Apotex Inc v Shire LLC 2021 FCA 52 Rennie JA: de Montigny, Gleason JJA affg 2018 FC 637 Fothergill J
2,527,646 / lisdexamfetamine [LDX] / VYVANSE / NOC
In Sanofi 2008 SCC 61 [67], the SCC endorsed the four-step Windsurfing / Pozzoli approach to the obviousness determination. The second step of this inquiry is to “Identify the inventive concept of the claim in question or if that cannot readily be done, construe it.” This initially led to a focus on identifying the inventive concept as a central part of the obviousness analysis. Unfortunately, with no clear test or methodology, identifying the inventive concept has proven to often be difficult, as the FCA noted in Bristol-Myers 2017 FCA 76 (see here). In Ciba 2017 FCA 225 the FCA remarked that the search for the inventive concept can be a “distraction” [76] that has resulted in “confusion” and “uncertainty” that can be reduced “by simply avoiding the inventive concept altogether and pursuing the alternate course of construing the claim” [77]: see here. Consequently, the recent decisions of the FCA “have downplayed the importance of the ‘inventive concept’ as an analytical tool in the context of an obviousness analysis,” and have instead focused the analysis “on the claims themselves”: TearLab 2019 FCA 179 [78]. In Vyvanse the pendulum has swung back towards a renewed emphasis on the inventive concept, with Rennie JA stating that “The inventive concept, properly construed and applied, remains the end point for the obviousness inquiry” [65].
It is not very satisfactory to imagine the jurisprudence of the Court swinging back and forth with the needs of the moment. Bristol-Myers, Ciba, and Vyvanse make sense in their own terms; the more difficult question is how to reconcile them. In this post I will focus on Vyvanse, which is straightforward in some respects but puzzling in others. In the next post I will argue that an approach to obviousness that focuses on the solution to the objective problem faced by the skilled person is the only way to reconcile Bristol-Myers, Ciba and Vyvanse. In a third post, I will argue that the Windsurfing approach is useless or worse and should be abandoned.
The 646 patent relates to the compound LDX and its use for treating ADHD. Amphetamines have long been used to treat ADHD, but they can be abused by snorting or injection. LDX provides a solution to that problem in the form of an abuse-resistant sustained release prodrug [31]: see here for more detail. The claims at issue included both compound and use claims [4], but for the purposes of the obviousness analysis it is convenient to focus on the compound claims, in particular Claims 1–5, to LDX as such and its salts [FC 34].
The key issue in respect of obviousness was the role of the inventive concept in step 2 of the Windsurfing framework [59]. At trial, Apotex argued that in light of Ciba the “inventive concept” should no longer be used in the obviousness analysis, which should be determined by reference to the claims alone [FC 116]. Fothergill J rejected this argument on the basis that Ciba did not hold that the inventive concept could not be used, but only that assessing obviousness on the basis of the claims was an “alternative course” that could be pursued if the search for the inventive concept proved difficult [FC 115, 117]. Fothergill J held that in this case the inventive concept could “be grasped without difficulty” [FC 117], finding that it was “a sustained release formulation of a therapeutically useful dose of amphetamine that is resistant to abuse” [FC 122], [60]. Consequently, he focused on the inventive concept in assessing obviousness. In particular, the key difference between the state of the art and the inventive concept was the compound LDX and its advantageous properties [FC 132].
Crucially, none of the asserted claims expressly incorporated the properties of sustained release or abuse resistance [63]. This set up the central question on appeal in respect of obviousness. Apotex argued that Fothergill J’s focus on the inventive concept was an error, that the focus of the inquiry must be on the “subject matter of the claim,” and that recourse to the disclosure is not allowed [61]. Apotex argued that “the claims in issue are limited to their bare formulae. . . excluding their beneficial properties or anything that is not an ‘essential element’ of the claim. The essential element of the claim in issue was LDX – the bare chemical compound, without its features or advantages, as those can only be found in the specification” [61, my emphasis].
Rennie JA rejected this argument [64], holding that “The inventive concept, properly construed and applied, remains the end point for the obviousness inquiry” [65]. He pointed out that the need to go beyond the claims is evident when the patent at issue claims a bare chemical compound, as in both Sanofi and Vyvanse itself.
[72] As in Sanofi, claims 1-5 of the patent in suit in this appeal are to bare chemical compounds. The essential element of each of these claims is simply the chemical formula itself which, standing alone, says nothing as to the “inventiveness” of the patent claims. As such, it is necessary to turn to the specification for amplification. The language of Sanofi is directly applicable:
[77] The inventive concept of the claims is not readily discernable from the claims themselves. A bare chemical formula in a patent claim may not be sufficient to determine its inventiveness.
Rennie JA noted that in Sanofi Rothstein J examined “the beneficial properties of the bare formulae,” in order to assess obviousness [73]. Consequently, he concluded that
[74] I cannot agree that the effect of section 28.3 of the Patent Act is to narrow the inventive concept to the essential elements of the claim itself.
I suggest that this analysis runs together two different points. The narrow point is whether the properties of the claimed compound may be considered in assessing obviousness. Apotex argued they could not be taken into account. Presumably, this means that a claim to a compound would be valid only if it was difficult to synthesize. Rennie JA firmly reject this, holding that the beneficial properties of the claimed compound may be considered in assessing obviousness. On the facts, Rennie JA held that the beneficial properties disclosed in the specification and relied on by Forthergill J were indeed part of the inventive concept, and Fothergill J had therefore made no error in finding the claimed invention to be non-obvious [84].
On that point, I agree wholeheartedly. The rule that the properties of a compound may be taken into account in assessing obviousness is well-established in the caselaw.1 There are at least two different ways in which a bare chemical compound might be inventive. It might be that the compound was easy to synthesize, and the inventive step lay in discovering its use, eg as an antibiotic. Or it might be that the use was obvious, and the compound was difficult to synthesize, as for example an enantiomer of a racemic compound; or, in the early days of genetic engineering, an artificial version of a human hormone such as insulin or EPO. If the properties of the compound cannot be considered in the obviousness inquiry, that would mean that only the second type of compound claims would be valid. This would be bad law and bad policy. The discovery of a new active pharmaceutical ingredient is widely considered to be the most beneficial pharmaceutical invention; if the properties of a compound could not be considered, such claims would be decimated in favour of formulation patents and a few new compounds that are difficult to synthesize. This would be a perverse result. Moreover, the caselaw clearly supports this conclusion.
