Monday, March 2, 2015

Is There a Requirement to Refer to Evidence of Demonstrated Utility in the Patent?

Les Laboratoires Servier v Apotex Inc / gliclazide (NOC) 2015 FC 108 Roy J
            2,629,670 / gliclazide / DIAMICRON MR

As noted yesterday, in Gliclazide Dosage Form Roy J dismissed Servier’s application for an order of prohibition primarily on the basis of non-infringement [147], though he went on to hold the patent to be obvious and lacking in utility. The findings of non-infringement and obviousness were legally straightforward, but in the utility analysis has an important novel development (as if the law of utility was not controversial enough as it is): Roy J held that the factual basis for demonstrated utility must be referred to in the patent.

Servier argued that the promised utility was demonstrated by two bioequivalence studies. However, neither of these studies was alluded to in the specification [213]. Roy J held that these studies cannot be considered in establishing demonstrated utility:

the tests Servier points to that are not referred to in the `670 Patent are not relevant for establishing demonstrated utility [216].

It is now established in Canadian law that when utility is based on sound prediction, the factual basis for that prediction must be disclosed in the specification. Roy J did not hold that the factual basis for demonstrated utility must be disclosed, but he did hold that it must be referred to. That is, the distinction between sound prediction and demonstrated utility is whether the data itself must be set out in the patent, or whether it is sufficient to merely reference the relevant tests.

Roy J's holding is problematic in terms of both precedent and policy. Certainly we have come a very long way from Consolboard [1981] 1 SCR 504, which nominally remains the leading case on disclosure of utility. Consolboard held that it is not generally necessary to disclose the invention’s utility; how can that be reconciled with a requirement to disclose studies demonstrating utility? There are some exceptions to Consolboard in that it is necessary to disclose the invention’s utility or advantages in new use or selection patent cases, but this has not previously been understood to require referencing the tests themselves, and in any event, the invention in Gliclazide Dosage Form does not fall into either of these categories

I believe this is only the second case to exclude evidence of demonstrated utility because it was not referenced in the patent. The first was the FCA's Viagra Infringement 2014 FCA 13 decision (Sharlow JA: Dawson, Mainville JJA), which, as discussed here, also states that testing must be disclosed in the sense that a skilled reader could have “discerned from the specification that [the invention] was tested” [26]. The FCA in Viagra Infringement cited no authority for this proposition.

The authority relied on by Roy J at [214] is the following obiter statement from the FCA decision in Apotex / latanoprost 2011 FCA 236 (Trudel JA; Sharlow, Stratas JJA):

[30] There is no requirement to prove demonstrated utility in the disclosure of the patent; so long as the disclosure makes reference to a study demonstrating that the patent does what it promises to do, this criteria is met (Pfizer Canada Inc. v. Novopharm Ltd., 2010 FCA 242, [FCA Viagra NOC] at paragraph 90).

(Hughes J in a purely obiter discussion in Pregabalin (NOC) 2013 FC 120, discussed here, also cited the same passage from FCA Viagra NOC to the same effect.)

The passage in question from FCA Viagra NOC is as follows:

[90] The appellant’s argument that Pfizer was required to include evidence of demonstrated utility in the patent disclosure is without merit. The requirements for demonstrated utility can be provided in evidence during invalidity proceedings as opposed to in the patent itself. So long as the disclosure makes reference to a study demonstrating utility, there do not appear to be any other requirements to fulfil section 2.

This last sentence was obiter, in the sense that no evidence was actually excluded on this basis, and the main point of the discussion was that it is not necessary to demonstrate utility in the patent. This passage in the FCA decision is essentially an affirmation of the holding of Kelen J at first instance 2009 FC 638 [82] that:

The Court finds that there is no requirement in patent law that evidence of the demonstrated utility of the patent must be included in the patent. It is sufficient that the patent states that the invention has been demonstrated to be useful, as the '446 Patent does by making reference to the clinical testing of the compound (Study 350), and that the patent-holder is able to show evidence of demonstrated utility if the validity of the patent is challenged.

Kelen J’s holding was that it is only necessary to state that the invention has been demonstrated to be useful, and that referencing the study is simply one way of making that statement. In context, it is reasonable to interpret the FCA’s holding as being to the same effect. This interpretation is the best way to reconcile the holding with Consolboard; the patent at issue in Viagra NOC was for a new use of a know compound, so it is accepted that the new use itself must be disclosed (it will necessarily be disclosed in the claim itself), and there is a very fine line between disclosing that the invention is useful and disclosing that it has been demonstrated to be useful. In any event, the point was obiter in the sense that the Viagra NOC decision did not exclude any evidence because it was not referenced.*

Compare this with the following statement by Barnes J in Atomoxetine 2010 FC 915 (my emphasis) aff’d 2011 FCA 220 [45]-[46] (Evans JA: Noel, Dawson JJA):

It seems to me that it is beyond debate in Canada that where a patentee asserts that the utility of its invention has been demonstrated, it need not assert its supporting evidence in the patent.

