Mayne Pharma Ltd & Anor v Debiopharm SA & Anor

In this action, the Claimants (“Mayne”) originally sought revocation of ten patents standing in the names of the Defendants (“Debiopharm”) and declarations that the marketing in the United Kingdom of a product manufactured in accordance with a Confidential Product and Process Description would not constitute an infringement of any one of those patents. By the time that evidence came to be exchanged, the scope of the dispute had narrowed to four of the specified patents, and in respect of two of those Debiopharm declined to file any evidence. Accordingly, before me the issues of infringement and validity were to be decided in respect of EP(UK) 0943331 (“’331”) and EP(UK) 1308454 (“’454”). For various reasons which it is not necessary to go into, the timetable for experiments and repetitions ran for longer than it should have done, and the results of the repetitions of Debiopharm’s experiments in relation to infringement of ’331 only became available during trial. It became apparent that it would be difficult to maintain the allegation of infringement in relation to this patent and it was formally withdrawn. Accordingly, this judgment deals with the validity of ’331 and with both infringement and validity of ’454.

Both patents are concerned with oxaliplatin, which is an anti-neoplastic platinum complex. ’331 is concerned with pharmaceutical formulations of oxaliplatin, and ’454 with a process for its preparation. By their respective priority dates, February 1998 and April 1996, oxaliplatin was an old compound. Its preparation had been referred to in USP 4,169,846 (Kidani) published 2^nd October 1979. Dr Cleare, Mayne’s expert, who was thoroughly familiar with the field, said that oxaliplatin could be viewed as the third generation platinum-based compound for use in cancer therapy. Three such compounds have been approved for clinical use: cisplatin in 1979, carboplatin in 1986 and oxaliplatin in 1996. Oxaliplatin is no longer protected by a patent.

Before turning to the patents themselves, it is convenient to deal with a question which was controversial at trial, the identity of the skilled addressee. Both these patents are what are sometimes called “second-tier” patents, in that they are concerned with improvements in formulation or in production method of a known active ingredient. On the face of it, therefore, they are addressed to persons already acquainted with the manufacture of the active ingredient and its pharmaceutical formulation. Mr Waugh QC, who appeared for Debiopharm, suggested that the addressee of the specification was a general organometallic chemist, represented by Professor Davies, who was Debiopharm’s witness on this part of the case. Mr Watson QC, for Mayne, suggested that the addressee of the patents had at least knowledge of other platinum-based anti-cancer compounds. The reason for these contrasting contentions lies in the question of the common general knowledge. I think this is one of those cases where it is important to bear in mind that those primarily interested in the invention will be employees of pharmaceutical companies interested in improving, or making for the first time, pharmaceutical compositions containing oxaliplatin. Chief among those would be those manufacturers who have already made oxaliplatin, carboplatin or cisplatin. Those seeking to enter this field for the first time do not provide what I can call an appropriate template for the skilled addressee, who must represent the attainments of those already in the field in which the invention is made. New entrants to a field may have clearer sight than those already in it, and lack the prejudices properly to be attributable to those with experience. At the same time, a new entrant into the field, albeit a specialist field within his or her general knowledge, will not possess the degree of experience which must also be attributed to the skilled addressee if a proper balance is to be held between the two extremes, of too much stiffness in refusing, and of too much easiness in admitting, any alleged inventive step. It is not sensible not to attribute to the skilled person the common general knowledge of those presently engaged in the manufacture and formulation of platinum-based pharmaceuticals.

Mr Waugh QC argues that since the specification enables a wider class of persons – organometallic chemists – to put the invention into effect, then that class of persons is the relevant class. I do not think this follows. After all, obviously the patent is in principle of interest to anybody, whether or not an organometallic chemist, who wishes to enter the field. That fact cannot be relevant to identifying the skilled addressee. It is not legitimate to draw the class of addressee so wide that the specific knowledge and prejudices of those most closely involved in the actual field with which the patent is concerned do not form part of the prejudices and attributes of the skilled person.

With that introduction, I can turn to the patents themselves. It is convenient to start with ’454.

The title of the invention is “Process of preparing platinum compound” and is said to relate to a process of preparing cis-oxalato (trans-l-1,2-cyclohexanediamine) Pt(II), which is oxaliplatin. Paragraph [0002] describes the general reaction scheme for the preparation of oxaliplatin. For present purposes, the scheme has the following important features:

The so-called trans dichloro material is reacted with silver nitrate to produce the so-called trans diaquo complex (also referred to as diaquo trans Pt DACH (Footnote: ¹)); and

The diaquo complex is reacted either with oxalic acid or with an oxalate.

The reaction is said to be disclosed in (inter alia) Kidani. It looks like this:

Paragraph [0003] describes the problem which the patent sets out to overcome. This is the problem of the formation of impurities by side-reactions. The principal side-reaction in question is the dimerization of the diaquo material. This competes with the reaction with oxalic acid, with the consequences stated in paragraph [0003]:

Paragraph [0004] points out that pharmaceuticals of this description are required to have high purity, and paragraph [0005] points out that the labour and cost of separation and removal of impurities is large and will involve the loss of a proportion of the desired product. Accordingly, as paragraph [0006] observes, it is desirable to depress the formation of impurities by shortening the reaction time while nonetheless increasing the yield. The solution to the problem is correctly stated in paragraph [0007] to be a simple one: it is to adjust the pH of the solution in which the reaction is taking place so as to increase the level of dissociation of the oxalate ion. This is clearly described in paragraphs [0009] – [0012]:

There follows a detailed description of the invention which adds little to that which I have quoted above. Paragraph [0016] says that the inventors have conducted a series of experiments under the hypothesis that in order to promote the coordination exchange reaction between the diaquo complex and the oxalic acid or the oxalate derivative, the degree of the dissociation of the oxalic acid or of the oxalate derivative is elevated to increase the concentration of the oxalate ion in the aqueous solution in which the reaction takes place, which, as paragraph [0015] points out, is strongly acidic (pH less than 1) before the addition of the oxalic acid. Paragraph [0017] states that the pH of that solution may be elevated by the addition of alkali solution and paragraph [0018] deals with the limits between which the pH of the alkali solution should fall after the addition of the alkali solution.

