England and Wales Court of Appeal (Civil Division) Decisions

IN THE HIGH COURT OF JUSTICE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE HIGH COURT OF JUSTICE
CHANCERY DIVISION (PATENTS COURT)
The Hon Mr Justice Floyd
HC 08 C00391

		Royal Courts of Justice Strand, London, WC2A 2LL
		17/11/2009

Lord Justice Jacob:

Introduction

1. Sandoz appeals Floyd J's decision ([2009] EWHC 996 (Pat)) that Leo's EP (UK) No. 0 679 154 is valid. By the time of trial it was common ground that if the Patent is valid, it is infringed by Sandoz's product. So only validity was in issue.
2. Antony Watson QC assisted by Piers Acland argued the case for Sandoz; Henry Carr QC and Andrew Lykiardopoulos that of Leo.
3. Claim 1 of the Patent is for a single chemical entity, calcipotriol monohydrate, a crystalline substance. The crystal lattice includes both calcipotriol and water.
4. Calcipotriol is a vitamin D analogue. It was first disclosed by an earlier Leo patent application WO 87/00834, published on 12^th February 1987, nearly six years before the priority date of the Patent (15^th January 1993). Example 5 of that application discloses calcipotriol and a process for making it. A paper by Calverley also published in 1987 contains the same information – in the Calverley paper the internal Leo designation MC903 is used for calcipotriol. MC903 is an anhydrous form of calcipotriol, also a crystalline material.
5. A patent application (WO 91/12807) (called "the acne use patent"), published on 5^th September 1991 (i.e. before the priority date) contains some examples. There are six product examples and one example reporting the use for acne treatment. Of the six product examples four are creams, one is a lotion, one a capsule. The only example reporting actual use and therapeutic effect relates to use of the lotion. Four of the product examples use MC903. Examples 1, 3 and 4 are recipes for making creams containing calcipotriol. Example 4 uses a milling process.
6. The medical use of calcipotriol is for the topical treatment of skin conditions (e.g. psoriasis or acne). The Patent acknowledges this. The 1987 application is to much the same effect, for instance mentioning psoriasis.
7. The Patent sets out the advantages of the monohydrate: superior stability and superior technical properties in the manufacture of crystal suspension formulations. These advantages are real and significant. It is not contended that anyone could have foreseen them.
8. The case against the Patent was that the hydrate (and a cream containing it) was not novel over example 4 of the acne use patent. Alternatively it was obvious simply because a "person skilled in the art" (in practice a team) would routinely have taken such steps such that the hydrate would have been produced and its unexpected beneficial properties discovered.
9. So the case is unusual. The ordinary obviousness attack consists of a contention that the skilled person would, using his technical knowledge, discern the invention from the prior art. The case here is that the skilled person would have come upon the invention (the hydrate and its benefit) without any expectation of successfully finding a better product.
10. That sort of obviousness attack should be scrutinised with great care. I do not say it could not succeed, but one must be very confident that the steps said to lead to the discovery of a new and beneficial product "by accident" as it were, were at the least, really likely, almost mandated. If you need to do research to find an invention then, for a finding of obviousness, that research must be of a kind which a skilled man would do, not which he might do.
11. Mr Watson advanced no less than five distinct attacks on the patent.

(1) Anticipation by the acne use patent.

(2) Obviousness over the acne use patent because it was obvious to conduct a full polymorph screen of the example 4 product, during which the monohydrate would have been discovered.

(3) Obviousness over the acne use patent because a product screen would have revealed the monohydrate.

(4) Obviousness over the acne use patent because it would have been obvious to vary example 4 so as to wet mill. That would have produced the monohydrate

(5) Obviousness in light of common general knowledge alone because experiments into crystallisation would have revealed the monohydrate.

Principles on Appeal

12. Apart from one point, which he sought to dress up as one of law, Mr Watson acknowledged that his appeal was a full frontal attack on the Judge's findings of fact, evaluations of the evidence and value-judgment conclusions. Although he was uncomfortable with the word, Mr Watson accepted that his attack amounted to one of "perversity" – that the Judge's conclusions were such that no judge could reasonably have reached them on the evidence.
13. Mr Watson had to put his case that high. An appeal on questions of this nature will only be allowed if this court is satisfied that "the decision of the lower court was wrong" (CPR Part 52.11(3)). If there was material upon which the findings could be made, that test is not passed.
14. This Court set out how that rule applies to patent appeals in more detail and some stringency in Merck's Patents [2003] EWCA Civ 1545 [2004] F.S.R. 332 at [62-71] per Buxton LJ and [71] per Sir Andrew Morritt V-C. I need not set it out again.
15. As regards obviousness, there is also the Biogen principle [1997] RPC 1 at p.45 per Lord Hoffmann. The Judge's conclusion must be reached as an overall assessment of the relevant evidence, both technical and non-technical. If he has made such an assessment, weighing all the factors involved, then it will be difficult indeed to show that he has made an error of principle.

