I.
Overview
[1]
This action for impeachment of the 2,139,653
patent (the ‘653 patent) and counter-claim for infringement pertains to a
compound known as esomeprazole. Esomeprazole is a proton pump inhibitor (PPI),
which is used in the reduction of gastric acid, reflux esophagitis and related
maladies. They are known, collectively, as GERD. Esomeprazole is sold in Canada as Nexium, in 20 and 40 mg strength tablets. It has proven to be a very successful
drug for the plaintiff AstraZeneca Aktiebolag.
[2]
Apotex Inc. sought to sell a generic version of
esomeprazole. Accordingly, it applied to the Minister of Health for a Notice
of Compliance (NOC) allowing it to do so. In response, AstraZeneca
brought an application for prohibition under the Patented Medicines (Notice
of Compliance) Regulations (SOR/93-133) (PMNOC Regulations)
prohibiting the Minister from issuing an NOC to Apotex. On June 30,
2010, Justice Roger Hughes, in AstraZeneca Canada Inc v Apotex Inc, 2010
FC 714 [Nexium NOC], dismissed the application for prohibition. Apotex
subsequently commenced to sell its generic version of esomeprazole,
precipitating these proceedings.
[3]
The validity of the ‘653 patent ultimately turns
on its proper interpretation, as informed by the following statement:
It is desirable to obtain compounds with
improved pharmacokinetic and metabolic properties which will give an
improved therapeutic profile such as a lower degree of interindividual
variation. The present invention provides such compounds, which are novel
salts of single enantiomers of omeprazole (emphasis added)
[4]
AstraZeneca contends that this is a mere
statement of a hoped for advantage or a goal that the compound may have
an “improved therapeutic profile,” and not a promise of such. In particular,
AstraZeneca contends that the use of the word “will” in
the patent supports its characterization of an “improved therapeutic profile” as merely a goal. “Will,” in
AstraZeneca’s submission, is prospective, and simply indicative of only a hope
or expectation. I will address this analysis in greater detail below –
I promise. Suffice to say, AstraZeneca seeks to circumvent the ordinary
meaning of “will” and a plain reading of the
phrase by truncating it – effectively reading out the phrase “which will give an improved therapeutic profile such as a
lower degree of individual variation.” This interpretation is not
consistent with the governing principles of patent utility nor is it consistent
with how the patent would be read by a skilled person.
[5]
Apotex, in response, contends that a plain
reading of the patent makes an explicit promise of an improved therapeutic
profile. In support of this position, Apotex relies on received principles
underlying the interpretation of the promise of the patent in the context of
utility. Accordingly, the ‘653 patent promises an improved therapeutic profile
– a promise that it fails to keep.
[6]
For the reasons that follow, AstraZeneca’s
action for infringement is dismissed, and Apotex’s counter-claim for a
declaration of invalidity is granted.
II.
Preliminary Issues
[7]
Prior to interpreting the ‘653 patent and
assessing its validity two preliminary issues merit discussion, namely, Apotex’s
view that AstraZeneca Canada lacks standing and secondly that AstraZeneca
Canada and AstraZeneca Aktiebolag should be precluded, in this trial, from re-litigating
the validity of its patent in the prior NOC proceeding.
A.
Standing of AstraZeneca Canada
[8]
The first preliminary issue is the standing of
AstraZeneca Canada. Under section 55(1) of the Patent Act, RSC, 1985, c
P-4, a patent-infringer is liable to “the patentee [in
this case, AstraZeneca Aktiebolag] and to all persons claiming under the
patentee.” As a consequence, the standing of AstraZeneca Canada turns on whether or not it qualifies as a person “claiming under” AstraZeneca
Aktiebolag.
[9]
Apotex contends that AstraZeneca Canada lacks
standing. In particular, Apotex asserts that neither the Further Amended Statement
of Claim nor the evidence support a finding that AstraZeneca Canada is a person
claiming under AstraZeneca Aktiebolag. In support of this assertion, Apotex
notes, correctly, that there is no express license agreement between
AstraZeneca Aktiebolag and AstraZeneca Canada.
[10]
In my view, AstraZeneca Canada has standing. More specifically, AstraZeneca Canada qualifies as a person claiming under the
patentee because there is an implied license between AstraZeneca and
AstraZeneca Canada regarding the sale of Nexium. However, prior to elaborating
on this finding, it is important to note the factual background underlying Apotex’s
surprisingly technical standing defence against its alleged infringement.
[11]
That factual background overwhelmingly supports
AstraZeneca Canada’s standing in this case. AstraZeneca Canada has sold Nexium in Canada for the last 13 years. During those 13 years, AstraZeneca
Aktiebolag has supplied AstraZeneca Canada with either bulk tablets or
pre-packaged Nexium except for two brief interruptions in supply where
AstraZeneca UK was a substitute supplier for approximately 3-6 months. In
turn, AstraZeneca Canada sold Nexium in Canadian markets, as its name
suggests, throughout those 13 years. Now, both AstraZeneca Canada and
AstraZeneca Aktiebolag are joined before this court seeking recovery for AstraZeneca
Canada’s losses caused by Apotex’s alleged infringement in the Canadian
market. For Apotex to claim that there was no implied license, no right
whatsoever, arising from a common understanding between AstraZeneca Canada and AstraZeneca Aktiebolag that AstraZeneca Canada was entitled to sell Nexium in Canada, strains credulity. Presumably, Apotex is of the opinion that AstraZeneca
Aktiebolag supplied AstraZeneca Canada with pre-packaged Nexium for some
purpose other than its sale – perhaps, for the profitable venture of storing
unsold pharmaceuticals. In this regard, Justice Rothstein’s remarks in Apotex
Inc v Wellcome Foundation Ltd, [2001] 1 FC 495 (FCA) at para 99 are on
point:
It is perhaps not uncalled for to observe that
this is not a case in which the alleged licensee is alone in advancing its
claim for patent infringement. Here, the patentee is also before the Court as a
co-plaintiff supporting the claim of GWI. It is difficult to conceive of what
more is necessary to prove the existence of a licence than to have the licensor
and licensee both attesting to the validity of the licence. Where both the
patentee and the person claiming under the patentee are before the Court, are
affiliated as being owned by the same parent and have an identity of interest
in the litigation--with the patentee supporting the person claiming under the
patentee--it is, to say the least, surprising that technical questions of
status to sue would be advanced as a defence to infringement.
[12]
On that basis alone, Apotex’s standing defence
rests on a weak foundation. Regardless, I will proceed to the merits of the
standing issue.
[13]
The substance of Apotex’s argument relates to
amendments to the pleadings of the Statement of Claim which, admittedly, result
in partial ambiguity with respect to the various relationships amongst the
AstraZeneca corporations. I described those amendments in an earlier ruling (Astrazeneca
v Apotex, T-1668-10, Reasons for Ruling, November 15, 2013). Relevant
to the issue of standing, those amendments re-characterized the relationships
between AstraZeneca Aktiebolag, AstraZeneca UK, and AstraZeneca Canada.
[14]
The Further Amended Statement of Claim, at
paragraph 5, state that AstraZeneca Aktiebolag is the exclusive holder of the
intellectual property rights to Nexium:
By reason of the grant of the patents, AstraZeneca AB has, in Canada, the exclusive right, privilege and liberty of making,
constructing, importing, exporting, using, offering for sale and selling to
others to be used, the invention claimed in the patents.
[15]
This exclusive right is confirmed in the
Distribution Agreement between AstraZeneca Aktiebolag and AstraZeneca Canada which authorizes AstraZeneca Canada, as a distributor of Nexium, to formulate and package
Nexium. However, the Further Amended Statement of Claim, surprisingly, do not
allege that AstraZeneca Canada has permission, express or implied, to sell
Nexium. In particular, paragraph 6(a) does not allege that permission:
AstraZeneca UK with the agreement of AstraZeneca AB can sell, and has sold, is a licensee of the patentee AstraZeneca AB in respect of the
patents and has sold and continues to sell
NEXIUM brand tablets containing (-)-omeprazole magnesium trihydrate to
AstraZeneca Canada who in turn distributes and sells the NEXIUM brand tablets
in Canada. (emphasis as in original).
[16]
It is from this omission of AstraZeneca Canada’s right to sell Nexium that Apotex grounds its defence that AstraZeneca Canada lacks
standing.
[17]
This lacunae in the claim is compounded by the absence of a plea of a licence,
express or implied, between AstraZeneca Canada and AstraZeneca Aktiebolag.
Rather, the pleadings make the bald statement that they are “related
companies,” In Apotex’s submission, the pleadings only support that
AstraZeneca Aktiebolag alone has the exclusive right to the invention claimed
in the patent.
[18]
The sole pleading of material fact in support of
AstraZeneca Canada’s standing is that AstraZeneca Canada sells and distributes
Nexium. Arguably, the Court is left to infer, from this, that a right
traceable to the patentee arises from the sale alone.
[19]
For their part, AstraZeneca Aktiebolag and
AstraZeneca Canada urge the Court to take a broad view of section 55(1) of the Patent
Act and the phrase “persons claiming under the
patent.” They also rely on evidence of the relationship between
AstraZeneca Aktiebolag and AstraZeneca Canada with respect to Nexium, and urge
a finding of an implied licence from AstraZeneca Aktiebolag to AstraZeneca Canada to sell Nexium in Canada, and thus, a right of AstraZeneca Canada to claim under the
patentee. As stated earlier, both the evidence, and common sense, support such
a finding.
[20]
The starting point in this analysis is the
decision of the Federal Court of Appeal in Signalisation
de Montréal Inc v Services de Béton Universels Ltée,
[1993] 1 FC 341 (FCA). In that decision, Justice Hugessen wrote, at paragraph
24:
In my view, a person “claiming under” the patentee is a person who derives his rights to use the patented
invention, at whatever degree, from the patentee. The right to use an
invention is one of the monopoly to which is conferred by a patent. When a
breach of that right is asserted by a person who can trace his title in a
direct line back to the patentee that person is “claiming under the patentee”. It matters not by what technical means the acquisition of the
right to use may have taken place. It may be a
straightforward assignment or a licence. It may, as I have indicated, be a
sale of an article embodying the invention. It may also be a lease thereof. What
matters is that the claimant asserts a right in the monopoly and the source of
that right may be traced back to the patentee (emphasis added)
[21]
In a more recent articulation of the test under
section 55(1), Justice Judith Snider examined whether the “right to use” the product could be traced back to the patent holder
and affirmed that Canadian jurisprudence has provided a broad interpretation of
“persons claiming under” the patentee which can
include the exclusive licensee, the non-exclusive licensee, the purchaser of a
patented article, and sales agents (Jay-Lor International Inc v Penta Farm
Systems Ltd, 2007 FC 358 at para 34).
[22]
The inquiry as to whether the right to use can
be traced back to the patentee is highly fact-dependent. In Apotex’s favour,
there is no express licence between AstraZeneca Canada and AstraZeneca
Aktiebolag. However, the existence of an express licence is not determinative
of whether a right may be traced back to the patentee. On the other side, in
AstraZeneca’s favour, it has sold Nexium for 13 years, though, as Apotex points
out, it too sells esomeprazole. Accordingly, the sale of a product alone is
similarly not determinative of whether a right may be traced back to the
patentee – there must be something more.
[23]
In this case, there is something more. Indeed,
a number of facts support the finding that AstraZeneca Canada’s right of use
can be traced back to AstraZeneca Aktiebolag:
1.
AstraZeneca Canada and AstraZeneca Aktiebolag
are both indirect subsidiaries of a common parent, AstraZeneca PLC, located in Sweden;
2.
AstraZeneca Aktiebolag, the owner of the ‘653
patent, is the principal source of supply to AstraZeneca Canada and globally;
3.
AstraZeneca Canada sought and obtained
regulatory approval to sell Nexium in Canada. The information in support of
the regulatory filing derived from AstraZeneca Aktiebolag – the holder of the
master regulatory file for Nexium;
4.
AstraZeneca Canada and AstraZeneca Aktiebolag
entered into a Formulation, Packaging and Distribution Agreement (Distribution
Agreement) in December 2000. In the Distribution Agreement, AstraZeneca Canada is defined as the “Distributor,”
and is granted non-exclusive rights to the “Products” which are defined to include Nexium. This agreement
addresses intellectual property rights in articles 24.1 and 24.2:
24 INTELLECTUAL PROPERTY RIGHTS
24.1 All intellectual property rights
relating to the Products shall remain the property of ASTRAZENECA at all
times. The Distributor shall not acquire any intellectual property rights
relating to the Products and shall only have permission to use such rights
granted to the Distributor under this Agreement.
24.2 The Distributor will inform
ASTRAZENECA of any infringement or suspected infringement of any of
ASTRAZENECA’s intellectual property rights in the Market which comes to the
notice of the Distributor. ASTRAZENECA will take all reasonable steps, at its
own expense, to prosecute infringers. The Distributor will give ASTRAZENECA
all reasonable assistance in such prosecution [emphasis added].
5.
From 2001-2008 AstraZeneca Canada packaged Nexium which it received from AstraZeneca Aktiebolag in bulk tablets, prior to sale
in Canada. In 2008, AstraZeneca Canada’s packaging facility in Mississauga was closed. The letter agreement between AstraZeneca Canada and AstraZeneca Aktiebolag dated December 12, 2007 stated that after closure, Nexium would be
supplied by AstraZeneca Aktiebolag to AstraZeneca Canada in finished packaged
form, and that AstraZeneca Canada would continue to act as the distributor.
Accordingly, after 2008, AstraZeneca Canada received pre-packaged Nexium from
AstraZeneca Aktiebolag for sale in Canada. Thus, AstraZeneca Canada has always
received its supply of Nexium (pre-packaged or in bulk) from AstraZeneca Aktiebolag,
except for a three month period in 2001 and a six month period in 2012, during
which AstraZeneca UK was the source of supply.
6.
According to the evidence of Ms. Elaine
Campbell, CEO of AstraZeneca Canada, AstraZeneca Canada has obtained the
consent of AstraZeneca Aktiebolag to file Form IV patent lists under the PMNOC
Regulations;
7.
Ms. Campbell testified that all of AstraZeneca Canada’s legal costs in respect of this litigation were being paid by AstraZeneca
Aktiebolag.
[24]
When assessed against this factual landscape,
AstraZeneca Canada’s right to use the patent may be traced back to AstraZeneca
Aktiebolag, the patentee. All rights of use of Nexium by AstraZeneca Canada are derivative, by an implied agreement, from AstraZeneca Aktiebolag. While there
is no express licence and no plea of licence, the conduct of the parties is
consistent with a finding of an implied licence granted by AstraZeneca
Aktiebolag. The Distribution Agreement grants AstraZeneca Canada permission to
use AstraZeneca Aktiebolag’s intellectual property rights “insofar as is
necessary to exercise the rights granted” under
the Distribution Agreement. These rights include the right to sell Nexium and
the obligation to assist AstraZeneca Aktiebolag in the civil prosecution of
possible infringement by others. Commencement of an infringement action by
AstraZeneca Canada falls within a reasonable interpretation of sections 24.1
and 24.2, and implicit to that is an acknowledgment of a right to recover
damages on behalf of the patentee for infringement. Consequently, AstraZeneca Canada is a person claiming under the patentee as required by section 55(2) of the Patent
Act and has standing in this trial.