Claim 1 is as follows:

Claim 2, which is alleged to possess independent validity, is to the process of Claim 1, wherein pH is adjusted to between 4.0 and 5.0.

The basic question that arises on this claim is whether it requires a distinct step consisting of the addition of an alkali solution. The question arises in this way: if the oxalate anion is provided by the addition (for example) of a near-neutral solution of some oxalate salt, and the result is a reaction proceeding at a pH between 4.0 and 5.0, or between 3.0 and 6.0, is that enough?

Let me ignore for the moment the words “and/or the oxalate derivative”. It is obvious that there is no difference between adding oxalic acid in an equimolar mixture with sodium hydroxide on the one hand and adding sodium oxalate on the other. So much is elementary chemistry, as the evidence established. At the date, the skilled man would immediately understand that this was exactly equivalent to simultaneously adding oxalic acid and sodium hydroxide to the solution of diaquo complex. Indeed, paragraph [0019] of the specification makes that quite clear. The specification exemplifies three suitable laboratory alkalis: potassium hydroxide, sodium hydroxide and lithium hydroxide. The claim appears to require addition of an alkali solution even when the “oxalate derivative” rather than oxalic acid is added. But there seems to have been no dispute that potassium oxalate, sodium oxalate and lithium oxalate are each properly to be regarded as oxalate derivatives.

The problem is acute because in its manufacturing process, Mayne’s supplier uses the ammonium salt of oxalic acid. In solution at the concentrations with which I am concerned, the solution of ammonium oxalate is mildly acidic (it has a pH of about 6.5).

One turns to the specification to ascertain what the purpose of the addition of the alkali is. The passages quoted above show that its purpose is to improve the dissociation of oxalic acid (or oxalate derivative) when added to the acidic solution of the diaquo species. A word of explanation is in order here. Acids are acid because in aqueous solution they lose a hydrogen ion, a process known as deprotonating. Acids fall, for present purposes, into two classes: strong and weak. Strong acids substantially dissociate in aqueous solution. Weak acids do not. In aqueous solution, a weak acid will go into equilibrium as shown:

Essentially, therefore, the purpose of the addition of alkali is to increase the available oxalate ion concentration. The downside, if I may call it that, is that the formation of impurities is favoured by a less acid reaction medium.

Mayne accepts that the skilled person would realise that the addition of an oxalate salt would be recognised by the skilled man as having the same practical effect as creating the oxalate from oxalic acid in situ by adding alkali. They submit, however, that unless the claim is construed so as to require the addition of an alkali solution as part of the process, it is tantamount to rewriting it as a process for preparing oxaliplatin by reacting an oxalate with the diaquo complex, characterised in that the pH at the beginning of the reaction is between 3.0 and 6.0 or 4.0 and 5.0. That is true, but it is not in itself an objection to construing the claim in this way: the question is simply the meaning of the claim in its context in the specification.

The speech of Lord Hoffmann in Kirin-Amgen Inc v. Hoechst Marion Roussel Ltd [2005] RPC 169 describes the task of the Court in purposively construing a patent specification in this way, in [34] - [35]:

At first sight, it is very strange that the specification expressly contemplates the simultaneous admixture of alkali solution to the oxalic acid as it is added to the solution of diaquo material, but appears not to contemplate a previous preparation of (for example) sodium oxalate alone. The answer, I think, lies in the fact that the specification always contemplates that pH would have to be adjusted, whether oxalic acid or ‘oxalate derivative’ is used as the source of oxalate ions: in other words, that equimolar quantities of oxalic acid and alkali solution would not be used. It is perhaps a straw in the wind, but paragraph [0002] of the patent does acknowledge U.S. 4,169,846 (Kidani) which discloses in example 4(i), read with example 3, the synthesis of oxaliplatin by the addition of potassium oxalate alone to the diaquo complex. Obviously, this disclosure, which I discuss in detail below, provides in itself a perfectly understandable explanation for the draftsman’s having included the express requirement of the presence of a distinct alkali-addition step. If he had not done so, there was a risk of anticipation. It must be accepted that the explanation set out in paragraphs [0003]ff for the elevated impurity level when reaction time is extended relates only to addition of the acid, there being in fact little discussion of the effect of the addition of alkali solution to a reaction mixture in which an oxalate derivative is used as the source of the oxalate anion, but I do not regard this as significant.

The example of the invention and the two comparative examples both exemplify oxalic acid as the source of the oxalate anion.

There is really very little clue in the specification as to the problem sought to be overcome by the addition of further alkali when the oxalate anion is provided by a salt of oxalic acid. The disclosure of the specification is largely concerned with the use of oxalic acid and with its neutralisation by the addition of a strong alkali, and in that context the emphasis throughout the specification on the distinct addition of an alkali solution (whether simultaneously with the addition of the oxalic acid or a short time after it) is entirely understandable.

I have come to the conclusion that the specification is simply not concerned with a system in which the pH of the solution prior to the addition of any alkali lies between the limits called for by the claims, because the oxalate ion is provided by a salt of oxalic acid. It is, in other words, a comparatively narrow invention, confined to the correction of processes in which the addition of oxalic acid or the chosen oxalate derivative to the diaquo complex produces significant quantities of undissociated, and so unreactive, oxalate. The addition of the alkali solution is the correction for this condition, and accordingly it is a distinct step in a process which would otherwise employ either oxalic acid alone or an oxalate derivative giving a solution whose pH lay outside the desired limits.

The method of synthesis employed to manufacture Mayne’s material is described in the confidential process and product description. It is not necessary at all to describe how this synthesis arrives at the diaquo material. The diaquo material is reacted with ammonium oxalate. Ammonium oxalate solution is itself mildly acid because of the presence of the ammonium ion, which will tend to associate with hydroxyl ions, leaving some free protons. For a kindred reason, potassium oxalate solution is mildly alkali: the oxalate ions tend to some extent to associate with protons, leaving some free hydroxyl ions. These are partial descriptions only of what is going on, but for present purposes it is only necessary to report that the solution added by the Claimants’ manufacturer is, in fact, slightly acid.

Oxaliplatin is precipitated.

The claim is, in my view, not infringed. The reasons for this conclusion really appear from my discussion of the interpretation of the specification, but the essential acidity of the ammonium oxalate solution added demonstrates that it is dangerous to read this claim any more widely than its clear and ordinary meaning.

