Monday, January 11, 2021

Reconciling the Two Essentiality Tests in Free World

Allergan Inc v Sandoz Canada Inc 2020 FC 1189 Crampton CJ

2,507,002 / silodosin / RAPAFLO / NOC

In this NOC decision Crampton CJ held that Allergan’s 002 patent was not invalid for obviousness, but it was not infringed by Sandoz’s silodosin formulation. These holdings turned largely on the facts, though Crampton CJ’s discussion raised several interesting legal issues, the most important relating to the interpretation of s 53.1, which permits the use of prosecution history in claim construction in certain circumstances. I’ll discuss this in a subsequent post. This post will provide the background and discuss claim construction, which was determinative of infringement. In particular, I’ll consider the question of how to reconcile the two apparently different tests for essentiality set out by the SCC in paragraph 55 of Free World 2000 SCC 66. Really this is a discussion of a point that was raised more directly by Locke J in Shire v Apotex 2016 FC 382, though I didn’t discuss it when I blogged about that case.

The 002 patent relates to a solid dosage form of silodosin. Silodosin was known to be useful in treating dysuria associated with a kind of benign prostate tumour. However, it was difficult to formulate in a solid dosage form, in part because of its potent adhesive properties. The 002 patent reflects the inventors’ solution to this problem. There were two independent claims at issue. Claim 1 was to a capsule comprising a granule “prepared by wet granulation” of silodosin and various excipients, with a specified dissolution profile [63]. Claim 6 was a method claim comprising granulating silodosin, and other excipients by a “wet granualtion process” [63]. The meaning of the terms was not disputed [64], nor was it disputed that the Sandoz product included all the specified ingredients [21]. The only dispute was as to whether the wet granulation elements were essential [64].

Crampton CJ’s starting point is that all elements are presumed to be essential: [46]. He then adopted at [67] the two-pronged approach to essentiality set out by the SCC in Free World 2000 SCC 66 at the beginning of para 55:

For an element to be considered non-essential and thus substitutable, it must be shown either

(i) that on a purposive construction of the words of the claim it was clearly not intended to be essential, or

(ii) that at the date of publication of the patent, the skilled addressees would have appreciated that a particular element could be substituted without affecting the working of the invention

The difficulty is that in the same paragraph, only two sentences later, the SCC quoted with approval the three part-test set out by Hoffmann J in Improver [1989] RPC 69, which was intended to summarize Catnic:

(i) Does the variant have a material effect upon the way the invention works? If yes, the variant is outside the claim. If no: –

(ii) Would this (i.e.: that the variant had no material effect) have been obvious at the date of publication of the patent to a reader skilled in the art? If no, the variant is outside the claim. If yes: –

(iii) Would the reader skilled in the art nevertheless have understood from the language of the claim that the patentee intended that strict compliance with the primary meaning was an essential requirement of the invention? If yes, the variant is outside the claim.

As Locke J pointed out in Shire v Apotex 2016 FC 382 [135], these two tests are not entirely consistent. The first prong of the two-part test corresponds roughly to the third Improver question, and the second prong corresponds roughly to the first and second Improver questions, but the two-part test is disjunctive while the Improver questions are conjunctive [135]. The problem, then, is to try to understand what the SCC intended. Locke J suggested that the SCC probably intended the two-part test to be conjunctive, so that it would be necessary to establish both elements in order to establish that the element at issue is not essential [137].

Crampton CJ acknowledged this uncertainty [48]. Fortunately, since Crampton CJ’s answers to both questions were the same, nothing turned on the issue, and he did not explore the matter further [48]. The point is nonetheless of general importance, and I since didn’t address in my post on Shire v Apotex, I’ll take the opportunity now.

As a preliminary point, in light of Whirlpool 2000 SCC 67, the companion to Free World, Crampton CJ glossed the first prong of the two-part test as requiring [67]:

(i) that on a purposive construction of the patent disclosure and claims as a whole, those elements were not intended to be essential (Whirlpool [2000 SCC 67] at paras 48, at 49(g))

That is, we may look to the entire specification, not just the claims, as the formulation in Free World [55] might suggest. This must be right: if all claim elements are presumed to be essential, and we can only look to the claims, we would inevitably have to conclude that the element is essential, so the first prong doesn’t make sense unless we can look to the specification as a whole. While Free World does not say this explicitly, it does refer to “purposive” construction of the claims, and this references the discussion in Whirlpool, which makes it clear in the passages cited by Crampton CJ, that a purposive construction is based on a “knowledgeable reading of the whole specification.”

