Novartis Ag v IVAX Pharmaceuticals UK Ltd

This is an action for infringement of two patents relating to the formulation of cyclosporin, which is a pharmaceutical of particular utility in producing immunosuppression in recipients of allogenic organ transplants. UK Patent 2,222,770 (“770” or “the Microemulsion Patent”) claims priority from September 1988. The second patent, UK Patent 2,380,674, claims priority from March 1998 and has been referred to as either “674” or “the Second Component Patent”.

The Defendant (“Ivax”) wishes to launch a cyclosporin formulation in the UK under the name “Equoral”. Both infringement and validity of both patents are in issue.

Before turning to the disclosure and the claims of this patent, I should consider the addressee of the specification. The patent is plainly directed to a pharmaceutical formulator wishing to provide a formulation of cyclosporin. By 1988, cyclosporin was a well-understood pharmaceutical, and in 1988 it was known to present difficulties in formulation. The patentees, the well-known Swiss pharmaceutical company Sandoz, were marketing a formulation of cyclosporin that attempted to overcome its almost complete insolubility in water. Pharmaceuticals that are insoluble in water may well be soluble to a greater extent in oils, and this fact was exploited by the existing formulation, “Sandimmun”. Because cyclosporin therapy will last potentially for the whole of the recipient’s life, it is desirable to formulate the drug so as to be as conveniently administered as possible. Oral administration is desirable, and Sandimmun was an oral formulation. It was based upon olive oil, a surfactant (Labrafil) and ethanol (alcohol). Sandimmun was what is called a pre-concentrate: when diluted with an aqueous phase (as for example the aqueous contents of the gut) it forms an oil-in-water (o/w) emulsion.

An emulsion is one example of a colloidal system. In very general terms, such systems consist of two “phases”: a discontinuous or dispersed phase and a dispersing or continuous phase. An oil-in-water emulsion comprises very small particles of oil dispersed through a continuous aqueous medium. Such emulsions are generally cloudy or milky in appearance and are generally stabilised, at least to some extent, so that they do not separate out, or at least separate out sufficiently slowly to make them useful. The stabilisation is achieved by the use of surfactants, mixtures of surfactants, or a surfactant and a co-surfactant. Surfactants are molecules which have both a hydrophilic and a lipophilic part, and form a monolayer between the oily phase and the aqueous phase. Such molecules are called amphiphilic. Co-surfactants are also amphiphilic molecules, but, as I understand the evidence, typically rather smaller than a surfactant: for example, ethanol, propanol, butanol and pentanol may be co-surfactants. These molecules also are amphiphilic, being soluble in water and in some lipids.

The relative proportions of the different phases present in an emulsion are commonly shown on ternary plots (Figures I and II of ’770 are examples) in which the relative concentrations yielding the desired emulsion are shown as areas within a triangle whose apices represent 100% of each component. In Figure I, for example, components 1.1, 2 and 3 are respectively the hydrophilic, lipophilic and surfactant components. Region A represents a region where the mixture works, and Region B a preferred region.

The cloudy appearance of an oil-in-water emulsion, and of other emulsions such as egg-yolk in olive oil, is caused by the size of the particles of the dispersed phase being sufficiently large to disperse light. As I understand the evidence, the droplet size of the dispersed phase in typical emulsions ranges from about 0.2 to 10 microns diameter. (This case is vexed by using three units for sizes in this general region: nanometres or nm = 10^-3 microns = 10 angstrom units = 10^-9 metres. The limit for optical visibility is about one wavelength of light, which at the blue end is about 400 nm, or 4,000 angstroms. The red end is about 700 nm. Any emulsion contains a distribution of particle sizes, and will tend to look bluish in reflected light (and reddish in transmitted light) as the centre of the distribution moves closer to sizes comparable with the wavelength of blue light.)

The scattering of light by colloids forms the basis for instruments that are capable of measuring the particle size distribution of the colloids. These instruments, in existence and standard pieces of laboratory equipment at the priority date, have various limitations to which I shall return in considering the experiments in this case. For present purposes, it is only necessary to note that the measurement of particle size distribution in colloidal materials was well-understood at the priority date.

At the priority date, cyclosporin was, as I have indicated, a generally accepted drug for immunosuppression. It was known to be nephrotoxic (liver-damaging) and the “therapeutic window” between the minimum effective dose and a toxic dose was not wide.

Chemically, the drug is unusual, being an 11-unit cyclic peptide. In 1988, it was known to be very hydrophobic. It could not be delivered in tablet form, and there was no commercial formulation for topical administration.

It is accordingly against the background of Sandimmun and the salient properties of the active ingredient summarised above that the disclosure of the patent in suit is to be assessed.

After describing cyclosporin itself and its actual and potential applicability, the specification identifies (page 5, final paragraph) the problems with formulations, such as Sandimmun, containing both ethanol and olive oil:

“First, the necessity to use oils or oil based carriers may lend the preparations an unpleasant taste or otherwise reduce palatability, in particular for the purposes of long-term therapy. These effects can be masked by presentation in gelatine capsule form. However, in order to maintain the cyclosporin in solution, the ethanol content has to be kept high. Evaporation of the ethanol, e.g. from capsules or from other forms, e.g. when opened, results in the development of a cyclosporin precipitate. Where such compositions are presented in e.g. soft gelatine encapsulated form, this particular difficulty necessitates packaging of the encapsulated product in an air-tight compartment, for example an air-tight blister or aluminium-foil blister-package. This in turn renders the product both bulky and more expensive to produce. The storage characteristics of formulations as aforesaid are far from ideal.”

