Leo Pharma (a/s Leo Laboratories Ltd) v Sandoz Ltd

As every schoolboy or girl knows, some chemical compounds form crystals: regular lattices of their constituent molecules. Crystalline materials are well known to be capable in some cases of forming crystal hydrates: arrangements in which one or more molecules of water per molecule of the compound are incorporated into the crystal lattice. This case is about a patent for a crystalline hydrate of a known vitamin D analogue, calcipotriol, and its use in a topical preparation for the treatment of psoriasis.

This judgment is arranged into the following sections:

Leo Pharma A/S, the first claimant, is the proprietor of European Patent (UK) No. 0 679 154 granted in respect of an invention entitled “New crystalline form of a Vitamin D analogue” (“the Patent”). Leo Laboratories Limited, the second claimant, is the exclusive licensee under the Patent. For most purposes it is not necessary to distinguish between them: I will refer to them both as “Leo”. The Patent has a priority date of 15^th January 1993. The Patent claims the monohydrate form of calcipotriol and pharmaceutical preparations containing that form.

The defendant, Sandoz Limited (“Sandoz”), markets an ointment and cream containing calcipotriol. An interim injunction was resisted by Sandoz, inter alia on the ground that there was no arguable case of infringement. Sandoz maintained that its cream contained only de minimis amounts of monohydrate – not enough to give the benefits of the invention. Mann J granted the injunction: see [2008] EWHC 541 (Pat). Although infringement has been in issue throughout, Sandoz now concedes infringement. Its only case now is that the Patent is invalid on the grounds of lack of novelty and obviousness. It counterclaims for revocation.

The parties are in dispute over the Patent elsewhere. The Dutch court, the District Court at the Hague (Judges Kalden, Van Peursem and Van Walderveen), has rejected Sandoz’ invalidity attack in a judgment dated 11^th February 2009. During the trial I was told that a German court had held the patent invalid on the basis of an obviousness attack different to the one pursued before me. I have not been provided with a copy of the German judgment. So, at the moment, it is one-all. If there were a supra-national court of the kind currently proposed, able to decide on disputes about the validity of patents for the whole of Europe, conflicting results of this kind between courts of first instance could be avoided.

Here, Mr Henry Carr QC argued the case for Leo with Mr Andrew Lykiardopoulos. Mr Antony Watson QC argued the case for Sandoz with Mr Piers Acland.

The Patent is not a complicated document. It explains that calcipotriol is a known vitamin D analogue which has proved to have very useful properties in the topical treatment of psoriasis: page 2 lines 5-9. The Patent explains that the manufacture of calcipotriol is described in an earlier PCT application WO 87/00834 (“the 834 application”).

The Patent states the advantages of the invention: superior technical properties in the manufacture of crystal suspension formulations and superior stability properties: page 2 lines 3-4.

At page 2 lines 10-11 the specification states that, due to the poor stability of calcipotriol in certain solutions, it is preferred to use crystal suspensions in creams and gels. This statement reflects Leo’s experience with its solution formulation of calcipotriol, which proved to have inadequate shelf life: but the statement would not have formed part of the common general knowledge.

The patent goes on to explain that, in the manufacture of crystal suspensions, a mechanical size reduction process such as micronisation or wet milling is necessary. The patent then says this at page 2 lines 16-23:

“In the case of calcipotriol a wet ball milling process has been used. However, it has turned out to be technically difficult to perform this process when using the anhydrous crystal form described in [the 834 application]. These crystals are not easily wetted and during the milling process they develop a stable foam which results in difficulties in obtaining a suitable small and uniform particle size.

It has now surprisingly been found that these technical problems can be avoided when a hitherto unknown crystalline form of calcipotriol, i.e. calcipotriol hydrate, is used instead of the anhydrous form. The hydrate is technically superior to the anhydrate; it is easily wetted and the wet ball milling process is running smoothly.”

Again, these statements reflect Leo’s experience, rather than the common general knowledge.

The hydrate is a monohydrate, that is to say it contains one molecule of water per molecule of calcipotriol. It can be made by dissolving calcipotriol in ethyl acetate or acetone and adding water. The process of creating the monohydrate directly in this way is illustrated in Examples 1 and 3. Example 2 illustrates a process for making it which involves seeding with monohydrate crystals to initiate crystallisation. Examples 4 and 5 illustrate the incorporation of calcipotriol hydrate into a cream and into a gel.

Of the five claims, each of claims 1 to 4 is said to have independent validity. Claim 1 is to the compound calcipotriol monohydrate itself. Claim 2 is to a “pharmaceutical composition containing” calcipotriol monohydrate. Claim 3 is a suspension cream and claim 4 a suspension gel, each containing calcipotriol as the monohydrate.

There are no disputes as to the proper construction of these claims. Both sides recognise that claim 1 extends to the monohydrate wherever it is found. Likewise, both sides recognise that the subsidiary claims, by referring to the monohydrate, must be requiring its presence in the relevant pharmaceutical formulation in crystalline form.

Sandoz contends that the Patent is invalid because (a) it is anticipated by a PCT Application with publication number WO 91/12807 and (b) because it is obvious in the light of that document (or alternatively the 834 application) and common general knowledge.

Each side called two expert witnesses. Neither side made any significant criticism of the manner in which these experts gave their evidence. That reflected my own view, which was that all the experts were doing their best to assist me.

Leo called Dr Poul Rasmussen and Dr Fritz Blatter. Dr Rasmussen was, in January 1993, Executive Vice-President of the Research and Development Department of Leo and a member of the Executive Management Committee. From 1984 the department in which he worked specialised increasingly in vitamin D compounds. Dr Rasmussen became Chairman of the Board of the Leo Foundation, which is the owner of the first claimant, in 2000, retiring in 2007. He was retained as an expert in July 2008. Thereafter he was re-appointed Chairman of the Board in November 2008 on the unexpected retirement of his successor.

Sandoz was naturally sceptical at the outset as to whether Dr Rasmussen would be able to give evidence unaffected by his past and continuing affiliation to Leo: but in the end, it was plain to everyone that Dr Rasmussen was tackling the task of acting as expert with a proper understanding of the duty he owed to the court, and discharging that duty appropriately. He was an impressive and very fair witness, making appropriate concessions even when he knew these were adverse to Leo’s case, whilst maintaining what were plainly his genuinely held and well reasoned opinions.

Dr Blatter joined Ciba-Geigy AG as Head of Laboratory in 1996, joining a group specialising in polymorphism and molecular crystals. He had no exposure to polymorphism before this date. He now works for Solvias AG and specialises in projects related to chemical development and analysis of new drug substances. Much of his evidence went to issues of non-infringement which were not pursued.

Sandoz called Professor Christopher Frampton and Professor Adrian Williams. Professor Frampton is a specialist in crystallography and solid state chemistry. Between 1992 and 2001 he was employed by Roche working on, amongst other things, solid state pharmaceutical development and analysis. More recently he has become the Chief Scientific Officer at Pharmorphix Limited, a company which specialises in polymorph screening. In assessing his evidence I have to bear in mind that he is a specialist in screening, and may have approached matters with a predisposition to pursuing screening programs which the relevant skilled team might not possess.

Professor Williams is the Professor of Pharmaceutics and Head of Department at the School of Pharmacy, University of Reading. His principal area of research is in the field of topical and transdermal drug delivery. His role in the case was in the attempt to repeat one of the examples in the prior art for the purposes of the anticipation attack.