Issue (1) Anticipation by Example 4 of the acne use patent.

16. There never was any dispute what it was that Sandoz had to prove: that the inevitable result of carrying out example 4 is that the monohydrate would result. The Judge set out the principles he had to apply at [56-58]. No error of law is suggested.
17. One would have thought that the task of proving this would be undertaken in a straightforward way. I set it out when at first instance in Synthon's Patent [2003] RPC 33 at [57]:

Given the "inevitable result" branch of the law of anticipation (see below) one might have thought (a) that a team of ordinary ability might have been engaged (one is concerned with the ordinary skilled man or team, not world champions) and (b) that the team concerned would simply have been given the Synthon patent and asked to carry out its teaching to make paroxetine mesylate crystals.

18. Experience shows that some parties attacking patents simply do not follow this straightforward path. Instead they depart from the prior art. Then, as here, an argument starts about the nature of the departure and whether it mattered. For the life of me I cannot understand why they do this. It inflates the costs and time, and seldom if ever does the defendant any good.
19. Here the experiment suffered from two vices. First the obvious precaution of ensuring a seed-free environment was not taken. On the contrary it was seeded (it is not suggested deliberately so) with monohydrate (Judgment [76]). Second the recipe of example 4 was simply not followed. Why was never explained. Example 4 uses MC903 which the acne patent tells you is made by the process of example 5 of WO/00834 also described in Calverley. Sandoz used calcipotriol made by a company called Teva, made by a process which was never disclosed and devised years after the Patent. That product had different impurities.
20. The Judge found (and this finding was not challenged) that "seemingly trivial matters such as impurities, may influence the result." He had evidence to support that from Dr Rasmussen (one of Leo's expert witnesses – "an impressive and fair witness" even though he was Chairman of the Board). Professor Frampton (one of Sandoz's expert witnesses) also accepted that impurities can have a "profound effect on the growth of a crystal."
21. Plainly as of the date of the Patent the skilled team could not have used the Teva process – the only published one was Calverley. So why the departure? Mr Watson suggested that there would be difficulties, though none had been offered in evidence. He took us to some cross-examination of Dr Rasmussen in which, probably out of desperation, he speculatively asked whether Leo had tried a repeat of example 4. They had not (and why should they, the onus lying on Sandoz?). But Dr Rasmussen also said it would take some time – he actually referred to administrative reasons rather than technical ones. It is not uninteresting that his first reason was the location of a seed-free environment – the very thing Sandoz had neglected.
22. The failure to use MC903 was one of the reasons why the Dutch court, Case No. 306029/HA 08-733), refused to accept a case of inevitable result (see 4.6). Mr Watson told us the Dutch decision "went off on a point" not available here. He was wrong at least in that respect.
23. The Judge did not think the failure to start with MC903 was in itself fatal. He thought that the skilled team would read the example as saying "start with the anhydrate compound howsoever made". But that meant, he reasoned, that Sandoz had to show that howsoever the anhydrous calcipotriol was made (and so with an undefined range of possible impurities) the inevitable result would be the monohydrate. The target was much bigger than if you had to use MC903. And Sandoz had simply failed even to attack that wider target.
24. Mr Carr, by a respondent's notice, said the Judge was wrong. The example said start with MC903 – so with that you had to start, as the Dutch court held. This is just the sort of argument that is let in by any departure from the recipe of the prior art and one of the reasons why unnecessary departure is to be so deprecated.
25. For present purposes I do not think it matters. If you have to start with MC903 the Sandoz experiment is largely irrelevant. If you do not, then the Judge is surely right. Mr Watson's "point of law" was devised to get around this. He submitted that if one starting material was shown to work, then somehow the onus shifted to the patentee to show other starting materials would not. I do not think any point of law is involved. The court must simply decide on the evidence as a whole whether the result is inevitable. There may be cases where the precise nature of the starting material is obviously technically immaterial. And there will be others where it may matter – as here where it is common ground that impurities can affect the result. If they may matter, it is clearly not for the patentee to conduct experiments to show that they in fact do. The onus lies on the challenger to show inevitable result.
26. The upshot of all this is that Sandoz's primary attack – proof by experiment of an inevitable result - failed miserably. There is no appeal.
27. However before the Judge, and resurrected before us, is Sandoz's fall-back position. It is said that from theoretical considerations, seeding would make no difference and is unnecessary: on the balance of probabilities the calcipotriol in the example 4 cream would inevitably turn to monohydrate even without seeding.
28. The Judge rejected this at [94]. Mr Watson submits he was wrong, so wrong that no judge could reasonably have come to that conclusion. His case was that his expert, Professor Frampton, advanced theoretical reasons why there would be this transformation and that Leo's witness, Dr Blatter agreed. So it would be perverse not to conclude that conversion of the crystalline form would happen.
29. The theoretical case runs as follows. The monohydrate is a pseudo-polymorph. If calcipotriol is dissolved in water and the thermodynamics are such that formation of the monohydrate is favoured then, because the system is in dynamic equilibrium with calcipotriol coming in and out of the water, when it comes out it will do so as monohydrate. It will do so rapidly and inevitably. You know that example 4 has the right thermodynamics (has the right level of water activity) because Sandoz's experiment at least shows that you get monohydrate, albeit with a seeded system. Even with seeding you would not get monohydrate if it were not the thermodynamically favoured form. I hope I have summarised Prof. Frampton's argument accurately.
30. Mr Carr advanced a host of reasons why there was ample material justifying the Judge's conclusion. I go to most of them seriatim.
31. In his second report Dr Rasmussen said that production of a crystalline form depends on whether it is thermodynamically preferred and whether the kinetics favour it. Prediction is not possible until both the thermodynamics and the kinetics are understood. He went on to give examples of the sort of thing that can affect the result – including rate of cooling, changes in milling conditions and so on. He was quite clear that it was his belief that it was the seeding which caused the formation of monohydrate in the Sandoz experiments – something he would not have said if he thought it would happen anyway. He referred to past experiments done at Leo on the subject, experiments done for the purpose of finding out and in no way done to prove something in litigation.
32. Then there is the Patent itself. This contains an example (2) of seeding to produce the monohydrate – so seeding was necessary there. If it was necessary, Prof. Frampton's theory is at fault.
33. In his second report §89 Prof. Frampton had said "it is very likely that the monohydrate would form, whether or not there are any seed crystals or other contaminants". But in cross-examination, faced with Dr Rasmussen's evidence, he gave way, accepting that you have to try to find out. The Judge quotes the relevant evidence at [93].
34. Mr Watson submitted that the position was recovered in re-examination, taking us to pp.461-2 of the Transcript. I am not impressed. First, without actually directly leading the witness, Mr Watson gave him a clear signal that he had given a damaging answer and what the "right" answer was supposed to be. He introduced his questions with the wholly unnecessary words:

Q. There were a couple of answers you gave, professor, that we need to put in context otherwise there will be a great debate about what you meant.

This is but a poorly concealed way of saying "you gave a wrong answer." He then took the professor to his evidence in chief, thereby showing him the "right" answer. He then read out the damaging question and answer and asked the witness what he meant by "predict." Eventually he got the answer he wanted – "monohydrate would form whether seeds were present or not."

35. The Judge was fully entitled to reject this evidence. Re-examination can be helpful when it merely clears up an uncertainty or brings out a fresh fact which puts a different complexion on things. But where an expert witness clearly says one thing in cross-examination and the opposite in re-examination the Judge is entitled to reject his evidence.
36. Much the same thing happened when Prof. Frampton was asked about Leo's internal milling experiments done as part of their ordinary research and not for the purposes of litigation. These showed that for identical conditions sometimes there was monohydrate, sometimes not. That could not happen according to Professor Frampton's theory. Prof Frampton suggested that the batches might be different in some way. But that does not fit his theory which is independent of the nature of a particular batch. Besides in one case calcipotriol from the same batch was used in two identical experiments – one produced monohydrate, the other not. As the Judge said (and was entitled to say) "the milling experiments show the appearance of the monohydrate is unpredictable."
37. Another explanation offered by Professor Frampton, also inconsistent with his theory of inevitability is that the milling experiments might have been on the "edge of water activity." But he had earlier said "as long as you above the critical water activity then the driver will be there and when you are on the critical water activity then you will get crystals of both anhydrate and monohydrate coming out."
38. Likewise Prof. Frampton's acceptance that impurities can have a profound effect on crystal growth is simply inconsistent with the assertion that the monohydrate will inevitably be formed.
39. Dr Rasmussen was pressed in cross-examination. He said he could not predict without experimentation (p.236). That was evidence which the Judge was entitled to accept.
40. Then there was the evidence of Dr Blatter. In his written evidence he was clear that he was unable to conclude that monohydrate would be formed – the vast bulk of the calcipotriol would be anhydrate because it is so insoluble. There would be little opportunity to nucleate and one could not predict when if at all it would happen.
41. Mr Watson's skeleton argument relied upon a part of what Dr Blatter had said in his second report as showing that Dr Blatter agreed with Professor Frampton:

If Professor Frampton's assumption is correct that in the relevant formulations the monohydrate is the thermodynamically preferred form of calcipotriol, then the statement in paragraph 90 of his report that the monohydrate will form in the cream/ointment "over a period of time" (even in the absence of seed crystals) is also correct. [Mr Watson's emphasis]

That was not a fair picture of the doctor's evidence. The passage went on:

However the kinetics of transformation cannot be ignored. The transformation of anhydrate to monohydrate can occur within a highly variable timeframe, because it depends on many parameters that affect the activation energy barrier for the transformation (as outlined above). It may be that the anhydrate is kinetically stable and that in the cream/ointment environment and under ambient conditions the amount of time needed for the conversion may exceed the shelf life of the product. Indeed it has been found for many different compounds that conversion to the more stable form does not readily occur in a seed-free environment; ritonavir, which is mentioned above, is an example of this.

42. The doctor was not accepting the thesis that conversion was bound to happen – a conversion which, if it happens at all, happens only beyond the shelf-life of the product is clearly as a practical matter one which does not happen. Inevitability was not accepted in the passage which was quoted.
43. Mr Watson then relied upon his cross-examination of Dr. Blatter. We were invited to construe it as though it were a document. I am bound to say two things. Firstly that is hardly a helpful approach when the Judge saw and heard the witness. Secondly I do not agree with Mr Watson's suggested construction anyway. The passage relied upon went like this:

Q. You expect it would come out as monohydrate and you have no reason whatsoever to believe that that will not happen in a short period of time, certainly in a time span of weeks or months?

A. I cannot know what the timescale exactly is. I cannot know.

Q. But take probabilities, have you any reason to believe it will not happen within that timescale?

A. That is difficult to say; very difficult. [Mr Watson's emphasis]

It was suggested that we should read this as Dr Blatter accepting that it would happen in a short period of time. Not so. He is saying he cannot say, one way or the other. Saying he had no reason to believe it will not happen does not equate to saying he believed it will happen. And from other passages we were shown it was clear that Dr Blatter thought a lot depended on whether and if so when, initial nucleation occurred.

44. Then there was the unsatisfactory position taken by Prof. Frampton about super-seeding. Example 4 has a very large amount of calcipotriol anhydrate in suspension and only a little calcipotriol in solution (because it is so insoluble). A large preponderance of seeds of a particular crystalline form can cause crystallisation of that form, even if another form, itself not present in the form of seeds, is thermodynamically favoured. So unless super-seeding could be ruled out, it was another reason why Example 4 would not inevitably produce the monohydrate.
45. Prof. Frampton sought to rule the possibility of super-seeding out. He said that it did not and could not happen in the case of pseudo-polymorphs. The Judge did not accept that, saying at [94] "I am not persuaded that the polymorph/polymorph transition is so different from the anhydrate/hydrate transition to enable one to rule out the presence of large quantities of anhydrate crystals in the suspension."
46. Mr Watson said the Judge had no basis for so holding. But he had lots. Prof. Frampton had not mentioned super-seeding in his written evidence at all. Yet there was occasion to. Dr Blatter had exhibited a paper (Dunitz and Bernstein) to his written evidence which said the formation of a critical nucleus is "decisive" and "may be delayed indefinitely." Prof. Frampton did not respond saying "but that only applies to polymorph/polymorph transitions". No papers showing Prof. Frampton's distinction was accepted science were produced. Dr Blatter had said that "where there is more than one possible crystal form, the form which crystallises is determined at the time of nucleation." When that was put to Prof. Frampton he suggested for the first time that was only true for polymorph to polymorph transition. But a little later he referred to super-seeding – and then said because the thermodynamically favoured form was monohydrate, only that could form. Super-seeding could not happen. "Your nucleation is enabled by the water". But water is not a nucleus. And Prof. Frampton accepted that smoke, dust or glass could act as a template for nucleation. So why not anhydrous calcipotriol? The whole reason why super-seeding works is because one has so much seed around. Moreover Sandoz's other expert, Prof. Williams clearly considered that super-seeding could happen with an anhydrate/hydrate transition.
47. Mr Carr had more but I need not go it. It is clear beyond peradventure that the Judge had ample material upon which to base his findings. We are light-years away from a perverse decision. The anticipation appeal is hopeless. It should never have been advanced.

