Promise me no Promises? The five-year anniversary of the Supreme Court's AstraZeneca decision

27 minute read
13 June 2022

From 2005 to 2017, numerous Canadian patents were invalidated on the basis of an approach to patent utility termed the "Promise Doctrine." Requiring any statement of advantage in a patent specification to have been demonstrated or soundly predicted by the patent's filing date lest the entire patent be invalidated, the Promise Doctrine created uncertainty for patent holders and placed Canada's approach to utility out of line with international norms. However, in the June 2017 decision of AstraZeneca v Apotex[1] the Supreme Court of Canada settled the Canadian patent litigation landscape by unanimously holding that the Promise Doctrine had no statutory foundation and was "not good law". We look back at AstraZeneca and review its impact in the five years since its release.

What was the Promise Doctrine?

Section 2 of Canada's Patent Act[2] requires that an invention be "useful." No level of utility is prescribed in the Act, and traditionally utility was seen as a low bar – so long as an invention was not devoid of utility, the statutory requirement was found to be satisfied.[3] Given this standard, it was rare for a patent to be invalidated on the basis of lack of utility.

Two primary events changed the dynamic in the early 2000s. First, the Supreme Court decision of Apotex v Wellcome[4] in 2002 imposed a requirement that utility be demonstrated or soundly predicted as of the patent's Canadian filing date. Then, in 2005, drawing on a statement made in a previous Supreme Court decision which itself was drawing on what became repealed UK law,[5] Canadian courts began to approach the utility question by asking whether any "promise of the patent" was demonstrated or soundly predicted by the filing date.[6]

This "uniquely Canadian" Promise Doctrine created a boon to those seeking to invalidate patents.[7]

The Promise Doctrine was particularly impactful as was often applied to invalidate an entire patent on the basis of an unsubstantiated "promise,"[8] even if the statement did not relate to the subject-matter of each of the patent's claims. The Promise Doctrine also had the result of invalidating patents even to useful inventions, as having proven utility in one respect would be of no moment to a patentee if a different "promise" was not substantiated.

Both of these prospects were realized in the lower courts' decisions in AstraZeneca, where all claims of the patent at issue were initially invalidated despite the claimed subject matter being useful as a proton pump inhibitor. Fatal to the entire patent was the fact that a second disclosed promise of "an improved therapeutic profile such as a lower degree of interindividual variation" – which was not part of the language of any of the asserted claims – had not been established by the filing date.[9]

The AstraZeneca v Apotex decision

The Promise Doctrine met its demise in AstraZeneca. The Supreme Court held that the Promise Doctrine was "incongruent with both the words and the scheme of the Patent Act," "punitive," "antagonistic to the bargain on which patent law is based" and that it "undermines a key part of the scheme of the [Act]."[10] The Promise Doctrine's primary faults were said to be: (1) that it conflated the utility requirement with a disclosure requirement by determining the utility standard by reference to promises expressed in the patent; and (2) that, where a patent contained multiple promises of utility, each needed to be fulfilled.[11]

Not only did the Promise Doctrine serve to invalidate otherwise valid patents, but it discouraged fulsome disclosure from inventors and thus was inimical to the goals of the patent system as a whole.[12]

After disposing of the Promise Doctrine, the Supreme Court set out the correct approach to utility and applied this framework to the facts before it.

In clarifying the utility standard needed to meet the Patent Act's requirement, the Supreme Court held that a demonstration or sound prediction of a scintilla of utility related to the nature of the subject-matter of the claims of the patent is sufficient.[13] Applying this, that the subject matter claimed in AstraZeneca's patent[14] was soundly predicted to be useful as a proton pump inhibitor meant that the statutory criteria was met. That the disclosure's "improved therapeutic profile" statement of advantage was not established by the filing date did not impact the utility analysis.[15]

While the ratio of AstraZeneca is definitive, the case raised several questions going forward. What would the scintilla standard look like in practice? How "related" does the claimed subject matter's utility need to be to be applicable? And what would be the impact of obiter statements made by the Supreme Court regarding the interplay of "overpromising" with insufficiency, overly broad claims, and s. 53 of the Act?[16] Five-years out, the landscape has largely settled, though some outstanding issues remain.