The broader holding is that “[t]he inventive concept, properly construed and applied, remains the end point for the obviousness inquiry,” [65], rejecting Apotex’s contention that “section 28.3 mandates a narrow, claim-based end point, focussed solely on the ‘subject matter of the claim’” [61]. Apotex essentially argued that a focus on the claim implies that the properties of the claimed compound cannot be taken into account. Rennie JA seems to have accepted that logic, and because it is clear that the properties of the claim can be taken into account, he rejected a focus on the claim in favour of a focus on the inventive concept. That is, Vyvanse appears to have implicitly proceeded on the basis that the choice is between (a) focusing on the claim and ignoring the properties or (b) focusing on the inventive concept and considering the properties. If that is the choice, then the latter option is preferable, as Rennie JA held. But I suggest that this is a false dichotomy, as there is another option. On the text of s 28.3 it is “[t]he subject-matter defined by a claim” that must not be obvious. In the case of a claim to a bare compound, the subject-matter defined by the claim is the compound itself. As Rennie JA emphasized at [96] the compound at issue, LDX, “cannot exist without [its] inherent properties: every molecule of LDX necessarily has these properties” [96].* Consequently, I would suggest that taking into account the properties of the claimed compound is entirely consistent with a focus on “the subject-matter defined by [the] claim,” which is the compound itself, including its properties. It appears from Rennie JA’s discussion at [23]–[24] that Shire made this argument. While Rennie JA did not pursue it, Shire’s argument strikes me as being essentially correct.
This may seem like a technical distinction, and to some extent, it is. In the great majority of cases, a focus on the inventive concept is helpful and appropriate. It is appropriate because the information provided by the specification is the quid pro quo for the patent, as I will discuss in my next post. Accordingly, a focus on the inventive concept implies a focus on whether the new information provided by the specification is non-obvious. But in my next post I will argue that this requirement is necessary, but not sufficient. In some cases, a patent may be invalid even though it provides new and non-obvious information. In those cases, a focus on the inventive concept—the information provided in the specification— does not work. Instead, it is necessary to focus on the subject-matter defined by the claim. Thus, I will argue that while a focus on the inventive concept is usually appropriate and helpful, the ultimate focus of the obviousness inquiry is the subject-matter defined by the claim. But that is a matter for subsequent posts; this post will focus on Vyvanse.
The focus on the inventive concept in Vyvanse makes sense in its own terms and on the facts of the case. But there is a real puzzle as to how to reconcile Vyvanse with Ciba 2017 FCA 225 and the related cases suggesting that it is not always necessary to identify the inventive concept.
As I read it, Rennie JA is saying that identifying the inventive concept is a necessary part of the obviousness inquiry. He began by noting that “It is at stage 2, where the Court is to ‘identify the inventive concept of the claim in question or if that cannot readily be done, construe it’, that the parties join issue. Distilled, the dispute before us is the end point of the obviousness inquiry” [59]. That is, the central issue was whether the end point of the obviousness inquiry is the inventive concept or the claim as construed. He rejected Apotex’s argument for “a narrow, claim-based end point, focussed solely on the ‘subject matter of the claim’” [61], and held that “[t]he inventive concept, properly construed and applied, remains the end point for the obviousness inquiry” [65]. Where there is no agreement on it, “the inventive concept needs to be construed” [67, my emphasis]. This seems to be saying that it is always necessary to identify the inventive concept as part of the obviousness inquiry. It is not always difficult to do so, but cases in which the task was easy do not stand for the proposition that it is unnecessary.
This is difficult to reconcile with Ciba. In Ciba Pelletier JA quoted with approval Jacob LJ’s remark in Pozzoli [2007] EWCA Civ 588 [19], that “if a disagreement about the inventive concept of a claim starts getting too involved, the sensible way to proceed is to forget it and simply to work on the features of the claim” [76], as well Jacob LJ’s statement in Actavis v Novartis [2010] EWCA Civ 82 [20] that the inventive concept “is a distraction almost as soon as there is an argument as to what it is.”2 Consequently, he held that unless the inventive concept can be “grasped without difficulty” the confusion and uncertainty that the search for it has generated
can be reduced by simply avoiding the inventive concept altogether and pursuing the alternate course of construing the claim. Until such time as the Supreme Court is able to develop a workable definition of the inventive concept, that appears to me to be a more useful use of the parties’ and the Federal Court’s time than arguing about a distraction or engaging in an unnecessary satellite debate.
This passage is a very clear statement that the inventive concept is not necessary to the inquiry, but is at best a handy shortcut that can and should be abandoned as soon as it causes trouble.
How then to reconcile Vyvanse with Ciba? Here is what Rennie JA had to say about Ciba. (It is perhaps worth noting that Rennie JA was on the Ciba panel, though Pelletier JA wrote the decision.)
[100] Ciba recognises that an inventive concept must be based on a claim, and not some vague paraphrase in the disclosure. Thus, if the identification about an inventive concept is not readily apparent, the judge should “simply work on the features of the claim” (Ciba at paras. 74-76). This ensures that obviousness is grounded in the claims themselves, a requirement discussed in both Sanofi and the Patent Act. Ciba does not address what should happen when, after examination of the claims construction, the inventive concept is still not “readily discernible”. Pursuant to Sanofi, that is when recourse to the specification is allowed (at para. 77).
The most obvious point from this discussion is that Ciba stands for the proposition it is necessary to focus on the particular claim. Ciba does emphasize that point [74]–[75], but that doesn’t address this issue of whether the claim itself or the inventive concept of the claim is the focus of the inquiry.
