The Promise of the Patent Must Be Explicit

Sanofi-Aventis v Apotex Inc / clopidogrel 2013 FCA 186 Pelletier JA, concurred in by Noël, Gauthier JJA; separate concurring reasons by Gauthier JA, rev’g 2011 FC 1486 Boivin J
            1,336,777 – PLAVIX

In Canadian law, utility is measured by the promise of the patent:

Where the specification does not promise a specific result, no particular level of utility is required; a "mere scintilla" of utility will suffice. However, where the specification sets out an explicit "promise", utility will be measured against that promise. The question is whether the invention does what the patent promises it will do.

[47, original emphasis] quoting Olanzapine (No 1), 2010 FCA 197 [76]

This doctrine has resulted in a number of pharmaceutical patents being invalidated, and in my article “The False Doctrine of False Promise,” (2013) 29 CIPR 3 (draft version available here), I argued that the promise doctrine is both unsound in principle and contrary to the Canadian Patent Act. The FCA is evidently of the contrary opinion, as the Court expressly and emphatically reaffirmed that “The promise of the patent is the standard against which the utility of the invention described in the patent is measured,” [47]: and see [48], quoted above. The FCA also expressly affirmed what I have argued is the most objectionable aspect of the promise doctrine: “An inventor whose invention is described in a patent which would otherwise be valid can nonetheless promise more for his invention than required by the Act so as to render his patent invalid” [54].

However, the FCA did significantly modify the doctrine, by emphasizing that the promise must be explicit, and that not every patent necessarily contains a promise (my emphasis):

[49] If the inventor does not make an explicit promise of a specific result, the test for utility is a “mere scintilla” of utility. If, on the other hand, the inventor makes an explicit promise of a specific result, then utility will be assessed by reference to the terms of the explicit promise.

[50] When this Court said at paragraph 80 of Olanzapine, cited above, that the promise of the patent must be ascertained, it should not be taken to have assumed that every patent contains an explicit promise of a specific result since, subject to what is said below with respect to selection patents, there is no obligation on the part of the inventor to disclose the utility of his invention in the patent. In Olanzapine, the Court was simply indicating that the firs step in assessing utility was to determine the standard against which utility will be measured. This requires the Court to construe the patent to determine if a person skilled in the art would understand it to contain an explicit promise that the invention will achieve a specific result. If so, the inventor will be held to that promise. If there is no explicit promise of a specific result, then a mere scintilla of utility will do.

Moreover, on the facts, the FCA reversed Boivin J on the basis that “He erred in law in reading into the ‘777 patent a promise for use in humans on the basis of inferences, in the absence of language at least as clear and unambiguous as that used to establish the advantages of the selection over the compounds of the genus patent” [66]. The FCA held that the patent made no promise [71], and therefore the fact that the invention had sufficient utility to support a selection patent was enough.

While the FCA Clopidogrel decision is formally consistent with the prior case law, the decision departs from the current practice both in the Court’s insistence that the promise must be explicit for it to serve as the touchstone for utility, and in emphasizing that not every patent has such a promise. The focus in previous cases has been to determine what the promise of the patent was, and not whether the patent contains a promise at all. So, in Olanzapine (No 1) the FCA said “The promise of the patent must be ascertained, [80] and “the promise of the patent is to be ascertained at the outset of an analysis with respect to utility” [93]. Clopidogrel implies that this is no longer entirely accurate, as it must first be ascertained whether the patent has a promise at all. Because the of need to ascertain the promise, if the language of the specification was not express, it would be finely parsed to extract a promise. Clopidogrel holds that this fine parsing is also misguided.

If Clopidogrel represents the law going forward, we should expect to see utility measured by the promise less often, and consequently, fewer utility attacks will be successful. The crucial question will be how explicit a promise must be in order to establish the standard for utility. Given that the doctrine was specifically affirmed, and given that the specification never has a statement beginning “This patent promises...”, I expect that direct statements of use such as “The compounds of this invention . . . are useful in the treatment of high blood pressure [and are] useful in the treatment of cardiovascular disorders and particularly mammalian hypertension” will be treated as an explicit promise: see Sanofi v Apotex / Ramipril 2009 FC 676 [122]. But where exactly the line of “explicitness” will be drawn is difficult to predict for any given patent. As I noted in my False Promise article at 41, “The root of the difficulty is that, under the promise of the patent doctrine, the utility of the invention is defined in the disclosure. The traditional purpose of the disclosure is to describe how to make and use the invention.” This disconnect remains, even if the promise must be explicit.

I said “if Clopidogrel represents the law going forward,” because it is not clear whether the views of this panel represent the views of the Court as a whole, particularly in light of Pfizer v Apotex / latanoprost (NOC) 2011 FCA 236 in which Trudel JA (Sharlow, Stratas JJA concurring), very aggressively implied a promise which was certainly not explicit in the patent. Latanoprost and Clopidogrel simply cannot be reconciled, and Clopidogrel made no attempt to do so; the Latanoprost decision was never mentioned. The question is whether we have a split in the FCA – the two decisions did not share any judges – or whether Clopidogrel represents a new consensus. If the former, then the outcome of promise cases on appeal will depend on the particular panel of the FCA. I note that the two other FCA decisions since Latanoprost endorsed a modest interpretation of the patent: Anastrozole (NOC) 2012 FCA 109 Evans JA: Sharlow, Dawson JJA (blogged here) and Donepezil (NOC) 2012 FCA 103 Mainville JA: Sharlow, Gauthier JJA (blogged here), in which the FCA remarked that “the jurisprudence does not permit an unescorted and unchaperoned romp through the disclosure” [57]. This suggests that Clopidogrel represents the culmination of a trend, but I am hesitant to draw firm conclusions; with a such a small sample size, it may be that this is not a trend at all, but a reflection of the composition of the particular panels.

