FibroGen v. Akebia: Arnold LJ addresses insufficiency, equivalence, second medical use claims, and much more

Courts, lawyers and legal scholars in many places across the globe have adjusted to the new normal that COVID-19 has imposed. It's far from ideal but with a range of special measures, from extended deadlines [here] to virtual hearings [here], we make do as well as we can. At the end of the day, we are fortunate to be able to continue our work despite the unfolding of a global crisis of near-Lovecraftian proportions and that, for many of us, it's business as usual—even in unusual circumstances. 

And so, this week Arnold LJ handed down his judgment in FibroGen v. Akebia [here] that displays his characteristic sharpness of reasoning and, to this young Kat at least, astonishing productivity. At 137 pages, it tackles a range of thorny patent law issues. Not all of them can be discussed here in detail. This post thus provides an "index" of the most interesting parts of the decision.

Background to case and overview of most pertinent points

Fibrogen holds a number of patents on enzyme inhibitors for treating various types of anaemia and related conditions. Akebia and Otsuka sought to revoke these patents to clear the way for an intended launch. Consequently, Astellas – Fibrogen's exclusive licensee – brought a counterclaim for infringement of the patents. The six patents that were subject of the proceedings were grouped into two patent families, each deriving from a common international application: Families A and B.

The outcome was that, for five out of six patents, all claims in the proceedings were invalid for insufficiency, and in addition the Family B claims were invalid because they were obvious over the priority document for Family A. Two claims – one from each family – survived the invalidity challenges, but they were not infringed by Akebia and Otsuka. Had the claims been valid, Akebia and Otsuka would have infringed the claims in issue from the Family A patents (except for the valid claim), but not the claims in issue from the Family B patents.

Here is a list of the paragraphs tackling the most relevant issues:

• 36: Discussing the proper way to instruct expert witnesses to avoid hindsight.
  
• 199-201: Discussing when individual papers can be considered common general knowledge: yes if a paper "has been cited 610 times" by the priority date, no if it is "simply too old and obscure".
  
• 204 ff: Discussing incorporation by reference in an application of patent and non-patent literature, and finding in 221 that "the skilled team would not necessarily follow up any of the cited publications" in the absence of "direction as to which would be worth following up".
  
• 283 ff: Discussing how the skilled person would read a passage in the description stating that "an aryl radical may be substituted by 1 to 5 substituents" and, in particular, whether they would understand that some limitation to potential substitutes was intended.
  
• 303: Discussing the interpretation of the expression "structural mimetic of 2-oxoglutarate" and finding that the skilled person would be uncertain what it meant, because neither the specification nor the cited publications specified its meaning.
  
• 344: Discussing inventive step and finding that the claimed compound would not be found through routine research, even though it would have been obvious to search for it.
  
• 347 ff: Discussing three types of insufficiency – "classical", "Biogen" and "Uncertainty" – and summarizing the case law on the latter two categories.
  
• 376 ff: Discussing Biogen-insufficiency and finding that claims are insufficient where a patent implicitly promises that substantially all compounds which satisfy the structural definitions in the claims will have the claimed therapeutic efficacy, whereas this is only demonstrated for five compounds.
  
• 453: Discussing the interface between claim amendments and equivalence and finding, that where a claim is limited to a single compound to save its validity, it would be contradictory that its scope of protection reaches well beyond that compound by way of equivalence, in particular to compounds covered by the original claim.
  
• 456: Discussing the principle of "disclosed but not claimed is disclaimed".
  
• 590 ff: Discussing infringement of a second medical use claim in a case where the infringing product is not yet marketed and the Summary of Main Product Characteristics (SMPC) is not finalized yet.

The IPKat takes advantage of the lockdown to catch up on its reading backlog (Image by fellow Kay Riana Harvey)

Two comments on claim construction from a Dutch perspective

It is impossible to do justice to all of these issues in a single short blog post, so I chose two that, from a Dutch perspective, are of particular interest [Cue to readers who are looking for a way to rid themselves of some mental Coronaspeck: this is a great opportunity for comments or even guest contributions on any of the outstanding issues!].

Claim scope: validity versus infringement. Paragraph 418 raises the interesting question whether, if infringement by equivalents is found, the same scope of claim should be adopted when considering any issue of validity. At first glance, it seems that the answer should be yes, following the reasoning of the well-known Gillette defence [after a 1913 House of Lords decision, excerpt here], known in Germany as the Formstein defence [after a 1986 Federal Supreme Court decision, here]. These doctrines state that a patent cannot be infringed by the prior art or obvious variations thereof: hence, the infringer either works outside of the scope of the patent, or the patent is invalid. 

But validity of a patent must be established at the priority date, whereas infringement – including by way of equivalence – is established at the date of infringement. This was confirmed by the Dutch Supreme Court in a 2014 decision [here], in which it held that there are two relevant points in time for claim construction: the priority date for validity and the date of infringement for scope of protection. Thus, equivalents of the claimed invention that became obvious over the prior art only after the priority date might infringe a patent, without being prejudicial to its validity. That would suggest that, in fact, the scope of claim including equivalents is not always decisive for validity, but only where the relation of those equivalents to the prior art has not changed since the priority date.

Disclosed but not claimed is disclaimed? The principle that, where the specification discloses several ways in which a particular technical effect can be achieved only one of which is claimed, the other ways to achieve the technical effect cannot amount to infringement ("disclosed but not claimed is disclaimed") seems settled in German case-law and is followed by Arnold LJ in this decision. Indeed, describing an embodiment but failing to claim it will normally be a strong indication that the patentee chose not to seek protection for that embodiment.

But this rule should not be applied mechanically and exceptions can arise. In this respect, two of this Kat's former colleagues have advanced the following hypothetical [Den Hartog & Blomme, BIE 2016/15 – in Dutch]: suppose a claim protects a nucleotide sequence with 600 base pairs, and the specification states that deviations in some of the pairs have no influence on the effects of the claimed amino acid. In such a case, it seems unreasonable to deny the patentee protection against sequences which deviate minimally from the claimed sequence.

Denying protection against effective equivalents on the basis of these types of statements would have the undesirable result of incentivizing overly careful and sparse drafting of patent applications, potentially hampering their disclosure function. Although the question has not been put to the Dutch Supreme Court in literal terms, two recent decisions – Resolution/AstraZeneca (2018) and Bayer/Sandoz (2016) – suggest that under Dutch patent law, "disclosed but not claimed is disclaimed" is not an absolute maxim.