England and Wales High Court (Patents Court) Decisions

IN THE HIGH COURT OF JUSTICE
CHANCERY DIVISION
PATENTS COURT

		Rolls Building Fetter Lane, London, EC4A 1NLL
		27 April 2015

MR JUSTICE ARNOLD :

Contents

Topic	Para
Introduction	1-3
Witnesses	4-10
Novartis' experts	5-7
The Defendants' experts	8-10
Technical background	11-33
Alzheimer's disease (AD)	11-12
Acetylcholine (ACh)	13-15
Treatment of AD with AChE inhibitors	16-19
Rivastigmine	20-21
The Exelon SmPC	22
Pharmacokinetic parameters	23-24
Transdermal therapeutic systems (TTS)	25-30
Development of a patch formulation	31-33
The Application	34-61
The Patent	62-68
The claim	69
The skilled team	70-71
Common general knowledge	72-93
Construction	94-102
Added matter	103-114
Obviousness	115-134
US031	116-122
The difference	123
Was it obvious?	124-134
Insufficiency	135-142
Infringement	143-149
Summary of conclusions	150

Introduction

1. The Claimants ("Novartis") claim that the Defendants in these two actions ("Focus", "Actavis" and "Teva") have infringed European Patent (UK) No. 2 292 219 entitled "Transdermal therapeutic system for the administration of rivastigmine" ("the Patent"). The Patent is directed to rivastigmine for use in a method of treatment of Alzheimer's disease ("AD") wherein the rivastigmine is administered by a transdermal therapeutic system, and in particular a transdermal patch. The Defendants deny infringement and counterclaim for revocation of the Patent on the grounds of added matter, obviousness and insufficiency. Although the Defendants challenged the claimed priority date of 1 December 2005 in order to contend that the Patent was anticipated by the application from which it derived, I do not consider that this challenge requires separate consideration from the added matter attack.
2. The Patent is one of a chain of divisionals deriving from International Patent Application No. WO 2007/064407 ("the Application"). The Patent has been opposed by no less than 13 opponents. On 19 February 2015 the Opposition Division of the European Patent Office issued a summons to oral proceedings in which it expressed the provisional opinion that the Patent was invalid on the grounds of added matter and insufficiency. The oral proceedings are due to be heard on 15 to 17 December 2015.
3. Other designations of the Patent, and equivalent national rights, have been extensively litigated in proceedings in other jurisdictions, namely Austria, Belgium, Denmark, Germany (Düsseldorf, Mannheim and Munich), Italy, the Netherlands, Poland, Portugal, Romania, Slovenia and Spain. Most of these proceedings have been applications for interim injunctions, although there have been final decisions on infringement in Mannheim and Portugal. There has been a striking diversity of outcomes, with some courts and tribunals finding infringement and others not on the same facts. Until now there has been no final decision on validity, although the Court of Appeal of the Hague and the Mannheim Regional Court have expressed the opinion that the Patent is probably invalid for added matter and the Düsseldorf Regional Court has expressed the same opinion about an equivalent utility model.

Witnesses

4. Each side called two expert witnesses, a clinician or neuroscientist and a transdermal formulator. There was a large measurement of agreement between the respective pairs of witnesses. In addition, Novartis adduced unchallenged factual evidence from Dr Nico Reum concerning the composition of Novartis' Exelon patch.

Novartis' experts

5. Professor Clive Ballard is Professor of Age Related Diseases and Co-Director of the Wolfson Centre for Age Related Diseases at King's College London, Director of the Biomedical Research Unit for Dementia and Honorary Consultant at the South London and Maudsley NHS Foundation Trust. He obtained a BMBS degree from the University of Leicester in 1987, a Masters in Psychiatry from the University of Birmingham in 1992 and an MD from Leicester in 1995. During this period he also completed his medical training at various hospitals in the Midlands and was a Lecturer in Psychiatry at Birmingham. Between 1995 to 2003 he was successively Senior Lecturer, Reader and Professor of Old Age Psychiatry at the University of Newcastle and Consultant in Old Age Psychiatry at Newcastle General Hospital. In 2003 he took up his current positions. He has held numerous other positions, including Director of Research for the Alzheimer's Society UK from 2003 to 2013. He has published more than 500 publications, including four books in the area of dementia and AD. He has also acted as a consultant to a number of pharmaceutical companies, including Novartis in relation to its Exelon patch.
6. Counsel for the Defendants submitted that Prof Ballard had adopted an overly cautious approach to the development of a new formulation of rivastigmine and had been inconsistent in his approach to the Patent and the prior art. I am sure that Prof Ballard was doing his best to assist the court, but I think there is force in both these points, and I shall take them into account in assessing the evidence.
7. Professor Marc Brown is Chair of Pharmaceutics at the School of Pharmacy at the University of Hertfordshire. He obtained a BSc in Medicinal and Pharmaceutical Chemistry from Loughborough University of Technology in 1990 and a PhD from the same institution in 1993. From 1993 to 1995 he was a Post-Doctoral Research Fellow, and from 1996 to 2006 he was a Lecturer and then a Senior Lecturer, at the Department of Pharmacy at King's College London. He took up his current position in 2006. In addition to his academic work, in 1999 Prof Brown co-founded MedPharm Ltd, a contract research organisation specialising in the formulation of topical and transdermal medicines, which he continues to be Chief Scientific Officer and Chief Operating Officer of. No criticism was made of his evidence.

The Defendants' experts

8. Professor Paul Francis is Professor of Neurochemistry at the Wolfson Centre for Age Related Diseases at King's College London. He obtained a BSc in Physiology and Biochemistry from the University of Reading in 1979 and a PhD in Neuroscience from the same institution in 1984. From 1982 to 1990 he was a Post-Doctoral Research Fellow, and from 1990 to 1995 he was an Honorary Lecturer, at the Institute of Neurology, Queen Square. In 1995 he became Senior Lecturer in Biochemistry and Molecular Biology at the United Medical and Dental Schools of Guy's and St Thomas' Hospitals, which merged with King's College in 2000. He was appointed Reader in 2004 and Professor in 2008. Since 2008 he has also been Director of Brains for Dementia Research, a partnership between Alzheimer's Research UK and the Alzheimer's Society. He has published 138 peer-reviewed articles and about 60 book chapters and reviews in this field. He has spoken at meetings and symposia for, and received research funding from, various pharmaceutical companies, including from Novartis in relation to rivastigmine.
9. Counsel for Novartis submitted that Prof Francis' field of expertise was less appropriate to this case than that of Prof Ballard, since Prof Francis had no clinical experience, nor had he carried out any pharmacokinetic studies. Furthermore, as at 2005, he had not done any research on the efficacy or side effects of AChE inhibitors in man. On the other hand, he was more knowledgeable about the biochemical mechanisms of action of such inhibitors than the relevant member of the skilled team. I agree that Prof Ballard had more directly relevant expertise than Prof Francis, but nevertheless I consider that Prof Francis was in a position to assist the court.
10. Professor Adrian Williams is Professor of Pharmaceutics and Head of the School of Chemistry, Food and Pharmacy at the University of Reading. He obtained a BSc in Combined Sciences, Biology and Chemistry in 1987 and a PhD in Pharmaceutical Technology from the University of Bradford in 1990. From 1990 to 2004 he was successively Lecturer, Senior Lecturer and Reader in Pharmaceutical Technology and Professor of Biophysical Pharmaceutics at the same institution. Since 2004 he has been a Professor at Reading. He has published 89 peer-reviewed articles and many reviews and book chapters, including a chapter entitled "Topical and transdermal drug delivery" in the well-known textbook Aulton's Pharmaceutics: The Design and Manufacture of Medicines (M.E. Aulton and K.M.H. Taylor eds, 4^th ed, 2013). He sits on the Scientific Advisory Board of MedPharm and has acted as consultant to pharmaceutical companies. No criticism was made of his evidence.

Technical background

Alzheimer's disease (AD)

11. AD is a progressive neurodegenerative disease that is the most common cause of dementia. There are approximately 750,000 dementia sufferers in the UK, of whom more than 520,000 have AD. Symptoms of AD include short-term memory loss, impairments of attention and concentration, loss of executive function, impairment of language ability, inability to eat or dress and other mood disorders such as depression, aggression and agitation. The disease is terminal, with a variable natural course of between 4-10 years from diagnosis to death.
12. The neuropathy of Alzheimer's disease was well known at the priority date. During the course of the disease, "plaques" (extracellular deposits of beta-amyloid peptide) and "tangles" (intracellular deposits of filaments composed of a hyper-phosphorylated protein known as tau) build up in the brain. To date, no effective treatment to address these underlying neuropathological signs of Alzheimer's disease has been approved.

Acetylcholine (ACh)

13. One of the known symptoms of AD is associated with degeneration of the cholinergic neurons that use the neurotransmitter acetylcholine (ACh). Degeneration of the cholinergic neurons leads to a loss of ACh in the brain. This loss of ACh was known to be a key underlying mechanism of the cognitive dysfunction associated with AD. ACh is active both in the central nervous system (CNS) and the peripheral nervous system (PNS).
14. ACh is synthesised within presynaptic cells (cholinergic terminals) by the enzyme choline acetyltransferase (ChAT) from choline and acetyl coenzyme A (acetyl-CoA). Once released, the acetylcholine may then interact with muscarinic receptors or nicotinic receptors on postsynaptic and presynaptic cells. The activation of post-synaptic muscarinic and nicotinic receptors leads to the activation of biochemical pathways or depolarisation of the target cell, and thus the propagation of the nerve impulse.
15. Acetylcholine is quickly inactivated by acetylcholinesterase (AChE), which breaks down acetylcholine to acetate and choline, thereby preventing ACh from binding to and activating acetylcholine receptors. (Another enzyme called butyrylcholinesterase is also sometimes involved.) One way of addressing the degeneration of the neurons which use ACh in AD patients was therefore to inhibit AChE, thus enhancing cholinergic transmission. Various cholinesterase inhibitors, including rivastigmine, had been approved for the treatment of AD before the priority date.