The post-AstraZeneca world

AstraZeneca has unsurprisingly joined the ranks of the most frequently cited patent law cases in the five-years since its release. About 54 reported decisions of the Federal and Provincial Superior courts have cited AstraZeneca as of the date of writing, with 31 of these being final or appeal decisions.

The initial decisions applying AstraZeneca were called upon to interpret the Supreme Court's two-step framework. One of the first reported cases to consider AstraZeneca was Bristol-Myers Squibb v Apotex (BMS Dasatinib),[17] an appeal wherein an allegation of inutility based on the Promise Doctrine had succeeded at first instance in a decision pre-dating AstraZeneca.

In BMS Dasatinib, the patent claimed the chemical compound dasatinib, as well as the compound for the use in the treatment of cancer. The Federal Court, in a decision pre-dating AstraZeneca, found that there was an overarching promise in the patent that dasatinib would be useful in treating a range of ailments and also in inhibiting enzymes from two different families of protein tyrosine kinases, or PTKs. The Court held that BMS had not met its burden to show that the overarching promised utility was either demonstrated or soundly predicted by the filing date, despite the work done by the inventors which showed that dasatinib inhibits PTKs.[18]

The decision was appealed, and AstraZeneca was released. In considering the bounds of AstraZeneca, Apotex argued that demonstration of the ability of dasatinib to inhibit certain enzymes in a test tube could not satisfy the scintilla of utility requirement. The Federal Court of Appeal disagreed, holding that:

Establishing that a compound has the ability to inhibit a biological target implicated in disease is doubtlessly a useful discovery. Here, it was known as of the relevant date that enhanced activity of PTK was involved in many diseases, as stated in the specification and confirmed in the evidence of several of the experts. Thus, discovery of a substance that acted to inhibit certain PTKs represented an important advance and certainly meets the minimal utility requirements that are now applicable following the decision of the Supreme Court in [AstraZeneca].[19]

Another early decision applying AstraZeneca was Pfizer v Apotex (Pfizer ODV).[20] The claims at issue covered a polymorphic form, "Form 1," of ODV succinate. The Federal Court rejected allegations of inutility, applying AstraZeneca in agreeing with Pfizer that Form I ODV succinate was useful as a "stable, solid state form of ODV succinate" and that this use was "directly related to the subject matter."[21] The Court rejected Apotex's argument that stability could not qualify as a relevant utility under AstraZeneca as, per Apotex's allegation, it is a physical property of the drug and not what the drug does as a practical matter (i.e. treat disease).[22] It was held that the practical usefulness of the drug as a stable solid form alone is sufficient utility under AstraZeneca.[23]

The application of AstraZeneca has remained generally consistent with that seen in BMS Dasatinib and Pfizer ODV. Overall, utility has been decided by the Court with reference to AstraZeneca at least 17 times since July 2017.[24] In only three of these decisions was a lack of utility found. Two of these dealt with utility after already finding the same claims to be invalid on another ground: in Swist v Meg Energy the claims found to lack utility were already held anticipated[25], while in Aux Sable v JL Energy the claims deemed to lack utility were previously held to be overbroad.[26] In only one decision, Safe Gaming v Atlantic Lottery, was utility the sole basis for any claim of a challenged patent to be declared invalid.[27]

This can be contrasted with the 2005-2017 era of the Promise Doctrine, where approximately 30 patents were found to lack utility.[28] Many of these decisions cited inutility as the sole basis to invalidate the entire patent.

The relative lack of success of inutility arguments post-AstraZeneca has not however dissuaded litigants from continuing to raise inutility. For instance, 12 of the post-AstraZeneca decisions in which utility has been considered have been released in 2020-2022, i.e. after parties have presumably had sufficient time to consider whether to pursue utility allegations in view of AstraZeneca and early post-AstraZeneca case law such as BMS Dasatinib and Pfizer ODV.

Attempts to resurrect the Promise Doctrine

In addition to ascertaining the bounds of AstraZeneca, Courts have also been faced with various attempts to resurrect the Promise Doctrine or some variation thereof. In the early cases which considered AstraZeneca, parties frequently scrambled to re-cast existing Promise Doctrine arguments as properly fitting within the AstraZeneca framework. The majority of these efforts were identified and rejected.

