The recent Court of Appeal decision in Generics (UK) Limited & Ors v AstraZeneca AB confirmed that the approach to plausibility is unchanged in the UK following G2/21: the ab initio plausibility test set out by the Supreme Court in Warner Lambert remains good law. Speculative inventions cannot be made good after filing – there must be sufficient support for the claimed invention in the application. Likewise for selection inventions – individual compounds picked out from a known Markush formula will be arbitrary and lack any inventive step if there is no support for that selection in the specification – theoretical or experimental.
The Court upheld Dr Tappin KC at first instance and dismissed all of AstraZeneca’s appeal grounds. The decision is final, with the UK Supreme Court refusing permission to appeal.
Background
The patent claimed dapagliflozin, a sodium-dependent glucose co-transporter protein (SGLT2) inhibitor, used to treat diabetes. The prior art (WO 128) disclosed a class of compounds said to be SGLT2 inhibitors, by way of a number of Markush formula. Dapagliflozin fell within this disclosure. The patent in issue did not contain any experimental results using dapagliflozin or theoretical passages directed to dapagliflozin and its use as an SGLT2 inhibitor.
The claimants, including Generics (UK) Limited, Teva, and Glenmark, challenged the patent on two grounds: (i) lack of inventive step, owing to a lack of technical contribution, and (ii) insufficiency. Both of these arguments stemmed from a common starting point - that the patent failed to make plausible that dapagliflozin was an effective SGLT2 inhibitor or useful for treating diabetes.
The High Court invalidated the patent on both grounds, finding that:
- the patent failed to make it plausible that dapagliflozin was an effective SGLT2 inhibitor or useful for treating diabetes; and
- the patent was an arbitrary selection from a prior broad disclosure (WO 128).
AstraZeneca appealed.
Key Issues and Findings
1. Plausibility
Under UK law, the test for plausibility is derived from the Supreme Court case of Warner-Lambert v Generics (UK) 2018 UKSC 56. This case established that plausibility is not an independent legal requirement but rather a threshold test for assessing both sufficiency and inventive step. Specifically, when a technical effect is claimed in a patent, there must be enough disclosure to make it positively plausible to achieve the stated technical effect. Merely asserting the existence of a technical effect is insufficient to meet this standard.
At appeal, AstraZeneca submitted that this "ab initio plausibility" standard is overly stringent. They urged the Court to consider the case in light of the European Patent Office (EPO) decision in G2/21, which they argued adopts a more lenient approach. Under G2/21, a technical effect claimed for inventive step need only be “encompassed by the technical teaching” and “embodied by the same originally disclosed invention.” In doing so, AstraZeneca contended that the Court should depart from its earlier decision in the case of Sandoz v BMS 2023 EWCA Civ 472, which applied and affirmed the Warner-Lambert standard.
AstraZeneca also submitted that the standard to assess plausibility must be different when applied to the inventive step and not sufficiency, as claim 2 of the patent was directed to the compound dapagliflozin per se. Therefore, AstraZeneca argued that the lower threshold of ab initio implausibility as set out in G2/21 should at least be applied to this claim.
The Court rejected AstraZeneca’s argument for a lower ab initio implausibility standard, which would only require that the claimed effect is not implausible, for the following reasons:
- Distinction Between Plausibility Standards: G2/21's attempt to unify the ab initio plausibility and implausibility standards has not eliminated their distinction. The Court stressed that reliance on hindsight to prove a technical effect absent from the patent is impermissible, and G2/21 did not warrant departing from the decision in Sandoz v BMS.
- Lack of Settled Case Law: The Court noted the absence of a consistent approach to interpreting and applying G2/21 across the EPO Boards of Appeal and European national courts. Although there is an “emerging consensus” against speculative patents, there is still no clear agreement on what defines a speculative patent, and the case law remains “some way from unanimous”. This creates a potential for divergence across Europe.
Application to AstraZeneca’s Patent
- The patent claimed that dapagliflozin was an SGLT2 inhibitor useful for treating diabetes but provided no experimental data or theoretical reasoning to support this assertion.
- References in the patent to an assay were characterised as bare methodological descriptions, not a “verbal statement of positive results.”
- Cross-referenced documents (WO 128 and four Tanabe papers) were considered, as the skilled team would review them. These documents generated doubt about dapagliflozin’s activity, particularly due to structural differences between dapagliflozin and other known SGLT2 inhibitors.
