Glaxo has lost its appeal against Beach J’s ruling that Apotex and Generic Partners did not infringe its sustained release paracetamol patent. In a battle between what the skilled addressee would understand and the plain, literal wording of the claim, the lawyers won out.

Interpreting the claim

Claim 1 of the patent is for a pharmaceutical composition with a bilayer tablet having both an immediate release phase of paracetamol and a sustained release phase wherein:

said composition has an in vitro paracetamol dissolution profile (as determined by the USP type III apparatus, reciprocating basket, with 250ml of 0.1M HCl at 37C set at a cycle speed of 15 strokes/min) with the following constraints:

(a) 30 – 48% released after 15 minutes,

(b) 56 to 75% after 60 minutes and

(c) more than 85% after 180 minutes.

The problem was in that reference to a “reciprocating basket”, as a USP type III apparatus had a reciprocating cylinder, not a basket.

At the relevant time, there were four types of apparatus for testing dissolution – a USP type I apparatus, a USP type II apparatus, a USP type III apparatus and a USP type IV apparatus. The four types of machines provided different hydrodynamics which, in turn, meant the user would obtain different dissolution results. If one used a different apparatus, therefore, one might get a different dissolution profile.[1]

As already noted, the USP type III apparatus did not have a reciprocating basket. “Pharmaceutical scientists” generally referred to the USP type 1 apparatus as a “basket”, the type II apparatus as a “paddle”, the type III apparatus as a “reciprocating cylinder” and the type IV apparatus as a “flow through cell”.

This reflected the different construction and working of the machines. To test dissolution using a type I apparatus, the composition was placed in a cylindrical mesh basket which was then placed inside a cylindrical vessel containg the dissolution medium. The basket then rotated around inside the cylindrical vessel.

In a type III apparatus, the composition is placed inside a glass reciprocating cylinder which is then moved up and down in a glass vessel containing the dissolution medium.

Apparently, it would be possible to modify a type III apparatus to use a “basket”, but there was no evidence anyone had ever done so.

Against this background, the evidence established that the skilled addressee would recognise that the reference to “basket” in the claim was a mistake.

The question is then how is the claim to be interpreted. GSK argued that the skilled addressee would simply read the reference to a “reciprocating basket” as a reference to the reciprocating cylinder that type III apparatus used. The Full Court considered that this went beyond a purposive construction and would involve impermissibly rewriting the claim.

After referring to Catnic[2] and Kirin-Amgen, the Full Court said at [109]:

It is important to note that Lord Hoffman [in Kirin-Amgen at [34]] was referring here to the meaning conveyed to the skilled addressee by the language used and was not directing himself to a situation in which the skilled addressee deduced that the language of the claim, although conveying to him or her a particular meaning, could never have been intended to mean what it conveyed.

Here, the skilled addressee knew what the term “basket” meant, but “deduced” it couldn’t be intended to mean that.

The Full Court then emphasised a point made in many cases: a patent is a document prepared by the patentee “in words of the patentee’s own choosing” and “the words will usually have been chosen upon skilled advice.”

First, if a mistake had been made, there was a procedure for amendment – which GSK had not invoked in this case. The availability of that procedure had the further significance that [s 115][s115] provided alleged infringers with some protection from damages before the amendments were allowed.

Secondly, in describing the device used for dissolution testing, GSK (at least on its case) was not trying to identify something new. So, the difficulties of describing something which had not existed before would not arise.[3]

Thirdly, the words had to have been put there to mean something. At [138]:

GSK’s infringement case can only succeed if the words “reciprocating basket” are either interpreted to mean “reciprocating cylinder” or simply ignored. Either approach involves an impermissible re-writing of the relevant claims.

And, really, that is the Full Court’s point. Athough one is supposed to interpet the claim through the eyes of the skilled addressee, reading claim 1 the way the skilled addressee apparently would have done did not make proper allowance for the function of the claims. At [139]:

This is a case in which the skilled addressee’s understanding of the claims, as found by the primary judge, does not make proper allowance for the function of the claims in defining the invention. Ultimately, it is for the court to decide the meaning of the claims. This is a case in which we think the language of the claim must be understood to mean what it actually says.