B.
Preclusion from Contesting Invalidity
(1)
Relationship between NOC Proceedings and
an Action for Infringement
[25]
The parties advance diametrically opposed
positions with respect to the legal effect of the decision of Justice Hughes in
the prior NOC proceeding addressing the same patent (the Nexium NOC).
AstraZeneca contends that the decision is neither binding nor instructive.
Further, it says that if the Court does consider it, the decision is wrong and
should not be followed.
[26]
Apotex contends that the Nexium NOC must
have some meaningful legal effect or consequence, otherwise the NOC
proceedings provide an empty remedy. In support, Apotex contends that the
doctrines of issue estoppel and abuse of process preclude AstraZeneca from
re-litigating the same issues in this proceeding as were previously determined
in the NOC proceeding. In the alternative, it says that comity requires
that the reasoning and result reached by a judge of this Court ought to be
followed.
[27]
The distinctions between the PMNOC
proceedings and patent infringement proceedings are well known. They differ in
form (an application as opposed to a trial) and in remedy (prohibition as
opposed to declarations, damages, or an accounting of profits). It is settled
law that decisions taken in the NOC proceedings are not binding on
infringement actions or to declare a patent invalid. NOC proceedings
were never intended to be a surrogate for a trial on infringement. In
consequence, a plea of res judicata will be struck:
This Court has been very clear on the fact that
section 6 proceedings are not adjudicative of the rights of the patentee. In Merck
Frosst Canada, supra at 319, Hugessen J.A. rejected the notion that
prohibition proceedings could be assimilated to an action of any kind:
The proceedings are not an action and
their object is solely to prohibit the issuance of a notice of compliance under
the Food and Drug Regulations. Manifestly, they do not constitute "an
action for infringement of a patent.
In these circumstances, it is idle to suggest
that any decision that this Court makes in these appeals could be used to
attack collaterally a judgment in an infringement action (Pfizer Canada Inc
v Apotex Inc, (2001) 11 CPR (4th) 245 at para 25).
[28]
In Apotex Inc v Pfizer Ireland
Pharmaceuticals, 2011 FCA 77 at paras 19 and 24 [Apotex sildenafil],
the Court of Appeal observed that there are nonetheless circumstances where the
interrelationship between the two proceedings can give rise to a remedy in
estoppel, and it is on these passages that Apotex predicates its argument that
AstraZeneca should be precluded from pursuing this action:
Even where a later proceeding involves issues
quite different from an earlier proceeding, it may be open to a judge to apply
the doctrines of issue estoppel or abuse of process in the later proceeding to
prevent a party from relitigating certain factual and legal issues decided in
the earlier proceeding: Danyluk v. Ainsworth Technologies Inc., 2001 SCC
44 (CanLII), [2001] 2 S.C.R. 460, 2001 SCC 44 [Danyluk] (involving issue
estoppel) and Toronto (City) v. Canadian Union of Public Employees
(C.U.P.E.), Local 79, 2003 SCC 63 (CanLII), [2003] 3 S.C.R. 77, 2003 SCC 63
[C.U.P.E.] (involving abuse of process). Danyluk and C.U.P.E.
both emphasize that these bars against relitigation are discretionary and that
the discretion must be exercised taking into account a wide variety of circumstances.
[…]
This court has repeatedly said that NOC
proceedings are quite different from subsequent infringement or impeachment
actions. In my view, there is scope for applying the bars of issue estoppel
and abuse of process in the later proceedings to prevent the relitigation of
subsidiary factual and legal issues in order to preserve judicial resources,
promote the integrity of the justice system, prevent inconsistent findings, and
prevent abuse. The difference between the NOC proceeding and later proceedings
is an important consideration for the judge in the later proceedings, along
with all of the other discretionary considerations discussed in Danyluk
and C.U.P.E. Simply put, Danyluk and C.U.P.E. can apply
in proceedings such as these.
[29]
There is support for Apotex’s argument, both in
the legal policy objectives served by the doctrines of issue estoppel and abuse
of process, and as well in the Regulatory Impact Analysis Statement (RIAS)
which accompanies the 1998 amendments to the PMNOC Regulations.
[30]
While a defence of res
judicata will be struck, the doctrines of issue
estoppel and abuse of process remain open for consideration in the discretion
of the trial judge. Thus, the re-litigation of an issue decided against a
party in an NOC proceeding “is generally not permissible” (Apotex sildenafil, at para 22). Justice Sexton
gave one example where he could foresee the application of issue estoppel or
abuse of process at paragraph 26:
Specific
applications of the principles in Danyluk and C.U.P.E. should await later cases. But, for clarity, I
offer one illustration. If a witness gives exactly the same evidence in both
proceedings, and the judge found the witness not to be credible in the NOC
proceedings, it may be open to the trial judge in the action to bar
relitigation of the witness’s credibility through issue estoppel or abuse of
process. On the other hand, if the witness gives different or additional
evidence at the action, the trial judge may be justified in reconsidering the
witness’s credibility. There may of course be other considerations as well.
Obviously, this is a discretionary matter. Suffice to say, the facts that will
inform the discretion are not known in a pleadings motion.
[31]
The application of these principles in any
particular case is discretionary and informed by the factual context, to which
I now turn.
(2)
Issue Estoppel
[32]
There is, in this case, a significant overlap
between the issues and the evidence led in the two proceedings. Justice Hughes
found the allegation of invalidity based on lack of sound prediction and
obviousness to be justified (Nexium NOC, at paras 94 and 137). Sound
prediction and obviousness are again in issue in this trial. Apotex contends
that AstraZeneca should not be able to re-litigate the same issues, with the
same evidence, before another judge of this Court.
[33]
Apotex contends that all of the criteria to
engage issue estoppel are at play here: the same parties, the decision which
creates the estoppel is final, and the same question has been decided.
However, as the Court of Appeal emphasized, the doctrine depends on similarity
in the substance of the evidence, whether it is contested, and the Court’s
assessment of its weight and credibility. The fact that the same witnesses
testified in respect of the same issues does not alone dispose of the matter.
[34]
Five of AstraZeneca’s key witnesses in the NOC
proceeding also gave evidence at trial. However, the scope of their evidence
at trial, in the form of expert reports, reply reports, sur-reply reports, was
broader. Additionally, the form of giving evidence, viva
voce, distinguishes the nature of the evidence.
The issues in respect of which the experts testified were not constrained by
the Notice of Allegation. There was much new evidence, and some key witnesses,
whose evidence may have informed the appreciation of the testimony of other
witnesses, did not testify.
[35]
I note, in particular, the extensive use at
trial of prior testimony to impeach the evidence of witnesses. This exercise
of confronting witnesses with apparent inconsistencies provided this Court with
an appreciation of the evidence and witnesses which was unavailable to Justice
Hughes. Importantly, credibility was not critical to Justice Hughes’
assessment of the evidence. Accordingly, this is not a case, as envisioned by
Justice Sexton, where a party seeks to re-coup before one judge credibility
lost before another.
[36]
Put otherwise, while Apotex focuses on the
similarities, they are overshadowed by the differences. Though many cards in
the deck are the same, they have been shuffled, considerably. Additionally,
witnesses and the evidence they give take their colour, in part, from the
Court’s appreciation of other witnesses. Dr. Bernard Kohl, one of the
inventors behind a key piece of prior art in this case, gave evidence in the NOC
proceeding, but not in this action. As will be seen, in some cases my
conclusions as to the expert opinion evidence have been affected by my
observations of their demeanour in court. To conclude the issue estoppel
analysis, the evidentiary record was not shown to be sufficiently similar and
therefore the doctrine does not apply.
[37]
It is in the context of this guidance that I
have considered the decision of Justice Hughes. It is informative and
instructive, but it remains a decision taken in a different context on a
similar, but nonetheless different, record. Given the clear language of the
Court of Appeal in Apotex sildenafil to the effect that the NOC
decision does not make validity and infringement res
judicata, I have reached my own conclusions based
on a different evidentiary record.
(3)
Abuse of Process
[38]
Apotex has a second bow in its quiver. It
contends that AstraZeneca has adopted positions in this infringement action
that are inconsistent with the positions it adopted in the NOC
proceeding, thus engaging the doctrine of abuse of process (Toronto (City) v
CUPE, Local 79, 2003 SCC 63, [2003] 3 S.C.R. 77). In CUPE, the Supreme
Court of Canada observed, at paragraph 52, that “relitigation
carries serious detrimental effects and should be avoided unless the
circumstances dictate that relitigation is necessary to enhance the credibility
and the effectiveness of the adjudicative process as a whole”. In this
case, Apotex argues that AstraZeneca abusively adopted new and contradictory views
with respect to two issues: (1) the promise of the patent (relevant to utility)
and (2) the motivation to separate the enantiomers of omeprazole (relevant to
obviousness). Though these are new and contradictory views, for the reasons
that follow, I do not consider them abusive and worthy of precluding
AstraZeneca from advancing its arguments.
[39]
First, Apotex contends that AstraZeneca has
modified its position in respect of the promise of the patent. In the NOC
proceeding, AstraZeneca took the position that there was no promise of utility
in the ‘653 patent specification. At paragraph 85 of his decision, Justice
Hughes wrote:
Nowhere in the patent, whether in the Examples or otherwise, is
any information given to the person skilled in the art as to whether, in fact,
the highly pure esomeprazole salt does give an improved therapeutic profile
such as a lower degree of interindividual variation. There is no evidence from
any witness to say that there is anything in the disclosure of the ‘653 patent
that would inform a person skilled in the art that the purified esomeprazole
salt fulfills this promise. Counsel for AstraZeneca argued that all that was
required was that an alternative to racemic omeprazole be provided not whether
it is an improvement. This argument ignores the promise of the patent as set
out in the portion recited above at page 1 that the resulting product would
provide “an improved therapeutic profile” (emphasis in original)
[40]
In contrast, before this Court, AstraZeneca
argues that the patent promises improved properties.
[41]
I do not see the promise advanced by AstraZeneca
in this case to be such a change in position that it could be considered
abusive. Admittedly, it is a variation of AstraZeneca’s position, presumably
in response to the decision of Justice Hughes in which that position was
rejected. However, the promise of the patent is ultimately a legal question
about which a party may strategically tailor their arguments through multiple
proceedings. AstraZeneca’s shift from no promise to a minor promise, while
still a shift, is not abusive, but rather, an argument that has strategically
evolved in the course of multiple proceedings. Moreover, it is a shift with
respect to a question of law informed by expert testimony, the substance of
which has changed between the NOC proceeding and this infringement
trial. Finally, as I discuss below, I ultimately accept Apotex’s version of
the promise of the patent over AstraZeneca’s version, thus resulting in the
‘653’s invalidity for lack of utility. As a consequence, abusive or otherwise,
AstraZeneca’s change of heart does not alter the outcome or reasoning of this
decision.
[42]
Apotex is correct to say that the promise of the
patent was an issue before Justice Hughes and now before this Court and that
the administration of justice is not enhanced by having differing
interpretations on an identical legal issue between the same parties. However,
as I described earlier, it is also a legal issue informed by expert testimony,
the substance of which is different between the NOC proceedings and this
infringement action. As a consequence, while a patent’s promise does not vary
depending on the day it is litigated, the evidence before the court charged
with interpreting that promise may vary, and did vary in this case. If
anything, to rely on all of Justice Hughes legal conclusions in the NOC proceeding
after several months of litigation and expert testimony in this infringement
trial would be the gravest error.
[43]
Second, AstraZeneca now disputes, in this
infringement trial, that a person of ordinary skill in the art (the skilled
person) would be motivated to separate the enantiomers of omeprazole – a point
which it conceded before Justice Hughes in the NOC proceeding. At
paragraph 132 of the Nexium NOC, Justice Hughes wrote:
There was no serious argument raised by
AstraZeneca that a person could and would be sufficiently motivated to make a
salt of the esomeprazole enantiomer.
[44]
Further, at paragraph 43 of the Nexium NOC,
Justice Hughes wrote:
AstraZeneca, in argument, put forward a portion
of Dr. Caldwell’s affidavit, an Apotex expert, and agreed with that portion as
far as it went. I refer to paragraphs 114 and 115 (Record, page 5616), noting
that the acronym PPI stands for “proton pump inhibitor”, the proton pump being
that which is found in the stomach that produces acid:
114. Omeprazole was a blockbuster
PPI – By May 1993, it was common knowledge that omeprazole was a very
successful drug that was useful to treat conditions that required the
inhibition of gastric acid secretion in humans. This fact was described in many
sources available to skilled persons including Lindberg et. al., “Omeprazole:
The first Proton Pump Inhibitor.”
115. Skilled persons were motivated
to resolve omeprazole to its enantiomers and study their respective properties
– Omeprazole was a drug that was known to be racemic. It was also known that
both of its enantiomers were active as PPIs. It was also known that the two
enantiomers of omeprazole might be metabolized differently.
[45]
By contrast, in this action, AstraZeneca took
the position that there would be no motivation to investigate differences in
the activity of the single enantiomers, no motivation to investigate toxicity,
and no motivation to assess pharmacokinetic differences (or, at least, a very
limited motivation to investigate these issues).
[46]
I consider this change of position to be more
problematic than AstraZeneca’s change of position with respect to the promise
of the patent. In particular, a changed view on the question of motivation is
problematic because it relates to a question of fact with respect to the actual
motivations facing research scientists in the early 1990s, rather than a
question of law such as the promise of the patent (though I recognize that the
question of motivation centers on the motivations of a legal creation: the
skilled person).
[47]
That being said, the focus of this Court is not
on the conduct of AstraZeneca, but on truth and fact finding. Put otherwise,
the focus of this Court, regardless of AstraZeneca’s adoption of inconsistent
positions, remains on the evidence before it. Save in egregious cases, which
this is not, relevant and otherwise compelling expert evidence should not be
excluded by reason of the conduct of the party who called the witness. While I
do ultimately agree with AstraZeneca’s position and find that the skilled
person was not motivated to investigate the enantiomers of omeprazole, that is
because I prioritized weighing the most credible and compelling evidence before
me (in the interest of truth-seeking) over disciplining the parties for their
inconsistent positions between the NOC proceeding and this trial.
AstraZeneca’s inconsistent positions speaks to its own integrity, not that of
its witnesses, and Dr. Armstrong’s compelling analysis of the incentives and
barriers facing the skilled person in 1993 displayed, as I discuss below, that
a limited motivation existed to investigate the enantiomers of omeprazole at
that time. In that sense, AstraZeneca’s “relitigation” of the motivation question
was, in the phrasing of CUPE, “necessary to
enhance the credibility and the effectiveness of the adjudicative process”
(at para 52).
[48]
Additionally, I note that motivation is only one
of many factors considered in the obviousness analysis, the majority of which,
motivation aside, favour AstraZeneca’s position on obviousness. I also note
that, in any event, the ‘653 patent is ultimately invalidated for lack of
utility, rendering the patent’s validity with respect to obviousness
insufficient to save its overall validity.
III.
Interpretation (Construction)
[49]
Having addressed the preliminary issues in this
case, I turn to the interpretation and assessment of the validity of the ‘653
patent.