Treatment of AD with AChE inhibitors

16. At the priority date, four drugs were being used to treat mild to moderate AD, three of which were AChE inhibitors (a fourth AChE inhibitor, tacrine, was rarely used due to its potential liver toxicity). Inhibition of AChE leads to increased levels of ACh, as the rate of breakdown of the neurotransmitter is attenuated. The benefits of AChE inhibitors were known to last for one to two years, so the period for their effective use was (and still is) relatively narrow.
17. The three AChE inhibitors in use at the priority date, all of which were orally administered, were as follows:

i) Donepezil (brand name Aricept). Donepezil was started at 5 mg (tablets) once daily, and after one month the dose was increased to 10 mg once daily. 5 mg once daily was the minimum effective dose.

ii) Galantamine (brand name Reminyl). Galantamine was started at 4 mg (tablets) twice daily. At four week intervals, each dose was increased by 4 mg, up to a dosage of 12 mg twice daily. 8 mg twice daily was the minimum effective dose.

iii) Rivastigmine (brand name Exelon). Rivastigmine was started at 1.5 mg (capsules or oral solution) twice daily. At intervals of a minimum of two weeks (although in practice four week intervals were commonly used), each dose was increased by 1.5 mg up to a dosage of 6 mg twice daily. 3 mg twice daily was the minimum effective dose.

18. At the priority date, donepezil was the market leader in the UK, with galantamine and rivastigmine generally considered second line treatments. This may have been partly because donepezil was approved before the other two, but donepezil was perceived to have a number of significant advantages over rivastigmine:

i) donepezil only required a single daily dose, rather than twice daily doses, which resulted in higher patient compliance and carer convenience;

ii) administration of donepezil started with a clinically therapeutic dose, whereas the initial dose of rivastigmine was sub-therapeutic, which both delayed the therapy and meant that patients might experience side-effects without efficacy in the early stages of treatment;

iii) rivastigmine had a more complicated titration regime; and

iv) rivastigmine was perceived to have worse side effects.

19. All three drugs were associated with mild to moderate cholinergic side effects such as nausea, vomiting and diarrhoea, which varied from patient to patient. Patients were titrated up gradually from lower doses to enable them to acquire tolerance to those side effects. If this was done, most patients could tolerate a therapeutically effective dose of the AChE inhibitor even if this was not the maximum therapeutic dose available. However, that tolerance was quickly lost if treatment was interrupted for more than a few days and the patient would need to be re-titrated. Serious adverse events were rare.

Rivastigmine

20. Rivastigmine is a pseudo-irreversible AChE inhibitor, unlike donepezil and galantamine, which are both reversible inhibitors. Reversible inhibitors mimic the structure of ACh, bind weakly and reversibly with the active site of the enzyme and compete with ACh for occupancy of the active site. Although the reversible inhibitors only bind to the active site for a relatively short time, the therapeutically useful reversible inhibitors have a relatively long plasma half-life and can therefore continue to interact with AChE, thereby blocking the binding of ACh over a prolonged period of time.
21. Rivastigmine binds to the active site weakly and then is hydrolysed by AChE, thereby forming a covalent bond between the enzyme and rivastigmine. This is similar to the process that occurs with the natural substrate ACh, save that the carbamylated enzyme after hydrolysis of rivastigmine is much more stable than after hydrolysis of ACh, leading to much longer inactivation of the enzyme. This explains why rivastigmine has a relatively long effect in vivo (a single dose inhibits ACh in the brain for about 9 hours), notwithstanding that its half-life in blood plasma is only about 1 hour. Rivastigmine was known to have low oral bioavailability: the bioavailability of rivastigmine after administration of a 3 mg oral capsule is approximately 36%.

The Exelon SmPC

22. The Summary of Product Characteristics (SmPC) for Exelon capsules, which dates from 2000, sets out much of the key information that was known about rivastigmine in 2005. The following passages are particularly relevant for present purposes (emphasis added):

"4. CLINICAL PARTICULARS

…

4.2 Posology and method of administration

…

Rivastigmine should be administered twice a day, with morning and evening meals. The capsules should be swallowed whole.

Initial dose: 1.5 mg twice a day.

Dose titration: The starting dose is 1.5 mg twice a day. If this dose is well tolerated after a minimum of two weeks of treatment, the dose may be increased to 3 mg twice a day. Subsequent increases to 4.5 mg and then 6 mg twice a day should also be based on good tolerability of the current dose and may be considered after a minimum of two weeks of treatment at that dose level.

If adverse effects (e.g. nausea, vomiting, abdominal pain or loss of appetite) or weight decrease are observed during treatment, these may respond to omitting one or more doses. If adverse effects persist, the daily dose should be temporarily reduced to the previous well-tolerated dose.

Maintenance dose: The effective dose is 3 to 6 mg twice a day; to achieve maximum therapeutic benefit patients should be maintained on their highest well tolerated dose. The recommended maximum daily dose is 6 mg twice a day.

Maintenance treatment can be continued for as long as a therapeutic benefit for the patient exists. Therefore, the clinical benefit of rivastigmine should be reassessed on a regular basis, especially for patients treated at doses less than 3 mg twice a day. Discontinuation should be considered when evidence of a therapeutic effect is no longer present. Individual response to rivastigmine cannot be predicted.

….

Re-initiation of therapy: If treatment is interrupted for more than several days, it should be re-initiated at 1.5 mg twice daily. Dose titration should then be carried out as described above.

…

4.4 Special warnings and special precautions for use

The incidence and severity of adverse events generally increase with higher doses. If treatment is interrupted for more than several days, it should be re-initiated at 1.5 mg twice daily to reduce the possibility of adverse reactions (e.g. vomiting).

Dose titration: Adverse effects (e.g. hypertension, hallucinations) have been observed shortly after dose increase. They may respond to a dose reduction. In other cases, Exelon has been discontinued (see 4.8 Undesirable effects).

Gastrointestinal disorders such as nausea and vomiting may occur particularly when initiating treatment and/or increasing the dose. These adverse events occur more commonly in women. Patients with Alzheimer's disease may lose weight. Cholinesterase inhibitors, including rivastigmine, have been associated with weight loss in these patients. During therapy patient's weight should be monitored.

…

As with other cholinergic substances, rivastigmine may cause increased gastric acid secretions. Care should be exercised in treating patients with active gastric or duodenal ulcers or patients predisposed to these conditions.

...

4.8 Undesirable effects

The most commonly reported adverse drug reactions are gastrointestinal, including nausea (38 %) and vomiting (23 %), especially during titration. ...

….

4.9 Overdose

…

Treatment: As rivastigmine has a plasma half-life of about 1 hour and a duration of acetylcholinesterase inhibition of about 9 hours, it is recommended that in cases of asymptomatic overdose no further dose of rivastigmine should be administered for the next 24 hours. In overdose accompanied by severe nausea and vomiting, the use of antiemetics should be considered. ...

…

5. PHARMACOLOGICAL PROPERTIES

5.1 Pharmacodynamic properties

…

Rivastigmine interacts with its target enzymes by forming a covalently bound complex that temporarily inactivates the enzymes. In healthy young men, an oral 3 mg dose decreases acetylcholinesterase (AChE) activity in CSF by approximately 40% within the first 1.5 hours after administration. Activity of the enzyme returns to baseline levels about 9 hours after the maximum inhibitory effect has been achieved. In patients with Alzheimer's disease, inhibition of AChE in CSF by rivastigmine was dose-dependent up to 6 mg given twice daily, the highest dose tested. …

…

Absorption: Rivastigmine is rapidly and completely absorbed, Peak plasma concentrations are reached in approximately 1 hour. As a consequence of the drug's interaction with its target enzyme, the increase in bioavailability is about 1.5-fold greater than that expected from the increase in dose. Absolute bioavailability after a 3 mg dose is about 36%±13%. Administration of rivastigmine with food delays absorption (t_max) by 90 min and lowers C_max and increases AUC by approximately 30%.

…

Metabolism: Rivastigmine is rapidly and extensively metabolised (half-life in plasma approximately 1 hour), primarily via cholinesterase-mediated hydrolysis to the decarbamylated metabolite. In vitro, this metabolite shows minimal inhibition of acetylcholinesterase (<10%). …

Excretion: Unchanged rivastigmine is not found in the urine; renal excretion of the metabolites is the major route of elimination. Following administration of 14C-rivastigmine, renal elimination was rapid and essentially complete (> 90 %) within 24 hours. … There is no accumulation of rivastigmine or the decarbamylated metabolite in patients with Alzheimer's disease."

Pharmacokinetic parameters

23. The pharmacokinetic parameters referred to in the SmPC are as follows:

i) C_max is a measure of the peak plasma concentration of the drug;

ii) t_max is the time at which C_max is reached;

iii) AUC or "area under the curve" is the area under the concentration-time curve. This reflects the actual body exposure to a drug after administration. AUC_24h is the AUC over a 24 hour period.

24. These are all measurements of the drug in the blood. What matters for a clinical effect is the measure of interaction at the active site and that is not informed solely by looking at these parameters. The pharmacokinetics of rivastigmine illustrate the point: although it has a short half-life in plasma of about an hour, its clinical effect lasts for longer because it binds covalently to AChE and the enzyme is only reactivated after that covalent bond is broken. Additionally, as rivastigmine has its effect in the brain, it is its level there that matters, not its level in plasma.

Transdermal therapeutic systems (TTS)

25. A transdermal therapeutic system (TTS), of which a transdermal patch is the most common form, is applied to the skin for delivering a drug into the blood stream. The first transdermal patch was approved in 1979. By the priority date, there were 28 FDA-approved transdermal patches on the market with about 10 different active ingredients. There were no TTSs available for the treatment of AD at the priority date, however.
26. Administration by transdermal means has a number of advantages over oral formulations:

i) it bypasses the gastro-intestinal (GI) system, thereby avoiding issues related to metabolism, degradation or absorption, and locally-mediated side effects, in the GI tract;

ii) it avoids the first-pass hepatic metabolism;

iii) it provides a means of once daily dosing when that may be impossible with an oral formulation;

iv) it provides a smoother delivery curve, avoiding the rapid fluctuation and peak levels of drug plasma concentrations that are seen with rapidly-absorbed drugs provided by oral administration, and will therefore improve tolerability if there are side effects associated with sharp/high peaks. This is illustrated by the following diagram in Professor Williams' first report:



v) a patch can readily be removed from a patient if required.

27. Administration by transdermal means also has a number of disadvantages:

i) only a limited number of active pharmaceutical ingredients (APIs) are suitable for administration by transdermal patch;

ii) TTSs are generally more expensive and time consuming to develop than oral formulations;

iii) the onset of treatment tends to be slower than with oral formulations; and

iv) there is the potential for local skin irritation.

28. The main features of a transdermal patch are as follows:

i) An outer backing layer. This is the layer that is most remote from the skin. It holds the patch together, protects it from rubbing off on clothes and prevents cross-contamination when handling the patch. The backing layer should be impermeable to the API and typically be impermeable to water.

ii) A preparation containing the API. This contains the API along with excipients.

iii) A release liner. The release liner is a detachable film that is impermeable to the API which is removed before the patch is applied to the skin.