The penultimate sentence suggests that Rennie JA considers Ciba to be a case in which the inventive concept was readily discernable from the claim. Another passage at [67] might be taken the same way: “To [construe the inventive concept]the judge is to first determine whether it can be identified from the previously completed claims construction exercise (Ciba at paras. 76-77).” If Rennie JA really did intend to explain Ciba as a case in which the inventive concept was readily discernable from the claim, it would have been preferable if he had said so more explicitly, as Ciba certainly does not read that way on its face.
Immediately after criticizing the search for the inventive concept as “an unnecessary satellite debate,” Pelletier JA in Ciba stated “As a result” — that is, as a result of the confusion relating to the inventive concept — “I turn to the construction of Claim 1” [78]. Now, after turning to what he described at [78] as a claim construction discussion, Pelletier JA concluded that
while conventional doses of flocculant were ineffective in solving the problems to which the invention was directed, doses in the ranges described in the patent, which must be taken to be outside the range of conventional doses, would have a beneficial effect” [85].
While Pelletier JA expressly referred to this as “[h]aving thus construed claim 1” [86], it arguably actually describes the inventive concept. It certainly doesn’t look much like the usual outcome of a claim construction exercise, which is to specify the precise meaning of contentious terms, such as “rigidify” (see eg Ciba FC [163]), or identifying which elements are essential. This might support an argument that while Pelletier JA said he was simply construing the claims, in fact he identified the inventive concept. Further, this conclusion was expressly based on the disclosure: “The expression ‘effective rigidifying amount’ is ambiguous, which allows reference to the disclosure to understand what is meant” [84]. Perhaps it could be said that Pelletier JA had recourse to the specification to construe the claim, and not to construe the inventive concept?
This might be taken to support the view that Ciba actually did identify the inventive concept. But when it came to actually assessing obviousness in respect of the invention at hand, Pelletier JA did not look to the nature of the invention he described at [85] (whether we consider that to be Claim 1 as construed, or the inventive concept). Instead, he considered the relevant prior art and identified the difference between the prior art and the elements of the patent in issue—quoting the novel elements directly from Claim 1 [91]. He then considered whether the differences represented by those two particular claim elements were obvious: [92]–[93] (yes, for the first difference), [94]–[95] (yes, for the second difference).
I don’t know where this leads. On its face, Ciba is a case in which the FCA addressed obviousness by focusing on the claims, which is at odds with the focus on the inventive concept in Vyvanse. Perhaps Rennie JA considered Ciba to be a case in which the inventive concept was readily discernable from the claims—perhaps on the view that the two claim elements addressed by Pelletier JA represented the inventive concept. I'll discuss that possibility in more detail below. But if that is indeed how Rennie JA interpreted Ciba, it would have been preferable if he had said so directly, as that is not apparent on the face of Ciba itself.
At the end of the day, Ciba is a case in which the FCA expressly said it would assess obviousness by looking to the claims as construed at stage 2 of the Windsurfing / Pozzoli test, and it actually did so. Rennie JA has given us a few hints as to how he would reconcile Ciba with a focus on the inventive concept, but the discussion is not satisfactory.
That leaves us with a puzzle. Ciba makes sense. Vyvanse makes sense. But how can they be reconciled? At a broad level, Vyvanse says that identifying the inventive concept is necessary, and Ciba says it’s hard. These are not incompatible positions. Ciba did not reject the use of the inventive concept entirely, but only “[u]ntil such time as the Supreme Court is able to develop a workable definition of the inventive concept,” as Pelletier JA put it in Ciba at [77]. Now, Pelletier JA suggested that the SCC should be the one to develop a workable definition, but it seems to me that the FCA is better placed to do so. The FCA has far more experience with patent cases than does the SCC and is consequently much better placed to know whether a given approach would be workable. That appears to be Rennie JA’s project in Vyvanse.
1.3 Identifying the inventive concept
A key question, then, is how the inventive concept should be “properly construed and applied” [65]? Rennie JA set out three “basic principles” [66]. While I agree entirely with the key holding in Vyvanse, that the properties of a compound may be considered in assessing obviousness, I am less comfortable with these broader principles.
Consider the first “basic principle” set out by Rennie JA:
[67] First, on occasion, the inventive concept may be “readily apparent” where there is agreement on it. If not, the inventive concept needs to be construed. To do that, the judge is to first determine whether it can be identified from the previously completed claims construction exercise (Ciba at paras. 76-77). Second, where it is not possible to fully grasp the nature of the inventive concept solely from those claims, the judge may have regard to the patent specification to determine if it provides any insight or clarification into the inventive concept of the claim(s) in issue (Sanofi at para. 77; AstraZeneca Canada Inc. at para. 31).
This seems to set up a novel three step approach to identifying the inventive concept, which is intended to synthesize the prior cases. Notice the language of “construing” the inventive concept, which I don’t think I have ever seen before.
I should begin by saying that what follows is a very long discussion of some quite brief remarks, involving considerable speculation as to what Rennie JA might have intended. No doubt the FCA will provide further clarification in future decisions. I think it is worthwhile to consider the possibilities nonetheless, given that the parties and the Federal Court are understandably in search of guidance regarding the inventive concept.
With that, let’s consider the steps in order.
1.3.2 Where there is agreement
First, on occasion, the inventive concept may be “readily apparent” where there is agreement on it.
This might seem straightforward—if the parties agree on the inventive concept, then that’s what it is. But in Bristol Meyers the FCA held that it was an error to adopt the inventive concept proposed by the parties: “On the basis of the arguments made to it by the parties, the Federal Court implicitly adopted a definition of the inventive concept which focussed on the properties of atazanavir bisulfate. This was, in my view, an extricable error of law that justifies our intervention” [74]. This means that the inventive concept may need to be construed by the court even when the parties agree on it.
1.3.3 Can the inventive concept be identified from the claim construction?
To do that, the judge is to first determine whether it can be identified from the previously completed claims construction exercise (Ciba at paras. 76-77). Second, where it is not possible to fully grasp the nature of the inventive concept solely from those claims. . .
