Friday, January 15, 2021

The Limits of Anticipation by Speculation

Teva Canada Innovation v Pharmascience Inc 2020 FC 1158 Kane J

2,702,437 / 2,760,802 / glatiramer acetate / Copaxone / Glatect / NOC

The patents at issue in this case relate to the use of glatiramer acetate for the treatment of multiple sclerosis [MS]. Glatiramer acetate was known to be effective in treating MS, with a standard dose of 20mg daily, administered by subcutaneous injection. The 437 patent claims the early use of glatiramer acetate for the purpose of delaying the development of MS, while the 802 patent claims a dosing regime of 40mg three times a week. The decision turned largely on the facts (to the tune of 936 paragraphs!), but a few points of interest were raised, particularly in the discussion of the 437 patent, which Kane J held to be novel, but invalid for obviousness. Kane J's holding that the 802 patent was not invalid for obviousness or lack of utility, and that it would be infringed by Pharmascience's 40mg product, turned on the facts.

MS is a disease of the central nervous system in which the patient’s own immune system attacks the myelin of the CNS. The early course of the disease is characterized by intermittent attacks, which include episodes of numbness, muscle weakness and spasticity, etc [73]. The disease also gives rise to lesions on the CNS resulting from destruction of the myelin sheath [71]. Patients who present with a single clinical attack having features typical of MS were considered to have a “clinically isolated syndrome” (CIS) [77]. CIS patients are at elevated risk for MS itself (also referred to as “clinically definite MS” or CDMS), but not all CIS is caused by MS, so CIS patients are not considered to actually have MS.

The inventive concept of the 437 patent is that glatiramer acetate could be used to delay the onset of clinically definite MS in CIS patents at risk of developing MS [569]. Accordingly, Claim 1 was to glatiramer acetate “for use in delaying the onset of clinically definite multiple sclerosis,” in a patient who presents with at least one lesion “suggestive of multiple sclerosis” and who is “at risk of developing” clinically definite MS, and “prior to development of clinically definite multiple sclerosis” [251].

The claim construction debate turned on the fact that around the time of filing, in 2007, the criteria for diagnosing MS were changing. The older Poser criteria relied on clinical signs, ie two clinical attacks. The new McDonald criteria were developed to take advantage of the use of MRI for diagnosis using direct detection of lesions characteristic of MS with less reliance on clinical signs. The result is that some patients who would not have been diagnosed as actually having MS using the Poser criteria, would be diagnosed as having it using the McDonald criteria. The 437 patent was apparently developed with the Poser criteria in mind and there was a long and involved argument as to whether the patients with MS according to the McDonald criteria would be included in the claim; if so, the claim would not be novel, as the use of glatiramer acetate for treating MS was known [305].

I must admit that I was having trouble following the argument and its significance, and I was relieved when Kane J stated that “In my view, far too much time and effort was devoted to debating the meaning of the terms, the evolving diagnostic criteria and the scope of the patients covered by the claims” [306] . All the experts construed the claims in accordance with their plain language, which did not specifically refer to either set of criteria by name, and “the claims mean what they say” [307]. Kane J held that on a plain reading of the claims they were directed to a patient who did not already have a confirmed diagnosis of MS, and so did not include patients who already met the McDonald criteria [329]. This reading seems eminently sensible to me. The result of this claim construction holding was to negate both anticipation based on the Pinchasi 2007 prior art [454] as well as the Gillette defence [396], [403].

In light of this conclusion on claim construction, the more interesting anticipation argument was based on the Karussis 2006 prior art. This was a report of an International Working Group for Treatment Optimization in MS [458]. The report suggested the use of therapies approved for treatment of MS itself, which prominently includes glatiramer acetate, as a possible treatment for patients at risk of developing MS (ie, those with CIS), but acknowledged that this was an inference and was not based on evidence [464], [465]. So, the Working Group concluded that “as [other therapies] and glatiramer acetate have been shown to be effective in [the most common form of diagnosed MS], and a patient with a CIS is highly likely to progress to MS, it is reasonable to propose early treatment with [therapies approved for treatment of MS] at the doses approved for MS, even though there is not yet evidence to support this approach” [465]. The question is whether this and similar statements anticipated the claims of the 437 patent.

