"This revocation action was brought by Lundbeck against Norpharma and its assignee Infosint (the proprietor of the patent, who had not registered the transfer with the UK IPO, leaving Norpharma as the registered proprietor). The patent related to a method to make 5-carboxyphthalide (5-cbx), the first intermediate in the manufacture of Lundbeck's blockbuster anti-depressant drugs citalopram and Escitalopram, involving a reaction of two common reagents in a solvent. This was claimed in claim 1, the distinguishing features of the claimed process over the prior art being one or more of (i) the temperature and (ii) the pressure at which the reaction was conducted, and (iii) the concentration of the solvent. The patent also contained a "reach-through" claim (claim 22), which claimed a process for the synthesis of citalopram in which the process of claim 1 was contained.
Lundbeck said that claim 1 lacked novelty and inventive step over the prior art; that claim 22 was obvious and insufficient. Infosint counterclaimed for infringement of the patent by Lundbeck's use of the processes in the UK, and by importation into the UK of direct products of claim 22 (i.e. citalopram) made outside the UK. Infosint's infringement case was based on over 40,000 pages of batch records produced by Lundbeck and its third party manufacturers. In addition to its non-infringement arguments, Lundbeck pleaded a prior user defence under s.64 (it had been using the process on a commercial scale for some 15 years prior to the publication of the patent application) and also argued s.68 (as a limitation to any damages and/or costs).
As for novelty, Floyd J was required to consider the law of overlapping ranges in relation to a novelty-only prior art citation under s.2(3). The only integer in issue was the concentration of the solvent in the prior art document. Floyd J found that absence of a concentration figure in the prior art could not amount to a disclosure of a range of concentrations from 1% to 99% . Accordingly, it was not necessary for him to decide whether to follow the EPO Technical Board of Appeal in Unilever T-666/89. [Readers who don't like the technical details should skip the coloured text]
Claim 1 was found to be obvious over the Forney prior art citations . Claim 22 was found to be obvious over the '513 patent and Forney. The approach taken in reaching the later conclusion is worth noting. Lundbeck said the skilled person would start with the '513 patent (a published method to make citalopram from 5-cbx) but would need a method to make 5-cbx; and his search for such a method would uncover Forney's work (with the obvious consequence that he would use the claimed process to make 5-cbx). The Judge noted early on in his Judgment that "it is impermissible to make so-called mosaics of individual citations (unless it would be obvious to do so)" . However, on the evidence, the Judge found that the skilled person seeking to work '513 would consult "the Beilstein handbook of Organic Chemistry … the standard reference for the preparation of organic compounds"  and would "obviously and inevitably find and read Forney" . Lundbeck also made an "aggregation" argument , which the Judge did not feel it necessary to deal with .
Claim 22 was not found to be insufficient, the Judge finding that the technical contribution of claim 22 was making citalopram via 5-cbx made by the process of claim 1 and accordingly the claimed monopoly did not exceed the contribution to the art . The Judge also held that despite the patent not disclosing all methods of making citalopram from 5-cbx, one could reasonably expect the invention to work with other methods and thus the claim was sufficiently enabled across its entire breadth .
Lundbeck had made 5-cbx before the priority date in both the UK (since 1995) and Denmark (since 1986) and Infosint accepted that Lundbeck had a section s64 prior user defence in relation to the operation of those processes. The judge was therefore not required to make detailed conclusions on the law on the prior user defence. However, the Judge did say that the prior user right in relation to importation of citalopram should only extend to citalopram made by the process of claim 22 and not citalopram however made - one relying on section s.64 may only do what is "substantially the same as what was done before" . A substantial change to the process may still result in a process within the scope of the claim - accordingly the s.64 defence did not extend to two pot processes (a variation devised by Lundbeck after publication of the patent to avoid infringement).
Although the patent was found to be invalid, Floyd J still dealt with s.68. Infosint had registered the assignment of the patent at the EPO - but this had been after its date of grant. Thereafter, Infosint's UK agents, although instructed to do so, had omitted to register Infosint as proprietor on the UK register. Floyd J found that, once a patent was granted, there was no deeming provision for the purposes of the Patents Act which applied to assignments registered at the EPO, and thus the assignment had not been registered for the purposes of the Act .
"The public should not be expected to check the EPO register to see whether by any chance an assignment of the patent was registered …".
Infosint argued that the "applicants" had done all that was practicable, and it was their agents who had erred. The judge did not agree, and accordingly Lundbeck could have relied on s.68 to limit damages (pre-29th April 2006) and costs (post-29th April 2006) had the patent been valid ".Thanks, Simon!