In Apotex v Shire, the Court characterized proposed pleading amendments as "a refusal to come to terms with and embrace the essence of the Supreme Court's teachings, and a fairly desperate attempt to shoehorn Apotex's promise allegations into each and every ground of invalidity known to law. The resulting pleading remains haunted by the ghost of the now defunct promise doctrine."[29] Remaining utility allegations were then dropped shortly before trial.

In Pfizer ODV, discussed above, in addition to its attempt to fit its Promise Doctrine arguments into the AstraZeneca utility framework, Apotex argued in post-hearing submissions that alleged "overpromises" made in the patent at issue violated the sufficiency of disclosure provision of the Act, subsection 27(3), and rendered the patent invalid. Brown J. rejected the argument, noting that the Supreme Court would have stated its intention had it desired for the Promise Doctrine analysis to be transferred from utility to insufficiency. The Court further held that such a result would have restored the underlying policy problems caused by the Promise Doctrine, and that AstraZeneca did not expand on the sufficiency of disclosure approach that was recently addressed in Teva v Pfizer.[30] Indeed, the Supreme Court has made clear on multiple occasions that utility and sufficiency are separate and distinct.[31]

A similar attempt to expand the Promise Doctrine into overbreadth and insufficiency was rejected in Hospira v Kennedy Trust, with the Federal Court stating that "it would be inconsistent to discard [the Promise] doctrine only to have it resurface under another principle without clear language to do so."[32] Likewise, in Apotex v Abbott Laboratories the Ontario Superior Court held that a party cannot simply raise promise-based pleadings under a different ground of invalidity, and that the Court should "whack the zombies dead once and for all."[33]

The rise of overbreadth?

Despite utility seemingly settling back into its traditional, less prominent role in patent validity, parties challenging patents continued to raise inutility alongside overbreadth. Some success was seen in 2019-2020.

In Les Laboratories Servier v Apotex,[34] the Federal Court accepted an overbreadth argument which was based on an alleged lack of utility across the claim scope. In that case, the invention disclosed was the arginine salt of perindopril and the evidence showed that studies had been conducted pre-patent filing using L-arginine, which demonstrated enhanced stability. The disclosure specified that L-arginine was preferred, but the claims did not include any limitation.[35] The Court held the claims, which were not limited to L-arginine,[36]overbroad as the inventors had not made or soundly predicted the usefulness of every arginine salt combination encompassed within their scope.[37]

In another overbreadth analysis, the Federal Court in Aux Sable Liquid Products v JL Energy Transportation drew elements from the specification and imported them as claim requirements that needed to be demonstrated or soundly predicted by the filing date – a hallmark of the Promise Doctrine.[38] The Court ultimately found that claims 9 and 10 were broader than the invention.[39]

This trend of invalidation via overbreadth via inutility arguments appears to have been short-lived. More recent decisions have rejected allegations that claims which do not establish a result or advantage across their scope are overbroad as being a proxy for the Promise Doctrine. One example of this was Eli Lilly v Apotex, where an allegation asking the Court to parse the disclosure, find a promise of a result, import the promise as a required claim element, and invalidate the claim for failing to meet the promise was rejected as being "very akin to the promise doctrine."[40] Very recently, the Court in Janssen v Sandoz rejected an overbreadth argument that was "essentially a restatement" of the utility allegation that had previously been addressed and rejected upon an application of AstraZeneca.[41]

It may be inevitable that overbreadth will take on increased prominence with the Federal Court of Appeal in Seedlings recently holding that overbreadth can be a distinct ground of invalidity (though overlap with other grounds of invalidity was recognized).[42] However, caution must be exercised, particularly when overbreadth is asserted as a Promise Doctrine proxy in attempts to invalidate claims that cover new, inventive, and useful subject-matter, but in respect of which some result or advantage is not proven across the scope of the claim.


AstraZeneca has firmly taken root as setting a clear and low utility threshold. This is consistent with the purpose of the utility requirement, which is to prevent the patenting of fanciful, speculative, or inoperable inventions. While parties continue to allege inutility and attempt to import the Promise Doctrine into other heads of invalidity, Courts have generally rebuffed these efforts, with obviousness and anticipation returning to being the primary gate keepers to patent invalidity.