Obviousness generally

48. Before going on to consider the four varieties of obviousness advanced, I would add a word about Sandoz's evidence. Prof. Frampton had no experience of vitamin D as of the date of the Patent and only a little from a much later time. He was a specialist in polymorph screening and the Judge politely de-valued the weight of his evidence, saying that "he may have approached matters with a predisposition to pursuing screening programs which the relevant skilled team might not possess." The Judge had ample basis so to hold. Indeed another Judge might have gone further – not only were there the errors I have detailed above but when Prof. Frampton read the documents showing how Leo actually found the monohydrate his predisposition was exposed because he wrongly assumed that Leo were engaged in some sort of screening process when they were not and the documents did not say they were.
49. It is noteworthy that Sandoz did not call any expert who was actually familiar with Vitamin D and its thousands (Leo alone had researched a 1,000) of analogues. Yet Sandoz clearly had access to such an expert – Mr Watson told us only a month before trial - but did not say they could not (or did not) find one earlier. It is a very striking thing that Sandoz did not produce a vitamin D analogue expert in a case about a vitamin D analogues.
50. The Judge found that all the witnesses were doing their best to assist. Accepting that to be so, it remains the case that in a number of respects Prof. Frampton's evidence was not accepted for good reason (see above). If it had been necessary (which it is not) for this Court to re-evaluate all the evidence, the inconsistencies and errors would have been something I myself might well have taken into account in assessing the reliability of his evidence on other points.

Issue (2): Obvious to do a full polymorph screen?

51. Mr Watson's case ran as follows. The Judge said:

[34] 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.

52. Given the possibility found by the Judge, Mr Watson submitted that a skilled team would carry out a full screen. A full screen is a systematic investigation, using a range of solvents and conditions, to discover whether a compound has polymorphs or pseudo-polymorphs, particularly hydrates. It is a polymorph hunt. And if it were carried out, the skilled team would probably find the hydrate.
53. Mr Watson accepted that there would, on the evidence, be no expectation of finding a hydrate, still less a hydrate which had better properties than the anhydrate. But, he submitted, the purpose of a full screen was not to find a better crystalline form. It was to find out if there were any other such forms – knowing there were none was as important as knowing about them if there were. It was, as he put it, a "win-win" investigation.
54. So, he submitted, the Judge fell into an error of principle in rejecting the full-screen case. What the Judge said was:

[137 … 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.

55. The error of principle was said to be that the Judge overlooked the fact that the motive for a full screen is not an expectation of success – it is simply to find out about other possible crystal forms.
56. Mr Watson had this difficulty, however, namely that the clear evidence was that no company in the vitamin D analogue field, actually did full screening at the time. Dr Rasmussen knew them, mentioning specifically Chugai, Teijin and Roche as well as Bristol-Myers-Squibb, Leo's licensees.
57. So in the real world full screening was not undertaken by real skilled teams. They did not work on Mr Watson's "win-win" basis. And there were good reasons – full screening was not required by the regulators, crystallising the analogues was very hard indeed if they crystallised at all (only a few out of 1,000 investigated at Leo crystallised) and past experience had not indicated it was worthwhile for any purpose.
58. Mr Watson sought to get round this by resorting to a notional skilled team which did not include anyone with practical experience of vitamin D analogues. He submitted that the Judge had made a finding at [34] that "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." That was not challenged. Given the "possibility" a full screen would be done.
59. Moreover he submitted, a notional skilled team would not necessarily include a person actually skilled in vitamin D analogues. The submission developed in stages:

(a) The Judge found at [21]:

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.

(b) The same goes for the prior art. Its audience includes teams of skilled people who might legitimately be interested in entering the field for the first time and they need not notionally have included a vitamin D specialist.

(c) Without such a specialist the novice notional team would get its knowledge about vitamin D and its analogues from the literature. There is nothing there about not doing full-screening.

(d) Moreover the Judge when he came to make findings of common general knowledge did not make any specific finding that not doing full-screening in this field was part of the common general knowledge.