29. In order to provide a patient with a fixed and reproducible dose of API over the prescribed period of application, the rate of release should be constant. This is achieved by ensuring that the API in the patch is saturated for the entire time of the application, or at least as close as possible to saturation for the entire period of application. Because of this, the design of transdermal patches means that not all the dose loaded into the patch is released during the application period. The excess of API in the patch is needed to maintain a constant rate of release over time.
30. A skilled transdermal formulator would be aware of the terms "loaded dose" (the amount of API put into the patch) and "released dose" (the amount of API that leaves the patch over a specified period) and their meanings. It is generally the released dose that is taken into account by clinicians.

Development of a patch formulation

31. Once an API is identified as being suitable for delivery by a patch, a target dose would be given to the formulator by the clinician and would generally be quantified in terms of a target AUC, typically the AUC_24h. The target AUC_24h will normally correspond to the AUC_24h for the oral dosage form.
32. Once the target dosage is selected, prototypes will be produced and tested in short term accelerated and real time stability tests and their performance will then be assessed based upon the amount of API released from the patch. This is tested in vitro in a test system such as a Franz diffusion cell and, ultimately, in vivo, usually in humans. The Franz diffusion cell test involves measuring diffusion of the test substance across a membrane. This may be a piece of human skin or an artificial membrane such as an EVA membrane.
33. In developing a generic patch formulation (i.e. a generic version of an existing patch), a key step is to ensure that the generic patch will release the same dose as the originator patch. This would be tested in the same way as described above – in vitro by Franz diffusion cell and ultimately in human volunteers.

The Application

34. The Application is entitled "Transdermal Therapeutic System". It begins with the following statement (at page 1):

"The present invention relates to Transdermal Therapeutic Systems comprising a backing layer, a reservoir layer and an adhesive layer, to Transdermal Therapeutic Systems having specific release profiles and their manufacture and use."

35. It is convenient to note straightaway that it is clear from page 4 of the Application that, as the skilled formulator would expect, the Transdermal Therapeutic System or TTS will also include what the Application refers to as a "detachable protective layer" (i.e. a release liner), but for the most part this feature of the TTS is ignored.
36. The Application goes on to identify eight objectives of the invention, including the provision of a TTS with "improved compliance, adhesion, tolerability and/or safety properties", the provision of a method of treatment that "substantially improves the efficacy and tolerability of rivastigmine" and the provision of a method of treatment and controlled-release formulations that "have substantially less inter-individual variation with regard to plasma concentrations of rivastigmine required to produce a therapeutic benefit without unacceptable side effects" (at pages 1-2). These objectives are said to be achieved by a TTS "as defined in claim 1 and the depending claims" of the Application (at page 2).
37. After introducing the Figures, the Application states (at page 2):

"Tests with active ingredients for the treatment of Alzheimer's disease have surprisingly shown that a line of silicone adhesive can be applied to a poorly adhesive reservoir matrix, thus significantly increasing the adhesive properties of the preparation without affecting the thermodynamic properties of the TTS, i.e. without reducing the release of active ingredient from the matrix and its permeation through the skin."

38. A first aspect of the invention of the Application is then set out (at page 3):

"The present invention provides TTS comprising a backing layer, a reservoir layer containing at least one active ingredient and a polymer, an adhesive layer comprising a silicone polymer and a tackifier.

A TTS according to the invention shows improved adhesive properties. Further, and very surprisingly, the so obtained TTS has essentially the same release profile when compared to a standard TTS."

39. This disclosure is not limited to rivastigmine, but applies to all of an extensive range of active ingredients set out at pages 4 to 6 of the Application.
40. The Application then sets out a second aspect of the invention (at page 3):

"The present invention is further related to a method for substantially improving the efficacy and tolerability of rivastigmine, comprising application of a TTS in the range of about 2 to 50 cm², said formulation providing a mean maximum plasma concentration of about 1 to 30 ng/mL from a mean of about 2 to 16 hours after application and an AUC_24h of about 25 to 450 ng.h/mL after repeated 'QD' (i.e., once daily) administration.

A TTS according to the invention quite surprisingly shows improved tolerability, particularly gastrointestinal adverse events such as nausea and vomiting, relative to equivalent levels of exposure (AUC_24h) of Exelon® capsule."

41. It should be noted that the pharmacokinetic data are not said to relate to the starting dose prescribed to a patient. Indeed, it can be seen from Table 1 of the Application that the highest figures in the ranges of C_max and AUC_24h of the TTS claimed (i.e. 30 ng/ml and 450 ng.h/ml, respectively) equate to the same C_max and three times the AUC_24h as the highest oral dose (6 mg bid i.e. bis in diem or twice daily).
42. These statements of the invention are reflected in the claims of the Application. Claims 1 to 14 relate to a TTS "comprising (a) a backing layer, (b) a reservoir layer comprising one or more pharmaceutically active ingredients and one or more polymers, (c) an adhesive layer comprising a silicone polymer and a tackifier". Claims 15 to 18 relate to a TTS providing rivastigmine C_max and AUC_24h values within the broad ranges disclosed in the specification.
43. After several pages of definitions, the Application proceeds to disclose a series of preferred and particularly preferred embodiments. Many of these concern aspects of the silicone adhesive layer and the reservoir layer, and other physical properties, of the three-layered patch aspect of the invention.
44. One of the preferred embodiments is described as follows (at page 7):

"In a preferred embodiment, the TTS provides a mean maximum plasma concentration of rivastigmine of 1 to 30 ng/ml from a mean of 2 to 16 hours after application with an AUC_24h of 25 to 450 ng.h/ml, particularly preferred, the TTS provides a mean maximum plasma concentration of rivastigmine of 2.5 to 20 ng.ml from a mean of 4 to 12 hours after application with an AUC_24h of 45 to 340 ng.h/ml."

45. Again, there is no suggestion in the text that these ranges are related to the starting dose, and for the reasons set out above this is not so.
46. After elaborating further aspects of the invention, the Application states (at page 9):

"Little has been published in detail on rivastigmine's biopharmaceutical properties in humans. It is rapidly and completely absorbed. We have found that it is metabolised mainly through hydrolysis by esterases, e.g. acetyl and butyryl cholinesterase and has a plasma half life of 1 hour. It is subject to pre-systemic and systemic metabolism. We now have found that a TTS containing rivastigmine may be produced with advantageous properties, e.g. better tolerability."

47. This is followed by consistory clauses which repeat the statements made previously about C_max, t_max and AUC_24h. Thus the better tolerability being referred to here is that produced by patches delivering plasma pharmacokinetic profiles in the claimed ranges previously discussed, and for the same reasons is not a disclosure about the starting dose given in a regimen.
48. The Application goes on (at pages 9-10):

"A person skilled in the art is familiar how to produce a TTS having the above defined plasma profiles. A person skilled in the art will appreciate that such plasma profiles may be obtained by varying, e.g.:

the composition of the first and/or second components, e.g., the nature and amount of excipients and/or active agent(s)

the type of the adhesive layer

the dimension of the patch

A TTS may be formulated with the following aspects in mind:

the time until the release of active agent (lag time or delay time)

the rate of release of the active agent (fast or slow)

the duration of release of active agent (long or short)

Reducing first-pass metabolism

Improve [sic] compliance of the patients

Reduce [sic] application intervals

Such aspects may be observed in standard in vitro dissolution tests, e.g., in water or if desired in body fluids, e.g., artificial gastric juices."

49. At page 11 the Application states:

"The TTS of the invention allows, e.g., the manufacture of once a day pharmaceutical forms for patients who have to take more than one dose of an active agent per day, e.g., at specific times, so that their treatment is simplified. With such compositions tolerability of rivastigmine may be improved, and this may allow a higher starting dose and a reduced number of dose titration steps.

A [sic] increased tolerability of rivastigmine provided by the compositions may be observed in standard animal tests and in clinical trials."

50. Three points should be noted about this passage. First, this is the only place in the entire Application where there is any reference to a "starting dose". Secondly, the reference to the "TTS of the invention" must be a reference to a TTS of the specified three-layer structure. Thirdly, this passage does not state that all rivastigmine-containing TTSs of the invention will have improved tolerability and will allow a higher starting dose and a reduced number of dose titration steps, but rather that some may do.
51. The Application then describes an example. In section I, the compositions of two patches referred to as TTS #1 and TTS #2 are set out (at page 12). TTS #1 has a reservoir layer with the following composition:

Rivastigmine (free base)	30 wt-%
Durotak® 387-2353 (polyacrylate adhesive)	49.85 wt-%
Plastoid® B (acrylate copolymer)	20.0 wt-%
Vitamine [sic] E	0.1 wt-%

52. TTS #2 has the same reservoir layer and an additional silicone adhesive layer with the following composition:

Bio-PSA Q7-4302 (silicone adhesive)	98.9 wt-%
Silicone oil	1.0 wt-%
Vitamine [sic] E	0.1 wt-%

53. Section II compares the adhesive forces of TTSs #1 and #2. TTS #1 is described as having a "low adhesive force", whilst that of TTS #2 is recorded as being about five times higher (at pages 12-13).
54. Section III compares the permeation through human skin and through an EVA membrane of rivastigmine from TTS #1 and TTS #2 in a Franz diffusion cell (at pages 13-14). The results are shown in Figures 2 (human skin) and 3 (EVA). I reproduce these below.

55. The Application comments that the slight differences between the permeation rates shown in Figure 2 are likely to be due to the use of a biological material and that Figure 3 confirms that the TTSs do not differ in their permeation properties. It is said to be surprising that the additional layer of silicone adhesive does not alter the permeation properties.
56. Section IV is entitled "Pharmacokinetic properties" and reports a small "dose proportionality study" into the pharmacokinetic properties of TTS #2 (at pages 14-18). The study is of four different sizes of TTS #2 (5 cm², 10 cm², 15 cm² and 20 cm²) and of the four standard oral doses of rivastigmine capsules (1.5, 3, 4.5 and 6 mg bid). The patients had mild to moderate Alzheimer's disease. The patients were put on 14 day ascending titration steps starting with the lowest dose. The study was a parallel-group study, i.e. there was no cross-over between the capsule and patch groups. Previous use of rivastigmine is not said to have been an exclusion criterion.
57. It is not said how many patients were enrolled in either arm of the study, and it is clear that not all of the patients had completed the study by the time that the Application was filed. The Application states that "the following number of patients completed each of the four periods, and were included in the pharmacokinetic evaluation":

58. The pharmacokinetics of rivastigmine were studied on the last (14^th) day of treatment at each dose level, save for the highest in each arm, where investigations were performed on the third day. The results are presented in Tables 1 and 2, and graphically in Figure 4. The AUC_24h data in Tables 1 and 2 may be summarised as follows:

Capsule	AUC24h	Patch	AUC24h
1.5 mg bid (3 mg)	12.3 + 7.41	5 cm2 (9 mg loaded dose)	45.6 + 16.6
3 mg bid (6 mg)	52.7 + 20.2	10 cm2 (18 mg loaded dose)	123 + 41.0
4.5 mg bid (9 mg)	90.4 + 45.1	15 cm2 (27 mg loaded dose)	226 + 85.5
6 mg bid (12 mg)	150 + 58.8	20 cm2 (36 mg loaded dose)	339 + 138

59. I reproduce Figure 4 below:

60. The study does not purport to be a clinical trial of tolerability or efficacy and does not report any qualitative or quantitative data on tolerability or efficacy. Nor does it state any conclusions about the (relative or absolute) tolerability of any of the tested formulations. Nor does it provide specific support for the plasma pharmacokinetic profiles set out in the body of the specification, although those ranges include within them all of the tested patches.
61. Section V consists of the bald statement that "TTS #2 shows improved pharmacological properties when compared with a capsule formulation as shown in standard animal test and in clinical trials" (page 15). No data, description or explanation is provided.