This raises the issue of what I have called “anticipation by speculation,” as raised in Hospira 2020 FCA 30 and Biogen v Taro 2020 FC 621, which held that similarly speculative statements were anticipatory. As discussed here and here, my view is that Hospira and Biogen v Taro stretched anticipation to the very limit, and perhaps beyond. In this case, Kane J held that Karussis 2006 was not anticipatory. This is not to say that Kane J’s holding is necessarily inconsistent with Hospira and Biogen v Taro, which Kane J considered and distinguished on the facts: [479]–[482]. I won’t go through all the particular grounds, though I will say that the distinctions she draw struck me as generally reasonable; as always in borderline cases, some of the distinctions to be drawn will necessarily be fine and turn on the particular facts. The broader point is that Kane J’s holding illustrates that anticipation by speculation can only be pushed so far and the line is somewhere in this general vicinity.

One general observation made by Kane J is worth pointing out (my emphasis):

[484] In my view, if Karussis 2006 is anticipatory art, what would prevent a group of experts in any field from having discussions and speculating on the positive outcome of a clinical trial in progress, developed and implemented by others, in order to later argue that their speculation that the trial in progress would be successful anticipated the claims of a patent that are based on such clinical trials. This approach seems to be inconsistent with the objectives of the experimental use exemption established in the common law.

[485] As noted by Teva, the law recognizes that there is no disclosure for the purposes of anticipation where a prior use is experimental. If a patent cannot be anticipated by its own clinical study, in my view, it cannot be anticipated by the speculative recommendations of a group of experts who discuss the state of MS treatment, note the lack of evidence, and note that trials are underway.

Kane J then went on to hold that Karussis 2006 did not in any event satisfy the enablement branch of anticipation: [486]–[493].

All that was for nought, so far as Allergan is concerned, as Kane J then went on to find the 437 patent invalid for obviousness on an obvious-to-try analysis. This is likely to be common in cases where a clear suggestion to try the invention can be found in the prior art, especially since the holding in Hospira 2020 FCA 30 that the state of the art for purposes of an obviousness attack includes all prior art. Exactly where we draw the line on anticipation is now less important, because if an obscure piece of prior art does not quite reach the level of anticipation, it can still be used as the basis for an obviousness attack. With that said, it does not appear that the Hospira rule made any difference in this case; while Karussis 2006 was not common general knowledge, it reflected several studies that were [573].

I’ll finish with two quick points on obviousness.

First, Kane J clearly stated that the mere fact that it is not self-evident that the invention will work, before any experiments are undertaken, does not mean that it is not obvious.

[593] Dr. Selchen’s opinion that the invention was not self-evident appears to conflate the obvious to try test with the factor or consideration that informs the test, i.e., whether it was self-evident that it would work, in the sense of being effective for CIS patients to delay the onset of MS and achieve the other claimed benefits.

I entirely agreed with Kane J on this, but it is a point that continues to pop up and cause confusion, as it did for Dr Selchen in this case.

Second, Kane J recognized and applied the “golden bonus” rule, usually traced back to Hallen v Brabantia [1991] RPC 195, 216 (CA), which says that if an invention is obvious for one reason, it does not become non-obvious merely because it has some non-obvious benefits (the "golden bonus"):

[594] Teva’s submission – that the therapeutic benefits or outcomes of the ‘437 Patent were not obvious or expected – does not diminish that the invention was obvious to try. In Janssen [2015 FC 184] at para 100 [see here], the Court explained, that “if a patentee obtains a workable formulation, the later discovery of one of its inherent characteristics does not add anything inventive to what had already been discovered: see Alcon Canada Inc. v Apotex Inc., 2012 FC 410 at para 45, [2012] FCJ No 1707 (QL).” In the present circumstances, once the POSITA pursued the use of glatiramer acetate for CIS patients, the claimed benefits would follow.

As Kane J notes, this is not an entirely new development in Canadian law, but it is nonetheless welcome to see the point expressly recognized and affirmed. I believe that this is the first time that the phrase “golden bonus” has actually been used in a Canadian case [510].

