Ranbaxy UK Ltd & Anor v Warner-Lambert Company

This is the judgment in two actions brought by Ranbaxy (UK) Limited (‘Ranbaxy’) and an action by Arrow Generics Limited (‘Arrow’) respectively against Warner-Lambert Company, (‘Warner-Lambert’). Warner-Lambert is the owner of European Patent (UK) 0247633 (’633) and of European Patent (UK) 0409281 (’281), which are concerned with atorvastatin, a cholesterol synthesis inhibitor of great commercial importance. The calcium salt is sold under the name Lipitor.

In action 2167 Ranbaxy seek a declaration of non-infringement in respect of ’633. Both Ranbaxy (in action 2059) and Arrow (in action 3986) seek to revoke ’281, which claims atorvastatin calcium. The issues in respect of ’281 are common to the latter actions, and the same evidence was relied on both by Ranbaxy and by Arrow.

So far as ’633 is concerned, the issue is one of construction and easily stated. Does the claim cover the single enantiomer which might be illustrated by formula (I) of the claim, or does it only cover racemic atorvastatin, that is the mixture of two enantiomers which Ranbaxy says is denoted by formula (I) on a proper interpretation of the formula in context. This issue is raised by Ranbaxy only. After the trial but before this judgment was complete, I heard and rejected an application by Arrow to be joined as a party to the action in respect of ’633.

So far as ’281, is concerned, the issues are (a) obviousness over the application for the ’633 patent (the ’633 Application) and (b) anticipation by Warner-Lambert’s application under the PCT number WO 89/07598 (the ’598 application).

I am only concerned with claim 1 of ’633 and with claims 1 and 2 of ’281. If claim 1 of ’633 is not infringed I understand that none of the others will be. Independent validity is only alleged in respect of claims 1 and 2 of ’281.

At the outset of the trial, I ruled that Ranbaxy could not rely upon certain documents, largely representations made to certain patent offices, in which persons acting for Warner-Lambert had made representations about the disclosure of ’633 much the same as that for which Ranbaxy now contend. I excluded the documents because I considered that they were not admissible on the question of construction.

This is an application for a declaration of non-infringement, after an acknowledgment was refused by Warner-Lambert. The letter seeking the acknowledgment described the acts that Ranbaxy intended to do as follows:

Claim 1 of the patent is as follows:

The case is concerned with stereochemistry. The specialist terminology is explained by the experts, and I will only repeat it so far as necessary to make what follows comprehensible, and I state the problem here without explaining the terminology. I have underlined the parts of the claim that lead to the allegation of infringement. The reason there is said to be no infringement is that although the two structural formulae (I) and (X) appear to show the R-(R*,R*) compound, it is common in organic chemistry to use the structural formula of a single enantiomer to denote the racemate. How such a structural formula is being used must be determined by reference to context. This much appears to be common ground. Given that the racemate is to be considered as a distinct compound from either of the two enantiomers to be found in it, Ranbaxy says that the specification makes it clear that the compound with which it is concerned is the racemate, and that in its context formula I is being used to refer exclusively to the racemate. It follows that the claim is limited to the racemate and that accordingly the enantiomer of interest does not fall within it. That is a statement of the argument, but before I discuss it I must set out the applicable legal principles, identify the skilled person to whom this document is addressed and set out the relevant part of the common general knowledge.

Kirin Amgen v Hoechst Marion Roussel Limited [2004] UKHL 46, [2005] RPC 9 (page 169) paragraphs [32] to [35] summarises the modern approach. Extracting short passages may tend to distort the overall meaning of this passage, but from it, and from the approval that it gives to the judgment of Jacob LJ in Rockwater Ltd v Technip France SA [2004] EWCA Civ 321, [2004] RPC 6 at [41] (and by inference also to the slightly different statement in Mayne Pharma v Pharmacia [2005] EWCA Civ 137 at [5]), I conclude that it is now clear that in deciding what the person skilled in the art would have understood the patentee to be using the language of the claim to mean, the court must approach the problem from the standpoint that the language chosen will be usually of critical importance. An over-meticulous analysis is one that is too willing to draw from a detailed analysis of the grammar, the punctuation and the particular words and phrases used inferences as to meaning that the words might support but which the skilled person would not draw, and it is the antithesis of giving to the words chosen in their context the meaning that the skilled person would give them. Carefulness is not to be equated with over-meticulousness.