With that in mind, we can turn to the question of how to best understand paragraph 55 of Free World. As Locke J also noted, it is unlikely that the SCC intended this difference [136]. So the question is how to read this paragraph in a way that best reconciles these statements in a manner that is harmonious with the decision as a whole, recognizing that perfect textual consistency will not be possible. In particular, in Whirlpool [45]–[50] the SCC stated expressly that it was affirming an established approach to claim construction, and essentiality in particular, so any interpretation of this paragraph must be consistent with the prior leading cases, especially those specifically approved by the SCC as examples of the essentiality analysis.

In Locke J’s view, the SCC likely intended a conjunctive test, so that to establish non-essentiality, a patentee would be required to satisfy both prongs of the two-part test [137], notwithstanding the conjunctive “or” used by the SCC. I agree with Locke J that the SCC probably did not intend a disjunctive test (see below). But I don’t think interpreting it as a conjunctive test is enough; I will suggest we also need to interpret “clearly not” as meaning “not clearly.”

The difficulty with simply interpreting the two-part test conjunctively is that would not be consistent with paradigmatic cases, such as McPhar [1956-60] ExCR 467 and Van Der Lely [1963] 80 RPC 61 (HL), which the SCC in Whirlpool cited with approval at [46], [47].

The first prong is important in a conjunctive test only if we have concluded under the second prong that the element obviously does not affect the way the invention works. Even so, under a conjunctive test, the element will only be non-essential if we can conclude that the element was “clearly not” intended to be essential. This implies some kind of positive indication in the specification that the claim element does not affect the way the invention works, which implies that the inventors turned their mind to the issue. But if the invention does not affect the way the invention works, and the inventors knew that when the drafted the patent, why did they include the contentious element in the claim in the first place? In cases such as McPhar and Van der Lely, the inventors included the contentious element because they never turned their mind to the alternative implementation; had they done so, they would have recognized the obvious and omitted the contentious claim element. Thus a conjunctive test is likely to exclude the possibility of a finding of non-essentiality exactly when it is typically used.

A closer look at Thorson P’s decision in McPhar [1956-60] ExCR 467 illustrates the point. The invention in McPhar was a portable electromagnetic mineral prospecting device with receiving and transmitting coils. The general concept was known, and the problem was ensuring frequency stability of the generator and hence of the electromagnetic waves. The prior art solution was to use a large generator with a governor. This stabilized the frequency but was very heavy and required a team to carry it in the field. The inventors solved the problem of lightweight frequency stability with a radical departure from the prior art, referred to as “load control” which loaded the generator with the transmitting coil itself.

At the same time, it was well known that the transmitting coil had to be oriented vertically. A convenient way of doing so was to suspend the coil to hang freely from a tripod, allowing gravity to hold it in the vertical position. This method – a means to “suspend said transmitting coil to hang freely” — was specified in the claim at issue. There was nothing new about this method (540). The defendant invented around the patent by bolting the transmitter coil rigidly to the top of a mast on a base plate and then adjusting the plate with leveling screws to achieve a vertical orientation. There was nothing new about this method either. It was inferior to the standard method that was described in the patent, as it was more difficult to set up, and was evidently only used to design around the patent.

Thorson P concluded that while a coil that was bolted rigidly from below could be said to be “suspended,” it could not be said to “hang freely” (521). The question therefore was whether the “hang freely” element was essential. Thorson P held that it was not. Neither verticality nor the means of achieving it were part of the inventive concept. It was well known that the transmitter coil had to be maintained in the vertical plane (521, 529, 540), and there were a variety of known means of doing so, including the means used by both the inventors and the defendants (540). While Thorson P did not, of course, explicitly address the various prongs of the modern test, he did specifically find that the “hang freely” element could be replaced “without making any real difference to the invention” (540), and it is implicit that this would have been readily apparent to a skilled person, as he found that there was nothing new or inventive in the means specified and that there were various other means that could have been used (540). So, it is clear that the second prong of the two-part test was satisfied.