[50]
Before addressing the grounds of invalidity,
three interpretive aides must be established: (1) identifying the skilled
person, (2) identifying the skilled person’s common general knowledge, and (3)
interpreting (i.e. constructing) the claims of the patent from the perspective
of that skilled person.
A.
The Skilled Person
[51]
The skilled person is a notional person used by
the courts to ensure that patents are read in an “informed” way. For the purpose of patent law, and as a
reflection of reality, patents are notionally addressed to a skilled person
rather than an ordinary member of the public. The skilled person is “deemed to be unimaginative and uninventive, but at the same
time is understood to have an ordinary level of competence and knowledge
incidental to the field to which the patent relates and to be reasonably
diligent in keeping up with advances.” Additionally, the skilled person
can come from a single discipline, or reflect a combination of multiple
disciplines, depending on the nature of the patent: Merck & Co v
Pharmascience Inc, 2010 FC 510 at paras 34-40 [Merck finasteride].
[52]
The parties substantially agreed on the
characteristics of the skilled person. The skilled person of the ‘653 patent
is a composite of:
•
An organic or medicinal chemist;
•
A pharmacologist;
•
A pharmaceutical formulator; and
•
A physician familiar with the pharmaceutical
treatment of excess gastric acid secretion and related diseases.
[53]
This composite skilled person (who is really a
combination of skilled persons, or a skilled team), embodies the science
incidental to the ‘653 patent. Throughout the trial, both parties advanced
witnesses who described how their specific expertise related to a proper
reading of the ‘653 patent in light of the ground of invalidity they discussed:
chemistry for obviousness and anticipation, pharmacology for utility, medicine
for clinical effectiveness, etc. At the end of testimony, the above list
emerged as the various fields touched upon by the ‘653 patent. Accordingly,
the skilled person’s expertise in this case is a combination of the fields
listed above. To hold otherwise, and limit the skilled person to a single
field, would result in a skilled person who, for example, is knowledgeable in
the chemistry necessary to juxtapose the ‘653 with prior art for the purpose of
anticipation, but ignorant to the pharmacology necessary to understand the
scope of the patent’s promised uses for the purpose of utility. I note, also,
that a composite skilled person in this case reflects the diverse team of
experts likely employed by pharmaceutical companies to develop and test drugs
like Nexium, making the use of a composite skilled person particularly
appropriate in the context of a pharmaceutical patent.
B.
The Common General Knowledge
[54]
The common general knowledge of the skilled
person is also, for the most part, undisputed between the parties. It
includes:
•
Stereochemistry;
•
The role of stereochemistry in drug action;
•
Omeprazole, in particular;
▪ its salts, and
▪ its mechanism of action;
•
Esomeprazole and its salts;
•
The use of salts to improve drugs;
•
General resolution techniques; and
•
The motivation to purify enantiomers.
[55]
Many scientific concepts are discussed
throughout this judgment. However, certain key concepts, which are central to
the ‘653 patent, and which fall within the common general knowledge of the
skilled person, merit an immediate discussion to ground the remaining
analysis. That being said, any science considered in the interpretation of the
‘653 patent or the assessment of its validity throughout this judgment was
either science within the common general knowledge of the skilled person or
science the skilled person would have encountered through reasonably diligent
research relating to the research questions before him (e.g. a literature
search).
[56]
In this case, the relevant aspects of the ‘653
patent related to enantiomers and their optical purity. As a
consequence, my discussion of the common general knowledge will focus on those
two key concepts. I will first discuss the basic chemical principles within
the common general knowledge of the skilled person relating to enantiomers and
their optical purity. Then, I will briefly discuss aspects of the ‘653 patent
within the common general knowledge of the skilled person relating to
esomeprazole.
[57]
Enantiomers are molecules that are structurally
similar (and which, as a consequence, have identical physical and chemical
properties) but that differ with respect to their orientation in
three-dimensional space. Their structural similarity comes from the same atoms
connected by the same types of bonds. However, these molecules, which are
comprised of multiple atoms and bonds, can rotate around those bonds in many
ways resulting in different three dimensional orientations. When those
orientations are non-super-imposable mirror images of each other, they are
labelled enantiomers. In other words, though they are similar
structurally, enantiomers cannot be oriented such that their reflections are
identical. A useful analogy is a right and left hand, which are similar
structurally (one palm, five fingers), but when oriented similarly (palms
facing down) and reflected in a mirror plane, are not identical because the
thumbs point in opposite directions, resulting in non-super-imposability.
[58]
As can be seen in the image above, some objects
(like a left and right hand), despite being structured similarly, cannot be
oriented in such a way so as to be super-imposable on one another. Those
non-super-imposable objects are described as chiral. Chiral objects
(such as enantiomers) may be described as “handed” because of the convenient
analogy of a left and right hand for describing chirality. By contrast, other
objects (like two flasks), can be oriented so as to be super-imposable on one
another. Those super-imposable objects are described as achiral.
[59]
The optical purity of a mixture relates to
the ratio of enantiomers in that mixture (i.e. the relative proportions of the
two enantiomers in the mixture). However, before elaborating on optical
purity, chemical purity must be explained. The chemical purity of a mixture
refers to the degree to which one substance is contaminated by others. Put
differently, chemical purity is inversely related to the extent of chemical
contamination. Accordingly, a highly contaminated mixture would have a low
chemical purity, and a lowly contaminated mixture would have a high chemical
purity. Molecules may be contaminated by a variety of substances. Of
particular interest to the ‘653 patent is the contamination by certain
enantiomers in a mixture.
[60]
Combining chemical purity and enantiomers
produces optical purity. While chemical purity more generally addresses the
degree of contamination in a mixture, optical purity addresses the degree of
contamination in a mixture caused by varying levels of the two different
enantiomers contained within that mixture.
[61]
In this particular field of science, there are
two different ways of expressing the optical purity of an enantiomeric
compound. One way is to express optical purity in absolute terms i.e.
how much of the whole compound is made up of the enantiomer in question. A
second way is to express optical purity in terms of the excess of one
enantiomer over the other, enantiomeric excess or “ee” i.e. how much more
there is of one enantiomer over the other enantiomer in the compound. For
example, with respect to the ‘653, the two enantiomers are the (+) and (-)
enantiomers of omeprazole, which may be labelled (+)-omeprazole and
(-)-omeprazole (with (-)-omeprazole being synonymous with “esomeprazole,” the
key compound claimed in the ‘653 patent). Assume a compound (the (-) enantiomer)
was described as having an absolute optical purity of 90%. This would mean
that the (-) enantiomer comprises 90% of the whole mixture, leaving the
(+) enantiomer to comprise the remaining 10%. Alternatively, 90% absolute
purity could be translated into 80%ee – an expression in terms of enantiomeric
excess. In other words, 80%ee means that there is 80% more of the (-)
enantiomer than the (+) enantiomer in the mixture. To do the math, 90%
absolute purity means 80%ee because 90% (the (-) enantiomer) – 10% (the (+)
enantiomer) = 80%(ee). The optical purity in dispute in this case is put in
terms of enantiomeric excess.
[62]
Optical purity can play a significant role in
the efficacy of pharmaceuticals. The human body is made up of chiral molecules
that may react differently to the “left” or “right” handed enantiomer of a
particular drug. Sometimes, this difference in reaction between the human body
and the different enantiomers of a drug is so small that it is of trivial
interest in the pharmaceutical context. For example, Advil (or Ibuprofen) is
sold as a 1:1 mixture of its two enantiomers (known as a racemic mixture,
or a racemate) because the difference in effect between the two
enantiomers is insignificant. By contrast, Thalidomide “opened
the world’s eyes” to the significance of chirality. One enantiomer of
Thalidomide reduced morning sickness in pregnant women, while the other had the
tragic consequence of causing birth defects in their children. Thus it may be
desirable, depending on the compound, for a pharmaceutical which is made up of
enantiomers to be separated into its “left” and
“right” handed enantiomers in order to maximize optical
purity and in turn optimize its therapeutic effect.
[63]
In sum, when the ‘653 patent refers to “new compounds with high optical purity” it is referring
to a compound (either a left or right handed enantiomer) with low contamination
from the other enantiomer; contamination which may have significant
implications for the therapeutic effect of the drug on the human body.
[64]
Moving now to the specific subject matter of the
‘653 patent itself, the skilled person would have known by May 28, 1993 that
omeprazole (the precursor of the drug in the ‘653) is a racemate, containing
equal amounts of (-)-omeprazole (esomeprazole) and (+)-omeprazole. The skilled
person would have also known that omeprazole is a proton pump inhibitor
(or, PPI), acting by blocking the proton pumps within the gastric parietal
cells, and is useful as an inhibitor of gastric acid secretion and for the treatment
of gastric acid-related diseases. Finally, the skilled person would have known
that omeprazole was a very safe and effective drug.
C.
Claims Construction
(1)
The Analytical Approach to Claims Construction
[65]
Claims construction is the interpretation of the
patent’s claims preceding the validity analysis. This preliminary
interpretation ensures that a patent’s validity is assessed purposively, rather
than through an inordinately rigid and technical approach. However, while
claims construction may significantly impact subsequent issues in patent
validity and infringement, it is not a result-oriented exercise. As a
consequence, the claims of the ‘653 patent must be constructed before issues
such as validity or infringement are considered Whirlpool Corp v Camco Inc, 2000
SCC 67 at paras 43 and 49-50, [2000] 2 S.C.R. 1067 [Whirlpool]. The basic
objective of patent interpretation is isolating the “essential elements” of the patent: Free World Trust v Électro Santé Inc,
2000 SCC 66 at para 31, [2000] 2 S.C.R. 1024 [Free World Trust].
[66]
The overall approach to claims construction is “purposive.”
Thus, constructing patent claims must look beyond a literal interpretation and
instead consider how the claims would be read by the skilled person, as of the
publication date, with the purpose and context of the patent in mind, and with a
“mind willing to understand” the specification
that is addressed to him (Whirlpool, at paras 43 and 48-49). To
determine how the claims would be read by the skilled person generally requires
the assistance of expert testimony regarding the skilled person’s knowledge and
for explaining technical terms. That being said, the court, and not the
experts, must interpret the patent: Whirlpool¸ at paras 45 and 57; Merck
finasteride, at paras 69-70.
[67]
More practically, the analytical approach to
constructing the claims of a patent places different weight on different
sections of the patent which requires an understanding of their respective
purposes. The patent as a whole (the specification) may be divided into
two discrete sections: the claims and the remainder, known as the disclosure.
In other words, combining the claims and the disclosure yields the
specification (these terms are occasionally used differently in other
judgments, but for clarity, I will consistently use them as I have defined them
above). The claims, not surprisingly, are central to the question of claims
construction, though the disclosure can play a role in interpreting the meaning
of the claims. The tension between reading the claims in the context of the
specification, and abstaining from rewriting the claims, is a long-standing
struggle in patent interpretation: Metalliflex
Ltd v Rodi & Wienenberger AG (1960), [1961] S.C.R. 117 at 122 [Metalliflex];
Whirlpool, at para 48, 49(f), and 52.
[68]
In essence, a two-step approach to constructing
claims can be employed: (1) Are the claims, when read in a principled manner,
and in the context of the entire specification, clear and un-ambiguous? If
they are, then the disclosure should not be consulted for the purpose of
qualifying the scope of clear and unambiguous claims. However, if the claims
are ambiguous, then a further question must be asked: (2) do the claims and
disclosure complement or contradict each other? If they complement each other,
then they can be read harmoniously (i.e. the claims may be qualified by the
disclosure). If they contradict each other, then the disclosure cannot be used
to resolve the ambiguity because the consequence would be to inappropriately
re-write the claims.
[69]
This distilled approach to claims construction
necessarily flows from the recognition in binding appellate authorities that
the specification may be reviewed when constructing the claims, but only to
clarify ambiguities in a way that is harmonious with those claims. For
example, in Pfizer Canada Inc v Canada (Minister of Health), 2007 FCA
209 at para 39, the Federal Court of Appeal observes that “[t]he claim of the patent which is to be construed by the
Court must be read in the context of the rest of the specification”. I
would add to this, however, that reference to the rest of the specification
cannot be used to expand the patentee's monopoly as expressed in the claim.
[70]
To the same effect, the Supreme Court in Whirlpool
recognized how the disclosure is an aide that provides context to
understanding the claims when constructing the claims, rather than being a
source for additional claims. First, at paragraph 48, the Supreme Court states:
“the scope of the monopoly remains a function of the
written claims but, as before, flexibility and fairness is achieved by
differentiating the essential features (“the pith and marrow”) from the
unessential, based on a knowledgeable reading of the whole specification”
(emphasis added). Then, at paragraph 49(f) the Court noted: “While the appellants express concern that ‘purposive
construction’ may open the door to extrinsic evidence of intent […] neither
Catnic, supra, nor O'Hara, supra, goes outside the four corners of the
specification, and both properly limit themselves to the words of the claims
interpreted in the context of the specification as a whole” (emphasis added). Finally, the Court, at paragraph 52,
affirmed the following statement of Justice Taschereau in Metalliflex,
at 122:
The claims, of course, must be construed with reference to the
entire specifications, and the latter may therefore be considered in order to
assist in apprehending and construing a claim, but the patentee may not be
allowed to expand his monopoly specifically expressed in the claims ‘by borrowing
this or that gloss from other parts of the specifications.’
[71]
The two-step approach to claims construction
that I provide above only permits qualification of the claims by the disclosure
if the claims, when read in the context of the specification, are ambiguous.
This pre-condition of ambiguity is implicit to the above statements from
appellate authorities which all affirm that the claims are the
subject of construction, whereas the disclosure merely aids in that
construction. Otherwise, to allow qualification in the absence of ambiguity
would necessarily entail importing “claims” from the disclosure. As the Court
explained in Whirlpool: “More recently, Hayhurst,
supra, at p. 190, cautioned that ‘[t]erms must be read in context, and it is
therefore unsafe in many instances to conclude that a term is plain and
unambiguous without a careful review of the specification’” (at para
52, emphasis added). Put differently, the conclusion that a claim is “plain
and unambiguous” precludes qualifying the claim with reference to the
disclosure, as long as that conclusion of lacking ambiguity is founded on an
informed reading of the patent as a whole.
(2)
Construction of the ‘653 Patent’s Claims
[72]
With the approach to claims construction
established, the essential elements of the ‘653 patent may now be identified.
An “essential element” of a patent is either: an
element which, if varied, would make a difference to the way in which the
invention works, or an element which is essential irrespective of its practical
effect according to the intent of the inventor, expressed or inferred from the
claims: Free World Trust, at para 31. However, the “intent of the
inventor” is construed from the claims
objectively, and is not an attempt to ascertain his subjective intent: Free
World Trust, at para 66.
[73]
I will first provide a brief outline of the
disclosure of the ‘653. No disputes relevant to the construction of the claims
arose between the parties with respect to qualifying the claims with reference
to the disclosure. As a consequence, I will only briefly outline the
disclosure before proceeding to the essential elements of the claims. Other
aspects of the disclosure, as they relate to the grounds of invalidity, are
discussed in greater depth later on.