This claim is couched in highly technical language. It uses the device of the chemical structural formula to convey its meaning. The structural formula does two things. It shows, in a highly schematic way, how the chemical bonds are located between different atoms of the molecule (or larger components, such as the benzene ring, which may be depicted in many ways. For the benzene ring, this patent uses Ph or.)

The problem in the present case is to interpret the formulae (I) and (X) in context. Quite apart from the myriad of different substitutions that are permitted by reference to the components X and R_1-4, the question is what possible 3-dimensional arrangements of the molecule are covered by the claim, that is, its stereochemical interpretation.

The title of the patent, together with the passage at page 1 lines 11-32 and the claims makes it clear that the patent is intended for those who will synthesise an active ingredient and formulate it for use in therapy as a hypolipidaemic or hypocholesterolaemic agent. The patent is therefore directed towards medicinal chemists with skills in organic synthesis. The evidence was that medicinal chemists often start their lives as organic chemists, and add the biological skills and knowledge to their degree-level skills in organic chemistry. Scrutiny of the document reveals that it is really only concerned with synthesis, and with describing the new class of compounds (page 4 line 3). Activity is described briefly (page 8 lines 20-34) and formulation is described in the most general and conventional terms (page 8 line 35 – page 9 line 14), as is dosage (page 9 lines 20-26). This is thus directed at the people employed by pharmaceutical companies to synthesise new active ingredients.

The principal expert called by Ranbaxy on this part of the case, Professor Clive, was an out-and-out synthetic organic chemist of wide experience. He was very careful in his answers. Dr Newton, who was called by Warner-Lambert, was a medicinal chemist of considerable eminence. There was an attempt, more by way of submission than on the basis of any evidence, to suggest that the perspective of a medicinal chemist and of a synthetic organic chemist on the disclosure of the ’633 patent would be different. This suggestion was made because Prof Clive’s approach, which was a very precise and painstaking approach to stereochemical descriptions, was unhelpful to the patentees and understandably they sought to diminish its significance by suggesting that it was in some way uncharacteristic of a medicinal chemist working in the pharmaceutical industry. I confess that to the extent that there was a difference between Prof Clive and Dr Newton, I considered it to be more a matter of personality than of substance. I shall deal with their respective positions as I deal with the various disputed points.

In any case, Dr Newton’s evidence was that a medicinal chemist is a synthetic organic chemist who has got additional training and, in particular, additional training in the biological aspects of pharmaceutical action. This case is about notation and context, and so both Dr Newton and Prof Clive were in a position to give relevant evidence on the points that arose.

’633 claims a priority date of 30 May 1986, to which Ranbaxy accepts it is entitled. This is the date at which the common general knowledge is to be assessed.

The tetrahedral arrangement of the covalent chemical bonds that a carbon atom may make with four other atoms means that if there are four different units at the ends of the four different bonds those units may be arranged in two different ways. When one says two different ways, it means that no amount of rotating will permit one arrangement to be superimposed on the other. They will be mirror images of each other, and are called enantiomers. This is shown in this diagram:

This is an attempt to show a tetrahedral structure. Conventionally the bonds depicted by single lines are in the plane of the paper: the bonds denoted by a solid wedge (a heavy line may also be used) are out of the paper and the bonds denoted by a dotted wedge (or dotted line) are into the paper. The picture shows two imaginary molecules with an asymmetric carbon atom C. A, B, X and Y denote different atoms or groups positioned at the apexes of the tetrahedron. Because the molecules are handed, they are called chiral, and the asymmetric carbon atom(s) they contain are called chiral centres. A molecule is an enantiomer if it cannot be mapped to its mirror image by rotations and translations alone.

Enantiomers do not differ in any of their physical properties (melting point, boiling point and so on) but one, differing in their effect on plane polarised light. One enantiomer will rotate the plane of polarisation to the left or anticlockwise, and the other will rotate it to the right. A (+) or d- is conventionally used to denote the enantiomer which rotates the plane of polarisation to the right, and a (-) or l- is conventionally used to denote the enantiomer that rotates the plane of polarisation to the left.