Thus the first problem with existing formulations that is identified is the problem caused by the presence of alcohol in the formulation. The second is variation between groups of patients and, even for a single patient, of bioavailability of the active ingredient:

“Bioavailability levels achieved using existing oral cyclosporin dosage systems are also low and exhibit wide variation between individuals, individual patient types and even for single individuals at different times during the course of therapy. Thus reports in the literature indicate that currently available therapy employing the commercially available Ciclosporin drink solution provides an average absolute bioavailability of ca. 30% only, with marked variation between individual groups, e.g. between liver (relatively low bioavailability) and bone-marrow (relatively high bioavailability) transplant recipients. Reported variation in bioavailability between subjects has varied from anything between one or a few percent for some patients to as much as 90% or more for others. And as already noted, marked change in bioavailability for individuals with time is frequently observed.”

On page 7 of the specification, the proposals already made to meet these problems are discussed, and it is observed that the overriding difficulty is the consequence of the inherent insolubility of the cyclosporins in aqueous media. The specification refers to the additional need for a topical delivery system for the drug, and sets out its solution to the problem at the foot of page 7:

“By the present invention there are provided novel cyclosporin galenic formulations in the form of a micro-emulsion pre-concentrate and/or based on the use of particular solvent media as hereinafter defined, which meet or substantially reduce difficulties in cyclosporin, e.g. Ciclosplorin, therapy hitherto encountered in the art. In particular it has been found that the compositions of the invention permit the preparation of solid, semi-solid and liquid compositions containing a cyclosporin in sufficiently high concentration to permit, e.g. convenient oral administration, while at the same time achieving improved efficacy, e.g. in terms of bioavailability characteristics.”

It is said (foot of page 8) that the invention of the patent in suit enables the preparation of compositions which are non-alkanol based, for example, free or substantially free of ethanol. The first aspect of the invention is stated on page 9:

“In a first aspect, the present invention specifically provides pharmaceutical compositions comprising a cyclosporin as active ingredient, which compositions are in the form of an “oil-in-water microemulsion pre-concentrate”.”

The key concept is thus a mixture (pre-concentrate) which will form a microemulsion when added to an aqueous medium. The passage which follows, which defines the term “oil-in-water microemulsion pre-concentrate”, is of particular importance:

“By the term “oil-in-water microemulsion pre-concentrate” as used herein is meant a system capable on contacting with, e.g. addition to, water of providing an oil-in-water microemulsion. The term microemulsion as used herein is used in its conventionally accepted sense as a non-opaque or substantially non-opaque colloidal dispersion comprising water and organic components including hydrophobic (lipophilic) organic components. Microemulsions are identifiable as possessing one or more of the following characteristics. They are formed spontaneously or substantially spontaneously when their components are brought into contact, that is without substantial energy supply, e.g. in the absence of heating or the use of high shear equipment or other substantial agitation. They exhibit thermodynamic stability. They are monophasic. They are substantially non-opaque, i.e. are transparent or opalescent when viewed by optical microscopic means. In their undisturbed state they are optically isotropic, though an anisotropic structure may be observable using e.g. x-ray technique.

Microemulsions comprise a dispersed or particulate (droplet) phase, the particles of which are of a size less than 2,000 A, hence their optical transparency. The particles of a microemulsion may be spherical, though other structures are feasible, e.g. liquid crystals with lamellar, hexagonal or isotropic symmetries. Generally, microemulsions comprise droplets or particles having a maximum dimension (e.g. diameter) of less than 1,500 A, e.g. typically from 100 to 1,000 A.

[For further discussion of the characteristics of microemulsions see, e.g. Rosof . . . ; Friberg . . . ; and Müller . . . . ]

From the foregoing it will be understood that the “oil-in-water microemulsion pre-concentrates” of the invention are galenic systems comprising a cyclosporin as active ingredient capable of forming an oil-in-water microemulsion, spontaneously or substantially spontaneously on contact with water alone.”

A clause repeating claim 1 of the patent, which is the only claim with which I am concerned, then follows. I set claim 1 out here for convenience:

‘1. A pharmaceutical composition comprising

a cyclosporin as active ingredient,

1)

a hydrophilic phase,

2)

a lipophilic phase, and

3)

a surfactant,

which composition is an “oil-in-water microemulsion pre-concentrate.”’

A very substantial amount of text follows. Its function is mainly to give examples of suitable materials, but there is also explanatory material dispersed throughout. The general term used for a composition falling within the claim is ‘a composition defined under (A)’ or ‘in accordance with (A)’. Important for present purposes is the explanation of the terms of the claim:

The main issues, surprisingly enough, are the meanings of the words ‘microemulsion pre-concentrate’ and ‘lipophilic phase’. To put them into context, it is necessary to say a little about Ivax’s proposed composition, Equoral.

Ivax want Equoral to be available in the UK as a 100mg/ml oral solution and 25mg, 50mg and 100mg soft gelatine capsules. The oral solution is intended for dispersion in a drink, and the capsules to be swallowed and to dissolve in the stomach where the active ingredient will be released. The (very) approximate proportions (w/w) of the various components in the solution are as follows:

It was accepted that one principal feature of all microemulsions, the one that makes a microemulsion transparent or slightly cloudy, rather than milky, is the fact that substantially all the dispersed phase particles are less than 200nm (2000Å, 0.2µ) in diameter. The results of dispersing this mixture in water have been the object of a considerable amount of investigation. The proper interpretation of the results has been the subject of much dispute. However, taking matters at their most favourable to the claimants, the results are as follows. As one would necessarily expect, the sizes of the particles in the dispersion are distributed over a substantial range. Taking a cut-off point for particle size of 2000Å (200nm, 0.2µ), the vast majority of particles by number falls below this limit. However (and this is important) less than 14% of the active ingredient is to be found in all the particles of size less than 2000Å: more than 86% of the burden of active ingredient is carried by particles having size >2000Å.