[74]
The ‘653 patent describes its invention as new
compounds with high optical purity, a process for their preparation, and their
use in medicine. Those compounds are the “novel salts of single enantiomers of
omeprazole.” By way of process, the ‘653 outlines how recrystallizing the
salts of partially separated enantiomers can result in enantiomers that have
optical purity of greater than 99.8%ee. Further, the ‘653 patent notes that
optically pure esomeprazole salts are resistant to racemization (the process
through which an enantiomerically pure mixture converts into an impure mixture
i.e. into a racemate).
[75]
The ‘653 patent acknowledges that the
enantiomers of omeprazole could be obtained by previous methods: (1) in
analytical scale using the High-performance Liquid Chromatography (HPLC)
technique taught by the Erlandsson paper (Erlandsson, P. et al Journal of
Chromatography, 532 (1990), 305-19) and (2) in preparative scale using a
technique involving a chiral auxiliary taught by the abandoned German Patent
Application No 40 34455 [DE 455] (see discussion of HPLC at paragraph 254).
[76]
The ‘653 patent states that its compounds have
improved pharmacokinetic and metabolic properties (understood by the skilled
person as including improved absorption, metabolism, distribution, and
elimination) which give an improved therapeutic profile (understood by the
skilled person as meaning a better or more predictable or consistent response
to the drug between people) such as a lower degree of interindividual variation
(understood by the skilled person as reduced variability between people with
respect to pharmacokinetics and pharmacodynamics). Additionally, the skilled
person would have understood that “improved,” in this context, refers to an
improvement over the precursor to esomeprazole: omeprazole. The ‘653 patent
also states that its compounds, similar to omeprazole, are effective gastric
acid secretion inhibitors and useful anti-ulcer agents.
[77]
The ‘653 patent provides various examples of how
to apply its method to obtain the single enantiomers of omeprazole. In
examples 10 and 11, the ‘653 patent exemplifies the preparation of
(-)-omeprazole having a purity of 94%ee and (+)-omeprazole having a purity of
98%ee, respectively. Most importantly, in examples 1-5, the ‘653 exemplifies
the preparation of the sodium and magnesium salts of the enantiomers of
omeprazole having purity ≥ 99.8%ee.
[78]
Having discussed the disclosure, I will now
proceed to the main focus of claims construction: the claims.
[79]
The ‘653 patent has 29 claims, of which claims
1, 2, 4, 6-8 and 25-27 are at issue (though the majority of evidence centred
solely on claims 7 and 8). Those claims may be loosely organized into two
groups: compound claims relating to the compound itself and its
composition (claims 1-2, 4-5 and 7-8) and use claims relating to how the
compound will be used (claims 25-27). A “cleansed” version of the relevant
claims, consisting of only their essential elements, is below.
[80]
The compound claims, in aggregate, relate to
optically pure salt forms of a specified chemical formula. More specifically,
the relevant compound claims may be constructed as follows:
•
Claim 1: An optically pure sodium,
magnesium, lithium, potassium, calcium or tetraalkylammonium salt of
esomeprazole.
•
Claim 2: A compound of claim 1 in solid
state form.
•
Claim 4: The sodium, magnesium or calcium
salts of esomeprazole of claims 1, 2 or 3.
•
Claim 5: The magnesium salt of esomeprazole
of claims 1, 2 or 3.
•
Claim 7: A compound of claims 1 to 6 having
an optical purity of 98% or greater.
•
Claim 8: A compound of claims 1 to 6 having
an optical purity of 99.8% or greater.
[81]
The use claims, in aggregate, relate to the use
of the claimed compounds in therapy and for the preparation of pharmaceutical
formulations for inhibiting gastric acid secretion and for the treatment of
gastric acid inflammatory diseases. More specifically, the relevant use claims
may be constructed as follows:
•
Claim 25: The use of the compounds of claims 1
to 8 in therapy.
•
Claim 26: The use of the compounds of claims 1
to 8 for preparation of a pharmaceutical formulation for inhibiting gastric
acid secretion.
•
Claim 27: The use of the compounds of claims 1
to 8 for the preparation of a pharmaceutical formulation for the treatment of
gastrointestinal inflammatory diseases.
[82]
With the claims constructed, I now proceed to
the grounds of invalidity for the ‘653 patent.
IV.
Utility
[83]
Apotex asserts that the ‘653 patent is invalid
because it was not useful. Utility is a requirement for an “invention” under
section 2 of the Patent Act. In essence, an alleged patent satisfies
the requirement of utility if, from the perspective of the skilled person as of
the filing date (May 27, 1994), its utility is demonstrated, or in the
alternative, if its utility is soundly predicted: Teva Canada Ltd v
Pfizer Canada Inc, 2012 SCC 60 at para 37, [2012] 3 S.C.R. 625 [Teva sildenafil].
[84]
A key concept underlying the utility analysis is
the “promise of the patent.” Given the centrality of utility in this case, a
review of its meaning is necessary.
A.
Legal Principles Regarding the Promise of the
Patent
[85]
The utility analysis is intimately connected to
the determination of the promise of the patent. Indeed, the promise of the
patent is “fundamental to the utility analysis”: Eli
Lilly Canada Inc v Novopharm Limited, 2010 FCA 197 at para 93 [Novopharm
Zyprexa].
[86]
Conceptually, the promise of the patent is the
yardstick against which utility is measured. Put differently, requiring that a
patent be useful begs the question: “useful for what?”
The answer to that question is the promise of the patent: Pfizer Canada Inc
v Mylan Pharmaceuticals ULC, 2011 FC 547 at paras 210-11 [Mylan Aricept].
In the words of the Supreme Court, inutility means “that
the invention will not work, either in the sense that it will not operate at
all or, more broadly, that it will not do what the specification promises that
it will do”: Consolboard Inc v MacMillan Bloedel (Saskatchewan) Ltd,
[1981] 1 S.C.R. 504 at 525. Thus, if a patent makes a promise, it will have
utility only if that promise is fulfilled.
[87]
Identifying the promise of the patent requires a
consideration of “the patent as a whole” (Novopharm
Zyprexa, at para 93). Claims construction, in contrast, only considers the
disclosure for broader context and to resolve ambiguities (discussed earlier).
Identifying the promise of the patent based on a review of the entire patent
specification, rather than the claims alone, is echoed in a recent article, Richard
Gold & Michael Shortt, “The Promise of the Patent in Canada and Around the World” (2014) 30:1 CIPR 35:
The majority tendency is […] to look to the
patent as a whole, including both the claims and the disclosure, in order to
construe the promise. As long ago as 1959, in a decision affirmed by the
Supreme Court of Canada, the Quebec Court of Queen’s Bench (Appeal Side) held
that an invention’s utility is to be assessed based on a holistic reading of
both the claims and the description:
[…]
The inevitable result of looking to the patent
specification as a whole is that the disclosure will furnish most promises,
since patentees are rarely required to discuss utility directly in the claims.
In most promise cases, the promise is found in an explicit statement in the
disclosure that explains the invention’s intended purpose, such as
“carboxyalkyldipeptides… are useful as inhibitors of angiotensin-converting
enzyme and as anti-hypertensive agents.
[88]
Indeed, the debate over the promise in this case
centres on a passage from the disclosure of the ‘653. While both parties have
differing perspectives on the scope of the promise flowing from the disclosure,
neither party disputes the relevance of the disclosure to the identification of
the promise of the ‘653.
[89]
The perspective adopted in ascertaining the
promise of the patent is that of “the [skilled person] in
relation to the science and information available at the time of filing”: Novopharm
Zyprexa, at para 93. However, while adopting such a perspective to
ascertain the promise of the patent may be assisted by expert evidence, it
remains a question of law in the exclusive province of the courts: Apotex
Inc v Pfizer Canada Inc, 2011 FCA 236 at para 17 [Apotex Xalatan].
To that end, in Mylan Aricept, at paras 218-24, Justice Hughes
reinforces the need for a clear demarcation of roles between experts and the
Court.
[90]
Finally, if the patent does not promise a
specific result, a “mere scintilla” of utility will suffice. However, if the
patent does promise a specific result, utility is measured against that
explicit promise. In essence, “[t]he question is whether
the invention does what the patent promises it will do”: Novopharm
Zyprexa, at para 76.
B.
The Utility Experts
[91]
The two utility experts (Dr. Urs Meyer for
Apotex, and Dr. Timothy Tracy for AstraZeneca) addressed the same mandates
regarding utility, namely:
a. To whom the ‘653 patent is addressed;
b. The subject matter to which the ‘653 patent and its claims relate;
c. The explicit promises (if any) made by the ‘653 patent regarding
useful properties of its subject matter; and
d. The demonstration or prediction of those properties as of May 27,
1994.
[92]
I first note that the utility experts were
refreshingly cooperative and credible. Both Drs. Meyer and Tracy were willing
to make concessions adverse to the interests of the party who called them,
reasonably qualified their answers, and generally came across as providing
honest and objective testimony for the assistance of the Court.
[93]
Additionally, there were areas of substantial
agreement within their expert reports and testimonies. In particular, both
experts essentially agreed on the skilled person of the ‘653 patent as a
person with knowledge in chemistry, basic and clinical pharmacology, and
pharmaceutical formulations and therapeutic use of such formulations to inhibit
gastric acid secretions and treat gastric-acid related disease. They also agreed
that the subject matter of the ‘653 patent was optically pure salts of
the enantiomers of omeprazole, described as novel compounds, having improved
pharmacokinetic and metabolic properties and high stability to racemization in
neutral and basic pH, a method to make them, and therapeutic uses.
[94]
Therefore, the key controversies relate to the
promise of the ‘653 patent regarding useful properties, and whether those
properties were demonstrated or soundly predicted.
C.
The Promise of the ‘653 Patent
[95]
The three potential promises found in the ‘653
patent include:
a. Use as a proton pump inhibitor;
b. Stability against racemization; and
c. Improved pharmacokinetic and metabolic properties.
(1)
Preliminary Issue: The “Promise” of Stability against Racemization
[96]
The second potential promise – stability against
racemization – is complicated. An initial hurdle is whether or not it is a
promise at all (unlike the hurdles facing the other promises, which centre on
their scope). AstraZeneca characterizes stability against racemization as a
promise, whereas Apotex characterizes it as an “integral aspect” of the third
promise (improved pharmokinetic and metabolic properties), rather than as a
promise in itself. As Dr. Meyer opined:
The pharmacologist would understand that
stability of the single enantiomers to racemization allows the single
enantiomers to be used therapeutically in humans to provide the promised lower
degree of interindividual variation. […] Thus, stability of the single
enantiomers to racemization is an integral aspect of the promise that they will
provide an improved therapeutic profile such as a lower degree of
interindividual variation. (Meyer Report, at para 92; emphasis added).
[97]
I agree with this characterization described by
Dr. Meyer. The promise of the patent, it must be recalled, is related to the
patent’s utility. Thus, the promise must be related to how the patent will
ultimately be used (assuming there is an explicit promise made, which both
experts agreed on). The patent in this case is not useful for possessing
the chemical property of being stable against racemization; it is useful as a
pharmaceutical drug in therapy. Stability against racemization merely enables
that use and is not a use in itself.
[98]
By way of analogy, think of “stability against
racemization” as a bridge over a river and “use in
therapy” as the other side of the river that you want to reach. If the patent
promises that it gets you across the river, and the intended means is a bridge,
then the promised use of the patent is still merely crossing the river; the
bridge is just the intended route. Scientifically speaking, the ‘653 patent
may still be useful, even if it were not perfectly stable against racemization,
if the period of racemization was so long that the drug was still useful in therapy
notwithstanding the eventual racemization that only meaningfully degraded the
compound decades later. Metaphorically speaking, if the patent still gets you
across the river (for instance, by boat), then the promise (crossing the river)
is still achieved, whether or not the intended means (a bridge) is ultimately
used.
[99]
Before leaving this point, I note that in the Nexium
NOC Justice Hughes makes the same observation in respect of the ‘653
patent:
It is important to distinguish between the
utility promised by the patent – “improved therapeutic profile, such as lower degree of
interindividual variation” – and the particular property that makes that possible “high
stability towards racemization” (Nexium NOC, at para 84; emphasis
in original).
[100] In sum, the only two promises of utility in this case are use as a
PPI and improved properties. I recognize that stability against racemization
is itself an “improved property,” but it is an intermediate property upon which
the ultimate use in therapy depends and is not useful in itself in the context
of the ‘653. Nevertheless, I will discuss the “promise” of stability against
racemization below in the event that a different interpretation of the promise
(a question of law) is found so that all relevant findings of fact are provided
in this judgment. For ease of writing, from this point on I will treat
stability against racemization as a promise. However, for maximal clarity, I
conclude the utility analysis with a section summarizing all of my key legal
and factual findings regarding the promised utilities of the ‘653 patent.
(2)
Common Ground Regarding All Three Promises
[101]
The first promise – the use of the ‘653 patent
as a proton pump inhibitor – is agreed to between the experts.
[102] Similarly, parts of the second and third promises are also agreed to
between the experts (besides the characterization of the second promise as a
promise, discussed above). However, both the second and third promises are
either extended by Apotex or truncated by AstraZeneca (depending on which
side’s characterization you prefer). Those extensions and truncations are the
subject of dispute between the witnesses with respect to the promise.
[103] Regarding the second promise, both experts agree to a promise of
stability against racemization with respect to chemical stability. However,
Apotex further asserts that the ‘653 patent promises enzymatic
stability, which AstraZeneca disputes.
[104] Regarding the third promise, both experts agree to a promise of
improved metabolic and pharmacokinetic properties. However, this promise is
interpreted by Apotex to include the qualification that such properties include
“an improved therapeutic profile such as a lower degree
of interindividual variation.” AstraZeneca disputes this purported
extension as well.
[105] These extensions – stability against enzyme-mediated racemization
and an improved therapeutic profile such as a lower degree of interindividual
variation – are the disputed promises.
(3)
Does the ‘653 Patent Promise Stability against
Enzyme-Mediated Racemization?
[106]
As stated earlier, the promise of the patent is
interpreted by reference to the entire patent including both the claims and the
disclosure: Novopharm Zyprexa, at para 93.
[107] The following portion of the ‘653 patent’s disclosure (from the “Detailed
Description of the Invention”) was central to the experts’ disagreement with
respect to a promise regarding stability against racemization:
[T]he optically pure salts are stable
towards racemization both in neutral pH and basic pH, which was surprising since
the known deprotonation at the carbon atom between the pyridine ring and
the chiral sulfur atom was expected to cause racemization under alkaline
conditions. This high stability towards racemization makes it
possible to use a single enantiomeric salt of the invention in therapy.
(‘653 patent, page 4, lines 16-22; emphasis added).
[108] The experts agree that the promise of stability against racemization
is limited, by the express language of the patent above, to circumstances of
neutral and basic pH.
[109] Dr. Tracy’s interpretation focuses on the earlier emphasized
portions in the passage above: that the stability of the salts was surprising
because “the known deprotonation […] was expected to
cause racemization.” Dr. Tracy argues that deprotonation – a chemical
transformation – links the claimed stability to chemical racemization.
Additionally, Dr. Tracy argues that the patent only claims stability against
chemical racemization because racemization is “only rarely” mediated by
enzymatic processes, and because the experiments described in the patent
examine chemical-mediated racemization.