There is, obviously, no addition of an alkali. Debiopharm argue that there is no difference between adding ammonium oxalate on the one hand and adding oxalic acid and ammonium hydroxide (ammonia solution) on the other. As I have said above, every chemist would realise that to the extent that an equimolar solution of alkali is added there is no chemical difference between the two. But there is no need to adjust the pH if the effect of the addition of the oxalic acid derivative (ammonium oxalate) results in a solution whose pH is already within the limits called for by the claim, and so Mayne’s supplier does not do so. Nor is the solution of ammonium oxalate alkali: it is slightly acidic (pH=6.4). It is not possible to consider a solution that is acidic as an alkali solution within the claim. I think that when an ‘oxalate derivative’ is used the claim will only be infringed if further addition of an alkali takes place. Otherwise, it is difficult to put any clear meaning on the words used. There are such ‘oxalate derivatives’: the potassium half-salt (HK(COO)₂) is an example provided by Professor Davies.

The allegation of infringement must accordingly be rejected, and I shall grant a declaration of non-infringement in a suitable form.

Certain of the attacks levelled by Mayne at this patent are attacks on the basis that Mayne’s construction of the claim is the right one. There are also attacks directed to the validity of the patent on the footing that Debiopharm’s construction (rejected above) is the right one and that the claim covers a process in which the addition of an “oxalate derivative” consisting of a salt of oxalic acid results in a reaction mixture which is within the pH limits of the claim. I shall consider first the attacks on the patent construed as I consider that it should be construed.

The attack on the basis of common general knowledge was a distinct attack. The starting point was the contention that in principle the reaction between the oxalate ion and the diaquo complex for the manufacture of oxaliplatin was common general knowledge.

Dr Cleare in his first report describes in some detail the place of oxaliplatin in the history of platinum-based anti-cancer drugs. He observes that oxaliplatin had been published by Kidani in the late 1970s and that oxaliplatin was approved for clinical use in 1996. He considers that by 1996 the platinum coordination complex pharmaceuticals represented an identifiable field. The field comprised particularly those families of compounds of which cisplatin, carboplatin and oxaliplatin were exemplars. As I have indicated above, I consider that it is right in this case to have regard to the fact that these compounds are established therapeutic agents having a particular field of application upon which a number of manufacturers concentrated at the priority date. In my view, it is appropriate to attribute to the skilled person a proper understanding of the chemistry of the platinum compounds. Even a new entrant to the profession would, in 1996, wish to synthesize oxaliplatin in order to employ it in a pharmaceutical preparation: there is no evidence of any other use for it. It would be essential, in my view, for such a person (I refer to a person, but of course the person may be an embodiment of a real world team) to know about such things as impurities, which are inevitable in the production of a pharmaceutical, and yield – particularly where expensive raw materials are involved.

It is the contention of Mayne that it is entirely obvious to optimise the production process for a pharmaceutical compound, to minimise impurities and maximise yield. The witnesses were agreed, subject to a point that concerned Professor Davies about ’012, to which I shall return, that this was indeed the case. I am in no doubt that the skilled person would also appreciate that the degree of dissociation of oxalic acid, and so the availability of oxalate anion to react with the diaquo complex, would be common general knowledge. It is basic chemical learning that where a soluble species partly dissociates in aqueous solution, it is possible to shift the resulting equilibrium by increasing or decreasing the concentration of one or other of the dissociated components of the compound. It was obvious that oxalic acid, a weak acid, would dissociate less in a strongly acid solution than it would in a neutral solution. It would also be obvious to the skilled person that the degree of dissociation of oxalic acid in aqueous solution can be controlled by controlling the pH of that solution by addition of alkali. It was also obvious to the skilled person that the rate of production of oxaliplatin depends upon the concentration of the oxalate anion.

Dr Cleare was of the view that the skilled person whom he had identified – one with experience of the chemistry of these platinum complexes – would know, in general terms at least, that the diaquo species was capable of reacting in aqueous solution to produce dimers and (for example) hydroxo complexes. These are impurities. The skilled person would also know that the risk of impurity creation would increase as pH increased. The dimer is produced certainly at pH > 3:

Professor Davies could not challenge the opinion of Dr Cleare as to the basic knowledge of the platinum-complex chemist, having no directly comparable experience, and the opinion was not successfully challenged in cross-examination. I find, therefore, that part of the common general knowledge of the skilled person was that the diaquo DACH was capable of dimerising, and of forming other impurities, in solution.

This, then, is the dilemma which faces the skilled person wishing to optimise his process. He knows that he will improve the dissociation of his oxalic acid if he increases pH: but he knows also that an increase in pH will increase the production of impurities by dimerisation (and other reactions) of the diaquo material. Dr Cleare was of the view that, given that the skilled person was aware of the fact that both the desired reaction rate and the rate of undesirable side-reactions were sensitive to pH, the skilled person would look for a pH that represented the best compromise. Dr Cleare took the view that it would be sensible to start around the second pKa of oxalic acid. (pKa, the dissociation constant of acids, is explained in the primer forming Exhibit 2 to Professor Davies’ report.) The second pKa is chosen because oxalic acid is a dicarboxylic acid, and the second pKa value is a measure of the extent to which the second carboxylic moiety is dissociated, the first being assumed to be wholly dissociated. If the pH is adjusted to be equal to the second pKa of oxalic acid, the degree of dissociation of the second carboxylic moiety will be 50%.

Professor Davies was vigorously opposed to this suggestion. It was on this issue that there was a serious divergence of opinion between the two experts. Dr Cleare maintained that it is standard in process chemistry to optimise the process by reference to a number of variables (pressure, temperature and so on) and pH is a prime parameter when the process is aqueous:

The contrary view was taken by Professor Davies (transcript 617 ff):

I have set out the whole of this passage, because I am invited to reject Professor Davies’ evidence on this point on the footing that it represents an extreme and illogical position and also that the evidence that the Professor gave on the same subject under cross-examination might be viewed as inconsistent (transcript 608-9):

It will be apparent from the first-quoted passage from Professor Davies’ cross-examination that the relationship between him and the cross-examiner was poor. I attempted to find some middle ground, but I remain surprised and unconvinced by the contention he advanced, neatly summarised in Debiopharm’s skeleton argument as follows:

I think this makes no sense. It is not even technically correct (pH 4-5 does not involve as slow a rate for production of oxaliplatin as pH 3). Put shortly, as pH increases, reaction rate increases but so does the production of impurities. But the two do not have the same dependency on pH (I will return to this subject in considering the disclosure of Gill & Rosenberg, which was not shown to be common general knowledge). The common general knowledge is therefore that there are two competing reactions, both pH dependent, but which have (or may have) different rates. There is or may be a trade-off. That is a matter for investigation: you cannot reject all pH values between 3 and 7 a priori. The argument proves too much: even pH 7 would not be obvious to use in the sense given to the word, because of excessive impurity production.