Wednesday, January 13, 2021

Prosecution History May Not Be Used to Rebut Representations Made by a Licensee

Allergan Inc v Sandoz Canada Inc 2020 FC 1189 Crampton CJ

2,507,002 / silodosin / RAPAFLO / NOC

Section 53.1 of the Patent Act permits the use of prosecution history in claim construction in certain circumstances. In this decision, Crampton CJ held that s 53.1 only permits use of prosecution history to rebut a representation made in litigation by a patentee, not representations made by a licensee. This is a truck-sized loophole in s 53.1, and Crampton CJ noted the prosecution history in question “provides a glaring example of the mischief that is implicitly permitted by the current wording” [135]. Nonetheless, Crampton CJ’s analysis of the provision is entirely compelling.

As noted in Monday’s post, the key issue on infringement was whether “wet granulation” was an essential element of the claims in question. Allergan argued it was not. Sandoz wished to rely on representations made during patent prosecution to rebut this assertion. The difficulty is that s 53.1 permits prosecution history to be introduced to rebut any representation “made by the patentee in the action.” Allergan is not the patentee; it is the exclusive Canadian licensee [1] and is the “first person” under the PM(NOC) Regulations [5]. The patent owner is Kissei Pharmaceutical and the patent was prosecuted on Kissei’s behalf [1], [3]. Kissei was also a defendant to this NOC action, joined pursuant to s 6(2), but crucially, Kissei took no position on the infringement issue, though it adopted Allergan's submissions relating to validity. (Kissei did not appear during the trial [7].)

So, Allergan’s argument was that the prosecution history could not be introduced because it can only be used to rebut a representation by the “patentee” and the patentee, Kissei had not made any representations. The question facing Crampton CJ was therefore whether s 53.1 should be interpreted to allow prosecution history to be used to rebut representations made by a licensee.

As noted, I find Crampton CJ’s reasons on this point to entirely convincing. I don’t have anything to add, so I will simply reproduce the key passage (with some references omitted):

[124] In my view, none of the arguments advanced by Sandoz can overcome the plain wording of subsection 53.1(1), a contextual reading of the Act or the jurisprudence in respect of the definition of the word “patentee” in section 2 of that legislation.

[125] It is trite law that “the words of a statute must be read ‘in their entire context and in their grammatical and ordinary sense harmoniously with the scheme of the Act, the object of the Act, and the intention of Parliament’”.

[126] The “chapeau” in subsection 53.1(1) plainly limits the scope of that provision to permitting certain written communications to be admitted into evidence to rebut any representation made by the patentee in an action or proceeding, in respect of the construction of a claim in a patent that is at issue in the action or proceeding. In the present proceeding, it is admitted that the patentee is the defendant Kissei, which has not made any representation to the Court with respect to the construction of the ‘002 Patent. Accordingly, in the absence of any clear indication elsewhere within the scheme or object of the Act that Parliament intended to word “patentee” to include a licensee of a patent, subsection 53.1(1) cannot be invoked in this proceeding.

[127] I do not agree with Sandoz’s position that a licensee falls within the meaning of the word “patentee”, as defined in section 2 of the Act, namely, “the person for the time being entitled to the benefit of a patent”. This position was specifically considered and rejected in Electric Chain Co of Canadas Ltd v Art Metal Works Inc, [1933] SCR 581 at 586-587 [Electric Chain]. The effect of that decision was that a licensee had no right to be a party to an infringement action in Canada. As a result, what is subsection 55(1) was added to the Act: American Cyanamid Co v Novopharm, [1972] FC 739 at paras 23-24 (FCA) [American Cyanamid]. That provision, which has undergone some minor amendments that are not germane for the present purposes, states:

“A person who infringes a patent is liable to the patentee and to all persons claiming under the patentee for all damage sustained by the patentee or by any such person, after the grant of the patent, by reason of the infringement.”

[128] It has since been confirmed that a person who is a licensee under a patent is a “person claiming under” the patentee within the meaning of subsection 55(1): Armstrong Cork Canada v Domco Industries Ltd, [1982] 1 SCR 907 at 914; American Cyanamid, above, at paras 31-32.

