Patents Court upholds validity of second medical use patent for glaucoma treatment

Written on 11 May 2021

This decision shows the court assessing patentability in a real-world practical context, recognising the uncertainties of R&D that only later may appear to have been obvious

The recent decision in Alcon v Accord & Aspire deals with a number of important patent law issues including novelty of therapeutic uses and the "obvious to try" assessment.

The trial was ultimately a revocation-only action (infringement was admitted by the generic manufacturers). The patent and supplementary protection certificate (SPC) rights had expired but the trial was necessary because interim injunctions had been obtained and accompanying cross-undertakings in damages provided. A validity ruling on the expired patent was therefore necessary to determine whether the patentee, Alcon, needed to pay damages on its cross-undertakings.

The patent in suit claimed a 1993 priority date and related to the use of travoprost (an ester prodrug of fluprostenol), a prostaglandin F2α analogue, for the treatment of glaucoma. Validity was attacked primarily on the basis of: novelty over a prior art patent application, EP800, and obviousness over a prior art scientific paper, “Phenyl substituted prostaglandin analogs for glaucoma” by Stjernschantz and Resul and published in "Drugs of the Future" (1992), referred to as Stjernschantz.

A number of interesting points of patent law and procedure arose in the judgment. Meade J supported the view that plausibility should form part of the law of novelty under the enablement requirement (first expressed in Merck v Ono), and applied a very thorough and well-reasoned "obvious to try" assessment. The judgment also contains further guidance on how to define the fields of expertise of the skilled team (this will be looked at in a separate Insight), following the recent development in this area in Illumina v Latvia MGI.

Novelty: therapeutic effect must be plausibly enabled

The novelty citation, EP800, disclosed combinations of "F series" and "E series" prostaglandin analogues for the treatment of glaucoma. The only reference to travoprost in EP800 was in a particular example where it was disclosed in combination with an E-series prostaglandin. However, there was no disclosure of any efficacy data.

The main claim in issue had been amended during the course of opposition proceedings before the European Patent Office to expressly disclaim the combination of travoprost and the E-series prostaglandin disclosed by the particular example in EP800. As EP800 did not contain an individualised disclosure of travoprost other than in combination with the disclaimed E-series prostaglandin, Meade J held there was no prior art disclosure of a composition falling within the claim and the novelty attack failed.

If this attack had succeeded, Alcon sought to address it with a conditional amendment. The claim before the court used open-ended wording of "comprising" (it is conventional in patent claim drafting to use this wording to mean "including"). The conditional amendment sought to delete the disclaimer and change the use of "comprising" to "consists of" (that is, closed drafting, such that it extends to travoprost only and no other elements). This effectively limited the claim to the use of travoprost as a monotherapy, so the disclosure of the combination in EP800 also failed to anticipate the amended claim.

Meade J also agreed with Alcon that, because there was no efficacy data in the prior art, the claimed therapeutic effect had not been made plausible (applying Merck v Ono), so any disclosure would not have been enabling. An alternative way of considering this kind of issue – which arguably requires a different threshold of disclosure – is to say that there would be no "disclosure" of the mental element required by the claims (adopted in Hospira v Genentech) as the mere proposal in the prior art to use a compound for a particular therapy would not disclose that the treatment is indeed efficacious.

Obvious to try?

One test frequently used to assess obviousness in the context of therapeutic-use claims is to ask whether the claimed invention would have been obvious to try with a reasonable expectation of success (although this is just one factor to be balanced against other relevant considerations; see Actavis v ICOS). This decision provides a clear well-reasoned demonstration of the "obvious to try" analysis, which came out in favour of the patentee, and re-iterates how important it is for parties to direct their experts to address properly all factors relevant to the test.

The skilled team

It was common ground that the "skilled team" would include a medicinal chemist and a pharmacologist (and both sides called experts from both fields), but the parties differed significantly over the characteristics of skilled pharmacologist and hence the skilled person's identity. In defining the characteristics of the skilled pharmacologist, the judge applied the approach set out in Illumina v Latvia MGI (following Birss LJ's detailed review of the law in this area). In doing so, Meade J observed the need to ground the assessment in reality by considering what teams existed at the priority date, and strike a balance of fairness to the patentee and to the public in not defining the skilled team too narrowly or too broadly. Meade J found there was a real but small and only recently established field in the relevant area. As is often the case, the definition of the skilled team was important when it came to identifying the uncertainties that prevailed in the field (as it was so newly established, these were significant).