Alternatively, the absolute configuration of the molecule can be worked out according to certain rules. The result of applying the rules is a decision either that the molecule is right-handed (denoted R) or left-handed (denoted S).

One enantiomer can only be synthesised in preference to the other if the stereochemistry is already present in the starting materials or if enantiomeric reagents are used. Unless the conditions for the reaction are stereospecific the result of the synthesis will inevitably be a 50/50 mixture of the two enantiomers of the chiral molecule. This is called a racemic mixture or racemate. There is no guarantee that the physical properties of the racemate will be the same as the physical properties of the individual enantiomers that go to make it up: for example the melting point may be higher, lower or the same. The racemate has no effect on plane polarised light, the mixture being precisely 50% (+) and 50% (-).

Where there are two asymmetric centres in a molecule, there are unsurprisingly four possible isomers. Looking just at the chiral atoms, both may be mirror images of each other, and the molecules form an enantiomeric pair. If one only is a mirror image of the other, then the molecules are called diastereoisomers. Thus each molecule of the four will have one corresponding enantiomer and two corresponding diastereoisomers. To distinguish the diastereoisomers terminology describing the relative position of significant groups may be used. If the significant groups are on the same side of a ring, it is called a cis- structure, and if on the opposite sides of a ring, it is called a trans- structure:

The foregoing covers the terminology and the visual conventions in structural formulae used in the patent. I do not understand that any of the foregoing is said not to be part of the common general knowledge of the skilled person. It provides the notational framework for the specification’s statements about the compounds of the invention.

Different enantiomers of a chiral molecule react differently with other chiral molecules. This is of particular importance in natural systems since enzymes, which are proteins responsible for all the chemical reactions carried out by the cell, are chiral molecules and are present only as a single enantiomer. I quote Dr Newton’s report:

I understood the evidence to be that the skilled person at the priority date would expect that where a drug was a chiral molecule, it was highly likely that only one of the enantiomers and diastereoisomers (if any) would be responsible for its pharmaceutical activity. This did not mean, and does not mean, that chiral drugs had to be administered as single enantiomers but there is undoubtedly a modern tendency to prefer single enantiomers where resolution of the racemate is practicable.

The other aspect of the common general knowledge which it is necessary to consider is the common general knowledge in respect of statins generally. By 1986, statins were a well-known class of compounds recognised as having a potential application for cholesterol-lowering drugs. The first statins, mevinolin and compactin, were natural products that existed as single enantiomers. Much work had been done on these compounds, and it was recognised that they had a 4(R)-trans- structure:

The document called Stokker (‘3-Hydroxy-3-methylglutaryl coenzyme A Reductase Inhibitors’ J Med Chem 1985 28, 347-358 also seems to have been accepted to be a document the skilled person would undoubtedly become aware of in doing any statin work, if not common general knowledge in the strict Beloit v Valmet sense, at the priority date. It reports a substantial statin study, and both Prof Clive and Dr Newton said that it disclosed that it was likely that all the activity lay in the 4(R)-trans- isomer.

It was also known that the active form of the molecule was the open-chain hydroxy acid formed by hydrolysing the lactone ring. This acid can readily form salts, for example the sodium salt.

After the acknowledgment of certain prior art, the patent starts its Description of the invention (page 2 line 11), with a statement of the invention. It is described as certain ‘trans-6-[2-(3- or 4-carboxamido-substituted pyrrol-1-yl) alkyl]-4-hydroxypyran-2-ones’. There follows a remark about EP-A-179559, which is said not to disclose the 3- or 4-carboxamido substitutions in the pyrrole ring with which this invention is concerned. These are the substituents labelled R₂ and R₃ in formula I. Then formula I is set out. There is no dispute that the lactone ring is shown with the trans- substitution, as is repeated at page 2 line 34.

Page 3 of the patent sets out the method of preparing the compounds of formula I. This is a general description, and appears to be set out for all the compounds that formula I covers (‘In another aspect of the present invention, there is provided a method of preparing the compounds of structural formula I above…’). The first stage (compound VIII to compound IX) results in a chiral compound. It is common ground that compound IX will be a racemate if synthesised according to this scheme.