The distribution of particle sizes is somewhat complex. As I understand the evidence of Professor Schurtenberger, it is necessary to centrifuge the dispersion of Equoral in acid or water in order to measure the smaller sizes in the distribution at all. Otherwise the big particles tend to hide the little ones. The results after centrifugation suggest a bimodal particle size distribution below 200nm, that is, there are two peaks in the distribution. Above 200nm, the distribution is approximately normal (i.e. bell-shaped). The observation levelled against Professor Schurtenberger’s procedure is that it appears that the distribution of particle sizes among the ranges covered by all techniques varies Pbetween centrifuged and uncentrifuged samples, suggesting that some of the small particles present in the centrifuged samples are artefacts of the centrifugation. He was not cross-examined, and I have anyway to decide this issue on the basis of techniques available in 1988.

Mr Alexander QC suggested that there was evidence that 14% active ingredient in the totality of particles of size <200nm was present in the dispersion. Mr Thorley QC observed that this figure had been calculated on the basis of the centrifuged samples, and accordingly assumes that no particles had been affected by centrifugation. For present purposes, I am prepared to accept the highest figure of 14%, or one-seventh. There is obviously force in Mr Thorley’s submission—Professor Attwood was emphatic and Professor Lawrence accepted that fragmentation of the Equoral particles on centrifugation was a possibility—but the fair conclusion is that in the particles of size < 200nm the experiments show that the quantity of active ingredient is certainly not more than 14% and is likely to be rather less than this.

The Equoral dispersion is opaque, even if one waits for a period longer than one would expect the dispersion to reside in the stomach. There was some suggestion that it was not optically isotropic (i.e. it displayed birefringence) but Ivax’s experimental results, although demonstrating a degree of birefringence both in chief and at the repeat, were barely consistent. There is some evidence of birefringence. Birefringence would be indicative of the presence of liquid crystals, but I was not satisfied that there were, in fact, substantial quantities of liquid crystals present, and certainly not enough to characterise the aqueous dispersion as being made up of liquid crystals.

Is this what the skilled person would recognise in 1988 as a microemulsion? The question has two aspects. First, was the use of microemulsions the sort of thing that can fairly be regarded as part of the common general knowledge of a pharmaceutical formulator? There was, as I have indicated, a real dispute on this question. Second, following the common general knowledge so far as available, and the explicit guidance given in the specification, what would the skilled person recognise as a microemulsion?

I have set out in summary the common general knowledge in respect of emulsions above. The basic knowledge possessed in 1988 by any skilled person who did know about microemulsions was not controversial: such emulsions are monophasic, optically isotropic, mixtures of two immiscible liquids. They are thermodynamically stable, and form spontaneously. They have a particle size of the dispersed phase < 200nm, and the particle size distribution for a particular composition is reproducible. They are transparent or translucent.

Professor Attwood, a distinguished scientist, undoubtedly knew about microemulsions and taught them to his undergraduate students as a method of formulating poorly soluble active ingredients. I was shown the contents of undergraduate lecture courses, final year projects and examination questions making this quite clear. Microemulsions were also described in the paper called ‘Bhargava’ at the trial. However, the impression I received is that this was more or less the sum total (with the extract from Remington) of the available publications. I was very struck by his description of his response to Neoral, the microemulsion formulation produced by Novartis:

‘A. I would not take it that far. I think my attitude at the time was, "Thank goodness somebody has used a microemulsion formulation at last" because I worked on it in 1972. I thought they were excellent systems. I was somewhat surprised that nobody had picked it up before. Now somebody has picked it up.’

Professor Lawrence did not consider that the use of microemulsions was common general knowledge at the date. After all, she observed that there were, in fact, no microemulsion formulations of other drugs at the date. On the whole, I think that Professor Attwood’s view was that formulators should have known about microemulsions, and that he had done his best to spread the word. But I do not think that this makes their use common general knowledge. It is well settled that the common general knowledge is knowledge that must be attributed to the skilled person, without which the latter may be taken not to be skilled. To it must be added any knowledge that every skilled person should be taken to acquire before he embarks on the problem to which the patent provides the solution, as, for example, the relevant properties of cyclosporin if the problem is to formulate cyclosporin. If microemulsions are common general knowledge, there must be an affirmative answer to the question ‘At the priority date, would the skilled person know that microemulsions were one of the alternatives available for the formulation of intractably insolvent active ingredients for oral administration?’ I am satisfied on this evidence that the answer to the question is negative.

So the skilled person must characterise an emulsion using the specification as a guide. It is not a good guide. The passage at page 11 quoted above is positively misleading: microemulsions do not possess ‘one or more’ of the listed characteristics, but all of them. Consideration of the references given (Rosof, Friberg and Müller) shows this to be the case. The references confirm the upper limit for particle size of 200nm. I shall refer to the ‘liquid crystal’ passage below.

The problem in this case is whether it is right to say that when Equoral is added to water, the result is two emulsions (one microemulsion and one non-microemulsion) or whether the result is an emulsion which while possessing some particles in the same size range as a microemulsion would have is in fact an emulsion which does not have the defining characteristics of microemulsions as described by the specification and as described by the references. Professor Lawrence considered that the former was permissible, while Professor Attwood said that it was not.

The specification does not suggest, subject to the point on liquid crystals, that any significant part of the size distribution of a microemulsion might lie above 200nm. Under cross-examination, Professor Lawrence said this (transcript page 161):

‘Q… I am trying to sort out what your opinion is. I am talking about the use of the word "microemulsion" now in the context of the 770 patent. I am asking whether it is your contention that the word as used in the 770 patent extends to cover something which is in part a microemulsion and in part contains particle sizes greater than 200nm.

