Novartis' patent on transdermal therapeutic system of Exelon Patch held invalid

21 May 2015

The Patents Court (Arnold J) has ruled that Novartis' patent to the transdermal therapeutic system of its Exelon Patch is invalid. The decision clears the way for the launch of 'generic' rivastigmine patches.

In his judgment in Novartis AG & Ors v Focus Pharmaceuticals Ltd & Ors [2015] EWHC 1068 (Pat) (27 April 2015), Arnold J found that Novartis' European Patent (UK) No. 2 292 219 entitled "Transdermal therapeutic system for the administration of rivastigmine" was infringed by the defendants' (Focus, Actavis and Teva) various rivastigmine patches, all of which were 'generic' versions of Novartis' Exelon patch.

However, the defendants challenged the validity of the patent, and Arnold J held it to be invalid for added matter and for lack of inventive step. Challenges of insufficiency did not succeed.

The claimed invention

Broken into integers, Novartis' claim 1 was to the following:

"[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 5cm2 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

Arnold J held that the patent was addressed to a skilled team which was interested in developing a new formulation for rivastigmine. The parties agreed that the team would contain a formulator skilled in the transdermal administration of drugs.

There was disagreement though as to whether the second member of the team would be a clinician working in the field of dementia, who would call upon a pharmacologist where necessary (Novartis' view), or a neuroscientist working in that field (the defendants' view). Arnold J's view was that nothing turned on it, so he went with Novartis' contention saying that clinical expertise was more directly relevant than pharmacological (or biochemical) expertise.

Common general knowledge

Arnold J referred to his review of the law on 'common general knowledge' (CGK) in KCI v Smith & Nephew [2010] EWHC 1487 (Pat) . He said that there was little, if any, dispute that everything described in the judgment as 'technical background' was CGK. This included the following:

Alzheimer's disease ("AD") is a progressive neurodegenerative disease that is the most common cause of dementia, with approximately 520,000 sufferers in the UK. At the priority date, four drugs ('active pharmaceutical ingredients' (APIs)) were being used to treat mild to moderate AD, three of which were AChE inhibitors. One of these was rivastigmine (brand name Exelon, developed by Sandoz/Novartis), and another was donepezil.

Rivastigmine was administered in capsules or oral solution at an initial twice-daily dosage of 1.5mg, rising to a maximum dosage of 6 mg. The minimum therapeutically effective dose was 3mg twice daily. Patients were 'titrated' up gradually from lower doses to enable them to acquire tolerance to side effects.

In contrast, donepezil only required a single daily dose.

All three AChE inhibitors were associated with mild to moderate cholinergic effects such as nausea, vomiting and diarrhoea. If there was a break in treatment of even a few days, tolerance to the drugs tended to reduce.

Key information that was known about rivastigmine included the following:

  • administration should be with morning and evening meals
  • administration with food
    • increased the area under the concentration-time curve (AUC), which reflected the actual body exposure to a drug after administration by approximately 30% - this was considered 'puzzling'.
    • lowered Cmax (a measure of the peak plasma concentration of the drug), and
    • delayed absorption (t) (the time at which Cmax was reached) by 90 minutes,
  • Administration by transdermal patch had a number of advantages over oral formulations, including: bypassing the gastrointestinal (GI) tract, thereby avoiding locally-mediated side-effects in the GI tract; enabling once-daily dosing when that may be impossible with an oral formulation, and providing a smoother delivery curve (see below). At the priority date there were 28 approved transdermal patches but none available for the treatment of AD

  • Administration by transdermal means also had some disadvantages, including: that only a limited number of APIs were suitable for such administration; that they were more expensive and time consuming to develop; that the onset of treatment was slower than with oral formulations; and that there was potential for local skin irritation.
  • A transdermal patch included an outer backing layer, typically impermeable to water, a preparation containing the API along with excipients, and a release liner which was removed before the patch was applied to the skin.
  • Once an API was identified as being suitable for patch delivery, a target dose would be given to the formulator, generally quantified in terms of a target AUC, typically the AUC in 24 hours (AUC24h). Usually, the target AUC24h would correspond to the AUC24h for the oral form. Prototypes would be developed and tested, in vitro and in vivo, and assessed upon the amount of API released from the patch.

There was a dispute, however, as to whether it was CGK that (i) the side effects of rivastigmine were caused by sharp peaks in drug levels, indicated by short tmax and high Cmax, and (ii) the recommendation to administer with food was given in order to improve tolerability.

The defendants contended that these points were CGK and cited 11 publications in support. Arnold J concluded as follows:

"[T]he 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 tmax and reduced Cmax which contributed to cholinergic side effects. The skilled person would also be aware, however, that there was no firm evidence to support this hypothesis."

