Novartis AG v Dexcel-Pharma Ltd

This is a claim by the Claimant (“Novartis”) for infringement of United Kingdom Patent No. 2 222 770 (“the Patent”). The specification of the Patent is entitled “Pharmaceutical compositions comprising cyclosporins”. The priority date of the Patent is 16 September 1988. There is no challenge to the validity of the Patent, the validity of which was upheld by Pumfrey J in Novartis AG v Ivax Pharmaceuticals UK Ltd [2006] EWHC 2506 (Pat). The sole issue is whether the Defendant’s (“Dexcel’s”) product Deximune infringes the Patent. Floyd J granted an interim injunction having concluded that there was a serious question to be tried as to whether Deximune infringed claim 1. As a result Dexcel has not yet marketed Deximune in the United Kingdom, although it has subsequently received authorisation to do so.

Although claim 1 of the Patent extends to cyclosporins as a class, the present case concerns the first of the cyclosporins to be isolated, cyclosporin A, referred to in the Patent as Ciclosporin. This is a naturally-occurring cyclic peptide produced by certain fungi. Like the parties, I shall refer to it simply as cyclosporin. Cyclosporin is used for a variety of indications, but its best known application is to produce immunosuppression in recipients of organ transplants. In 1988 cyclosporin was known to be a challenge to formulate, in particular because it is almost completely insoluble 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 marketed by Sandoz (now Novartis) under the trade marks Sandimmun and Sandimmune. 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(e) was an oral formulation. It was based upon olive oil, a surfactant (Labrafil) and ethanol. Sandimmun(e) was what is called a pre-concentrate: when diluted with an aqueous medium (such as a drink or the aqueous contents of the stomach), it forms an oil-in-water emulsion.

The Patent discloses an improved formulation which is marketed by Novartis under the trade mark Neoral. It differs from Sandimmun(e) in being a microemulsion pre-concentrate rather than an emulsion pre-concentrate.

Novartis called two experts, Professor Jayne Lawrence and Dr Peter Griffiths. Professor Lawrence is Professor of Biophysical Pharmaceuticals at King’s College, London. She obtained a BSc in Pharmacy in 1981 and then qualified as a pharmacist. In December 1985 she received a PhD for research into synthesising and characterising the aggregation of micelle-forming surfactants. By the priority date of the Patent she had had experience with making and characterising microemulsions. Since then her research has been particularly focussed on the synthesis of amphiphilic molecules, the characterisation of aggregates formed by such molecules and the use of such aggregates, including microemulsions, as drug delivery vehicles. In 1993 she spent six months on sabbatical working in the Respiratory Product Development Section of what is now GlaxoSmithKline. Professor Lawrence also gave evidence on behalf of Novartis in the Ivax case.

Dr Griffiths is a Reader in Physical Chemistry at Cardiff University. In September 1991 he obtained a PhD for research into Nuclear Magnetic Resonance (NMR) measurements of the diffusion of polymer mixtures in solution. Since then his research has focussed on a wide range of polymer and surfactant based systems using NMR as the principal characterisation tool.

Dexcel called one expert, Professor Julian Eastoe. He is Professor of Chemistry at the University of Bristol. He obtained a BSc in Chemistry in 1986. He was awarded a PhD for work on microemulsions in 1990. Since then his research has centered on colloid and interfacial science with a particular interest in microemulsion chemistry. In addition he has consulted for pharmaceutical companies on formulation issues.

Although Dexcel suggested that Professor Lawrence had become close to Novartis and Novartis suggested that Professor Eastoe was overzealous, I am satisfied that all three experts did their best to assist the court.

More importantly, counsel for Novartis submitted that Professor Lawrence had had more experience of pharmaceutical formulation, and had approached the case more from the perspective of a formulator, than Professor Eastoe, who had approached it more from the perspective of a chemist. Furthermore, counsel submitted, Professor Eastoe’s reading of the Patent was strongly coloured by his own understanding of what constituted a microemulsion. I agree with both these points.

Counsel for Dexcel submitted that both Professor Lawrence and Professor Eastoe had much greater expertise in microemulsions than the person skilled in the art to whom the Patent is addressed would have had at the priority date. I also agree with this, and have borne it carefully in mind when considering the evidence.

A patent specification is addressed to those likely to have a practical interest in the subject matter of the invention, and such persons are those with practical knowledge and experience of the kind of work in which the invention is intended to be used. The addressee comes to a reading of the specification with the common general knowledge of persons skilled in the relevant art, and he (or, once and for all, she) reads it knowing that its purpose is to describe and demarcate an invention.

In the present case there is no dispute between the parties as to the identity and attributes of the person skilled in the art to whom the Patent is addressed. The addressee is a pharmaceutical formulator who is interested in formulating cyclosporin. Such a person would have a degree in pharmacy or a related scientific discipline such as chemistry and experience in pharmaceutical formulation.

The common general knowledge of the addressee falls into two fields, namely the properties of cyclosporin and colloidal systems.

I have already summarised this aspect of the common general knowledge to the extent necessary for the purposes of this case in paragraph 2 above.

Before describing colloidal systems, it is first necessary to explain some of the terminology used in this field.