The Patent is addressed to those with an interest in manufacturing vitamin D analogues: and, at least so far as the subsidiary claims are concerned, preparations of such analogues for pharmaceutical use. However, the category of skilled persons so defined is not limited to those with previous extensive experience of research into vitamin D analogues: the Patent is plainly intended to be addressed as much to such persons as to others, who may be entering the field for the first time. In the latter case, of course, the addressee would need to familiarise him or herself, so far as necessary, with the common general knowledge in the field.

The task of making a new API, or new pharmaceutical formulation would be given to a team made up of members from different disciplines. Clearly, in the present case, the team would include a medicinal chemist and a formulator.

The skilled team would need to be equipped with knowledge of what was generally known about vitamin D analogues and pharmaceutical preparations containing them. To the extent that the skilled team did not include a vitamin D expert, it would familiarise itself with the generally accepted teaching from standard reference works such as Martindale, Kirk-Othmer and the British Pharmacopoeia.

The skilled team would know as part of its common general knowledge that vitamin D compounds were biologically potent, and sensitive to light, oxygen and heat.

The skilled team would also be aware of the vitamin D preparations that were already on the market at the priority date. It would know that all such preparations, whether for systemic or topical use, contained the vitamin D analogue in solution.

The skilled team would know of calcipotriol, and that it had been used in a solution ointment called Dovonex marketed by Leo for the treatment of psoriasis.

The common general knowledge would include the basics of crystallisation as follows. A compound is said to exist in polymorphic form if it crystallises into more than one arrangement in the crystal lattice. Some compounds form crystals into which, in addition to the compound itself, a fixed proportion of solvent is incorporated as well. These are known as solvates, or hydrates if the solvent is water. If one molecule of water is included per molecule of the compound, it is called a monohydrate; if two it is called a dihydrate and so on. Solvates and hydrates are sometimes referred to as pseudopolymorphs, to distinguish them from true polymorphs, which are the solvent-free forms.

It was well known and accepted that different crystal forms of the same compound (whether true polymorphs or solvates or hydrates) can have different physicochemical properties. These properties include solubility, dissolution rate, stability and processing characteristics. In consequence the bio-availability of the drug can be dependent on the crystal form as well.

In general, crystallisation is an empirical process. It is not possible to predict in advance whether a particular compound will crystallise, or whether it will form an amorphous solid. A given compound may fail to crystallise even where similar compounds are known to crystallise. If a given compound crystallises in one way, it is not possible to predict what if any other forms it might exist in, or whether it will form solvates. Inducing a compound to crystallise for the first time, or into a new crystalline form, involves choosing the correct conditions which themselves cannot be predicted in advance. Seemingly trivial matters such as impurities, may influence the result.

Leo contend that the skilled team would be aware that vitamin D compounds exist in dynamic equilibrium with the corresponding pre-vitamin, making them particularly difficult to isolate and purify, particularly given the sensitive nature of the compounds. Dr Rasmussen suggested that, in the light of these difficulties, the approach of those working in the vitamin D field was to stick with the first crystal form discovered, and not deliberately to seek out alternative forms.

I accept that vitamin D analogues presented some difficulties in crystallisation. Thus, hundreds of known vitamin D compounds were characterised as amorphous. To some extent this is due to the fact that these compounds were being described at an early stage of their development. Nevertheless, I accept Dr Rasmussen’s evidence that crystallisation of vitamin D compounds presented non-trivial problems. It is important not to get these out of proportion however: the evidence establishes that if a cogent rationale existed for performing crystallisation experiments, the skilled team would not be deterred from conducting them.

Leo also contended that the skilled team would be aware that hydrates were rare amongst vitamin D analogues and that true polymorphism amongst vitamin D analogues was unknown. They relied on a database of known vitamin D analogues compiled for Chemical Abstracts to show this.

Sandoz counter by saying that a significant proportion (more than 50%) of the known commercial vitamin D analogues were hydrates, and that polymorphism amongst drug molecules generally was very common, even if unknown for vitamin Ds.

In my judgment the skilled team would approach vitamin D analogues in the knowledge that the formation of hydrates was a possibility, but without any pre-conceived notion of its likelihood. I reach this conclusion based on the fact that the skilled team would be aware of the formation of hydrates both generally and in the case of the commercial vitamin D analogues. But in the case of any individual vitamin D analogue, the skilled team would not be able to predict whether it would form a hydrate or not: it could only attempt to establish the answer by experiment.

Similarly, the skilled team would know as part of its common general knowledge that polymorphism was a possibility, even if it had not been observed in vitamin D analogues to date.

I think it is artificial to suppose that the skilled team would rely too heavily on the database of vitamin D analogues to determine whether to look for solvates and polymorphs. In an empirical field, if the skilled person thought it relevant or worthwhile to consider alternative crystal forms and was sufficiently motivated to do so, he or she would be far more likely to conduct a crystallisation experiment than to decline to do so based on a review of prior compounds, which would be a poor guide to future crystallisation performance.

A polymorph screen involves crystallising an API from a variety of solvents and solvent mixtures and characterising the resulting crystals for evidence of polymorphism or pseudopolymorphism.

Sandoz contends that it was common general knowledge in 1993 that it was advisable and to an extent mandatory to carry out a polymorph screen in respect of an API which one hoped to market.

Leo contend that, whatever may have been the practice elsewhere in the pharmaceutical industry, it was not the practice in the vitamin D field to go out and look for polymorphs or conduct polymorph screening. Instead, those in the industry merely kept a watch on their process and storage conditions for signs of polymorph formation.

It was plainly not the universal practice to conduct a polymorph screen. Leo did not do anything which could properly be described as polymorph screening, and Dr Rasmussen said that he was not aware of any of his competitors doing so. Professor Frampton said it would be dangerous to assume that other companies had necessarily done such a screen.

Nevertheless, well before 1993, standard formulation textbooks were teaching with varying degrees of emphasis that, in the preformulation stage of pharmaceutical development, one should actively look for polymorphs. For example in Lachman & Lieberman “The Theory and Practice of Industrial Pharmacy” published in its third edition in 1986, the authors state:

“During pre-formulation, it is important to identify the polymorph that is stable at room temperature and to determine whether polymorphic transitions are possible within the temperature range used for stability studies and during processing (drying, milling, etc.). … To screen for additional polymorphic forms of a particular drug, bridging solvents, supersaturated solutions, supercooled melts and sublimination have proven useful.”

It is noteworthy that the authors distinguish between studying the possibility of transition within the processing and storage conditions (described as “important”), and screening for additional polymorphic forms (which is said to have “proven useful”). Another general text, “Pharmaceutics: the Science of Dosage Form Design” by Michael Aulton, published in 1988, goes rather further. He teaches that, prior to the development of any major dosage form:

“it is essential that certain fundamental physical and chemical properties of the drug molecule and other derived properties of the drug powder are determined. This information will dictate many of the subsequent events and possible approaches in formulation development. This first learning phase is known as pre-formulation.”

Under the heading “polymorphism” within the pre-formulation chapter, Aulton indicates that one should ask:

“(a)

How many polymorphs exist?

(b)

How stable are the metastable forms?

(c)

Is there an amorphous glass?

(d)

Can the metastable forms be stabilized?

(e)

What is the solubility of each form?