He went further:

"I would add that, even if [this] point …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."


Arnold J referred to the principles applicable to the construction of patent claims as summarised by Jacob LJ in Virgin Atlantic Airways Ltd v Premium Aircraft Interiors UK Ltd [2009] EWCA Civ 1062, [2010] RPC 8.

The defendants argued that the claim should be construed as restricted to the administration of rivastigmine via a transdermal therapeutic system (TTS) having the structure and composition specified in integers [3], [4] and [5] of the claim.

Arnold J disagreed. In his view, the defendants' construction did not engage with the language of the claim, and in particular the language "the starting dose is that of" in [3]. Arnold J said that the natural meaning of those words was that the method of administration involved a starting dose which was the same as that of a TTS having the specified characteristics. If the patentee had intended to claim the administration of rivastigmine via a TTS having the structure of example #2, then those words would be redundant.

Instead, Arnold J accepted Novartis' construction, which was that the claim had three components:

  1. to rivastigmine for use in treating dementia or AD;
  2. the rivastigmine being administered via a TTS; and
  3. the 'starting dose' of rivastigmine administered by TTS being the dose released by a reference TTS which was specified in integers [3], [4] and [5] (i.e. a reference patch described in example #2).

Arnold J said that although the benefits of the invention were obtained by using the structure of TTS #2, the claim was not limited to use of a TTS having such a structure.

As regards the meaning of 'starting dose', Arnold J accepted Novartis' contention, as consistent with usual terminology, that this referred to the dose of rivastigmine with which treatment of a patient was started. This was quantified in terms of the mass of the drug released over the period of application of the TTS, typically 24 hours (e.g. 4.6mg/24 hrs). The defendants' contention that 'starting dose' meant AUC24h was rejected, AUC being a measure of bioavailability, not dose.

The parties agreed that the claim was not limited by any particular number of titration steps, any particular level of efficacy or any particular level of side effects.

Added matter

Arnold J referred to Jacob LJ's explanation of the law regarding added matter in Vector Corp v Glatt Air Techniques Ltd [2007] EWCA Civ 805, [2008] RPC 10 and Kitchin LJ's summary in Nokia v IPCom [2012] EWCA Civ 567, [2013] RPC 6. He noted also Jacob LJ's finding in Napp v Ratiopharm [2009] RPC 18 that a claim does not disclose subject matter merely because it is wide enough to cover that subject matter. He said that the key question was whether the patent presented the skilled team with information about the invention which was not directly and unambiguously derivable from the application.

Arnold J described the invention disclosed in the application for the patent as having two main aspects:

"The first concerns a three-layer TTS. The second aspect concerns a TTS providing Cmax and AUC24h values of rivastigmine within the broad ranges disclosed in the claimed application."

He described the description of the patent (a divisional) as broadly similar, but with a number of differences:

"…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]).

Secondly, the extensive definition of "active ingredient" contained in the Application has been deleted from the Patent.

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.

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 effect of this, Arnold J concluded, was that the patent informed the skilled reader for the first time that:

  1. the invention lay in the selection of a particular starting dose for rivastigmine administered via a TTS for the treatment of AD;
  2. the dose delivered by the 5cm2 TTS of example #2 should be used as the starting dose; and
  3. this starting dose may be obtained using a TTS which does not have the structural and compositional features disclosed in the application.

The judge noted that Novartis really had no answer to the first of these points.

Regarding the second point, all that the reader of the application was explicitly told about the starting dose was that the TTS of the invention may allow a higher starting dose. The skilled clinician would appreciate that whether this could be achieved would depend on whether the higher starting dose was tolerated by patients, but there was no information in the application about the level of side effects experienced by patients.

This meant that the claim in the patent was an intermediate generalisation, because it took the feature of the starting dose delivered by a 5cm2 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.

Regarding the third point, Arnold J noted that previously, the structure and composition of the TTS were presented as the core of the invention. It was irrelevant that although the claim covered a starting dose obtained by TTSs having different structural and compositional features, it did not disclose it as such. That was not the test for added matter.

Consequently the patent was invalid for added matter.


On the law, Arnold J referred to Windsurfing v Tabur Marine [1985] RPC 59, restated by Jacob LJ in Pozzoli v BDMO [2007] EWCA Civ 588, [2007] FSR 37, and noted Kitchin LJ's summary of the correct approach to the fourth question in MedImmune v Novartis [2012] EWCA Civ 1234.

At the trial, the defendants relied on a single piece of prior art, a US patent, which disclosed in example 4 a TTS having the same structure and composition as that of the TTS of example #2 in the patent in suit, but with double the loaded dose per cm2.