If the skilled person will without invention investigate the position between pH 3.0 and pH 6.0 for any purpose, to find that there is between them some desirable range (pH 4.0 to 5.0), that cannot as a matter of common sense involve an invention that the law rewards with a patent, either as to the wider or the narrower range.

‘Syntheses, Kinetics, and Mechanism of Formation of Polynuclear Hydroxo-Bridged Complexes of (trans-1,2-diaminocyclohexane)platinum (II)’ J Am Chem Soc 1982, 104, 4598-4604 (‘Gill & Rosenberg’) is a paper reporting an examination of the stability of the diaquo platinum (II) DACH complex in solution and the formation of the dimer and trimer complexes. It is not about oxaliplatin, which is not mentioned. What it discusses is the impact of pH on the formation of dimer and trimer complexes (compounds II and III) and the dihydroxo complex (compound Ib). The way the compound was studied was to use ¹⁹⁶Pt nuclear magnetic resonance to investigate ‘[t]he effect of changes in concentration, pH, ionic strength, and temperature on the rate of formation of the dimer…’ In the experiments reported, the hydrogens of the aquo and hydroxo groups were replaced with deuterium (chemically identical to hydrogen, but with different physical qualities) for the purpose of other studies, and the experiments were in consequence conducted in D₂O rather than water with pD (the corresponding measure of concentration of deuterons to pH) being measured as an indication of acidity. Thus before each run, the pD of the solution in D₂O was adjusted with sodium deuterioxide rather than sodium hydroxide, and so on.

The paper reports that the diaquo Pt(II) DACH complex is unconditionally stable at pD of 3.1:

Quite apart from Dr Cleare’s view that this document had value in showing in detail that which the skilled man would have in mind in general terms as a matter of common general knowledge, the question is whether the invention is obvious having regard to its disclosure. The starting point for this is the process for making oxaliplatin itself by reaction of the diaquo DACH with the oxalate anion. Dr Cleare pointed out that the reaction would certainly proceed at a low pH but would take a long time. Long time for reactions can give rise to impurity problems. Dr Cleare accepted that a pH of 2.7 (equivalent to pD of 3.1) as disclosed in the document would work (Transcript 325 to 326):

The foregoing answer was to a question of mine. Mr Waugh QC followed it up:

Dr Cleare said that he could not say whether or not a pH of 2.7 would be optimum. He accepted that Mr Waugh QC was correct when he suggested that the paper did not discuss impurity formation in the diaquo DACH/oxalate system, but contended that that fact in itself only required the optimum pH to be investigated. He said this of the corresponding statements in his second report:

It was on the basis of the Gill & Rosenberg paper that Professor Davies had said that the skilled person would not go above pH of 2.7 or below a pH of ~7. Dr Cleare’s evidence was given against his background of great familiarity with the platinum complexes and cisplatin and carboplatin in particular. I think that for this reason I prefer the evidence of Dr Cleare on this issue. The invention is obvious as to both pH ranges, in the light of Gill & Rosenberg.

The foregoing conclusion renders any further consideration of the obviousness of this patent otiose, but I should nevertheless consider the case advanced on the basis of EP 0 136 012 (“’012”) (Kidani et al) because it does incidentally throw light on the issues discussed above.

’012 is a patent concerned with the synthesis of a variety of platinum (II) DACH complexes. Generally speaking, the compounds disclosed are analogues of oxaliplatin using monocarboxylic and dicarboxylic acids other than oxalic acid. The dicarboxylic acids are described in class (B) of the claim and at page 1 line 54. The disclosure extends to using mixtures of monocarboxylic acids for adding different acids to the two legs of the complex. The synthetic route is via reaction with (trans-l-DACH Pt(II)) (NO₃)₂ in aqueous solution, which is the diaquo material. So the reaction is analogous to the reaction the subject of the patent in suit. Each of the reactions described in the examples (page 6–page 8) with the exception of that of example 1 involves the addition of the acid to the diaquo species in aqueous solution followed by the addition of sodium hydroxide, the purpose of which is explicitly stated in examples 2, 3, 4, 5, 6, 7 and 12 to be to adjust the pH of the solution to stated values, variously, pH 4, 5 and 7. Example 1 shows the addition of NaOH to the carboxylic acid before addition to the diaquo DACH Pt(II) solution, which of course forms the carboxylate salt before addition.

Mayne say that this sort of pH adjustment is obvious to use for exactly the same reason as it is, they say, obvious to use it for oxalic acid. Professor Davies suggested in cross-examination (contrary to his previous evidence that the pH was acid-specific) that these examples were intended to determine the optimum pH for all carboxylic acids. This possibility can be immediately rejected as a matter of language. There is no suggestion that these are anything other than examples acid by acid to demonstrate synthesis of the corresponding complexes. No use is made of any results showing durations of the reactions: some are comparatively short (overnight in example 6) or longer (4 days for example 2) or very long (2 months for example 3). There is no obvious correlation between pH and yield, although generally lower pH meant longer time (Dr Cleare’s point about available carboxylate ion). Professor Davies said that the only teaching of the document was to use pH 7, or at least that this would be what is sometimes called the ‘take-home message’, but as it happens two of the lowest yields were accompanied by the highest pH which again one would expect in general terms. But in no way can this be considered a proper optimisation study for any of the acids, let alone all acids.

I regard this document as tending to confirm the view I had otherwise formed on the basis of Dr Cleare’s evidence that there is no magic in “below pH 2.7 or above pH 7”. The skilled man adjusts pH as he thinks appropriate to trade off between yield and time to completion. In this connection, a great deal was made of the existence of other process variables all of which might be adjusted. This is obviously correct, but pH is of particular importance. In ’012, only two variables are mentioned: temperature, which is never changed, and pH.