The next stage in the synthesis (page 3 lines 34-53) is described as producing a compound of formula X. Formula X in fact depicts a single trans enantiomer, but again it is common ground that the result of the synthesis described will be a racemate consisting of the two trans enantiomers.

At page 4 line 3, the description returns to the inventive class of compounds, this time with reference to the bridging group X in formula I. The bridging group is an alkyl chain and the preferred one is ethylene (page 4 line 7). After this short discussion, the stereochemistry of the compounds is expressly discussed in a short but significant paragraph at page 4 lines 8 to 12:

At page 4 line 21, the description of the preferred reaction sequence which ‘is used to prepare compounds of the present invention’ is described. The specification turns first to the synthesis of compound VIII. This is not chiral and need not be further considered. Reaction sequence II on page 7 sets out in detail the steps from compound VIII to compounds X and I. This sequence forms part of the invention (claim 8), and I have outlined it above. In talking of compound X, the specification states (page 7 line 51) that it contains ‘a predominance of the desired R*, R* configuration at carbon atoms three and five which bear the hydroxy groups’. I have not discussed the R* notation under the heading of common general knowledge. The notation is used to describe pairs of atoms that are relatively R to each other. It is disambiguated by adding the appropriate absolute value, so that R(R*,R*) is the same as (R,R). The specification explains that it is the (R*,R*) configuration that gives the desired trans- configuration to the resulting lactone.

The results of comparing ‘representative examples of compounds in accordance with the present invention’ are set out in Table 1. The preparation of each of these compounds is described in Examples 1, 3 and 4 of the patent, and it is common ground that each of these Examples will produce a racemate.

Finally, and this is a small point, that may be neutral, when the draftsman comes to put a name to a compound rather than a structural formula, in claims 3, 4 and 5, only in claim 3 does he include the (±) symbol to denote the racemate.

Mr Kitchin QC submits that this reaction scheme, the only scheme given for reaching ‘compounds of the present invention’ having formula I and which undoubtedly yields the racemate, will stand for all intents and purposes as a definition of what the patent means by formula I. It means the racemate. This consideration is reinforced by the consideration that Dr Newton agreed that from a medicinal chemistry standpoint, racemates and their component enantiomers ‘are totally different compounds’. The reason is that the biological response to one of two enantiomers of a racemate is completely different.

Against this, it is necessary to consider the common general knowledge:

The skilled person knew that compactin and mevinolin were potent anti-cholesterolaemics, and were single (R, R) enantiomers;

He knew from Stokker that it was likely that all active compounds of this description would be single trans- enantiomers, the (R,R) enantiomers, the (S,S) enantiomers being likely to have no activity;

He knew from his common general knowledge that a racemic mixture can be resolved into its component enantiomers. Professor Clive under cross-examination described this as a standard procedure, but not mindless. He put it like this (transcript pages 329-330):

So we have a patent which describes formula I as the compound of the invention, and tells the skilled person to make the compound of the invention by a method which will only produce a racemate. When it comes to the admittedly thin description of the treatment of human beings, it says nothing about resolution, or about the near certainty that only one enantiomer composing the racemate will matter. When it compares compounds of the invention with the prior art compactin, what it compares are, in fact, racemates. In the claim it uses a formula which may also be used to denote the enantiomers of interest in accordance with an acknowledged convention that permits it to be used to denote racemates.

Against this background, it is not surprising that Mr Kitchin QC relied strongly upon Lord Hoffmann’s remark that a specification is not a document inter rusticos for which allowances must be made. But this is an invention in a field that is well investigated. The skilled person’s common general knowledge extends to the overall characteristics of the claimed compounds, which are analogous to existing compounds that also possess those characteristics. It is also a feature of communications between skilled persons that the writer will feel no need to set out material which he can safely assume the other is aware of. In the case of a patent, which is a document directed to a person skilled in the art, the writer can assume that the reader possesses the common general knowledge. In my view, a proper approach to construction of this claim to ask why the patentee, who has covered a two-element composition for use in a drug, would wish not to cover one element of that composition which any reader would know was the effective element and which could be isolated using routine techniques. Of course, clear words would be conclusive, but there are no clear words. I can see that to ask this question would not be justified if the skilled reader did not have a clear understanding that the claimed material would have a component that was new and useful, but there is every difference between on the one hand an omission that was surprising but for which the patentee might have reasons, and on the other an omission that is both surprising and would in the eye of the reader immediately deprive the patent of any commercial effect.

