Novartis AG v Generics (UK) Ltd (t/a Mylan)

Obviousness is a single, fact-sensitive statutory question. Matters such as whether a route was obvious to try, the expectation of success, the available avenues of research and the effort required are relevant considerations, rather than independent legal hurdles.

An invention may be obvious although the outcome of a routine experiment is unpredictable. The experiment must be prompted by rational technical considerations, rather than mere curiosity or a hope that something useful might emerge. Where prior art identifies a promising racemic pharmaceutical compound, selecting it for development and routinely resolving its enantiomers may therefore lack an inventive step despite uncertainty about the precise benefits.

Novartis owned a supplementary protection certificate and its basic patent protecting rivastigmine, the negative enantiomer of a known racemic compound used to treat Alzheimer’s disease. Mylan sought revocation so that it could market a generic product.

Floyd J, in [2011] EWHC 2403 (Pat), held claims 1–9 invalid for lack of inventive step and revoked the patent and certificate. The prior Weinstock publications identified the racemate RA 7 as a promising treatment but did not propose resolving it into its enantiomers.

The appeal concerned whether the judge had erred in finding it obvious to select RA 7 for development, resolve it, and formulate the negative enantiomer as a pharmaceutical for treating Alzheimer’s disease.

1. Appeal dismissed. Kitchin LJ held that the judge had directed himself correctly and was entitled to find the claimed invention obvious. Lewison LJ and Munby LJ agreed.

2. Obviousness remained a single statutory question requiring evaluation of all relevant circumstances. Whether a route was obvious to try, the expectation of success, the number of possible avenues, the necessary effort and the duration and complexity of experimentation were interacting considerations. They were not successive legal hurdles. The principles in Conor v Angiotech, [2008] UKHL 49, and MedImmune v Novartis, [2012] EWCA Civ 1234, were applied.

3. The law did not require it to be manifest or more or less self-evident that an experiment would work before an invention could be found obvious. That formulation could be appropriate on particular facts, as in Saint Gobain v Fusion-Provida, but it was not a universal test. An entirely routine experiment could support obviousness even where its outcome was unpredictable.

4. The Weinstock publications presented RA 7 as one of the most promising compounds and provided a plausible basis for its proposed therapeutic use. The judge had considered the remaining stages of drug research and was entitled to find it obvious to investigate RA 7 through those stages and take it into development. Continuing to synthesise other compounds and developing disclosed promising compounds were not mutually exclusive courses.

5. Resolution was driven by rational technical considerations, not idle curiosity. Enantiomers commonly differed in activity, metabolism, transport and toxicity. The skilled team would have recognised possible differences in metabolism, penetration of the blood-brain barrier and stereo-specific side effects. Resolution was routine and chemically straightforward. Unpredictability in chiral chemistry did not preclude obviousness.

6. Once RA 7 had been resolved, formulating rivastigmine for the treatment already taught by the prior art was obvious. The judge’s findings were supported by the evidence, disclosed no error of principle and afforded no proper basis for appellate interference.

1. Court of Appeal (Civil Division): By [2012] EWCA Civ 1623, unanimously dismissed Novartis’s appeal and upheld the finding that claims 1–9 lacked an inventive step.
2. High Court, Patents Court: Floyd J, in [2011] EWHC 2403 (Pat), held claims 1–9 of the basic patent invalid for lack of inventive step. His consequential order revoked the patent and supplementary protection certificate.