The exercise of construing the inventive concept itself branches depending on whether the inventive concept can be identified from the “claims construction exercise” or equivalently “solely from those claims.” How are we to know whether the inventive concept can be grasped solely from the claims? Rennie JA does not tell us. The only authority he cites is Ciba, and, in my view, Ciba does not help.
Ciba did not really say that the first step is to determine whether the inventive concept can be identified from the claims. Pelletier JA said “[t]here may be cases in which the inventive concept can be grasped without difficulty,” but he did not say, or even suggest, as I read it, that there were cases where the inventive concept could be grasped without difficulty from the claims. The cited paragraphs of Ciba are where Pelletier JA counseled that when the inventive concept caused difficulty, it should be avoided entirely in favour of the “alternative course of construing the claim” [77]. As discussed above, it is possible that Rennie JA is interpreting Ciba as a case in which the inventive concept could be discerned solely from the claims. I find it difficult to read Ciba itself that way, even apart from the express statements in Ciba that it was not considering the inventive concept at all. The closest thing I can find in Ciba to a statement of the inventive concept is [85], quoted above. This conclusion was expressly based on consideration of the disclosure, which was discussed in [84]. In any event, whether or not Ciba can be understood as a case in which the inventive concept can be discerned solely from the claims, Ciba certainly did not give advice on how to identify the inventive concept from the claims.
Nor did Rennie JA provide any other examples of cases where the inventive concept can be determined solely from the claims. Perhaps the suggestion is that the inventive concept may be grasped solely from the claims when the claims expressly contain elements corresponding to the inventive concept.3 In my view, this suggestion is not tenable. Even when the claims do expressly contain an element corresponding to the inventive concept, the claims very commonly also contain elements that do not correspond to the inventive concept, and I cannot see how it is possible to tell which is which from an examination of the claims alone.
So, in this case, amphetamines were known to be useful in treating ADHD, and the problem faced by the inventors was that they were susceptible to abuse [5], [66e,f]. Previously, “formulations were the primary, perhaps the only, method used to reduce a drug’s potential for abuse” [FC 130]. No prodrugs had been developed as a means of reducing abuse potential, and without actual development and testing, it was not obvious that using a prodrug form would be effective [FC 131]. The decision to pursue a prodrug strategy was the key decision leading to the discovery that a prodrug was an effective abuse resistant medicine for treating ADHD. This led to Fothergill J’s conclusion that the inventive concept was “a sustained release treatment of amphetamine with a reduced potential for oral, intravenous and intranasal abuse” [FC 132]. So, on the facts of this case, the suggestion might be that an example of a claim where the inventive concept of “a sustained release formulation of a therapeutically useful dose of amphetamine that is resistant to abuse” would be readily apparent from the claims would be a hypothetical Claim 52 to “A pharmaceutical composition comprising one or more pharmaceutically acceptable additives and an amount of from 25 to 75 mg of LDX providing a mean T1/2 (hours) from 9.68 ~ 1.43 to 10.3 ~ 1.7 of amphetamine when orally administered to a human subject, for decreasing abuse of amphetamines or salts thereof in a subject in need thereof.”
Now consider a hypothetical twist on these facts. Suppose prodrugs were commonly used for abuse resistance and extended release and that LDX was known to be the prodrug of a particular amphetamine, so that it was obvious that LDX would provide an abuse resistant sustained release substitute for amphetamines for treating ADHD—but actually synthesizing LDX turned out to be very difficult. In that case, the key question in the obviousness inquiry would presumably be whether a chemist of ordinary skill would have been able to synthesize LDX, not whether a skilled person would have considered a prodrug as a solution to the problem of abuse resistance. In that case, the inventive concept, it seems to me, would be the method of synthesis of LDX, not its advantageous properties. In this hypothetical scenario, Claim 52 would be entirely permissible and appropriate; the inventor would no doubt claim LDX as such, but would likely also want to claim the main commercial use. But in that scenario, it seems to me that the inventive concept cannot be grasped solely from Claim 52; the claim itself is the same in the two scenarios, but the inventive concept is different.
More generally, the fact that an element appears in a claim does not tell us whether that element is related to the inventive concept. If the claim is to a formulation comprising a specific vaccine and extract of the Chilean soapbark tree, how can we tell from the claim whether the vaccine is the clever part and the soapbark tree extract is a standard adjuvant; or whether using soapbark tree extract as a vaccine adjuvant is the clever part; or whether the insight is that soapbark tree extract is an effective adjuvant for that specific vaccine? As another example, if a claim is to a specific salt of a particular compound, the clever bit is usually the discovery of the properties of the compound, but occasionally the compound was known and developing an appropriate salt form was the challenge. How can we tell from looking at the claims alone?
In the absence of examples in Vyvanse itself, I am grasping at straws here. I have speculated that perhaps the suggestion is that the inventive concept may be grasped solely from the claims when the claims expressly contain elements corresponding to the inventive concept, but I don’t know whether that is really what Rennie JA had in mind. I don’t know that Rennie JA would have said that my Claim 52 is an example of a claim where the inventive concept could be grasped solely from the claim. Perhaps there is some other scenario that hasn’t occurred to me where the inventive concept is readily apparent from the claims. Or perhaps my analysis of the putative Claim 52 is wrong. In any event, as it stands, I can’t think of any example where the inventive concept is reliably discernable solely from the claim.
1.3.4 When is recourse to the entire specification permitted?
On to the third step:
3) Where it is not possible to grasp the inventive concept solely from the claims, the judge may have regard to the specification.
As a matter of the sentence structure, the phrase “where it is not possible to fully grasp the nature of the inventive concept solely from those claims” seems to be setting a condition precedent on recourse to the specification. This suggests that recourse to the specification to identify the inventive concept is not permissible unless it cannot be identified from the claims.
If that is indeed what Rennie JA intended, I see two problems. First, as just discussed, it’s not clear (at least to me) how to decide whether the inventive concept can be grasped solely from the claims. If we can only have regard to the specification if the inventive concept cannot be grasped from the claims, but it is not easy to decide whether that is true, then there is likely to be dispute as to whether this requirement is satisfied. The “satellite debate” as to the nature of the inventive concept has proven troublesome enough; we don’t need a satellite around that satellite, debating whether the inventive concept can or cannot be identified from the claims alone.