[129] What is instructive for the present purposes is that while the Act was amended to permit a licensee to sue for infringement, the definition of “patentee” was not amended following the interpretation that it was given in Electric Chain, above.

[130] Moreover, given that Parliament included the words “the applicant for a patent” in clause 53.1(1)(b)(i), but not in the “chapeau” of subsection 53.1(1), it can be inferred that (i) Parliament was aware of the distinction between a patentee and a person who is not the patentee, and (ii) Parliament decided to strictly limit the scope of the “chapeau” to a person who is a patentee.

[131] This interpretation of Parliament’s intent finds some support in the legislative history. In particular, in the submission that IPIC made to the [Senate Banking, Trade and Commerce Committee], discussed at paragraph 123 above, IPIC noted that the language of section 53.1 created a “loophole” that would permit “a patentee to circumvent the operation of this section by acting through a licensee”: IPIC Submission, above, at 12. To address this “loophole”, IPIC recommended that subsection 53.1(1) be amended to include “persons claiming under the patentee”. The specific amendment proposed by IPIC is the underlined wording in the chapeau of that provision:

53.1 (1) In any action or proceeding respecting a patent, a written communication or any part of such a communication, may be admitted into evidence to rebut any representation made by the patentee or a person claiming under the patentee in the action or proceeding as to the construction of a claim in the patent if […]

[132] Ultimately, IPIC’s recommendation was not accepted, and Bill C-86 was passed without any change to the language of subsection 53.1(1). This legislative history provides additional support for the view that, at the time Parliament added subsection 53.1(1) to the Act, it was aware of the distinction between a patentee and a person who is not a patentee, yet it chose to limit the scope of section 53.1 to representations made by patentees.

While Crampton CJ did not rely primarily on the last point, I must say that I find it particularly compelling; this very issue was before the responsible Parliamentary committee, along with a very simple fix, and it was not implemented. The inference that it was considered and rejected is strong, which implies that the “loophole” was specifically intended.

[133] In summary, the plain and ordinary meaning of the language in subsection 53.1(1), together with a contextual reading of the Act and the jurisprudence discussed above, support the view that the word “patentee” in subsection 53.1(1) does not include a licensee. The legislative history also provides some additional support for this interpretation. Sandoz has not identified any contextual consideration to support the alternative interpretation that it has advanced.

While the considerations reviewed by Crampton CJ are textual and contextual, rather than purposive, and there is evidently some kind of purposive argument to be made that this loophole frustrates the operation of the provision, Crampton CJ’s textual and contextual analysis is so entirely compelling that I cannot see that any other conclusion would be possible without an impermissible usurpation of the legislative role.

Sandoz also argued that the prosecution history was admissible entirely apart from s 53.1, relying on the FC decision in Distrimedic 2013 FC 1043 at [210]: [116]. Crampton CJ rejected this argument, in part on the basis that Distrimedic is distinguishable [117]. More importantly Crampton CJ also observed that in light of the bar on the use of prosecution history set out in of Free World 2000 SCC 66 [66], Distrimedic has been “overcome” by the enactment of section 53.1 [115], [118]. I understand Crampton CJ to be saying that the use of prosecution history is now confined to the circumstances set out in s 53.1. This strikes me as undoubtedly correct. Any use of prosecution history in Canada now turns entirely on the interpretation of s 53.1 and any subsequent legislative amendments.

Crampton CJ therefore concluded that “given that no representation has been made by the patentee (Kissei) of the ‘002 Patent in the present action, the file prosecution history is not admissible in evidence in this action. It is barred by the prohibition against extrinsic evidence: Free World, above, at para 66" [133].

Crampton CJ’s holding was driven inexorably by the legislation; he did not necessarily welcome it. On the contrary, he observed “in passing,” that the “file prosecution history in question provides a glaring example of the mischief that is implicitly permitted by the current wording of subsection 53.1(1)” [135]. For those interested in looking at the prosecution history themselves, the documents 02507002 - 2011-08-25 - PRO and 02507002 - 2011-01-19 - PRO emphasize wet granulation to overcome various objections.

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].

