Monday, January 27, 2020

Overbreadth Sole Basis for Holding Claims Invalid

Seedlings Life Science Ventures, LLC v Pfizer Canada ULC 2020 FC 1 Grammond J
             2,486,935 / Auto-Injection of Medication

In Seedlings Grammond J held four claims of the 935 patent to be invalid on the sole ground of overbreadth. In my draft paper on Overbreadth in Canadian Patent Law, I argued that overbreadth is almost always redundant, in the sense that the same facts and analysis that support overbreadth also support a finding of invalidity on some other ground as well. Overbreadth has been applied as a truly independent ground of invalidity in just four cases. The first was Amfac (1986), 12 CPR (3d) 193 (FCA), in which the FCA upheld a decision striking down the key claim in the patent on the sole basis of overbreadth. The doctrine then lay unused for over thirty years, until last year, when it was invoked in two FC decisions, Servier v Apotex 2019 FC 616 (here) and Aux Sable 2019 FC 581 (here). In those decisions, while overbreadth was applied as an independent doctrine, it was not the sole basis for striking down the claims at issue. Seedlings is the third recent case to apply overbreadth as an independent doctrine, and in this case four claims were held invalid solely on that basis.

The SCC decision in BVD [1937] SCR 221 provides a good example of the typical use of overbreadth. The patent at issue disclosed a method of making semi-permeable fabrics using thermoplastic fibres woven into a yarn with cotton fibres, then fusing the fibres with heat and pressure. The use of thermoplastic in the form of a yarn was “the very essence of the invention” (228, 230). The claims, however, specified a fabric which “contains” a thermoplastic with no mention that it was in the form of yarns. The Court therefore held that the prior art Dreyfus patent, which disclosed a similar process using thermoplastics in the form of sheets rather than yarn, “constitutes a complete anticipation of the claims of the respondent,” unless those claims could be limited to a thermoplastic yarn (232-33). The Court held that the claims could not be so limited, and consequently, “[t]he claims in fact go far beyond the invention. Upon that ground the patent is invalid” (237). Thus the claims were overbroad because they had omitted an essential element, namely that the thermoplastic should be woven into the yarn. As the SCC explained, the situation was much the same as that in Erickson’s Patent:

In Erickson’s Patent case [(1923) 40 RPC 477 (CA)], it was held that the patentee had failed so to limit his first claim as to confine it to that which was the novelty (if any) of the invention, and that accordingly the claim was so wide as to render the patent invalid.

It is because the claim lacked novelty that, “accordingly,” it was too wide.

When used in this way, overbreadth is not a distinct doctrine; to say the claim is overbroad is just a manner of explaining why it lacks novelty. The term is typically used when the patent discloses subject-matter that is novel, non-obvious etc, but claims more broadly, as opposed to those cases in which the patent claims exactly what is disclosed, but it turns out that that subject-matter is not new. So, in Erickson’s Patent Warrington LJ at 487 introduced the argument as follows (my emphasis):

It is said that the Specification, when properly construed, claims in its first Claim more than the Patentee can by any means maintain is a novelty; in other words, that he has not limited himself to the new matter which he says is the feature of his invention, and that if he obtains a monopoly for the Claim in the way he has framed it he would be claiming a monopoly to which he is not entitled.

I propose to deal with the matter solely upon the first ground of attack, which I have already mentioned, namely, that the Claim is so wide as to include matters to which it is impossible to attribute any novelty at all.

In Erickson’s Patent, as in BVD, the claims were overbroad because they were anticipated; and conversely, they were anticipated because they were overbroad.