Second, it is not clear what principle underpins this rule. Rennie JA made it clear that in some cases it is permissible to have recourse to the disclosure to identify the inventive concept and on the facts of the case he relied very extensively on the disclosure: see eg [67], [70], [71], [72], [74], [79], [80]–[84]. If the court cannot have reference to the disclosure when the inventive concept is clear from the claims, that means there is a sharp distinction: in some cases we cannot have recourse to the disclosure at all, and in other cases we must rely heavily on the disclosure. Given that the claims and the disclosure are both part of the specification, I have considerable difficulty seeing what principle might underpin such a distinction.
As authority for this proposition Rennie JA cited Sanofi [77] and AstraZeneca 2017 SCC 36 [31]. Sanofi says (my emphasis):
[77] The inventive concept of the claims is not readily discernable from the claims themselves. A bare chemical formula in a patent claim may not be sufficient to determine its inventiveness. In such cases, I think it must be acceptable to read the specification in the patent to determine the inventive concept of the claims. Of course, it is not permissible to read the specification in order to construe the claims more narrowly or widely than the text will allow.
Should this statement, that is permissible to read the specification to determine the inventive concept when it is not apparent from the claims, be understood to mean that it is otherwise impermissible to read the specification?
I think not. The SCC cites no authority for this particular proposition, and I am not aware of any prior decisions at all in which the court held that it was impermissible to read the specification to determine the inventive concept. On the contrary, the prior leading cases all routinely have recourse to the specification to determine the inventive concept; none of the leading cases suggest that there is first step of determining whether the inventive concept can be determined from the claims alone. In Windsurfing, the inventive concept was discerned directly from the disclosure without any attempt to first discern whether they could be grasped from the claims alone.4 This was even though the inventive elements were expressly included in the claim, albeit accompanied by various other elements necessary to describe a functioning sailboard, so Windsurfing is at least a candidate for being a case in which the inventive concept could be grasped solely from the claims.5 In Pozzoli, the EWCA remarked that “it is generally better to construe the claim” before determining the inventive concept, though construing the claim is of course not the same as determining the inventive concept, but “we heard argument in a different order at our insistence” [38]. That is, the EWCA began by determining the inventive concept before even construing the claim, much less considering whether the inventive concept can be determined solely from the claims. Given that the SCC in Sanofi approved the analysis set out in Windsurfing and Pozzoli, prima facie, Sanofi should not be interpreted in a manner that is inconsistent with those decisions.
Rennie JA also cited AstraZeneca [31], which says:
[31] That is, the Promise Doctrine requires the identification of promises based on a review of the entire specification, i.e. both the claims and the disclosure. Generally, an analysis regarding issues of validity, such as novelty or non-obviousness, focuses on the claims alone, and only considers the disclosure where there is ambiguity in the claims (Sanofi-Synthelabo). This is in accordance with this Court’s direction that claims construction precedes all considerations of validity. The Promise Doctrine, by contrast, directs courts to read both the claims and the disclosure to identify potential promises, rather than the claims alone, even in an absence of ambiguity in the claims.
Taken on its own, the statement that “Generally, an analysis regarding issues of validity, such as novelty or non-obviousness, focuses on the claims alone,” simply means that the ultimate question is whether the subject-matter defined by the claims is new, useful and non-obvious. This is well established and enshrined in the Act in the case of novelty and obviousness. Under the promise doctrine, validity was assessed in terms of the statements made in the specification, entirely apart from the claimed subject-matter. The promise doctrine was wrong to do so and that is the main point of this paragraph.
The sole source of the view that the inventive concept must first be discerned from the claim alone is the statement that validity “only considers the disclosure where there is ambiguity in the claims. This is in accordance with this Court’s direction that claims construction precedes all considerations of validity.” There are two problem with using this as authority for the proposition that recourse to the specification is permissible to determine the inventive concept only where it is not possible to identify the inventive concept solely from the claims (if indeed that is what Rennie JA intended). First, the SCC’s suggestion that claim construction only considers the disclosure where there is ambiguity in the claims is not really correct. While there is an ongoing debate on this point, the current trend is that the claim construction exercise must always consider both the disclosure and the claims. In practice, the courts will always consider the disclosure in construing the claims where it is relevant, as even the judges who express the view that recourse to the disclosure is permissible only if the claims are ambiguous, or who don’t want to have to grapple with the issue, will essentially always find the requisite ambiguity. I’m not aware of any cases which refused to look at the disclosure when doing so would have made a difference. However this debate ultimately plays out, the view that recourse to the disclosure is permissible only when the claims are ambiguous is certainly not a fundamental principle of claim construction. It may not even be the law, and if it is, it is of marginal significance.
Second, even if it were well settled that recourse to the disclosure is permissible to construe the claims only when the claims are ambiguous, the question of whether it is only permissible to consider the disclosure to determine the inventive concept when the concept cannot be grasped from the claims is a distinct question, and there is at most a loose analogy between the two. While my own view is that recourse to the disclosure to construe the claims should always be allowed, the contrary view, that it is permissible only when the claims themselves are ambiguous, does have a logical rationale, namely that a party doing a freedom to operate search should be entitled to rely solely on the claims as they appear clear on their face, to save the time and expense of a purposive interpretation in light of the specification, particularly if it is necessary to review large numbers of patents. It is difficult to see any corresponding argument related to the inventive concept; if a party is considering a validity challenge it does not seem unreasonable to ask them to read the disclosure.
In summary, a putative rule that recourse to the disclosure to determine the inventive concept is permissible only if it is not apparent from the claims themselves is inconsistent with the leading authorities, namely Windsurfing and Pozzoli. The main basis for such a rule would be a loose analogy with the SCC’s statement of a disputed claim construction rule in a case that had nothing to do with the inventive concept or even obviousness. In the absence of a clear principled basis for the rule, this is not an adequate basis for introducing a new rule that departs from the leading cases. The question of whether recourse to the claims is permissible only when the claims are ambiguous is the very definition of a sterile debate. It has caused considerable confusion in the law, with many pages and much argument devoted to it, and yet it never affects the outcome. It would be very unfortunate to see it imported into the Windsurfing test for obviousness, which is confusing enough as it is.
