FibroGen Inc. v Akebia Therapeutics Inc. and another company; Astellas Pharma Inc. v Akebia Therapeutics Inc. and other companies [2021] EWCA Civ 1279

What are the practical implications of this case?

If Arnold J’s approach to insufficiency for want of plausibility had stood, it would have meant that pharmaceutical claims would have to enable substantially all of the compounds satisfying the structural definition as to the therapeutic efficacy. Equally, insufficiency for undue burden meant that the skilled team must be able to identify substantially all compounds covered by the claim without undue burden.

What the Birss LJ judgement does, is propose tests by which an invention, described using general terms, and which patent discloses a principle of general application, can reasonably be expected to work for anything falling within the scope of the claims. For plausibility (which the judge considers is best described as ‘reasonable prediction’) there is a three step test (although there is potentially a further, separate functional limitation in the first step namely, treatment of eg a disease). First, identify what falls within the scope of the claimed class; second, ask what is it intended to do; and third, whether it is possible to make a reasonable prediction the invention will work with substantially everything falling within the claimed scope. For undue burden, claims with a functional feature, or a mix of those and a structural feature, must make possible, without undue burden, the identification of compounds which satisfy the relevant test (eg treatment of a disease). It is not necessary to identify all or substantially all compounds that satisfy that test.


What was the background?

FibroGen owned two families of patents (A and B) relating to the treatment of anaemia associated with the kidney: Chronic Kidney Disease (‘CKD’ – family A) and Anaemia of Chronic Disease (‘ACD’ – family B). The A patents predated the B patents and constituted full prior art.

In family A the claims in issue comprised a heterocyclic carboxamide compound (selected from a named group), that inhibited hypoxia inducible factor (HIF) prolyl hydroxylase (PH) enzyme activity, in a medicament for increasing endogenous erythropoietin, in the prevention etc of anaemia associated with CKD. The family B patents had claims for the same compounds but were for treatment of ACD.

Akebia had a rival HIF-PH inhibitor (vadadustat), which was in clinical trials. They brought proceedings for revocation of the A and B patent families, met with a quia timet infringement action relating to the trials.

At trial, Arnold J concluded that the A patents were not obvious (over prior art called Epstein) but were invalid for insufficiency for lack of plausibility and undue burden. Akebia’s vadadustat infringed on the basis of normal construction. The B patents were obvious in light of the A patent’s original application and failed for insufficiency for the same reasons. Even if valid, there was no infringement (as the clinical trials were aimed at CKD).

FibroGen appealed the insufficiency findings for both families, and Akebia for the infringement finding.


What did the court decide?

On the two insufficiency bases, Birss LJ considered Arnold J’s approaches were wrong, and overturned them. In order to do so, Birss LJ considered all of the law on insufficiency, both UK cases (including the Supreme Court decisions in Regeneron v Kymab and Warner Lambert v Generics), the EPO decisions, and (importantly) the decision of the Bundesgerichtshof (‘BGH’) in Dipeptidyl-Peptidase-Inhibitoren.

In agreeing with the BGH’s reasoning that a claim to ‘everything that works’ without saying how to achieve that is unacceptable, Birss LJ also agreed with the BGH that a claim to the use of any compound with a functional feature (eg act as an inhibitor) for a therapeutic purpose (lower blood glucose), does not claim everything that works. Such language could also cover compounds not yet found, so for Arnold J to consider the skilled person must identify substantially all compounds covered by the claim meant he was wrong.

Birss LJ also clarified, based on EPO cases, that the second paragraph of 6.6.9 of the 9th Edition of EPO case law text book should be given a narrow reading. The term ‘without undue burden’ in that context refers to the ability of the skilled person to identify compounds within a structural class as having a functional feature ie  with a functional feature, without undue burden, it must be possible to find compounds satisfying the test (eg acting as an inhibitor).

In deciding that the law of undue burden does not require the skilled person to identify substantially all compounds covered by the claims, this begged the question of how many compounds does the patentee need to disclose? The answer is some, but the key is that the skilled person, without undue burden, must be able to identify more beyond those stated in the patent. Moreover, the skilled person has to be able to work anywhere within the claim to apply the tests without undue burden, and identify if it is a claimed compound.

Birss LJ considered the A (and B) patents were not insufficient, and so were valid and infringed. He upheld Arnold J’s conclusion that the B patents were obvious and so invalid.


This article was first published by Lexis®PSL on 27 August 2021