A. The patent does talk about having systems with other components in which are greater than that particular size. It is quite hard to divorce the two things. If you are asking me about how it defines the microemulsion patents and microemulsion which is 82(a) and (b), I agree, it does not say anything about a part there, but in the patent as a whole, it does imply that there may be mixed systems.’

This I take to be an acceptance that the possibility is not referred to by the passage on page 11 (the references are to her first report). When the point was revisited, she said this:

‘Q…If it was written like that and you had something with particles less than 2000 angstrom and particles more than 2000 angstroms, whether you would call the resultant product a microemulsion or not?

A. I think there is probably a misunderstanding here. I would call a microemulsion coexisting with something else anyway. I would never call it, if it had two populations, a microemulsion per se. That would just be totally wrong. I would say that the particles that are microemulsion particles and particles that are something else. Does that answer your question?

Q. It is the answer you have given to my question. Let me add one further rider to that description. Assume that it is said that the particles are of a size less than 2000 angstroms hence their optical transparency. You have added into it optical transparency. Would you then be prepared to call a mixture of 2000 angstrom and smaller particles and 2000 angstrom larger particles a microemulsion or not?

A. Absolutely not. I would not call it a microemulsion. I thought a microemulsion was something else so the same thing still stands. If I could prove that there are particles of less than 2000 angstroms that are thermodynamically stable, isotropic, have the correct components making them a form of microemulsion droplets.

Q. Would you call the system a microemulsion?

A. No, I have just said that it would be whatever it was. For example, a suspoemulsion. You get suspensions and emulsions. You could have suspomicroemulsions. It is appropriate for whatever components are in there.’

While, therefore, there are microemulsion-sized particles in an Equoral dispersion, that dispersion is not a microemulsion.

The second question is whether Equoral contains a lipophilic phase. Professor Attwood says that the polyglycerol (3) oleate is not appropriately described as the lipophilic phase for the purposes of the invention: it is not hydrophobic, but an amphiphilic surfactant. On the footing ethanol is the hydrophilic phase, as it appears to be, the evidence is that the polyglycerol (3) oleate is soluble in the hydrophilic phase, which does not comply with the indication at page 15 of the specification. The miscibility of the oil and water phases does not sit easily with the suggestion that the emulsion is to be an oil-in-water emulsion.

Many materials, (including the Miglyols which are the preferred lipophilic phases of the patent) are miscible with ethanol. But one of the advantages of the invention is to enable the use of ethanol to be avoided and to use other hydrophilic materials that are less volatile. It seems clear from the patent (page 5) that in Sandimmun the ethanol was a carrier medium and that the cyclosporin was partitioned between the ethanol and the olive oil. Mr Alexander’s submission that to construe hydrophobicity as to require non-miscibility with ethanol would exclude the preferred embodiment of the patent in which Miglyol is used is without substance. Miglyol was preferred in a formulation one of whose claimed advantages is the elimination of the use of ethanol, although it must be accepted that the patent does contemplate the use of ethanol as the hydrophilic phase as a non-preferred alternative.

Furthermore, the patent says (page 15) that the lipophilic material is a solvent which will be ‘devoid or substantially devoid of surfactant function’. Polyglycerol (3) oleate is an amphiphilic surfactant and is certainly not free of surfactant function.

Mr Alexander submits that there is no need to give the word ‘lipophilic’ such a narrow meaning. In my judgment, there is every reason to give the word its natural meaning when the underlying physical system (the microemulsion) is not part of the common general knowledge and the skilled person is relying upon the patent itself, and the references it contains, to construct his view of how a microemulsion is to be characterised and how it works. The claim calls for distinct hydrophilic and hydrophobic phases and a surfactant. If one surfactant is to be taken to be the hydrophobic phase, why not the other two also? They will form aggregates in aqueous medium in which the active ingredient is held: but they will all act together, as Professor Attwood emphasised. I conclude that it is artificial and incorrect to construe the word ‘lipophilic’ otherwise than as meaning substantially immiscible with the hydrophilic phase and destined to form the dispersed phase in the resulting oil-in-water microemulsion. It follows that in Equoral the active ingredient is carried by the mixture of surfactants and to some extent the ethanol as well. There is no separately identifiable lipophilic phase.

Once, therefore, it is established that the Equoral pre-concentrate produces a dispersion in which the average particle size exceeds 200nm, it follows that it is not a microemulsion pre-concentrate. It is a pre-concentrate that produces both microemulsion-sized particles and others. I think that the fact that so little of the active ingredient (≤14%) was carried in the microemulsion-sized particles goes some way to support this view. The fact that it is impossible to identify a lipophilic phase distinct from a mixture of amphiphilic surfactants having differing hydrophilicity/hydrophobicity indexes means that the ternary mixture called for by the claim is not present. The allegation of infringement fails.

Validity is challenged on the basis of US 4,146,499 (Rosano) and the common general knowledge. The common general knowledge included Sandimmun. In summary, the problem is to improve on Sandimmun so as to improve its qualities, as for instance stability of the resulting dispersion or bioavailability. Rosano is a general description of how to formulate microemulsions. It contains no suggestions relating to formulation of cyclosporin. Unless the skilled person sees that a microemulsion is the way to go, I cannot see how Rosano supports an allegation of obviousness. It is not a document that the skilled person will greet with ‘That gives me what I want’ unless he already appreciates the merits of a microemulsion formulation. This is a good example of hindsight reasoning. It is not surprising that Ivax concentrated at trial on the case based on common general knowledge alone.