As Grammond J noted, and as BVD illustrates, “[u]sually, a finding of overbreadth flows from the fact that an essential element of the invention is missing from the claims” [173, citing BVD at 235]. The key question is how to determine what is essential. If, as I have argued above, overbreadth is no more than a way of saying the invention is invalid on some other ground, then overbreadth is determined in the same way as the other relevant ground of invalidity. In BVD, for example, the SCC spent more than two pages detailing how the prior art patents anticipated the patent at issue, if the claims were not limited to the use of woven thermoplastic yarns (230-330). Similarly, in Erickson’s Patent, the invention related to the nozzle for an oil burner such as were used in the boiler of a steamship. The fuel has to be filtered for impurities and then atomized before entering the burner. Nozzles known in the prior art used a burner with a filter separate from the atomizer, and Warrington LJ noted that the novelty (if any) in the patentee’s nozzle lay in using a nozzle with perforations that acted both as a filter and an atomizer (490). The nature of the prior art and the putative point of novelty were not disputed; the real question was whether the patentee had drawn his claims so broadly as to encompass the prior art, and the Court of Appeal found that he had.

Thus in both BVD and Erickson’s Patent, as well as in almost all cases in which overbreadth is applied, an element is considered essential if it is necessary to ensure that the claimed invention is indeed an invention within the meaning of s 2 of the Act, which is to say that it is new, useful and non-obvious subject-matter. If the patentee has disclosed subject-matter that is new, useful and non-obvious, but fails to include the point of novelty in the claims, then it follows both that the claims are broader than the invention disclosed, and the claims are anticipated. Essentiality in such cases is determined in the same manner as novelty or non-obviousness, by comparison of the claimed invention with the prior art.

The problem arises when essentiality is determined solely from an inspection of the specification. That is the approach taken in Amfac, in which the FCA stated (197, emphasis added):

the issue involves solely the construction of the patent and…no reference to the prior art is either required or proper in this case.

The FCA in Amfac cited no authority for this proposition. In my paper I suggest that this holding turned ultimately on a misreading of BVD by Strayer J at trial. In any event, the problem is that it is not normally possible to determine what is new, useful and non-obvious, solely by an inspection of the specification. Novelty and non-obviousness both turn on a comparison between the prior art and the claimed invention. In Amfac, the FCA instead determined which elements were essential by examining the disclosure to see what elements were emphasized as being important. There are two problems with such an approach.

First, very often, the inventor does not know what they have invented. That is why cascading claims are used. The inventors of Viagra®, for example, thought that what they had discovered was that selective PDEV inhibitors were effective for the treatment of erectile dysfunction when administered orally, and they claimed accordingly. It turns out, however, that the active ingredient in Yin Yang Huo (aka Horny Goat Weed), a traditional Chinese herbal medicine used for treating impotence, is icariin, a PDEV inhibitor. The broadest claim to the use of PDEV inhibitors was therefore anticipated: Ex parte Pfizer Inc, No. 2009-004106, 2010 WL 532133 (BPAI Feb 12, 2010). The narrower claim to the use of sildenafil in particular, for treating ED, was novel, but the inventor did not know the true point of novelty at the time of filing.

Second, and perhaps more importantly, identifying the essential elements of the invention amounts to defining the true invention. The problem is that the purpose of the disclosure specification is to disclose the invention, not to define it. Defining the invention is, by s 27(4), the role of the claims. The whole reason that claims were introduced — by patent applicants themselves — was because of the arbitrary consequences of the courts attempting to discern the nature of the invention from a general specification that was intended to serve the dual purpose of both disclosing and defining the invention: see Fletcher Moulton LJ British United Shoe Machinery Co (1908), 25 RPC 631 at 650, noting that “These two things – the delimitation of the invention, and full practical directions how to use it – are in their nature almost antagonistic”; and see Siebrasse The False Doctrine of False Promise, (2013) 29(1) CIPR 3 at 44-46.

The problem is acute when the specification discloses both a basic invention and an improvement, or at least some bells and whistles. If the disclosure emphasizes the advantages of the improvement, a court looking solely to the disclosure to identify the essential elements might take the features of the improvement to be essential, and consequently strike down the claim to the basic invention as being overbroad — even though the basic invention is itself new, useful, not obvious, and properly disclosed. That is what happened in Amfac. The invention at issue in Amfac was a system for slicing raw potatoes into french fries by using a high velocity water stream to force the potatoes through a cutter box containing a series of knives set at right angles. That basic invention was claimed, but there was also a separate claim to a refinement of the basic system which additionally diverted the irregular outer slabs for other uses. The FCA decided that diverting the outer slabs was essential, and consequently invalidated the claim to the basic invention for overbreadth — even though the basic invention was new, useful, non-obvious and fully disclosed.