(e) So this notional novice team, unaware that real teams had never found it necessary to full-screening would do it because full-screening is the default position for pharmaceuticals generally.

60. I am quite unable to accept these submissions. They fly in the face of reality. Once the Judge found (as he did) that it was not universal practice to conduct a polymorph screen and that a skilled team would not regard such a screen as mandatory, it seems to me that there was ample material for the Judge to reach the conclusion he did. He properly took all the relevant factors into account, weighed them carefully and cannot be faulted in any way. Mr Watson's submissions involve no more than taking some bits of the Judge's findings out of context and overlooking others. That will not do.
61. And the "novice team" point utterly lacks reality. It supposes this team seizes upon example 4, makes the cream, and then decides to do a general screen. But when? Surely only after a research programme to find out if the cream worked (the acne use patent offers no proof). Other creams might be investigated too. Somewhere down the line, if the line proved worthwhile (which was not shown) at best a screen might be done.
62. Even then it was not proved that the hydrate would inevitably be found. Dr Rasmussen accepted it probably would, which is not the quite the same thing. And Sandoz themselves did not initially find the hydrate in their own product, initially contending (on the basis of some experiments) that it was not present.
63. This case is much too speculative. The hydrate is simply not ob via – in the way - there was ample material for the Judge so to find. The full-screen obviousness attack fails.

Issue (3) Obviousness because a product screen would have revealed the monohydrate

64. The Judge held:

[131] .. 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.

This is a much more limited investigation than a full-screen. It is, as the Judge said:

[138] … 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.

65. The Judge went on to hold that it was not proved that such a limited investigation would reveal the anhydrate. Mr Watson says he was wrong. He particularly relied upon Dr Rasmussen's acceptance, quoted at [140], that monohydrate would probably be found in a screen. But it is quite clear that the sort of screen he was being asked about was a full-screen – he referred to acetone and water – solvents not used for example 4.
66. To prove this case it would be necessary to establish that the limited investigation actually necessary would reveal the hydrate. That simply was not undertaken. Sandoz's whole case was based on full-screening. You cannot get answers about that to read on to limited screening. Mr Watson complained that Leo never said that a limited screen would not reveal the anhydrate. But it was not for Leo to say that unless and until the opposite was clearly advanced by Sandoz, which it was not.
67. Moreover one again cannot but be impressed by Sandoz's own initial failure to find the hydrate in their own product – which had a limited screen of some sort.
68. The Judge had ample material for his conclusion. The appeal on this point fails too.

Issue (4): Obvious to wet-mill when making example 4

69. The Judge held that example 4 only disclosed dry milling. There is no challenge to that. But it is said that an obvious variant is wet-milling and if one wet-mills one will get the hydrate. So one would bump into the invention almost by accident.
70. The Judge accepted that wet-milling was indeed an obvious variant. But he did not accept that it would produce the hydrate. Sandoz did not try to prove it did by direct experiment. Reliance was placed on what was done at Leo. But at Leo it only happened 50% of the time. And Dr Rasmussen did not accept that wet-milling would necessarily produce the hydrate. He said: "this is something which is, something you have to investigate". That is miles from obviousness.
71. The Judge dealt with the argument at [154]:

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.

72. I cannot begin to see a fault here, still less one of principle.

Issue (5): Obviousness over common general knowledge

73. The case here is simply that routine crystallisation experiments would have produced the hydrate. The Judge rejected this in a single paragraph:

[155] 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.

74. Prof. Frampton had advanced an argument that the skilled man, having got the anhydrate, would conduct some sort of crystal growing experiments to see if larger crystals could be obtained for the purpose of X-ray diffraction examination and that the hydrate would be found in the course of that.
75. To support his argument he relied on how Leo had found the hydrate – as I have already said he was mistaken about the purpose. So that support fell away.
76. As to the attempt to grow larger crystals, there was no evidence this was done in the vitamin D field and the nature of the experiments was not identified. I cannot discern the beginnings of an error of principle here.

Conclusion

77. This appeal falls to be dismissed. I only add a couple of points. This is yet another case where validity has to be assessed by several national courts. We have reached the same result as that in Holland at first instance (where the argument was in part different). The Bundespatentgericht has gone the other way – but working on different prior art, prior art which Sandoz in this country abandoned. Different results in different countries based on different cases is, of course, explicable. It is an unfortunate state of affairs, curable only by a single European Patent Court.

Lord Justice Patten:

78. I agree.