[110] In contrast, Dr. Meyer’s interpretation focuses on the later
emphasized portions in the passage above: that the use of the salt in therapy
depends on its stability. Dr. Meyer argues that the link of the claimed
stability to a use in therapy necessitates both chemical and enzymatic
stability towards racemization. However, Dr. Meyer also admits that this link
only necessitates enzymatic stability if the therapy claimed includes “an improved therapeutic profile such as a lower degree of
interindividual variation” (which relates to the third disputed promise).
As a consequence, the merit of Dr. Meyer’s view of the promise regarding
enzymatic stability is predicated on the assumption, and his interpretation,
that the patent also promises an improved therapeutic profile. Put
differently, according to Dr. Meyer, if the ‘653 patent promises an improved
therapeutic profile, then it must also promise enzymatic stability.
Alternatively, if the ‘653 patent does not promise an improved therapeutic
profile, then it does not promise enzymatic stability.
[111] These logical deductions are agreed to by Dr. Tracy. He
acknowledges the link between stability and therapy, but limits that therapy to
the salt’s effectiveness as a PPI, which precludes a promise of enzymatic
stability. In sum, both experts agree that the promises of stability against
racemization and therapeutic use go hand in hand. Either (1) the patent
promises both an improved therapeutic profile and stability against
enzyme-mediated racemization, or (2) the patent promises neither.
[112] With that relationship between the second and third promises
established, I now consider the third promise upon which the scope of the
second promise depends: an improved therapeutic profile.
(4)
Does the ‘653 Patent Promise an Improved
Therapeutic Profile such as a Lower Degree of Interindividual Variation?
[113] This question distils to the proper interpretation, in light of the
patent as a whole, of the following passage from the ‘653 patent:
It is desirable to obtain compounds with
improved pharmacokinetic and metabolic properties which will give an
improved therapeutic profile such as a lower degree of interindividual
variation. The present invention provides such compounds, which are novel
salts of single enantiomers of omeprazole. (‘653 patent, page 1, lines 18-22;
emphasis added).
[114] Apotex interprets the emphasized portion above to be an explicit
promise. In contrast, AstraZeneca interprets the emphasized portion above as a
“goal” that falls short of an explicit promise. I will first outline the
difference between “goals” and “promises” in the
jurisprudence. Then, I will dispose of the arguments raised by AstraZeneca in
this regard which, in turn, will explain why the emphasized portion above is
best characterized as a promise of the ‘653 patent.
[115] There is a difference between the goals that a patent hopes to
address, and the outcomes that a patent promises will occur. In AstraZeneca
Canada Inc v Mylan Pharmaceuticals ULC, 2011 FC 1023 [Mylan Arimidex],
I observed that “not all statements of advantage in a
patent rise to the level of a promise. A goal is not necessarily a promise”
(at para 139). This distinction between goals and promises has been affirmed
by the Federal Court of Appeal (see e.g. Apotex Inc v Sanofi-Aventis,
2013 FCA 186 at para 67 [Sanofi-Aventis Plavix]).
[116] Differentiating goals and promises is a question of
characterization. Thus, before interpreting whether or not the ‘653 patent’s
reference to an improved therapeutic profile is a goal or a promise, goals must
be distinguished from promises in the abstract.
[117] Goals merely describe “a hoped-for advantage
of the invention” (Mylan Arimidex, at para 139). For example, in
Mylan Arimidex, I found that an object clause, beginning with “it is a particular object of the present invention to,”
merely described a goal that the patent strived to achieve rather than a
promised outcome. Similarly, in Sanofi-Aventis Plavix, at paras 55-67,
Justice Pelletier found the inference of a promise of therapeutic utility based
on indirect references to the use of the drug in humans (e.g. references to
human diseases and dosages that potentially correspond to use in humans) was
insufficient to substantiate a promise and merely alluded to potential uses.
In sum, promises are explicit and define guaranteed or anticipated results
from the patent (depending on whether the promise is demonstrated or soundly
predicted), whereas goals merely relate to potential uses for the
patent.
[118] AstraZeneca advanced two arguments in support of an improved
therapeutic profile being merely a goal (which in turn supports the truncated
promise). First, it argued that the use of “will”
in the patent supports the characterization of an improved therapeutic profile as
a goal rather than promise. Second, it argued that the absence of clinical
trials from the specification supports this characterization. Neither argument
can succeed.
[119] First, AstraZeneca argued that the use of “will” is prospective and therefore indicative of an
expectation or goal rather than a promise. I cannot accept that argument.
Promises (and goals) are both inherently prospective. Thus, the use of
prospective language such as “will” does not
undermine the promissory nature of this passage. The fact that promises are
forward looking should be self-evident, though I also note that such an
understanding of a promise, and how the word “will”
conveys that understanding, is additionally supported by the jurisprudence (Consolboard,
at 525: “what the specification promises that it will
do” ; emphasis added) and the dictionary definition of the word “promise” (Oxford Concise Dictionary, 12th
ed, sub verbo “promise”: “a declaration or assurance that one will do something
or that a particular thing will happen”; emphasis added). No
persuasive explanation was provided by AstraZeneca as to why the skilled person
would understand “will” differently than its
ordinary meaning.
[120] There are much clearer terms that express the view advanced by
AstraZeneca. Had the patent stated that such compounds “may”
or “could” give an improved therapeutic profile,
then the argument that such statements referred merely to a goal would be more
compelling. The same cannot be said of “will.” Will does not convey a low threshold of potential
outcomes, but to the contrary, a high threshold of probable or certain
outcomes that will occur, which in turn, suggests that such outcomes are
promised by the patent.
[121] Second, AstraZeneca argued that the skilled person would
characterize an improved therapeutic profile as merely a goal because such a
promise could not be substantiated in a patent that does not contain clinical
studies. In other words, will, must, in the absence of clinical
studies, mean may. However, this argument is predicated on the skilled
person assuming that, when reading a patent, all promises made within the
patent are also demonstrated therein. This assumption is not supported by the
evidence. Moreover, extrapolating on such an assumption would effectively
nullify the possibility for a patent ever being found invalid for a lack of
utility.
[122] Based on the evidence, and on the plain and ordinary meaning of the
language used, the skilled person would not read down the unequivocal language
of the patent promising an improved therapeutic profile because of the absence
of clinical studies. The evidence of Dr. Meyer (for Apotex) is more compelling
in this regard.
[123] First, Dr. Meyer is clear that, in his view, the language of the
patent “unambiguously” promises an improved therapeutic profile:
Q. Okay. Let me back up. The first part of
that language talks about it being desirable to obtain compounds with improved
pharmacokinetic and metabolic properties. I can pause there. Do you agree
that's what the language says?
A. Yes, that's desirable to have these
compounds.
Q. And then the last sentence says that the
present invention provides such compounds?
A. Yes, but the pharmacokinetic and
metabolic properties are then identified with such unambiguous language as “will
give an improved therapeutic profile, such as a lower degree of interindividual
variation,” so it's clearly qualified what these pharmacokinetic and metabolic
properties are. That's how I understood the
patent. (Meyer Cross, Trial Transcript, Vol 9, p 1458 at line 28 – p 1459 at
line 18; emphasis added).
[124] Second, Dr. Meyer clearly explains how the absence of clinical
studies does not change his perspective about the promise of an improved
therapeutic profile:
I thought that if you say “will
give,” that this was quite clear that these compounds will have these
properties. These compounds, you know, that the
present invention provides, will have these properties. It's clear that these
are properties in therapy, in vivo.
You know? Even if you have in vitro data and all that, these are properties that relate to therapy and
to therapy of patients. (Meyer Cross, Trial
Transcript, Vol 9, p 1462 at lines 3 – 11; emphasis added).
[125] By contrast, Dr. Tracy (for AstraZeneca) reads out the qualification
of an improved therapeutic profile, appears to restrict his analysis to a
subset of the claims (as opposed to the patent as a whole, as he should have),
and provides an unpersuasive explanation for doing so:
Q. And then in 92, you give the promise only
for claims 1 to 8. Right?
A. In 92, yes.
Q. Can you tell us how it came to be that in
answering the same question you immediately came to claims 1 to 8 as opposed to
the patent as a whole?
A. From what I was stating here, claims 1 to 8
dealt with stability toward chemical mediated racemization, proton pump
inhibitors and pharmacokinetic and metabolic properties that were in claims 1
through 8. I did not address anything beyond that.
Q. I gather that. You don't have separate
sections dealing with the promise of claims 9 to 24 or the promise of claims 25
to 29. You limit yourself to claims 1 to 8, and I want to know why it is that
you came to do that given that your mandate wasn't to do that.
A. I addressed it from the concept of the
person, of a pharmacologist, and I felt the others related to others than a
pharmacologist. I used the pharmacologist as the skilled person.
[…]
Q. […] you didn't look at claims 9 through the
rest of them to figure out what the promise was because that's not what the
pharmacologist would have read?
A. That's correct.
Q. Are pharmacologists not concerned with
interindividual variability?
A. Yes, they are.
Q. Doesn't claim 28, one of the claims
you didn't direct yourself to and didn't consider as part of the promise, you
say, doesn't it talk about using a compound in the manufacture of
medicament with a lower degree of interindividual variation in plasma
levels?
A. Yes, it does.
Q. Wouldn't that be exactly what a
pharmacologist would be interested in?
A. It would be one of the concepts, yes.
Q. So it can't be that you excluded yourself
from considering that claim as part of the promise because a pharmacologist
wouldn't consider it.
A. No. I did not consider it.
(Tracy Cross, Trial Transcript, Vol 15, p 2440
at line 3 - 2442 at line 1; emphasis added).
[126] Both sides agree that the skilled person in this case included a
pharmacologist. And Both Dr. Meyer and Dr. Tracy consider interindividual
variability relevant to the field of pharmacology. Yet, in Dr. Tracy’s
opinion, he disregards those sections of the patent that refer to reduced
interindividual variation. As a consequence, I prefer the evidence of Dr.
Meyer who considered the patent as a whole, and who concluded that reduced
interindividual variability was unambiguously provided for in the ‘653 patent.
Indeed, the language of the ‘653 patent clearly provides for such a promise.
It describes “compounds” with improved properties “which will give an improved therapeutic profile such as a
lower degree of interindividual variation” as “desirable”
and then states that “[t]he present invention provides
such compounds” – an unambiguous statement that the ‘653 promises to
provide compounds with a lower degree of interindividual variation.
(5)
Additional Grounds for Rejecting AstraZeneca’s
Proposed Promise
[127] The evidence aside, the interpretation advanced by AstraZeneca is
unpalatable because it is tautological, promotes perverse incentives for
innovators, and fails to appreciate the distinction between disclosure and
utility requirements under the Patent Act.
[128] First, AstraZeneca’s approach to utility is tautological. On a high
level, the promise is the yardstick against which utility is measured for the
purpose of demonstration. Yet, AstraZeneca proposes a backwards approach that
establishes that benchmark based on what can ultimately be demonstrated in the
patent. To circumscribe the scope of the promise based on what is demonstrated
in the patent makes it impossible to ever conclude that a patent is invalid for
lack of utility. No matter how broad a promise (e.g. this drug cures cancer),
it would always be read down to a narrower promise based on what was
demonstrated. Such an approach would run contrary to the policy objectives of
patent law which serve to create consistency and clarity in the bargain struck
between innovators and the public. Instead, unequivocal promises in patents
could in no way be relied upon and would be subordinate to more complex
questions of demonstration within the patent.
[129] Moreover, such an approach would perversely encourage patentees to
over-promise. The broadest possible promise would potentially provide the
greatest protection against invalidity based on obviousness and anticipation
(though I recognize that there are differences between the legal constructions
underlying those inquiries, such as the promise of the patent and the inventive
concept). Simultaneously, that overbroad promise would present no concern with
respect to invalidity based on inutility because any promise that cannot be
demonstrated will be read down to what can be demonstrated.
[130] Finally, such an interpretation disregards how patent law currently
accommodates scenarios in which the promised utility is not demonstrated in the
patent. I discuss the disclosure requirements with respect to utility in
greater depth later on, but for now I will simply state that it is not in
dispute that disclosure is not required for the demonstration of utility. As a
consequence, it is illogical to interpret the scope of a patent’s promise based
on the assumption that the evidence of demonstration will be disclosed in the
patent. While extrapolating on the legal implications of AstraZeneca’s
interpretation is not directly relevant to how the skilled person would read
the ‘653 patent, I find it hard to accept AstraZeneca’s perspective on the skilled
person when it would result in nullifying pivotal doctrines of patent law.
[131] As a final note, AstraZeneca’s argument with respect to clinical
studies is also inconsistent with its interpretation of another promise in the
‘653 patent. AstraZeneca advances that the patent promises improved properties
that are either demonstrated or soundly predicted by studies that, like
clinical studies, are also absent from the specification. But how can the
patent promise improved properties when the studies demonstrating or soundly
predicting such properties are not disclosed? By AstraZeneca’s own reasoning,
the skilled person, without such studies, would not view a promise of improved
properties as credible. Needless to say, such reasoning is not the proper approach
to the construction of the promise.
[132] I find that the ‘653 patent promises improved metabolic and
pharmacokinetic properties which will give an improved therapeutic profile such
as a lower degree of interindividual variation. As a corollary of this promise
(as explained in an earlier section), I therefore also find that the ‘653
patent promises stability against enzyme-mediated racemization – a precondition
to such improved pharmacokinetic and metabolic properties.
[133] In summary, the unambiguous wording of the patent supports three
promises (subject to my caveats about the promise of stability against
racemization described earlier):
a. Use as a proton pump inhibitor;
b. Stability against racemization (both chemical and enzyme-mediated)
in neutral and basic pH; and
c. Improved pharmacokinetic and metabolic properties which will give an
improved therapeutic profile such as a lower degree of interindividual
variation.
D.
Demonstration and Sound Prediction of Utility
[134] Having established the promises outlined above I turn to whether or
not those promises were demonstrated or soundly predicted. I will consider the
demonstration and sound prediction of both the full and truncated versions of
the promises in the event that my interpretation of any of the promises is
subsequently reconsidered.
(1)
The Evidence Regarding Demonstration and Sound
Prediction of Utility
[135] The ‘653 patent does not contain studies or references to studies
that were relevant to the issue of demonstration or sound prediction of utility
(except with respect to the promise of use as a PPI, which is not in dispute).
Rather, the evidence at trial in this regard consisted of several internal
studies performed by AstraZeneca.
[136] The two key studies were a human liver microsomal study and two
human blood plasma re-analyses. In addition, studies performed in rats were
also considered. These studies were not disclosed in the ‘653 patent.
However, as I explain below in the discussion of proper disclosure, this has no
bearing on the ultimate legal and factual conclusions. Regardless, I note the
absence of the studies from the patent here in the event that their disclosure
becomes relevant on appeal.
(2)
The Law on Demonstration and Sound Prediction of
Utility
(a)
Differentiating Demonstrated and Soundly
Predicted Utility
[137] Demonstrated utility is based on whether, as of the filing date,
there was proof that the patent worked as it promised. In the words of the
Federal Court of Appeal: “what amounts to demonstrated
utility would be evidence that establishes that the embodiment at issue does in
fact work in a manner that gives rise to the advantages stated in the patent”
(Eurocopter v Bell Helicopter
Textron Canada Ltée, 2013 FCA 219 at para 147 [Eurocopter]).