“Lipophilic”, “hydrophilic” and “amphiphilic”. A hydrophilic compound is one that is water-loving, i.e. soluble in water. A lipophilic compound is oil-loving, i.e. soluble in oil. Lipophilic compounds may also be described as hydrophobic, i.e. water-hating. An amphiphilic compound contains distinct hydrophilic and lipophilic regions within its structure. It is not possible to draw a sharp boundary between hydrophilic, lipophilic and amphiphilic compounds, however. The terms hydrophilic, amphiphilic and lipophilic are relative, and depend on the context in which they are used. Furthermore, the terms are not mutually exclusive. An amphiphilic compound can correctly be classified as lipophilic if lipophilic properties are dominant in the relationship with water molecules. Likewise, an amphiphilic compound can be classified as hydrophilic if hydrophilic properties dominate. A hydrophilic compound may contain a lipophilic part, like an alkyl chain, as long as the hydrophilic properties dominate, for example ethanol. Likewise, a lipophilic compound may contain a part that interacts favourably with water although the lipophilic parts dominate. Whether a given molecule will act as, say, a hydrophile or an amphiphile will depend on the other molecules present in the system under consideration.

“Surface active”, “surfactant” and “cosurfactant”. A surface active molecule is an amphiphilic molecule that exhibits surface activity (i.e. tends to locate itself at surfaces and interfaces and thereby alter the nature of that surface or interface). A surfactant is a molecule that exhibits surface activity, but which also self-assembles (or aggregates) in solution (either spontaneously or on addition of energy to the system). Although all surfactants are amphiphilic molecules, not all amphiphilic molecules are surfactants. For example, DNA is an amphiphilic molecule, but it has no significant surface activity and is not a surfactant. Surfactant function can also be dependent on the other components in a formulation. Thus some components are surfactants in some solvents, but only surface active in other solvents. A cosurfactant has surface active properties (i.e. it can lower surface tension and therefore contribute to the effect of the surfactant), but is not able to form aggregates on its own.

“Phase”. The word phase is used in chemistry to denote a state of matter, or a part of a system, which is characterised by uniformity in chemical composition and physical properties and which possesses a clearly defined boundary surface or interface.

Colloidal systems. A colloidal system is an intimate mixture of two phases, one of which, called the dispersed phase (or colloid), is distributed in a finely divided state through the second phase, called the dispersion medium (or dispersing medium). The dispersion medium may be a gas, a liquid, or a solid, and the dispersed phase may also be any of these, with the exception of one gas in another. There is no sharp line of demarcation between solutions and colloidal systems or between colloidal systems and suspensions. When the particles of the dispersed phase are smaller than about l nm in diameter, the system begins to assume the properties of a true solution. When the dispersed particles are larger than 10,000 nm, separation of the dispersed phase from the dispersing medium becomes rapid and a true suspension (as opposed to a colloidal suspension) is formed.

Colloidal systems can result from the interaction of surfactant molecules to form aggregates. These aggregates can vary in size and may take on a variety of different shapes including spherical, cylindrical and lamellar sheets. There are a number of different types of aggregates, including micelles, vesicles (also known as liposomes), emulsions and microemulsions.

Emulsions. Emulsions are intimate mixtures of two immiscible phases, typically (but not exclusively) an oily phase and an aqueous phase. They generally have a milky or cloudy appearance, and are typically stabilised (insofar as they are stable) so that they do not "separate out" by a surfactant, a surfactant and cosurfactant or a mixture of surfactants, which form a monolayer (barrier) between the oily phase and the aqueous phase. There are two main types of emulsions: oil droplets dispersed in water, known as "oil-in-water" emulsions (where water is the "continuous phase" and oil is the "dispersed phase"), and water droplets dispersed in oil, known as "water-in-oil" emulsions (where water is the "discontinuous phase" or "dispersed phase" and oil is the "continuous phase"). The dispersed phase droplet size ranges from around 200 to l0,000 nm in diameter. The droplets are generally, but not exclusively, spherical. An emulsion may comprise droplets of a range of sizes which vary each time the emulsion is made.

Microemulsions. Like emulsions, microemulsions are intimate mixtures of two immiscible phases, typically (but not exclusively) an oily phase and an aqueous phase. They are typically stabilised by a surfactant which may be present in combination with a cosurfactant. As with emulsions, microemulsions may be oil-in-water microemulsions or water-in-oil microemulsions. Despite these similarities, microemulsions differ from emulsions in a number of respects. It is not appropriate to go into the characteristics of microemulsions at this stage, however, since it is common ground that (as Pumfrey J concluded in the Ivax case at [27]) the skilled person in 1988 would know very little about microemulsions. They were almost unheard of in the pharmaceutical field, and Neoral was the first drug to be formulated in a microemulsion when it was launched in 1992.

After describing cyclosporin, its actual and potential applicability and the class of cyclosporins, in the last paragraph on page 5 and the first paragraph on page 6 the specification identifies two problems with formulations, such as Sandimmun(e), which contain both ethanol and olive oil:

Thus the first problem with existing formulations that is identified is the problem caused by the high ethanol content of the formulation. The second problem is the variation in bioavailability of the active ingredient between patients, and even for a single patient over time.