Structural Similarities
AstraZeneca argued that structural similarities between dapagliflozin and other known SGLT2 inhibitors (e.g., phlorizin) were sufficient to predict that dapagliflozin would have similar effects, thereby providing evidence that the claimed effect was plausible.
The Court rejected this argument, noting that:
- Small structural differences can significantly impact biological activity.
- The patent provided no reasoning or data to support the assertion that dapagliflozin’s structure would result in SGLT2 inhibition.
- AstraZeneca’s patent would fail to meet even the ab initio implausibility standard. The absence of key structural features in dapagliflozin, would have given the skilled team legitimate reason to doubt its efficacy as an SGLT2 inhibitor, rendering the patent implausible under any standard.
2. Arbitrary Selection
In response to the Claimants' argument that dapagliflozin was an arbitrary selection from the broad class of compounds disclosed in prior art (WO 128) and therefore not inventive, AstraZeneca contended that the selection was not arbitrary. They argued that the patent made it plausible that dapagliflozin had the claimed technical effect (SGLT2 inhibition for treating diabetes), which was not made plausible by WO 128. As such, AstraZeneca maintained that this selection constituted a technical contribution, setting it apart from the prior art.
The Court applied the arbitrary selection doctrine, as established in Dr Reddy’s v Eli Lilly, and held that the patent was invalid for a lack of technical contribution, for the following reasons:
- WO 128 already disclosed a broad genus of compounds, including dapagliflozin, as potential SGLT2 inhibitors.
- The patent did not show any special advantage of dapagliflozin over other compounds in WO 128, such as Example 12.
- Simply narrowing the claim to a single compound and re-stating the same property (SGLT2 inhibition) does not constitute a technical contribution.
3. Interim Injunction and Appeal
An interim injunction was in place by the time of the Court’s decision, restraining generic launch. AstraZeneca sought to maintain the injunction pending its permission to appeal application. The Court refused, concluding that AstraZeneca’s appeal had no real prospect of success. However, the Court did exercise its discretion to grant a short 14-day extension of the injunction to preserve the status quo and allow AstraZeneca to make an urgent application to the Supreme Court. Permission to appeal was refused by the Supreme Court just 2 weeks later, on 31 July 2025.
Key Takeaways
1. Plausibility as a Safeguard Against Speculative Patents
The judgment reinforces the principle that patents must disclose enough to make their claimed technical effects plausible at the time of filing. This ensures that the patent system rewards genuine technical contributions rather than speculative claims.
2. Arbitrary Selection
A selection from prior art is inventive only if the selected compound demonstrates a specific, plausible technical benefit over its peers. Merely isolating and re-asserting the same property is insufficient. Importantly, the patent specification must disclose enough to make that technical benefit plausible at the priority date — it cannot be established solely through later evidence.
As an aside, at [145], Arnold LJ postulated that the patentee’s decision to omit experimental data from the patent may have reflected a deliberate filing strategy adopted at that time; noting that BMS (the original applicant for the patent in issue) had been found to have withheld data in respect of apixaban when filing its apixaban patent, in the same time period. Choosing to keep such data secret is a dangerous filing strategy, as the Court has confirmed that the patentee cannot rely on such data if the plausibility threshold his not met: patents must stand on their own disclosure at the time of filing.
3. Evidence of Plausibility
The Court clarified that skilled readers may consider cross-referenced documents when assessing plausibility. However, these documents may increase doubt if they suggest structural or functional issues with the claimed invention.
The Court also reiterated that post-published evidence cannot be used to establish plausibility. The patent must stand on its own disclosure at the time of filing.
Implications for Patent Drafting and Litigation
- For Patent Applicants: Include sufficient data or reasoning in the patent application to support the claimed technical effect. Avoid relying on post-filing evidence or speculative assertions.
- For Litigants: The case provides a clear framework for challenging patents on plausibility and arbitrary selection grounds. It highlights the importance of scrutinising the patent’s disclosure and critically reviewing how well the patent achieves its aims.
Conclusion
The Court of Appeal’s decision cements two key principles:
The ab initio plausibility test remains the standard for sufficiency and inventive step in the UK, with G 2/21 not altering this.
A selection from prior art is inventive only if it demonstrates a genuine technical advance.
By upholding these standards, the Court ensures that the patent system remains a tool for promoting genuine innovation rather than speculative monopolies, thereby reinforcing the fundamental patent bargain: a fair exchange of exclusive rights for meaningful public disclosure. Practitioners should heed the judgment’s lessons on robust disclosure and the risks of arbitrary selection, and the procedural challenges surrounding interim injunctions.