Against the case based on the common general knowledge, the patentee put a submission that can be summarised in the contention that had the alleged invention been obvious, one or more formulators would have made the invention earlier than it was, in fact, made. The earliest priority date claimed for ’770 is 1989, and Sandimmun had been on the market since 1982/3. So, says the patentee, the 6 years or so must be explained away.

A difficulty with arguments of this sort is that one cannot be confident of arriving at a fair answer without being aware of the history of the developments as they in fact took place. There was no evidence of the history in this case. A patentee will invoke industry’s failure to act because it will be assumed that had the alleged invention been obvious it would have been made by someone else, and in some cases this assumption is entirely justified — see Vickers v Siddell (1890) 15 App Cas 496, 7 RPC 292, a good example where the invention contributed to safety in forges and the need for safety would have contributed a strong incentive. Where a patent like the present one is under consideration, one must be careful before assuming that the skilled person is free to do the obvious thing. Until patent protection on the active ingredient (or the use of the active ingredient in a pharmaceutical preparation) has expired, the skilled person will be less inspired to do further formulation work, unless he or she is an employee of the relevant patentee. Where there is relevant patent protection, in other words, the actual behaviour of the industry cannot provide any sort of reliable guide to help provide a rational resolution of the question whether the invention is obvious or not. The behaviour of the patentee’s employees may help (see Pharmacia v Merck [2002] RPC 410). As Sir Donald Nicholls V-C said in Mölnlycke v Procter & Gamble [1994] RPC 49, 113 secondary evidence of this kind is merely an aid in assessing the primary evidence of inventive step, which will be provided by the expert witnesses. Often it will help in eliminating the baleful effects of hindsight.

Genuine technical improvement is nevertheless relevant. Technical merit may exist in an obvious development from a non-obvious starting point, but it does give rise to the question why the patentee did not start with the allegedly obvious development. Neoral is undoubtedly an important development over Sandimmun, and the evidence to this effect was accepted entirely by Professor Attwood.

The evidence was relied on in two ways. First, as I have said, the defendants said that the skilled person, confronted with Sandimmun an emulsion pre-concentrate, would know of his common general knowledge that cyclosporin was a large, insoluble molecule and would know that an emulsion was an obvious way to administer it. If Sandimmun turned out to have unpredictable pharmacokinetics, it was likely that poor absorption was the problem, and the obvious way of approaching the problem of poor absorption was to reduce the size of the emulsion particles.

Professor Attwood was challenged on the proposition that reduction of particle size and the corresponding increase in droplet surface area would be thought to increase absorption in the gut. Once one looks at the rival candidates for the mechanism whereby the active ingredient passes into the wall of the gut and so into the bloodstream, it would be surprising if surface area was irrelevant to any of them: and the larger the surface area, the better. But this is a complex area, even now, and general propositions of this sort are probably unwise. In this area, there was genuine doubt whether droplet size did effect absorption from the gut, as the paper called Tarr & Yalkowsky, published 1989, showed.

If obviousness is to be demonstrated by an argument of this kind, it is essential for the defendants to show that it was obvious to the skilled person that the particle size of the emulsion could be reduced, and that the obvious way to reduce it was to formulate a microemulsion. Essentially for the reasons I have given in considering the common general knowledge, I think they have failed to show this. In dealing with the disclosure of Bhargava (which was relied on as showing the common general knowledge, but not as a pleaded prior publication) Professor Attwood gave these answers:

‘Q. But an oil in water microemulsion is quite hard to produce.

A. It is. It is just relating one to the other.

Q. If you look at the phase diagram at the bottom of the page, Fig. 1, you will see a rather small section of the phase diagram. A skilled person would take from that that there was a pretty narrow window of opportunity that one had.

A. They would recognise it as a challenge, yes.

Q. That phase diagram, which is a ternary diagram, involves a concentration of the surfactant, the primary surfactant, being constant. A skilled person would take from this that both a primary surfactant and a cosurfactant were required; is that right?

A. Yes.

Q. They would also take from this, would they not, that in order to prepare an oil in water microemulsion, it was necessary to follow a four-step procedure which is set out in that little section.

A. I would agree, but I do not think the steps are particularly difficult, are they? They are just mixing things with other things.’

Two points can be made here. If microemulsions are not common general knowledge, this is really research work. Second is the professor’s use of the word ‘challenge’. It is easy to give too much weight to soundbites, but the answer was given in respect of a skilled person either reading Bhargava or at least equipped with knowledge equivalent to what Bhargava discloses. Deprive the skilled person of that crutch, and the answer acquires additional force.

On the other hand, if I am wrong in my finding that microemulsions were not part of the formulator’s common general knowledge at the date, I think that I would not be deterred from a finding of obviousness by the professor’s use of the word challenge. The broadest claims are just claims to microemulsion pre-concentrates, and invention would have to be found in particular formulations or particular components. As it is, I am satisfied that the attack on the basis of Sandimmun and the common general knowledge fails.

US 4,146,499 (Rosano) is relied on as the starting point for a second attack on the ground of obviousness. Rosano was put to Professor Lawrence as a document considered by a skilled person who was faced with the task of improving the formulation of cyclosporin at the priority date. Mr Alexander QC submits that Professor Attwood accepted that Rosano:

This summary is, I think, fair. While I do not attach so much importance to the obvious need to avoid excipients and other components that are pharmaceutically acceptable, I do accept that Rosano does not, on a fair reading, suggest that a microemulsion is a potentially valuable method of administering cyclosporin. The obviousness attack on the basis of Rosano likewise fails, and the patent is valid.

The invention of this patent lies in what is referred to the second component. It is not well drafted:

‘As discussed in [770 and GB 2 257 359] the cyclosporins present highly specific difficulties in relation to administration generally and galenic composition in particular, including in particular problems of stability, drug bioavailability, and variability in inter– and intra–patient dose response.