The first prong of the two-part test asks if anything in the specification indicates that the “hang freely” element was “clearly not” intended to be essential. Thorson P never undertook this inquiry; his analysis was based entirely on what is effectively the second prong of the test: see 540-42. This in itself is an indication that a conjunctive two-part test is not consistent with McPhar. Moreover, as I read it, there was no such indication in the 484,515 patent. The patent described the “principle feature” of the invention as consisting of various features, including “a support for the transmitting coil to enable it to be suspended to assume a vertical position but adjustable in azmuthal [sic] position” (p3). But as noted, Thorson P held “suspended” to include the variant at issue, so this doesn’t say that the “hang freely” element is or is not essential. The drawings show an embodiment in which the transmitting coil hangs freely, and the discussion of that embodiment in the specification includes some corresponding language (though not much). Because this is only a discussion of a preferred embodiment, it cannot be taken to imply that the “hang freely” element is essential, but it certainly does not clearly indicate that it is not essential. So, while the “hang freely” element is barely discussed, alternative means of achieving verticality are not discussed at all. In the absence of any discussion of alternative means, I can’t see how it could be said that the specification indicates that the “hang freely” means is “clearly not” essential.

So, if the two-part test is conjunctive, McPhar would be considered to have been wrongly decided. Since the SCC expressly approved McPhar, it seems very unlikely that this is what the SCC intended.

I’ll suggest that one more tweak would save the conjunctive test: we could read “clearly not” in the first prong to mean “not clearly.” That is, we might interpret the first prong as requiring that there was no indication that the element was intended to be essential, instead of requiring a clear showing that it was not intended to essential. In effect, this reads the first prong as being exactly the same as the third Improver question. That view is entirely consistent with McPhar. It is trivial to conclude that the “hang freely” element was not clearly intended to be essential, as it was barely discussed at all.

The main argument against this reading is textual. It would require two changes to the two-part test: we would have to read “or” to mean “and” and we would have to read “clearly not intended to be essential,” as meaning “not clearly intended to be essential.” But as Locke J noted, it is simply not possible to read the paragraph consistently while adhering strictly to the text. Something somewhere has to give, at least a little; if we’re willing to read “or” as “and,” reading “clearly not” as “not clearly” is only a modest step.

On the other hand, there are two important arguments in favour of this interpretation. First, it is consistent with the prior caselaw that was approved in Free World and Whirlpool, such as McPhar. Because the SCC stated expressly that it was affirming the traditional test established in that caselaw, this is a strong point in favour of this interpretation.

Further, on that interpretation the two part test reduces exactly to the Improver questions, thus eliminating the inconsistency between the two tests set out in [55]. This is an important consideration. As Locke J pointed out:

[136] It seems unlikely that the SCC intended this apparent difference. Its decision does not acknowledge any inconsistency between the Improver questions and its own test for determining essentiality. Nor does the SCC suggest any disapproval of the Improver questions. In fact, the SCC clearly relies in Improver.

I entirely agree. I’d also point out that in the very same sentence that it articulated the two-prong test, the Court explained the second prong by reference to Improver. The inescapable conclusion is that the SCC intended to articulate the same test in slightly different ways. We therefore need to tweak one of these tests slightly to make them consistent with each other. But we also know that the SCC intended its test to be consistent with the prior cases, and we know that the SCC considered the Improver questions to be consistent with the prior cases: “they encapsulate the heart of Lord Diplock's analysis, and have been endorsed in subsequent English cases” [56]. So, in my view, it makes sense to tweak the two-part test to read it in a way that is consistent with the Improver questions.

Now, as Lord Hoffmann himself pointed out in Kirin Amgen [2004] UKHL 46 [66], the Improver questions are a “mangle,” and it would be very reasonable for the SCC to have sought a simplified way of making the same point. I suggest that the two-part test in Free World was intended as a simplification of the Improver questions, and the best interpretation of the two-part test is one that is consistent with the Improver questions.

To a considerable extent the SCC was successful in simplifying the Improver questions. The first two Improver questions rely on extrinsic evidence, while the third is purely extrinsic—it is essentially a matter of construction. So the two-prong test helpfully reduces the Improver questions to an intrinsic prong (the first) and an extrinsic prong (the second).

Further, the second prong of the two-part test, which asks whether a skilled person would have appreciated that a particular element was not material, strikes me as being a consistent simplification of the first two Improver questions. The idea is that a skilled person starts off expecting that every claim element is essential, but if they appreciate that one element makes no difference at all, then they would consider it to be non-essential. The first Improver question asks whether the variant in fact has a material effect; if the answer is that the variant in fact has a material effect, it is reasonable to conclude that a skilled person would have appreciated this, and would not have thought (wrongly) that it was immaterial. But even if variant in fact had no material effect, a skilled person would not necessarily appreciate that fact; for example, it might well require considerable investigation to determine that two approaches that seemed different on the surface ultimately worked through the same mechanism. That is the point of the second Improver question. So, the second prong of the two-part test is the ultimate inquiry, and this can be broken down into the two separate inquires reflected in the first two Improver questions.

