Is a Dosage Range Unpatentable as Being a Method of Medical Treatment?

Bayer Inc v Cobalt Pharmaceuticals Company / drospirenone (NOC) 2013 FC 1061 Hughes J
            YAZ / drospirenone / 2,179,728

As noted in my overview post, the Drospirenone decision is notable primarily as a development in the law relating to the patentability of methods of medical treatment, as Hughes J held four of the five claims at issue to be invalid on this basis.

Claims 1, 2, 6,7 and 8 were at issue [124]. All but Claim 8 were held to be invalid [162]. Claim 1, which is representative of the invalid claims, was as follows:

1. Use of an oral dosage form comprising an estrogen selected from

2.0 to 6.0 mg of 17-estradiol and

0.015 to 0.020 mg of ethinylestradiol;

and a gestagen selected from

0.05 to 0.075 mg of gestodene,

0.075 to 0.125 mg of levonorgestrel,

0.06 to 0.15 mg of desogestrel,

0.06 to 0.15 mg of 3-ketodesogestrel,

0.2 to 0.3 mg of norgestimate,

>0.35 to 0.75 mg of norethisterone,

0.1 mg of drospirenone to a drospirenone dose equivalent to 0.075 mg of gestodene, and

0.1 mg of cyproterone acetate to a cyproterone acetate dose equivalent to 0.075 mg of gestodene;

for contraception for a female of reproductive age who has not yet reached premenopause, by administration of the form of dosage for 23 or 24 days, beginning on day one of the menstrual cycle, followed by 5 or 4 pill-free or placebo pill days, for a total of 28 days in the administration cycle.

Claim 2 restricted the estragen to ethinylestradiol, Claim 6 restricted the gestagen to one of the two listed in Claim 1, and Claim 7 was similar to Claim 1, but with somewhat more restricted dose ranges [158].

Claim 8, held to be valid, was as follows:

8. Use according to claim 1, whereby the estrogen is present in a dose of 20 µg of ethinylestradiol or an equivalent dose of 17-estradiol and the gestagen is present in a dose of 75 µg of gestodene or an equivalent dose of levonorgestrel, cyproterone acetate or drospirenone.

As noted by Hughes J, in contrast with the first four claims, “Claim 8 is restricted to a single dosage (not a range) of one of two estrogens; and a single dosage, not a range, of one of three gestagens” [159].

After referring to his Cobalt / zoledronate (NOC) 2013 FC 985 decision (blogged here) as summarizing the law, Hughes J held Claim 1, 2, 6 and 7 invalid for the following reasons (my emphasis):

[160 In the present case, all claims are clearly expressed in terms of use for a contraceptive. All claims except claim 8 provide for a range of dosages for one or both of the estrogen and gestagen components.

[162] The point, however, is not whether a commercial product is provided with fixed dosages and regimens. The point is, what do the claims say? All claims at issue are use claims, not product claims. All but claim 8 claim the use as a contraceptive of a two-component drug with each component to be selected from a choice of components, and with each component to be furnished at a dosage within a range of dosages. Claims 1, 2, 6 and 7 are not proper subject matter for a Canadian patent, as they do not claim a vendible product; they provide for a choice to be made by those prescribing or providing contraceptive drugs to choose between a variety of components and a variety of dosage ranges. Only claim 8 survives, as it is directed to a single dosage of each of two compounds.

Hughes J evidently wrote Cobalt / zoledronate with Drospirenone in mind, so the contrast between Claim 8 and the other four claims is directly reflected in the second pair of contrasting claims he identified at [91]-[92] of that decision:

• the substance X in the form of a 5 mg tablet for the treatment of Y [patenable]

• the use of substance X in a dosage range between A and B for the treatment of X [unpatentable].

From this, it follows directly that Claim 8 is patentable while the others are unpatentable. Drospirenone is helpful in explaining why Hughes J is of the view that claims to a dosage range are unpatentable, namely that “they provide for a choice to be made by those prescribing or providing contraceptive drugs to choose between a variety of components and a variety of dosage ranges” [162]. The implication is that by providing for a choice to be made, this implicates the professional skill of the prescribing doctor: (and see [161]).