See similarly Olanzapine (No 1) 2010 FCA 197 [74] (Layden-Stevenson J.A.: Sharlow, Nadon JJA), and Viagra NOC, 2010 FCA 242 [80] (Nadon JA: Blais CJ, Trudel JA( (rev’d but not on this point 2012 SCC 60), saying that “Evidence [of demonstrated utility] beyond that set out in the specification can, and normally will, be necessary.” These statements can be formally reconciled if we suppose that there is a difference between “asserting” evidence and “referring” to the evidence. But this kind of distinction is merely post-hoc formalism unless there is some principle underlying such a distinction.

There are also a number of cases holding utility to be demonstrated based on evidence that was not referenced in the patent. For example, in the leading FCA case on disclosure, Atorvastatin (NOC), 2008 FCA 108 Nadon JA: Linden, Ryer JJA rev’g 2007 FC 91,von Finckenstein J, the FCA held:

[56] The Applications Judge was wrong in interpreting the disclosure requirement of subsection 27(3) of the Act as requiring that a patentee back up his invention by data. . . . Allowing Ranbaxy to attack the utility, novelty and/or obviousness of the 546 patent through the disclosure requirement unduly broadens the scope of an inventor’s obligation under subsection 27(3) and disregards the purpose of this provision.

[62] I also conclude that the fact that the 546 patent does not provide a justification as to why the calcium salt of atorvastatin is the preferred embodiment of the invention does not render the disclosure insufficient. . . . Pfizer was not required to include in the 546 patent data which supports its statement that the calcium salt of atorvastatin is the preferred embodiment of the invention, nor was it required to explain why the calcium salt was the preferred embodiment.

On its face this passage could be read narrowly, as saying only that the data itself need not be disclosed. However, the patent itself did not refer to the relevant studies. The patent did disclose some experimental data, but von Finckenstein J held that utility had not been established because it did not disclose data related to the calcium salt in particular [86], [94], and it was on this point that he was reversed. Neither the FCA nor von Finckenstein J considered whether the studies related to the calcium salt were referred to, but so far as I can tell from my own reading of the patent, 2,021,546, there was no such reference.

Similarly, in Rosiglitazone (NOC) 2011 FC 239, Hughes J [96, [97], citing Viagra NOC, held that “[t]here is no requirement for a patent to demonstrate utility in the disclosure so long as the Court finds it to be proven when challenged in Court. The utility in question is to be determined, on the evidence before the Court, on the basis of what was done at Beecham prior to September 2, 1988.” The patent (1,328,452) did disclose murine tests on one of the compounds, but the disclosed tests were not on rosiglitazone. It is clear that the rosiglitazone tests that formed the basis for demonstrating utility were not disclosed: see [82], [99], [102] - [117].

In Bauer Hockey Corp. v. Easton Sports Canada Inc., 2010 FC 361, aff’d 2011 FCA 83, Noël JA: Létourneau, Trudel JJA, Gauthier J held that the utility of three promises [295] had been demonstrated without any discussion of whether the evidence of that utility was referred to. So far as I can see from reading the specification (2,302,953), there is no reference at all to studies demonstrating utility; the disclosure is entirely concerned with how to make the invention.

Nor can Roy J’s holding be reconciled with Rennie J’s Esomeprazole decision, 2014 FC 638, (blogged here) in which Rennie J noted that “it is not in dispute that disclosure is not required for the demonstration of utility" [130]. He then went on to hold that far from disclosure being required even for demonstrated utility, the enhanced disclosure requirement is confined to “a subset of sound prediction case” [151], in particular, it is “limited to the context of `new use' patents, assuming such a utility disclosure requirement exists at all” [141]. He concluded that one of the promises, stability against enzyme-mediated racemization, was satisfied based on “studies that were not disclosed in the ‘653 patent” [190]

And then of course there are Gauthier J’s remarks in Plavix FCA questioning whether a heightened disclosure requirement is appropriate at all.

In my reading, the weight of authority is contrary to the proposition that it is necessary to refer to evidence of demonstrated utility in the patent. In support, we have Roy J’s decision in this case, and two obiter statements from the FCA. Against, there is the FCA non-obiter holding to the contrary in Atorvastatin (NOC), as well as several FC decisions.