[1] AstraZeneca v Apotex, 2017 SCC 36 [AstraZeneca]. Discussed here.

[2] Patent Act, RSC 1985, c P-4 [Act].

[3] e.g. Monsanto v Commissioner of Patents, [1979] 2 SCR 1108 at 1122 (If the inventors have claimed more than what they have invented and included substances which are devoid of utility, their claims will be open to attack. But in order to succeed, such attack will have to be supported by evidence of lack of utility.)

[4] Apotex v Wellcome, 2002 SCC 77.

[5] Consolboard v MacMillan Bloedel, [1981] 1 SCR 504 at 525 [Consolboard], citing Halsbury's Laws of England, (3rd ed.), vol 29, at p 59. The "False Promise" provision in the United Kingdom's Patent Act was removed in the 1977 revision.

[6] See e.g. Aventis Pharma v Apotex, 2005 FC 1283 and Pfizer v Apotex, 2005 FC 1205.

[7] AstraZeneca at para 32.

[8] Typically a statement of result or advantage.

[9] AstraZeneca v Apotex, 2014 FC 638 at paras 80-81, 214-215, 218, aff'd 2015 FCA 158, rev'd in AstraZeneca.

[10] AstraZeneca at paras 36, 51.

[11] AstraZeneca at para 37.

[12] AstraZeneca at para 51.

[13] AstraZeneca at paras 52-57.

[14] Said to be "optically pure salts of the enantiomers of omeprazole". See AstraZeneca at para 61.

[15] AstraZeneca at paras 59-63.

[16] AstraZeneca at para 51.

[17] Bristol-Myers Squibb v Apotex, 2017 FCA 190 [BMS Dasatinib].

[18] Bristol-Myers Squibb v Apotex, 2017 FC 296.

[19] BMS Dasatinib at para 40.

[20] Pfizer v Apotex, 2017 FC 774 [Pfizer ODV].

[21] Pfizer ODV at para 340.

[22] Pfizer ODV at para 341.

[23] Pfizer ODV at para 341.

[24] As per internal student research. The authors are indebted to Eslam Mehina, Hannah Im, Ayushi Thakur, and Justin Kim for their assistance in gathering much of the data cited in this article.

[25] Swist v MEG Energy, 2021 FC 10.

[26] Aux Sable Liquid Products v JL Energy Transportation, 2019 FC 581 [Aux Sable].

[27] Safe Gaming System v Atlantic Lottery Corporation, 2018 FC 542.

[28] Either by way of action or under the previous scheme of the Patented Medicines (Notice of Compliance) Regulations in which allegations of patent invalidity were decided by way of application.

[29] Apotex v Shire, 2017 FC 831 at para 8.

[30] Pfizer ODV at paras 356, 360, 363, 364, citing Teva v Pfizer, 2012 SCC 60 [Viagra].

[31] Consolboard, Viagra, and AstraZeneca.

[32] Hospira v The Kennedy Trust, 2018 FC 259 at para 258.

[33] Apotex v Abbott Laboratories, 2018 ONSC 5199 at para 28.

[34] Les Laboratories Servier v Apotex, 2019 FC 616 [Laboratories Servier].

[35] Laboratories Servier at paras 174-179.

[36] Laboratories Servier at para 179.

[37] Laboratories Servier at paras 219, 240-241.

[38] Aux Sable at paras 56-69.

[39] Aux Sable at paras 74.

[40] Eli Lilly v Apotex, 2020 FC 814.

[41] Janssen v Sandoz, 2022 FC 715 at para 250.

[42] Seedlings v Pfizer, 2021 FCA 154.

NOT LEGAL ADVICE. Information made available on this website in any form is for information purposes only. It is not, and should not be taken as, legal advice. You should not rely on, or take or fail to take any action based upon this information. Never disregard professional legal advice or delay in seeking legal advice because of something you have read on this website. Gowling WLG professionals will be pleased to discuss resolutions to specific legal concerns you may have.

Related   Life Sciences