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 in the paragraph bridging pages 7 and 8:

The specification goes on to describe other advantages of the invention, and in particular at the bottom of page 8 it states that the invention enables the preparation of compositions which are non-alkanol based, for example, free or substantially free of ethanol.

In the middle of page 9 the first aspect of the invention is stated:

The following passage on pages 9-10 defines an ‘oil-in-water microemulsion pre-concentrate’:

This passage is followed by a consistory clause corresponding to claim 1, in which compositions of the invention are designated as “(A)”.

In last two paragraphs on page 11 the specification states:

On page 12 the specification states:

A pharmaceutically acceptable C_1-5 alkyl or tetrahydrofurfuryl di- or partial-ether of a low molecular weight mono- or poly-oxy-alkanediol; or

1,2-propyleneglycol.”

In the second paragraph on page 15 the specification states:

In the penultimate paragraph on page 16 the specification states:

In the penultimate paragraph on page 21 the specification states:

In the second paragraph on page 22 the specification states:

The next paragraph on page 22 of the specification says that compositions in accordance with the invention may also comprise a thickening agent.

The last paragraph on page 24 is a consistory clause corresponding to claim 48, in which compositions of the invention are designated as “(B)”.

The last two paragraphs on page 28 of the specification state:

Pages 39-44 of the specification describe a number of examples of compositions of the invention. Example 1.1 contains 180 mg Glycofurol 75 (a tetrahydrofurfuryl alcohol polyethylene glycol ether described on page 13 as being preferred for use as the hydrophilic phase) and 90 mg Miglyol 812 (a fractionated coconut oil comprising caprylic-capric acid triglycerides described on page 15 as especially suitable for use as the lipophilic phase). Example 1.2 contains 180 mg Glycofurol 75 and 78 mg Miglyol 812. Example 1.3 contains 200 mg Glycofurol 75, 60 mg Miglyol 812 and 19 mg ethanol. Example 1.4 contains 100 mg Glycofurol 75 and 75 mg Miglyol 812.

Pages 44-48 of the specification describe bioavailability studies of compositions in accordance with the invention in dogs and in humans. In the human trial, described at pages 47-48, Composition I in accordance with the invention, which is the same composition as Example 1.1, is compared with Composition X, namely Sandimmun(e). Figure 3 plots the bioavailability of Composition I, and Figure 4 the bioavailability of Composition X. Comparison between the two shows a reduced variation between subjects in Figure 3. It can also be seen that Composition I gives a well-defined peak in blood concentration at about 1.5-2 hours.

The only claims which it is necessary for me to consider are claims 1 and 48.

Claim 1 is as follows:

Claim 48 is as follows:

Deximune is a pharmaceutical composition comprising cyclosporin as active ingredient. It is intended to be administered in a gelatine capsule. It consists of the following components:

cyclosporin (8.42%);

ethyl lactate (27.95%);

tricaprin (14.14%);

macrogolglycerol hydroxystearate (Cremophor RH 40) (14.14%);

lecithin (soy phospholipid) (7.07%);

sorbitan (mono)oleate (Span 80) (14.14%);

polysorbate 20 (Tween 20) (14.14%).

Ethyl lactate is an ester formed from lactic acid and ethanol.

Tricaprin is a medium chain fatty acid triglyceride which comprises at least 90% C10 chains with smaller amounts of C6, C8, C12 and higher chains. Its melting point is 33^oC, and so it is a solid at ambient temperature if ambient temperature is taken to be 20 or 25^oC.

Macrogolglycerol hydroxystearate, lecithin, sorbitan (mono)oleate and polysorbate 20 are all surfactants.

There are two main issues on infringement. First, does Deximune comprise a hydrophilic phase and a lipophilic phase? Secondly, is it a microemulsion pre-concentrate? The first issue is essentially one of construction, while the second issue depends partly on construction and partly on the technical facts.

There is no dispute that Dexcel will infringe claim 48 pursuant to section 60(2) of the Patents Act 1977 if Deximune forms a microemulsion falling within that claim upon administration.

The task for the court when construing a patent claim is to determine what the person skilled in the art would have understood the patentee to have been using the language of the claim to mean: see Kirin Amgen Inc v Hoechst Marion Roussel Ltd[2004] UKHL 46, [2005] RPC 9 at [30]-[35]. In that case the list of principles to be found in the judgment of Jacob LJ in Technip France SA’s Patent[2004] EWCA Civ 381, [2004] RPC 46 at [41] was approved subject to one point.

Novartis contends that ethyl lactate constitutes the hydrophilic phase and tricaprin constitutes the lipophilic phase required by the claims.

Dexcel does not dispute that tricaprin is lipophilic. Indeed, it is a member of the class of preferred substances for use as the lipophilic phase identified in the third paragraph on page 15 of the Patent.