The Patent

62. When publishing the B1 specification of the Patent, the EPO introduced a number of typographical errors in the text of the specification when processing the Druckexemplar. The EPO has since issued a corrected B9 specification of the Patent. Unfortunately, it also introduced a change into the paragraph numbering. I shall refer to the text of the B9 specification, but using the paragraph numbering of the B1 specification.
63. The specification begins with the statement (at [0001]):

"The present invention relates to rivastigmine, in free base or pharmaceutically acceptable salt form, for use in a method of preventing, treating or delaying progression of dementia or Alzheimer's disease, wherein the rivastigmine is administered in a Transdermal Therapeutic System and the starting dose is as defined in claim 1."

64. It goes on to identify the same eight objectives as the Application. These are said to be achieved by a TTS "as defined in the enclosed claims [sic]" (at [0014]). The specification then continues:

"[0015] Embodiments of the present invention are set out in the enclosed independent claim.

[0016] In one aspect, the present invention provides rivastigmine, in free base or pharmaceutically acceptable salt form, for use in a method of preventing, treating or delaying progression of dementia or Alzheimer's disease, wherein the rivastigmine is administered in a TTS and the starting dose is as defined in claim 1."

65. The remainder of the description of the Patent is broadly similar to that of the Application, but there are a number of differences. The principal differences are as follows. First, whereas the Application said that "[t]he present invention provides" a TTS comprising a backing layer, a reservoir layer and an adhesive layer, and so on, the Patent instead refers to "one embodiment [of] the present disclosure" of the Patent doing so, to "a TTS according to the disclosure" or to "a TTS as used in the invention" etc (see, for example, [0019], [0020] and [0022]). Similarly, whereas in the Application there was reference to preferred embodiments having particular characteristics in relation to e.g. the reservoir or silicone adhesive layer, in the Patent the corresponding passages now refer to preferred embodiments in which the TTS comprises such a reservoir or silicone adhesive layer (see, for example, [0032] to [0043]).
66. Secondly, the extensive definition of "active ingredient" contained in the Application has been deleted from the Patent.
67. Thirdly, the passage concerning better tolerability (at [0049] corresponding to page 9 of the Application quoted in paragraph 46 above) is no longer followed by the consistory clauses concerning the specified pharmacokinetic profiles.
68. Fourthly, the passage referring to the starting dose (at [0057] corresponding to page 11 of the Application quoted in paragraph 49 above) now refers to the TTS "used in", rather than "of", the invention.

The claim

69. The Patent contains a single claim, claim 1. Broken down into integers, this is as follows:

"[1] Rivastigmine for use in a method of preventing, treating or delaying progression of dementia or Alzheimer's disease,

[2] wherein the rivastigmine is administered in a TTS and

[3] the starting dose is that of a bilayer TTS of 5 cm² with a loaded dose of 9 mg rivastigmine,

[4] wherein one layer: has a weight per unit area of 60 g/m² and the following composition:

- rivastigmine free base 30.0 wt %

- Durotak® 387-2353 (polyacrylate adhesive) 49.9 wt %

- Plastoid® B (acrylate copolymer) 20.0 wt %

- Vitamin E 0.1 wt %

[5] and wherein said layer is provided with a silicone adhesive layer having a weight per unit area of 30 g/m² according to the following composition:

- Bio-PSA® Q7-4302 (silicone adhesive) 98.9 wt %

- Silicone oil 1.0 wt %

- Vitamin E 0.1 wt %."

The skilled team

70. A patent specification is addressed to those likely to have a practical interest in the subject matter of the invention, and such persons are those with practical knowledge and experience of the kind of work in which the invention is intended to be used. The addressee comes to a reading of the specification with the common general knowledge of persons skilled in the relevant art, and he (or she) reads it knowing that its purpose is to describe and demarcate an invention. He is unimaginative and has no inventive capacity. In some cases, such as the present one, the patent may be addressed to a team of persons having different skills.
71. In the present case it is common ground that the Patent is addressed to a skilled team which is interested in developing a new formulation for rivastigmine. There is also no dispute that the team will contain a formulator skilled in the transdermal administration of drugs. Novartis contend that the other member of the team is a clinician working in the field of dementia, and in particular AD, whereas the Defendants contend that he is a neuroscientist working in that field. The dispute is less sharp than this might suggest, however, since Novartis accept that the clinician would call upon a pharmacologist where necessary to supplement his existing understanding of pharmacology and the Defendants accept that the skilled team would call upon a practising clinician as required. Furthermore, there is relatively little dispute as to the common general knowledge of the skilled team, and the disputes do not turn upon the precise qualifications of the second member of the team. Accordingly, I do not think anything turns upon those qualifications. Since I consider that clinical expertise is more directly relevant than pharmacological (or biochemical) expertise, however, I shall refer to the second member as the clinician.

Common general knowledge

72. I reviewed the law as to common general knowledge in KCI Licensing Inc v Smith & Nephew plc [2010] EWHC 1487 (Pat), [2010] FSR 31 at [105]-[115]. That statement of the law was approved by the Court of Appeal [2010] EWCA Civ 1260, [2011] FSR 8 at [6].
73. There is little, if any, dispute that everything I have set out in the technical background section of this judgment formed part of the skilled team's common general knowledge. There is, however, a substantial dispute as to the state of the skilled clinician's common general knowledge with regard to the manner in which the tolerability of rivastigmine could be increased (i.e. its side effects reduced).
74. The starting point is that it is common ground that the skilled clinician would be aware from the passages in the Exelon SmPC I have underlined in paragraph 22 above that (a) rivastigmine should be administered with food and (b) administration of rivastigmine with food delayed t_max by 90 minutes, reduced C_max by 30% and increased AUC by 30%. It is also common ground that the finding that food increased the AUC was puzzling.
75. The Defendants contend that it was common general knowledge that:

i) the side effects of rivastigmine were caused by sharp peaks in drug levels, indicated by short t_max and high C_max; and

ii) the recommendation to administer with food was given in order to improve tolerability.

Novartis dispute this.

76. In support of their contention, the Defendants rely upon eleven papers which had been published by the priority date, a number of which are review articles. I shall take them in chronological order.
77. The first is R. Anand, G. Gharabawi and A. Enz, "Efficacy and safety results of early phase studies with Exelon (ENA-713) in Alzheimer's disease: an overview", J. Drug Dev. Clin. Pract., 8, 109-116 (1996). This is a publication by authors from Sandoz (now Novartis). Under the heading "Tolerability and efficacy" the paper refers to three studies. The first study compared 4 mg/day and 8 mg/day with placebo, and found that the drug was well tolerated with high completion rates. The second study compared bid and tid (three times daily) dosing in the range 6-12 mg/day with placebo, and found that 12 mg/day could be tolerated by the majority of patients, but completion rates were lower than in the first study. The third study compared doses of up to 12 mg/day with placebo, and found that doses up to 12 mg/day were well tolerated, with the highest completion rate of the three studies. After summarising the second study, the paper states (at 113):

"However, a major difference between this study and a third study reported by Sramek at el [16] discussed here in the following pages is that, in the former case, the drug was given before food, whereas in the Sramek et al study, it was given after food. Data from other studies (Sandoz Pharma Ltd Basel, data on file) indicate that food has an appreciable effect on the plasma levels of ENA-713. Giving the drug after food reduces the C_max by approximately 30%, with a concurrent increase in AUC of approximately 30%. This could well point to a mechanism influencing intolerance to the drug, and in phase III studies currently underway patients are instructed to take the drug with food."

78. It is common ground that the data on file referred to in this passage formed the basis for the corresponding passage in the SmPC. The Defendants rely upon the last sentence of this passage. On its face, it is clearly a speculative statement, but it does suggest that there may be a link between administering rivastigmine after food, reducing C_max and increasing tolerability of the drug.
79. The second paper is R.J. Polinsky, "Clinical pharmacology of rivastigmine: a new-generation acetylcholinesterase inhibitor for the treatment of Alzheimer's disease", Clin. Therapeutics, 30, 634-647 (1998), another publication by an author from Novartis. Under the heading "Pharmacokinetics in humans" and the sub-heading "Absorption", this states (at 640):

"Coadminstration of rivastigmine with food slows absorption (t_max increased by 1.4 to 1.6 hours, maximum concentration [C_max] reduced) but increases bioavailability by 30% compared with the fasting state. Although this is only a moderate food effect, it is recommended that rivastigmine be administered in conjunction with food, since the propensity for gastrointestinal side effects may be related to high peak plasma concentrations.¹³"

Reference 13 is "Data on file. Novartis … 1996". Again, on its face, this is a speculative statement, but it does suggest that there may be a link between administering rivastigmine after food, increasing t_max and reducing C_max and increasing tolerability of the drug.

80. The third paper is C.M. Spencer and S. Noble, "Rivastigmine: A review of its use in Alzheimer's disease", Drugs & Aging, 13(5), 391-411 (1998). This states (at 400 and 405-406 respectively):

"Concomitant food slows the absorption of rivastigmine and results in a decrease in C_max values of approximately 30% and an increase in AUC values of about 30%.²³ Thus rivastigmine should be administered with food, which improves tolerability (Sections 5 and 6)."

"The tolerability of higher doses of rivastigmine was improved by administration of the drug after food and addition of antiemetics to the treatment regimen³² (administration with food decreases the C_max and increases the AUC of rivastigmine (Section 3))."