Professor Clive’s view was that if one is going to use the name, or structure, of a single enantiomer in a document to denote a racemate, then you cannot use the same name or structure to denote the enantiomer as well. That seems to me to be obvious, although I imagine that it must be possible for context to change within a document and meanings to change accordingly. I have in mind a document divided into two parts, one headed ‘Preparation of the racemate’ and the other headed ‘Properties of the enantiomer of interest’ in which the same formula was used throughout. The change in context must, I think, be sufficient to change the meaning of the formula. The author of a document is obliged to explain what he is doing.

In the ’633 patent, it is absolutely clear from context throughout that formula (I) is being used to denote a racemate. In my judgment, every time the skilled person sees formula I or formula X he will see it with eyes that tell him that in that racemate, there is a single enantiomer that is the effective compound, and that he can resolve the racemate using conventional techniques to extract that enantiomer. When one comes to claim 1, which echoes the purpose of the invention with its conventional reference to pharmaceutically acceptable salts, he will, in my judgment, continue to see the formulae in this light. In my view, the claim covers the racemate and the individual enantiomers.

In the result, therefore, I refuse the declaration sought.

Claim 1 of the ’281 patent is to the hemicalcium salt of [R-(R*,R*)]-2-(4-f1uorophenyl)-,-dihydroxy-5-(1-methylethyl)-3-phenyl-4-[(phenylamino)-carbonyl]-lH-pyrrole-1-heptanoic acid. This can be conveniently referred to as atorvastatin calcium. It is the commercial material of interest.

It is accepted that the invention of claim 1 is entitled to priority from 21 July 1989. Two grounds of invalidity are raised: anticipation by WO 89/07598 (’598) and obviousness in the light of the application for ’633 (’633A).

On this part of the case, evidence was given by Dr Cunningham for Ranbaxy and Arrow. Dr Spargo gave evidence for Warner-Lambert. No criticism could be made of either. I can first deal with the allegation of anticipation.

’598 is a section 2(3) citation which means that it is not available to support an allegation of obviousness. These circumstances emphasise the importance of not drifting away from the strict principles of anticipation, which are set out in General Tire v Firestone [1972] RPC 457 at 485. For a claim to be anticipated by a prior disclosure, the prior disclosure must contain a clear description of, or clear instructions to do or make, something that would infringe the patentee’s claim if carried out after the grant of the patentee’s patent. If, on the other hand, the prior publication contains a direction which is capable of being carried out in a manner which would infringe the patentee’s claim, but would be at least as likely to be carried out in a way which would not do so, the patentee’s claim will not have been anticipated, although it may fail on the grounds of obviousness. A signpost, however clear, upon the road to the patentee’s invention will not suffice. The prior inventor must be clearly shown to have planted his flag at the precise destination before the patentee. This formulation is sometimes glossed as requiring that the invention be the ‘inevitable result’ of carrying out the directions of the prior publication. This is the test applied in the European Patent Office as well:

’598 is an application for ways of making, particularly, formula I, which is the same as formula I of ’633 save that group X is explicitly ethyl, -CH₂CH₂-. It is particularly interested in formula Ia:

On page 43, the application continues (line 15):

It follows that the material claimed in claim 1 is an expressly specified salt (calcium) of the preferred isomer of one of the three materials explicitly specified. If one is in any doubt, it is easy to compare the final structural formula on page 12 of ’281 against formula XII on page 40 of ’598. They are identical, save that in ’281 the calcium salt, and in ’598 the acid, are shown. In fact, the synthetic route described in ’598 actually produces a racemate. But this time, the precise enantiomer (4R,6R) is specified. This notation means the same thing as the [R-(R*,R*]… used in respect of the acid in claim 1 of ’281. The evidence (which I have already discussed) was that resolution to obtain the enantiomers was common general knowledge. It is no answer to an allegation of anticipation that the specification gives clear and unmistakable directions to use the common general knowledge to produce a specific material.