Thus, the two prong test simplifies the Improver questions by separating them into intrinsic and extrinsic inquiries and by posing a single overarching question that captures the two separate extrinsic inquiries in the Improver questions.

So far, so good. But another problem is that the Improver questions are difficult to follow because of their flow-chart structure (“If yes, the variant is outside the claim. . . . If no, the variant is outside the claim. . . . If yes, the variant is outside the claim”). I suspect the SCC was looking for a cleaner structure. The third Improver question is framed in positive terms—in effect, does the specification say it was essential?—and is (more or less) conjunctive with the first two. So it looks like the SCC flipped the question—does the specification say it was not essential—and made it disjunctive, with the idea that a double negative gets you back to the same place we started. Unfortunately, that’s not quite true. So the SCC started with a mangle and added another mangle, leaving us with a mangle2, so to speak.

That is a speculative reconstruction of how the two-prong test ended up incorporating a “or,” rather than an interpretation of the passage as such. But the essential point is Locke J’s observation at [136], that the SCC could not have intended to articulate two inconsistent tests in the very same paragraph. We have to interpret the two tests in a consistent manner. As a matter of strict text, the two are not consistent, so something has to give. If we consider the two-prong test was intended to simplify the structure of the Improver questions while maintaining the substance, then the question is how to tweak the two-prong test to it actually achieves that aim. I would suggest that the minimal change is to read the test conjunctively, as Locke J suggested, and also reading “clearly not” as “not clearly.” Just one of the tweaks alone will not suffice.

Finally, I’ll turn to the issue of why we should not read the two-prong test disjunctively. On a disjunctive reading, the first prong is redundant if the second prong has been met. The difficulty arises when it would not be apparent to a skilled person that the variant works the same way. In this case for example, dry processes are in fact materially different from wet granulation because electrostatic charges are increased, more lubricants are needed which reduces dissolution time, etc [106]. Further, the formulation requires a combination of many different excipients which interact in unpredictable ways [206]–[208], so that even if wet granulation made no difference this would not have been apparent to the skilled person. On a disjunctive reading of the two-prong test wet granulation might nonetheless be non-essential under the first prong, if the specification stated clearly that wet granulation was not essential, even though it in fact had a material effect on the way the invention worked. Fortunately, on the facts the specification actually reinforced the same conclusion: [81]–[83], [88]. But a disjunctive test implies that in some cases, an element may be non-essential because the specification says so, even though the skilled person would know that the variant in fact has a material effect.

There are two problems with this. First, there is no support whatsoever for this in the case law. I have, at one time or another, reviewed every case cited by the SCC in Free World and Whirlpool, and many others besides, and I never seen a case where an element that did in fact have a material effect on the invention was found to be non-essential. As noted, the SCC in Free World and Whirlpool expressly stated that it was reaffirming a long-standing doctrine, albeit with some clarification and elaboration, not creating an entirely new principle which had never before been applied.

The second problem is related. If the element in fact has a material effect on the way the invention works, and yet the element is considered non-essential under the first prong, then the claim would almost certainly be too broad in one way or the other; ie it would claim subject matter that is not useful, or anticipated, or not enabled by the disclosure. I am confident that there are no prior cases in which a finding of non-essentiality led to the claim being construed in a way so as to result in invalidity. To back away from these tests for a second, an element is considered “essential” to the claim because it is, well, essential to the invention. Many of the early cases referring to “essential” elements were decided in the context of the old “pith and marrow” approach to infringement, where the question was really whether the infringer had taken the inventive concept. The essential elements were those that defined the pith and marrow of the invention, and an element that was in fact central to the invention would never be considered non-essential.

In many cases, such as this one, the first prong simply leads to the same result as the second prong. This is normal; when an element in fact has a material effect on the way the invention works, it is not very surprising that the disclosure would say so. But a disjunctive test implies that the intrinsic examination of the disclosure can make a claim non-essential, entirely independently of the extrinsic evidence. There simply aren’t any cases that I am aware of in which an element was found to be non-essential even though it had a material effect on the way the invention worked. It is inconceivable that the SCC intended to signal a departure from prior law simply by the use of the word “or”. For that reason, I agree with Locke J that the two-prong test should not be read disjunctively.

In summary, even though the two tests set out by the SCC in paragraph 55 of Free World are not consistent in purely textual terms, a purposive analysis of the decision means that we have to read them in a consistent manner. I would suggest that they best way to do this is to read the two-part test as being consistent with the Improver questions, and the minimal way to so this is to read the test conjunctively and to also read “clearly not” as meaning “not clearly.”

No comments:

Post a Comment