In July this year, Wragge & Co's Intellectual Property team reported on the German Appeal Court decision on Eli Lilly's patent for the anti-psychotic drug olanzapine. The German court granted an injunction despite the Federal Patent Court having revoked the patent in suit a year earlier for lack of novelty.

The validity of the EP(UK) patent on olanzapine has now also been considered by the UK High Court. On 13 October 2008 the UK High Court held the patent to be valid. This decision is interesting as it analyses the UK law on "selection patents" and continues an emerging UK trend of following European Patent Office jurisprudence.


Olanzapine is an anti-psychotic drug used in the treatment of schizophrenia and has fewer side effects than other treatments. It is the world's top-selling schizophrenia drug.

Dr Reddy's Laboratories (UK) Limited challenged the validity of the EP(UK) patent for olanzapine on the basis of lack of novelty and inventive step. There was also an attack on the sufficiency of the disclosure in the patent.

The key cited prior art is summarised as follows:

  1. '235 Provisional' patent – this claimed a wide class of compounds having useful central nervous system activity. This class was defined with reference to a general formula where certain groups could be substituted. The substitutions necessary to result in the olanzapine compound were among those specifically mentioned as possibilities for each of the variables in the preferred group. However, the general formula was wide enough to encompass some 1019 compounds;
  2. '235 Patent' – this document takes the above disclosure further, highlighting a particular preferred class. This class would again include olanzapine. This document also identified a compound, flumezapine, as particularly active;
  3. 'Chakrabarti 1980' – this paper disclosed a further preferred class of compounds which had been synthesised. "Compound 6" was ethyl olanzapine (olanzapine is a compound where an ethyl group is substituted for a methyl group). However, Compound 6 was not discussed in a list of highlighted compounds;
  4. 'Chakrabarti 1989'– this was a paper which followed up on the work conducted and reported on in Chakrabarti 1980. It focused on a particular candidate selected from the Chakrabarti 1980 paper, flumezapine. This paper states that compounds in the class of flumezapine (thiopenes) and flumezapine itself were of interest (as opposed to alternatives which included the class to which olanzapine belonged). The structures of flumezapine and olanzapine are identical save that flumezapine has a fluorine group in the first phenyl ring.

The main challenges on the validity of the patent were based on the above disclosures.


Floyd J rejected the attack on the novelty of the patent. In doing so, he considered in some detail the status of disclosures in "Markush"-type (chemical class) formula.

Floyd J considered the leading cases from the EPO Technical Board of Appeal. The EPO deals with this area of patent law on a regular basis and has developed a doctrine that a chemical class disclosure does not necessarily take away the novelty of an individual compound falling within the class. The EPO treats the individual compound as not having been disclosed if, in order to arrive at the specific combination of the substituted groups, a choice has to be made from two (or more) lists of some length. So, the prior disclosure of a class of compounds did not take away the novelty of a claim to a specific compound unless the specific compound is disclosed in an "individualised form". Floyd J said that a general formula covers a large number of compounds. The reader's attention focuses on the compounds actually described and the remainder of the class is no more than "a theoretical penumbra around those compounds". Such a prior document, in Floyd J's view, cannot be said to contain a clear description of, or clear instructions to do or make, something which would infringe the patent claim. A combination of substitutions is required to the prior disclosed general formula before the compound could be regarded as "unalterably established" and so the description is not clear. Therefore, a general formula to a compound, where multiple groups of that compound can be substituted from lists, will not normally take away the novelty of a subsequent claim to an individual compound.

Floyd J also considered the UK law on selection patents under the Patents Act 1949. In particular he considered the judgment of Maugham J in I.G. Farbenindustrie's Patent (1930) in which it was held that for a selection patent to be valid it must:

  1. be based on some substantial advantage to be secured by the use of the selected members; and
  2. the whole of the selected members must possess the advantage in question; and
  3. the selection must be in respect of a quality of a special character which can fairly be said to be peculiar to the selected group.

Floyd J then considered the House of Lords' decision in E.I. Du Pont de Nemours & Co's (Witsiepe's) Application [1982]. He felt that Du Pont was not saying that the only way in which a selection patent could be new when compared to the prior disclosed class is if the specification satisfies the criteria laid down in the Farbenindustrie's Patent case. Floyd J concluded that there is nothing in Du Pont which precludes him from following the EPO approach on selection patents, which is what he proceeded to do.

Applying the EPO jurisprudence, Floyd J found that the 235 Provisional disclosure provided no individualised disclosure of olanzapine: one would need to make numerous choices about the various substitutions to the general formula. Chakrabarti 1980 was also held not to be novelty destroying as one would need to pick Compound 6 and then modify it in a particular way, i.e. change the ethyl group to a methyl group without modifying the rest of the structure. For novelty, the question to ask is whether olanzapine is disclosed in the document; not whether one could obtain olanzapine from the document.

In terms of obviousness, Floyd J held that the patent was inventive over the prior art. The Chakrabarti disclosures were the starting points for assessing obviousness:

  1. Chakrabarti 1980 provided no individualised teaching of the structural substituents of olanzapine – i.e., there is no disclosure of the methyl-olanzapine. The evidence indicated that the skilled person would be likely to focus on the five compounds highlighted in the paper (which did not include Compound 6) rather than embark on a more systematic review by synthesising further compounds. This is in fact what Eli Lilly did in relation to flumezapine (which was Compound 9 in the Chakrabarti 1980 disclosure). Further, there was no suggestion in Chakrabarti that the compounds disclosed therein were a solution to the problem (i.e. finding a new anti-psychotic for treating schizophrenia which had fewer side effects than the then-existing product). Finally, there was no teaching in Chakrabarti 1980 that changing the ethyl group on Compound 6 to a methyl group would make any difference. In fact, there was teaching in this disclosure to the effect that there would be no purpose in making this alteration. Floyd J felt that the skilled person would not use Compound 6 as the starting point for exploring structural variations. The patent was not obvious in light of Chakrabarti 1980; and
  2. Chakrabarti 1989's focus was on flumezapine and there was no basis provided in this disclosure for abandoning the halogen substituent in the phenyl ring (which would be required to obtain olanzapine).

The patent therefore withstood the attacks on novelty and inventive step. The insufficiency attack also disappeared given Floyd J's decision on the above.


This case is interesting on a number of levels. Firstly, Floyd J looks to the EPO jurisprudence on the issue of chemical classes and selection patents and follows the doctrines developed there rather than old law in the UK. There is a growing trend in the UK for the courts to follow the EPO in relation to pharmaceutical and biotechnology patents. This is generally regarded as a welcome development.

It is also reassuring to see the UK court coming to the same conclusion, for pretty much the same reasons, as the German court (albeit the infringement Appeal Court of Dusseldorf). The Dusseldorf court's decision in May this year caused uproar so it will be interesting to see what the outcome of the appeal in the German validity case will be – but that decision is still three to five years away.

Finally, it is worth noting that the commercial success argument (as a secondary indicia of non-obviousness) was poorly received. The evidence run on this issue did not assist Floyd J in deciding the issue of obviousness and he stated that arguments on commercial success in cases like this are "more or less doomed to failure".