On the whole, the FCA Clopidogrel decision is a very welcome development which should restrain some of the worst excesses of the false promise doctrine. However, restraining those excesses is as far as the FCA was willing to go. This is the most extensive consideration of the promise of the patent doctrine that the FCA has undertaken. That the Court reaffirmed the doctrine at the same time that it significantly restricted it, means that there is no question but that the promise of the patent doctrine will remain part of Canadian law, at least until the SCC might choose to deal with it. That is unfortunate, as I remain of the view, expressed in my False Promise article, that the doctrine is unjustifiable in any version; but as the saying goes, half a loaf is better than none.

Restrictive Interpretation of the Obvious to Try Test

Sanofi-Aventis v Apotex Inc / clopidogrel 2013 FCA 186 Pelletier JA, concurred in by Noël, Gauthier JJA; separate concurring reasons by Gauthier JA, rev’g 2011 FC 1486 Boivin J
            1,336,777 – PLAVIX

There is a great deal packed into this FCA decision holding Sanofi’s clopidogrel patent to be valid and infringed. The NOC proceedings respecting clopidogrel had culminated in Sanofi 2008 SCC 61 (referred to by the FCA as Plavix), in which the SCC had held that the ‘777 patent was neither anticipated nor obvious. Apotex subsequently sought a declaration of invalidity, which was consolidated with Sanofi’s infringement action. In the FC, Boivin J held the patent invalid for lack of utility and as being obvious. The FCA has now reversed on both these points. The FCA decision raises all of the most difficult doctrines in Canadian pharmaceutical patent law; the promise of the patent, selection patents, and the “obvious to try” analysis are all discussed at length, and the requirement that the factual basis for sound prediction be disclosed in the patent is touched on in Gauthier J’s concurrence. This post deals with the obviousness issue.

Time to Abandon the Doctrine of Selection Patents?

Hoffman-La Roche Ltd v Apotex Inc / valganciclovir (NOC) 2013 FC 718, Kane J
            2,154,721 VALCYTE

An issue discussed at length by Kane J in valganciclovir was whether the ‘721 patent is a selection patent [133]-[178]. Kane J acknowledged that “A selection patent is like all other patents; the same principles will apply” [140], but it is not clear that this was true in her analysis. She stated that “Although it is not essential that I make a finding that the ‘721 is or is not a selection patent, as there is no attack on utility, the characterisation as a selection patent will inform the analysis, particularly with respect to anticipation and obviousness” [134]. This implies that it would be essential to decide whether the ‘721 patent was a selection patent if utility were in issue. Moreover, while she concluded that the ‘721 patent was not a selection patent, and she stated she would have come to the same results with respect to anticipation and obviousness even if it had held that it was a selection patent, her analysis of those issues was substantively affected by her characterization, which, as she put it would “inform the analysis” [134], [140].

The Inventive Concept as the Focus of the Obviousness Analysis

Hoffman-La Roche Ltd v Apotex Inc / valganciclovir (NOC) 2013 FC 718, Kane J
            2,154,721 VALCYTE

Yesterday’s post discussed anticipation. This post discusses obviousness, which was the other main validity issue. Ultimately, obviousness turned largely on the facts, but the analysis raised an important question as to whether the focus of the obviousness inquiry should be the invention as claimed, the inventive concept of the specification, or the inventive concept of the individual claims.

Valganciclovir is the mono-L-valine ester of ganciclovir [32]. Ganciclovir is an antiviral nucleoside which effective in the treatment of certain herpes viruses [4]. Acyclovir is another antiviral nucleoside which effective in the treatment of certain herpes viruses. Both acyclovir and ganciclovir suffered from poor bioavailability when taken orally. In the prior art the L-valyl ester of acyclovir, valaciclovir, was the acyclovir prodrug with the best oral bioavailability. (As I understand it, an “L-valyl” ester and an “L-valine” ester are the same.) The question was whether the good bioavailability of the L-valyl ester of acyclovir made it obvious to develop the same ester of ganciclovir in order to improve bioavailability. Kane J held it did. This conclusion turned largely on the facts.

Genus Anticipates Species

Hoffman-La Roche Ltd v Apotex Inc / valganciclovir (NOC) 2013 FC 718, Kane J
            2,154,721 VALCYTE

Roche’s ‘721 patent claims the compound valganciclovir in various forms. Valganciclovir is the mono-L-valine ester of ganciclovir [32]. It is a pro-drug of ganciclovir, and has better bioavailability than ganciclovir when orally administered. Ganciclovir was known in the prior art as the leading drug for the treatment of certain herpes viruses, particularly cytomegalovirus [4]. The main issues were whether the ‘721 patent was invalid for anticipation and obviousness. Kane J held that they were.

Kane J’s anticipation analysis is, in my view, very problematic. The issue was whether the ‘721 patent was anticipated by EP0375329. EP 329 disclosed and claimed the amino acid esters of ganciclovir. It was not disputed that mono-L-valine ester was encompassed with the approximately 500,000 compounds described by EP 329, and the question was whether the disclosure was sufficiently specific to anticipate. It was not disputed that valganciclovir was not disclosed in any of the examples of EP 329 [231]. Kane J did not expressly identify the anticipatory reference, and the only reference to valine and mono-esters that I have been able to find in EP 329 is the following (p 3):

The amino acid acyl residue of the above compounds [the general formula of the disclosed compounds] according to the invention may be derived for example from naturally ocurring [sic] amino acids, preferably neutral amino acids i.e. amino acids with one amino group and one carboxyl group. Examples of preferred amino acids include aliphatic acids, e.g., containing up to 6 carbon atoms such as glycine, alanine, valine and isoleucine. The amino acid esters according to the invention includes the mono- and di-esters of the compound of formula (I). The amino acids may be D-, L-and DL amino acids, with the L-amino acids being most preferred.

Examples of preferred compounds of formula (I) above include those of Examples 1 to 6.

The above-mentioned physiologically acceptable salts are preferably acid addition salts derived from an appropriate acid, e.g ., hydrochloric. sulphuric. phosphoric. maleic, fumaric, citric, tartaric, lactic or acetic acid. The above-defined amino acid esters of formula (I) and their salts which are hereinafter referred to as the compounds according to the invention. . .