(f)

Will a more soluble form survive processing and storage?”

In a review article in 1990, Borka and Haleblian wrote:

“The discovery of polymorphism among pharmaceutical substances … initiated a growing interest in this field. The synthetic and analytic departments of leading pharmaceutical companies nowadays carry out systematic work to detect polymorphism of their drugs and to find intelligent applications of this phenomenon. Drug registration documents submitted by leading pharmaceutical companies to regulatory bodies will today, almost without exception, have a section on polymorphism when describing the physico-chemical properties of the active substance. The statement that the substance exhibits no polymorphism is equally important as exhibiting polymorphism. After the discovery of the first cases of polymorphism with dramatic differences in biological activity between two forms of the same drug (e.g. chloramphenicol palmitate) no pharmaceutical manufacturer could neglect the problem.” (emphasis supplied)

The passage emphasised in the extract from Borka is important. The immediate objective in looking for alternative crystal forms is not research conducted solely or even primarily to find something better. The objective is to establish that when you make and sell the product it is not going to change into something different. That is why establishing that polymorphs or pseudopolymorphs do not form is as important as establishing that they do. Nevertheless, as we shall see when we come to the relevant regulatory guidelines, what is said to be important is to demonstrate stability under the actual process and storage conditions. To look more widely for polymorphs would not be regarded by the skilled team as mandatory.

The Borka review article also included a long but not exhaustive list of various pharmaceuticals which had reported crystal polymorphism. It is fair to point out, as Leo do, that none is a vitamin D compound.

I conclude that the skilled team would know in 1993, as part of its general knowledge, that full polymorph screening was an available step to take in the pre-formulation phase of the development of a dosage form. That does not mean that it would be obvious to conduct such a full polymorph screen in all circumstances. The skilled team would consider in any individual case how difficult such experiments were, and the prospects of finding useful results. However, the skilled team would appreciate the importance of checking for the possibility of polymorph or solvate formation under the actual process and storage conditions proposed.

Both the European and the United States regulatory authorities issued guidelines on the information needed for registration of pharmaceuticals.

Volume III of “The Rules Governing Medicinal Products in the European Community” contains detailed guidelines on the quality, safety and efficacy of medicinal products for human use. In the 1989 edition there is a section giving guidance on the type of information required for the control of new active ingredients. Under the heading “Development Chemistry” there appears the following statement:

“4.1

Evidence of chemical structure

A scientific discussion of the chemistry of the active substances molecule should be given and should include, where applicable, unequivocal proof of structure, configuration, conformation and potential isomerism….

4.2.3

Polymorphism

Where relevant, the presence of polymorphic forms and the methods of detection and control should be discussed, or their absence confirmed.” (emphasis supplied)

Leo argued that the words “where relevant” limited the generality of the guidance. I agree that these words limit the application of the guidance to cases where it is plausible that polymorphs might be formed under the processing and storage conditions used. The same would be understood to apply to solvates and hydrates.

The Guideline for Submitting Supporting Documentation in Drug Applications for the Manufacture of Drug Substances issued in 1987 by the FDA contained the following in relation to solid-state drug substances:

“By the time of an NDA [New Drug Application] submission, the applicant should have established whether (or not) the drug substance exists in multiple solid-state forms, whether these affect the dissolution and bioavailability of the drug product, and whether particle size is important for dissolution and bioavailability of the drug product. It is not necessary to “create” additional solid-state forms by techniques or conditions irrelevant to the synthetic process.

The applicant should provide information describing how and why it has been concluded that

(a)

a change in the solid-state form does not occur when the drug substance is manufactured and stored according to the NDA directions; or

(b)

different forms occur but do not result in a bio-availability problem; or

(c)

polymorphism, solvation, or particle size has an important effect on bioavailability….

II.G.1 Polymorphism

Some drug substances exist in several different crystalline forms (“polymorphs”), due to a different arrangement of molecules in the crystal lattice, which thus show distinct differences in their physical properties. The same drug substance may also exist in a noncrystalline (amorphous) form. These various forms differ in their thermodynamic energy content, but not in composition. One of the critical factors affecting polymorphism (or solvation) is the choice of final solvent and isolation conditions in the synthesis…

II.G.2 Solvation (including hydration)

Conditions used in manufacture and/or storage of the drug substance may result in the isolation or formation of a solvated or hydrated drug substance. …” (emphasis supplied)

Dr Rasmussen again pointed to the words I have emphasised in the above extract. Again I accept that these words limit the generality of the work that needs to be done. The same words make it clear that where it is plausible, having regard to the product and process, that polymorphs might be formed, the necessary investigations should be undertaken and the data supplied. Paragraph II.G.2 makes particularly clear that one should think about whether the conditions used in the storage of the drug will lead to the formation of a hydrate.

The European and United States guidance would have been matters which would have been well known to the skilled team in 1993. The documents indicate the studies that the skilled team would know are likely to be required by the regulator: it does not necessarily follow that wider studies might not, in some circumstances, be an obvious step to take.

Wet milling was used in the manufacture of suspension formulations. It would be used for a number of reasons:

Sandoz contends that the Patent is anticipated by the disclosure of Example 4 of a PCT Application with publication number WO 91/12807. Because it relates to the use of vitamin D analogues for the treatment of acne, WO 91/12807 was referred to as “the acne use patent”.

The law of novelty is stated by Lord Hoffmann in Smithkline Beecham plc’s (Paroxetine Methanesulfonate) Patent [2005] UKHL 59, [2006] RPC 10. In summary, and for present purposes:

In the present case the relevant subject-matter contained in the prior art document is said to disclose the invention of the Patent, not because it describes calcipotriol monohydrate in terms (which it does not), but because it is said that carrying out the disclosure will inevitably produce the monohydrate. This type of disclosure can be called disclosure by inevitable result.

It is not necessarily fatal to a case of disclosure by inevitable result that it is possible to carry out the teaching of the prior art in more than one way. However, where that is the case, the court must be satisfied that each way of performing the invention would, on the balance of probabilities, produce the same result.

The acne use patent was published on 5^th September 1991 and claims the use of a range of vitamin D analogues for use in the treatment of acne, both systemically and by topical application.

At page 3 line 34 to page 4 line 25, the acne use patent gives examples of vitamin D analogues considered suitable for use in pharmaceutical preparations of the invention. One category of these is the class of compounds described in the 807 application:

“in particular the compound designated MC 903 (example 5 in the [807 application]) (confer also Calverley, M., Tetrahedron 43, 4609-4619 (1987);”

There is no dispute that the reference here to MC 903 is to calcipotriol anhydrate. So much is clear from the publications which are cross-referred.

The examples of the acne use patent include three cream solution examples, Examples 1-3. Two of these use MC903. Example 5 is a solution lotion, which is used in Example 7 in a clinical study, said to show encouraging results. Example 6 is a suspension in a capsule for systemic use.

Example 4, which is of particular relevance for this case, is a suspension cream. The example teaches the manufacture of an emulsion to which an aqueous suspension of MC903 is added. The preparation of the aqueous suspension includes the instruction:

“Mill MC903 to particle size below 5 μm and suspend in an aqueous solution of disodium hydrogenphosphate”.

The parties dispute whether the instruction to “mill” is a disclosure of dry milling or wet milling. Leo contend that, as the calcipotriol would be in the form of a dry powder, the instruction to mill it is an instruction to mill in the dry state. If wet milling were meant, the next part of the phrase “suspend in an aqueous suspension” would not be apposite: the wet milling would already have created an aqueous suspension, and the appropriate instruction would be to mix the two.