Dexcel disputes that ethyl lactate is hydrophilic. It says that, if anything, it is slightly lipophilic. As Professor Eastoe accepted in cross-examination, however, it is miscible with water and can properly be described as a hydrophilic solvent. Indeed, Dexcel itself described ethyl lactate as a hydrophilic solvent in its International Patent Application No WO 00/40219 filed to protect Deximune. As Professor Eastoe pointed out, ethyl lactate is also lipophilic: it has a log P_oct value of 0.06, indicating a slight preference for octanol over water. Although Professor Lawrence considered that the log P_oct value was potentially misleading because lipophilicity depends on the oil, she accepted that ethyl lactate had some lipophilic character. I conclude that ethyl lactate can accurately be described as “hydrophilic”, at least in some systems.

Even assuming that ethyl lactate is hydrophilic, however, Dexcel disputes that ethyl lactate functions as the hydrophilic phase and that tricaprin functions as the lipophilic phase. Dexcel’s case is simple. It is common ground that ethyl lactate and tricaprin are miscible in the absence of other components. Accordingly, Dexcel contends that they do not fulfil the express requirement of non-miscibility set out in the second paragraph on page 15 of the specification (quoted in paragraph 31 above). Thus they do not constitute distinct phases.

At the hearing before Floyd J, counsel for Dexcel submitted that this was an issue of claim construction, and hence of law, which had already been concluded in Dexcel’s favour by the Court of Appeal on appeal from Pumfrey J in the Ivax case [2007] EWCA Civ 971. In his judgment Floyd J held that it was seriously arguable that the court trying this case would not be bound by the construction of the claims adopted in Ivax: [2008] EWHC 1266 (Pat), [2008] FSR 31 at [15]-[23]. Moreover, he also held at [33]-[34] that it was at least arguable that the issue of claim construction which arises here was not decided in Ivax. Before me, counsel for Dexcel did not contend that the issue had been determined in Ivax; but he did rely upon the reasoning of Pumfrey J and the Court of Appeal as being persuasive.

In the Ivax case Novartis contended that the lipophilic phase in Ivax’s Equoral product was constituted by polyglycerol-3-oleate (Pg3o). Pg3o is an amphiphilic surfactant. At first instance Pumfrey J held at [36] (emphases added):

In the Court of Appeal Jacob LJ held (emphasis added):

I agree with the provisional view of Floyd J that the issue which Pumfrey J and the Court of Appeal decided was whether it was necessary to have a lipophilic phase distinct from the surfactant at all, not when it was necessary to have lipophilic and hydrophilic phases present, and that it is the latter issue which arises in the present case.

As Novartis has proved by experiment, in an aqueous dispersion, ethyl lactate and tricaprin are largely immiscible and form two distinct phases, i.e. a tricaprin phase and an ethyl lactate-and-water phase. Furthermore, as Novartis has also proved by experiment, when Deximune is diluted 1:5 in water at 25^oC, the majority of the cyclosporin is associated with microemulsion-sized particles, i.e. it is carried by the dispersed phase.

Novartis contends that claim 1 does not require that there be distinct hydrophilic and lipophilic phases present in the pre-concentrate, and that it is sufficient for there to be hydrophilic and lipophilic components present in the pre-concentrate which will be in separate phases in the microemulsion. Novartis also argues that claim 48 merely requires that there be separate hydrophilic and lipophilic phases in the microemulsion, regardless of whether they are present in the pre-concentrate. Accordingly, Novartis contends that Deximune falls within claim 1 and in any event produces a microemulsion falling within claim 48.

In my judgment Novartis is correct, for the following reasons.

The starting point is that it is clear from the specification that the pre-concentrate need not itself be a microemulsion. Rather, the pre-concentrate must form a microemulsion when adequate water is added to it. Thus the specification states that a microemulsion pre-concentrate is “a system capable on contacting with, e.g. addition to, water of providing an oil-in-water microemulsion” (page 9). The microemulsion pre-concentrates of the invention are “capable of forming an oil-in-water microemulsion, spontaneously or substantially spontaneously on contact with water alone” (page 10). “Microemulsions [are] obtained on contacting the ‘microemulsion pre-concentrate’ compositions of the invention with water or other aqueous medium”, “to obtain a microemulsion, adequate water will be required”, “on contacting with water, the ‘microemulsion pre-concentrate’ compositions of the invention are capable of providing microemulsions” and “microemulsions obtainable from ‘microemulsion pre-concentrates’ of the invention, e.g. at dilutions as indicated above” (all page 11). The last paragraph on page 28 (quoted in paragraph 35 above) says the same thing. The patent contemplates in the middle of page 11 that the pre-concentrate “may contain minor quantities of water or otherwise exhibit fine structural features characteristic of microemulsions”; but equally it may not.

Given that the pre-concentrate need not itself be a microemulsion, no technical reason has been identified by Dexcel as to why the hydrophilic component and the lipophilic component must constitute distinct phases in the pre-concentrate as opposed to in the microemulsion. On the contrary, as Professor Lawrence pointed out and Professor Eastoe accepted, from the perspective of the formulator, it would be desirable to have a homogenous pre-concentrate.

Furthermore, as Professor Lawrence said and Professor Eastoe accepted, the skilled reader of the Patent would understand that, given that water will dominate the aqueous phase of the dispersion, what is critical is that the lipophilic phase should be immiscible with the hydrophilic phase on dispersion in water.