The mono-L-valine ester is never singled out, and the mono-esters and valine are independently enumerated as part of separate lists. There was no disclosure of any special advantages that the mono-L-valine ester might have over the other amino acid esters that were disclosed.

Compare this with Sanofi 2008 SCC 61, aff’g 2006 FCA 421, in which the question was whether clopidogrel bisulfate, claimed in 1,336,777, was novel over the disclosure of 1,194,875. The ‘875 patent disclosed over 250,000 racemic compounds, of which only 21 individual compounds were specifically identified [FCA 9]. Clopidogrel bisulfate is the dextro-rotatory isomer of one of the specifically identified racemates, namely derivative 1, which was disclosed as Example 1 of the ‘875 patent [SCC 3, 6], [FCA 9]. Not only did the ‘875 specifically disclose the racemate, it also expressly stated at p 1 that it encompassed the enantiomers:

Ces composés comportant un carbone assumétrique peuvent exister sous la forme de 2 énantiomères. L’invention concerne aussi bien chacun des énantiomères que leur mélange.

In Sanofi, of course, the SCC held that clopidogrel bisulfate was not anticipated by the ‘875 disclosure. With respect, I find it impossible to distinguish Sanofi from this case.

Kane J held that:

[234] EP 329 discloses both the mono- and bis-ester, although it prefers and exemplified only the bis-esters. On a balance of probabilities, the POSITA would read EP 329 and understand that it was teaching both mono- and bis-esters with both promising improved bioavailability over ganciclovir.

With respect, it is not correct to say that EP 329 “discloses” the mono-ester, at least if that word is used in the sense of the two-part test for anticipation set out in Sanofi, which requires disclosure and enablement. It is entirely true that EP 329 taught that both the mono- and bis-esters of the amino acid esters would have improved bioavailability over ganciclovir, but it taught this in respect of the entire class. The question is whether EP 329 taught that the mono-L-valine ester specifically has particularly good bioavailability, and the answer is that it does not. That ester is not singled out in any way from the other amino acid esters disclosed in EP 329.

It is also true is that EP 329 describes the mono-ester, and that its disclosure encompasses the mono-ester, just as it was true that in Sanofi the ‘875 patent described and encompassed the enantiomers. But one the major holdings in Sanofi is that a species is not anticipated by a genus simply because the genus encompasses the species. While Sanofi set out a two-part test for anticipation, on the facts, the ‘777 patent was held to be novel on the first part, disclosure:

[41] Since the ‘875 patent did not disclose the special advantages of the dextro-rotatory isomer and of its bisulfate salt, as compared to the levo-rotatory isomer or the racemate and their salts, or the other compounds made and tested or otherwise referred to in the ‘875 patent, the invention of the ‘777 patent cannot be said to have been disclosed and therefore it cannot be said to have been anticipated.

It was not even suggested by any of the experts for Apotex in this case that EP 329 disclosed any special advantages of the mono-L-valine ester as compared with the other esters disclosed by that document.

Kane J’s analysis appears to go astray on two points. She understood Sanofi as “rejecting the stringent test which required that the exact invention had already been made and disclosed” [202], and similarly, she relied on Hughes J’s summary of the law of anticipation in 2008 FC 1359 [75] that “The disclosure does not have to be an ‘exact description’ of the claimed invention. The disclosure must be sufficient so that when read by a person skilled in the art willing to understand what is being said, it can be understood without trial and error” [209]. It is true that in Sanofi the SCC said that “in light of recent jurisprudence, I am of the respectful opinion that the applications judge overstated the stringency of the test for anticipation that the ‘exact invention’ has already been made and publicly disclosed” [23]. However, the SCC did not reject the law as set out in Beloit and General Tire, which it described as having been accepted in Canada “without reservation” [22]. What was new in Sanofi was to explicitly separate disclosure and enablement. The SCC described this as “is a refinement of the approach set out in Beloit.” The test for disclosure remains stringent: “No trial and error is permitted” [32]. It is in the test for enablement – which, as the SCC pointed out, was not at issue in Beloit [28] – that the stringency is relaxed. Some trial and error is permitted, though “the skilled person must still be able to perform or make the invention of the second patent without undue burden” [33].

The second point emphasized by Kane J is that the civil burden of proof applies rather than the criminal burden [211], [209.6], [207]. This is no doubt true, but Kane J seems to have understood this, combined with the purportedly relaxed standard for anticiaption set out in Sanofi, to mean that the claimed invention is anticipated if a skilled person following the prior art might produce the claimed invention. This is not what the civil standard implies. The point that the civil standard applies arose ultimately from Actavis UK Ltd v Janssen Pharmaceutica NV [2008] EWHC 1422 (Pat) [cited by Hughes J in 2008 FC 1359 [69]]. In Actavis the claimed invention was the use of RSSS isomers of specified blood pressure drugs to potentiate other, non-RSSS, blood pressure drugs [53]. The anticipation point arose because the prior art compound asserted to anticipate contained the RSSS isomer, as well as the SRRR isomer and the other stereoisomeric pair of RSRR and SRSS [82]. It was established that when a medicine is made from only RSSS and SRRR the former will potentiate the latter, but there was some question on the facts as to whether RSSS would potentiate SRRR when it was combined with the other stereoisomeric pair as well [83]. It was on that point that Floyd J held that the civil standard of proof was to be applied. This was not a case where the prior art disclosed a variety of compounds, and the question was whether one picked at random would fall within the claims. Rather, there was one compound, which either did or did not have the claimed property, and it is that question which is to be decided on the balance of probabilities.

I cannot draw a straight line from these two points to Kane J’s conclusion in respect of anticipation, but between the two she seems to have understood the test for disclosure as being relatively relaxed. Qualitative descriptions of a test as being “relaxed” or “stringent” are not always a reliable guide, and however we characterize the test for anticipation, Sanofi and Roche/ valganciclovir (NOC) are very similar in their facts, and cannot be reconciled on this point.