That also appears to be what happened in Seedlings. Recall that the invention at issue relates to an auto-injector, primarily intended for injecting epinephrine to treat anaphylaxis. The best known auto-injector is the EpiPen. The original EpiPen was on the market from the late 1980s until 2009 [10]. The original EpiPen was relatively bulky — a cylinder about 15 cm long and 2.5 cm in diameter — which made it inconvenient to carry. Also, the needle remained exposed after use, which is a particular concern in light of risk of transmission of blood-borne diseases [11]-[12]. Seedlings’ auto-injector, known as the LifeCard, was the inventor’s solution to these problems [13]-[14]. The LifeCard is flat and much smaller than the original EpiPen — the size of a thick credit card — so it is more easily carried and used [63]. It also has a shield which moves forward once the injection is complete to protect the needle [19].

The overbreadth analysis turned on three other elements that are not mentioned anywhere else in the decision:

[174] I agree with Pfizer’s submission that there are at least three essential elements of the invention that are not mentioned in the claims asserted by Seedlings: a syringe carrier, a flat reverse syringe or collapsible bellows and a shared latch locking mechanism. This renders all the asserted claims overbroad, and thus invalid.

The claims invalidated solely on the ground of overbreadth are claims 40, 59,60 and 62. These claims were not obvious ([138]), nor were they anticipated ([102], [115]), nor did they lack utility [163], nor were they insufficiently disclosed ([186]). Thus, claims to subject-matter that was an invention as defined by s 2 of the Act, namely a new and useful manufacture or improvement thereof, and that was fully disclosed, were held to be invalid on the sole ground of overbreadth.

Grammond J gave three main reasons for his conclusion that these elements were essential[175].

[176] First, they are shown in all the embodiments disclosed in the patent. The patent does not teach how to make the invention without these elements. Moreover, Mr. Taylor testified that all but the most preliminary designs incorporated these three elements (October 22, 2019, pp. 171–181, 208–210).

[177] Second, these elements interact with one another and are at the core of the mechanism of the device. In particular, the syringe carrier allows for a smoother distribution of forces on the syringe body, while also unlocking the needle shield after injection. The shared latch locking mechanism allows for the safe positioning of the needle shield both before and after injection. Thus, the device would work differently if they were replaced by other components: Free World Trust, [2000 SCC 66,] at paragraphs 55–57. Replacing those components is not something that the skilled person could do without having recourse to inventiveness.

[178] Third, these components and their arrangement are entirely original. Nothing like this is disclosed in the prior art. Flat syringes are extremely rare. The syringe carrier is an original component that replaces the collet that is typically found in previous auto-injectors. The shared latch locking mechanism is not found in any of the devices that were the focus of discussion at trial or in the expert reports.

With respect, the effect of this analysis is to construe a preferred embodiment as being the true invention.

Grammond J’s first point is that the patent did not teach how to make the invention without these embodiments. But that is irrelevant. As the FCA explained in Apotex v AstraZeneca 2017 FCA 9 [79] (and see also Gilead Sciences 2017 FCA 161 [19]):

It is well established in patent law that when one claims a new and inventive product, an inventor is only required to enable the person skilled in the art to work the invention. He or she need only describe one method or process for making it.

An invention might require two parts to be fastened together. It is enough that the patent discloses the use of a nail for that purpose; it is not necessary to teach how to fasten the parts with a screw, or velcro, or a rivet, etc.

