Anticipation by Speculation Redux

Biogen Canada Inc v Taro Pharmaceuticals Inc 2020 FC 621 Manson J

           2,562,277 / fampidrine sustained release / FAMPYRA / NOC action

In Biogen v Taro, Manson J held the 277 patent, owned by Acorda and licenced to Biogen, to be anticipated and obvious, ultimately because the inventors had themselves disclosed the invention in a publicly available financial filing prior to the claim date. The novelty analysis raised the issue of anticipation by speculation that was recently addressed by the FCA in Hospira 2020 FCA 30 (here), and the obviousness discussion is the first clear application of the FCA’s holding in Hospira that the state of the art for the purpose of an obviousness attack includes all prior art, not just that which would be reasonably discoverable. Manson J rejected the argument that the claims were directed to unpatentable methods of medical treatment. This post provides an overview and deals with anticipation.

MS is a chronic disease of the nervous system. Impaired walking is one of the most commonly reported MS symptoms [12], to the extent that a standard measure of the MS patient’s disability is scored in part on the patient’s ability to walk [13]. The 277 patent relates to the use of sustained release fampidrine for improving walking in a person with multiple sclerosis (MS) [109].

Acorda conceived the idea of developing fampidrine SR as a treatment for MS. It conducted a small (36 subjects) phase 2 clinical trial, referred to as MS-F201, using doses from 10-40 mg bid, with pre-established endpoints related to mobility, including a timed 25 foot walk. The study failed, in the sense that most of the pre-established endpoints did not show any statistically significant difference as compared to placebo [33]. After re-analysis (less politely, data mining), Acorda discovered that MS-F201 showed a nominally significant difference in walking speed (as opposed to the pre-established endpoint of walking time) [34]. Acorda therefore planned a further phase 2 study, MS-F202, to evaluate walking speed [34].

In an effort to garner public investment, Acorda filed a financial document, referred to as “Acorda S-1,” with the SEC [121]. Acorda S-1 was prior art against the 277 patent [121]. Acorda S-1 disclosed the results from the MS-F201 trial, as well as the protocol for the planned MS-F202 trial [36, 122], that had not yet been undertaken.

The key question is whether Acorda S-1 anticipated the 277 patent. This raised the same question of anticipation by speculation that was addressed by the FCA in Hospira 2020 FCA 30 (here):

[126] The main thrust of Biogen’s novelty argument is that use claims, such as those at issue in the 277 Patent, are inextricably linked with their utility such that they are inventive because they convey “new knowledge to effect a desired result”. In this case, the claimed invention is the knowledge that 10 mg bid of fampridine SR results in a statistically significant and clinically meaningful improvement in walking in MS patients with walking disability. Disclosure of the MS-F202 protocol alone, without disclosure of its results, does not satisfy the disclosure requirement of the anticipation analysis.

Manson J rejected this, holding that disclosure of MS-F202 study design was anticipatory even if it had not actually been performed as the relevant time (citations omitted):

[138] Despite the lack of results from the MS-F202 study, the disclosure requirement is satisfied if performing what is described in the prior art reference would necessarily result in infringement. Performing the MS-F202 study protocol would necessarily result in infringement of the 277 Patent, and hence the disclosure requirement is satisfied for claims 17, 18, 31, and 32.

In this respect, as Manson J noted, the reference to MS-F202 “is similar to the allegedly anticipatory art in Hospira [2020 FCA 30]” [133]. I agree with Manson J that the cases are similar on this point, and I think he was bound to follow Hospira. (My sense is that Manson J agreed with the Hospira holding, as he cited it as “similar” rather than binding.)

With that said, I had and continue to have considerable difficulty with the holding in Hospira, which I described as “anticipation by suggestion.” The general thrust of Biogen’s argument is that a patent can’t be granted for a new use for a known product unless the use is demonstrated or soundly predicted: “The public should not be expected to pay an elevated price in exchange for speculation” Wellcome / AZT 2002 SCC 77 [37], even if that speculation subsequently turns out to be correct. If the prior art was not enough to support a patent, then it should not be enough to anticipate. So, on “the ‘unquestionable authority’ of Lord Westbury in Hills v. Evans [(1862), 45 ER 1195]” (Sanofi 2008 SCC 61, [24]), applying the law to the facts,

[The prior art Heard specification] unquestionably, according to the rules which we have endeavoured to ascertain, is not such information as will vitiate a subsequent patent, for it is not such information as will be sufficient to support a patent.

If an experiment is necessary to support a patent, it is perfectly clear that the results are required; a proposal to carry out the experiment is not enough: Wellcome / AZT 2002 SCC 77 [72]. So, as the SCC noted in Wellcome / AZT at [64], in Beecham v Bristol Labs [1978] R.P.C. 521 (HL), “claims in respect of a semi-synthetic penicillin were invalidated as being little more than an announcement of a research project.” (p. 570).