[138] By contrast, soundly predicted utility is based on the three part
test set out by the Supreme Court of Canada in Apotex Inc v Wellcome
Foundation Ltd, 2002 SCC 77 at para 70, [2002] 4 S.C.R. 153 [AZT]: (1)
there must be a factual basis for the prediction, (2) the inventor must have at
the date of the patent application an articulable and sound line of reasoning
from which the desired result can be inferred from the factual basis, and (3)
there must be proper disclosure.
(b)
The Meaning of Proper Disclosure
[139] Before turning to the demonstration and sound prediction of utility
I will first dispose of the issue of proper disclosure. To be clear, proper
disclosure is only engaged in this case in the alternative. For proper
disclosure to influence the final outcome in this case both my interpretation
of the promise and my conclusions on demonstration and sound prediction would
have to be incorrect. Regardless, in order to ensure a complete record, I am
committed to making all legal and factual conclusions. Furthermore, given the
recent evolution in the jurisprudence on the issue of proper disclosure in the
context of sound prediction it deserves more detailed consideration.
[140] The question of proper disclosure is engaged in this case with
respect to the second and third promises. The promises of stability against
enzyme-mediated racemization and an improved therapeutic profile are, according
to AstraZeneca, soundly predicted because of studies that were internal to
AstraZeneca and not disclosed in the ‘653 patent. I conclude that the patent
promises reduced interindividual variation, and that none of the studies,
disclosed or otherwise, demonstrate or soundly predict such utility. However,
if the validity of the patent were to depend on the sound prediction of the
second or third promises, Apotex further submits that such a sound prediction
is invalid in law because it fails to satisfy the requirement of proper
disclosure.
[141] The law on proper disclosure with respect to utility is unsettled.
However, my reading of the jurisprudence suggests that the requirement for
proper disclosure of utility is limited to the context of “new use” patents,
assuming such a utility disclosure requirement exists at all. As a
consequence, I would rule that the failure of AstraZeneca to disclose the
studies it relies upon for a sound prediction in this case would be irrelevant
because the ‘653 patent is not a new use patent.
[142] By way of outline, my view that “proper disclosure” of utility
applies only to new use patents is based on my reading of AZT, together
with the Supreme Court’s recent obiter remarks in Teva sildenafil, which I
conclude overturn previous Federal Court of Appeal decisions alleging such a
disclosure requirement in all cases of sound prediction. Notably, my reading
of AZT finds support in Justice Gauthier’s recent comments at the
Federal Court of Appeal in her concurring remarks in Sanofi-Aventis Plavix
– a decision released after Teva sildenafil.
[143] The starting point for the requirement of a valid patent is the Patent
Act. In particular, with respect to the disclosure of utility, sections 2
and 27(3) are of interest.
[144] Utility and disclosure are addressed separately in the Patent Act.
Thus, there is no statutory basis for combining the two to create a disclosure
requirement for utility. Section 2 – which addresses utility – provides that
an “invention” by definition is “useful” and makes no mention of such utility
needing to be disclosed. Section 27(3) – which addresses disclosure – provides
that “[t]he specification of an invention must”
disclose several things, none of which include the demonstration or prediction
of an invention’s utility. Read together, there is no statutory basis for a
requirement to disclose either the factual basis or the sound line of reasoning
required to support a sound prediction of utility. As I will point out below,
this reading of the Patent Act was affirmed by a unanimous decision from
the Supreme Court of Canada in 2012.
[145] The next authority to be considered with respect to characterizing a
“proper disclosure” is the Supreme Court of Canada’s decision in AZT.
This decision is the jurisprudential basis for the view that there is a
disclosure requirement for utility in cases of sound prediction. However, as I
will show, such an argument as advanced by Apotex, does not account for the
Supreme Court’s subsequent remarks in Teva sildenafil.
[146] To begin, Justice Binnie’s discussion in AZT of proper
disclosure is best characterized as a general rule with an exception. First,
he describes the general rule:
Normally, it is
sufficient if the specification provides a full, clear and exact description of
the nature of the invention and the manner in which it can be practised […] It
is generally not necessary for an inventor to provide a theory of why the
invention works. Practical readers merely want to know that it does work and
how to work it. (at para 70, emphasis added)
[147]
Then he describes the exception:
In this sort
of case, however, the sound prediction is to some
extent the quid pro quo the
applicant offers in exchange for the patent monopoly. (at para 70, emphasis
added)
[148] Finally, he elects to not elaborate on this exception because it
makes no difference in AZT and would therefore be obiter dicta:
Precise disclosure requirements in this regard
do not arise for decision in this case because both the underlying facts (the
test data) and the line of reasoning (the chain terminator effect) were in fact
disclosed, and disclosure in this respect did not become an issue between the
parties. I therefore say no more about it.
[149] The question of proper disclosure therefore turns on the
scope of the exception flagged by Justice Binnie in AZT. Put
differently, when Justice Binnie says that the sound prediction is the quid pro quo “in this sort of case” (at para 70),
what sort of case is he referring to? The Federal Court of Appeal interpreted
this to mean in cases of sound prediction:
The decision of the Supreme Court in AZT is
particularly significant to the disposition of this appeal. […] As was said
in that case (para. 70): “the sound prediction is to some extent the quid pro quo the
applicant offers in exchange for the patent monopoly.” In sound
prediction cases there is a heightened obligation to disclose the
underlying facts and the line of reasoning for inventions that comprise the
prediction. (Elli Lilly Canada Inc v Apotex Inc, 2009 FCA 97 at para
14; emphasis added)
[150] More recently, the Federal Court of Appeal confirmed this same
interpretation of AZT in Novopharm Ltd v Eli Lilly and Co, 2011
FCA 220 at paras 47-51 – though that affirmation was phrased in terms of
judicial comity as opposed to a full consideration of the issue (at para 50).
[151] Justice Binnie’s comments in AZT do not support an enhanced
disclosure requirement in all cases of sound prediction. I say this for two
reasons. First, it is clear from Justice Binnie’s reasoning that “this sort of
case” is a subset of sound prediction cases and not a reference to all sound
prediction cases. As he writes, “[i]n this sort of case,
the sound prediction is to some extent the quid pro quo the applicant offers in
exchange for the patent monopoly” (at para 70).
By implication, there are other “sort[s] of case[s]”
where the sound prediction is not the quid pro quo offered by the applicant.
[152] Second, and even more critically, limiting “this sort of case” to
new use cases, rather than sound prediction cases generally, is consistent with
the rationale provided by Justice Binnie. In a new use case (which AZT was),
there may be an enhanced disclosure requirement because utility is the only
thing being offered in exchange for the patent monopoly since the compound
itself was previously disclosed. Theoretically, without such an enhanced
disclosure requirement in new use cases, a new use patent could consist of a
single sentence alleging a new use and a reference to a prior patent disclosing
the compound to which the use attaches. None of the research or studies
supporting that new use would have to be disclosed. While new uses can be of
tremendous importance (see AZT), such seemingly sparse patents would
fairly raise concerns for the court when evaluating the bargain between
innovators and the public. That Justice Binnie was emphasizing new use cases
and not sound prediction cases in general is further supported by his earlier
comments in AZT at paragraph 56 where he expressly described the “new
use” as the “gravamen” (i.e. the essence or the quid
pro quo) of the invention in that case.
[153] As I noted earlier, this reading of AZT is supported by
subsequent appellate authorities from the Federal Court of Appeal and the
Supreme Court. In Sanofi-Aventis Plavix, Justice Gauthier observed:
In contradistinction with the situation in AZT,
where the invention claimed was the new use/utility and thus the quid pro quo for the grant of the monopoly
was a full disclosure in respect of such utility, the public here received all
the information necessary to make and use clopidogrel. (at para 135)
[154] Further, in Teva sildenafil, at para 37, Justice Lebel makes
specific reference to the purported heightened disclosure requirement in sound
prediction cases. In the course of rejecting such a requirement based on his
own interpretation of AZT, he writes:
[A]ll that is required to meet the utility
requirement in s. 2 is that the invention described in the patent do what the
patent says it will do, that is, that the promise of the invention be fulfilled
[…].
That the invention must be useful as of the
date of the claim or as of the time of filing is consistent with this Court's
comments in AZT, at para. 56:
Where the new use is the gravamen of the invention, the utility
required for patentability (s. 2) must, as of the priority date, either be
demonstrated or be a sound prediction based on the information and expertise
then available. If a patent sought to be supported on the basis of sound
prediction is subsequently challenged, the challenge will succeed if ... the
prediction at the date of application was not sound, or, irrespective of the
soundness of the prediction, “"[t]here is evidence of lack of utility in respect of some of
the area covered.”[Italics in original; underlining added.]
Nothing in this passage suggests that utility
is a disclosure requirement; all it says is that “the utility required for
patentability (s. 2) must, as of the priority date, either be demonstrated or
be a sound prediction”. Utility can be demonstrated by, for example, conducting tests, but
this does not mean that there is a separate requirement for the disclosure of
utility. In fact, there is no requirement whatsoever in s. 27(3) to disclose
the utility of the invention: see, e.g., Consolboard, at p. 521, per
Dickson J.: “I am further of the opinion that s. 36(1) [now s. 27(3)] does not
impose upon a patentee the obligation of establishing the utility of the
invention”. (At paras 38-40)
[155] Notably, Justices Lebel and Rothstein were part of both unanimous
decisions of the Supreme Court of Canada in Teva sildenafil and AZT.
This supports the interpretation of AZT as not creating an enhanced
disclosure requirement in all cases of sound prediction.
[156] Admittedly, Justice Lebel’s remarks in Teva sildenafil were obiter dicta because
sound prediction was not “the main issue” on
appeal (at para 36) and because “in any event, Pfizer disclosed the utility of sildenafil”
(at para 41). Still, the Supreme Court’s view that this secondary topic
deserved such explicit treatment places this commentary within the “wider circle of analysis which is obviously intended for
guidance and which should be accepted as authoritative,” or, at the very
least, makes the remarks “commentary, examples or
exposition that are intended to be helpful and may be found to be persuasive”
(R v Henry, 2005 SCC 76 at para 57, [2005] 3 S.C.R. 609).
[157] Since Teva sildenafil, the Federal Court of Appeal has made
no binding remarks regarding the issue of disclosure in the context of sound
prediction. The only Federal Court of Appeal decision to discuss this issue is
Eurocopter. However, the discussion of disclosing utility is obiter dicta (see Eurocopter,
at para 159 and the dismissal of a motion to reconsider Eurocopter, 2013
FCA 261 at para 18). Further, the discussion of disclosing utility in Eurocopter
falls short of expressly reaffirming a disclosure requirement of utility in
sound prediction cases. The Court of Appeal observed that “where the factual basis is reliant on data which does not form
part of the common general knowledge, then disclosure in the specification may
indeed be required to support a sound prediction” (at para 153; emphasis
added).
[158] I am compelled to follow the Supreme Court’s remarks in Teva sildenafil
and the interpretation of AZT endorsed by Justice Gauthier in Sanofi-Aventis
Plavix. As a final note, I would also add that Professor Siebrasse’s
remarks on this very question provide further support to this interpretation.
In his paper, “Must the Factual Basis for Sound
Prediction be Disclosed in the Patent?” (2012) 28:1 CIPR 39, Professor
Siebrasse concluded that:
the requirement that the factual basis for a
sound prediction of utility must be disclosed in the patent itself, is unsound
in both law and policy. There is no basis in the text of the Patent Act, in
legal principle, or in practice, for a distinction between utility based on
sound prediction and demonstrated utility.
[159] Having reviewed Professor Siebrasse’s article, I generally agree
with his observations, and echo Justice Gauthier’s view in Sanofi-Aventis
Plavix, at para 132, that the article identifies negative policy
consequences of an enhanced disclosure requirement for sound prediction.
[160] In conclusion, I am of the view that there is no enhanced disclosure
requirement in all sound prediction cases. Utility and disclosure are treated
separately under the Patent Act, and consequently, should be treated
separately in the jurisprudence as well.
[161] I will now consider the legal tests related to demonstration and
sound prediction of utility for each of the promises made by the ‘653 patent
without regard to the disclosure of utility. I will, however, for completeness
of the evidentiary record, note which sound predictions depend on undisclosed
studies.
(c)
Application of the Law on Demonstration and
Sound Prediction of Utility to the Promises of the ‘653 Patent
(i)
Demonstration and Sound Prediction of Use as a
Proton Pump Inhibitor
[162]
Both experts agree that the claimed compounds
would be useful as proton pump inhibitors.
[163] For his part, Dr. Tracy opined that the use of the claimed compounds
as a proton pump inhibitor was either demonstrated or soundly predicted (this
aspect of his evidence was unclear) based on the Erlandsson paper, which taught
that the activities of the omeprazole enantiomers were the same. Omeprazole
itself had already been extensively studied and proven to be useful as a PPI.
It follows that Erlandsson’s conclusion regarding the omeprazole enantiomers
having identical activity either demonstrated or soundly predicted the usefulness
of the enantiomers as PPI’s as well.
[164] Dr. Meyer, while less certain than Dr. Tracy in this regard,
nonetheless also conceded that the use of the claimed compounds as a proton
pump inhibitor was soundly predicted based on omeprazole’s known mechanism of
action. Through omeprazole’s mechanism of action, its enantiomers are
converted into an achiral sulfenamide – the inhibitory active agent in the
parietal cell. As explained earlier, chirality is the primary attribute
distinguishing the enantiomers of omeprazole. Consequently, the transfer into
an achiral sulfenamide eliminates that difference, leaving the
enantiomers with predictably equal effect.
[165] In my view, the knowledge that omeprazole is itself useful as a PPI,
and that its enantiomers have identical activity, supports a sound prediction
of utility. The use of the enantiomers as a PPI was, however, not directly
demonstrated. Without studies directly on the enantiomers themselves such
evidence falls short of demonstration. Rather, such a use is extremely
likely given the relationship between the enantiomers and the racemate.
[166] With respect to this first promise, there is no issue regarding
proper disclosure. Dr. Tracy’s opinion, that the use of the claimed compounds
as PPI’s was soundly predicted, is based on the Erlandsson paper which is
expressly referred to in the ‘653 patent. Moreover, Dr. Meyer’s opinion is
based on omeprazole’s mechanism of action, which the parties agreed formed part
of the common general knowledge of the skilled person and consequently need not
be disclosed in the ‘653 patent (see Eurocopter, at para 153).
(ii)
Demonstration and Sound Prediction of Stability
Against Racemization
[167] As discussed earlier, this second promise is directed at two types
of racemization: stability against chemical-mediated and enzyme-mediated
racemization. The experts agreed that the promise of stability against
chemical-mediated racemization is limited to circumstances of neutral and basic
pH.
[168] Regarding chemical-mediated racemization, the experts agreed that
this promise was demonstrated as described on page 21 of the ‘653 patent with
respect to basic pH conditions. The experts also agreed that demonstration in
basic conditions in turn supported a sound prediction of stability against
chemical-mediated racemization in neutral conditions because basic conditions
were a “worst case” scenario for racemization.