This reading of the Patent is entirely consistent with the purpose of the lipophilic phase identified by Jacob LJ, namely to carry the cyclosporin and to form the oil phase of the oil-in-water microemulsion. As Pumfrey J put it, the lipophilic phase is “destined to form the microemulsion”. It follows that I agree with the provisional view of Floyd J that:

In my judgment, Dexcel’s construction is flawed because it is based on linguistic analysis rather than consideration of the technical purpose of the elements of the claim. Dexcel relies on two main points in support of its construction. The first is that the word “phase” has a well-understood meaning which I have set out above. Counsel for Dexcel submitted that the patentee had deliberately chosen to describe the elements in question as phases, and not merely as components, solvents or substances, and that the skilled reader would consider that the patentee must have done so for good reason. The second is the passage stating that the phases must be non-miscible. Counsel for Dexcel submitted that the skilled reader would be bound to take this teaching at face value since he would know little about microemulsions. I am unimpressed by both arguments. The skilled reader would come to the Patent with the common general knowledge that I have set out above, and thus would know how an emulsion is formed. The specification tells him everything he needs to know about microemulsions and how they are formed. He would not disregard the patentee’s use of the word phase or the statement that the hydrophilic and lipophilic phases are non-miscible, but reading that statement in the context of the specification and having regard to the inventor’s purpose he would understand that non-miscibility was required in the microemulsion but not in the pre-concentrate.

This is reinforced by consideration of claim 48. If one takes a purely linguistic approach to it, this claim is technical nonsense, because it requires the presence of both a hydrophilic phase and water. Counsel for Dexcel submitted that the addressee would resolve this conundrum by concluding that the requirements for a hydrophilic phase and a lipophilic phase referred to the pre-concentrate. In support of this he relied upon the penultimate paragraph on page 28 (quoted in paragraph 37 above) as showing that the hydrophilic phase and the lipophilic phase were the same in both. In my view the addressee would reach the opposite conclusion. He would understand that what matters is that the microemulsion contains a hydrophilic component which is miscible with water and a lipophilic component which is non-miscible with both the hydrophilic component and the water.

The nearest Professor Eastoe came to providing a technical justification for Dexcel’s interpretation was in his second report, where he said:

He explained this to me as follows:

I am not persuaded by this evidence, for the following reasons. First, Professor Eastoe only advanced this justification in his second report, indicating that it had not occurred to him by the time he completed his first report. If it only occurred to Professor Eastoe second time round, I think it unlikely that it would occur to the addressee, who as noted above would know much less about microemulsions than Professor Eastoe, at all.

Secondly, Professor Eastoe went on to make it clear that in his view the Patent did exclude the second option. That view was based, first, on the use of the word “phase” and, secondly, upon the expression “microemulsion pre-concentrate”, which he interpreted as a microemulsion which had been pre-concentrated. Counsel for Dexcel accepted, however, that claim 1 is not limited to a pre-concentrate which is a microemulsion. Furthermore, claim 48 is clearly not limited to a microemulsion which is formed from a pre-concentrate that is itself a microemulsion.

Thirdly, at best Professor Eastoe’s evidence provides a reason why the skilled reader might think it was rational for the patentee to want to restrict the claims to a pre-concentrate which is a microemulsion, namely that such a pre-concentrate would be more certain to give a microemulsion when diluted with water than a pre-concentrate which was not a microemulsion. Once it is accepted that the claims are not so restricted, however, his reasoning supports the view that what matters is the behaviour of the hydrophilic and lipophilic phases after the pre-concentrate is diluted with water.

Finally, I would add two points. The first is that Novartis proved by experiment that the hydrophilic and lipophilic components specified in Examples 1.1, 1.2, 1.3 and 1.4 (see paragraph 38 above) were miscible in isolation. This evidence was not before the courts in the Ivax case. If Dexcel’s construction were correct, it would mean that those Examples, which include especially preferred embodiments of the invention, were outside the claims. That would be unusual and surprising. Counsel for Dexcel submitted that this was irrelevant because the specification does not tell the skilled reader this and he would not know from his own common general knowledge. In my judgment the skilled reader would be likely to find this out if he attempted to repeat the Examples. In any event, I consider that this is a technical fact which it is legitimate for the court to take into account at least to the extent of confirming a construction arrived at for other reasons.

The second is that in a carefully-reasoned judgment dated 2 July 2008 the District Court of The Hague held at paragraphs 4.64-4.72 that “the requirement of a separate hydrophilic and lipophilic phase has any meaning only in the microemulsion stage when it has come into water”, and therefore concluded that Deximune satisfied the corresponding requirements of the claims of the Dutch equivalent of the Patent.

It is common ground that (as Pumfrey J concluded in Ivax in the first sentence at [28]), since the addressee would know very little about microemulsions, he would have to rely upon the specification as a guide to characterise them. The key passages for this purpose are those on pages 9-10 (quoted in paragraph 27 above) and the last paragraph on page 11 (quoted in paragraph 29 above).

Counsel for Dexcel, supported by Professor Eastoe, submitted that a microemulsion has to possess all the characteristics listed in these passages, and that the most important is transparency. Counsel for Novartis, supported by Professor Lawrence, submitted that it was not necessary for a microemulsion to possess all of those characteristics, although it would be expected to exhibit many of them, and that the most important criterion is that of particle size.