Reference 23 is "Anand R. Clinical expert report … 1997 (Data on file)". Reference 32 is Anand et al. This time the link is firmly stated, but pursuing the reference would reveal that it was speculative.

81. The fourth paper is N.R. Cutler and J.J. Sramek, "Tolerability profiles of AChEIs: a critical component of care for Alzheimer's disease patients", Int. J. Geriatic Pyschopharm., 1 (Supp), S20-S25 (1998). This states (at S23):

"Ingestion of food with medication can ease the gastrointestinal distress associated with the AChEIs. Administration with meals is highly recommended to improve tolerability of high doses of donepezil¹⁹ or rivastigmine.¹⁰ … "

Reference 10 is the paper by Sramek et al discussed in Anand et al. Although this links improved tolerability with administration with food, it does not refer to t_max or C_max.

82. The fifth paper is M. Weinstock, "Selectivity of cholinesterase inhibition: clinical implications for the treatment of Alzheimer's disease", CNS Drugs, 12(4), 307-323 (1999). This states (at 317):

"The incidence and intensity of the cholinergic adverse effects was greater with drugs such as physostigmine, tacrine and rivastigmine that reach the CNS relatively quickly than metrifonate, eptastigmine and donepezil and could be significantly reduced by a slow dose titration⁹⁹ or sustained release preparations.¹⁰"

Reference 99 is Cutler and Sramek. Again, on its face, this is a speculative statement, but it does suggest that there may be a link between smoothing out the peaks in the curve and increasing tolerability of the drug.

83. The sixth paper is B.P. Imbimbo, "Pharmacodynamic-tolerability relationships of cholinesterase inhibitors for Alzheimer's disease", CNS Drugs, 15(5), 375-390 (2001). This states (at 382-383):

"The rate of absorption determines the onset of the pharmacodynamic activity and may influence the occurrence of cholinergic-mediated adverse effects. Drugs that are rapidly absorbed may cause abrupt stimulation of the cholinergic system, with the consequent onset of cholinergic adverse effects. Conversely, cholinesterase inhibitors that are slowly absorbed may allow the organism to adapt to the cholinergic stimulation, with consequent reduced risk of cholinergic toxicity.

… The slow rate of absorption or rate of onset of pharmacodynamic activity of donepezil and eptastigmine agree with their good cholinergic tolerability profile. Other compounds, such as physostigmine and rivastigmine, have a short t_max CHE (0.5 to 1.5 hours) and increased tendency to generate cholinergic adverse effects."

This makes a firm link between short t_max and increased side effects, but no data or reference is provided to substantiate the statement.

84. The seventh paper is M.W. Jann, K.L. Shirley and G.W. Small, "Clinical pharmacokinetics and pharmacodynamics of cholinesterase inhibitors", Clin. Pharmacokinetics, 41(10), 719-739 (2002). This states (at 736):

"Another factor is the drug absorption rate, since cholinesterase inhibitors that are rapidly absorbed may cause a rapid stimulation of the cholinergic system with adverse effects closely following.¹⁰ Administering the drug with food can lower the amount of gastrointestinal effects by delaying the t_max."

Reference 10 is Imbimbo. Again, on its face, this is a speculative statement, but it does suggest that there may be a link between food, increasing t_max and increasing tolerability of the drug.

85. The eighth paper is D.B. Hogan and C. Patterson, "Treatment of Alzheimer's disease and other dementias – review and comparison of cholinesterase inhibitors", Can. J. Neuro. Sci., 29(4), 306-314 (2002). This states (at 309):

"Slower titration and administering rivastigmine with food appears to decrease the risk of GI side effects. Galantamine is also best taken with meals. If nausea and vomiting occur with rivastigmine or galantamine, first ensure that they are being taken with food, as this will decrease the C_max and delay the T_max of both agents.^{48, 49}"

"Rivastigmine reaches the CNS more rapidly and has a quicker onset of pharmacological effect. This may explain the seemingly greater frequency of GI AEs seen with it.^{4, 68}"

Reference 48 is to what I understand to be the Canadian equivalent to the SmPC. Reference 4 is Weinstock. Reference 68 is Imbimbo. Again, these statements suggest that there may be a link between food, increasing t_max and reducing C_max and increasing tolerability of the drug.

86. The ninth paper is F. Inglis, "The tolerability and safety of cholinesterase inhibitors in the treatment of dementia", I.J.C.P. Supp., 127, 45-63 (2002). This states (at 50 and 56 respectively):

"Rivastigmine is taken twice-daily with full meals (to reduce cholinergic side effects caused by rapid absorption), starting treatment with a low dose that is escalated slowly."

"Administration with meals is strongly recommended for those ChEIs whose absorption rate is affected by food. This can delay the absorption and lower the peak plasma concentration of ChEIs, which may reduce the likelihood of acute cholinergic events."

Again, these statements suggest that there may be a link between food, increasing t_max and reducing C_max and increasing tolerability of the drug.

87. The tenth paper is G.T. Grossberg, "Cholinesterase inhibitors for the treatment of Alzheimer's disease: getting on and staying on", Curr. Therapeutic Res., 64(4), 216-235 (2003). This states (at 220 and 220-221 respectively):

"It is thought that the occurrence of these AEs [adverse events] reflects the rapid increase in central ACh levels after oral intake.³⁷"

"In addition to slow dose escalation, administration of some ChE inhibitors with a full meal also may reduce the incidence of centrally mediated cholinergic AEs.^12,41  Administration of these agents with a meal delays drug absorption and lowers the peak plasma and brain concentrations, which reduces the likelihood that the patient will experience acute AEs.  For rivastigmine and galantamine, it is recommended that the drug be taken with morning and evening meals.¹²"

Reference 12 is the Physicians Desk Reference. Again, on their face, these are speculative statement, but they do suggest that there may be a link between food, increasing t_max and reducing C_max and increasing tolerability of the drug.

88. The last paper is D.G. Wilkinson et al, "Cholinesterase inhibitors used in the treatment of Alzheimer's disease", Drugs & Aging, 21(7), 453-478 (2004). This paper was co-authored by Prof Francis. It states (at 464-465 and 471 respectively):

"The shorter t_max (time to reach maximum concentration) and half-life of rivastigmine (necessitating more frequent administration) may lead to rapid increases in acetylcholine levels, contributing to the incidence of centrally mediated gastrointestinal adverse events ...¹¹⁸"

"Many of the adverse effects (e.g. nausea, diarrhoea and bradycardia) caused by the use of cholinesterase inhibitors are related to the stimulation of cholingeric neurons in the periphery. … Large fluctuations in cholinesterase inhibitor plasma concentrations and the attendant proportional AChE inhibition may contribute to an increase in the incidence of adverse effects.  Cholinesterase inhibitors with shorter half-life require more frequent administration and are, thus, more likely to have large peak trough differences in their plasma concentrations, in particular, before steady state is achieved . … Furthermore, cholinesterase inhibitors that are rapidly absorbed may abruptly stimulate the cholinergic system, thereby causing cholinergic adverse effects.¹⁰⁴"

Reference 118 is Weinstock. Reference 104 is Imbimbo. Again, on their face, these are speculative statements, but they do suggest that there may be a link between food, increasing t_max and reducing C_max and increasing tolerability of the drug.

89. Prof Francis' evidence in his first and third reports was that the skilled person would have understood that the side effects of rivastigmine were caused by sharp peaks in drug levels, indicated by short t_max and high C_max, and that the recommendation to administer with food was given in order to improve tolerability. He cited eight of the papers considered above in support of this opinion. Although he was cross-examined at some length on these matters, I did not understand him to resile from that opinion. He did, however, qualify it in two respects. First, he accepted that there was no firm evidence for these propositions, because no comparative trial had been carried out which showed lower side effects when rivastigmine was administered with food than without food, and that the statements in the papers involved some speculation. Secondly, he accepted that the skilled person did not have any expectation, as a result of the food effect, that the lowest titration dose could be omitted.
90. Prof Ballard took issue with Prof Francis' opinion in his second report. Again, he was cross-examined at some length on these matters. Counsel for Novartis made certain criticisms of the cross-examination, in particular that Prof Ballard was not taken to Anand et al. I am unimpressed by this point, since Prof Ballard was well aware of Anand et al and referred to it himself. The upshot of the cross-examination was as follows:

"2 Q. I know you have expressed your reservations about the Anand

3 paper, but it does rather look as though the proposition that

4 Cmax and tmax are implicated in side-effects for rivastigmine

5 was common wisdom at the time. You may say it is misguided,

6 but you would accept that it was the common wisdom?

7 A. No, I do not think I would accept that. I think I would

8 accept that there are a number of review articles that have

9 made statements speculating that food may reduce Cmax and that

10 that may be related to tolerability. So I think based on the

11 available evidence, what reviewers are citing generally is

12 that that is a reasonable hypothesis and in clinical practice

13 it might be sensible to give rivastigmine with food, based on

14 that hypothesis. I do not think they are saying anymore than

15 that.

16 Q. Okay. Perhaps we can put it like this. The common wisdom was

17 that it was a reasonable hypothesis that Cmax and tmax were

18 implicated in side-effects?

19 A. I would agree with that with one word different. Contributed

20 to I would agree with.

21 Q. Right. Of all the papers that we have been looking at, the

22 thing that is particularly singled out as of utility in

23 reducing the side-effects is things to address Cmax and tmax.

24 That is correct, is it not?

25 A. I think it is one specific thing that is mentioned in most of

2 the studies, which is administration with food, which is based

3 on good clinical practice and would be done for a wide variety

4 of agents."

91. Counsel for Novartis reminded me in his closing submissions that the fact that a statement is made in a widely read paper, or in a number of papers, does not demonstrate that it was common general knowledge: the statement must be both generally known and generally accepted as a good basis for further action. On the other hand, the greater the number of papers a statement is made in, particularly where many of them are review articles, the more likely it is to be common general knowledge. Furthermore, what counts as a good basis for further action depends on the context and the further action in question.
92. Taking all of the evidence into account, my conclusions are follows:

i) It was generally accepted that rivastigmine should be administered with food.

ii) As Prof Ballard pointed out, and Prof Francis accepted, this is common practice for many drugs, and there are a number of different potential reasons for doing it.

iii) In the case of rivastigmine, the skilled person would be aware that it was a reasonable hypothesis that administration with food increased the tolerability of rivastigmine and that this was because it increased t_max and reduced C_max which contributed to cholinergic side effects. The skilled person would also be aware, however, that there was no firm evidence to support this hypothesis.