Miscellaneous Issues in Lovastatin Damages

Merck & Co Inc v Apotex Inc / lovastatin damages 2013 FC 751 Snider J
            1,161,380 MEVACOR

Yesterday’s post provided an overview of the facts and issues in Lovastatin damages, and discussed the issue of whether the availability of a non-infringing alternative is relevant to the assessment of damages.* This post describes the remaining issues.

Pre-Expiry Replacement Sales – Deductibility of Royalties Paid to MACI
A subsidiary issue in respect of pre-expiry sales was whether Merck Canada’s lost profit damages should include the 8.5% royalty normally payable by Merck Canada to MACI [122-142]. (Again, quantum was agreed.) If Merck Canada had made the lost sales, it would have been required to pay that royalty to MACI, thus reducing its profit. Apotex argued that it should be therefore be deducted, to put Merck Canada in the same position it would have been in but for the sales [126]. Merck argued that it would be required to pay the royalty to MACI on the damages award, so if the royalty was deducted from its lost profit damages, it would not be made whole; it would in effect pay the royalty twice, once through the deduction in damages, and then again to MACI.

Snider J held for Apotex on this point [142]. The crux of her reasoning was that

[130] In my view, the MACI Royalty Agreement is clear and unambiguous, creating no obligation to pay upon receipt of lump sum damages. Since there is no obligation to pay any royalty on the an award of legal damages, the MACI Royalty is properly characterized as an expense that would have been incurred in the hypothetical scenario but saved because of the infringement. As such, it should be deducted in the calculation of Merck Canada’s lost profits.

This implies that if the contract does specify that a royalty is to be paid on an award of legal damages, then it will be added to the damages. The lesson here is straightforward: if the licensee intends to pay royalties on damages to the licensor, the contract should say so expressly.

On this point Merck also argued that the royalty was payable as a matter of law, on the basis of the surrogatum principle, applied by courts in the tax context, that lump sum damages take on the character of the interest settled. Snider J rejected this argument on the basis that this was a principle specific to tax law, which could not create obligations between the parties to the contract [141]. Put another way, if the contract had expressly specified that the royalty was not payable on a damages award, then even if the damages award were nonetheless treated as a receipt on sales for tax purposes, this would not create an obligation on the licensee to pay the royalty to the licensor.

Reasonable Royalty
Apotex argued that Merck was entitled only to a reasonable royalty on the pre-expiry replacement sales [15]. Given her conclusion that Merck was entitled to its lost profits on those sales, Snider J did not have to, and chose not to, calculate the reasonable royalty that would be owing in the event she was wrong that Apotex could not raise the non-infringing alternative defence [148]. However, she did describe two “conceptual elements” of the approach that she would apply in making that calculation, in the event that it was referred back to her on appeal.

The first element was that the reasonable royalty should be calculated on the basis of a one-time negotiation on the eve of the first infringement [155], [156 - 62]. The reason the date of the hypothetical negotiation matters is that the bargaining position of the parties might change in the interim: see eg Applied Medical Resources Corp. v US Surgical Corp 435 F.3d 1356 (Fed Cir, 2006). In this case that means that the reasonable royalty on the Phase 3 shipments from Blue Treasure, which commenced in March of 1998, would be assessed at the time of production of the first batch, CR0157, in 1996, notwithstanding that Merck was awarded lost profits for this batch, and not a reasonable royalty.

The second element is “the use of a framework taking into account the hypothetical licensee's maximum willingness to pay (MWP) and hypothetical licensor's willingness to accept (MWA) methodology” [155], with the assumption that the parties with equal bargaining power would then split the difference equally [168]. This is essentially the Nash Bargaining Solution to a negotiation. (Compare Suffolk Tech. LLC v. AOL Inc., No. 1:12cv625 (E.D. Va. April 12, 2013 (Dkt. No. 518), rejecting the use of the Nash Bargaining solution under Daubert, as not sufficiently tied to the facts of the case; and VirnetX Inc. v. Cisco Systems, Inc., Case No. 6:10-cv-00417-LED (Doc. No. 745), accepting the use of the Nash Solution.) The assessment of MWP and MWA is the central question when using this method. Snider J held that the patentee’s MWA sets a floor on the reasonable royalty; so, if the patentee’s MWA is greater than the infringer’s MWP, the royalty would be set equal to the MWA, notwithstanding that that is more than the infringer would have been willing to pay [174].

Reasonable Royalty – Pre-Expiry Export Sales
Merck acknowledged that it would not have made the post-expiry sales made by Apotex, and Merck claimed only a reasonable royalty. The quantity of sales was not disputed, and the parties agreed that the reasonable royalty would split the difference between Apotex’s profits costs using the infringing and non-infringing processes; that is, the parties would split equally the additional profits made available by the use of the patented invention as compared with the best available alternative [179]. Snider J used the same methodology to award a reasonable royalty in respect of post-expiry sales of infringing product.

Lost Profits – Post-Expiry Ramp-Up
This head of damages is often referred to as “springboard” damages. A generic cannot normally make sales instantaneously on expiry of the patent, as some time is required to physically ship product to pharmacies and so on. It is established in UK and US law that springboard damages are recoverable, and Snider J acknowledged that they are likely recoverable in Canadian law as well [205]: and see Siebrasse et al “Damages Calculations in Intellectual Property Cases in Canada”, (2008) 24 CIPR 153, §2.4. She nonetheless denied Merck’s claim for two reasons.

The first was that Merck did not provide adequate notice that springboard damages would be claimed. Snider J concluded that the issue had been raised for the first time in Merck’s opening argument [212], so that Apotex did not have adequate notice, and she would have denied the claim for that reason alone [215].

Snider J also held that Merck’s evidence establishing springboard damages was not adequate. One key problem is that Apotex did make substantial non-infringing sales prior to patent expiry, and so it would have had distribution networks in place [226]. Consequently, Merck's assumption that Apotex’s ramp-up in the hypothetical post-expiry world in 2001 would have been the same as it’s actual ramp-up in 1997 was not sound [221].