I conclude that this is a clear case of anticipation of claim 1 of ’281. ’598 gives specific directions to make the three preferred enantiomers, one of which falls within the claim.

So far as claim 2 is concerned, Warner-Lambert surprisingly suggest that’598 does not disclose pharmaceutical compositions of atorvastatin. This is incorrect. Claim 15, which claims a method of making compound Ia, concludes with this feature:

It is occasionally said that there cannot be clear and unmistakable directions to do something which is described as optional. I do not agree: to describe the thing as optional is to describe the thing. It is rather like the disclosure of something as adjustable: it necessarily also discloses something that is not adjustable—see Gillette v Anglo-American (1913) 30 RPC 465. Claim 2 is also anticipated.

There is a dispute between the parties as to the right approach to obviousness in this case. Warner-Lambert take as their exemplar the decision of the Technical Board of Appeals 3.3.1 of the European Patent Office on the present patent. The Examining Division had rejected the application in the light of a document (US 4,681,893) whose disclosure is substantially identical to ’633A upon which Ranbaxy and Arrow rely before me. This document discloses in Example 2 the sodium salt of the atorvastatin racemate. Warner-Lambert submitted new claims (the claims with which I am concerned) and submitted that the claim to the hemicalcium salt of the R-trans enantiomer was not obvious. This argument was successful, for reasons appearing clearly in the following extract from the decision of the Board:

In view of this state of the art, the problem underlying the present application as submitted by the Appellant consists in providing a hypocholesterolemic compound having improved handling properties, in particular improved hygroscopicity and solubility.

As a solution to this problem the present application proposes the hemicalcium salt of the particular R-enantiomer as defined in claim 1.

To support his submission that the alleged improvement is achieved by the claimed invention, the Appellant referred to his experimental report filed on 20 June 2000. That test report comprises experimental data about the hygroscopicity and the solubility of the hemicalcium salt of the R-enantiomer according to the claimed invention, on the one hand, and of the sodium salt of the racemate of that enantiomer according to example 2 of [’633A], on the other….The Board is satisfied that the problem underlying the patent in suit as defined in point 4.2 above is successfully solved by the claimed subject-matter.

Finally, it remains to be decided whether or not the proposed solution to the problem underlying the patent in suit involves an inventive step.

Before me, the question was examined both on a problem-solution basis and on the approach set out in Windsurfing [1985] RPC 59. I shall set out the facts relating to the common general knowledge in the art, and then return to the rival analyses.

Much of the common general knowledge I have already considered. By 1989, resolution of racemates with pharmaceutical activity was well established. So far as this family of statins is concerned, the skilled addressee (the team to whom I have already referred) would have known that the activity would reside in the R,R enantiomer, and would also have known that the active form was not the lactone but the open chain acid and its salts.

It was also known that the salts would be likely to be more soluble, and so possibly more bioavailable, than the free acid. Dr Spargo accepted that the skilled person would investigate the salts as well as the free acid. The problem confronting the skilled addressee is to get the active ingredient into a pharmaceutical composition suitable for administration. A solid dosage form requires a solid active ingredient: the obvious choices are the salts.

To test salts for their properties, it is standard practice for the skilled person to carry out a ‘salt screen’. In the case of an acid, this would include the standard sodium, potassium and calcium salts. These are the most commonly used salts. A review article by Berge & ors. from 1977 called ‘Pharmaceutical Salts’ in the Journal of Pharmaceutical Science showed that these three cations were the commonest by an order of magnitude, as Dr Spargo accepted, and they are all obviously ‘pharmaceutically acceptable’. The screen would on Dr Cunningham’s unchallenged evidence include a test for hygroscopicity, rate of dissolution and solubility. The hygroscopicity test is carried out in a controlled humidity chamber, or otherwise using the standard equipment of any formulation laboratory. The whole process is entirely routine.

Remembering that the foregoing is the knowledge that the skilled person will necessarily bring to bear on the disclosure of ’633A, one can turn to compound 1 of Table 1, which is atorvastatin lactone racemate, and Example 2, which describes making atorvastatin racemate sodium. Atorvastatin lactone racemate is made according to Example 1. These are the two leading examples of the specification. Dr Cunningham’s evidence was that compound 1 of Table 1 was of particular interest, a conclusion with which Dr Spargo agreed.