Anticipation by a single example in Kidani’s oxaliplatin disclosure, US patent 4,169,846 (‘Kidani’) published 2 October 1979, is alleged. The allegation is by way of a ‘squeeze’ on construction. Mayne say that if the patent is construed widely enough to catch an adjustment of pH to within pH 4-5 by the addition of an oxalate salt (ie oxalic acid to which an alkali has been previously added) the synthesis of oxaliplatin described in this patent anticipates, or certainly renders obvious, the patent in suit. As I have construed the claim, what is disclosed by Kidani does not anticipate the claim. But if I am wrong in the construction I place upon the claim, this allegation becomes relevant.

Kidani is generally concerned with synthesis of a number of complexes from the diaquo DACH Pt(II) species. Example 4 is as follows:

This is oxaliplatin. Example 3 is the preparation of an analogue, the cis isomer. The preparation proceeds from the dichloro material

Dissolve 3g of the dichloro material in 500 ml of water, boiling to complete dissolution;

Add 2.6 g AgNO₃ (i.e. twice molar);

Stir for 3 hours in the dark;

After cooling repeatedly filter using a 5C filter paper supplied by Toyo Roshi Kaisha Ltd until filtrate became transparent;

Concentrate under reduced pressure to 100 ml;

Add 1.3 g of potassium oxalate and allow mixture to stand for 8 hours.

Concentrate under reduced pressure to obtain white precipitate;

Recrystallise from aqueous solution.

There was no evidence to suggest that this did not work for the cis compound, as stated. There was, however, evidence to suggest that it did not work without adjustment of the conditions for the trans isomer. Mayne performed an experiment, and a repeat, and obtained oxaliplatin. Debiopharm did not perform any experiment in chief, but sought to rely upon a repetition performed in 2002 at the ETH in Zurich by Professor Ward at the instance of Debiopharm’s patent department. There was a later experiment in reply, performed by a Dr Marinetti at a laboratory of the CNRS in Paris. Professor Ward gave evidence: Dr Marinetti did not.

As I assess the disproportionate amount of evidence adduced on the issue of this experiment, the only real problem is caused by the low solubility in water of the trans dichloro material compared with that of the cis dichloro material. It is not possible to dissolve 3g of the former in 500ml of water, even with prolonged boiling. Professor Ward’s solution was to add more water. The Mayne solution was to ignore the problem, and to slurry the undissolved material in the silver nitrate solution (after a short, 10 minute, boil), knowing that as the reaction with the chloride ion took place to form the completely insoluble silver chloride more of the dichloro material would necessarily dissolve.

It was an odd feature of the Mayne protocol that there was the 10-minute boil under reflux after the addition of the dichloro material to the water which can have achieved little or nothing in producing dissolution. As far as I can see, there was no point in this boil at all from the technical point of view, and I formed the view that it was essentially colourable (Footnote: ²). Somebody, somewhere, must have formed the view that the experiment would look more like a repetition if it contained a boiling step, no matter how pointless. I do not know who did this because Dr Cleare did not design the protocol and its author was not called to give evidence. All that needed to be said was that ‘slurrying is a well known technique of dealing with sparingly soluble materials, and when we decided to slurry no effort to achieve dissolution was necessary’.

I am satisfied that this is the only relevant departure from the express protocol of Example 3, and that it is entirely justified by the problems caused by the relative insolubility of the trans-l material. I find that slurrying is a well-known, common general knowledge technique for carrying out reactions providing that the further dissolution of the relatively insoluble material can be encouraged by removing the relevant ion from solution, as happens here with the removal of the chloride ion by the precipitation of the silver chloride. Dr Cleare explained it very well from the perspective of a practical chemist (transcript 213 ff):

Nothing can be set against this to suggest that slurrying is not a common general knowledge technique.

I should add that the question of slurrying aside, Professor Davies made three criticisms of Mayne’s experimental protocol in paragraphs 27 ff of his report. These were (i) molar concentration of silver nitrate incorrect; (ii) temperature of solution at the time the silver nitrate was added; and (iii) corrections for use of potassium oxalate monohydrate. Under cross-examination, it was clear (i) that the molar concentration of silver nitrate was correct to 2 significant figures; (ii) the temperature of the solution at the stage of addition of the silver nitrate was within the common general knowledge and (iii) it was obvious that a molar quantity of potassium oxalate was to be used, which needed to be corrected from the value in Kidani if anhydrous material was not available and a hydrate was being used instead. These objections had no substance.

The failure of Professor Ward’s experiment is difficult to explain, because it has not been repeated. It has two departures from the experiment described in the patent. The first is the use of a great excess of water, because Professor Ward dissolved the trans-l material. The second was the use of celite rather than a paper filter. Celite was used also by Dr Marinetti. Professor Ward could not find a paper that worked. Celite is a purified diatomaceous earth. Some grades contain calcium salts: calcium oxalate is highly insoluble and so any reaction with oxalate ion would strip the oxalate ion out. Dr Cleare tentatively identified the compound precipitated in Dr Marinetti’s experiment, which also used celite as the filter medium, as calcium oxalate and so produced a rationale for the failure of the experiments involving celite filtered material which I thought was convincing. Of course, lack of success with celite does not have to be explained: it is a failed departure from the experiment as described, and the cure is to select a filter paper that will catch the silver chloride, something Professor Ward said was sometimes difficult. Mayne obtained the paper described in the Example, and the paper Mayne obtained worked. Thus, although both Professor Ward and Dr Marinetti had tried filter papers that did not work, that does not affect the problem unless I am satisfied that the celite they used was in all respects (including possible reactivity) the equivalent of paper. This was not done.

I should finally add that there is no evidence in the experiments carried out by Mayne by way of workup, which I ordered to be disclosed by an interlocutory judgment of 10 February 2006, of any failure . Although some of the notebook pages are not clear, oxaliplatin seems to have been produced and there is certainly no indication of any attempt to zero in on a narrow set of acceptable conditions, or anything of that sort.