Counsel for Dexcel also relied on what Pumfrey J went on to say at [28]:

With respect, I do not agree with this. As Professor Lawrence pointed out, the cited references are not consistent with each other in describing the characteristics of a microemulsion. Nor do they confirm the upper limit for particle size of 200 nm (i.e. 2,000 Å).

Rosoff states at page 405-406:

Friberg states at page 317:

Müller states in paragraph 2.1 (in the agreed translation from the German):

In my view the skilled reader of the specification would appreciate that, when the patentee said that microemulsions possess “one or more” of the following characteristics, he was reflecting a degree of uncertainty in the scientific literature at the time. Accordingly, the skilled person would understand that a microemulsion will generally display all the listed characteristics, but that he should not leap to the conclusion that a microemulsion is not present if one or more are missing. A judgment is required based on all the criteria. Moreover, the judgment must be one which takes into account the function of the microemulsion, namely to carry the cyclosporin (or at least most of it). As Jacob LJ put it at [16]:

This inevitably means that claims 1 and 48 have fuzzy edges. I do not accept, as counsel for Dexcel submitted, that it means that the claims are so unclear that they cannot be infringed. In support of this submission he cited The Procter & Gamble Co v Peaudouce (UK) Ltd[1989] FSR 180 at 198, Milliken Denmark A/S v Walk Off Mats Ltd[1996] FSR 292 at 300-301 and Scanvaegt International A/S v Pelcombe Ltd[1998] FSR 786 at 797. I consider that the specification is sufficiently clear to enable the question whether a composition is or is not a microemulsion to be posed, although answering the question in borderline cases may be difficult.

Before turning to the individual characteristics, I must deal with one more general point. The specification states on page 11 that “to obtain a microemulsion, adequate water will be required”. It goes on to say that “a dilution of 1:1, e.g. 1:5 p.p.w. (‘microemulsion pre-concentrate’: H₂O) or more will generally be appropriate”. I think the words “or more” have been misplaced in this sentence, and should follow “1:1”. In my judgment this sentence does not constitute a promise that the pre-concentrate will, still less a requirement that it must, produce a microemulsion at a dilution of 1:5, still less a dilution of 1:1. In the bioavailability study described on page 44 of the Patent, the composition was diluted 1:40. I consider that the question on infringement of claim 1 is whether the pre-concentrate is suitable for producing a microemulsion when it is diluted to the extent that occurs during use.

“Non-opaque or substantially non-opaque … i.e. transparent or opalescent”. It appears that the word “opalescent” derives from Müller. It was not a term of art in 1988. One of the meanings given in the Oxford English Dictionary is “milky iridescence”. Professor Lawrence equated it with “translucent”, while Professor Eastoe said that he would not use the word “opalescent”, but accepted that microemulsions which were close to their instability boundary and exhibited a blue tinge due to Tyndall scattering could be called translucent. In my judgment this passage in the specification clearly does not require the microemulsion to be transparent, it may be translucent.

Furthermore, as Jacob LJ pointed out in Ivax at [20], the skilled reader of the specification would understand that the composition as a whole may not be transparent or even translucent since it goes on to say that the composition may contain a thickener:

“Formed spontaneously or substantially spontaneously”. On the evidence, this characteristic is quite an important characteristic of microemulsions. It does not exclude the use of gentle agitation.

“Thermodynamic stability”. The reference to thermodynamic stability on page 9 has to be read together with the statement on page 11 that “microemulsions exhibit thermodynamic stability, that is they will remain stable at ambient temperatures … over periods of at least 2 hours, more preferably at least 4 hours, most preferably at least 12 to 24 hours”. From a strict scientific point of view, either a dispersion is thermodynamically stable or it is not; and if it is, it will be stable indefinitely. Professor Lawrence’s reading of the specification was that it did not require true thermodynamic stability, but stability over a period of at least 2 hours and preferably up to 24 hours. As she pointed out, this is consistent with the conditions in vivo for the formulations of the Patent, since the peak absorption takes place at about 1.5-2 hours. Professor Eastoe accepted that the author of the Patent was using the expression “thermodynamic stability” in a practical sense for formulating a composition, rather than in its strict scientific sense, although he considered that the aim was stability over 24 hours. In my judgment it is sufficient for the composition to be stable for at least 2 hours.

“Monophasic”. It is common ground that this refers to the composition having a single phase, or being homogenous, on a macroscopic scale.

“Optically isotropic”. It is common ground that this means that the composition is not birefringent.

“Particles which are of a size less than 2,000 Å”. Counsel for Dexcel pointed out that the specification does not instruct the reader to measure the particle size of the composition, let alone tell him how to do it. In my judgment it is implicit that the skilled person is expected to measure the particle size distribution of the dispersion. As to how, it was common ground between Professor Lawrence and Professor Eastoe that a well-known method in 1988 was light scattering, although there were other methods.