The disclosure is certainly not a disclosure of wet milling. The only real possibilities are that the skilled person would understand that the disclosure is silent on the topic of milling, leaving it open to the skilled person to choose the method of milling or that it is teaching dry milling specifically. On balance, I think Leo is right about the disclosure. There is enough of an indication in the subsequent part of the phrase for the skilled person to understand that dry milling is intended. Wet milling might be an obvious variant: but it is not disclosed.

Sandoz’ case is that although the skilled person would carry out Example 4 of the acne use patent using anhydrous calcipotriol as the starting material, the monohydrate would inevitably form in the course of manufacture or storage. They seek to demonstrate this in two ways. Firstly, Sandoz relies on experiments carried out under the supervision of Professor Williams. The experiments are a purported repetition of Example 4 of the acne use patent. Secondly, Sandoz says that, even if the experiments do not demonstrate that the inevitable result of carrying out Example 4 is calcipotriol monohydrate, the effect of other factual and expert evidence in the case is that it is so established. Mr Watson called this “experiment-free anticipation”.

Sandoz’s Notice of Experiments carried out a purported repetition of Example 4 of the acne use patent and yielded calcipotriol monohydrate. On that basis, Sandoz contended that claims 1, 2 and 3 were anticipated. The raw material used was taken from the batches of material used by Sandoz in its commercial production.

Leo, through their experts, launch two main attacks on this experiment. Firstly, they say, there has been no attempt to carry out Example 4 of the acne use patent according to its teaching. That teaching involved making calcipotriol anhydrate by the method of Example 5 of the 834 application or Calverley, which are cross-referred for that purpose. Not only have Sandoz not done this, but the method by which the raw material was made was not part of the Notice of Experiments. So Leo have not had the opportunity of witnessing how it was made. Different methods of making the raw material will create different impurity profiles, which could have an effect on the eventual outcome of the experiment.

Secondly Leo say that, far from seeking to mimic as closely as possible the teaching of the 834 application, the raw material for these experiments came from the same plant which produced a number batches of calcipotriol which contained significant quantities of monohydrate. Moreover, the batches were produced at points in time close to the time when the contaminated batches were produced. This, they say, makes it likely that the batches of calcipotriol used contained trace quantities of monohydrate which acted as seeds for the growth of monohydrate crystals. The presence of seeds would plainly make a difference to the outcome of the experiment. The experiment does not show what would happen in the absence of seeds.

I deal with the second objection first, because in my judgment it is dispositive of this way of putting the anticipation attack. It is not in dispute that if seeds of the monohydrate are present in the starting material, they would have the capacity to cause a conversion of the product to monohydrate in the aqueous environment of the cream. Seeds act as a template for further crystal growth. If seeds were present, one cannot tell for certain that one would have got the same result in the absence of seeds. Ultimately the point gives rise to a dispute of primary fact: were seeds present in the calcipotriol starting material, or was it pure anhydrate?

There is direct evidence of the analysis of the starting material. But the analysis techniques used have a limit of detection. So the analysis does not prove that monohydrate was absent. It could have been present in amounts below the limit of detection. The limit of detection is said by Sandoz to be 0.7%; but it could be higher or lower. The precise level does not matter because seeding could be caused by tiny amounts.

The batches used for the Notice of Experiments were produced by Teva in June 2008. Batches produced by Teva in April, May and June 2008 contained between 5 and 15% monohydrate. Two monohydrate-containing batches were made just prior to the batch used for the experiments in the Notice. Another was made between the batches used for the Notice and the repetition of the Notice in the presence of Leo’s representatives.

Dr Rasmussen’s view in his report was expressed as follows:

“Sandoz have accordingly admitted that some batches of API used in their commercial product contained monohydrate. Clearly, the amount of monohydrate in these batches was above the limit of detection … Any pharmaceutical manufacturing process will produce chemical impurities and, where appropriate, physical impurities (i.e. different crystal forms). The amount and nature of the impurities produced depend upon the manufacturing process used. A manufacturing process for calcipotriol anhydrate would not necessarily produce any monohydrate. However, if you do detect monohydrate in one or more batches, this indicates that the process does produce monohydrate, and it is likely that monohydrate will always be present, albeit below the limit of detection in some cases. Accordingly, it cannot be assumed that where no monohydrate is detected, it is not present at all. …

I believe that the finding of monohydrate in the cream prepared in Sandoz’ experiment can be explained by the presence of monohydrate crystals in the starting material.”

In cross-examination Dr Rasmussen explained that, based on his experience, the fact that one batch comes out without the level of impurity above the level of detection does not really tell you that you have eradicated the impurity.

Professor Frampton said in his first report that it was indeed possible that there were seeds in the material used in the experiment. He accepted that to get to the bottom of the matter you would have to look at the batch records: but these were not available.

I hold that there were seeds present in the starting material used for the Sandoz experiments. The evidence as to proximity of the experimental batches to batches which contained monohydrate, coupled with Dr Rasmussen’s evidence, is sufficient to make it more likely than not that the batch used for the experiment did contain at least some trace amounts of the impurity, in the absence of evidence to the contrary. Sandoz could have made the MC903 itself and demonstrated the precautions it took to eliminate seeds. It was established that a “clean room” experiment was at least a possibility. By contrast there was nothing more that Leo could have done to establish the presence of seeds in Sandoz’ material, given that there was also likely to be a limit of detection on its equipment as well.

An experiment with seeded MC903 is not a repetition of Example 4. The variant on Example 4 is one which it is established has the ability to affect the result. It follows that the anticipation attack based solely on the experimental repetition of Example 4 of the acne use patent fails.

It is not therefore strictly necessary to deal with the first of Leo’s’ objections, which raises a rather more finely balanced question. Example 4 of the acne use patent teaches the skilled reader to use MC903. MC903 is a shorthand for a particular Vitamin D analogue, identified by reference to the characteristics given in the earlier 834 application and the Calverley paper.

Leo’s position is that Example 4 should be carried out with calcipotriol made precisely by the Calverley process. Anything else is not a reproduction of Example 4. Sandoz submits that Example 4 can be carried out with any pure source of MC903, whether made by the Calverley process or not.

I think Sandoz is correct: the skilled person would not understand from Example 4 that only the Calverley process is to be used. The reference to the 834 application is to identify the anhydrate compound, rather than its method of manufacture.

However the consequence of that approach to the disclosure of Example 4 is that it covers a range of processes for making MC903, all of which will have different impurity profiles. As it is Sandoz who asserts that the inevitable result of Example 4 is that the monohydrate is formed, and as the impurity profile is a factor which may affect the result, Leo submit that it is for Sandoz to establish that using material with a different impurity profile would not affect the result. Sandoz submits that, by the Notice of Experiments and Repetition, the burden has shifted to Leo to show that the impurity profile has in fact caused the formation of monohydrate.

Had it been necessary to decide the point, I would have preferred Leo’s submissions. Once it is accepted that Example 4 is indifferent to the origin of the starting material, and covers a range of different processes, Sandoz must show that the necessarily differing impurity profiles would not influence the result. They have not done so. On the other hand, had the correct reading of Example 4 been that only the Calverley process was intended, and had Sandoz done an experiment to show that the Calverley process produced monohydrate, then I think Sandoz would have been correct that the burden would have shifted to Leo, at least to show that the Calverley process is sufficiently variable for the impurity profile to affect the result.