[169] I agree. Chemical stability in a worst-case scenario (basic
conditions), while not a direct demonstration of chemical stability in a
less-than-worst-case scenario (neutral conditions), amounts to a sound
prediction of stability in neutral conditions. Therefore, the promised utility
of chemical-mediated racemization was satisfied by the ‘653 patent.
[170] Regarding enzyme-mediated racemization, neither expert claimed that
this promise was demonstrated. Such a demonstration would be unlikely, if not
impossible, because of the absence of studies in humans.
[171] Whether or not this promise was soundly predicted is more
controversial for two reasons: (1) because the experts’ opinions depend on the
AstraZeneca rat studies that were not disclosed in the ‘653 patent (therefore
putting the issue of “proper disclosure” in play), and (2) because the experts
were in less agreement on this issue.
[172] A preliminary critique of Dr. Tracy’s opinion on sound prediction
regarding enzymatic stability is that his written opinion is located within his
expert report in the section addressing the demonstration of utility rather
than the sound prediction of utility and thus was not the subject of a discrete
analysis. This critique was best displayed during his cross-examination:
Q. […] Paragraphs 111 through 113. This is
your analysis of Dr. Meyer's promise of in vivo racemization. Correct?
A. Correct.
Q. Now, you start paragraph 111 saying that you
agree with Dr. Meyer […] that the in vivo racemization study in the rat determined that the enantiomers of
omeprazole undergo minimal racemization in vivo in the rat, and you say you agree with that conclusion.
Q. Is the rest of that which follows here an
analysis of sound prediction or of demonstration?
A. Sound prediction.
Q. So it's an analysis of sound prediction,
notwithstanding that it's in your demonstration section?
A. Correct.
Q. Where in paragraphs 111 through 114, or any
paragraphs that come before, do you set out what your understanding is of a
sound prediction?
A. I don't.
Q. And where in paragraphs 111 through 114 do
you say that the analysis of in vivo racemization was based upon was to answer the question of
prediction, not demonstration?
A. I do not say that.
(Tracy Cross, Trial Transcript, Vol 15, p 2461
at line 28 – p 2462 at line 28).
[173] A later section in Dr. Tracy’s report which is directed at sound
prediction (beginning at para 178 of his report) only addresses the promise of
an improved therapeutic profile because, in his view, stability against
chemical racemization (Dr. Tracy’s truncated promise) was demonstrated. As a
consequence, some doubt is cast on the extent to which Dr. Tracy turned his
mind to the substance of a “sound prediction” in the context of stability
against enzyme-mediated racemization. To be clear, this is not, in itself,
fatal to AstraZeneca’s argument. If the content of Dr. Tracy’s opinion,
regardless of the heading it is under, amounts to a sound prediction, then it
is still evidence that tends to prove a sound prediction.
[174] The above issue aside, the evidence of Dr. Tracy and Dr. Meyer
supports a sound prediction of stability against enzyme-mediated racemization.
[175] A sound prediction requires (1) a “factual
basis for a prediction” and (2) “an articulable
and sound line of reasoning” from that prediction to an inference of the
desired result (Wellcome, at para 70).
[176] The “factual basis for the prediction” is
not in dispute – both experts provided ample evidence regarding the AstraZeneca
studies relevant to a sound prediction of stability against racemization.
[177] Thus, the key issue is whether or not those studies support an “articulable
and sound” line of reasoning. I will first expand
on how the case law describes a sound prediction, and then apply that standard
to the evidence in this case.
[178] In AZT, at para 71, the Supreme Court of Canada made clear that a sound prediction is a question of fact that is case specific:
It bears repetition that the soundness (or otherwise)
of the prediction is a question of fact. Evidence must be led about what was
known or not known at the priority date, as was done here. Each case will turn
on the particularities of the discipline to which it relates.
[179]
Second, the Supreme Court was clear about
a lower limit that falls below what is necessary for a “sound
prediction”:
There is no doubt that care must be taken that
the doctrine is not abused, and that sound prediction is not diluted to include
a lucky guess or mere speculation. The public is entitled to obtain a
solid teaching in exchange for the patent rights. (AZT, at para 69;
emphasis added).
[180] Third, the Supreme Court, while outlining the history of the
jurisprudence regarding sound prediction, tacitly approved of Pigeon J’s
phrasing in Monsanto Co v Commissioner of Patents, [1979] 2 S.C.R. 1108,
which defined an upper limit that exceeds the requirements for a “sound
prediction”:
I have quoted again the passage quoted by the
[Patent Appeal] Board because I consider the last sentence of the paragraph of
some importance as it does clearly indicate what is meant by a “sound
prediction.” It cannot mean a certainty since it does not exclude all
risk that some of the area covered may prove devoid of utility. (AZT,
at para 62 citing Monsanto, at 1117; emphasis in original).
[181] Thus, a sound and articulable line of reasoning falls somewhere
above a “lucky guess” or “mere speculation” and somewhere below “a certainty.”
This characterization of the upper and lower limits on a sound prediction has recently
been applied by the Federal Court of Appeal (see e.g. Apotex Xalatan, at
para 33). To be clear, though, these statements do not establish the precise
boundaries of a sound prediction. In this case, having passed the first hurdle
of mere speculation, what threshold must AstraZeneca actually meet to satisfy a
sound prediction?
[182] The most recent Federal Court of Appeal decision to provide
precision to the meaning of an articulable and sound line of reasoning is Novopharm Zyprexa.
In that decision, at paragraph 85, the Federal Court of Appeal interpreted AZT
as providing that a sound prediction “requires a prima facie
reasonable inference of utility.” The situating
of a prima facie reasonable inference between mere speculation and certainty provides
the clearest guidance to the proper approach to be taken in this case.
[183] The evidence of Dr. Tracy and Dr. Meyer supports a prima facie reasonable
inference from the AstraZeneca studies that the enantiomers of omeprazole would
be stable against enzymatic racemization in humans.
[184] In his expert report, Dr. Tracy opined that the in vivo racemization
studies in rats demonstrated “minimal racemization” of
the omeprazole enantiomers in rats. Further, he stated that “the physiological pH conditions in rats do not materially
differ from humans” and that “enzyme mediated
racemization/interconversion was typically more extensive in rats than humans.”
On that basis, Dr. Tracy concluded that “the expectation
would be that there would be even less enzyme-mediated interconversion/racemization
in humans.” In my view, this amounts to a prima
facie reasonable inference.
Testing showed minimal racemization in rats, a species that is similar in this
regard to humans, and which is typically known, if anything, to experience more
enzyme-mediated racemization than humans. Such reasoning is sound. Moreover,
such reasoning, which is based on evidence, knowledge, and reason, is well
above “mere speculation.”
[185] Apotex was critical of this opinion and asserted that Dr. Tracy
equated sound prediction with an “expectation.” While it is true that Dr.
Tracy used the word “expectation” in his explanation above, isolating that word
from the rest of his opinion would be an incomplete analysis. His
“expectation” that the enantiomers of omeprazole would be stable against
racemization was reasonably inferred from the factual basis of the rat studies
and his scientific knowledge of physiology. As a consequence, regardless of
whether he labelled his perspective as an “expectation,” or a “prediction,” or
a “sound prediction,” it amounted to a sound prediction.
[186] It is critical when assessing expert evidence, particularly in the
context of patent litigation, to not be beholden to the specific wording put
forth by experts. It is all too easy for an expert to claim that they “soundly
predict” a specific use, or that an invention was “obvious,” in an attempt to
strengthen their evidence before a judge who must apply legal tests which use
that same terminology. In that same vein, the court should not weigh too
heavily when an expert fails to use the precise wording from the jurisprudence
in their opinion. Rather, the court must look beyond this language to the
substance of the expert’s evidence to form its own opinion on these issues.
[187] Dr. Meyer’s evidence corroborates this sound prediction. During
examination-in-chief, Dr. Meyer accepted that enzyme-mediated racemization in
humans was “relatively unlikely.” Furthermore,
while Dr. Meyer insisted that the skilled pharmacologist would demand human
studies for a prediction, his explanation makes clear that the standard he
applied for a sound prediction was too high:
Q. Why at the level of the skilled
pharmacologist would he want to see it in humans to make the prediction?
A. To be certain. I think to be absolutely
certain the study has to be done in humans.
(Meyer Chief, Trial Transcript, Vol 8, p 1399
at lines 6 – 11; emphasis added).
[188] The jurisprudence is clear: “certainty,”
or “absolute certainty,” is above the threshold
required of a sound prediction. As a consequence, Dr. Meyer’s claim that a
“prediction” of use in humans requires studies in humans is not consistent with
the legal standard for a sound prediction. To be clear, I am not merely
rejecting Dr. Meyer’s evidence because of his choice of words (as I cautioned
against just above). Rather, these words accurately reflect the substance of
his overly stringent approach to sound prediction. In substance, Dr.
Meyer’s evidence consisted of the skilled person preferring clinical
studies in humans to studies in rats to support predictions of utility. That
view is entirely reasonable, if not clearly correct. However, the doctrine of
sound prediction does not require that the skilled person be able to make the best
possible prediction, only a sound one. Accordingly, Dr. Meyer applied too
stringent a standard in this regard.
[189] The second critique of Dr. Tracy’s opinion advanced by Dr. Meyer is
that Dr. Tracy’s opinion was “obviously the opinion of an
expert in this field who has done these studies himself, not as the normal
skilled person.” The only possible elaboration on this assertion
provided by Dr. Meyer consisted of pointing out how Dr. Tracy’s analysis was
not contained in the AstraZeneca research documents themselves. Presumably,
this was because the researchers at AstraZeneca may have lacked the expertise
that Dr. Tracy possesses, although this point was not established in evidence.
Without a more particular description of how Dr. Tracy specifically relied on
knowledge or skills beyond the capabilities of the skilled person, such a vague
criticism cannot succeed.
[190] In conclusion, the promise of stability against enzyme-mediated
racemization, like the promise of stability against chemical-mediated
racemization, is satisfied by the ‘653 patent through a sound prediction. I
note that this sound prediction is supported by studies that were not disclosed
in the ‘653 patent.
(iii)
Demonstration of the Full Promise of an Improved
Therapeutic Profile
[191]
Both experts agreed that the promise of an
improved therapeutic profile was not demonstrated by the filing date.
I agree. The insufficiency of the studies relied upon for only a
sound prediction of such utility, explained below, explains with even greater
force the insufficiency of those studies with respect to demonstration. As a
consequence, this promise is not demonstrated in the ‘653 patent
(iv)
Sound Prediction of the Full Promise of an
Improved Therapeutic Profile
[193] The evidence of Dr. Meyer (for Apotex), in opposition to a sound
prediction of an improved therapeutic profile, was most compelling. In
particular, Dr. Meyer’s evidence was more comprehensive and instructive than
the evidence of Dr. Tracy and ultimately persuaded me that the internal
AstraZeneca studies were insufficient to make a sound prediction across an entire
patient population of an improved therapeutic profile.
[194] Dr. Meyer testified that the limited data in the two studies – three
human livers and six plasma re-analyses – could not form the basis for a sound
prediction of utility that extrapolates across an entire patient population.
In response, Dr. Tracy asserted that such a limited sample was sufficient, but
was unable to point to any prior art that had relied on a similarly small
sample for the purpose of extrapolating across an entire population, either
with respect to the three livers or the six plasma re-analyses. Absent any
such examples, or other compelling evidence addressing the validity of such an
extrapolation, I am more persuaded by Dr. Meyer’s view that the studies in
question are insufficient to support a sound prediction of utility across an
entire patient population. In other words, this critique alone is sufficient
to refute a sound prediction of utility.
[195] In conclusion, the expert testimony with respect to utility supports
Apotex’s claim that the internal AstraZeneca studies neither demonstrate nor
soundly predict the full promise of an improved therapeutic profile. The ‘653
patent is therefore invalid because of a lack of utility.
(v)
Demonstration of the Truncated Promise of
Improved Properties
[196] Even if I accepted the truncated promise advanced by AstraZeneca,
there would still be no demonstration of utility. In the end, both experts
agreed that “improved pharmacokinetic and metabolic
properties” were not demonstrated by the internal AstraZeneca studies.
Dr. Meyer consistently held this view. While Dr. Tracy initially opined
otherwise in his expert report, on cross-examination, he ultimately conceded
that the studies did not demonstrate improved pharmacokinetic and metabolic
properties.
[197] In closing argument, AstraZeneca attempted to rescue these
admissions from Dr. Tracy by advancing that, while the studies independently
fail to demonstrate utility, they succeed in demonstrating utility when read
together. I cannot accept this argument for two reasons. First, Dr. Tracy
conceded that the plasma re-analysis work was not a demonstration, but was a
prediction. Second, no explanation was provided for how the gaps in the
microsomal studies were filled by the plasma reanalysis studies or vice versa.
As Apotex argued in closing:
What Astra says is this. […] it acknowledges Tracy's admission that the human liver microsomal studies did not demonstrate the truncated
promise. But, it says, this admission is of no moment because Dr. Tracy relied
upon the combined results of the microsomal studies and the plasma reanalysis
in providing a demonstration, an opinion of demonstration.
Respectfully, this is not correct. Dr. Tracy
admitted that the plasma reanalysis work was not a demonstration at all, but
it was a prediction. The result is that each of the studies, in my
submission, even if both relied upon, failed to provide a demonstration of
the truncated promise. At no point did Dr. Tracy provide an opinion that
the two studies only when taken together provide a demonstration.
(Trial Transcript, Vol 31, pg 4945 at lines 10 – 27; emphasis added).
[198] I agree with these submissions. Bare assertions about how two
studies when combined demonstrate utility without any elaboration on how those
studies complement one another in that regard is unconvincing. Thus, both
experts were of the view, and I am as well, that the AstraZeneca studies do not
demonstrate the truncated promise of improved pharmacokinetic and metabolic
properties.
(vi)
Sound Prediction of the Truncated Promise of
Improved Properties
[199] Two questions underlie the sound prediction of the truncated promise
of improved properties: (1) what amounts to “improved
pharmacokinetic and metabolic properties?” and (2) were such properties
soundly predicted?
[200] First, the interpretation of improved properties. I did not address
this interpretation when dealing with demonstration above because both experts
agreed, independent of that interpretation, that no such properties could be
demonstrated. However, greater detail on the meaning of “improved
pharmacokinetic and metabolic properties” at this stage is required
because both its meaning and its sound prediction are in dispute between the
experts.
[201] Though I have typically used “esomeprazole” to mean the (-)
enantiomer of omeprazole in this judgment, for the purpose of clarity in the
ensuing discussion, I will use (-)-omeprazole and (+)-omeprazole since both are
implicated.
[202] Dr. Tracy interpreted “improved
pharmacokinetic and metabolic properties” as (-)-omeprazole undergoing a
slower metabolism and exhibiting a greater area under the curve (AUC) than
(+)-omeprazole and the racemate (omeprazole). In other words, Dr. Tracy viewed
only (-)-omeprazole, rather than both enantiomers, as the improved compound.
[203] However, Dr. Tracy’s evidence with respect to sound prediction
should be given less weight because of his understanding of the promise of
improved properties.