’633A also contains a general disclosure in respect of all the ring-opened compounds to the effect that they may be used as salts (page 7 line 54 to page 8 line 4):

There is, in my view, a clear teaching that the lactone racemate may be hydrolysed to produce the acid racemate, and that the acid may be used to react to form metal salts, calcium being one of those expressly mentioned. Furthermore, it describes how to make the sodium salt by reacting the lactone directly with sodium hydroxide in a 1:2 mixture of tetrahydrofuran/water.

What, then, is the difference between this and the invention of the patent in suit? The first is that ’633A discloses compound 1 as a racemate, and not as an enantiomer, and the second is (so far as the example is concerned) the salt is the calcium and not the sodium salt, and (so far as the general teaching is concerned) calcium is chosen from a list of seven inorganic cations.

Is the difference obvious? In my judgment it most certainly is. The resolution of the racemate was common general knowledge at the date. It was becoming preferred in the industry generally, and this particular resolution did not, on the evidence, involve any work that was not common general knowledge. Seven salts are specifically described. How can it be inventive to use one?

The answer to this problem raises a difficulty which, in English law, has been more or less laid to rest. There is no doubt that where there is a disclosure of a class of compounds, it is possible to select from that class a subclass united by a common feature distinguishing it from the rest: these are so-called selection patents. A fortiori, it is possible to select a single compound having advantageous properties from a class— see IG Farbenindustrie AG’s Patents (1930) 47 RPC 289. But underlying this principle is the necessity for the prior disclosure to be a disclosure of a class, rather than a disclosure of the individual members of that class as distinct entities. The difference is explained by Lord Wilberforce in E I Du Pont de Nemours (Witsiepe’s) Application [1982] FSR 303. This was an opposition under section 14 of the Patents Act 1949, and the only issue was prior publication. Although a patent could be successfully opposed under the 1949 Act if the invention was ‘clearly obvious and lacking in inventive step’ that question was not in issue in this case. Lord Wilberforce explains that a selection invention will not be prior published if (1) all the selected members of the class possess the advantage (2) the later specification discloses what that advantage is and (3) the prior publication of the wider class does not refer to that advantage. He says this:

In other words, the selected class may still be obvious, but the nature of the advantage will be one of the factors to be taken into account in assessing obviousness. Witsiepe was a case under the 1949 Act, but the same principles have been accepted by the Court of Appeal in a 1977 Act case, Hallen v Brabantia [1991] RPC 195, and it is too late to query that. In any event, it seems to me that the principle is a sound one. Unless the later patent states what the advantage possessed by the selected class is, it is merely an arbitrary selection among things already disclosed, and will lack novelty.

I think that the belief that the law of selection is concerned with obviousness to be a misconception. Obviousness only becomes relevant if the later patent is not anticipated, and the obviousness of the selected class will be decided according to the normal principles. It will no doubt help the patentee to repel an allegation of obviousness if he can point to a statement of the advantage possessed by the selected class, but I do not believe it to be essential, as I believe Lord Wilberforce makes clear.

The EPO view is stricter. As expressed in T198/84 Hoechst/Thiochloroformates [1985] OJEPO 209, it seems to be to the effect that a newly discovered effect can never add novelty to a narrower class if the class is otherwise old. The claim was to a method of making thiochloroformates using a particular catalyst in the range 0.02 to 0.2 mol%. The prior art was a disclosure of the process with the same catalyst present in the range 0-100 mol%. The Board held that there was no anticipation: in [7] they say this:

I read this as saying that so far as the EPO is concerned, there must be no disclosure of the selected class, either as to its individual members or as to the class as a whole if the invention is to be new. To give an example, if the disclosure of the seven inorganic cations was to be construed in context as a disclosure of the class but not of the individual members, the calcium salt would be new. If it were a disclosure of the individual salts made having those cations, the invention would be old. The advantage possessed by the selected class or individual over the prior art class merely confirms the conclusion to be drawn from considering the prior art disclosure as a whole.