Leo also relied on a number of other aspects of Sandoz’ protocol. I was not persuaded that any of them was any better than the starting material points. In the light of the conclusions I have come to on the two main points, it is not necessary to burden this judgment with a more detailed disposition of the remainder.

Recognising that I might come to the conclusion that the attempted repetition of Example 4 did not take it home, Sandoz’ alternative approach is to prove by means of expert and other evidence what would happen, on the balance of probabilities, if Example 4 were carried out in the absence of seeds of the monohydrate.

One thing which does come out of Sandoz’ experiment, irrespective of the seeding point, is that the conditions in the Example 4 cream are thermodynamically favourable to the production of the monohydrate. The monohydrate would not have formed in Sandoz’ experiments if the amount of water, the so-called water activity, were not above the critical level for monohydrate formation. By the end of the trial all the experts had agreed that the water activity was sufficiently high for the monohydrate to be the thermodynamically favoured form.

Acceptance that the monohydrate was the thermodynamically favoured form in the Example 4 conditions is not enough to show that it will inevitably be formed. For the monohydrate to be formed it must be kinetically favoured as well. Crystallisation may be favoured, but never get started due to kinetic factors. Sandoz sought to pull together a number of areas of evidence to show that the monohydrate was the kinetically favoured form in the unseeded environment as well.

Firstly, Sandoz relied on Professor Frampton’s evidence. His view was that the thermodynamics acted as a powerful kinetic driving force. He recognised that at least in crystallisations involving different polymorphs, the presence in a suspension of a large amount of the thermodynamically less stable form might prevent the formation of crystals of the more stable form. This phenomenon, known as super-seeding, was not, according to Professor Frampton, observed with an anhydrate/hydrate transition. Here the molecules were not simply rearranging themselves inter-se, they were incorporating water molecules into their structure. The further away one was from the critical water activity, the more highly favoured kinetically the formation of the hydrate.

Secondly, Sandoz relied on the evidence of Dr Blatter for Leo under cross-examination:

Thirdly Sandoz relies on the fact that conversion to the monohydrate is known to occur in a number of seed-free environments:

Leo respond in the following way:

Further, Leo point to the results of their own milling experiments. These were conducted under differing conditions: sometimes they showed the formation of monohydrate needles, and at other times they did not. These, Leo say, are some evidence that water activity alone is not enough to result inevitably in the formation of monohydrate.

Leo also rely on some answers of Professor Frampton, given in relation to predicting which form would be produced:

“Q. I want to show you what Dr. Rasmussen has said about your view that, in the macroscopic environment of cream and ointment, the monohydrate form is more stable and formation of a monohydrate is likely to take place even in the absence of seeding. So stay in the same bundle, same tab, and read to yourself paragraphs 25-27, please.

A. (After a pause) Yes.

Q. So the general point that he is making there is that one simply cannot predict without, as it were, carrying out an experiment, whether in fact in an aqueous environment, the anhydrate will convert to the monohydrate. One has to try to find out. Do you agree with that, or not?

A. Yes.”

In relation to the milling experiments he was also asked this

“Q. Indeed. Let's assume that the needles were monohydrate. This shows that with calcipotriol, it was not possible to predict in advance on wet milling whether there would be conversion to monohydrate. It depends on the raw material that was being used; correct?

A. It would appear so.”

In my judgment, the evidence does not establish that the inevitable result of a seed-free repetition of Example 4 would produce a cream containing calcipotriol monohydrate. The environment in the Example 4 cream is not sufficiently similar to any of the allegedly analogous systems to enable one to predict, on the balance of probabilities, that monohydrate would be the result. I am not persuaded that the polymorph/polymorph transition is so different from the anhydrate/hydrate transition to enable one to rule out the impact of the presence of large quantities of anhydrate crystals in the suspension. The milling experiments show that the appearance of the anhydrate is unpredictable.

It follows that the allegation of anticipation on the basis of Example 4 of the acne use patent fails by the experiment-free route as well.

A patent will be invalid for lack of inventive step if the invention claimed in it was obvious to a person skilled in the art having regard to the state of the art at the priority date.

The correct approach for the fact-finding tribunal is summarised in the judgment of Jacob LJ in Pozzoli v BDMO SA, [2007] EWCA Civ 588; [2007] FSR 37 at [23].

“In the result I would restate the Windsurfing questions thus:

(1)

(a) Identify the notional "person skilled in the art"

(b)

Identify the relevant common general knowledge of that person;

(2)

Identify the inventive concept of the claim in question or if that cannot readily be done, construe it;

(3)

Identify what, if any, differences exist between the matter cited as forming part of the "state of the art" and the inventive concept of the claim or the claim as construed;

(4)

Viewed without any knowledge of the alleged invention as claimed, do those differences constitute steps which would have been obvious to the person skilled in the art or do they require any degree of invention?”

In St Gobain v Fusion Provida [2005] EWCA Civ 177 Jacob LJ explained the role of “obvious to try” in the assessment of inventive step:

“Mere possible inclusion of something within a research programme on the basis you will find out more and something might turn up is not enough. If it were otherwise there would be few inventions which were patentable. The only research which would be worthwhile (because of the prospect of protection) would be in areas totally devoid of prospect. The “obvious to try” test really only works where it is more-or-less self evident that what is being tested ought to work”

This and other authorities on “obvious to try” were considered in the judgment of Jacob LJ in Conor v Angiotech [2007] RPC 20; [2007] EWCA Civ 5. In the House of Lords, [2007] UKHL 49; [2008] RPC 28 at [42] Lord Hoffmann explained:

“In the Court of Appeal, Jacob LJ dealt comprehensively with the question of when an invention could be considered obvious on the ground that it was obvious to try. He correctly summarised the authorities, starting with the judgment of Diplock LJ in Johns−Manville Corporation's Patent [1967] RPC 479, by saying that the notion of something being obvious to try was useful only in a case in which there was a fair expectation of success. How much of an expectation would be needed depended upon the particular facts of the case. As Kitchin J said in Generics (UK) Ltd v H Lundbeck A/S [2007] RPC 32, para 72:

“The question of obviousness must be considered on the facts of each case. The court must consider the weight to be attached to any particular factor in the light of all the relevant circumstances. These may include such matters as the motive to find a solution to the problem the patent addresses, the number and extent of the possible avenues of research, the effort involved in pursuing them and the expectation of success.””

Lord Justice Jacob’s phrase “more-or-less self-evident that what is being tested ought to work” is explained by Lord Hoffmann as a “fair expectation of success”, with the degree of expectation depending on the facts of the case. I have had these principles in mind when approaching the question of obviousness in this case.

Sandoz put their case of obviousness in quite a variety of different ways:

It seems to me that points (iii) – (v) are all slight variations on a more general attack: that given the disclosure of an aqueous suspension cream containing calcipotriol (e.g. Example 4 of the acne use patent) it was obvious to find and use the monohydrate. I will deal with this case first, and return to the remaining points thereafter. Before doing so I need to summarise the history of the way in which Leo arrived at the invention, as this formed an important part of Dr Rasmussen’s reasoning as to why the invention was not obvious.

Calcipotriol was first synthesized by Leo in 1985 and was disclosed in the 834 application and the Calverley paper in 1987.