I see two problems with this analysis. First, there is a distinction between a patent which requires the use of professional skill to practice the invention, and one in which professional skill may be exercised, but is not required, while practicing the invention. In Drospirenone, Hughes J held the claims to be useful, which implies that any combination with the specified range would work as an effective contraceptive. Professional skill might be exercised in practising the invention, to reduce side effects, for example, but in principle professional skill is not necessary to practice the invention effectively. Consider an analogy to a selection patent, in which the genus patent specifies that a broad range of compounds is effective for a particular purpose. It might well require professional skill, whether that of a doctor or engineer, depending on the patent, to get the most out of the invention; and if sufficient skill is required in choosing the best species, this might even be the basis for a valid selection patent. But the fact that professional skill is normally exercised in the practice of the genus patent, does not make the genus patent invalid. In this case, the patent does not “provide” for a choice to be made, in the sense of requiring a choice to practice the invention effectively, but rather it allows for a choice to be made. Put another way, elsewhere in this decision, Hughes J accepted that the word “about” is not vague and implies a range of about 10% around the specified value [103-06]. It would be possible to cover the entire range specified by Claim 1, by claims framed similarly to Claim 8, except to dosages “about” a single dosage. While this would require a very large number of claims, the difference between this multiplicity of specific dosage claims, and a single claim to a range of dosages, is purely formal. In summary, many patentable inventions allow for the exercise of professional skill in their implementation, and this alone does not make the claims invalid. While Claim 1 certainly allows for the exercise of professional skill, it does not seem to me to require it.

Moreover, even the suggestion that a claim is invalid if it requires the exercise of medical skill, is very difficult to reconcile with Wellcome / AZT 2002 SCC 77 which upheld a Claim 22 to the use of “an effective amount” of AZT. Surely choosing an effective amount of AZT implicates the professional skill of a doctor or pharmacist just as much as choosing from a specified range. Indeed, more skill is required by the AZT claim, because, as just noted, in this case any selection within the range specified by the ‘728 patent would work, whereas in the AZT claim, professional skill must be exercised to select an effective amount. (See also my post on Cobalt / zoledronate, making a similar point.)

More broadly, what is wrong with a claim which requires the exercise of professional skill? Many claims are of this nature, and the question is normally treated as one of sufficiency. It is well established that a claim requiring the exercise of professional skill is not invalid for that reason, so long as undue experimentation is not required. As discussed in my previous post, there may be a sense that the exercise of professional skill is inherently nebulous, but if that is the real objection, it should be dealt with directly, by invalidating for insufficiency or ambiguity, claims where the professional skill required by the claim is of a nature that is not described or cannot be reliably replicated.

If the objection relates to the exercise of medical skill in particular, we must ask what distinguishes medical skill from other professional skills. One answer is that there is a policy argument to be made that a physician should not be prevented from treating her patient to the best of her ability by fear of a patent action. But if that is the real objection, the better response is to address it directly, by providing a defence for physicians, as in the US under 35 USC § 287(c)(1). That is surely a policy decision for the legislature; and in any event, this policy goal is not aided by Hughes J’s distinction between dosage ranges and a single dosage; to the extent that the compositions encompassed by any of the Claims are “used” by the prescribing physician, allowing claims to specific dosages exposes physicians to liability; Claim 8 would be infringed by a physician who determines that the dosage specified by that claim is appropriate for her patient.

Ultimately, it seems to me that the distinction between a dosage range and a specific dosage is a formalism which lacks a a clear policy rationale, and is inconsistent with Wellcome / AZT.

