same time, even though it was clear both to a skilled person and to Sanofi in particular that the enantiomer can have different properties, it is also clear than the enantiomers do not always have superior properties. For example, in the case of one of the prior compounds separated by Sanofi, the enantiomers had no better activity than the racemate [426], and in the other the enantiomer had better activity but equally poor toxicity [416]. So while it was common knowledge that the enantiomers can have different activity, it was perhaps not evident, at least to Sanofi, that the likelihood that they would have better activity was more than a “mere possibility” [SCC 66] and rose to the level of being “very plain” or “more or less self-evident” [SCC 65].

This question “If it was so obvious, why was it not done before?" is not purely rhetorical. In discussing the second branch of the obvious to try test Boivin J found that there were idiosyncratic reasons why Dr Bardoc had pursued some less obvious approaches to separation before trying the “classic” method which was ultimately successful [781]. This question may well have a satisfactory answer, but, as the FCA put it in Beloit (1986) 8 CPR(3d) 289 at 295

It is so easy, once the teaching of a patent is known, to say “I could have done that”; before the assertion can be given any weight, one must have a satisfactory answer to the question, “Why didn’t you?”

Boivin J directly addressed this question in his analysis of the second branch of the “obvious to try” test, but not, as I read it, in the first branch. In a a previous post I have noted that “secondary” evidence, such as the actual course of conduct, is often given more weight than is evidence that aims to reconstruct the state of mind of the inventor. In Beloit, for example, the trial judge had found the invention to be obvious based on expert evidence, but the FCA reversed, focusing on the actual course of conduct. Here, the SCC in Sanofi focused on the actual course of conduct, while Boivin J relied primarily on reconstruction of the reasoning of a hypothetical skilled person in light of expert evidence. Expert evidence is of course an essential part of any obviousness inquiry, but Boivin J’s analysis would have been more satisfactory if he had explained why, if attempting separation was obvious, it had not been attempted sooner.


  1. Interesting article. Please consider the following situations:
    1) Lets say there are 10 experiments of equal probability, where there is a high probability that at least one test/method from the 10 would yield the desired result (e.g. success of test GROUP is 70%, where success of each test is 7%). Can it be argued that it is "obvious" and "self evident" to try this CLASS of tests? Lets assume that all the tests are easy to do and "not arduous".
    2) Is the notion of "self-evident" in relation to the "TRY to obtain invention" or "OBTAINING the invention" itself. The language in Sinofi and Pfizer Canada Inc. v. Apotex Inc. (F.C.A.) seem to flip-flop on this a bit in the language. What I mean is, if you have 10 possible approaches to "try", where 2 of them have 40% chance each of giving desired result (as estimated before experiments), while the other 8 tests share a 20% chance of success, it can be said that the two top approaches are "obvious to try" (I would certainly give a 40% chance experiment a shot). HOWEVER, no one can say that either of these promising tests would cause the "invention [to be] more or less self-evident." (see para 65: "more or less self-evident that what is being tested ought to work") - given the likelihood of success is less than 50% in each case.
    3) Depending on your response to (2), if there are incredible riches awaiting to anyone who discovers an invention, wouldn't an experiment with a marginal rate of success, be elevated to the "obvious to try" category? If someone told me that I would get $10 billion dollars, I would certainly try all possibilities, even the ones that only have a 1% chance of success.

    Looking forward to your insight.

  2. 1) It can be obvious to try a class of experiments, as for example routine salt selection by screening of standard pharmaceutically acceptable salts: see eg ratiopharm inc. v. Pfizer Ltd. / amlodipine besylate 2009 FC 711, which I think is a good example of a correct application of the obvious to try test to find an invention obvious.

    2) The language does flip-flop a bit, but my view is that it must be self-evident to try to obtain the invention; if must be self-evident that to obtain the invention, then, as you point out, one could almost never say that an invention is obvious in an unpredictable art, no matter how obvious it was to try. So, when the SCC in Sanofi says that the “obvious to try” inquiry should be considered [62], it was saying that the invention need not be obvious to obtain. However, simply because it is obvious to try (ex ante 40% chance of success), and it actually succeeds once tried, does not mean it is necessarily obvious. It may turn out that there were unexpected hurdles in the way of ultimate success (the ex ante prediction of 40% chance of success was actually optimistic).

    3) This is a very good point, which was also made by Laddie J in Lilly Icos Ltd. v. Pfizer Ltd., [2000] EWHC Patents 49 ¶ 106 in the course of holding the invention in question to be obvious, aff’d with approval of this analysis [2002] EWCA Civ 1 ¶ 67-68:

    Whether something is obvious to try depends to a large extent on balancing the expected rewards if there is success against the size of the risk of failure.

    The FCA in Pfizer Canada Inc v Apotex Inc / sildenafil (NOC) 2009 FCA 8 expressly considered this approach at ¶ 42-45 and rejected it, in my view correctly. The risk / reward analysis advanced by Laddie J is compelling as a purely textual interpretation of the word “obvious.” However, it is at odds with a purposive interpretation of the obviousness requirement, which serves to screen out inventions which do not require the lure of the patent. If an approach is almost certain to succeed if tried, and yet will cost $100 million to try, then the lure of a patent is needed, so such an invention should not be considered “obvious” on a purposive interpretation. The purposive interpretation, and the FCA conclusion, is supported by SCC case law. In Halocarbon [1979] 2 SCR 929, the SCC noted “Very few inventions are unexpected discoveries. Practically all research work is done by looking in directions where the "state of the art" points.” The Court went on to say in effect that the “patient searcher” is entitled to their reward. Similarly in Sanofi 2008 SCC 61 ¶ 37 noted that “prolonged or arduous trial and error would not be considered routine.” Note also that Laddie J himself, speaking extra-judicially, has recognized the irrationality of the implications of his argument: see Conor v Angiotech [2008] UKHL 49 ¶ 48.