Merck US
Merck US was the patent owner, but it was not the licensor, as it had granted the exclusive right to grant licenses to MACI. However, Merck US did make a profit by supplying Merck Canada with lovastatin, and consequently it’s profits were reduced by Apotex’s infringement. The quantum of Merck US’s lost profits was agreed, but Apotex argued that it should be restricted to nominal damages because of the exclusive licence granted to MACI. Snider J rejected this argument, noting that in the liability portion she had held that Merck US had standing as the patentee to bring an action [237]. She noted that notwithstanding the licence to MACI, Merck US remained the patentee, and there was nothing in the agreement that could be read to exclude Merck US from claiming damages due to infringement [244].

*As noted in yesterday’s post, I consulted for Apotex on the damages portion of this action.

Non-Infringing Alternative Not Considered in Assessing Damages

Merck & Co Inc v Apotex Inc / lovastatin damages 2013 FC 751 Snider J
            1,161,380 MEVACOR

Lovastatin damages raises fundamental issues as to the nature of causation in patent damages; it is the damages equivalent of Monsanto v Schmeiser 2004 SCC 34. The specific question is whether the availability of a non-infringing alternative should be considered in assessing damages for patent infringement. Snider J's answer is a resounding “No.” I consulted for Apotex in this case,* and consequently my discussion of the issues will be primarily descriptive rather than analytic.

Facts
Lovastatin damages is the damages phase of a bifurcated action. In the first phase, Lovastatin liability 2010 FC 1265, Snider J held that the ‘380 patent was valid and infringed by Apotex. (See this post for more background on the liability decision.)

The ‘380 patent is a process and product-by-process patent to lovastatin when made with the micro-organism A. terreus [see generally liability 28-39]. In anticipation of obtaining a compulsory licence, Apotex, through its subsidiary Apotex Fermentation Inc (AFI) developed an infringing process for the production of lovastatin, AFI-1, that used A. terreus. Subsequently, in anticipation of the compulsory licencing regime being replaced by the PM(NOC) Regulations, AFI developed a non-infringing process, AFI-4, which used a different organism, C. fuckelii. Some lovastatin was produced by AFI in Canada using both processes. The infringing Canadian-made lovastatin was used primarily for regulatory and experimental purposes, and so was exempt from liability [liability 625ff]. Apotex then decided to shift production to China, where the lovastatin was made by Blue Treasure, a joint venture with Chinese partners. Apotex transferred know-how related to both the AFI-1 and AFI-4 technologies to Blue Treasure. Once it became clear that Apotex would proceed under the NOC Regulations rather than by compulsory licence, Apotex insisted that Blue Treasure use the non-infringing AFI-4 technology for shipments to Canada. (The lovastatin made by the AFI-1 process was to be sold in China or other foreign markets [liability 297]). However, this was more expensive than the infringing AFI-1 process, and Snider J ultimately found that Blue Treasure had been boosting its profits by using the cheaper infringing process to make the lovastatin that it delivered to Apotex. Ultimately, most of the product sold by Apotex (71% by volume) was infringing [9]. More specifically, the AFI-4 production took place in three phases. In Phase 1 lovastatin was produced by AFI in Winnipeg. The first batch of AFI-4 lovastatin shipped by AFI to Apotex – “batch CR0157" – was contaminated with AFI-1 product, and was therefore infringing, but the remaining Phase 1 production was non-infringing. In Phase 2, from mid-1997 to January 1998, Blue Treasure produced 70 batches of lovastatin, which were entirely non-infringing. In Phase 3, from March 1998 to March 2000, Blue Treasure shipped infringing lovastatin [8]. Thus there was a relatively small initial infringing shipment, then a substantial period of non-infringing supply, followed by the majority of the infringing supply. Most of the infringing supply was sold prior to the expiry of the patent, but some was sold post-expiry.

Merck’s Supply Arrangements
The plaintiff’s supply arrangements are relevant to some of the issues. The plaintiff Merck Canada Inc (Merck Canada) was a licensee under the patent and it supplied the Canadian market. (Note that Merck Canada was known as Merck Frosst Canada Ltd at the time of the liability litigation, and it is referred to as Merck Frosst in the liability decision.) The plaintiff Merck & Co Inc (Merck US) was the owner of the patent, but Merck US had granted an exclusive licence to Merck and Company, Incorporated (MACI), and Merck Canada paid MACI an 8.5% royalty. MACI was not a party to the litigation. While Merck Canada paid a royalty to MACI, the supply chain agreement required Merck Canada to purchase its lovastatin from Merck US, so Merck US also profited from Merck Canada’s sales.

Heads of Damages
Merck Canada claimed lost profit damages on tablets that it would have sold to replace the infringing Apo-lovastatin tablets (“Pre-Expiry Replacement Sales”).With respect to post-expiry sales, Merck argued that Apotex would have faced a ramp-up process, and so would not have sold as many tablets post-expiry if it had not infringed. Merck Canada claimed lost profits on the sales it would have captured in this ramp-up period (“Post-Expiry Ramp-up Sales”). A reasonable royalty was claimed in the alternative. Merck US claimed for its lost profits on its supply of lovastatin API to Merck Canada in respect of the same tables, both pre- and post-expiry [11].

Merck also claimed a reasonable royalty on sales it conceded it would not have made, namely infringing Apo-lovastatin tablets sold into the export market pre- and post-expiry (Export Tablets); and infringing Apo-lovastatin tablets sold domestically after the '380 Patent expired, outside the ramp-up period (the Post-Expiry Replacement Tablets) [12].

Merck Canada also requested that its lost profits award include an amount to reflect the 8.5% royalty payable to MACI [13]. Presumably this is why MACI was not a party; on Merck’s view, Merck Canada was required to pay the royalty on the damages collected by Merck Canada in the event of success in litigation, so that MACI would have suffered no loss [127].