The second point, that the device would work differently if these elements were replaced by other components, is true for any preferred embodiment. Consider claim 40, which “describes an auto-injector with a flat housing, that is front-actuated and that has an actuator that also serves as a needle shield. That is the ‘inventive concept’ of this claim” [139]. Recall that the problem that the inventor sought to solve was that the original EpiPen was bulky and difficult to handle, and after use the potentially contaminated needle was left projecting. The inventor solved this problem with a flat housing and a needle shield. Grammond J held that was not obvious [138]. The use of a syringe carrier, a flat reverse syringe and a shared latch locking mechanism were only one embodiment; perhaps the preferred embodiment. It is very well established that “a patentee who has discovered a new principle is entitled to … claim all modes of carrying it into effect”: Unilever (1995) 61 CPR (3d) 499 (FCA), 514 (FCA), citing Whitford J in American Cyanamid v Berk Pharma [1976] RPC 231 at 257-8 (ChD) (reviewing the cases) and quoting Badische Anilin und Soda Fabrik v Levinstein (1883), 24 Ch D 156:

But when a patent is taken out for a new result not known before, and there is one process described in the patent which is effectual for the purpose of arriving at that new result at the time when the patent is taken out, the patentee is entitled to protection against all other processes for the same result, and no person can, without infringing upon his patent, adopt simply a different process for arriving at the same result. That is stated in a great number of cases and in great detail.

Thus it is clear law that the inventor is entitled to claim any means of implementing the basic invention. The basic invention in respect of claim 40 is a flat injector with a needle shield, and the inventor is not limited to the particular implementation that he envisaged.

The point that “[r]eplacing those components is not something that the skilled person could do without having recourse to inventiveness,” is more interesting. It may seem intuitive that the scope of the claim should not encompass embodiments that could not be practised without inventive ingenuity. The difficulty with that proposition is that any pioneer invention would soon be overtaken by general advances in technology or improvements to the invention itself. A claim to the wheel should encompass a wheel made of bronze, even if the inventor only knew of wood and discovering bronze required inventive ingenuity. Similarly, a claim to the wheel should encompass the spoked wheel, even though devising spokes would have required inventive ingenuity. As Kitchin LJ explained in for the English Court of Appeal in Regeneron v Kymab [2018] EWCA Civ 67 [265] (and see [249]):

[T]he law does not require a patentee to enable each and every embodiment of a claimed invention. As the authorities recognise, a claim may encompass inventive improvements of what is described and a specification is not insufficient merely because it does not enable the person skilled in the art to make every such invention. It is important, however, that any such improvement is still a way of working the original invention.

In this case it is evident that the embodiment with the three contentious elements remains a way of working the original invention of Claim 40. Moreover, to the extent that one might dispute this view of the law, the debate is a matter of insufficiency. This illustrates a distinct reason for avoiding overbreadth, apart from the fact that it may result in striking down a claim to an invention that is new, useful, non-obvious and sufficiently disclosed. A considerable body of doctrine has been elaborated in respect of each of the statutory grounds of invalidity addressing difficult questions such as these. To treat a matter as being one of overbreadth, when it is really an issue of insufficiency, risks neglecting the case law that already exists to provide guidance, both as a matter of law and as a matter of policy.

The third point, that the three elements at issue are entirely original, is beside the point, except to say that they might themselves have supported an independent improvement patent. That the patent disclosed a valuable invention which it failed to claim, is no reason to strike down the separate valuable invention that it did claim.

Thus the effect of Grammond J’s analysis is to construe the key elements of the preferred embodiment as being essential to the invention. If Claim 40, to a flat injector with an actuator that serves as a shield, had been obvious, and the only inventive aspect of the disclosure had been the particular implementation, with a syringe carrier, a flat reverse syringe or and a shared latch locking mechanism, then Claim 40 would have been overbroad (and obvious). But Grammond J held Claim 40 to be not obvious. The particular embodiment with those features is, at most, the best mode. The result is that Grammond J struck down a claim that was new, useful, not obvious and fully disclosed, simply because the inventor had also disclosed the best mode he knew of implementing that invention.