Leo first developed a solution cream in which the calcipotriol anhydrate was dissolved in a water-in-oil cream. Successful clinical trials were undertaken, and published in 1988, but the specific cream tested was not sufficiently stable to give a long enough shelf life for a commercial product. Calcipotriol was not sufficiently stable in this particular solution.

Leo then decided to develop an ointment. The calcipotriol was dissolved in a buffered propylene-glycol/water phase and then dispersed into a fat phase to produce the ointment. This solution ointment was launched in the UK in 1991 and is still on the market under the brand name “Dovonex”.

Despite the problems encountered with the solution cream, Leo still wanted to market a cream in addition to the ointment. From around 1989, Leo worked on formulating a suspension cream. It was thought within Leo that this would be likely to be more stable than a solution.

Leo investigated milling as a way of obtaining the requisite sized particles of calcipotriol for a reproducible suspension, but encountered problems with wet milling. The principal problem was the formation of foam and occasionally threads or needles. These problems persisted for about 3 years. Attempts were made to avoid the problems by changing the milling and process conditions. One theory put forward was that the threads that appeared resulted from crystal growth of the anhydrate crystals, because anhydrate also exhibited a needle-like crystal habit.

In February 1991, Erik Hansen re-crystallised some calcipotriol for his own purposes. This was unconnected to the milling problems being experienced with the calcipotriol anhydrate being used for the cream formulation: he was merely seeking to characterize the compound more fully. This work was done in order to try to create larger crystals suitable for a single X-ray analysis. The material was evaluated and designated ETH 2164. It was the monohydrate. The results of this experiment were not communicated to the production team, however. An attempt was made in the cross-examination of Dr Rasmussen to suggest that Erik Hansen was setting out to make the hydrate. Dr Rasmussen rejected it, and I do as well. There was no technical reason to try and make a hydrate for the particular purpose which Hansen had in mind.

In August 1992 Leo analysed a sample of the needles that had been produced in the wet milling process over the last 3 years to see if anything could be learned from this analysis. It was established that these needles were identical to the ETH 2164 identified by Erik Hansen. The first reaction to this was to change the milling process. Subsequently, contamination with monohydrate was put forward as a cause of the problem and attempts were made to ensure that all milling components were free of monohydrate.

It was then suggested that, rather than altering the milling process, the solution to the problem lay in using the monohydrate from the outset.

As a result, Leo switched from using calcipotriol anhydrate to calcipotriol monohydrate and in so doing solved the production problems.

Sandoz puts its obvious case under this head in three ways. Firstly, Sandoz submits that the skilled team would know that, in the course of submitting an application for regulatory approval for a calcipotriol aqueous suspension cream (such as Example 4 of the acne use patent), it could expect to be required by the regulatory authority to check whether the API formed a hydrate. Secondly, it would be obvious to conduct a polymorph screen in the course of pre-formulation studies on an API such as calcipotriol. Thirdly, the skilled team would be bound to check for the presence of hydrates in any event, in the course of routine stability studies. In each of these cases Sandoz submits that the skilled team would encounter the monohydrate and incorporate it into the cream.

Whilst most of the evidence and argument centred around Example 4 of the acne use patent, I have kept in mind that both the acne use patent and the 834 application contain general disclosures of suspensions of calcipotriol.

It is important to realise at the outset that it is not and cannot be asserted by Sandoz that it was obvious that calcipotriol would form a hydrate. Far less could it be asserted by Sandoz that it was obvious that calcipotriol would form a hydrate which had better properties than the anhydrate for inclusion in a suspension cream or gel, or for wet milling. Sandoz’ case is based on the suggestion that, faced with the disclosure of the acne use patent, the skilled team would embark on a programme of work which would lead them, without invention, to such a product. To analyse that case correctly I must first determine what programme or programmes of work, in point of fact, I consider it was obvious for the skilled team to have undertaken. Secondly I must reach a conclusion as to what the skilled team would have found as a consequence of undertaking that programme.

Sandoz relies on the history of the interaction between Leo (and its licensee BMS) with the United States Food and Drug Administration (FDA) in relation to an application for regulatory approval of Leo’s ointment formulation in which the calcipotriol was in solution. In September 1992, the FDA reviewer produced a list of deficiencies in the application. One of these was put as follows:

“Powder diffraction data and their interpretations were not submitted for the nine batches tested… We recommend that you conduct similar studies on [calcipotriol] obtained by the recrystallization from various solvents (e.g. water, methanol, ethanol, ethyl acetate, hexane).”

Leo responded to BMS saying that it would provide powder diffraction data on the batches but it doubted whether recrystallisation studies could be completed in the time frame. Leo suggested that BMS respond to the effect that:

“crystal modifications of calcipotriol is of no significance for the present formulation .. [as it] is present in a solution”.

BMS responded to the FDA on October 2^nd 1992 saying that the powder diffraction studies on the nine batches indicated no variations in crystalline form. BMS relayed Leo’s suggestion that, as the drug was dissolved in the drug product, crystal modifications were not likely to be a concern.

The FDA did not accept BMS’ and Leo’s justification for not performing crystallisation studies and continued to insist on a solvent screen, re-listing the exemplified solvents. BMS had had a discussion with two chemists at the FDA. The discussion was reported to Leo as follows:

“Both [chemists] were quite firm that such data are necessary to provide assurance that [calcipotriol] reveals no polymorphic forms that might affect solubility in the ointment vehicle.

They were not dissuaded by the fact that the active was fully dissolved in the vehicle, since they are of the opinion that a less soluble polymorph might not fully dissolve, and indeed might provide seeds for recrystallization in the finished ointment. They insisted on data from material recrystallized from both polar and nonpolar solvents to provide assurance that polymorphs will not arise by inadvertent use of an incorrect solvent.”

Soon afterwards Leo commenced crystallisation experiments using (1) acetone (2) propanol (3) ethylacetate and (4) di-ethyl ether. Two solvents (acetone and diethyl ether) were not used in the API synthesis.

Sandoz submits that it is significant that the FDA was concerned with a solution product, yet went so far as wanting to know whether other polymorphs could be created by other solvents that might somehow get into the manufacturing process. It submits that, if that is the case in relation to a solution product, then if an undertaking were seeking to market calcipotriol in an aqueous suspension cream (or an aqueous suspension gel), the FDA, and corresponding regulatory bodies would, as an absolute minimum, have wanted to know whether the API could form a hydrate. More importantly, the skilled team considering the development of such a cream would know that they could expect such a requirement.

Dr Rasmussen’s evidence in relation to this proposition was as follows:

“Q. We may have touched on this but let's take it fully frontally. The FDA if they were being asked to approve a suspension cream of an oil in water emulsion, calcipotriol, and they were wanting to know about polymorphs -- first of all I think you obviously agreed they could not be fobbed off with "It is all in solution"?

A. That would be messing around with the FDA and you do not want to do that.

Q. The FDA asking about solvents that might inadvertently get into the system and create a different polymorph, the moment they saw that the anhydrate was in an aqueous environment, they would want to know about whether hydrates formed, would they not?

A. Yes, in the suspension?

Q. Yes.”

Later, Dr Rasmussen was also asked if this would be the case if seeking to market a cream according to Example 4 of the acne use Patent:

“MR. JUSTICE FLOYD: Mr. Watson is asking you to assume that you were going forward with example 4, making it up into a suspension and so on, and submitting it to the FDA.