Pre-Expiry Replacement Sales – Non-Infringing Alternative
The main issue with respect to the pre-expiry sales, and indeed the main issue in the case in respect of both the space devoted to it by Snider J, and the quantum at stake, was whether Apotex was able to raise the defence that it had a non-infringing alternative available to it, namely the capacity to use the AFI-4 process to supply the market [30-120]. Apotex argued that but for the infringement, it would have produced the same amount of non-infringing material using the AFI-4 process, which it would have sold for the same price, and so Merck would have lost exactly the same market share had Apotex not infringed [33]. Merck’s lost profits were therefore not caused by the infringement. (The exception was with respect to the initial batch CR0157; Apotex conceded that it would not have been able to use its AFI-4 process to make that batch, and Merck was therefore entitled to its lost profits [145].) (Note that the quantum of Merck’s profits, had it made the pre-expiry replacement sales, was agreed.)

This was resolved as a pure question of law. In addition to the legal question of whether the non-infringing alternative (NIA) defence is available as a matter of law, the point turns on the factual issue of whether Apotex could have made and sold non-infringing tablets in place of the infringing tablets. Merck also argued that the fact that Apotex had breached its undertaking in the NOA that it would not infringe the patent, should have consequences in the assessment of damages in the infringement action [34]. SniderJ resolved these issues in Apotex’s favour [37-39]. That is, Snider J held that as a matter of fact, but for the infringement, Apotex would have made the same sales it in fact made, but using the non-infringing process. "The availability of the NIA defence at law is therefore the determinative question” [40].

The leading UK case is United Horse Shoe & Nail Co v Stewart & Co, (1888), 5 RPC 260 (HL), which holds that the availability of a non-infringing alternative is irrelevant to the calculation of damages. As Snider J notes, subsequent UK law affirms this position: [63]-[69]. US law, on the other hand, is equally clearly to the contrary: an available non-infringing alternative must be considered [92]. Canadian case law on point consists only of Domco Industries (1983), 76 CPR (2d) 70 rev’d on other grounds, (1986), 10 CPR (3d) 53, accepting United Horse Shoe, and Snider J’s own obiter comments in and Jay-Lor 2007 FC 358 approving the relevant passage from Domco [71], [73]. As Snider J noted, this “may not constitute the highest authority” [74].

Given the state of Canadian authority, a key question was the impact of the SCC decision in Schmeiser 2004 SCC 34, in which the SCC stated that in calculating an accounting of profits, “[a] comparison is to be made between the defendant's profit attributable to the invention and his profit had he used the best non-infringing option” [102] (emphasis added). Does this principle apply also to damages? In my article, “A Remedial Benefit-Based Approach to the Innocent-User Problem in the Patenting of Higher Life Forms” (2004), 20 CIPR 79, esp at 91-95, which was cited by the SCC on this point in Schmeiser, I argued that there is one general principle of causation which applies across all areas of law. Snider J disagreed. In her view, the applicable principles in respect of patent damages are different from those in an accounting [90]. Snider J acknowledged that the Schmeiser principle was also implicitly applied by the SCC in Cadbury Schweppes [1999] 1 SCR 142 [55], and by Snider J herself in [107] in Teva Ramipril, 2012 FC 552 in assessing damages under s 8 of the NOC Regulations, but she held that the principles of patent damages are different from these as well [55], [108]. The question is why patent damages should be different from these other monetary remedies. Snider J’s answer is that it is necessary to ignore the non-infringing alternative in order to provide adequate deterrence: [188]-[119].

My view, expressed in article “A Remedial Benefit-Based Approach,” is that United Horse Shoe ignores the generally applicable causation requirement, and it “must therefore be considered to be clearly in error” (p95). Similarly, in my co-authored article, “Damages Calculations in Intellectual Property Cases in Canada”, (2008) 24 CIPR 153, we noted that “United Horse-Shoe seems to contradict the approach to causation taken by the Supreme Court of Canada in the context of accounting of profits and equitable compensation remedies for intellectual property.” Snider J discussed these articles at some length [98]-[106], but concluded that on this point I am “‘the lone voice in the wilderness’ supporting the adoption of the [non-infringing alternative] defence” [106]. This is a very complementary criticism, as the reference is to John the Baptist saying "I am the voice of one calling in the wilderness, 'Make straight the way for the Lord.'" While I would like to be thought of as the lone voice preaching the true gospel, that is only true if the wilderness is confined to the backyard of Canadian patent damages. Outside of patent law, I am in the excellent company of the Supreme Court of Canada in Schmeiser and Cabdury Schweppes, and in the patent field, but outside of Canada, I am in the company of Chief Judges Rader (Grain Processing 185 F.3d 1341 (Fed Cir 1999) and Easterbrook (Grain Processing VIII 979 F.Supp. 1233 (ND Ind1997)), and the leading US academic on patent remedies, Thomas F. Cotter, Comparative Patent Remedies: A Legal and Economic Analysis (2013) at 187ff.

*I consulted only on the damages portion of this case, not the liability action. I wrote my post on the Relevance of Non-Infringing Alternatives to Damages before being approached by Apotex.

First Application by PAB of New Guidelines On Patentable Subject Matter

Re Application 2,306,540, CD 1334, Pitney Bowes / Postal Tax Information System
Re Application 2,344,781, CD 1336, Progressive Directrac / Insurance Monitoring System

These two decisions from the PAB and Commissioner of Patents are the first to address patentable subject matter since the FCA Amazon.com decision, 2011 FCA 328, and since the revised Guidelines on Examination Practice Respecting Computer-Implemented Inventions PN2013-03, released in March. As discussed in my previous post, the Guidelines repudiated the earlier contribution approach, which asked whether inventive contribution was itself patentable subject matter, in favour of an approach which asks whether the invention as claimed is patentable subject matter. The effect of this change is to allow a broader range of valid claims. So, on the contribution approach, a computer implemented invention in which the inventive contribution lay in an ingenious algorithm would not be patentable unless there were some ingenuity in its implementation, whereas under the more recent approach, it will be patentable if the invention is implemented in a machine or other system which is statutory subject matter, regardless of where the inventive ingenuity is found.