93. I would add that, even if point (iii) was not common general knowledge, I consider that it would have been an obvious step for the skilled team, at the outset of a project to develop a new formulation of rivastigmine one of whose objects was to improve its tolerability, to undertake a short and focussed literature search into factors affecting the tolerability of rivastigmine. This would have thrown up some, if not all, of the papers considered above, from which the skilled clinician would draw the same conclusions.

Construction

94. The general principles applicable to the construction of patent claims were summarised by Jacob LJ in Virgin Atlantic Airways Ltd v Premium Aircraft Interiors UK Ltd [2009] EWCA Civ 1062, [2010] RPC 8 at [5].
95. Novartis' construction of the claim is that it has three components. First, it is a claim to rivastigmine for use in treating dementia or AD. Secondly, the rivastigmine is administered via a TTS. Thirdly, the "starting dose" of rivastigmine administered by the TTS is the dose released by a reference TTS which is specified in integers [3], [4] and [5] of the claim. Thus the TTS may have any structure or composition providing it can be used to deliver the same starting dose.
96. The Defendants' construction of the claim is that it is restricted to the administration of rivastigmine via a TTS having the structure and composition specified in integers [3], [4] and [5] of the claim, that is to say, a 5 cm² patch having the structure and composition of TTS #2.
97. In my judgment Novartis' construction is the correct one. Although counsel for the Defendants advanced a number of arguments in support of the Defendants' construction, none of them really engaged with the language of the claim, and in particular the words "the starting dose is that of" in integer [3]. The natural meaning of those words is that the method of administration involves a starting dose which is the same as that of a TTS having the specified characteristics. If the patentee had intended to claim administration of rivastigmine via a TTS having the structure of TTS #2, then those words would be redundant. In those circumstances, I do not propose to consider the Defendants' arguments at any great length. The two main ones were as follows.
98. First, counsel for the Defendants argued that the message conveyed by the specification was that the benefits of the invention were obtained by use of the three-layer structure of the TTS, and in particular the structure of TTS #2. I accept that the skilled team would understand from the specification that the three-layer structure of TTS #2 would enable them to achieve a starting dose which was claimed to be beneficial. I do not accept that the skilled team would understand that the claim was limited to the use of a TTS having that structure. On the contrary, it seems to me that the specification is replete with passages which indicate the opposite.
99. Secondly, counsel for the Defendants argued that construing the claim as contended for by Novartis would mean that it was invalid on the grounds of added matter and/or insufficiency. As he accepted, however, this is not a case where the patentee is relying upon equivalents to support a construction of the claim that extends beyond the primary, acontextual or literal meaning of the relevant expression, where that may be a relevant consideration. If the consequence of Novartis' construction is that the patent is invalid, then that is the consequence.
100. As for the meaning of the expression "starting dose", Novartis contend that the starting dose means the dose of rivastigmine with which treatment of a patient is started. The dose may subsequently be increased, for example, by applying a TTS of twice the area or two TTSs at once, or in any other way. I did not understand the Defendants to dispute this, but in any event I accept it. As counsel for Novartis submitted, it is necessary to keep in mind that the skilled team come to the Patent knowing that oral rivastigmine therapy comprises a process of titration from a starting dose, which is used for a period of 2-4 weeks, to a target maintenance dose in a number of steps. It follows the skilled person will have no difficulty in appreciating that the "starting dose" is the released dose administered during the first period of treatment. This construction is also supported by [0057], which makes reference to a "higher starting dose and reduced number of titration steps".
101. It is plain that the starting dose cannot be equated with the loaded dose in the TTS, since integer [3] of the claim distinguishes between the starting dose and the loaded dose. Novartis contend that the starting dose is to be quantified in terms of the mass of the drug released over the period of application of the TTS, typically 24 hours. The Defendants contend that the starting dose means, or at least could mean, the dose which actually finds its way into the blood plasma as measured by the AUC_24h. Again, I have no hesitation in accepting Novartis' construction. It is normal for transdermal patches to be described by reference to the amount of drug released over a period of time (thus both Novartis' Exelon patches and the Defendants' patches are described as "4.6 mg/24 h" patches). Neither Prof Ballard nor Prof Brown had ever come across patches being described in terms of the dose measured by reference to AUC_24h. Furthermore, AUC is not a measure of dose, but of bioavailability. There is nothing in the specification to suggest that the starting dose is to be measured by AUC_24h rather than in the conventional way.
102. Before leaving the question of construction, it is convenient to note that it is common ground that the claim is not limited to methods of treatment having any particular number of titration steps or any particular level of efficacy or any particular level of side effects.

Added matter

103. The law with regard to added matter was explained by Jacob LJ in Vector Corp v Glatt Air Techniques Ltd [2007] EWCA Civ 805, [2008] RPC 10 at [4]-[9]. As he held in Napp Pharmaceutical Holdings Ltd v Ratiopharm GmbH [2009] RPC 18 at [98]-[99], a claim does not disclose subject matter merely because it is wide enough to cover that subject matter. Further explanation of the law, particularly with regard to claim broadening and intermediation generalisation, was provided by Kitchin LJ in Nokia Corp v IPCom GmbH & Co KG (No 3) [2012] EWCA Civ 567, [2013] RPC 6 at [52]-[60]. The key question is whether the Patent presents the skilled team with information about the invention which is not directly and unambiguously derivable from the Application.
104. In the present case, as counsel for Novartis stressed, I am concerned with a Patent which was granted as a result of a divisional application. As counsel for Novartis submitted, it follows that it is neither surprising, nor objectionable, that the Patent has been confined to part of the subject matter contained in the Application. The Defendants contend, however, that, at least if the claim is construed as Novartis contend and I have accepted, the amendments made to the Application during prosecution leading to the grant of the Patent do not simply confine the Patent to part of the subject matter contained in the Application, but amount to the disclosure of a new invention altogether.
105. I have set out the disclosure of the Application in some detail above. In summary, it discloses an invention which has two main aspects. The first aspect concerns a three-layer TTS. The second aspect concerns a TTS providing C_max and AUC_24h values of rivastigmine within the broad ranges disclosed and claimed in the Application. I have also set out the disclosure of the Patent above. In summary, it discloses an invention in which rivastigmine is administered via a TTS with a starting dose which is the same as that of a reference patch, namely a 5 cm² TTS #2 patch.
106. The Defendants' added matter attack is a wide-ranging one, but in essence they contend that the Patent presents the skilled team with information that is not directly and unambiguously derivable from the Application in three main ways. First, the skilled team is informed for the first time that the invention lies in the selection of a particular starting dose for rivastigmine administered via a TTS for the treatment of AD. Secondly, the skilled team is informed for the first time that the dose delivered by the 5 cm² TTS #2 should be used as the starting dose. Thirdly, the skilled team is informed for the first time that this starting dose may be obtained using a TTS which does not have the structural and compositional features disclosed in the Application.
107. It is convenient to consider Novartis' answers to these contentions in reverse order. So far as the third point is concerned, counsel for Novartis treated this as an allegation of claim broadening. He pointed out that claim broadening during prosecution was permissible provided that it did not add subject matter. He submitted that, although the claim covered a starting dose obtained by TTSs having different structural and compositional features, it did not disclose such TTSs. It is true that neither the claim, nor any other passage in the Patent, discloses any particular different structure or composition of TTS other than TTS #2. But in my judgment that does not meet the Defendants' point. What the Patent tells the skilled team for the first time is that it is the starting dose delivered by the TTS that matters, not the structure or composition of the TTS, whereas previously the structure and composition of the TTS was presented as the core of the invention.
108. As to the second point, counsel for Novartis treated this as an allegation of intermediate generalisation, namely that the Application disclosed a particular starting dose only in the context of TTS #2 and that it was not legitimate to take that feature from the specific embodiment in the example without the other features of the embodiment. He submitted that it would be clear to the skilled team that any TTS that released the same amount of drug as a 5 cm² TTS #2 would have the same therapeutic effect irrespective of the structure and composition of the patch. In support of this submission, he relied upon the following passage in his cross-examination of Prof Williams:

" 12 Q. Now, if the skilled team looked at this document in 2005, they

13 would be aware that the patch being described, the TTS2 patch,

14 and indeed the TTS1 patch, delivered a certain dose to the

15 skin at a certain rate.

16 A. Yes.

17 Q. And in so far as the starting dose produced a certain effect

18 on the body; okay, which was deemed by the clinician to be

19 beneficial, it would be apparent that you could make other

20 patches also delivering the same dose at the same rate that

21 would do the same.

22 A. Yes.

23 Q. And reading this document the skilled person would have no

24 doubt that there would be other patches that could produce the

25 same effect with the same starting dose.

2 A. That is correct.

3 Q. And they would not think that the effect can only be achieved

4 with that particular design of patch, that particular TTS2

5 design of patch.

6 A. This particular composition, no."

109. In my judgment, this evidence does not assist Novartis for three reasons. First, the use of the word "apparent" in the second question fails to distinguish between what the document would disclose to the skilled formulator expressly or impliedly and what would be obvious to the skilled formulator.
110. Secondly, the second question required the witness to assume that the starting dose produced a certain effect on the body which was deemed by the clinician to be beneficial. As this demonstrates, however, the enquiry depends at least in part on how the clinician would interpret the disclosure of the Application. As counsel for Novartis noted in his closing submissions, Prof Francis did not feel able to express any opinion as to whether, "in so far as the patent … teaches there are benefits to administering a patch at a particular starting dose, in that it delivers a particular starting dose in a certain way", "these benefits are peculiar to … the particular patch described in the patent". This is the wrong question, however. The issue concerns the disclosure of the Application, not that of the Patent. Furthermore, the real question is what the Application would teach the skilled clinician about the starting dose and its benefits.
111. All that the skilled reader of the Application is explicitly told about the starting dose is that the TTS of the invention may allow a higher starting dose, and hence a reduced number of titration steps. As the skilled clinician would appreciate, whether this can be achieved will depend on whether the higher starting dose is tolerated by patients. Looking at the data in the example, specifically Tables 1 and 2 and Figure 4, the skilled clinician will see that the 5 cm² TTS #2 delivers an AUC_24h which is approximately equivalent to that of the 3 mg capsule bid. The skilled clinician will also see that the patients in the oral arm of the study were started on 1.5 mg bid whereas the patients in the TTS arm were started on the 5 cm² TTS #2, and in that sense the latter received a higher starting dose than the former. But the skilled clinician would note that, although the Application asserts that TTSs of the invention have better tolerability relative to equivalent levels of AUC_24h of Exelon capsule, there is no data to show that the 5 cm² TTS #2 was better tolerated as a starting dose than the 3 mg capsule bid would have been. The nearest one gets to this is the information about the numbers of patients who completed each of the four periods; but there is no information about the level of side effects experienced by those patients.
112. The third reason follows from the second reason. As the discussion in the preceding paragraph illustrates, the claim in the Patent is an intermediate generalisation because it takes the feature of the starting dose delivered by a 5 cm² TTS #2 stripped of its context in the example when it would not be clear to the skilled team that that feature was generally applicable or that the other features of the example were inessential to the invention.
113. Turning to the Defendants' first point, counsel for Novartis really had no answer to this. In my judgment it encapsulates the fundamental objection to the Patent when compared with the Application.
114. Accordingly, I conclude that the Patent is invalid on the ground of added matter.