Wednesday, January 6, 2021

Lear v Adkins Should Not Be Followed

Loops LLC v Maxill Inc 2020 ONSC 5438 Lederer J: Swinton, Penny, JJ revg 2020 ONSC 971 Templeton J

            2,577,109 / Toothbrush

In Monday’s post I reviewed Lederer J’s decision in this case, holding that a no-challenge clause in a settlement agreement is enforceable, notwithstanding the decision of the US Supreme Court in Lear Inc v Adkins 395 US 653 (1969) which abolished the doctrine of licensee estoppel in US law. In Tuesday’s post I explained that the doctrine of licensee estoppel is not based on estoppel at all, but is based entirely on contract law. Absent an express agreement to the contrary, the licensee’s obligation to pay royalties will not normally be contingent on the validity of the patent. This is partly because the courts will not normally imply a warranty of validity into the agreement and partly invalidity will not normally constitute complete failure of consideration, but might instead give rise to a counter-claim for damages. In this post I will step back and consider the bigger question: as a matter of policy, should Lear v Adkins be followed in Canada? Recall that while Lederer J was of the view that holding a no-challenge clause in a settlement agreement to be unenforceable was a leap too far [64], but he left open the possibility that it might be adopted in more incremental steps. That is the question I will focus on.

Tuesday, January 5, 2021

There Is No Such Thing as Licensee Estoppel

Loops LLC v Maxill Inc 2020 ONSC 5438 Lederer J: Swinton, Penny, JJ revg 2020 ONSC 971 Templeton J

            2,577,109 / Toothbrush

In Lear Inc v Adkins 395 US 653 (1969) the US Supreme Court abolished the doctrine of “licensee estoppel,” which (more or less) prevents a licensee from challenging the validity of the licensed patent. Monday’s post summarized the decision of Lederer J for the Ontario Divisional Court holding that a “no-challenge” clause in a settlement agreement is enforceable, Lear v Adkins notwithstanding. In this post I will argue that there is really no such thing as licensee estoppel in Canadian law. More precisely, the rule at issue is simply based on contract law and it is not helpful to characterize it as based on an estoppel.

First, what exactly is the rule? The precise nature of the licensee estoppel rule was not at issue in this case, as the question was only whether the express no-challenge clause was enforceable. Lederer J at [61] did quote from Asturiana (1979) 55 CPR(2d) 131 (Ont HCJ) referring to the consistent Anglo-Canadian authority “that a licensee is not permitted to dispute the validity of the patents.” This is not exactly right because there are many circumstances in which a licensee is permitted to dispute validity. It is more accurate to say that a licensee cannot raise invalidity of the patent as a defence to an action for payment of royalties owing under a licence agreement, except in the case of fraud, unless the agreement contains an express covenant as to validity of the patent. (Asturiana was an action for payment of royalties, so the quoted statement is correct in that context.)

Now, in some respects the rule is very strict. In particular, the licensee cannot raise invalidity as a defence to an action for payment of royalties even when the patent has actually been held invalid in other proceedings: Mills v Carson (1893) 10 RPC 9 (CA); African Gold Recovery Co Ltd (1897) 14 RPC 660 cited with approval in Anderson v EJ Shepard Ltd (1930), 66 OLR 105 (CA) [20]; Duryea v Kaufman (1910), 21 OLR 165 (OHCJ) [23] (in obiter).

On the other hand, it is subject to a major exception. While a licensee cannot raise invalidity as a defence to an action for payment of royalties under the licence, a licensee, even a current licensee, may always raise invalidity as a defence to an action for infringement: this was in fact the holding in Duryea v Kaufman, the leading case that firmly established licensee estoppel in Canadian law with a thorough review of the English cases; see also Bayer AG v Apotex Inc (1995) 60 CPR(3d) 58 (OGD) 73 (“a licensee will not be estopped from alleging that the patent is invalid if the action brought against it is for infringement of the patent”) aff’d (1998) 82 CPR (3d) 526 (Ont CA) 532-33; Fuel Economy v Murray (1930) 47 RPC 346 (Ch) at 353 (“an estoppel cannot arise in an action for infringement”) aff’d 358 (CA). So, for example, a current licensee may argue invalidity if the patentee alleges the licensee has infringed by carrying out activities outside the scope of the licence, whether by falling outside of the permitted uses (Apotex v Tanabe (1994) 59 CPR(3d) 38 (OGD) 49; Bayer AG v Apotex Inc) or outside the geographic scope of the licence (Fuel Economy v Murray). A defendant claiming a right by licence may plead invalidity in the alternative: Duryea [35], [41]-[42]. A former licensee whose licence has come to an end by revocation, lapse of time, or any other method, may raise invalidity as a defence to an action for infringement in the same manner as if it had never had a licence: Duryea [35]; Crossley v Dixon (1863), 10 HLC 293, 308. Conversely, a former licensee may not raise invalidity as a defence in an action for unpaid royalties owing under the agreement that was formerly in effect: Crossley v Dixon (1863), 10 HLC 293, 308; Philco Products Ltd v Thermionics Ltd [1943] SCR 396, 419.