It is also very disappointing to see the apparent resurrection, both in this decision [162] and in Cobalt / zolendronate and the case-law discussed therein, of the long-discredited “vendible product” test for patentable subject matter. This test was created by Morton J in Re GEC's Application, (1942) 60 RPC 1 at 4, where he stated that “a method or process is a manner of manufacture if it (a) results in the production of some vendible product or (b) improves or restores to its former condition a vendible product or (c) has the effect of preserving from deterioration some vendible product to which it is applied.” The list is curiously specific, if it is an attempt to reconcile difficult cases on narrow factual grounds – though Morton J did not cite specific authority for this proposition, referring merely to “cases cited to me”. But it is more bluntly described as legal gerrymandering, aimed at reaching a particular result on the facts. There is no unifying principle, either stated or apparent. While the phrase “vendible product” is repeated, it is not truly a unifying concept, as not every invention that affects a vendible product is patentable; only certain effects are within the scope of the rule. In particular, on the facts of the case, Morton J held that using a known compound for extinguishing a fire did not fall within the rule as stated, apparently because preserving a vendible product from fire does not amount to “preserving from deterioration.” This invisibly fine distinction was evidently developed for the purpose of allowing Morton J to hold the invention unpatentable. The application at issue was for a patent on a method of extinguishing incendiary bombs by means of a concentrated aqueous solution of a known substance (zinc chloride). Legal realism provides the most obvious explanation for the rule: the case was decided in the middle of World War II, just after the Blitz, and Morton J did not want the inventor to be able to hold London to ransom. While this is no doubt a reasonable concern, it does not warrant re-writing the law of patentable subject matter. In Canada today, this concern could be addressed by s 19 of the Act.

Regardless of the explanation for Morton J’s decision in Re GEC's Application, it is no longer good law. Morton J held that the claims in question on the basis that they were “nothing more than claims for a new use of an old substance, such use not being, in my view, a manner of manufacture.” This holding, which was made immediately after his statement of the vendible production rule, and which is apparently his application of that rule on the facts, is directly inconsistent with Shell Oil, [1982] 2 SCR 536, the leading SCC decision on patentable subject matter, which upheld the patentability of a new use for an old compound. Even prior to that, the vendible product rule itself has been effectively expunged from the law by the penetrating analysis of the High Court of Australia in NRDC (1961) 102 CLR 252 (HCA). For more discussion of the vendible product rule, see my article, The Rule Against Abstract Claims: History and Principles, 26 CIPR 205 at 224-25 (draft version here).

YAZ Overview

Bayer Inc v Cobalt Pharmaceuticals Company / drospirenone (NOC) 2013 FC 1061 Hughes J
            YAZ / drospirenone / 2,179,728 – 2,382,426

Cobalt sought an order of compliance allowing it to market a drospirenone + ethinylestradiol combination birth control product. Two patents were listed on the register. Hughes J granted the order of prohibition based on the ‘426 patent (which doesn’t expire until 2020), but dismissed it in respect of the earlier expiring ‘728 patent on the basis that the claims which were arguably infringed were unpatentable methods of medical treatment, as well as non-infringement. This decision is significant primarily as a development in the law relating to patentability of methods of medical treatment. Another point worth noting is that the promise of the patent respecting utility was modestly construed for both patents. The decision also raised an evidentiary point of interest. Other issues turned on the facts. This post discusses the evidentiary point and the promise of the patent.

Evidentiary point
Claim 31 of Bayer’s ‘426 patent was to a composition of drospirenone with a specified dissolution profile and the question of fact was whether Cobalt’s product fell within the specified parameters. The problem for Bayer was that Cobalt refused to provide samples to Bayer for testing: “Bayer complains that, despite a motion that it brought to compel Cobalt to produce samples, Cobalt refused to do so; and the Court would not compel it to do so” [33] (and see [66]). Despite this lack of direct evidence, Hughes J held in Bayer’s favour on the infringement issue:

[68] Given that Cobalt is obliged in its Notice of Allegation to provide sufficient information so that Bayer can come to grips with the allegations made; and, given that Cobalt has supplied no sample tablets nor any evidence as to the dissolution parameters of its tablets, I must conclude that Cobalt’s allegations as to non-infringement of claim 31, and dependent claims, of the '426 patent are not justified.