Perhaps I’ve devoted too much space to this point. It is far from clear that Rennie JA really intended to say that recourse to the disclosure to determine the inventive concept is permissible only if it is not apparent from the claims themselves. I don’t want to read too much into a couple of sentences, and after all, in this case Rennie JA did have recourse to the disclosure to determine the inventive concept, so the statement is strictly obiter.
1.4 What is the inventive concept?
Much confusion has arisen because the leading cases, including Windsurfing, Pozzoli, and Sanofi, use the term “inventive concept” as though it has some definite meaning which is clearly understood, when in fact it is ambiguous and is used in a variety of ways: Bristol-Myers [64]. All of the concerns I’ve raised about the Vyvanse framework ultimately turn on two questions: What is the inventive concept? How do we know when we’ve found it?
The second “basic principle” identified by Rennie JA is that “while an inventive concept is an attribute of the claims, it differs from claims construction” [68]. On its face, this simply says that identification of the inventive claim concept is different from the claim construction exercise, as is illustrated by the example of a bare chemical compound claim. I’m not sure anything more was intended. The statement that the inventive concept is “an attribute of the claims” is a bit puzzling. In the context of a bare chemical compound claim, the main point of Vyvanse is that the unexpected properties of the compounds may comprise the inventive concept, and those strike me as an attribute of the claimed subject-matter rather than an attribute of the claims themselves—perhaps that is what Rennie JA meant. But in some cases of a bare chemical compound claim, the inventive part might be discovering how to synthesize a compound that was already known to have useful properties, and I’m not sure the method of making a compound can be said to be an attribute of the compound itself, given that different methods of synthesizing a compound are commonly subsequently discovered. In any event, the statement that the inventive concept is “an attribute of the claims” doesn’t purport to be a definition of the inventive concept and, as noted, the main point of this second principle is simply that identification of the inventive concept is different from claim construction.
Rennie JA also states that the inventive concept is not the same as the essential elements of the claim itself [74]. So, in the case of a claim to a compound, as in Vyvanse, the only essential element of the claim is the compound itself, while the inventive concept lies in its properties. To equate the inventive concept with the essential elements of the claim “conflates the claims construction exercise with the identification of the inventive concept” [74]. I agree with all of this, but it only tells us what the inventive concept is not; it doesn’t tell us what the inventive concept is. Rennie JA also remarks that the inventive concept “is an element of [the obviousness] determination. Its purpose is to help determine what, if anything, makes the claim, as constructed, inventive” [76]. This tells us what it does, but not what it is.
The closest thing to an express definition of the inventive concept are two references to the “solution taught by the patent” in Bristol-Myers (my emphasis):
[76] As noted in Bristol-Myers, Sanofi did not change the substantive law of obviousness by implication, and the term “inventive concept” is not materially different than the previously used term of “solution taught by the patent” (Bristol-Myers at paras. 65-68, 75).
[84] Unlike the situation seen in Bristol-Myers, these beneficial properties were the “solution taught by the patent” claim. They explain the source of the motivation to pursue the solution (Bristol-Myers at para. 75).
In my next post I will argue that the approach taken in Bristol-Myers is indeed the best approach. But that requires considerable unpacking. While the cited paragraph of Bristol-Myers does say that “the solution taught by the patent” is the same as “the inventive concept,” is also says that it is the same as “what is claimed” [65]. As we’ve seen, in Vyvanse Rennie JA makes it clear that the inventive concept is distinct from what is claimed, so there is a puzzle of how to reconcile Vyvanse and Bristol-Myers at that level. I’ll return to Bristol-Myers in my next post, but for now it is enough to say that Vyvanse itself does not provide a clear definition of what is meant by the “inventive concept.”
1.5 How to identify the inventive concept
That’s all Vyvanse says about what the inventive concept is. The next question is how to go about determining it. Rennie JA does not provide a set methodology, apart from the basic principles discussed above, though he did approve Fothergill J’s approach, which he described as being based on “an analysis of the claims as informed by the specification” [71] (and similarly [70]). A point repeatedly emphasized by Rennie JA is that the specification may be used to help identify the inventive concept, both in general statements of principle—see eg [67], [70], [71], [72], [74], [79]—and in determining the inventive concept of the claims at issue [80]–[84]. Indeed, everything that Rennie JA relied on in determining the inventive concept was “rooted in the specification” [82].
Rennie JA made it abundantly clear that it is permissible to go beyond the claims, to the disclosure, in determining the inventive concept, at least when the inventive concept is not apparent from the claims. The question then arises as to whether it permissible to go beyond the specification. In particular, it is permissible to consider the prior art in comparison with the specification, or is the inventive concept to be found solely in the specification, without regard to the prior art?
Rennie JA’s focus on the specification might be taken to suggest that the inventive concept is to be found solely in the specification. However, I think that would be reading too much into it. Rennie JA did not expressly address this question, and as I read it, his focus on the specification was simply a response to Apotex’s argument that the proper approach was an exclusive focus on the claims, and that recourse to the disclosure was not allowed [61]. Consequently, I don’t think Vyvanse should be read as saying that the inventive concept is to be found entirely from an examination of the specification; the point is simply not addressed. There is an open question as to whether the inventive concept is to be found solely in the specification, or whether its identification requires a comparison between the specification and the prior art.