Now, while this rule is commonly referred to as an estoppel, it is not based on any recognized form of estoppel. Indeed, in every case I have found in which a recognized estoppel has been argued, it has been rejected: see Curtiss-Wright Corporation v Canada [1968] 1 ExCR 519 (rejecting estoppel in pais and promissory estoppel); Cropper v Smith (1884), 1 RPC 81 (CA) (rejecting arguments based on estoppel of record, estoppel by deed and estoppel in pais) aff’d on this point (1885) 10 AC 249 (HL) at 259. In Clark v Adie (No 2) (1877), 2 AC 423, 435-36 (HL) Lord Blackburn drew a loose analogy with the position of a tenant of lands who is his estopped from denying his lessor’s title so long as the lease remains in force, and the tenant has not been evicted from the land, he is estopped from denying that his lessor had a title to that land. However, Lord Kenyon in Hayne v Maltby (1789) 100 ER 665 at 666, rejected the analogy.

This should perhaps not be very surprising. Generally speaking, an estoppel is an equitable doctrine that supplements contract law by giving effect to some representation that is not supported by consideration, such as in promissory estoppel when a representation by one party to an agreement leads the other party to suppose that the agreement will not be strictly enforced, even though no consideration may have been given for the forebearance: John Burrows Ltd [1968] SCR 607 at 614-16; Curtiss-Wright Corporation v Canada [1968] 1 ExCR 519 at 530; and see generally John D McCamus, The Law of Contracts, (2005), Ch 8. It is sometimes said that the licensee estoppel arises “from the fact of the relationship itself,” (Fox on Patents 3rd ed Vol I, 618; Fox 4th at 320) but this is clearly not true; a licensee may raise invalidity as a defence if there is an express guarantee that the patent is valid, notwithstanding that the nature of the relationship is otherwise the same. In any event, it is very difficult to see how the act of entering into an agreement could in itself give rise to an estoppel making it inequitable to insist on the terms of the agreement.

In Lear the US Supreme Court at 656 explained that

The theory underlying this doctrine is that a licensee should not be permitted to enjoy the benefit afforded by the agreement while simultaneously urging that the patent which forms the basis of the agreement is void.

Perhaps this was the theory in US law, but it cannot be the explanation for the Anglo-Canadian rule; it is perfectly clear, as discussed above, that when sued for infringement a current licensee may indeed enjoy the benefit of the licence agreement while simultaneously urging that the patent is void.

So, we have a bit of a puzzle. Licensee estoppel is an estoppel that is not related to any other estoppel, that does not prevent either party from relying on the strict terms of the contract, and that operates only in respect of a claim brought on the contract. In short, it looks a great deal like a contract doctrine.

And indeed it is. The basis for the “estoppel” doctrine is really very simple. A licensee cannot raise invalidity as a defence to an action for unpaid royalties under a licence because the validity of the patent is normally immaterial to the obligation to pay under the contract. This has nothing to do with estoppel; it is purely a matter of contract. The licensee is contractually obliged to pay and unless otherwise provided in the contract, that obligation is not contingent on the validity of the contract. The “estoppel” such as it is, does not arise from the relationship between the parties, it arises from the terms of the contract. See Aidan Robertson, “Is the Licensee Estoppel Rule Still Good Law? Was It Ever?” 10 EIPR 373 (1991) making much the same point.

