(1) Teva UK Ltd (2) Teva Pharmaceutical Industries Ltd v Leo Pharma A/s

An invention is not obvious merely because it would be worthwhile to include a possible solution in a research programme. An obvious to try case requires a fair expectation of success, assessed in light of the real knowledge, practices and prejudices of the skilled person.

Where prior art identifies only one of many possible non-aqueous solvents, and there is no reasonably optimistic expectation that it will solve a formulation problem, finding that it does so may constitute an inventive advance. Long-felt need, prior failure to find the solution, and the significance of the resulting product are relevant circumstances in the obviousness assessment.

An amendment adds matter only if it teaches the skilled person something about the invention which could not be learned from the application as filed.

LEO appealed against Birss J’s decision in the Patents Court, [2014] EWHC 3096 (Pat), revoking two patents for an ointment combining calcipotriol and betamethasone with the solvent Arlamol E.

The judge had held the patents obvious over Turi, a 1978 United States patent which disclosed Arlamol E in a corticosteroid formulation. He did not determine Teva’s alternative challenges of added matter and insufficiency.

The combination solved a practical incompatibility problem: the two active ingredients could not be used together in aqueous products because each required stability at a different pH. The central issue was whether, at the priority date, the skilled formulator would have had a fair expectation that using Arlamol E in a non-aqueous combined ointment would succeed.

1. Appeal allowed unanimously. Sir Robin Jacob, with whom Lady Justice King and Lord Justice Kitchin agreed, held that the judge had made errors of principle in the obviousness analysis. The Court therefore reconsidered obviousness independently and reversed the revocation order.

2. The real skilled formulator would have understood the need for a non-aqueous system, but that was necessary rather than sufficient. Water contamination and apparent-pH effects meant that a formulator could not predict whether any particular non-aqueous solvent would stabilise both active ingredients. Testing candidate solvents would be a research project, not an exercise with a fair expectation that Arlamol E would work.

3. The judge had wrongly treated the notional skilled person as less conservative than real formulators. The skilled person is attributed the real practices and prejudices of the relevant field. Turi disclosed Arlamol E only for a mono-active corticosteroid product. It did not identify it as an outstanding candidate for the dual-active formulation or give a reasonably optimistic expectation of success. Mere possible inclusion in a research programme was insufficient for obviousness.

4. The judge also failed to weigh material secondary circumstances. The need for a combined product had existed since at least the mid-1990s, yet no one had produced a stable combined ointment before LEO. The eventual product was a significant advance because it improved patient compliance. Those matters supported invention.

5. The added-matter challenge under Article 123(2) of the EPC failed. The amended combination was disclosed within the application: the two actives were identified as appropriate, a non-aqueous solvent was preferred, Arlamol E appeared first in a specified list, and an example used that combination. Omitting an optional anti-oxidant added no new information. The insufficiency challenge was also rejected because the patents enabled the skilled person to make the claimed ointment.

• Court of Appeal (Civil Division): Allowed LEO’s appeal and reversed the finding that the patents were obvious: [2015] EWCA Civ 779.

• High Court, Chancery Division, Patents Court: Birss J revoked the patents for obviousness over Turi. The alternative added-matter and insufficiency grounds were not determined: [2014] EWHC 3096 (Pat).