This emphasis on the specification raises another puzzle. The third “basic principle” identified by Rennie JA is that “[i]t is the ‘inventive concept’ of the claim in question which must be considered, not some generalised concept to be derived from the specification as a whole” [69], quoting Unilever v Chefaro [1994] RPC 567 (Pat).6 How is this to be squared with Rennie JA’s emphasis on the specification in identifying the inventive concept? Rennie JA’s answer is to emphasize unity of invention:
• “a single inventive concept must flow through a patent, but each claim’s specific inventive concept may be different” [77];
• “it is important to remember that a single, overarching inventive concept connects every claim of a patent, with its genesis usually in the independent claim(s)” [86];
• "Despite the singular common inventive concept, '[d]ifferent claims can, and generally will, have different inventive concepts.' These are, in effect, inventive concepts that are stitched on, or bound to, the single, common concept" [87];
• As a practical matter, the inventive concept of particular claims need not be explicitly considered so long as the “single common inventive concept” is found to be non-obvious [87];
• “each claim can give rise to its own inventive concept, and the inventive concepts of the various claims may overlap or replicate each other. This is permissible, provided they are joined or unified by an over arching, single inventive concept” [88].
These points were directed primarily to Apotex’s argument that Fothergill J’s identification of the inventive concept was faulty “because it made the inventive concept of some of the claims ‘necessarily redundant’” [85]. In that context, Rennie JA’s point that there is typically a single overarching inventive concept establishes that “the potential for redundancy is inherent in the inventive concept exercise” [88]. If there is a single overarching inventive concept joining all the claims, it should not be very surprising if in some cases the inventive concept of two different claims was exactly the same. In this case some claims defined LDX verbally, as “L-lysine-d-amphetamine” while others defined the same compound visually, as a structural formula: [FC 75]. This example is a bit unusual, in that it is not just the inventive concept of these claims that is the same, but the very subject-matter, which is simply described in two different ways. A more typical example is when an inventor discovers that a new compound is useful in treating a particular disease and the patent includes claims to the compound itself, as well as claims to the compound in different forms: in this case, for example, there are separate claims to a 20mg, 30mg, 40mg, 50mg, 60mg and 70mg dosage form (Claims 25–30). Commonly, there are separate claims to a tablet form, injectable form, perilingual form etc. In such cases the inventive concept is exactly the same, and the various claims are intended to facilitate establishing infringement. So far as this point is concerned, it is enough to point out that in practice patents very often have a single overarching inventive concept; it doesn’t matter whether there is a strict legal requirement to that effect. On the redundancy point I agree entirely with Rennie JA’s analysis.
But how does this explain the principle that it is the inventive concept of a particular claim that is at issue? Rennie JA raises the point in [69], but he doesn’t directly answer it. He says that Forthergill J “did not determine the inventive concept based on some ‘generalized concept’; rather it was based on a reading of the claims informed by the specification” [70] and then, after reviewing the inventive concept on the facts of the case, he concluded by saying “In sum, a single inventive concept must flow through a patent, but each claim’s specific inventive concept may be different” [77]. This asserts that different claims may have different inventive concepts even if a single inventive concept flows through the patent, but it is not really an explanation of how that is possible. If a single inventive concept flows through the entire patent, then this suggests that the inventive concept of each claim is the same. The unity of invention principle addresses the redundancy point because it explains why two different claims may have the same inventive concept; it does not explain the inventive concept point, which turns on whether two different claims may have different inventive concepts. If anything, the unity of invention principle seems contrary to the idea that different claims may have different inventive concepts. Do the separate claims to a 30mg form and a 40mg form have different inventive concepts? That does not seem intuitive. It would imply that the dosage amount is an element of the inventive concept, and that seems to conflate the essential elements with the inventive concept. More generally, how do we know whether the different claims have the same inventive concept or different inventive concepts? If different claims have different inventive concepts, how do we reconcile that with the fact that the administrative unity of invention rule means that a patent will normally have a single inventive concept? I’ll say that I don’t find Rennie JA’s explanation of how different claims can have different inventive concepts to be entirely satisfactory.
I'll suggest that the answer is that it is not really correct to say that a single inventive concept “must” flow through a patent. The Act is express that the unity of invention requirement is procedural and administrative rather than substantive; while s 36(1) provides that “[a] patent shall be granted for one invention only,” it goes on to say that “in an action or other proceeding a patent shall not be deemed to be invalid by reason only that it has been granted for more than one invention.” Historically, it was not unusual for applicants to put more than one invention in a single application as a means of avoiding fees, and the purpose of the unity of invention requirement in the Act is simply to ensure that the fees are adequate in light of the time the examiner must spend on the application. Thus it is clear that a valid patent may have more than one entirely different invention concept. Because of the administrative unity of invention requirement, patents normally appear to have a single unifying inventive concept when granted. But when the applicant, and the examiner, are wrong as to what has actually been invented, it may turn out that different claims have different inventive concepts. I'll discuss this at more length in my next post.
A final problem is that the holding that “[t]he inventive concept, properly construed and applied, remains the end point for the obviousness inquiry” [65] is not entirely easy to reconcile with the text of s 28.3, which provides that it is “[t]he subject-matter defined by a claim” that must not be obvious. On its face, this indicates that it is what is claimed that is the end point for the obviousness inquiry. Rennie JA remarked that “Section 28.3 of the Patent Act does not displace the common law test for obviousness” [65]. With respect, that is doubtful, given that the FCA in Hospira 2020 FCA 30 [86] held that s 28.3 had changed the common law, so that now all prior art may be considered in an obviousness attack, not just that which would have been discoverable in a reasonably diligent search. Moreover, if it was not intended to change the law, it presumably reflected what the legislators understood the prior law to be. None of this means that s 28.3 changed the common law with respect to the end-point of the obviousness inquiry, but it does mean that the textual argument cannot simply be dismissed out of hand.
The Vyvanse decision is an ambitious attempt to develop a framework for assessing obviousness. The central holding is that “The inventive concept, properly construed and applied, remains the end point for the obviousness inquiry” [65], and the inventive concept cannot always be determined from the claims. I agree with these points. But the decision leaves a number of questions unanswered:
● Is the ultimate focus of the obviousness inquiry the inventive concept (as indicated by Vyvanse) or what is claimed (as suggested by Ciba and the Act)?
● What is the inventive concept?
● When, if ever, can the inventive concept be determined from the claims?
● Is the inventive concept to be ascertained solely from the specification?
● When do different claims have the same inventive concept? When do they have different inventive concepts?
I’ll try to address these puzzles in my next post.
