Dr Reddy's v Boehringer Ingelheim: empagliflozin SPC falls for lack of plausibility

Patents Court invalidates Jardiance compound patent and SPC but upholds crystalline form patent.
The Patents Court has held that Boehringer Ingelheim's basic patent for empagliflozin, the active ingredient in its type 2 diabetes drug Jardiance, was always invalid because the specification did not make plausible that the compound would inhibit SGLT. The supplementary protection certificate (SPC), due to run until November 2029, falls with it.
In Dr Reddy's Laboratories (UK) Ltd v Boehringer Ingelheim International GmbH [2026] EWHC 2944 (Pat), Meade LJ, sitting in the Patents Court, revoked EP 1 730 131 but upheld EP 1 888 552, which claims a crystalline form of empagliflozin, as valid while in force. Dr Reddy's had sought to clear the way for a generic launch and had been restrained by a preliminary injunction granted in November 2025.
No disclosure of testing
EP 131 contained a broad Markush formula and successively narrower preferred classes, culminating at [0054] in a list of 15 "particularly preferred" compounds, one of which was empagliflozin. It described SGLT assays and stated that compounds "may for example have" IC50 values below 1000, 200 or 50 nM, but reported no results.
Applying Warner-Lambert, as interpreted in Sandoz v BMS (Apixaban) and Generics v AstraZeneca (Dapagliflozin), Meade LJ rejected Boehringer's argument that the funnel of preferences disclosed that the 15 compounds were its best and had been successfully tested. That was "detective work", not disclosure. The preferences could equally rest on medicinal chemistry theory, ease of synthesis or patenting strategy, and synthesis of the compounds did not imply testing. Yields and mass spectra were given, but no IC50 data.
The "glucose argument", that structural similarity to phlorizin and T-1095 made activity plausible, also failed. Boehringer's own medicinal chemistry expert accepted that activity could not be predicted from general structure, and a pharmacophore argument had been abandoned during trial.
The judge rejected Dr Reddy's contention that in vitro activity was insufficient without in vivo data. Because SGLT inhibition was known to be specifically involved in diabetes, a plausible in vitro effect would have sufficed; requiring more would set the bar too high.
Contingent findings
Noting that the Supreme Court might yet revisit Warner-Lambert in light of G 2/21, Meade LJ made findings in case the law changes. He found positive reason to doubt empagliflozin's activity, given the unusual size, polarity and non-planar conformation of its tetrahydrofuranyloxy substituent. He also made deliberately narrow findings, on Boehringer's Civil Evidence Act notice, that testing yielding IC50s below 50 nM had been recorded before filing, while making no finding on the reliability of those tests.
Crystal form patent survives
EP 552's plausibility of SGLT inhibition stood or fell with EP 131. However, the PXRD and DSC data, read together, plausibly disclosed a substantially single crystal form with improved stability. Evidence from Dr Reddy's crystallography expert was directed to absolute certainty rather than plausibility and could not assist.
Applying Human Genome Sciences v Eli Lilly, the judge held that industrial application was satisfied even without plausible therapeutic activity. A stable crystal form of a compound put forward as a candidate SGLT inhibitor would at least facilitate testing it. That contribution was modest but real, unlike stabilising a substance for no purpose. Infringement of EP 552 was admitted.
The judge also declined Dr Reddy's post-draft invitation to rule on further EP 552 claims, describing it as not a proper use of the Re A and L procedure.
Mark Vanhegan KC, Adam Gamsa and Seaghan Davey (Mishcon de Reya) for Dr Reddy's; Thomas Mitcheson KC, Stuart Baran and Thomas Lunt (Kirkland & Ellis International) for Boehringer.
.png&w=3840&q=60)











