In brief
The UK Court of Appeal handed down its decision in Generics v AstraZeneca on 16 July 2025. It found that AstraZeneca's patent for the diabetes drug dapagliflozin (marketed under the brand name FORXIGA) was invalid because the patent did not make it plausible that the compound could be useful to treat diabetes.
Importantly, the Court of Appeal considered that:
- The attempt by the European Patent Office's Enlarged Board of Appeal (EBA) in its decision in G 2/21 to reconcile the "ab initio implausibility" versus "ab initio plausibility" tests, espoused in separate lines of authority, had failed.
- It did not make sense to apply a different standard depending on whether the validity challenge was based on a lack of inventive step or insufficiency.
- Permitting a patentee to rely on information that was not in the patent application, and did not form part of the common general knowledge at the priority date (and was therefore not available to the notional skilled person), in order to support validity was inconsistent with the first to file system.
- Neither the Technical Boards of Appeal of the EPO nor national courts in European Patent Convention (EPC) member states had achieved a settled approach following G 2/21.
- It remained bound to follow the approach of the Supreme Court in Warner-Lambert v Generics (pregabalin) and the Court of Appeal in Sandoz v BMS (apixaban) on the quality of information that a patent needs to contain in order to justify a claimed monopoly.
Background
The case related to AstraZeneca's patent EP1506211 entitled "C-aryl glucoside SLGT2 inhibitors and method" (the Patent). Claim 2 was for the compound dapagliflozin per se. Claim 15 was for the use of dapagliflozin in the manufacture of a medicament for treating or delaying the onset of diabetes.
The closest prior art (international patent application WO'128) disclosed and claimed large classes of compounds said to be SLGT2 inhibitors, useful for the treatment of diabetes (by inhibiting the reabsorption of glucose filtered by the kidneys), and contained 18 worked examples. Dapagliflozin was within the most preferred class, and one of the exemplified compounds had a very similar structure. WO'128 also described the same assay for checking the activity of the compounds as the assay in the Patent, but, like the Patent, disclosed no test results.
At first instance, Glenmark contended the Patent was invalid for a lack of inventive step and/or insufficiency because:
- it did not make it plausible that dapagliflozin is an SGLT2 inhibitor, a selective SGLT2 inhibitor, or useful for the treatment of diabetes; and
- it did not make a technical contribution over WO'128, but merely made an arbitrary selection from the class of compounds disclosed therein.
The trial judge held the Patent invalid on both grounds because it did not contain anything more than bare assertions that dapagliflozin was useful for the treatment of diabetes, and AstraZeneca had not advanced any case that it was any different to any of the other compounds of WO'128. We discussed the Patents Court decision, including its validity findings, in our earlier article.
Court of Appeal decision
AstraZeneca urged the court to adopt the approach of the EBA in G 2/21 (Sumitomo). In the EPO, a validity attack based on the paucity of information in a patent may be considered under inventive step or insufficiency. If the claim includes a utility limitation, then the attack must be considered a challenge to sufficiency. A second medical use claim (like claim 15 of the Patent), which implicitly requires that a compound has efficacy to treat a disease, is an example. On the other hand, if the claim does not include a utility limitation, then the attack must be considered as a challenge to inventive step. A claim to a compound per se (like claim 2 of the Patent) is an example.
The EBA noted that, to overcome an insufficiency challenge to a second medical use claim, information in a patent had to make it credible that the drug was useful for the claimed therapeutic application. If that threshold was met, then the patentee could rely on post published information (i.e. information that was not available to the notional skilled person at the priority date because it did not exist or was secret) to meet the challenge. However, the same threshold did not apply if the attack was based on inventive step. In such a case, the test was less onerous and was met if a patent "taught" that the subject matter of the claim had a relevant technical effect, and the notional skilled person would understand it to have that technical effect. However, the EBA did not further articulate what this less onerous test required.
The Court of Appeal concluded that, while the EBA in G 2/21 had clearly intended to get away from debates over plausibility in the context of inventive step, and had attempted to reconcile ab initio plausibility and ab initio implausibility, the distinction between them was inescapable. It pointed out that claim 2 of the Patent was significantly broader than claim 15, because it covered all uses of dapagliflozin, while claim 15 was limited to the use of dapagliflozin for the treatment of diabetes. It observed that, as a matter of basic principle, a broader claim to monopoly requires more justification than a narrower claim, so it did not make sense for the patentability of claim 2 to be subject to less stringent criteria than claim 15. Accordingly, the court was not persuaded that G 2/21 justified it departing from Sandoz v BMS.
The court also reasoned that patent laws worldwide operate on a first to file basis, which makes the priority date (the cut-off date for any common general knowledge and prior art information that can be relied on in an assessment of novelty, inventive step, or insufficiency) critical to questions of validity. The first to file system would be undermined if applicants could add "matter" to their application after the priority date or rely on information that was not present in the priority application nor part of the common general knowledge to support patentability.
AstraZeneca also contended that the repeated references in the Patent to dapagliflozin "as an SGLT2 inhibitor useful for the treatment of diabetes" should be understood as a verbal description of the results of performing an experiment using the assay, even though no numerical results were given. The court rejected this as a bootstraps argument, and that such statements were mere assertions unsupported by any experimental results.
Finally, while the Court of Appeal recognised that harmonisation among EPC member states was an appropriate goal, it was not yet possible to say that the Boards of Appeal in the EPO or national courts had adopted a settled view as to how the test in G 2/21 ought to be interpreted and applied. Accordingly, this goal did not justify the court departing from Sandoz v BMS.
Key takeaways
The Court of Appeal considered that the reasoning of the EBA of the EPO in G 2/21 did not justify any change to the approach UK courts must take when assessing whether the information in a patent made a technical effect plausible, whether the attack is framed as an inventive step or insufficiency challenge.
The court refused AstraZeneca's application for permission to appeal to the Supreme Court on the basis that the application had no prospect of success. Nevertheless, it extended temporary injunctions restraining Glenmark from launching its dapagliflozin products for 14 days (to 4pm on 30 July 2025), to give AstraZeneca the opportunity to make an urgent application to the Supreme Court for permission to appeal, and continuation of the injunctions "holding the ring".