Now, Mayne’s experiment showed the effect of the addition of the potassium oxalate was to bring the pH of the reaction mixture from less than pH 3 to pH 4.5 consistently. Accordingly, if I am wrong on construction of the claim, and the claim would cover the addition of potassium oxalate as well as oxalic acid followed by potassium hydroxide, the impact of the experiment has to be considered.

Dr Cleare was not challenged on his suggestion that the general scheme of the reaction (convert dichloro to diaquo and react with carboxylic acid) was well known. It is analogous to the preparation of both cisplatin and carboplatin, which differ in the dichloro-Pt(II) complex employed, but for both of which the reaction with silver nitrate is employed to produce the relevant diaquo complex.

First, I am satisfied that the document discloses oxaliplatin. Of that, there is no doubt. Second, I am satisfied on the evidence that Example 4 is enabling, in the sense that it discloses a method of synthesis of oxaliplatin that will enable a skilled person wishing to achieve success rather than failure to make oxaliplatin without an undue expenditure of time and effort and without undue experimentation. The classic statement of the test for insufficiency stated in Valensi v BRC [1973] RPC 337 at 377 (CA) is apposite here:

Aldous J’s statement of the same principle in the context of the 1977 Act was approved by the Court of Appeal in Mentor v Hollister [1993] RPC 7 at 14:

Finally, the EPO Guidelines to Examination (C-IV para 7.3a):

In this case, the answer to the question whether the Kidani disclosure is enabling or not is perhaps rendered rather clearer by considering that at the priority date corresponding reactions had been carried out many times (I have quoted Dr Cleare’s evidence on this under cross-examination above) and the scheme was well known. This passage in Kidani is relied on in other Debiopharm/Sanofi/Tanaka publications as setting out a synthetic route to oxaliplatin: paragraph [0002] of ’454 itself (oddly), and paragraph [0002] of ’331 are two examples and also the two patents in suit in this action. Precisely the same words are used in another patent granted to Tanaka (EP 0 567 438) at page 11 lines 35 to 40 to provide a synthetic route to oxaliplatin for the purposes of a comparative example. It would be surprising if these disclosures were non-enabling. The only indication in the case is Professor Ward’s experiment. This is open to the basic objection that it uses a filter material which may be reactive in a relevant way (celite) and that once failure had been obtained on the first repetition it seems as though no attempt at all was made to get it to work. Indeed, Professor Ward’s second attempt was identical to his first.

I conclude that the evidence favours the conclusion that Kidani is an enabling disclosure. Thus, on the wider construction of the claim that I have rejected, claims 1 and 2 are anticipated.

This is an article concerned with certain Pt(IV) complexes with anti-tumour activity. One of them (trans-l-DACH trans-dihydroxo(oxalato)platinum(IV)) may be synthesised by a route in which oxaliplatin is an intermediate, and the allegation is that the route to oxaliplatin described in the paper anticipates.

Dr Cleare describes the route as follows ([77] of his first report):

The variations on the ‘standard’ method are the use of the tetraiodoplatinate in place of the tetrachloroplatinate and the use of silver sulphate rather than silver nitrate. I was told that silver iodide is more insoluble than silver chloride. There is no challenge to Dr Cleare’s summary and I accept it. Again the crucial feature of the synthesis is the reaction of the diaquo species with the oxalate salt (sodium oxalate) rather than oxalic acid. Accordingly, the disclosure will have the same ‘squeezing’ effect on the construction of the claim as Kidani if the pH of the solution after adding the sodium oxalate solution is within the prescribed ranges. This Mayne sought to show by experiments.

The starting point for the experiments performed by Mayne was the sulfato complex, trans-l-DACH sulfatoplatinum(II)H₂O. The material used for the experiment in chief was admitted by Debiopharm following the repetition to be the sulfato material. There is no substantial criticism of these experiments, a rather faint suggestion that there may have been some impurity being met by the fact that no attempt to identify such an impurity had been undertaken. All the experiments bar two showed a pH within the limits of claim 2 on addition of the oxalate, the two in question having stood for an hour before the addition (Dr Cleare explained that the diaquo material deprotonates on standing, thus decreasing pH). Of the work-up experiments, those called A and B2 are the ones where the range of claim 2 was probably not met—all the experiments met claim 1. B1, B3, B4 did meet both ranges. A and B2 are the two experiments where the diaquo complex solution was allowed to stand before addition of the oxalate salt.

Had the claim covered addition of the oxalate salt alone, without further addition of alkali solution, Khokhar would have anticipated both claims.

Kidani and Khokhar disclose syntheses in which the oxalate salt added to the diaquo species is an alkali metal salt (potassium and sodium respectively). I am asked to find that the substitution of the ammonium oxalate was obvious in the light of these citations.

Of course, at the very basic level, once use of the oxalate of a monovalent metal is published, it is difficult to see what could be inventive in the use of ammonium oxalate. Dr Cleare was quite clear about this (transcript 180 ff). Use of ammonium oxalate might well have disadvantages, because of the tendency of the ammonium ions to deprotonate at higher pH to give ammonia which will in turn form coordination complexes with the platinum ions and lead to impurities. It is certainly not the cation of first choice. Nobody suggested (they could not suggest) that it would not work, or anything like that. It is obvious as a second choice in the light of Kidani and Khokhar.

This patent, which claims priority from 25 February 1998, is apparently directed to the stabilisation of solutions of oxaliplatin. Claim 1 is to

There is a corresponding method claim (claim 18) which calls for the addition of an effectively stabilising quantity of oxalic acid or an alkali metal salt of oxalic acid to a solution of oxaliplatin and a process claim (20) which adds nothing.

Mayne sought a declaration of non-infringement in respect of this patent and Debiopharm counterclaimed for infringement. Both sides adduced experimental evidence, the repeats of the Debiopharm experiments taking place during the early part of the trial. The results of the repetitions made it impossible for Debiopharm to maintain an allegation of infringement on the basis they had adopted, and Mayne’s entitlement to a declaration of non-infringement in suitable terms is conceded. The only issue is validity.

Validity is challenged on the basis that the claims are impossible to construe and the specification is accordingly insufficient in failing to define the invention clearly enough and completely enough for it to be performed by a man skilled in the art. To the extent that they cover the idea of stabilising oxaliplatin by the addition of oxalate ion to reduce the probability of decomposition of oxaliplatin by departure of the oxalate ion and consequent formation of impurities, they are said to be obvious. Claims 5, 12, 13, 14 and 18 are alleged to be independently valid on the footing that claim 1 fails for anticipation.