Grammond J gave the following policy justification for the overbreadth doctrine (my emphasis):

[171] Nevertheless, overbreadth sometimes has an independent role to play. Inventors who have a new, non-obvious and useful invention cannot, through the creative use of language, claim something that they have not invented. Justice Ian Binnie of the Supreme Court of Canada colourfully illustrated this principle in Free World Trust, at paragraph 32: “It is not legitimate, for example, to obtain a patent for a particular method that grows hair on bald men and thereafter claim that anything that grows hair on bald men infringes.” Although Justice Binnie made this remark in the context of a discussion of the principles of claims construction, the principle also justifies the doctrine of overbreadth. In his paper quoted above, Professor Siebrasse refers to this situation as the “roads to Brighton” problem, that a 19th-century English judge described as follows: “it would [not] be reasonable to say that if one man has a road to go to Brighton by Croydon another man shall not have a road to go to Brighton by Dorking” (Curtis v Platt, (1863) 3 ChD 135). In a nutshell, a patent must describe a particular method of achieving a goal; it cannot claim the goal itself. (See also Wellcome Foundation, [ 2002 SCC 77] at paragraphs 82–83.)

There are two difficulties with this. The first is that notwithstanding the SCC dicta in Free World Trust, the emphasized proposition is not good law, on the authority of Lord Hoffmann’s decision for the unanimous House of Lords in Lundbeck [2009] UKHL 12 [Lundbeck], which was adopted by Gauthier JA in her decision for the unanimous FCA in Apotex Inc v AstraZeneca 2017 FCA 9 [81]-[89]: see my Overbreadth paper at 20-26.

Now it is true that the roads to Brighton problem is difficult, and might well have gone the other way, as it initially seemed to have done in Biogen [1996] UKHL 18. But in any event, the roads to Brighton problem is irrelevant to the case at hand. The roads to Brighton problem arises when the inventor has discovered one method of achieving a known desideratum (finasteride for growing hair on a bald man), and claims all methods of achieving that desideratum (any compound for growing hair on a bald man), even those which owe nothing to the inventor’s contribution. In this case, the desideratum was a safer and more convenient auto-injector. The inventor discovered one method of achieving that goal, namely an injector with a flat shape and a actuator that serves as a shield. The inventor, in Claim 40, claimed the method which he had discovered—not every method of achieving a safer and more convenient auto-injector. The roads to Brighton problem simply does not arise in this case.

Grammond J cited my paper as suggesting that the overbreadth doctrine has no statutory basis [168]. He responded that overbreadth

stems, by implication, from the scheme of the Patent Act, and in particular section 27(4), which requires “a claim or claims defining distinctly and in explicit terms the subject-matter of the invention for which an exclusive privilege or property is claimed.” If the subject-matter of the invention must be defined, it must be defined accurately and not in an overly broad manner. Thus, the doctrine of overbreadth flows by necessary implication from the provisions of the Patent Act. . .

This is a good justification for the overbreadth doctrine, but, with respect, only the redundant form of that doctrine. An invention, under the Act, is new, useful and non-obvious subject matter. Indeed it is true that the claim must define new, useful and non-obvious subject-matter or it will be overbroad — because it will also encompass subject matter that is not new, or obvious, or lacking in utility.

In any event, my primary objection to overbreadth is not in respect of its statutory basis or lack thereof. My argument was that because of its uncertain statutory basis

the courts should be very cautious. It is not enough to apply the doctrine and declare the claim invalid. Before invalidating a claim solely for overbreadth, the court should satisfy itself that there is some principled reason for invalidating a claim to an invention which is new, useful, non-obvious and sufficiently disclosed.

I fail to see any good reason for striking down the Claim 40, to an invention which Grammond J had expressly held to be new, useful, not obvious and sufficiently disclosed.

I must say that I am surprised that Pfizer, an innovator pharmaceutical company, chose to make this argument. In my view, the emergence of overbreadth doctrine has the potential to destabilize Canadian patent law in the same manner as the promise doctrine, and for much the same reason, namely that in the Amfac version of overbreadth, validity will turn on a subjective parsing of the disclosure. No doubt there is a lot of money in the EpiPen market; but I’m sure there is even more in the pharmaceutical market. If this doctrine becomes entrenched, it will be difficult for Pfizer to turn around and complain to the USTR about a doctrine which it introduced. And if overbreadth doctrine does end up on the Special 301 Watch List, Canada will have a ready answer: “You asked for it.”

No comments:

Post a Comment