Interpreting patent claims: Actavis v Eli Lilley

It's now 13 years since Lord Hoffmann told us definitively how to interpret patents claims, in his opinion in Kirin Amgen v Hoechst Marion Roussel [2005] 1 All ER 667. The courts had spent a lot of time pondering how to do what should have been a pretty simple task: judges have always been in the business of finding the meaning of legal documents, and patent claims should never have been much different. Or should they? For many years the courts have eschewed (nice word) a literal interpretation and chosen instead to look at the claims as practical documents written for practical readers.
I have never bought into that view (but who cares what I think?). Patent claims are usually written by highly skilled and trained patent attorneys for other patent attorneys, members of the patent bar and ultimately specialist judges to read. Why should it be necessary to adopt a different rule about interpretation from that which governs the interpretation of, say, conveyances?
I suppose that, centuries ago, when patents for inventions first came before the courts, the judges were able to say "we know how to interpret documents like this: we look at the wording." Only later did the fiction arise that the claims were anything other than highly specialised documents written by and for experts. The idea that patent claims are addressed to some hypothetical individual skilled in the art strikes me as nonsense: what ordinary person, however skilled they might be in the art, ever reads patent claims?
The courts pursued this strange idea through a series of cases, tying themselves in more and more knots. Catnic, for one, should probably have been a professional negligence case - which is not to say that the patent agent who write the claims was necessarily negligent, but it would have been helpful if the courts had considered the point. Then in the Improver case Hoffmann J formulated a series of questions for the judge to ask, which turned into the Protocol Questions when the Protocol on Interpretation of Article 69 (of the EPC), originally dating from 1973, became recognised as the governing document, and in due course Lord Hoffmann put his own Improver Questions out of their misery.
In Activis UK Ltd v Eli Lilley & Co [2017] UKSC 48 (12 July 2017) the claim in issue was a Swiss-style claim for pemetrexed disodium, with vitamin B12, used in the manufacture of a medicament for treating cancer. Actavis sought a declaration that it would not infringe by marketing drugs containing other pemetrexed compounds (for example, with potassium instead of sodium). Arnold J at first instance gave the declaration, and the Court of Appeal upheld him, both following (as they pretty well had to) Kirin Amgen. The Supreme Court, however, allowed Actavis's appeal.
Lord Neuberger (who, remember, is a Oxford chemistry graduate) held that when a court has to consider a possible infringement by a variant, there are two questions that it must ask:

• First, does the variant infringe the claim at issue as a matter of interpretation? If it does, that's all the court has to do.
• If not, ask whether the variant infringes because it varies only in a way, or in ways, which are not material. Put another way, is the variant equivalent to the claimed invention?

Lord Neuberger said that the Improver or Protocol Questions remain useful in finding an answer to the second question, but needed to be changed slightly. They now require the judge to consider

1. whether the variant has a material effect on the way the invention works;
2. whether the notional addressee, on learning what the variant does, would consider it obvious that it would achieve substantially the same result in substantially the same way as the invention; and
3. whether the notional addressee would have uderstood from the language of the claim that the patentee intended strict compliance with the primary meaning to be an essential requirement of the invention.

The courts had previously set their faces against introducing a doctrine of equivalents into UK patent law, but the amendment to the Protocol which introduced Art.2 made this untenable. Potassium was, in effect, equivalent to sodium for the purposes of the patent and Actavis would infringe if they marketed the products about which they asked in the first place.
The further the courts move away from a literal interpretation of patent claims, the less certainty there is for those, like Actavis, who want to operate in the field but also don't want to infringe. One argument - perhaps the argument - against a literal interpretation is that it prevents a patent claim being read in the light of technological developments. But is it right that a patentee who has chosen to specify a sodium compound rather than using broader language (if, of course, they could get away with it) should receive such an extensive, and flexible, monopoly? Perhaps the correct approach would be to regard the discovery that a potassium or some other compound would do the job as leading to an invention in its own right.

Rounding conventions: what is the value of 1?

Smith & Nephew Plc v Convatec Technologies Inc [2015] EWCA Civ 607 (24 June 2015) is a patent case which turns on how to read a numerical range. The patent related to wound dressings in which silver - a known antimicrobial agent - was applied to the fibres. To prevent discolouration, the patent specified the use of a binding agent, at a concentration between 1 per cent and 25 per cent.

The trial judge, Birss J, interpreted this (using the conventions that I thought I had remembered from school maths lessons) to mean between 0.95 per cent and 25.5 per cent. But the Court of Appeal knew better (or at least they picked up on the argument).

Of course, like all interpretations of patent claims, the right approach is that set out in the Protocol on Interpretation of Article 69 - which requires us to consider what the person skilled in the art would think the claim was intended to mean.

It came down to whether the right approach was a "whole number" approach or a significant figure approach. The whole number approach would say that 1 per cent means all figures greater than or equal to 0.5 per cent and less than 1.5 per cent - or to put it another way, all values that round to 1 per cent when expressed in a whole number. The significant figure approach was explained by Kitchin LJ, who started by noting that the rules formed part of the common general knowledge: 

i) non-zero digits are always significant;
ii) zeros between non-zero digits are always significant;
iii) leading zeros are never significant; if a decimal point appears in a number then trailing zeros are significant (before or after the decimal point);
iv) in the absence of a decimal point, trailing zeros are not generally significant unless stated otherwise either expressly or with a bar over the zero.

Applying these rules, 1 per cent indicates a range from 0.95 per cent to 1.5 per cent. It is asymmetric because the values that round to 1 must comprise two significant figures: if there's only one significant figure, you can't round it up or down. This is how Birss J got to the conclusion that the defendants' 0.77 per cent concentration did not infringe. He reckoned that the skilled person would use the significant figures approach. The Court of Appeal preferred to think that the skilled person would assume that the values were expressed to the nearest whole number, so there was an infringement.

Why, oh why, didn't the person who drafted the claims (and the description, of course, as it has to support the claims) either write "up to 25 per cent" or "1.0 to 25 per cent" (especially when the teaching of the patent was that as little as 0.01 per cent would do the trick)? I'm sure they had their reasons, just as the patent agent who wrote the Catnic patents had his reasons for writing "vertical" (or was it "vertically"?). However, in neither case are the reasons known, while in both of them the effect was significant, indeed determinative, in the patentee's favour in both cases, though only just in the earlier one. Cases like these cause me to reflect on the received wisdom about patent claims, that they are technical documents written for technical people and not suited to the sort of legal scrutiny that a conveyance should get (I forget the precise quote). But in fact they are complicated legal documents (often incredibly so) written by highly-trained professional patent agents (soi disant attorneys, indeed) and to my mind that means that meticulous verbal analysis (there, I remembered the words!) is entirely appropriate. And the Protocol is wrong!



'via Blog this'