As a matter of English law, ’281 could not be a valid selection patent because no advantage other than that claimed by the disclosed materials is claimed for the invention of claim 1. But since for obvious reasons anticipation is not in issue, questions affecting selection do not need to detain us. The real difficulty is caused by certain aspects of the problem-solution approach. I start with the description of this approach from the Guidelines:

There are two possible difficulties with this approach (there may be as many, or more, with Windsurfing, I do not know). The first is its concentration on the closest prior art, which must stem from a belief that if an invention is not obvious in the light of the closest prior art it cannot be obvious in the light of anything further away. This runs the risk of offending against the principle that a skilled man must be permitted to do that which is obvious in the light of each individual item of prior art seen in the light of the common general knowledge. There is an illustration of this risk in the present case. The second is that the reformulation of the problem can obscure that which is objectively obvious. But as stated, I cannot see this approach producing a different result from that produced by a Windsurfing analysis in the vast majority of cases. Where results differ, I suspect that it is because of the importance that a judge following the Windsurfing approach will give to the common general knowledge.

I am not particularly concerned about the EPO result in the present case. This was an appeal in the course of prosecution, with all the risks that hearing only the patentee entails. I am, however, rather more concerned about the methodology. First, I do not see why Example 2 of ’633A was taken to be the closest prior art. Why was a disclosure of the racemic hydroxy acid, together with the list of possible cations, not considered? Why does the patentee appear to have been allowed to pass over the greater part of the teaching of the document in relation to possible cations? Again, I suspect the answer to this question lies in the exigencies of in-prosecution appeals, and in the precise terms of the objection taken by the Examining Division.

Second, I do not understand the reformulation of the problem. There was, objectively, no problem with the sodium racemate until internal documents showing what may have been a tendency towards gel formation in the sodium salt were disclosed to the EPO. But this tendency was discovered at the same time as the calcium salt was tested. One of the documents in the case was the so-called ‘Racemate Report’. This document recorded what was on Dr Spargo’s evidence a preliminary salt selection study of the atorvastatin racemate, carried out before the priority date. It appears that the calcium and n-methylglutamine salts showed promise. Data contained in this report and in one other underlay a letter to the EPO from Warner-Lambert’s representative stating that the sodium salt showed poor hygroscopicity and a tendency to become sticky which was not shared with the calcium salt. This led to the recharacterisation of the ‘problem’ recorded in the part of the decision I have quoted. But there was no objective problem: the calcium salt was being tested alongside the sodium salt. We find ourselves in the strange position that if the sodium salt had been satisfactory, there would have been no invention in going for the calcium instead, but since it was not, there was an invention. It seems to me that if redefinition of the problem is permitted on the basis of any advantage which was known to the patentee before the priority date but not referred to in the specification or which is discovered after the priority date then there is a substantial risk that the reformulation will result in a finding of non-obviousness: how can one solve an (objective) problem that one did not know existed?

It is for this reason that in this jurisdiction after-discovered advantages are highly unlikely to be capable of supporting inventiveness, for the reasons given by Jacob J in Richardson-Vicks [1995] RPC 568 and T867/95 RICHARDSON-VICKS/Cough/cold mixtures (TBA 3.3.4). The later decision emphasises in [14] that reformulation of the problem can be allowed “provided the skilled man could recognise the same as implied in or related to the problem initially suggested”.

I do not detect in this case any real difference in the substantive law between the jurisdictions. It is true that the problem-solving techniques may be different, but I do not see either in this case or in others differences that cannot be attributed to differences of appreciation, rather than principle. I return to the present case.

Having now been presented in ’633 with a new lactone (formula I) and a new open-chain acid (formula X) what does the skilled person do? The answer is that he carries out preformulation testing for such things as absorption in vivo (which is why the lactone was rejected) and he will include a salt screen for his acid, to see which of the salts shows a potentially desirable combination of features. In my view, once the evidence establishes that this work will be done and involves no inventive endeavour, the various salts that are screened will be obvious. He is screening for pharmaceutical formulation. Claims 1 and 2 are obvious.

In the result, I refuse the declaration of non-infringement sought by Ranbaxy in respect of ’633, and I find ‘281 invalid for anticipation and obviousness. I will hear counsel on the order if it cannot be agreed.