With that said, authority is not everything. Principle is important as well. Sometimes a split in the court signals an important matter of principle, and a minority view may eventually become a majority position. But a putative requirement to reference evidence of demonstrated utility fares even worse in terms of principle than it does in terms of authority. There are two related problems.

The first is that this doctrine implies a distinction between sound prediction, where it is necessary to disclose the data itself, and demonstrated utility, where it is only necessary to reference the data, rather than disclose it. No principled basis for this distinction is evident. The rationale provided by Roy J is as follows:

[215] No one suggests that the patentee must extol the virtues of its discovery. But without any reference to studies that will show, once they have to be produced, the existence of the promised utility, how is the public to know that utility is demonstrated?

Compare this with the rationale provided for disclosing the factual basis for sound prediction set out by the FCA in Atomoxetine:

[51] Indeed, if disclosure in the patent of the factual basis of the prediction of utility was not required for sound prediction, it would be difficult to see what Lilly could be said to have given to the public, in exchange for the grant of the monopoly, that it did not already have. When utility is based on sound prediction, disclosure of its factual foundation goes to the essence of the bargain with the public underlying patentability.

Except for the reference to sound prediction, there is really no difference between these rationales.

Secondly, Roy J’s rationale does not support a requirement to refer to the studies demonstrating utility. Roy J asks rhetorically, if the studies are not referred to “how is the public to know that utility is demonstrated?” But even if the studies are referred to, how is the public to know that utility is demonstrated? If the data itself is not disclosed, then the public cannot know that utility is demonstrated, in the sense that the public cannot judge for itself whether the experiments establish utility. Consider the disclosure at issue in Viagra.

In man, certain especially preferred compounds have been tested orally in both single dose and multiple dose volunteer studies. Moreover, patient studies conducted thus far have confirmed that one of the especially preferred compounds induces penile erection in impotent males.

In Viagra 2012 SCC 60 [40]-[42] the SCC held that the utility of sildenafil was adequately demonstrated and disclosed. If there is any requirement to refer to studies demonstrating utility, the SCC decision establishes authoritatively that it is satisfied by the above statement. (The FCA’s holding in Viagra Infringement that tests that were not referred to could not be considered was in the context of discussing utility of the second individually claimed compound, not sildenafil.)

How much information does the above statement give us? Which of the following would satisfy a requirement to refer to the relevant studies?

1)        Moreover, patient studies conducted thus far have confirmed that one of the especially preferred compounds induces penile erection in impotent males.

2)        Patient studies conducted thus far have confirmed that one of the compounds induces penile erection in impotent males.

3)        Studies have confirmed that the invention induces penile erection in impotent males.

4)        Studies have confirmed that the invention is useful.

5)        The invention is useful.

Statement (5) adds nothing more than we already know from the simple fact that the patentee filed a patent application, knowing that it would have to provide data to establish the utility of the invention if challenged. If the patentee files an application, the patentee must believe that it has sufficient evidence to establish utility. If statement (5) is enough, the requirement to refer to evidence of demonstrated utility is entirely redundant. If statement (5) is not enough, then we have to draw a line somewhere between statement (1) and (5). I can see no principled basis for drawing such a line.

Clarification of the state of the law is urgently needed. Perhaps it is time for the FCA to adopt an en banc review procedure, as in the US. We cannot rely on the SCC to straighten out every tangle. The SCC only hears about 60 cases a year, including many as of right criminal appeals, and it has an understandable focus on constitutional and criminal law cases. With what remains it has to deal with every other area of private law. The SCC simply does not have the ability to deal with every confused technical area of law, no matter how important that area might be to some parties. Moreover, the SCC is not a specialized patent court. The FCA has more patent law expertise than the SCC, and is more familiar with its own jurisprudence. It is understandable that confusion and even conflicts will arise as cases are decided by different panels of the FCA. But given that such splits with the FCA are inevitable, it seems to me that some institutional mechanism for resolving those splits is desirable.

*ADDENDUM: A commenter has pointed me to Pharmascience v AstraZeneca / NEXIUM (NOC) 2014 FCA 133 (Pelletier JA: Sharlow, Mainville JJA) in which the FCA explained its decision in Viagra NOC as indeed holding that “All that was required to establish that utility was demonstrated as opposed to soundly predicted was a reference in the patent to a study demonstrating utility” [40]. This does add more weight to the view that only evidence referred to in the patent can be considered in demonstrating utility. Again, this statement in this case was obiter in the sense that it was not actually applied to exclude any evidence. If anything, this decision makes a stronger case for the need for en banc review, as it implies a more even split as to whether such disclosure is required.


  1. See also: Pharmascience Inc. v. Canada (Health), 2014 FCA 133 at paras. 39-40.

  2. Thanks - I've added an addendum to the post.