A. And the question was?

MR. WATSON: The question was, there would be no ifs and buts, you would have to, as part of your dossier, prepare information about the existence of hydrates.

A. Yes.”

So far as more general polymorph screening is concerned, Sandoz submits that in 1993 this was an obvious step to take in the Vitamin D field.

Finally as regards stability studies, Dr Blatter accepted that in a situation where you are planning to put an anhydrate into an aqueous formulation, it was “elemental” (in context it was clear he meant “elementary”) to check, for stability reasons, whether the hydrate could form. At the conclusion of his evidence he returned to this point and qualified it by reference to an incident in which an experienced PhD chemist had not appreciated that his problem was that a hydrate was forming in an aqueous suspension.

Leo’s response on the question of what (if any) programme would be undertaken is as follows. Firstly, they say that Example 4 is a “paper” example. They rely on the fact that, at the priority date, all calcipotriol preparations were in solution. There is no evidence that Example 4 was ever performed. Indeed Dr Rasmussen said that it had not, to his knowledge ever been used. When Professor Williams started to attempt to replicate it, he ran into problems with his initial attempts at very small quantities in a pestle and mortar. All this, say Leo, is a very long way from establishing that one would get as far as an application for regulatory approval for an Example 4, or any other suspension cream.

Secondly, Leo submit that the evidence of Dr Rasmussen on which Sandoz relies was predicated on the assumption that one is telling the FDA in terms that the API is an anhydrate. In fact, at the priority date, there was only one compound, that made by the Calverley process. There would have been no need to call that compound an anhydrate: that is a term which is used to distinguish from hydrated forms, and no such forms were known. In consequence, one cannot assume that the FDA would have been encouraged to invite the putative applicant to test for hydrates.

Thirdly, Leo submit that there is an oddity about Sandoz’ case in the following sense. Sandoz’ initial position was that its suspension cream did not infringe because it used the anhydrate. Plainly the regulators had not required Sandoz to investigate the formation of hydrates and required them to use the more stable form.

As to polymorph screening, Leo’s case is that it was not done in the vitamin D field, either by Leo, or, at least to Dr Rasmussen’s knowledge, by others.

The check for hydrates would, say Leo, not routinely be performed: particularly so where the only known form was MC903 made by the Calverley process, which was not a hydrate.

Taking the three suggested programmes of work in turn, I am first of all not persuaded that the skilled team would undertake full polymorph screening. There would be little, if any, motivation to do so in the light of what was known about Vitamin D analogues, and no regulatory requirement to do so. To the extent that the FDA was pressing for more work in the case of Leo’s solution product, I think it was going further than the skilled team would expect or the guidelines required. Leo certainly considered at the time that it was going too far. I was not persuaded by Professor Frampton that work of this nature would be undertaken with anything approaching the necessary “fair expectation of success”, particularly in the light of the problems inherent in any work on crystallisation with Vitamin D compounds. His evidence, relying on the three patents for the known commercial hydrates was a research programme that would be carried out in the hope of finding out something valuable, but with no particular expectation of success. On any view, that is not a case of obviousness. His cross-examination, at one point, went like this:

“Q. So you say knowledge of those three compounds would further suggest to the skilled person that it was possible that calcipotriol would form a hydrate, although this would not necessarily be the case?

A. You know, if you see from reading these three documents that all of these compounds form hydrates, knowing something of the solid form where water will hydrogen bond to hydroxy groups, there is always a chance that it will do it for a similar compound. And I am not saying that crystal structures of this complexity are predictable and I am not saying that it will definitely form, what I am saying is it is worth trying to make this material by recrystallising from water, to see if you can drive the hydrate. Because, you know, given the paucity of solid forms with Vitamin D compounds, it may be a golden nugget by which to move your process on.

Q. But in trial, you would have no particular expectation as to whether you would succeed or not?

A. No. No, but it is certainly worth trying.”

However, in my judgment the evidence does establish that the skilled team would, at some stage in the development of an aqueous suspension cream of calcipotriol based on Example 4 of the acne use patent or otherwise, realise that it would have to investigate the formation of hydrates. It would be obvious to do this as part of a routine check in the course of stability studies or in anticipation of a regulatory request. My reasons follow.

Firstly, the evidence established that there were good reasons for seeking to develop the Example 4 cream. Suspension creams were well known, and the skilled team would recognise Example 4 as a known type of formulation with known advantages. It would have no reason to reject it as a paper example as Dr Rasmussen said it was in fact: in that respect Dr Rasmussen was not in the same position as the hypothetical addressee. Professor Williams’ early difficulties would not in my judgment have led to the project being abandoned, far less being abandoned without thought of what data would be needed to take it forward.

Secondly, Leo’s and BMS’ dealings with the FDA are not only good evidence of what would have happened if the skilled team had progressed the project as far as an application for approval. They are practical confirmation of what the skilled team would expect from its knowledge of the relevant guidelines and the regulatory climate. In particular, the skilled team would appreciate from an early stage that it would be necessary to ensure that the drug did not form a hydrate either in the course of manufacture or in storage. It is not surprising that both Leo’s experts accepted this proposition given that the calcipotriol is being suspended in an aqueous environment.

Thirdly, I do not think that Dr Rasmussen’s evidence (that he would expect the regulator to require assurance that there were no hydrates formed from calcipotriol in an aqueous suspension cream) was predicated on any false premise. His acceptance of that was inevitable given the Guidelines. If so, there are good reasons to establish whether or not hydrates are formed at an early stage.

Fourthly, I formed the impression that Dr Rasmussen accepted the force of these points, but clung to the fact that Leo did not proceed in this way at the time. I am conscious of the need to place the question of obviousness in its real-life context. But historical evidence of this kind needs to be kept in its place: see Molnlycke v Procter & Gamble [1994] RPC 49 at 113 lines 8-15. Leo approached the suspension cream after having successfully marketed a solution ointment and unsuccessfully developed a solution cream. With these products, they could tenably take the position with the regulators that investigations of polymorphs were unnecessary. By the time they came to the suspension cream they had a history of working with the anhydrate and plainly, and perhaps understandably, overlooked the fact that the new environment required them to reassess their attitude. As a result they ran into problems of their own making with wet milling which may well have required some ingenuity to solve. Dr Rasmussen accepted that this was a cautionary tale. Given what is said in the textbooks and guidelines which were common general knowledge at the priority date, it could, at the time, have been cited as one. None of this is typical of the condition in which the ordinary skilled team approaches the development of Example 4.

I have not ignored Leo’s history: but in the end, nothing in Leo’s history persuaded me that it was technically anything other than obvious to proceed to check for the formation of hydrates in the process contemplated for use.

Finally, I am not persuaded that anything is to be inferred from the fact that Sandoz proceeded with an anhydrate product, apparently without objection from the regulator. I have no direct evidence about this, or what may have passed between applicant and regulator in relation to the product. The guidelines, and the history of Leo’s dealings with the FDA, of which I do have evidence, are, to my mind, much more compelling.

This, in my judgment is the crucial question. I approach this question on the basis that the programme is a limited one designed to check for the likelihood of formation of hydrates in a process such as Example 4 of the acne use patent, and subsequent storage of the suspension cream. I have rejected the suggestion that a more wide-ranging, research-based project aimed at uncovering polymorphs and pseudopolymorphs of all kinds would be undertaken. Sandoz submits that the skilled team undertaking a check for hydrate formation, even of the more limited kind, would discover the hydrate.