This explains what might otherwise be counter-intuitive aspects of the doctrine. It is why the actual invalidity of the patent, established in different proceedings, has not afforded a defence. It is not because the courts have buried their heads in the sand, like ostriches, resolutely ignoring the fact of invalidity; rather, the courts have in effect said, “Yes, we know the patent is invalid, but that doesn’t affect your obligations under the contract.” That is why fraud is an exception; contracts induced by fraud are not enforceable as a matter of general contract law doctrine. That is why an express provision is an exception; it is no exception at all, as the question is simply a matter of interpreting the contract, and an express term will always trump an implied term. That is why a licensee in good standing can always raise invalidity as a defence to an infringement action; invalidity is always material to the licensee’s obligation to pay damages for infringement, even when it is not material to the licensee’s obligation to pay royalties. That is why licensee estoppel is not related to any other established form of estoppel: it is not an estoppel, but purely a matter of contract. When an allegation of invalidity by the licensee is struck, it is simply on the basis of the procedural rule that allegations that can have no effect on the outcome of the case should be struck as a waste of judicial resources.

The more interesting question is why, as a matter of contract law, the obligation to pay royalties does not turn on the validity of the agreement (unless of course, the agreement itself so provides). The answer is most clearly articulated in the early cases, decided before the matter was settled simply by reference to authority. It follows from two principles. First, a warranty or guarantee of validity will not normally be implied into a licence agreement; absent express language or special circumstances, either a licence or assignment is presumed to be transferring only such rights as the letters patent conferred upon the patentee: Electric Fireproofing Co of Canada v Electric Fireproofing Co (1910) 43 SCR 182 at 184-85 (Davies J), 189-90 (per Duff J), 195 (Anglin J) (assignment); Vermilyea v Canniff (1886), 12 OR 164 (OntHCJ), 169 (licence); Duryea (1910), 21 OLR 165 [10]-[14] (reviewing the cases); Anderson v EJ Shepard Ltd (1930), 66 OLR 105 (Ont CA) [19] (licence); Rymland v Regal Bedding Co Ltd (1966), 51 CPR 137 (Man CA) 141 (licence). This is because the validity of the patent turns on objective considerations—the early cases dealt primarily with anticipation by a third party publication—and not on information peculiarly within the knowledge of the patentee, and therefore the licensee is generally as well placed as the patentee to assess the validity of the patent: Hall v Conder (1857) 140 ER 318, 327; Taylor v Hare (1805) 127 ER 461, 462; Lawes v Purser (1856) 119 ER 1110, 1112; Vermilyea v Canniff (1886), 12 OR 164 (OntHCJ), 169; Rymland v Regal Bedding Co Ltd (1966), 51 CPR 137 (Man CA) 141. As the exception that proves the rule, if the patentee had actual knowledge of invalidity, this would likely constitute fraud that would enable the licensee to raise invalidity even in an action for payment under the contract: see Lawes v Purser (1856) 119 ER 1110, 1112. Moreover, this is not a strict rule as such; it is simply the usual outcome of interpreting the contract in light of the fact that the licensee is normally in as good a position to assess validity as the licensor. Thus the most accurate statement of the law is that “Courts are not prepared to readily extend the meaning of doubtful words so as to convert them into an express warranty of the validity of a patent”: Rymland 141.

Second, invalidity of the patent does not normally constitute complete failure of consideration, as even an invalid patent may effectively confer the benefit of exclusivity until such time as it is challenged: Vermilyea v Canniff at 169; Lawes v Purser (1856) 119 ER 1110, 1112; Taylor v Hare (1805) 127 ER 461, 462. So, in the early English case of Taylor v Hare, the licensee had practised the invention for 6 years before discovering prior art said to anticipate, and sought on that basis to recover all the royalties that had been paid to date. The court rejected the action, with Heath J stating at 462 “It might as well be said, that if a man lease land, and the lessee pay rent, and afterwards be evicted, that he shall recover back the rent, though he has taken the fruits of the land.” A licence agreement may also involve a transfer of know-how or confidential information, or more than one patent, so that it would be unwarranted to read in a provision absolving the licensee of the obligation to pay on the basis of the invalidity of one patent. Consequently, even when there is language in the agreement relating to the validity of the patent, the courts will not interpret it as absolving the licensee of the obligation to pay royalties (unless it is expressly to that effect); it is more likely to be interpreted as a warranty, breach of which might be the basis for a counter-claim in damages: see eg Hall v Conder (1857) 140 ER 318, 323; Gillard v Watson (1924) 26 OWN 77 (OCA).