Professor Eastoe, having noted that there was no support in the cited references for the particle size limit of 2,000 Å, said this about it (footnote omitted):

I find this persuasive, except that as I have said I think that it is clear that the Patent does not insist upon transparency as a characteristic of microemulsions. It follows that, just as some haziness may be tolerated, so may the presence of a few larger particles. Thus I agree with the view apparently expressed in the following exchange during argument in the Court of Appeal in the Ivax case:

Given my construction of these expressions, Deximune does contain both a hydrophilic phase and a lipophilic phase.

Both Novartis and Dexcel carried out a number of experiments with a view to proving that Deximune does, or does not, form a microemulsion upon dilution. Before considering what these show, it is necessary to return to the question of what the appropriate dilution is.

The patient information leaflet for Deximune, states “Swallow the medicine whole with a small amount of water”. Professor Lawrence’s evidence, which I accept, was that this would lead to a dilution of around 1:20 even without taking into account the contents of the stomach, a dilution of over 1:30 for a fasted stomach and considerably more for a fed stomach.

This conclusion has important consequences when assessing the experiments, since Dexcel’s experiments were carried out at dilutions which did not exceed 1:10, and in some cases were less than that, whereas Novartis’ experiments included 1:20 and 1:50 dilutions. In my judgment these dilutions are more representative of the conditions of use than those employed by Dexcel.

A further point is that Novartis carried out experiments at 37^oC and in 0.1N hydrochloric acid as well as at 25^oC and in water. This was an attempt to mimic gastric conditions. Counsel for Dexcel submitted that this was contrary to the teaching of the Patent, which is that the composition should be assessed at ambient temperature and says nothing about employing acid conditions. Counsel for Novartis submitted that these conditions were appropriate for assessing whether a microemulsion would be formed in vivo as opposed to in vitro. I am inclined to agree with counsel for Novartis, but little turns on this because the experiments at 37^oC and in 0.1N hydrochloric acid did not produce very different results to those at 25^oC and in water.

“Non-opaque or substantially non-opaque … i.e. transparent or opalescent”. Novartis’ Experiment 1 was a visual assessment of Deximune at various dilutions and under the two sets of conditions mentioned above over a 24 hour period. In my judgment the photographs of the 1:20 and 1:50 dilutions show compositions which can properly be described as “translucent”, although they do display some haziness. Moreover, Professor Lawrence gave evidence, which I accept, that the photographs tend to exaggerate the degree of haziness of the samples. Although Professor Eastoe said that he would not call them translucent, he accepted that reasonable people could. I would not characterise the 1:5 dilutions as translucent, however. To my eyes these are substantially opaque.

Dexcel’s Experiment 1 was also a visual assessment at various dilutions over a 24 hour period. As already stated, I consider that Dexcel’s Experiment 1 was flawed in not including any dilution greater than 1:10. In addition, Professor Lawrence was critical of the design of the experiment, which involved taking photographs against a black background and under varying light conditions. I agree with these criticisms: I found the photographs of the repetitions, at least as reproduced in the trial bundle, much less helpful than the photographs from Novartis’ Experiment 1. Professor Lawrence’s evidence, which I accept, was that when she attended the repetitions she was able to see black lettering through the dispersion in the 1:5 and 1:10 dilutions. Having considered the photographs both of the original experiments and of the repetitions in the light of the expert evidence, I am satisfied that the 1:10 dilutions can just about be described as translucent.

I conclude that, when diluted to the extent that it is in use (namely, at least 1:10), Deximune forms a dispersion that is substantially non-opaque. This is consistent with Dexcel’s own description of Deximune in Example 10 of its PCT application as forming an “almost clear” dispersion.

Dexcel’s Experiment 5 was an assessment of Deximune diluted at 1:1 and 1:3 dilutions by means of light microscopy. This experiment is even less representative of the in use conditions than Dexcel’s Experiment 1, but it is of some assistance. At 1:1 dilution, one sees a number of larger liquid droplets which look as though they have a tendency to cluster. At 1:3, the droplets have become smaller and much more widely spread. Professor Eastoe accepted these photographs showed a fairly material reduction in the size of the large particles. His evidence was that one could not necessarily extrapolate that to greater dilutions since one only had two points through which to draw a line. Nevertheless, these photographs are consistent with the size and number of the larger particles reducing as one increases the dilution.

Professor Lawrence’s evidence was that, based on all the experiments, she believed the reason why the visual assessments of the samples showed the degree of cloudiness that they did was that they contained a small quantity of particles of size larger than 2,000 Å which were responsible for the slightly cloudy appearance, in addition to a microemulsion. I found this evidence persuasive.

“Formed spontaneously or substantially spontaneously”. There is no dispute that the dispersions formed by Deximune are formed substantially spontaneously.

“Thermodynamic stability”. Novartis’ Experiment 1 shows a slight increase in the cloudiness of the 1:50 dilution over 24 hours, but this is not noticeable at 2 hours or 4 hours. In the case of the 1:20 dilution it is hard to see any difference even at 24 hours. Likewise, it is hard to see much difference in the 1:10 dilution in Dexcel’s Experiment 1 over 24 hours.

Novartis’ Experiment 2 was a size analysis of dilutions of Deximune by means of dynamic light scattering (DLS), also known as photon correlation spectroscopy (PCS), under the same conditions and over the same time periods as Novartis’ Experiment 1. Professor Eastoe accepted that some of the samples showed stability. I consider that the results for the 1:20 and 1:50 dilutions show considerable stability up to 24 hours.