These decisions illustrate the approach set out in the Guidelines. In CD 1336 the inventive concept appears to be the idea of basing insurance rate on driver risk factors derived from real-time monitoring, rather than on risk categories, and providing insurance premium information back to the driver. (I say this appears to be the inventive concept, because, as is often true in judicial decisions as well, the inventive concept as stated by the PAB is basically a paraphrase of the claims, rather than an identification of the point of inventiveness which sets the invention apart from the prior art.) As I understand it, the system would in effect say to a driver “Your GPS system shows you have been speeding a lot this month. Unless you start driving at the speed limit, your premiums will go up by $X.” This idea might be characterized as a business method, but that characterization was irrelevant to the PAB’s analysis:

82. The monitoring of operating characteristics limitation being essential, the claimed subject matter not being merely abstract, and the subject matter not being otherwise excluded from patentability, the panel finds that method claims 1-9 and system claims 10-17 are directed to statutory subject matter.

That is, regardless of whether the inventive concept is a business idea, the claim is to a system, which is patentable subject matter, and not merely to the idea in the abstract. The decision in CD 1334 is very much to the same effect. (The claims at issue in CD 1336 were nonetheless refused on the basis of obviousness, while those in CD 1334 were allowed, subject to clarifying amendment.)

In my post on the FCA Amazon.com decision I argued that the crucial paragraph 63 contained two different and contradictory ideas. Consequently:

If the Patent Office wishes to refuse a patent in which the inventive step lies in a business idea, it will focus on the first aspect of [63], and say that "the only inventive aspect of the claimed invention is the [business idea]" which is to be assimilated to a mathematical formula, and is therefore unpatentable. On the other hand, if the Patent Office wishes to grant the patent, it will focus on the second aspect of [63] and says that on a purposive construction the business idea is "only one of a number of essential elements in a novel combination," just as software patents were granted on the basis that the invention pertains "to a useful system" and "more than calculations are involved."

In CD 1336, the PAB focused on the second aspect of [63], saying “The feature 'determining the cost of insuring', although a mathematical calculation "is not the whole invention but only one of a number of essential elements in a novel combination,'" with the result that the claimed invention was patentable subject matter. It appears from the Guidelines and these two decisions, that CIPO has settled on this interpretation of Amazon.com. As I argued in my post on the Guidelines, this is the correct approach, as it is consistent with the SCC Shell Oil decision.

How Abstract is Your Idea?

Ultramercial, Inc v Hulu, LLC 2010-1544 (Fed Cir 2013) on remand from 132 S.Ct. 2431 for further consideration in light of Mayo v Prometheus 132 S.Ct. 1289.

Ultramercial v Hulu (No 2) is the latest engagement in the long-running battle between the US CAFC and the USSC over the subject-matter exclusion of “laws of nature, physical phenomena, and abstract ideas.” The outcome of this round is that business methods remain patentable in the US, and the key to the validity of such patents will be the level of generality at which the basic idea underpinning the invention is described.

As I have explained in various articles and posts (here and here), starting with the 1948 Funk Brothers decision, the USSC has understood the rule against patenting abstract ideas and laws of nature as an exclusion of claims for inventions in which the inventive contribution lies in the discovery of a law of nature. So, as Breyer J put it Prometheus,

the claims inform a relevant audience about certain laws of nature; any additional steps consist of well understood, routine, conventional activity already engaged in by the scientific community; and those steps, when viewed as a whole, add nothing significant beyond the sum of their parts taken separately.

That is, the inventive concept must be found in the application of the law of nature, and not in the discovery of the law of nature. This is contrary to the Anglo-Canadian orthodoxy, which holds that the inventive contribution may lie in either the idea or the application, so long as the claim itself is to the application. So, in Hickton's Patent (1909), 26 RPC 339, 348 (CA) Fletcher Moulton LJ said that

invention may lie in the idea, and it may lie in the way in which it is carried out, and it may lie in the combination of the two; but if there is invention in the idea plus the way of carrying it out, then it is good subject-matter for Letters Patent.

Affirming this as part of Canadian law, the SCC in Shell Oil, [1982] 2 SCR 536 554, stated

A disembodied idea is not per se patentable. But it will be patentable if it has a method of practical application.

The US Federal Circuit has long held the same view as the Anglo-Canadian courts, and the story of the US law of patentable subject matter is that of the efforts of the Federal Circuit to bend the USSC jurisprudence into a more acceptable shape. The USSC position is incoherent, as “[t]he "laws of nature" are an essential part of the working of many and probably most patented inventions,” (Harvard Mouse 2002 SCC 76, [87], Binnie J), and consequently there has always been some ambiguity in the USSC case law to be exploited. Moreover, while Funk Bros, Benson and Flook represent the purest expression of the USSC’s reasoning, Diehr more closely approaching the orthodox position, and has always been the Federal Circuit’s favourite USSC decision.

Ultramercial, v Hulu (No 2), is our first look at how the Fed Cir will deal with Prometheus. The Court’s opinion, written by Rader J, is very much in the traditional Federal Circuit mold. In a passage that had me wondering if he had been reading Shell Oil, Rader CJ explained at slip op 14 that

a claim is not patent eligible only if, instead of claiming an application of an abstract idea, the claim is instead to the abstract idea itself. . . . The inquiry here is to determine on which side of the line the claim falls: does the claim cover only an abstract idea, or instead does the claim cover an application of an abstract idea?

Rader CJ rejected the notion that a claim will be unpatentable if there is an abstract idea at its core (slip op 16):

a claim may be premised on an abstract idea and, indeed, the abstract idea may be of central importance to the invention—the question for patent eligibility is whether the claim contains limitations that meaningfully tie that abstract idea to an actual application of that idea through meaningful limitations.