In order to meet these and related difficulties, in GB patent publication no. 2 222 770 and 2 257 359A, galenic compositions are disclosed comprising a cyclosporin as active ingredient and which take the form of, inter alia, an emulsion, e.g. microemulsion, or emulsion, e.g. microemulsion, pre-concentrate. Such compositions typically comprise 1) a hydrophilic component, 2) a lipophilic component, and 3) a surfactant.

In accordance with the present invention it has now surprisingly been found that particularly stable emulsion e.g. microemulsion, or emulsion, e.g. microemulsion, pre-concentrate galenic compositions with difficultly soluble active agents having particularly interesting bioavailability characteristics and reduced variability in inter– and intra–subject bioavailability parameters, are obtainable using at least one component selected from the group consisting of (i) fatty acids, (ii) N-methyl pyrrolidone, and (iii) a transesterification product of hydrogenated palm kernel and/or hydrogenated palm oil with polyethylene glycol 15000, hereinafter referred to as “secondary component”.’

I am not sure what ‘galenic’ means (the dictionaries are unhelpful), but it probably has the same general sense as ‘pharmaceutical’. The claim is to “a composition in form of an emulsion or microemulsion pre-concentrate for oral administration…” and the first question is what kind of pre-concentrates are covered. ’770 I have discussed above: it is restricted to microemulsions. GB 2 257 359 is concerned both with emulsion and micro-emulsion pre-concentrates, microemulsions being preferred (see page 11). Finally, the patent itself contains a remarkable definition:

‘An “emulsion, e.g. microemulsion, pre-concentrate is defined in this specification as being a composition which spontaneously forms an emulsion, e.g. microemulsion, in an aqueous medium, for example, in water…’

Other references to spontaneous formation all relate to microemulsions: as ’770 says, one of the features that distinguish microemulsions from emulsions is that they form spontaneously (Footnote: ¹). All this leads Ivax to submit that the claim is to be construed as extending to microemulsions and their pre-concentrates only. I do not accept this. The words are clear, and to the extent that they cover pre-concentrates, they cover any pre-concentrate that spontaneously forms either an emulsion (if that is possible) or a micro-emulsion.

The whole of the claim is as follows:

‘A composition in form of an emulsion or microemulsion pre-concentrate for oral administration comprising

1)

a cyclosporin or macrolide, and a carrier medium comprising

2)

a second component selected from the group consisting of

(i)

fatty acids

(ii)

N-methyl pyrrolidone

(iii)

a transesterification product of hydrogenated palm kernel and/or hydrogenated palm oil with polyethylene glycol 1500

3)

a lipophilic component, and

4)

a surfactant

with the proviso that when component 2) consists of N-methyl pyrrolidone, the composition is in the form of a microemulsion pre-concentrate.’

The description of the specification, which is reflected in claim 2, is that the lipophilic component may be transesterified ethoxylated vegetable oil, mixed mono-, di- and triglycerides, propylene glycol mono- and di- fatty acid esters, and esterified compounds of fatty acid and primary alcohol. The real question in the present case is really whether the second component has to be added to the mix, or whether it is sufficient for infringement that it be present as an impurity in one of the other elements, such as the fatty acid esters. For example, the specification accepts that the material Maisine, a mixed mono-, di- and tri-glyceride, contains 1% free oleic acid. Oleic acid is a preferred second component, but the specification does not suggest that this is relevant. Similarly, the polyglycerol fatty acid esters, which are exemplified on page 15 as potential surfactants, will contain levels of impurity including unreacted fatty acid. For example, the two Decaglyn materials contain lauric and myristic acids as impurities. These are respectively the C12 and C14 saturated fatty acids (see transcript 254).

Every reference to the second component refers to its being ‘used’, and the specification describes as a further alternative aspect of the invention (albeit not claimed) ‘a process for the production of a pharmaceutical composition as defined above, which process comprises bringing (1) the second component; (2) the lipophilic component; and (3) the surfactant into intimate admixture and adding the active agent…’.

I do not find any support in the specification for a claim whose scope extends to materials of the kind described in ’770, which contain as a recognised impurity a material suitable for use as a ‘second component’. I accept the contention that the words cannot cover amounts of these materials merely present as impurities. If they do cover such materials present as impurities, the impact on validity is significant and I discuss it below.

Four points of distinction from the claim are relied on. These are (i) Equoral is not a microemulsion pre-concentrate (ii) nor is it an emulsion pre-concentrate, being a liquid crystal dispersion (iii) it has no second component, and (iv) it has no lipophilic component. Bearing in mind the extensive cross-referencing between ’770 and this patent, together with the way in which the invention is described by reference to ’770, I will not give the phrase ‘lipophilic component’ in this claim a different meaning from that which I gave it in discussing ’770, even though I consider that by the priority date of the patent microemulsions had become part of the common general knowledge. I cannot simply ignore the passage on page 15 of ’770, given that the patent relies (page 8) on the disclosure of ’770 for a wide disclosure of suitable lipophilic components.

It is not suggested that the ~1% fatty acids present in Equoral are present otherwise than as impurities in the other components. The claim is not infringed for this reason. While I consider that Equoral is not a microemulsion, I do not think that it is an emulsion as that term is used in the specification and in accordance with its usual sense. As I have indicated, the evidence as to its liquid crystal nature was not satisfactory — it displayed little birefringence in the repeat experiments — but I find that the term ‘emulsion’ does not cover it, because there is no lipophilic component, as that phrase must be construed. So the allegation of infringement fails.

I shall deal first with the allegations of anticipation that depend upon my construction of the claim as to ‘second component’ being incorrect. These are based on the prior publication of ’770 and ’359A, although the latter adds little, and upon the prior sale of Neoral.