In the absence of evidence, the answer must turn on the burden of proof. Hughes J summarized the rule as being that “with respect to the second person’s [generic’s] allegations of non-infringement, the first person [innovator] bears the burden of proving that such allegations are not justified” [32], which would seem to imply that Cobalt should prevail. But more precisely, the generic must present “sufficient evidence to give the allegations in its NOA an air of reality”: see eg 2013 FC 573 [37]. Similarly, in 2011 FCA 215 [15], quoted by Hughes J, the FCA affirmed that “The determination must turn on whether there are allegations by the second person sufficiently substantiated to support a conclusion for administrative purposes (the issuance of a NOC) that an applicant's patent would not be infringed if the second person's product is put on the market” (FCA emphasis). So, Hughes J’s holding on this point illustrates that a mere statement that its product will not infringe does not amount to sufficient substantiation of the allegation of non-infringement to give the allegation an air of reality. However, it appears that if the generic does supply the product for testing, this is sufficient to substantiate the allegation of non-infringement, even if the innovator does not actually test it: see 2005 FC 1205 rev'd on related but distinct grounds 2007 FCA 209.

Promise of the Patent
While the obviousness issue in this case turned on the facts, a brief review is necessary to understand the utility issue. Drospirenone is acid-labile and would degrade (isomerize) in the presence of stomach acid, and micronization could be expected to exacerbate this problem. Consistently with this, degradation of micronized drospirenone was a serious problem in in vitro tests, and consequently Bayer initially worked on an enteric coated tablet. It turned out that micronized drospirenone nonetheless had good bioavailability in vivo. Hughes J held on the facts that the good bioavailability was non-obvious [87].

The ‘426 patent contained the following statement:

Without wishing to be limited to any particular theory, it appears that the in vitro dissolution rate of drospirenone is connected to the dissolution rate in vivo resulting in rapid absorption of drospirenone in vivo on oral administration of the compound. This is an advantage because isomerization of the compound in the gastric environment and/or hydrolysis in the intestine is substantially reduced, leading to a high bioavailability of the compound.

That is, the patent speculated that the reason for the good bioavailability was that fine particles of drospirenone were absorbed more quickly than they degraded. Hughes J indicated that Cobalt argued that this passage should be construed as a promise, so that the patent would lack utility unless this theory could be shown to be the correct explanation for the observed bioavailability [97]. (I must say that the passage from Cobalt’s NOA set out at [89] seems to me to be making the more modest argument that the studies relied on were not sufficient to show that a composition with the claim dissolution profile would be an effective contraceptive, but of course I do not have the advantage of the full record.)

Hughes J declined to read this statement as a promise. He quoted at length from the FCA’s recent Clopidogrel decision 2013 FCA 186 (blogged here), emphasizing that the promise must be explicit [94] and that “Courts should not strive to defeat otherwise valid patents” [93]. He concluded that the quoted passage from the patent was not a promise, but rather an effort to explain the mechanism of action [98].

Hughes J’s emphasis on the need to avoid being overly aggressive in construing the promise is, in my view, welcome, but his holding with respect to the ‘426 patent is not especially significant, as it seems to me that the passage in question would probably not have been construed as a promise even prior to Clopidogrel.

The promise analysis of the ‘728 patent is more interesting. The ‘728 patent claimed a low-dose contraceptive composition, with a 23-day dosage cycle. In effect, a lower dose is achieved by using a longer period of administration. The specification contained the statement that “The advantages of a combination preparation for oral contraception [with the claimed characteristics] can be characterized as follows” with four specific advantages listed [120]. The passage concludes by noting that “In summary, an intake [of the claimed composition] can produce the above-mentioned advantages. . .” [120]. This is very much the kind of statement that might be taken as a promise of utility. Hughes J did not construe this as a promise, finding that it was included in the patent to support non-obviousness, rather than as a promise of utility [120]:

[152] That list should not be elevated to a “promise”; it is simply an observation as to advantages expected to be achieved. As expressed by Pelletier JA in Sanofi-Aventis v Apotex Inc, [Clopidogrel] 2013 FCA 186 at paragraph 67, in alluding to a possibility, an inventor is not promising a result to be achieved; a goal is not necessarily a promise.