So, although the patent specification is addressed to, and to be understood through the eyes of the person skilled in the art, the Full Court qualified that approach in light of other policies gleaned fron the Act.

Fair basis

Apotex and Generic Partners also lost their appeals against the trial judge’s ruling that GSK’s patent was fairly based and there had been no failure to disclose the best method.

On fair basis, the claims were consistent with the consistory clauses, but Apotex argued the body of the specification showed that the invention was narrower than the broad consistory clauses. This appears to have been an attempt to read the claims down to two specific formulations discussed in the specification, Formulations C and D.

A key point was whether the trial judge had impermissibly taken into account information in an FDA report to ascertain if the claims travelled beyond the disclosure in the specification. Apotex argued this was excluded by the High Court’s decision in the first Lockwood decision, where it had said at [48]:

If all that is essential in assessing a fair basing objection is recourse to the contents of the specification, there is no call, for example, for an examination (except on construction questions) of common general knowledge (which is essential when considering an objection based on want of an inventive step), or of prior art (which is essential when considering novelty (s 7(1))) …

The Full Court, however, rejected this attack; concluding that the information in the FDA Report (which was common general knowledge) informed how the skilled addressee would understand the claims. At [166], the Full Court said:

What is critical to the pharmacokinetic behaviour of the many formulations within the claims is the dissolution profile (or release rate) of the formulation. The primary judge accepted that the FDA Report recognised that a variation of ±10% percentage points in the release rate was acceptable to the FDA even where no IVIVC had been established. This provides evidentiary support for the finding that the skilled addressee would know that various formulations within the claims apart from Formulations C and D were likely to have similar pharmacokinetic properties. This also provides a complete answer to Apotex’s argument that the skilled addressee (equipped with the common general knowledge) would approach the Patent with an understanding that there would be no reason to think that other formulations within the claims would have a similar pharmacokinetic profile to Formulations C and D in the absence of any established IVIVC.

The Full Court also rejected Apotex’ argument that a claim could not be fairly based unless the specification explained why the claims worked. Making it clear that they were dealing only with the position before the Raising the Bar Act reforms, the Full Court said at [170]:

Of course, it is important to note that s 40(2)(a) requires that the complete specification “describe the invention fully”. A complete specification may still “describe the invention fully” without explaining why the invention works. After all, the inventor, who presumably believes that the invention described works, may not understand why it works. But this does not prevent him or her from obtaining patent protection for the invention.

Best method

For best method, the argument built on the Servier ruling to argue there had been a failute to disclose the best method because the specification did not disclose the particular grade or viscosity of the high viscosity HPMC or the granulation end points used to make Formulations C and D.

At [192], the Full Court accepted that there could be a failure of best method if information was withheld even though it could be ascertained by routine experiment. The Full Court rejected the best method attack, however, finding that the informatin omitted was inessential manufacturing and production information. According to the Full Court at [201]:

It does not follow merely because the patent applicant uses a particular manufacturing process or a particular excipient in formulating its commercial embodiment that it will form part of the best method. The patent applicant may have adopted a particular process, or used a particular excipient, for reasons that are associated with its own particular circumstances rather than because it believes that they reflect the best method. The best method known to the patent applicant may be one that allows for the optimisation of a formulation by the skilled addressee rather than one that adheres to one specific formulation that the patent applicant seeks to commercialise.

Those preparing specifications might want, first, to note the reservation that the Full Court was not dealing with the “new” post-Raising the Bar regime. Secondly, [192] appears to carry with it the warning that, if one leaves something out, one does so at one’s own peril.

GlaxoSmithKline Consumer Healthcare Investments (Ireland) (No 2) Limited v
Generic Partners Pty Limited
[2018] FCAFC 71 (Middleton, Nicholas and Burley JJ)


  1. At [50], the Full Court explained, “There is no established correlation between dissolution results measured using the different USP apparatus. Sometimes small differences are measured, other times the measured differences are large. Differences in measured dissolution profiles as between each of USP type I, II and III apparatus may not be predictable.”  ?
  2. Catnic Components Limited & Anor v Hill & Smith Limited [1982] RPC 183 (HL).  ?
  3. Referring to another aspect of the construction problem adverted to by Lord Hoffmann at [34] in Kirin-Amgen. Leading of course to the irony that GSK did in fact describe something new!  ?