1.9 Footnotes
1) Rennie JA cites several recent decisions of the FCA which took into account the properties of the claimed subject-matter [74], and the point has long been established. Essentially the same point, that the claims themselves do not necessarily define what is inventive, was made by Jenkins J in May & Baker (1948) 65 RPC 255 (Ch). (This is the same May & Baker that is famous for being the genesis of the doctrine of sound prediction, but on another issue.) The patent claimed processes for the manufacture of certain sulpha drugs. The processes themselves were known, as were the starting materials, and there was no inventive step in making the drugs in question by the claimed processes (295); if a skilled person were seeking to make the drugs in question, it would have been obvious to them to use the claimed processes (279). The “essence of the invention” was not the process itself, but the usefulness as antibiotics of the compounds thereby produced:
my great secret, my discovery, is that these particular known substances A and B when combined do not merely produce a new substance answering the chemical description C (which according to accepted chemical theory was a foregone conclusion) but produce in the shape of C a remarkably valuable drug. (295)
This discovery could not be discerned from the claims, which made no mention of the usefulness of the compounds, but only the compound itself and the process for making it. This reasoning was quoted with approval by the SCC in Comm’r of Patents v Ciba [1959] SCR 378, 383 in holding that processes for the production of useful compounds were new. The same passage was also relied on in Shell Oil [1982] 2 SCR 536, 546, in holding that a new use for a known compound is patentable.
Note that at one time compound claims were considered controversial, on the basis that the scope of the claims should be limited to the compound when produced by a particular process. Even under that restriction, there was no requirement that the process itself be inventive—the validity of the claim rested on the properties of the compound, not the difficulty of manufacture: Comm’r of Patents v Ciba [1959] SCR 378. That restriction was eventually lifted. In the UK, this was on the recommendation of the Swan Committee Report (Cmd 7206), on the basis of the argument that “the real invention lies in the discovery of a new substance, with new and useful properties, and that the process of manufacture often involves little novelty in itself”: see Lundbeck [2008] EWCA Civ 311 [43]–[46].
2) Pelletier JA mistakenly attributed this statement to Lord Hoffman in Conor Medsystems v Angiotech [2008] UKHL 49; this does not affect the point.
3) This is suggested by Apotex’s argument. The inventive concept in this case was “a sustained release formulation of a therapeutically useful dose of amphetamine that is resistant to abuse” [102]. Apotex argued that Fothergill J erred in holding that the inventive concept could be grasped without difficulty, without an analysis “of whether any of the claims’ inventive concepts could be readily identified from the wording of the claim itself” [62], and without explaining why it was readily apparent [63]. Apotex argued that none of the asserted claims related to sustained release or abuse resistance, though there were claims were the sustained release profile (Claims 16-21 ) and or abuse resistance (Claim 45) were elements, and consequently, “the specific inventive concept(s) of some of the claims were rendered redundant, giving rise to a palpable and overriding error” [63]. This implies that Apotex’s argument was that the inventive concept of those particular claims could be grasped directly from the claims. (Whether that results in impermissible redundancy is a different question.)
4) See 74, stating “we respectfully agree with the learned judge that the inventive concept of the patent is the free-sail concept,” and at 66 holding that the inventive concept “ cannot be expressed better than by adopting the language of the learned judge when he said: ‘The specification is disclosing the concept and method of constructing a wind-propelled craft carrying a sail which is mounted on a mast free to move in any direction. . . .” (My emphasis).
5) The inventive concept was the “free-sail” concept, reflected in the following element of GB1258317 Claim 1: “an unstayed spar connected to said body means through a joint which will provide universal-type movement of the spar in the absence of support thereof by a user of the vehicle” (63). Other elements included eg “a sail,” “a pair of arcuate booms” etc.
6) Both Rennie JA in Vyvanse and Pelletier JA in Ciba cite Unilever as being a decision of the EWCA. It is in fact a decision of Jacob J, as he then was, in the Patents Court. The confusion likely arises because in Pozzoli [2007] EWCA Civ 588 [17], Jacob LJ, by that time elevated to the Court of Appeal, quoted Unilver, remarking “[A]s I pointed out in Unilever. . .”, without indicating it was a Patents Court decision.
*UPDATE: Rennie JA’s observation is supported by US law, in which as I understand it, the properties of a compound are considered to be part of the compound itself, so that the question is not whether the compound was structurally obvious, but whether it would have been obvious that the compound had the particular beneficial property. The leading case is In re Papesch 315 F2d 381 (Fed Cir 1963), in which the claimed compounds were homologues of prior art compounds which could be prepared using to standard methods. The examiner had refused the claims on the basis that in view of the known general relationship of homology, the disclosure of a chemical structure renders its homologues obvious (385) and the Board affirmed. The CCPA reversed, saying that the decision of the board “rests on one fundamental error of law, namely, the failure to take into consideration the biological or pharmaceutical property of the compounds.” “From the standpoint of patent law, a compound and all of its properties are inseparable; they are one and the same thing. The graphic formulae, and the chemical nomenclature, the systems of classification and study such as the concepts of homology, isomerism, etc., are mere symbols by which compounds can be identified, classified, and compared. But a formula is not a compound and while it may serve in a claim to identify what is being patented, as the metes and bounds of a deed identify a plot of land, the thing that is patented is not the formula but the compound identified by it. And the patentability of the thing does not depend on the similarity of its formula to that of another compound but of the similarity of the former compound to the latter. There is no basis in law for ignoring any property in making such a comparison" (391).
At paragraph 27 of Apotex v. Janssen, 2021 FCA 45, a different panel seems to have adopted a similar position to Rennie JA ruling that advantages not specifically claimed can be taken into account when assessing obviousness.
ReplyDeleteThanks for pointing that out - that's helpful for my next post as well.
DeleteVery interesting and helpful series of posts. To your point at 1.3.1, [67] of Vyvanse suggests that Rennie JA interpreted Sanofi "... or if that cannot readily be done, construe it [the inventive concept]" rather than "construe it [the claim in question]". In Ciba it seems to have been interpreted the other way.
ReplyDelete