In summary, there is no rule of licensee estoppel, because there is no rule and it is not an estoppel. While a licensee cannot normally raise invalidity as a defence to an action for payment of royalties, this is simply because the obligation to pay royalties under the licence is not normally conditional on validity.

Monday, January 4, 2021

No-Challenge Clause in Settlement is Enforceable

Loops LLC v Maxill Inc 2020 ONSC 5438 Lederer J: Swinton, Penny, JJ revg 2020 ONSC 971 Templeton J

            2,577,109 / Toothbrush

This interlocutory motion concerns the enforceability of a “no-challenge” clause in a settlement agreement. It also indirectly raises the question of whether Canadian law should follow the decision of the US Supreme Court in Lear Inc v Adkins 395 US 653 (1969) and abolish the doctrine of “licensee estoppel,” which (more or less) prevents a licensee from challenging the validity of the licensed patent. At first instance Templeton J held that the no-challenge clause at issue was not enforceable (see here). In this appeal, Lederer J, for a unanimous panel of the Ontario Divisional Court, reversed, holding that a no-challenge clause in a settlement agreement is indeed enforceable according to its terms. At the same time, he left the door open to the possibility that Lear might be followed in a different context as part of the incremental evolution of the law; my sense is that he did not particularly favour such a development, but was simply concerned not to decide any more than was necessary to dispose of this motion.

In this post, I will summarize Lederer J’s decision. I agree entirely with his reasons, both in the substantive analysis and in his balanced approach to legal change. In two subsequent posts I will address issues that are raised by the decision, but which go beyond what was required by this appeal. My next post will review the Anglo-Canadian law and argue that there is no such thing as licensee estoppel; more precisely, “estoppel” is a very unhelpful label for a doctrine grounded entirely in contract law and not on any kind of equitable principle. In my third post I will argue that Lear is not sound as a matter of policy and should not be followed in any context, even as a part of a more incremental development of the law.

Facts

Loops and Maxill are in the toothbrush business. In 2012 Loops sued Maxill in Canada for infringement of Loops’ 109 patent. Two years later they reached a global settlement, which contained a “no-challenge” clause stating that Maxill would not “directly or indirectly assist any person attacking the validity” of either the Canadian 109 patent or the corresponding US Patent 8,448,285 [3]–[5], [ONSC 971/15]. Maxill then created another toothbrush to try to design around the 109 patent. In 2015 Loops sued Maxill in Canada over the new toothbrush, this time for breach of the Agreement [7], [ONSC 971/20]. In 2017 Loops also sued Maxill in the US, alleging the same new toothbrush infringed the US 285 patent [9]. Shortly thereafter, a US subsidiary of Maxill, Maxill (Ohio), brought an action seeking a declaratory order that the 285 patent was invalid [10]. The US actions were consolidated in Washington [11]. The result of the US consolidation is that Maxill itself was attacking the validity of the US patent, notwithstanding the no-challenge clause in the settlement agreement; under Lear, as interpreted in the US 9th Circuit (which includes Washington), the no-challenge clause is unenforceable [ONSC 971/62]. Loops then brought this Canadian motion for an interlocutory injunction prohibiting Maxill from challenging the validity of the US 285 patent in the US litigation, on the basis that this was a breach of the no-challenge clause in the Agreement [ONSC 971/1]. In effect, Maxill is seeking a partial anti-suit injunction to enforce the no-challenge clause in the settlement agreement.

Because the injunction sought involved the enforcement of a restrictive covenant, this required Loops to make out a “strong prima facie case” on the merits [15]. In the decision under appeal, Templeton J held that Loops had failed to make out a strong prima facie case, for three reasons, the first two of which are largely peculiar to the case. The Divisional Court reversed on all points.