Dexcel’s Experiment 4 is an NMR experiment on a 1:1 dilution of Deximune over a 24 hour period. Professor Eastoe’s evidence was that the experiment showed that the dispersion was not stable. Dr Griffiths disagreed with this, and suggested that there were two alternative interpretations of the results. In the event counsel for Novartis did not cross-examine Professor Eastoe on the alternative interpretations, because Professor Eastoe accepted that the results from the 1:1 dilution experiment could not be extrapolated even to a 1:2 dilution, let alone greater dilutions. It is therefore unnecessary for me to express any view on what, if anything, the 1:1 dilution experiment shows.

I conclude that the evidence demonstrates that the 1:10, 1:20 and 1:50 dilutions are sufficiently stable to qualify as thermodynamically stable as interpreted above.

“Monophasic”. Novartis’ Experiment 1 and Dexcel’s Experiment 1 both suggest that the dispersions of Deximune are monophasic.

Dexcel’s Experiment 6 consisted of centrifugation of 1:1, 1:3 and 1:5 dilutions of Deximune and Neoral. In each case the result was a deposit (described as a “pellet” by Professor Eastoe and a “thin film” by Professor Lawrence). Dexcel contends that this experiment shows that Deximune produces a dispersion which is biphasic rather than monophasic. Again this experiment is flawed in not examining dilutions greater than 1:5. In any event Professor Lawrence’s evidence, which I accept, was that she believed the deposit to be due to the small fraction of larger particles present in the dispersion. Counsel for Dexcel relied on Professor Lawrence’s acceptance that the skilled person would not have had experience of microemulsions coexisting with larger particles; but it was not put to her that the lack of such experience would lead the skilled person to interpret the experimental evidence in a different way to that which she had done on the basis of her (admittedly post-1988) experience.

“Optically isotropic”. There is no dispute that the dispersions formed by Deximune are optically isotropic.

“Particles which are of a size less than 2,000 Å”. Novartis’ Experiment 2 shows that in the 1:20 and 1:50 dilutions 100% by number and from 96% to 99.4% by volume of the particles are contained in a sharp peak centred at from 17.29 nm to 33.66 nm. The particles in this peak do not exceed 100 nm, let alone 200 nm.

Dexcel relied upon the fact that the instrument used in the Novartis experiment, the Malvern Zetasizer Nano ZS, did not exist in 1988. I am not impressed by this point. The basic principles upon which the machine operates were well-known at that date, although some operational refinements have been introduced over the years. Furthermore, the same instrument was used by Dexcel’s expert at the time of the interim injunction application, Dr Müllertz. Still further, the results obtained by Novartis are consistent with the average particle size of 25-50nm (obtained by a different method) for Deximune dispersions reported in Example 10 of Dexcel’s PCT application.

Dexcel also relied on the fact that Professor Lawrence was personally unfamiliar with all the details of how the machine operated. Again, I am not impressed by this. She was familiar enough with it to be able to interpret the results, as was Professor Eastoe.

More substantially, Dexcel relied on the evidence of Professor Eastoe that the samples in question were too turbid to permit accurate light scattering analysis, particularly if one were restricted to 1988 technology. (Indeed, Professor Eastoe said that he had planned to conduct light scattering experiments, but abandoned the plan when he saw how cloudy the samples were. Of course, he was looking at samples which were only diluted at up to 1:10.) I was troubled by the fact that some of Professor Eastoe’s reasons for expressing this opinion only emerged in re-examination. Be that as it may, my conclusion having reviewed the evidence of both experts is that (i) the difficulties would not have been insuperable using 1988 technology at the higher dilutions and (ii) there is no material doubt about the accuracy of the measurements obtained during Novartis’ experiments, which it may be noted were both consistent and repeatable.

Overall, I find that when diluted to 1:20 or more, Deximune forms a microemulsion within the meaning of the Patent. Although the evidence is slightly less clear cut, on balance I consider that Deximune forms a microemulsion when diluted at 1:10 as well. Thus Deximune does produce a microemulsion in the conditions of use.

I would add two points. First, my conclusion is consistent with the fact that Dexcel itself described Deximune as forming a microemulsion in its PCT application, in a corresponding US patent and in a European application. Professor Domb, the inventor, gave evidence that this was an error on someone else’s part which he had failed to spot. Be that as it may, it appears that someone else thought that it was correctly described as forming a microemulsion. Furthermore, Deximune was also described as forming a microemulsion in a Public Assessment Report issued by the Danish Medicines Agency which was provided in draft to Dexcel to comment on.

Secondly, the Dutch court concluded at paragraphs 4.73-4.91 of its judgment that the product did not literally infringe the corresponding requirement in the claims of the Dutch patent because the product contained a small proportion of particles whose size exceeded 2,000 Å, but infringed on the ground of equivalence since the product exhibited only insubstantial differences from the claimed composition. The court’s factual assessment is very similar to mine, although its legal analysis is slightly different. Both lead to the same conclusion.

The marketing of Deximune would infringe claims 1 and 48 of the Patent.