It is accepted that a small quantity of fatty acid will be present in any of the mono-, di- and tri-glycerides. Myritol (page 16 of ’770) and Maisine (page 6 of 359A) will contain corresponding fatty acid impurities. Cremophor EL (’770 page 17) has an acid number of 2. The calculations were not entirely satisfactory, but I think that it was established that before the priority date it is likely that Neoral contained up to about 0.48% by weight free fatty acid, and that a mixture according to the disclosure of ’359 about 0.47%. I am not myself happy about the impact of this upon claim 10, which I find difficult to construe:

‘a composition according to any preceding claim wherein the ratio of the cyclosporin or other macrolide : second component : lipophilic component : surfactant is 1 : 0.1 to 10 : 1 to 10 : 1 to 10 on the basis of weight.’

This appears to include a mixture of 1 part cyclosporin, 0.1 parts second component, 10 parts lipophilic component and 10 parts surfactant. This means that the weight fraction of second component is at the lowest extreme 0.1/21 or 1/210, which is 0.48%. This is very close indeed to the calculation at X25 of 0.47% for ’359. It is the same as the maximum figure for Neoral, based on the calculations made by Professor Lawrence in her third report. It follows, I think, that claim 1 would be bad, and probably claim 10 also, but I prefer not to express a concluded view on that subject, if only because there is force in Mr Alexander’s submission that it has not been proved that any Neoral in fact had that content of free fatty acid.

A further stand-alone citation, relevant only if I am wrong on construction, is PCT WO98/10747 (Stuchlik), a citation which can be relied on as novelty-destroying only by virtue of section 2(3) of the Patents Act 1977. Example X of this document, which is said to be of a material which is a pre-concentrate forming liquid crystals, undoubtedly falls within the claims if the claims cover a system which is neither an oil-in-water emulsion pre-concentrate nor a water-in-oil emulsion pre-concentrate. The claim is undoubtedly restricted to materials forming an “emulsion, e.g. microemulsion” on admixture with water. The disclosure of ’770 is, as I have said, incorporated by reference on the characteristics of a microemulsion (page 3) and this, Mr Alexander QC said, clearly covered a liquid crystal system on the basis of a passage on page 10. Professor Attwood considered that it was simply incorrect to consider that the term “microemulsion” encompassed liquid crystals, which did not possess any of the other features of microemulsions. Professor Lawrence considered that the term did, indeed, include liquid crystals which were small enough. To this extent, it may be that the specification has written its own dictionary: the term is used to cover liquid crystals that are small. On admixture with the aqueous phase Stuchlik’s material forms a liquid crystal mesophase: in the gut, this makes what is called a “very fine suspension”. The size of the suspension is described in a short sentence:

‘The size distribution of the particles resulting from the complete dispersion of the product in an aqueous phase advantageously ranges to a maximum of 1000nm.’

This is larger than the particle size (<200nm) to be expected in a microemulsion. So there is no anticipation. If there were infringement of ’770, the answer would be different.

I should just add this. One of Mr Alexander’s points on Example X was that it was one only of a number of examples, others of which had no effect on the novelty of the claims. That, he submitted, is ‘not directions which are “clear” or “unmistakeable”’. I reject this. If a patent specification contains an example of its invention which is complete enough to be performed and, when performed, always falls within the later claim, it anticipates. The fact that non-anticipatory matter is also disclosed is irrelevant. There are clear and unmistakeable directions to perform all the examples of the invention. The fact that choice between the examples may be exercised does not make the individual example, and the group of directions it contains, any the less clear or unmistakeable. The position is not analogous to a single example containing one of more choices if the way the choices are exercised will affect whether the result of performing the example falls within the claim. The latter case is generally considered to be incapable of supporting an allegation of anticipation by inevitable result, for the very good reason that the prior inventor has not planted his flag within the later inventor’s claim. In the former case, he has done so.

I shall approach this question along the lines set out by the Court of Appeal in Windsurfing International v Tabur Marine [1985] RPC 59. Neoral is the first starting point. The evidence is that the composition of Neoral would be known to the skilled person at the priority date. It is legitimate for this purpose to hypothesise a skilled person as wishing to improve, or vary, an existing product. It will be remembered that the whole purpose of these formulations is to get cyclosporin, substantially insoluble in aqueous solution, into the patient. Neoral is a pre-concentrate which will produce an oil-in-water microemulsion on addition to an aqueous medium, so that the cyclosporin is not precipitated. It is formulated in accordance with ’770, and it contains alcohol. The suggestion made by Mr Thorley QC is that it is obvious to carry out a solvent screen to find good solvents for cyclosporin, and, in particular, solvents which are suitable for incorporation in a micro-emulsion pre-concentrate. Neoral includes ethanol which will act as a solvent of cyclosporin, but the objective is not necessarily to replace the alcohol but to improve the dissolution of the cyclosporin.

A word of explanation is necessary at this point. The law is that if a claim covers anything that is obvious, the claim is bad. It does not matter that the claim also covers things that are meritorious, and far from being obvious. To be valid, it must cover only things that are not obvious, and the patentee’s remedy is to amend a claim which covers more than it should. A claim formulated using the word ‘comprise’ is not restricted to covering only the listed components, because there is an established international usage in this context to the effect that ‘comprise’ means ‘include’, a usage that can only be displaced by clear words. The EPO Guidelines put it like this:

‘4.13 "Comprising" vs. "consisting"

While in everyday language the word "comprise" may have both the meaning "include", "contain" or "comprehend" and "consist of", in drafting patent claims legal certainty normally requires it to be interpreted by the broader meaning "include", "contain" or "comprehend". On the other hand, if a claim for a chemical compound refers to it as "consisting of components A, B and C" by their proportions expressed in percentages, the presence of any additional component is excluded and therefore the percentages should add up to 100% (see T 759/91 and

T 711/90, both not published in OJ).’