But how to deal with the USSC decisions? Diehr was duly relied upon, but Prometheus had to be addressed. Here Rader CJ engaged with the preemption theory that has emerged as the main rationale advanced by the USSC for its view that laws of nature cannot form the basis for a valid claim. As originally explained in Benson:

It is conceded that one may not patent an idea. But in practical effect that would be the result if the formula for converting BCD numerals to pure binary numerals were patented in this case. The mathematical formula involved here has no substantial practical application except in connection with a digital computer, which means that if the judgment below is affirmed, the patent would wholly pre-empt the mathematical formula and in practical effect would be a patent on the algorithm itself expalin

This is misguided and contrary to Anglo-Canadian law. As the SCC explained in Burton Parsons [1976] 1 SCR 555 in which “[t]he objections raised against the claims really are that, except those pertaining to some specific embodiments of the invention, the others are so framed as to cover every practical embodiment,”

If the patent is to have a practical value, it must cover all [formulations of the new type of cream] which can yield the desirable result . . . [I]f, in order to guard against possible invalidity, some area is left open between what is the invention as disclosed and what is covered by the claims, the patent may be just as worthless as if it was invalid. Everybody will be free to use the invention in the unfenced area.

Nonetheless, Rader CJ was forced to accept the preemption argument, no doubt against his better judgment. He correctly recognized that “[t]he concern . . .which has become clearer through the Supreme Court’s more recent precedents, is to deny patentability to an idea itself, rather than an application of that idea.”

[T]he Supreme Court has stated that a claim is not meaningfully limited if it merely describes an abstract idea or simply adds “apply it.” . . . We also know that, if a claim covers all practical applications of an abstract idea, it is not meaningfully limited.

The question, then, is defining the level of generality of the abstract idea that underpins the patent. The patent at issue in this case, U.S. Patent No. 7,346,545, claims in very broad terms “A method for distribution of products over the Internet,” by requiring a consumer to view an advertisement before accessing copyrighted content. The problem, as described in the patent, is that

Recently, the widespread use of the Internet has made it possible to distribute and share intellectual property in its digital form, worldwide, beyond the control of the intellectual property rights holders. . . [S]everal factors encourage the growing violation of intellectual property using the Internet. . . . Young people are becoming accustomed to getting music for free and have shown little guilt about the fact that they are breaking the law.

At the same time, advertisers today are facing very unpleasant realities in conventional advertising space. More specifically: 1) the proliferation of cable channels and attendant "channel surfing" during commercials breaks; 2) consumer habits of taping shows and then fast-forwarding through commercials during playback; 3) new hard-disk TV recorders that can scan out commercials. Further, Internet banner ads have hit new lows in responses.

On this description, the idea behind the invention is specific to the problem of the Internet, and the basic idea is that “upon logging on to an Internet web site where intellectual property products, such as musical composition or written articles, are for sale via download, a consumer may choose to obtain the right to download selected intellectual property products ('IP products') by viewing or participating in an advertiser's message or commercial, rather than by paying for the intellectual property products with cash or credit card.” If that idea is described at that level, then it appears to be preempted.

But according to the Federal Circuit, “The abstract idea at the heart of the ’545 patent, which the district court properly identified, is ‘us[ing] advertising as an exchange or currency.’” Described at this higher level of generality, the idea is clearly not preempted, because, as Rader CJ emphasized, the claim is tied in many ways to an Internet implementation.

Whether the USSC will view an Internet implementation as “meaningful” post-solution activity remains to be seen. But for the time being at least, it seems that the key to getting business method patents in the US will be to describe the basic idea at a high level of generality, so that the application in the patent is a specific instance of that idea.

First Interpretation of Rule 3.1 re Grace Period for Failed Attempt to Pay Fees

Karolinska Institutet Innovations AB v Attorney General of Canada, 2013 FC 715, Hughes J

This decision of Hughes J is the first interpretation of the grace period provided for by Rule s 3.1(1) when a “clear but unsuccessful” attempt is made to pay a fee. Hughes J’s decision was a fairly straightforward application of s 12 of the Interpretation Act, which requires that an enactment be given an interpretation “as best ensures the attainment of its objects.”

On September 4 and 15, 2008, the Applicants filed patent applications under the PCT in Sweden and the US. The crucial dates for late entry into the Canadian national phase under Rule s 58(3) (42 months, rather than 30 months), were therefore 4 and 15 March 2012. The letter requesting national phase entry was filed electronically 29 Feb 2012. However, due to a clerical error in office of the Canadian agent, the total fee was sufficient to cover only the basic fee (10(a)) and the second year maintenance fee (30(a)(i)), and not the late payment fee (11); the total tendered was $250, when it should have been $450. CIPO consequently refused entry into the national phase, on the basis that s 58(3) requires payment of fees prior to the expiry of 42 months from the priority date.

The decision concerned the application of Rule s 3.1(1) which provides for a grace period

if, before the expiry of a time limit for paying a fee set out in Schedule II, the Commissioner receives a communication in accordance with which a clear but unsuccessful attempt is made to pay the fee.

In such a case, the Commissioner is required to (“shall”) send a notice requesting payment of the deficiency: s 3.1(2). The Commissioner did not do so.

CIPO argued that it was not “clear” that the Applicant intended to pay the late payment fee. Hughes J’s description of CIPO’s argument was very brief: “Counsel relies on the words ‘the fee’ in Rule 3.1” [36].

Treating the interpretation of s 3.1(1) as a matter of first instance [29], Hughes J appealed to s 12 of the Interpretation Act [37], and held, in effect, that “the fee” should be interpreted to mean the fee that was payable for the thing that the applicant was clearly trying to do [38]. Thus Rule 3.1 was applicable, and the Commissioner’s decision not to apply the grace period was unreasonable.

It seems to me that Hughes J’s decision was straightforwardly correct. I find it very difficult to think of an interpretation that would support CIPO’s position. Perhaps CIPO was of the view that there must be a clear attempt to pay the full fee, and the provision is only operable if the attempt is unsuccessful for technical reasons? I don’t see any reason why the requirement of a subjective intent to pay the full fee should be read into this Rule. In any event, I am just speculating as to what CIPO’s position might have been. I wonder if what happened is that someone at CIPO simply forgot about Rule 3.1(2), and so sent out a refusal instead of a request for payment, and this litigation was an attempt to justify that error post hoc?

Thanks to Alan Macek's IPPractice for making this decision available before it has been posted on the Federal Court website.