I should briefly consider the notional addressee of this patent. Obviously the patent is concerned with formulation of what was by 1998 an old active ingredient, which was in clinical trials. It is therefore directed at a formulation chemist, who must be assumed to call upon the services of a process chemist and a regulatory expert as needed. Dr Cleare, who was a research chemist, a process chemist and had assisted formulators gave the evidence of Mayne on this patent, while Professor Allwood gave unchallenged evidence for Debiopharm, while Professor Davies was cross-examined on the questions falling within his expertise.

The main problem is to work out what the phrase ‘stable’ means in its context in the claim, and to decide whether ‘an effective stabilising amount of a buffering agent selected from oxalic acid or an alkali metal salt thereof’ requires that the effect of the buffering agent is to introduce oxalate ion to a level in excess of that which would be expected in oxaliplatin solution alone. Put another way, the question is whether all stable solutions according to claim 1 must be made according to claim 20.

[0014] sets out the context of the disclosure. Oxaliplatin is administered by injection, and at the preclinical and clinical test stage is provided by lyophilized powder reconstituted in water or glucose solution and diluted. The need is to provide oxaliplatin solutions rather than lyophilized powder preparations of the drug, so as to avoid an expensive lyophilization step, errors in reconstituting the pharmaceutical including use of the wrong solvent and incomplete dissolution and the increased risk of microbial contamination. [0015] states that oxaliplatin in solution degrades to a number of identified impurities: diaquo-DACH-Pt(II) (the diaquo material I have discussed extensively above), the dimer also discussed above, and a Pt(IV) species which is either the same as or an isomer of the trans-l-DACH trans-dihydroxo(oxalato)platinum(IV) which is the object of the preparation described in the Khokhar article above. [0015] proceeds to identify the object of the invention in the following terms:

It is not necessary to consider the very extensive data given in the patent relating to stability to autoclaving of the various solutions made according to the invention, but to turn to the stability studies in [0078] ff and the accompanying tables. It is of great importance in considering this material also to note the comparative example 18, which is buffer-free and whose results are set out in Tables 8 and 9.

As I have indicated above, the reaction for the production of oxaliplatin from the diaquo DACH material is like this:

What the tables record is the level of diaquo DACH in solution at the outset of the test, i.e. shortly after dissolution. This must reflect a dissociation of the oxaliplatin, assuming it is pure in the first place:

Dr Cleare says that any chemist with an understanding of the basic chemistry of the platinum complexes will appreciate that these near equilibria represent the degradation of oxaliplatin in water. While oxaliplatin and the diaquo-DACH remain in solution, the equilibrium moves slowly to the left as the diaquo-DACH forms the hydroxo species (not shown) and ultimately the dimer, which is insoluble and will come out of solution. It seems to me that the fact that this is a very slow progressive degradation is supported by tables 4, 5, 6, 7, 8 and 9 of the patent. In each of these tables, the initial concentration of diaquo-DACH and dimer appear to depend upon the initial concentration of oxalate ion, remaining reasonably steady thereafter (some concentrations of diaquo-DACH and dimer even decrease slightly with time). It is to be noted that the specified impurities remain constant even in the comparative example (table 8) while the unspecified impurities seem no worse than in some of the other examples (2, 3, 14, 15 and 16) although the conditions are not really comparable. Finally, it is difficult to see any helpful contrast between comparative Example 18 (table 8) and Examples 1 and 8 (tables 5) which are allegedly according to the invention, save possibly in respect of the unspecified impurities.

Dr Cleare takes these examples and says that they provide no guidance as to what ‘stable’ can mean.

The last clue as to what is meant by stable is in [0010] which states that the solution disclosed in WO 96/04904 (Ibrahim), a Debiopharm patent, which consists of a solution of oxaliplatin in water at a concentration in the range of 1 to 5 mg/ml and pH in the range of 4.5-6 is stable. This document adopts the similar, but possibly more stringent, definition of stable as the patent, of stability over 3 to 5 years at room temperature or refrigerator temperature (see page 4).

In my judgment, the only working definition for ‘stable’ that this document provides is the one in [0016], i.e. pharmaceutically stable over prolonged periods of storage of 2 years or more. There is no basis in the remainder of the document for any other view. It follows that stability is achievable without use of the invention, and Ibrahim suggests that many aqueous solutions are stable. (Footnote: ³) Unfortunately none of the Examples in the patent serve to demonstrate anything other than that (1) impurity level is established at the outset and (2) it does not thereafter change, or changes very slowly indeed.

The reason I cannot take any other view is that Professor Davies accepted that Example 18 was stable but that Example 8 was not. In other words, you cannot spell out of the examples any better view of what the patent is concerned with. I reject the contention, in the words of Mr Waugh QC’s closing submissions, that

This fails as a definition for two reasons. First, as paragraph [0016] makes clear, the patent is concerned with pharmaceutical stability over time. If the pharmaceutical is acceptable at the beginning of the period of two years, and acceptable at the end of two years, it is stable as that term appears to be used in the patent. The precise impurity level does not matter. Second, the prior art is said to include stable compositions: see the reference to Ibrahim in paragraph [0010]. If the patent had been concerned with formulations as stable as the prior art but with lower impurity levels, it would have said so, but the whole emphasis is on stability.

I have wondered whether paragraphs [0073] and [0074] of the specification, where the very poor stability of oxaliplatin in various standard aqueous buffers is discussed, has any bearing on the meaning of ‘stable’ in this context. Does it just mean ‘better than other well known buffers’? Such an approach could be consistent with [0015]

I turn to the phrase “effective stabilising amount…”. Implicit in these words is a comparison, and one is confronted with the same problem. In fact, the addition of the oxalate ion, whether in the form of oxalic acid or in the form of an alkali metal oxalate, does not seem to contribute much, if at all, to stability, but it does reduce initial levels of the diaquo-DACH material and of the dimer. The initial level remains, as I have said, more or less constant in the majority of the examples (Example 8 is an exception).