Sandoz rely on their cross examination of Dr Rasmussen as follows:

“Q. It is the scientific fact that if you recrystallize the anhydrate in the presence of active water it readily converts to the hydrate?

A. Not necessarily. In several of the experiments in these documents seeding was necessary to get the crystals. It depends I think on the actual volume of the mixture, the ratio between the solvent and so on. At least in quite a lot of the experiments in this file, seeding was used to get it. So when you have it once, you can get it by seeding the next time.

Q. Yes, but also there was no problem getting it without seeding from a number of solvents, with a number of ratios of water?

A. No.

Q. I mean, you are not suggesting if a polymorph screen had been carried out on calcipotriol, that the hydrate would not have been found?

A. It would probably have been found, yes.”

The following day he replied as follows:

“Q. I thought you agreed that you are not suggesting that the monohydrate would not have been found in the screen?

A. It is likely it would because as we also agreed yesterday, acetone and water are pretty common solvents in crystallization experiments.”

Dr Rasmussen’s acceptance of the likelihood of finding the monohydrate was on the basis of a polymorph screen. He answered, on the second occasion, specifically by reference to acetone and water, a combination of solvents now known to produce the monohydrate. He was never asked precisely what tests he would have performed if asked by the regulator to check for hydrate formation in the process of Example 4 and in subsequent storage (or any other process), and whether those tests would necessarily have produced the hydrate. Sandoz’ witnesses did not proffer any evidence as to this either. I think it is wrong to infer that Dr Rasmussen would have accepted that acetone or acetone and water would have been used in these more limited circumstances. Acetone is not used in the Example 4 process.

It is tempting to speculate about whether the regulators would have insisted on a more wide ranging investigation, using other solvents, but I think it would be wrong to do so. I think the evidence supports the view that the skilled team would expect to do no more than satisfy the regulator that in the conditions of the contemplated process there was no evidence of the formation of hydrates. This would not involve using a wide range of solvents.

I do not think that the evidence provides an adequate basis for the conclusion that a hydrate check of the kind suggested would, more probably than not, uncover the hydrate. The Patent shows that formation of the hydrate can be triggered by the methods taught from ethyl acetate or acetone and the addition of water, or by seeding. In contrast, there is no evidence to suggest that the former would be part of the limited check for hydrates. To use seeding would, on the other hand, presuppose some monohydrate had already been obtained, which it would not.

It is true that the evidence I have reviewed under the heading of lack of novelty shows that there is a variety of sets of conditions under which the monohydrate will form, such as in gels and in milling: but I am not in a position to hold that any of these would be likely to be tried as part of the test for hydrates. In an empirical field such as this, assumptions as to what would be found by an experimental programme are not justified. The problem is accentuated where the scope of the alleged experimental programme is imprecisely defined.

In the end, therefore, I was not persuaded that the step from the acne use patent to the inventive concept of any of the claims of the Patent would have been obvious on any of these approaches.

Sandoz argued that, to the extent that some of the more minor criticisms of the protocol adopted by Professor Williams were made out, then they were obvious variants of the procedure of Example 4 of the acne use patent. However, as I have held that, on the balance of probabilities, there were seeds present in the starting material, this ground of objection falls away.

Sandoz contends as follows:

It is important to note that Sandoz’ argument for obviousness by this route is not based on the skilled team having formed any prior intention to see whether there is an alternative crystal form to the anhydrate. Unlike the argument I have already rejected, it is an argument which depends on the skilled team stubbing its collective toe on the invention in the course of making a trivial alteration to Example 4.

Leo contends:

The evidence establishes to my satisfaction that wet milling is an obvious variant of Example 4. Furthermore, I was not persuaded that the problems encountered by Professor Williams with his pestle and mortar would be reproduced if proper wet milling equipment was used. Dr Rasmussen did not suggest that in fact this would give rise to any problems. Of course wet milling might produce crystals of monohydrate and consequent foaming: but that is not a point which can assist Leo: by that point monohydrate would have been made. Points (i) to (iii) do not persuade me that the skilled person would have abandoned Example 4. I also do not think that Leo’s point (v) is a good one: Dr Rasmussen accepted that if the crystals of monohydrate were produced in the milling process they would end up in the cream or other product that was being produced.

The decisive question is whether the monohydrate would in fact be produced when the skilled team carried out wet milling in Example 4. Dr Rasmussen explained that the Calverley process was varied by Leo over the years until it was locked down for the commercial production process. Throughout the period some of the batches produced monohydrate needles and others did not. His cross-examination concluded in this way:

“Q. Could I suggest, based on the Leo experience, that three different APIs, as the manufacturing developed, gave monohydrate during milling. As best you can say to my Lord is, the assumption is that if you carried out example 4 with any pharmaceutical grade calcipotriol anhydrate on a regular basis, you will achieve monohydrate. Leo did it. Is there any reason to believe that any pharmaceutical grade calcipotriol anhydrate will not likewise hydrate while being wet milled?

A. Yes, I think there is because some of these batches do not do exactly that, so this is something which is, something you have to investigate.

Q. Yes, but since it happened 50% of the time with three different APIs, different variants of the API, I am suggesting to you that the likelihood is that if you used pharmaceutical grade calcipotriol anhydrate to carry out example 4 on a commercial scale, on a regular basis, as best one can predict you would get monohydrate from time to time. Is that not the best estimate you can make on the evidence available?

A. That is the estimate for the Leo batches here, yes.

Q. And have you any reason to suggest that other pure calcipotriol anhydrate is going to behave differently?

A. I have no knowledge of that, no.

Q. Is there any reason to suggest it is going to behave differently?

A.

No.”

Mr Watson also asked Dr Rasmussen how one would prevent the monohydrate forming in wet milling. His reply was that one would have to test the production batch-by-batch for the presence of monohydrate, something which of course one could not do if one did not already know it existed. But I do not think this advances the argument. It does not help to establish the underlying fact on which this obviousness case must be based, namely that the consequence of using wet milling is that one will stumble on the invention.

Although wet milling is an obvious step to take, it is plain that wet milling does not guarantee production of the monohydrate. Leo’s experiments demonstrate that Leo’s batches produced monohydrate about half the time, but it does not follow that another manufacturer developing its process conditions in another way would encounter the monohydrate at all. Dr Rasmussen, in the passage cited above, answered that he simply did not know what would happen.

In my judgment, it is not established that the invention was obvious on this basis. I do not think it fair to hang as much on Leo’s milling experiments as Sandoz wish to do. The form in which a crystallisation occurs, is inherently unpredictable. There is also the potential for a wide variation in processing conditions which could either favour or militate against the formation of the anhydrate. Taking these factors together, it seems to me to be impossible to conclude on the balance of probabilities that the skilled team pursuing Example 4 would encounter the monohydrate. If that is the conclusion for Example 4 of the acne use patent, it seems to me that the same conclusion applies to the 834 application and Calverley.

Sandoz also argued that routine crystallisation experiments would lead to the formation of the hydrate. I do not think this ground adds anything to the other grounds. The nature of the experimental programme suggested is neither established nor sufficiently clearly defined to enable a conclusion as to what it would have uncovered.

No claim of the Leo patent is invalid. The action succeeds and the counterclaim will be dismissed.