Obviousness

115. The structured approach to the assessment of allegations of obviousness first articulated by the Court of Appeal in Windsurfing International Inc v Tabur Marine (Great Britain) Ltd [1985] RPC 59 was re-stated by Jacob LJ in Pozzoli v BDMO SA [2007] EWCA Civ 588, [2007] FSR 37 at [23]. The correct approach to the fourth step in a case such as the present was recently summarised by Kitchin LJ, with whom Lewison and Moore-Bick LJJ agreed, in MedImmune Ltd v Novartis Pharmaceuticals Ltd [2012] EWCA Civ 1234 at [90]-[93].

US031

116. At trial the Defendants relied on a single item of prior art, namely United States Patent No. 6,335,031 entitled "TTS containing an antioxidant" published on 1 January 2002 ("US031").
117. US031 discloses the administration of (S)-N-ethyl-3[-1-dimethylamino-ethyl]-N-methyl-phenyl-carbamate (the chemical name for rivastigmine), referred to as "compound A", by a transdermal patch. Having pointed out that compound A is useful in inhibiting acetylcholinesterase in the central nervous system, e.g. for the treatment of AD, US031 states that it has been found after exhaustive testing that compound A is susceptible to degradation, particularly in the presence of oxygen (at column 1 lines 14-28). Accordingly, US031 proposes the use of a pharmaceutical composition comprising compound A and an anti-oxidant (at column 1 lines 29-39). The pharmaceutical composition is preferably used for transdermal application (column 4 lines 31-32). A transdermal device for administering compound A may comprise a pharmaceutical composition containing compound A, a backing layer, an adhesive and a release-liner (column 4 lines 33-39). The transdermal device may be conveniently formed in continuous sheets and cut into patches of any desirable size or configuration before use (column 5 lines 60-63).
118. In terms of dosing, US031 teaches as follows (at column 6 line 46 – column 7 line 14):

"The transdermal devices of the invention in general have, for example, an effective contact area of pharmaceutical composition on the skin of from 1 to about 80 square centimeters, preferably about 10 square centimetres, and are intended to be applied at intervals of about once every 1 to 7 days, preferably 1-3 days. Compound A is well tolerated at a dose of 36 mg in free base form in up to 80 cm² of patches according to the invention containing 36 mg compound A from which 12 mg was absorbed. Compound A may, for example be administered at a dose of 8 mg in a patch of ca. 10 cm², once every day. …

… The exact amounts of compound A to be administered may depend on a number of factors, e.g. the drug release characteristics of the compositions, the drug penetration rate observed in vitro and in vivo tests, the duration of action required, the form of compound A, and for transdermal compositions the size of the skin contact area, and the part of the body to which the unit is fixed. The amount of and, e.g. area of the composition etc. may be determined by routine bioavailability tests comparing the blood levels of active agents after administration of compound A in a composition according to the invention to intact skin and blood levels of Compound A observed after oral administration of a therapeutically effective dose of the compound.

Orally, the Compound A is well tolerated at an initial dose of 1.5 mg twice a day orally and the dose may be stepped up to 3 mg twice a day in week 2. Higher doses are possible, for example 4.5 mg twice daily and even 6 mg twice daily. Tolerability is seen to be even better for the transdermal device, wherein 24 mg were absorbed in 24 hours."

119. US031 does not disclose any data to support the statement in the last sentence of the passage quoted. Furthermore, it is convenient to note at this point that Prof Ballard's opinion was that the skilled clinician would not regard it as plausible. He did not make the same suggestion about the figures of 8 and 12 mg mentioned earlier, however.
120. Example 4 is a dual layer patch. The first layer has the following composition per unit (10 cm²):

Compound A	18 mg	30%
Polymer	29.94 mg	49.85%
Methacrylate	12 mg	20%
(-tocopherol	0.06	0.1%
Total	70 mg	100%

121. The second layer has the following composition per unit (10 cm²):

Bio-PSA Q7-4302	29.67 mg	98.9%
Silicone oil	0.3 mg	1.0%
(-tocopherol	0.03 mg	0.1%
Total	30 mg	100%

122. There is no dispute that the structure and composition of the TTS disclosed in Example 4 is the same as that of TTS #2 in the Patent. Nor is there any dispute that it would be a very straightforward matter for the skilled formulator to make various patches of differing size having this structure and composition, including a 5 cm² patch with a loaded dose of 9 mg rivastigmine (half the unit size in US031, which has a loaded dose of 18 mg rivastigmine per 10 cm²).

The difference

123. On the construction of the claim which I have adopted, the essential difference between the disclosure of US031 and the claim is that US031 does not disclose the use of a starting dose which is that released by the 5 cm² TTS #2.

Was it obvious?

124. Although the evidence ranged more widely, the dispute on obviousness is quite a narrow one. The starting point is that there is no real dispute as to the following points which arise out of the skilled team's common general knowledge:

i) The skilled team would have been motivated to develop a formulation of rivastigmine which addressed the disadvantages of rivastigmine compared to donepezil identified in paragraph 18 above. In particular, the skilled team would have been motivated to develop a formulation which enabled once daily administration.

ii) The skilled team would have known that a transdermal patch would be likely to enable once daily administration to be achieved.

iii) The skilled team would have ascertained, if necessary by routine testing, that the properties of rivastigmine made it suitable for administration by a transdermal patch.

iv) In developing a transdermal patch for rivastigmine, the skilled team's starting point would have been to seek to develop a patch which delivered an AUC_24h which matched that of an existing oral formulation, namely Exelon capsules.

125. The issue is as to what the skilled team would do, against that background, when presented with US031. There is no dispute that it would be obvious to make a series of patches having the structure and composition of Example 4 which delivered AUC_24h values matching those of the existing oral formulations. The dispute is as to what dose the skilled team would select as the starting dose when administering rivastigmine by means of such patches.
126. Novartis contend that the skilled team would match the AUC_24h values delivered by each of the daily doses in the oral regimen, i.e. 3 mg (1.5 mg bid), 6 mg (3 mg bid), 9 mg (4.5 mg bid) and 12 mg (6 mg bid), and would then administer a starting dose using the patch which matched that of the starting dose in that regimen. There is no dispute that, if the skilled team followed that approach, they would administer a starting dose which is less than half of that released by the 5 cm² TTS #2. Furthermore, they would follow the same number of titration steps as with the oral regimen.
127. The Defendants contend that it would be obvious to try the dose released by the 5 cm² TTS #2 as the starting dose in a small scale clinical trial (i.e. of the kind reported in the Patent). The Defendants put their case in two ways. The first is that US031 expressly teaches the skilled reader that the size of the patch "may be determined by routine bioavailability tests comparing the blood levels of active agents after administration of compound A in a composition according to the invention to intact skin and blood levels of Compound A observed after oral administration of a therapeutically effective dose of the compound [emphasis added]." The skilled team would know that the lowest therapeutic oral dose of rivastigmine was 6 mg daily, and so they would take that as the lowest dose to match. The routine bioavailability tests would produce data reflecting that shown in Tables 1 and 2 and Figure 4 of the Patent. That would lead the skilled team to choose the dose released by the 5 cm² TTS #2 patch as their starting dose, since it delivers an AUC_24h approximately equal to that of the 6 mg daily oral dose. They would then find that (as is now known to be the case) it was well tolerated.
128. The second way is that, if the skilled team were concerned about tolerability, they would appreciate that the patch would deliver rivastigmine with a smoothed out plasma profile, and hence a longer t_max and lower C_max than the oral formulation. Indeed, the data in the Patent show that the C_max of a patch matched to the AUC_24h of a 3 mg bid dose would be lower than the C_max of the 1.5 mg bid dose. Based on the skilled clinician's common general knowledge about the food effect, they would think it reasonably likely that the dose released by the 5 cm² TTS #2 would be sufficiently well tolerated to be administered as the starting dose. Certainly, they would think that this was sufficiently likely to warrant a small scale clinical trial given that it would enable a therapeutic dose to be administered immediately and would eliminate a titration step.
129. Novartis' response to the Defendants' first way of putting their case is that the skilled team would read US031 knowing that oral administration of rivastigmine involved titration from a therapeutically ineffective initial dose up to a therapeutically effective maintenance dose and would read the statement quoted in that light. Thus they would not interpret it as suggesting that the sub-therapeutic dose should be omitted.
130. Novartis' response to the Defendants' second way of putting their case is that the skilled team would have no expectation that increasing t_max and reducing C_max would improve the tolerability of rivastigmine to the extent that it would remove the need for the sub-therapeutic initial dose. Accordingly, the skilled team would have no expectation that the dose released by the 5 cm² TTS #2 would be sufficiently well tolerated to be administered as the starting dose. Furthermore, given that rivastigmine is generally administered to elderly and vulnerable patients, the skilled team would proceed with caution. Therefore they would replicate the titration regime of the oral formulation. This is supported by what in fact happened when an extended release oral formulation of galantamine was developed: the sub-therapeutic starting dose was retained.
131. More generally, Novartis rely on the fact that, prior to the Patent, no one in the art had suggested that administration of the sub-therapeutic dose of rivastigmine could be avoided by the use of an extended release formulation. On the other hand, there is no evidence that US031 had been widely read before the priority date.
132. In assessing these contentions, it is important to note that, as Prof Ballard accepted, it was routine at the priority date when developing a new formulation to perform dose titration studies to determine the appropriate dose for initiating therapy, the maintenance dose and the maximum dose. Furthermore, the skilled clinician would be aware that there was unlikely to be a unique dose which was suitable for each of these purposes: rather, the appropriate dose was likely to be a matter of judgement, depending on the balance between efficacy and side effects.
133. It is also important to remember that, while the side effects of rivastigmine could be unpleasant for patients, they were generally not severe. The inventors of the Patent were not put off trying the dose released by the 5 cm² TTS #2 as the starting dose in their study by the potential side effects, and there is nothing to suggest that they were taking a risk that the skilled team would not have been prepared to countenance. On the contrary, a paper by G. Lefevre et al, "Pharmacokinetics and pharmacodynamics of the novel daily rivastigmine transdermal patch compared with twice-daily capsules in Alzheimer's disease patients", Nature, 83, 106-114 (2008), published after the priority date, which describes more fully the trial reported in the Patent, indicates that the inventors followed exactly the reasoning advanced by the Defendants (at 106):

"The incidence of centrally induced cholinergic gastro intestinal side effects with rivastigmine has been associated with the high maximum plasma concentrations (C_max) and short times to C_max, (t_max) provided by oral administration. Measures that prolong t_max and reduce C_max such as the administration of rivastigmine capsules with food, may improve tolerability of cholinesterase inhibitors.^8,9 For a given level of exposure, the transdermal administration of rivastigmine, by providing continuous delivery of drug with reduced fluctuations in plasma levels (i.e., lessening the rapid rise and fall of drug concentration), prolonging t_max and achieving a lower C_max is expected to reduce side effects. This may also offer additional therapeutic advantages over oral administration, such as access to higher doses, with the potential to improve compliance and treatment effects."