Because the statement in the ‘728 patent might easily have been construed as a promise, Hughes J’s holding on this point is significant in indicating that Clopidogrel may mark a real shift in approach. Of course, it shouldn’t be surprising that an FCA decision would influence the Federal Court, but there is still a great deal of subjectivity involved in assessing the promise of the patent, so the enthusiasm with which the FC embraces Clopidogrel will strongly influence its real impact. With that said, one decision does not make a trend. Moreover, the significance of the decision on this point is diminished because it was not determinative. While Hughes J rejected the utility attack, he held the key claims invalid for lack of patentable subject matter. A more stringent test of the impact of Clopidogrel will come when validity turns on the construction of the promise.

How (Not) to Cure an Inadvertent False Promise in a Granted Patent

F. Hoffmann-La Roche AG v. Commissioner of Patents 2013 FC 1001 Roy J

            Capecitabine / XELODA / 2,103,324

As reaffirmed in the recent FCA decision in Sanofi-Aventis v Apotex / clopidogrel 2013 FCA 186, the promise of the patent doctrine is now firmly established in Canadian law. If a patent promises a particular utility, utility will be measured against that promise: ibid [47]. However, there is no requirement that a patent contain a promise, and if it does not, the invention need have a “mere scintilla” of utility: ibid [49]. In the future, applicants may be able to draft patents that will avoid making promises, but what can be done about patents that have already been granted, but not yet litigated? This was the question faced by Hoffmann-La Roche, the applicant in this motion. While Roy J refused the application, the decision suggests the possibility that s 53(2) of the Act might be available to rectify a false promise in litigation.

Non-Obviousness Affirmed on the Facts in Teva / Zoledronate

Teva Canada Ltd v Novartis Pharmaceuticals Canada Inc / zoledronate (NOC) 2013 FCA 244 Gauthier JA: Evans, Near JJA aff’g 2013 FC 283 Hughes J
            ZOMETA ACLASTA / zoledronate / 1,338,937

In a brief decision from the bench, the FCA has affirmed Hughes J’s decision granting an order of prohibition in favour of Teva on the basis of deference in respect of his appreciation of the evidence.

This NOC proceeding initially concerned two patents, one to zoledronate (‘937) and a separate patent to a broad class of compounds including zoledronate (‘895). At first instance Hughes J held the ‘895 patent invalid for lack of utility, but the ‘937 patent was held to be valid and an order of prohibition was granted accordingly: see my post here. Teva appealed only the holding respecting the ‘937 patent, on the basis that Hughes J had erred in his obviousness analysis, arguing that Hughes J had misapplied the Windsurfing / Pozzoli test accepted by the SCC in Sanofi 2008 SCC 61 [3]. Teva argued that Hughes J set the bar for non-obviousness too high, based on his statement that “one cannot raise the bar too high in respect of obviousness. Research ought to be rewarded, not discouraged” [FC 161]. The FCA pointed out at [6] that this statement reflected a similar statement made by Rothstein J in Sanofi at [64]. The FCA then went on to emphasize the deferential standard of review for factual findings in respect of obviousness [11], and noted that “The Judge is presumed to have considered all the evidence before him” [12]. In the circumstances, the FCA concluded that there was no basis for interfering with Hughes J's conclusion.

Are Swiss Form Claims a Matter of Substance?

Novartis v Cobalt / zoledronate (NOC) 2013 FC 985

            2,410,201 – ACLASTA

It is now well-established in the Federal Courts that methods of medical treatment are not patentable, but how to define such unpatentable methods is very much an open question. In Cobalt / Zoledronate Hughes J, has reviewed and summarized the Supreme Court and Federal Court jurisprudence on this point, before concluding that all the claims at issue were unpatentable. In my view, to the extent his conclusions of law are supported by  the jurisprudence, they do not justify his conclusion on the facts that the Swiss form claims are to unpatentable subject matter.

Opening the Door to Punitive Damages in Patent Law?