The reason, as Lord Westbury explained at 1201, is that a proposal “adds nothing to the real stock of practical knowledge of mankind, and ought not to derogate from the validity and the benefit of a subsequent invention.” This all flows from the principle that an invention is information, and to anticipate, the information provided by the prior art must be equal to that provided by the patent. This is the point of the requirement of enabling disclosure. As Lord Hoffmann stated in Merrell Dow [1995] UKHL 14 [28]:

An invention is a piece of information. Making matter available to the public within the meaning of [the statutory novelty provision] therefore requires the communication of information.

Again, per Lord Westbury at 1200, “the information as to the alleged invention given by the prior publication must, for the purposes of practical utility, be equal to that given by the subsequent patent,” and “the prior knowledge of an invention to avoid a patent must be knowledge equal to that required to be given by a specification.” Statements to that effect are too numerous to cite in this post.

To the same effect, is the statement of Viscount Dunedin in Pope Appliance [1929] AC 269, 276: the general rule “may be expressed thus: Would a man who was grappling with the problem solved by the patent attacked, and having no knowledge of that patent, if he had had the alleged anticipation in his hand, have said, ‘That gives me what I wish?’”; and at 282 “Does the man attacking the problem find what he wants as a solution in the prior so-called anticipations?” When the prior art is a suggestion to carry out an experiment, the answer to these questions must be “no,” because the experiment remains to be carried out before the problem is solved.

Returning to the case at hand, if the patentee could not have been granted on the basis of the information disclosed in Acorda S-1, then Acorda S-1 should not anticipate, because Acorda S-1 does not put the public in possession of the invention. The rule is that infringement if after constitutes anticipation if before; but I don’t see that that rule implies that Acorda S-1 anticipates: actually performing the MS-F202 protocol would infringe; but proposing the MS-F202 protocol would not.

With that said, I’ve consulted with some US colleagues and it looks like the US Fed Cir adopts much the same approach as Hospira: see Rasmusson v SmithKline Beecham Corp., 413 F.3d 1318 (Fed Cir 2005) and Bristol-Myers Squibb Co. v. Ben Venue Laboratories, Inc. 246 F.3d 1368 (Fed Cir 2001). The US cases don’t make sense to me either, but there they are.

While Biogen v Taro illustrates and applies the Hospira principle of anticipation by speculation, it is not a particularly clear example because there are a number of problems with the prior art disclosure.
So, Acorda S-1 reported the results of MS-F201 as follows:

The clinical trial demonstrated that doses up to 25 mg twice a day were well tolerated, and were associated with statistically significant improvements in walking speed and leg muscle strength.

A key element from the claim was to the use of fampridine SR at a dose of 10mg bid in treating a patient with MS “for improving walking in a statistically significant way” [109], and Acorda S-1 on its face disclosed that element [128]. However, so far as I can tell, the MS-F201 trial didn’t really show a statistically significant improvement in walking speed — it just showed that if you parse a data set enough ways, you’ll eventually find some correlation with p<0.05. I expect that’s why Manson J referred to it as showing a “nominally” significant difference in walking speed [34]. In any event, Acorda S-1 expressly stated that the results were statistically significant, and it’s not clear to me whether the data mining behind that result was reported in Acorda S-1.

Further, the essential elements of the claim also included the use of fampridine SR for a time period of at least two weeks at a unit dose of 10 mg bid [109], and it seems that Biogen was trying “a detailed parsing” of the claim elements in an effort to show that the two-week dosing period and the 10 mg bid dosage were not disclosed [129]. But the 10 mg bid dosage was disclosed by the MS-F201 trial, which also included a 10mb bid regime, so it was only the two-week time period that was not disclosed [131]. It is apparently at this point that the MS-F202 trial became relevant. But the MS-F202 trial also failed in terms of predefined endpoints [40] and “showed no meaningful difference in walking speed between the 10 mg, 15 mg and 20 mg bid doses” [43], so I’m not sure what Biogen’s basis was for arguing that MS-F202 revealed anything more than MS-F201.

Thus, while Biogen v Taro illustrates and applies the Hospira principle of anticipation by suggestion, it doesn’t really advance the law on this point. The facts don’t provide a particularly clean illustration of the issues because both the MS-F201 and MS-F202 studies were a fiesta of data-mining. I’d almost be inclined to say that they weren’t anticipatory because a skilled person wouldn’t believe anything that was said in either of them; but that’s not an issue that was raised in argument. I’ll end this post with one more quotation from Lord Westbury in Hill v Evans at 1201:

Now it is an admitted fact that that proposition [in the prior art] is untrue; it is a clear result, therefore, that the proposition would only mislead the individual who relied upon it.