(Reference 8 is Jann, Shirley and Small. Reference 9 is a post-priority date paper.)

134. In those circumstances, I conclude that it would have been obvious to try the dose released by the 5 cm² TTS #2 as the starting dose in a small scale clinical trial for both the reasons advanced by the Defendants. So far as the first reason is concerned, while it is true that US301 does not in terms instruct the skilled team to omit the sub-therapeutic dose, it cannot be inventive to do exactly what it does say. So far as the second reason is concerned, I consider that, having regard to the skilled team's motivation and the relative ease with which a small study could be carried out, the skilled team would have had a sufficient expectation of success to warrant trial. Accordingly, I conclude that the Patent is invalid for lack of an inventive step.

Insufficiency

135. A patent will be insufficient if the skilled person is unable to carry out the claimed invention given the description of it in the specification and common general knowledge (sometimes called "classical insufficiency"). A patent may also be invalid on the ground of insufficiency as a result of either ambiguity or excessive breadth of the claims. I reviewed the law with regard to classical insufficiency and excessive claim breadth most recently in Idenix Pharmaceuticals Inc v Gilead Sciences Inc [2014] EWHC 3916 (Pat) at [465]-[468]. I reviewed the law with regard to ambiguity in Sandvik Intellectual Property AB v Kennametal UK Ltd [2011] EWHC 3311 (Pat), [2012] RPC 23 at [119]-[120].
136. The Defendants contend that the Patent is insufficient in all three ways. I can deal with the first way very briefly, since this is only advanced as a squeeze on obviousness. The Defendants' primary case is that it was plausible in the light of the skilled team's common general knowledge that the dose released by the 5 cm² TTS #2 would be well tolerated as a starting dose. But the Defendants say that, if that was not the case, then the Patent contains no data which makes this any more plausible, and therefore it is insufficient. Since I have accepted the Defendants' primary case, this objection does not arise.
137. Turning to excessive breadth of claim, the Defendants contend that, if the claim extends to any TTS that delivers a starting dose which is that of the 5 cm² TTS #2, which is the effect of Novartis' construction, then the claimed benefit of tolerability is not plausibly obtained across the breadth of the claim. In particular, the Defendants contend that the claim would cover TTS formulations with a "burst release" profile, that is to say, a rapid initial release of drug which then tails off; but the Patent does not make it plausible that such formulations would have the tolerability of the 5 cm² TTS #2.
138. Although counsel for Novartis advanced a number of answers to this contention, it is sufficient to focus on one. This is that, as Prof Williams accepted, someone seeking to put the invention of the Patent into effect by making a transdermal patch which delivered the same starting dose as the reference patch would have no reason to choose a burst release formulation. As counsel for Novartis submitted, a patent is not insufficient merely because it is possible to imagine a way of implementing the claim badly if in practice that is not something that the skilled person would do.
139. As for ambiguity, the Defendants contend that, if the claim extends to any TTS that delivers a starting dose which is that of the 5 cm² TTS #2, which is the effect of Novartis' construction, then the Patent does not enable the skilled team to determine whether any other TTS delivers the same starting dose or not. Not only does the specification not disclose the starting dose delivered by the 5 cm² TTS #2 (as to which, see below), but more fundamentally the specification does not identify any test, and in particular any statistical test, for determining what counts as "the same" starting dose. Depending on the precision and repeatability of whatever experimental technique is used, if the 5 cm² TTS #2 delivers a dose of 4.6 mg/24 h, then a dose delivered by another TTS of (say) 5.2 mg/24 h may or may not be regarded as the same. (On the other hand, the accuracy and reproducibility of the experimental technique are not concerns, since the issue is one of comparative testing of two TTSs by the same laboratory.)
140. Counsel for Novartis submitted in his closing submissions that this point had been raised in a manner which was procedurally unfair to Novartis. The Defendants pleaded in their Grounds of Invalidity that the skilled person would not be able to establish without undue burden whether or not a TTS fell within the claim, since "neither the starting dosage … of a TTS with the attributes set out in claim 1, nor the starting dosage of a TTS which did not have those attributes, are ascertainable". Neither in their Grounds of Invalidity, nor in their experts' reports, nor in their skeleton argument, however, did the Defendants raise any point concerning the absence of any statistical test in the Patent. The point first emerged when counsel for the Defendants cross-examined Prof Brown. Counsel for Novartis raised no objection at that juncture. At the conclusion of the cross-examination I queried whether the point had been pleaded. Counsel for the Defendants submitted that it had been, and relied on the passage quoted above. Counsel for Novartis did not submit to the contrary, and proceeded to cross-examine Prof Williams on the point. In those circumstances, I do not consider that it is open to Novartis to object that the point is not pleaded. On the other hand, I do consider that it is legitimate for Novartis to rely upon the timing and manner in which the point was raised as excusing the absence of a more rigorous treatment of the point in the evidence. Furthermore, neither Prof Brown nor Prof Williams was an expert in statistics, and it was clear from their evidence that neither was very comfortable discussing the point.
141. Turning to the merits of the point, there is no dispute that the specification does not address this question: the Patent simply assumes that the skilled team can determine, using their common general knowledge, whether or not a TTS delivers the same dose as the 5 cm² TTS #2. It is plain that this would involve doing a number of repeat measurements of the doses delivered by the two TTSs using the same apparatus, and then performing a statistical analysis. Although Prof Brown accepted that, "if you have not defined the test that you are going to do to show that they are the same, then you cannot show that they are the same", he also gave evidence that "you just have to apply normal statistics to it". Prof Williams explained that the kind of statistical test that would be appropriate would depend on the nature of the data, and in particular whether it was Gaussianly or non-Gaussianly distributed. Having regard to the considerations outlined in the preceding paragraph, I find that the effect of their evidence was that the skilled team would be able to select an appropriate statistical test to determine whether or not the TTSs delivered the same dose. Accordingly, I am not satisfied that the Defendants have demonstrated that the claim is ambiguous.
142. Accordingly, I conclude that the Defendants have not established that the Patent is insufficient.

Infringement

143. It is admitted by each of the Defendants that they are offering to supply and supplying transdermal patches containing rivastigmine in the UK for the treatments of AD. In particular:

i) Focus and Actavis are supplying and offering to supply 5 cm² and 10 cm² patches branded "Voleze" which release 4.6 mg/24 h and 9.5 mg/24 h of rivastigmine respectively.

ii) Teva is supplying and offering to supply 5 cm² and 10 cm² patches branded "Rivatev" which release 4.6 mg/24 h and 9.5 mg/24 h of rivastigmine respectively.

144. The Defendants' rivastigmine patches are identical save for the branding, and are made by the same manufacturer, SK Chemicals. There is no dispute that the SK patches are generic versions of Novartis' Exelon patches (i.e. Novartis' Exelon patches were used as the reference product for SK's generic marketing application) or that they are bioequivalent to the Exelon patches.
145. Dr Reum's unchallenged evidence is that the 5 cm² Exelon patch has the structure and composition of the 5 cm² TTS #2 as required by the claim and releases a dose of 4.6 mg/24 h, which the SmPC for the Exelon patches indicates is the starting dose.
146. Similarly, the SmPCs for the Defendants' products each indicate that treatment is to be started with the same dose of 4.6 mg/24 h. For example, the Voleze SmPC states:

"Initial dose

Treatment is started with 4.6 mg/24 h.

After a minimum of four weeks of treatment and if well tolerated according to the treating physician, this dose should be increased to 9.5 mg/24 h, which is the recommended effective dose."

147. Novartis contends that infringement by sale of the 5 cm² Voleze patch arises because that patch is plainly for use for the claimed indication wherein rivastigmine is administered in a dosage regime with the claimed starting dose. Novartis contends that infringement by the sale of the 10cm² patch arises because that patch is plainly for use for the claimed indication wherein the rivastigmine is administered in a dosage regime with the claimed starting dose. In this case the 10 cm² patch is the maintenance dose rather than the starting dose, but it is submitted that it nevertheless falls within the scope of the claim.
148. As counsel for Novartis pointed out, none of the Defendants made any attempt to demonstrate that, notwithstanding the description of their products in their SmPCs, their 5 cm² patches in fact released a different dose to the Exelon 5 cm² patch. Furthermore, Prof Williams accepted that he had no reason to doubt that the SK products delivered the same starting dose as the Exelon patch. In any event, a SK regulatory document in evidence presents the results of a comparative analysis of SK and Exelon 9.5 mg/24 h patches using Franz diffusion cells with human skin and EVA membranes and concludes that the two patches exhibit "an equivalent permeation profile", although it is clear from the data that they were not precisely identical. Prof Williams did not know what statistical analysis had been performed to arrive at this conclusion, but I see no reason not to take the conclusion at face value.
149. Counsel for the Defendants submitted that Figure 3 of the Patent indicated that the dose released by a 5 cm² TTS #2 over 24 h would be 2.75 mg (5 times 550 (g). Prof Brown accepted that this was an accurate interpretation of the graph, but it does not necessarily follow that the SK 5 cm² patches do not release the same dose as the 5 cm² TTS #2 when measured using the same apparatus and applying an appropriate statistical analysis as discussed above. Accordingly, I do not consider that this evidence undermines the conclusion to be drawn from the evidence discussed in the preceding paragraphs. I therefore conclude that, if the Patent was valid, the Defendants would have infringed it.

Summary of conclusions

150. For the reasons given above, I conclude that:

i) the Patent is invalid on the ground of added matter;

ii) the Patent is invalid since the claimed invention lacks an inventive step over US031;

iii) the Defendants have not established that the Patent is insufficient;

iv) if the Patent was valid, the Defendants would have infringed it.