When one is confronted with a claim which requires ‘an effective stabilising amount’ of a material, it must be possible to design a test which can answer the question ‘Have I used such an amount or not?’. There will always be problems on the edges of claim, but it should, in general, be possible to know what the test is. If one cannot identify the test on the basis of the disclosure, then I think that the disclosure is insufficient (see Milliken v Walk-Off Mats [1996] FSR 292 (Pinocchio units) and Scanvaegt v Pelcombe [1998] FSR 786: see also the Guidelines C-II paragraph 4.10:

We are dealing here with parameters (stabilising quantity as opposed to insufficient quantity) and the Guidelines are characteristically trenchant on this subject (C-III para 4.7a):

If to my question ‘Have I used an effective stabilising amount or not?’ the specification’s answer is ‘You don’t have to add any at all’, which is the only possible answer in the light of Ibrahim and its acknowledgment in [0010], in my judgment it follows that the invention the subject of this claim is insufficiently disclosed, and the same goes for the other claim that contains no express limitations on quantities (claim 18). I prefer to interpret claims 5, 12 as dependent on 5, 13 and 14 simply as claims to aqueous solutions of oxaliplatin containing the specified concentration of oxaliplatin and/or molar concentration of oxalate ion. These are not insufficient on my approach.

Finally, I should return to the question whether the claim requires a distinct addition of oxalic acid or alkali metal oxalate. In my judgment it plainly does. If it does not, the patent is hopelessly insufficient, and not only because no clear guidance is given to determine what is or is not an ‘effective stabilising amount’ if it can just appear when the active ingredient dissolves. Debiopharm’s experiments show clearly that on dissolution of pure oxaliplatin in water sufficient dissociation takes place that the concentration of free oxalic acid is about 6.7 × 10^-5M, which is within the ranges claimed in claims 5(a) to 5(c). If this is an ‘effective stabilising amount’ the claim must be talking about added oxalate ion to avoid anticipation by Ibrahim or a further insufficiency.

An attack is made on the basis of the common general knowledge and on the basis of two publications, Ibrahim and US 5,455,270 (Kaplan). I disregard Ibrahim, which is relevant only if the claim covers oxaliplatin solutions to which no additional oxalate ion has been added, whether in the form of oxalic acid or of the alkali metal salt. For the reasons I have given above, I do not think this is a tenable construction. Kaplan is relied on by Mayne to strengthen an argument of obviousness in the light of the common general knowledge.

The argument on common general knowledge proceeds as follows. The addressee of the specification includes a formulation chemist and a process chemist. Assuming that they are confronted with a long-term stability problem (it is not likely, it seems, that they will be) what do they do? First, Professor Davies accepted that he would expect the formulation chemist to look to the process chemist for insight into degradation mechanisms. Second, he agreed that at the date an analysis would follow broadly the route outlined by Dr Cleare in paragraphs 143 to 149 of his first report, which essentially sets out the scheme I have outlined above for degradation via the dimer. He also agreed that cisplatin (which, for the reasons I have outlined above, I consider would be known to the team) does a similar thing but in hours rather than years.

The mechanism for cisplatin is as follows:

Dr Cleare said that the solution in the case of cisplatin was to dissolve it in a saline solution, where the presence of chloride ions would shift the equilibrium to the left. Given the concentration of cisplatin used, physiological saline was enough. The chloride ion is the ‘leaving group’ in this equilibrium, and the general principle is known as Le Chatelier’s principle. Using the same logic with oxaliplatin, one would add oxalic acid or an oxalate, since the oxalate is the ‘leaving group’. This would tend to shift the equilibrium in favour of oxaliplatin and reduce the concentration of the diaquo species and so the rate of formation of the hydroxo species and, ultimately, the dimer. In other words, the patent is merely an application of a well-known principle.

Dr Cleare said that the toxicity of oxalic acid was irrelevant: one would be adding it in concentrations comparable with those in which it appeared on dissolution of oxaliplatin. Indeed, at the same concentration it should halve the concentration of diaquo species. Nonetheless, this would not matter unless the concentration of diaquo species determines the rate of formation of the dimer. If it does, then the rate of dimer formation will be affected by the addition of small quantities of oxalic acid, as the patent suggests it is. The contention advanced by Mr Waugh QC in paragraph 203 of his concluding skeleton, that substantial amounts of oxalic acid are required, cannot be justified on any application of Le Chatelier’s principle. The degradation reaction is so slow it is a quasi-equilibrium, and the matter can be approached in that way.

Professor Davies took a completely different view. His view was predicated on the assumption that oxaliplatin is stable in water (reply report paragraph 54 concluding sentence and paragraphs 57, 59 and 62). This is tantamount to saying the invention has no point. This does not seem to me to be relevant, and in any event his cross-examination covered the matter from about page 667 onwards. He accepted that the common general knowledge view of the stabilisation of cisplatin was that it was an application of Le Chatelier’s principle, and in the end I found that Dr Cleare’s view of the patent in suit was not as tainted with hindsight as Professor Davies suggested. On the whole I preferred Dr Cleare’s evidence. I find this patent invalid in the light of the common general knowledge.

Kaplan is said to show one final example of the application of this principle, this time with carboplatin, by use of the leaving group of carboplatin (1,1 cyclobutanedicarboxylic acid, or CBDCA) or a salt to stabilise the active ingredient in solution. This is another application of Le Chatelier’s principle so far as Dr Cleare is concerned. For Professor Davies, it was an example of prevention of nucleation of the degradation product. I think there is little doubt that Kaplan himself did not see his invention as an application of Le Chatelier’s principle, but the argument by analogy remains a strong one: if CBDCA works for carboplatin, why doesn’t oxalic acid work for oxaliplatin?

Kaplan was referred to in two emails of the 28 and 29 March 1996. It is said that ‘a small amount of sodium oxalate could act as a buffer and keep the equilibrium to the left [my emphasis]. The same idea came out of a discussion in Nick’s office a few days ago.’ This is Le Chatelier-speak, if I may put it like that. I do not pay any regard to this, because I am satisfied on Dr Cleare’s evidence that the patent in suit is an application of the principle, and because on its own the email does not exclude the possibility that there was an inventive insight into what was happening in Kaplan. As it is, though, the patent is on the evidence more or less superfluous. If the problem with which it purports to deal were a real problem the principle underlying the solution that it advances is an obvious one. It is, however, claimed in such a way that the claims are insufficiently defined to the extent I have described. It must be revoked.