It was not disputed that it would be straightforward for the skilled formulator to make various patches of differing size having the disclosed structure and composition, including a 5cm2 patch with the loaded dose of the TTS #2. Nor was there much dispute about motivation i.e. to get once daily administration using a patch, or that the starting point for the skilled team would be to develop a patch which matched the AUC24h of the existing oral formulation of Exelon capsules.

Novartis contended that the skilled team would administer a starting dose using a patch which matched that of the starting dose in the oral regimen (i.e. less than half that released by the 5cm2 TTS #2), and then titrate it up in a similar way to the oral regimen.

The defendants contended that it would be obvious to try the dose released by the 5cm2 TTS #2 as the starting dose in a small scale clinical trial, because:

  1. The US patent taught that the size of the patch could be determined using routine bioavailability tests; these would reveal that the dose released by the 5cm2 TTS #2 patch delivered an AUC24h approximately equal to that of the 6 mg daily oral dose (i.e. the minimum therapeutically effective dose). They would then find that it was well tolerated.

On this approach, Arnold J noted that while the US patent did not in terms instruct the skilled team to omit the sub-therapeutic dose, "it cannot be inventive to do exactly what it does say".

  1. If the skilled team were concerned about tolerability, they would appreciate that delivery by a patch would smooth the plasma profile and hence provide a longer tmax and lower Cmax than the oral formulation. Based on CGK about the food effect, the skilled team would think it reasonably likely that the dose released by the 5cm2 TTS #2 would be sufficiently well tolerated to be administered as the starting dose.

On this approach, Arnold J commented that "[i]ndeed, the data in the Patent show that the Cmax of a patch matched to the AUC24h of a 3mg bid dose would be lower than the Cmax of the 1.5mg bid dose". He also noted that while the side effects of rivastigmine could be unpleasant, they were generally not severe and the inventors were not put off trying the claimed dose as the starting dose by the potential side effects. Nor was there anything to suggest that they were a risk which the skilled team would not be prepared to countenance.

Arnold J concluded that for both the reasons advanced by the defendants, it would have been obvious to try the dose released by the 5cm2 TTS #2 in a small scale clinical trial. The skilled team would have had a sufficient expectation of success to warrant doing so. Consequently the patent was invalid for lack of inventive step.


Arnold J referred to his review of the law with regard to classical insufficiency and excessive claim breadth in Idenix v Gilead [2014] EWHC 3916 (Pat) , and with regard to ambiguity in Sandvik v Kennametal [2011] EWHC 3311 (Pat).

The defendants ran an obviousness/insufficiency squeeze, saying that if it was not plausible that the claimed dose would be tolerated, the patent contained no additional data which made this more plausible. Having succeeded in showing the patent was obvious (i.e. that the claimed invention was plausible), this fell away.

The defendants also argued that if the claim was construed as Novartis contended, extending to any TTS that delivered a starting dose which was that of the 5cm2 TTS #2, then it was insufficient. This was because 1) it would cover 'burst release' formulations which would not have the required tolerability and 2) the patent did not enable the skilled team to determine whether any other TTS delivered the same starting dose or not.

These challenges failed too. On the first point, Arnold J said "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".

On the second point, Arnold J held that on the expert witnesses' evidence, the skilled team would be able to select an appropriate statistical test to determine whether the same TTS was delivered. This was not explored in much detail owing to the lack of pleading and scant preparation of the experts in relation to it.


Infringement was dealt with briefly, each of the alleged infringing patches being generic versions of Novartis' Exelon patch. Arnold J concluded that if the patent had been valid, the defendants would have infringed it.


This decision underlines the risks of 'adding matter' in the course of the conceptual development of an invention, after the filing of the application for the patent. The risk would tend naturally to be greater for divisional applications, such as the patent in this case, where there can be a relatively long interval in time between the filing of the application and the grant of the patent.

Interestingly, both Novartis and the defendants seem to have presented their cases on obviousness on the basis of the test being whether the skilled person 'would' have made the invention. The finding that there was no inventive step reflects in part the level of expertise of the pharmaceutical 'skilled team' in this case. Before the Exelon Patch, there had never been an AD therapy delivered by a TTS, but this did not make the development of the first patch inventive.

Pat has also written an article in Bio-Science Law Review titled " Novartis' patent on a transdermal therapeutic system for the administration of rivastigmine found invalid, clearing the way to 'generics' of the Exelon Patch". 

NOT LEGAL ADVICE. Information made available on this website in any form is for information purposes only. It is not, and should not be taken as, legal advice. You should not rely on, or take or fail to take any action based upon this information. Never disregard professional legal advice or delay in seeking legal advice because of something you have read on this website. Gowling WLG professionals will be pleased to discuss resolutions to specific legal concerns you may have.