Eurocopter v. Bell Helicopter Textron Canada Limitée 2013 FCA 219 Mainville JA: Noël, Trudel JJA aff’g 2012 FC 113 Martineau J
            2,207,787

Punitive damages are very rarely awarded in Canadian patent cases, yet in his trial decision in Eurocopter, Martineau J held that Eurocopter was entitled to punitive damages, essentially on the basis that Bell had intentionally copied helicopter landing gear which it knew to be patented [FC 431-33]. My post on this aspect of Martineau J’s decision argued that this holding is very problematic, as it is likely to having a chilling effect on desirable challenges to invalid patents. The FCA has now affirmed Eurocopter’s entitlement to punitive damages.

Line of Reasoning Supporting Sound Prediction Must Be Disclosed in the Patent

Eurocopter v. Bell Helicopter Textron Canada Limitée 2013 FCA 219 Mainville JA: Noël, Trudel JJA aff’g 2012 FC 113 Martineau J
            2,207,787

The doctrine of sound prediction of utility is routinely applied in the context of pharmaceutical and chemical inventions, but rarely – perhaps never, before Eurocopter – applied to a mechanical invention. Nonetheless, the FCA Eurocopter decision is a significant development in the law of sound prediction, which affirms and emphasizes what I consider to be the most problematic aspects of the doctrine.

Given the novelty of applying sound prediction in the mechanical context, Eurocopter argued that the doctrine doesn’t apply to mechanical inventions at all [144]. The FCA rejected this, on the ground that there is no reason in principle why the doctrine should not apply [146]. It may generally be easier to establish demonstrate utility in the context of mechanical inventions, but I cannot think of any reason why a mechanical patentee should be precluded from establishing utility on the basis of sound prediction if it cannot establish demonstrated utility. As Snider J noted in Imatinib 2013 FC 141 [164], sound prediction “is a way of showing that an invention is useful when the invention has not been directly demonstrated to work. Its introduction into Canadian law was not, as I understand it, to give a crushing hammer to those who challenge patents.” (See more here.)

Why then was Eurocopter arguing that the doctrine does not apply? The answer is that sound prediction has become a potent weapon for attacking patents, even though Snider J was right to say that it was not originally intended as such. The reason for this is the requirement that the factual basis for sound prediction must be disclosed in the patent if, and only if, utility is based on sound prediction. This point has previously been affirmed by the FCA in the Raloxifene 2009 FCA 97 [15] and Atomoxetine 2011 FCA 220 [46-47] decisions. The consequence is to draw a sharp distinction between demonstrated utility and sound prediction, with more stringent requirements for the latter. I have argued that this distinction is inconsistent with the Act, as there is no statutory basis for this distinction; there is only one utility requirement in the Act, not two requirements with two different criteria.. Again, as Snider J stated in Imatinib “Sound prediction is not a free-standing statutory requirement” [164].

No Sign of Changed Approach to the Promise Doctrine in First Post-Clopidogrel FCA Decision

Eurocopter v. Bell Helicopter Textron Canada Limitée 2013 FCA 219 Mainville JA: Noël, Trudel JJA aff’g 2012 FC 113 Martineau J
            2,207,787

While the promise of the patent doctrine and the requirement that the factual basis for sound prediction be disclosed in the patent are most closely associated with pharmaceutical patents, the FCA Eurocopter decision is significant on both issues, particularly when contrasted with the recent FCA Clopidogrel decision 2012 FCA 186. This post deals with the promise doctrine, while the next will deal with the factual basis issue. An overview of the facts in Eurocopter is provided in my last post.

In Clopidogrel decision, as discussed here, the FCA re-affirmed the promise doctrine, but it also adopted a relatively restrictive approach to construing the promise, emphasizing repeatedly that the promise must be “explicit,” and also noting that it should not be assumed that every patent has a promise. The question remains as to whether this change in tone will generate a real change in practice. Eurocopter is the first FCA decision since Clopidogrel to consider the issue. In Eurocopter the FCA affirmed, with little discussion, a fairly aggressive pre-Clopidogrel style of interpretation of the promise. Since Eurocopter was argued before Clopidogrel was released, this may simply reflect the fact that counsel would not have argued the Clopidogrel requirement for an express promise, and the FCA did not feel inclined to review the trial judge’s construction of the promise on its own initiative. But it is disquieting that the FCA simply accepted Martineau J’s construction, without even reference to the Clopidogrel approach.