The prior art: Stjernschantz

The Stjernschantz paper disclosed a body of work investigating structure-activity relationships for various synthetic prostaglandin analogues (showing different activities and side effects in various animal models using different substituents). The claimed compound, travoprost, was not disclosed in the paper, but latanaprost, a structurally different prostaglandin F2α analogue, was put forward as a particularly promising compound with positive activity data and reduced side effects demonstrated in a number of in vivo animal models.

The paper disclosed that latanaprost was undergoing Phase III trials at the priority date after successfully completing Phase II trials (although this information was also held to be common general knowledge (CGK)). The paper put forward a hypothesis about why certain chemical substitutions may result in increased activity and reduced side effects.

The question was: whether it was obvious to use travoprost to treat glaucoma instead of any of the compounds described in the paper?

The defendants' attack

The defendants argued that the paper showed that latanoprost bound well to a particular prostaglandin receptor (the FP receptor) and that this binding was responsible for its biological activity (reduced intra-ocular pressure) and side effects were mediated by binding to different prostaglandin receptors. The skilled addressee would then consider further FP receptor agonists for the same purpose and, as travoprost was known to be a potent and selective FP receptor agonist, the defendants argued it would have been obvious to try travoprost for treating glaucoma.

There were a number of uncertainties in the CGK and disclosed in the prior art (many of which were a function of the field having been newly established), which included a lack of any "gold standard" animal model to test therapeutic activity and side effects, and a suggestion that an increase in activity was also positively correlated with certain side effects. This limited the confidence of the skilled team that progress could be made to improve or maintain activity but reduce side effects. The judge was careful to qualify the extent to which fluprostenol (the free acid form of travoprost) formed part of the CGK; its use in therapy was not known (it had veterinary uses unrelated to glaucoma, which did not form part of the CGK) and it was not mentioned in Stjernschantz. The judge also considered the importance of the overall teaching of the paper, which was largely focussed on investigating structure-activity relationships.

Obviousness assessment

With these points in mind, Meade J held that "the natural way for the skilled team to approach obvious developments from Stjernschantz would be to consider further prostaglandin analogues, altered in ways concretely reasoned out from the structure activity work described. This would be logical and routine and in keeping with the approach of the paper. It is suggested in the penultimate paragraph of the paper."

The judge went on to consider other obvious ways of following the teaching of the prior art (noting, of course, that more than one approach can be obvious); for example, whether it would have been obvious to use compounds with a significantly different structure to latanaprost but with similar activity (as the defendants contended). Meade J decided that the paper was very much focussed on structure-activity relationships, so its teaching pointed away from trying other structurally different compounds based on activity. It was not argued that using fluprostenol would be consistent with any of the structure-activity work reported in Stjernschantz. Overall, the prospects of using fluprostenol were considered uncertain with merely a hope (rather than a positive expectation) of success.

The defendants' expert evidence was not compelling and, in particular, their pharmacology expert failed to deal with why the skilled team would consider using fluprostenol in the first place, did not include a proper analysis of the prospects of success, and did not take account of the nature of the work done and suggestions made in the prior art.

OC comment

This decision emphasises the need for experts to address all factors relevant to the obviousness assessment in their evidence.

Of particular note is the care taken by Meade J to characterise accurately the practical content of the prior art paper and how the skilled team would realistically implement its teaching. It is welcome to see the court really emphasising the practical nature of the task of assessing obviousness.

Although the case law has often highlighted the dangers of avoiding hindsight reasoning, there have been occasions when undue weight has been placed on a particular sentence or passage of the prior art document, removing that content from its proper context.

It is refreshing to see the court properly engage with the teaching of a prior art citation and how this would have been understood and acted upon by the skilled team at the priority date. This demonstrates that the Patents Court continues to place patents in a real-world practical context and recognises the uncertainties involved in research that only later may appear to have been obvious.

If you have any questions about this article, please contact Tim Harris or Chris Pratt to discuss how we can assist you.