Claim 1 of ’674 does not exclude the presence of other elements, and in particular other solubilisers for cyclosporin including ethanol. In any event, this is made clear in the body of the specification:

‘The second component may include further hydrophilic co-components, for example lower alkanols such as ethanol. These co-components will generally be present in partial replacement of other components of the second component. While the use of ethanol in the compositions is not essential, it has been found to be of particular advantage when the compositions are to be manufactured in soft gelatine, encapsulated form. This is because storage characteristics are improved, in particular the risk of active agent precipitation following encapsulation procedures is reduced. Thus the shelf life stability may be extended by employing ethanol or some other such co-component as an additional ingredient of the second component. In a further alternative aspect the ethanol may comprise 0 to 60 % by weight of the second component; preferably 20 to about 55% by weight and more preferably about 40 to 50 % by weight. Small quantities of liquid polyethylene glycols may also be included in the second component.’

This is in remarkable contrast to ’770, one of whose benefits was said to be the removal of the undesirable ethanol from soft gelatine capsule formulations (see pages 5 and 8), but this inconsistency is not really material.

As put by Mr Thorley to Professor Lawrence, the question was whether it was obvious to look at the lipophilic element of the formulation with a view to improving the solubility of cyclosporin in it, this being the difference between Neoral and a formulation covered by the claim (transcript 259):

‘Q. So you are regarding this as being a microemulsion exercise simply without regard to the ability to solubilize cyclosporin.

A. No, I guess the point that if I was trying to improve on a formulation which already exists, you are right, one of the first things I do is to have a look at the solvents. I have no problem with that. I do not believe that from what I was trying to formulate n-methylpyrrolidone would be one of them.

Q. But some sort of fatty acid would be.

A. Yes, it would be.’

Mr Thorley then cross-examined Professor Lawrence about the use of n-methylpyrrolidone, a solvent for cyclosporin, described in the document called Woo. Professor Lawrence was emphatic about the non-obviousness of the use of this material (which is hydrophilic) but her further answers are most instructive (transcript 261-2):

‘Q. Can I suggest that if you read this patent in 1998 when you were seeking to optimise the Neoral formulation, that would encourage you to include n-methylpyrrolidone in your list of potential additional solvents to look at. You pause.

A. I still think that if I was going back and reformulating, I would not be looking at in particular hydrophilic. I would be looking to see what I could do with the lipophilic phases. The problem is that when you dilute the microemulsion pre-concentrate in vivo, you have got to keep that solution still [i.e. the solution of cyclosporin in lipophilic material]. The solution probably, when it is dispersed, the n-methylpyrrolidone is probably not associated with the droplets [of cyclosporin in lipophilic material forming the dispersed phase of the emulsion]. I would say that it is probably a dangerous ploy to go with. My instinctive reaction is that I probably would not for exactly that reason.

Q. That is what drives you back into the fatty acids and more lipophilic ----

A.

Exactly.’

(My understanding of the explanations in square brackets)

On this evidence, the use of n-methylpyrrolidone may well not be obvious (this depends on a proper view of Woo’s disclosure) but employing fatty acids certainly is, as a matter of the common general knowledge for the purpose of improving, or attempting to improve, the solubilization of cyclosporin in the lipophilic phase.

This hypothesis does not require the skilled person to contemplate the substitution of the ethanol present in Neoral with a fatty acid. It is clear from the evidence of both experts that one would concentrate on the phase which appeared to be the one carrying the cyclosporin, that is, the lipophilic phase. I got the clear impression that alteration of the hydrophilic phase was not a first-order investigation, but it does not really matter. The use of additional fatty acids, in a Neoral formulation, to improve the dissolution of the cyclosporin was obvious in 1998 in the light of the common general knowledge.

I was less impressed with the argument based on Woo, which discloses a topical formulation of cyclosporin in which the solubilising component is n-methylpyrrolidone. Woo discloses oil-in-water emulsion pre-concentrates which are intended to form emulsion gels on the addition of water that are suitable for topical application. The omission of the gelling agent results in a pre-concentrate that is obviously capable of forming an oil-in water emulsion. The difference between this disclosure and the claim is thus the use of the gelling agent. Is it obvious to throw the gelling agent away if one wishes to make a pre-concentrate for oral use? Perhaps it is, but this is not quite enough. Where n-methylpyrrolidone is the solubilizing agent, the claim requires that the pre-concentrate is a microemulsion pre-concentrate: nothing is said in Woo about microemulsions. I do not find this surprising where the formulation is for topical administration. I can find no evidence that Woo compositions free of gelling agent form microemulsions, and it follows that if the claim is to be obvious the argument must be that the formulator wishing to improve Neoral reads Woo, notes that n-methylpyrrolidone is a good solvent of cyclosporin, and resolves to improve his Neoral by adding that material to the formulation. I found Professor Lawrence’s reasons that I have quoted above for not taking this course persuasive, but of course they threw her back on the lipophilic component and the obvious addition of fatty acids.

Mr Alexander submits that in 1998 there was no motive to improve Neoral. Absent some evidence that Neoral met all medical needs, I do not accept this. The object is to ensure that the cyclosporin does not precipitate in aqueous environments: the object is therefore to improve its solubility in the lipophilic phase as much as possible. This seems to me to be quite obvious.

Accordingly claim 1 of this patent as it stands is invalid. I will hear counsel on the appropriate relief.

’770 is plainly valid, but not infringed by the Equoral formulation. ’674 is not infringed but is anyway invalid.