[204] The first problem with Dr. Tracy’s interpretation of the promise of
improved properties is that it does not reflect the specification of the ‘653
patent. In particular, his interpretation incorrectly limits the promise to an
improvement of (-)-omeprazole over (+)-omeprazole and the racemate, rather than
an improvement of both enantiomers over the racemate.
[205] Dr. Tracy’s basis for limiting the promise to (-)-omeprazole stems
from it being the only enantiomer referred to in the claims:
[“The present invention provides such compounds, which are novel salts
of single enantiomers of omeprazole”] doesn't specify which enantiomer
of omeprazole. It simply says, “Single enantiomers of omeprazole,” but if one were to look at the
claims later in the patent, the claims describe minus omeprazole which is the
[(-)-enantiomer] in this case.
I believe the skilled person would take from
that, then, that the invention is providing compounds being [(-)–omeprazole],
the [(-)-omeprazole] enantiomer, which would have those improved
pharmacokinetic and metabolic properties. (Tracy Chief, Trial Transcript, Vol
15, p 2344 at lines 17-27)
[206] However, it would only be reasonable to limit the entire
specification with reference to an implication from the claims that the patent
promises improved properties for (-)-omeprazole alone if there was ambiguity in
the specification about what the patent was promising. That is not the case
here. Viewed in its entirety, the ‘653 patent is clearly directed at both
enantiomers being an improvement over the racemate. As a consequence, there is
no need to “read between the lines” and inordinately weigh the significance of
the claims to discern the ‘653 patent’s purportedly hidden (and narrower)
promise. Beyond this assertion, Dr. Tracy provided no evidence in support of
his view that an implication from the claims should trump the consistent
language throughout the specification.
[207] There are multiple reasons for interpreting the ‘653 patent as
promising improved properties for both enantiomers of omeprazole:
a.
the ‘653 patent states that the invention,
without qualification, provides for “novel salts of the
single enantiomers of omeprazole” (‘653 Patent, p 1 at lines
21-22; emphasis added);
b.
the ‘653 patent makes reference throughout to
both the (+) and the (-) enantiomers (see e.g. ‘653 Patent, p 2 at lines 17-22;
p 3 at lines 4-30);
c.
the examples in the patent provide methods for
the synthesis of both enantiomers (‘653 Patent, pp 10-15);
d.
the stability studies in the patent are
performed on both enantiomers (‘653 Patent, p 21); and
e. there is no mention in the patent of (-)-omeprazole having improved
properties over (+)-omeprazole.
[208] Thus, Dr. Tracy’s interpretation that the promise only applies to
the (-) enantiomer of omeprazole and its improvements cannot be supported by
the specification when read as a whole, as is required when interpreting the
promise of the patent.
[209] A second problem with Dr. Tracy’s interpretation of the promise is
that the properties he advances may not actually be an improvement at all. Dr.
Tracy argues that the improved properties are slower metabolism and an
increased AUC. Keeping in mind his view that reduced interindividual
variability is merely a goal (and hence not promised), the improved properties
he advances must still, at the very least, be consistent with such a goal.
Yet, by his own admission, the improved properties he advances (increased AUC)
would result in increased interindividual variation amongst slow
metabolizers:
Q. In fact, if you were to increase
bioavailability in slow metabolizers, all other things being equal, that would
lead to an increase in interindividual variation between rapid metabolizers and
slow metabolizers?
A. If you had higher areas under the curve for
poor metabolizers, then it would as compared to rapid metabolizers there would
be increased difference between the two.
Q. And, therefore, an increase in variation across
the patient population?
A. Comparing the two groups.
(Tracy Cross, Trial Transcript, Vol 16, p 2504
at lines 17-28).
[210] Furthermore, Dr. Tracy conceded that there were other possible
interpretations of improved properties (see: Tracy Cross, Trial Transcript, Vol
16, p 2505 at line 1 – pg 2507 at line 18). Such interpretations, which are
consistent with the expectation of reduced interindividual variation, ought to
be preferred.
[211] To the extent that Dr. Tracy’s interpretation of the promise
differed from what was actually promised, his testimony with respect to the
demonstration and sound prediction of these properties is of little weight.
That combined with Dr. Meyer’s evidence about the maximal prediction that could
be supported by limitations in the AstraZeneca studies precludes a sound
prediction of improved properties.
[212] For his part, Dr. Meyer did not advance an alternative
interpretation of improved pharmacokinetic and metabolic properties (recall
that, in his view, such properties were expressly qualified in the patent as
reduced interindividual variation). Rather, he opined that, at most, the
AstraZeneca studies could form the basis for a prediction of (-)-omeprazole
having slower metabolism than (+)-omeprazole:
Q. What is your opinion as to whether there is
a sound basis to predict the Tracy promise, if I can use that, not an improved
therapy profile but rather an improved pharmacokinetic or metabolic property
relative to omeprazole? […]
A. This is a question I really kind of
struggled with because I said myself what if you take the whole papers in
the AstraZeneca documents what is the minimal prediction you can make from
these results? Can you actually make a minimal prediction? And I think
the minimal prediction there would be that there probably will be some
differences in metabolic rate if this is an advantage.
If this is improved, we don't know either, but the
prediction would be there is slower metabolism of [(-)-omeprazole] than
[(+)–omeprazole] to some degree in vivo. Again, it is not clear if this
would result in an improved therapeutic profile. I think that prediction could
be made. (Trial Transcript, Vol 8, p 1441 at line 25 – p 1442 at line 20;
emphasis added)
[213] This minimal prediction, which mirrors Dr. Tracy’s misinterpretation
of the promise (namely, improvements in (-)-omeprazole over (+)-omeprazole),
fails to accord with the promise in the ‘653 patent (namely, improvements of
both enantiomers over the racemate) and therefore displays how the AstraZeneca
studies, regardless of the interpretation of improved properties, cannot form a
factual basis in support of a sound prediction of the improved properties
promised in the ‘653 patent.
E.
Summary of Utility Analysis
[214] In my view, the promises of the ‘653 patent were; (1) use as a PPI;
and (2) an improved therapeutic profile such as a lower degree of
interindividual variation.
[215] The demonstration and sound prediction of these promises may be
summarized as follows: promise (1) was soundly predicted and promise (2) was
neither demonstrated nor soundly predicted. As a consequence, the ‘653 patent
is invalid because of inutility, namely, the lack of demonstration or sound
prediction with respect to promise (2).
[216] Again, although not necessary to do so, to ensure completeness of
the record, I addressed the following promises as well:
- stability against chemical-mediated racemization;
- stability against enzyme-mediated racemization; and
- improved pharmacokinetic and metabolic properties.
[217] With respect to the demonstration and sound prediction of these
additional promises, I concluded as follows: promise (3) was soundly predicted,
promise (4) was soundly predicted, though based on a study that was not
disclosed in the ‘653 patent, and promise (5) was neither demonstrated nor
soundly predicted.
[218] Inutility is itself fatal to the validity of the ‘653 patent. I
will nevertheless for completeness of the record, consider the other grounds of
invalidity.
V.
The Differences between Novelty and Obviousness
[219] There are two key differences between novelty and obviousness: (1) the
body of prior art that they consider and (2) the threshold of ingenuity (in the
case of obviousness) or effort (in the case of novelty) that must be met when
moving from that prior art to the patent in question (see: Rothmans, Benson & Hedges Inc v Imperial Tobacco Ltd (1993), 47 CPR (3d) 188 at 197-99).
[220] With respect to the body of prior art considered, novelty considers
whether a single prior art reference makes the patent old whereas obviousness
considers whether the state of the art (i.e. multiple prior art references)
makes the patent obvious. Put differently, novelty asks if the invention was
already discovered by a single prior art reference whereas obviousness asks if
the invention would have been self-evident to discover given the state of the art:
Merck finasteride, at paras 181-82.
[221] With respect to the threshold that must be met to move from this
prior art to the patent in question, novelty considers a lower threshold
relating to effort: whether or not a single prior art reference teaches the patent
in question without “undue burden” (Apotex Inc
v Sanofi-Synthelabo Canada Inc, 2008 SCC 61 at para 33, [2008] 3 S.C.R. 265 [Sanofi-Synthelabo
Plavix]). In contrast, obviousness considers
a higher threshold relating to ingenuity: whether or not the state of the art
teaches the patent in question without the need for an “inventive step” or
“degree of invention” (Sanofi-Synthelabo Plavix, at paras 33 and 67).
Requiring an inventive step to get from a single prior art reference to the
impugned patent is determinative of the impugned patent being novel with
respect to that prior art reference. In other words, inventive steps exceed
the threshold of undue burden (Sanofi-Synthelabo Plavix, at paras 33).
[222] With these key differences established, I turn to an analysis of
novelty and obviousness. However, as a preliminary observation I note that the
science involved was, at the risk of understatement, complex. Both parties
retained experts fit to the challenge. In particular, Dr. Stephen G. Davies
(Chairman of the Department of Chemistry at Oxford), Dr. Eric N. Jacobsen
(Chair of the Department of Chemistry & Chemical Biology at Harvard), and
Dr. Rick L. Danheiser (AC Cope Professor of Chemistry and former Associate and
Acting Department Head of Chemistry at MIT) are pre-eminent experts, holding
senior positions at prestigious academic institutions. They are all
recognized globally for their research and teaching, their publication records,
and the role they play on editorial boards of leading scientific journals. Each
has received awards recognizing their contributions in the course of shaping
aspects of modern chemistry. They provided the court with a wealth of
knowledge underlying the dispute, and, suffice to say, the amount of
instruction these experts provided to the court was likely equivalent to an
undergraduate course in chemistry, if not the whole degree.
VI.
Novelty
[223] Apotex asserts that the ‘653 patent is invalid because what the
patent claims was not new, or rephrased in positive terms, because it was
“anticipated.” Novelty is a requirement for a valid patent under sections 2
and 28.2 of the Patent Act. In essence, an alleged patent is
anticipated if the skilled person, before the patent claim date (May 28, 1993),
and with reference to a single prior art reference, could have performed the
patent without “undue burden.”
[224] The Supreme Court of Canada established a refined two-part test for
anticipation in Sanofi-Synthelabo Plavix, at paras 30-33. For a patent
to be anticipated there must be, from the perspective of the skilled person:
(1) prior disclosure and (2) enablement from that prior disclosure.
A.
Prior Disclosure
(1)
The Law of Prior Disclosure
[225] To satisfy the first part of the novelty test – prior disclosure –
Apotex must demonstrate that there is at least one prior art reference which,
on its own, discloses subject matter that would, if performed, result in
infringement of the ‘653 patent: Sanofi-Synthelabo Plavix, at paras 25
and 28. Put differently: what infringes
later, anticipates earlier.
[226] The scope of prior disclosure is governed by section 28.2 of the Patent
Act which sets out the relevant dates for disclosure depending on its
source. In this case, Apotex did not advance any evidence of disclosure
originating directly or indirectly from the applicant (section 28.2(1)(a))
or originating from a third-party’s pending patent application (sections
28.2(1)(c) and (d)). As a consequence, Apotex can only prove
prior disclosure if it can show that the subject-matter of the ‘653 patent was
made “available to the public before the claim date”
by any person other than AstraZeneca (section 28.2(1)(b)).
[227] The only prior disclosure alleged by Apotex with regard to novelty
is DE 455, which was filed on November 8, 1990 and published on May 14, 1992.
DE 455 relates to a method for separating the enantiomers of a broad class of
compounds known as pyridylmethylsulphinyl-1H-benzimidazoles. The method uses a
strategy of preparing diastereomers by introducing another chiral molecule, a
chiral auxiliary. Unlike the enantiomers, the diastereomers have different
physical properties which may allow separation. The method requires several
steps: preparation of diastereomers by reacting a chiral auxiliary with a salt
of a racemic substituted benzimidazole; chromatography to separate the mixture
of diastereomers; crystallization to purify the diastereomer; hydrolysis in
strong acid to remove the chiral auxiliary group; and neutralization of the
hydrolysate using strong base. The ‘653 makes the binding admission that DE
455 may be used to separate the enantiomers of omeprazole in a preparative
scale (Shire Biochem Inc v Canada (Minister of Health), 2008 FC 538 at
para 24).
[228] I conclude that DE 455 does not disclose the subject matter of the
‘653 patent because it does not contain an essential element of the ‘653
patent, namely, a specified optical purity of 99.8% enantiomeric excess.
[229] There can only be prior disclosure if every essential element in the
‘653 is found within DE 455. A prior disclosure, upon performance, results in
infringement (Sanofi-Synthelabo Plavix, at para 25). Infringement, in
turn, requires infringement of each essential element (Free World Trust,
at paras 31 and 68). It therefore follows that the absence of a single
essential element of the ‘653 from the DE 455 disposes of the issue of prior
disclosure, and in turn, anticipation (see Elli Lilly and Co v Apotex,
2009 FC 991 at para 397).
(2)
No Prior Disclosure of 99.8% Enantiomeric Excess
[230]
The ‘653 patent specifies a significant
threshold of optical purity. As discussed earlier in the “Claims Construction”
analysis, claim 8, when read by the skilled person in light of the disclosure,
describes a compound having an optical purity of 99.8%ee or greater. Thus, an
essential element of the ‘653 patent is a compound having that high degree of
optical purity.
[231] The DE 455 does not specify a degree of optical purity with
sufficient detail to disclose the degree of purity specified in the ‘653.
Admittedly, the DE 455 describes a high degree of optical purity: “optically pure,” “configurationally
homogeneous,” “enantiomerically pure,”
etc. These descriptions are unqualified, and as a consequence, Apotex argued
that the DE 455, by providing for an “optically pure” compound, previously
disclosed the high degree of purity specified in the ‘653 patent. However,
prior disclosure must be assessed from the perspective of the skilled person.
From that perspective, the DE 455 did not disclose the same degree of purity as
the ‘653 patent.
[232] On this point, Dr. Armstrong (for AstraZeneca) convincingly
described how “optically pure” under the old nomenclature (during the time of
the DE 455) meant “as pure as we can make it… according
to the techniques at hand.” From that perspective, the DE 455’s
reference to optical purity cannot be read as a prior disclosure of 99.8%ee or
greater. Rather, such an unqualified reference to optical purity merely
discloses that, at the time, the DE 455 disclosed the greatest possible purity
attainable by its methods, not an objective benchmark.
[233] According to Dr. Jacobsen (for Apotex), “optically pure” lacks a
clear meaning to the skilled person. Nevertheless, Dr. Jacobsen claimed that a
sufficient degree of purity is disclosed because of the context of the DE 455.
In particular, Dr. Jacobsen claimed that the skilled person would understand
“optically pure” as meaning the same extent of purity as that found in the ‘653
patent because it would be desirable from a commercial perspective. This
interpretation is unconvincing. The mere fact that high optical purity is desirable
is not equivalent with an explicit benchmark of 99.8%ee.
[234] Furthermore, the evidence of Dr. Danheiser (for Apotex) is
particularly unconvincing with respect to the prior disclosure of sufficient
purity. He was instructed by Apotex to assume that “enantiomerically pure” is
equivalent to 99.8%ee. As a consequence, he appears to not have turned his
mind to the question of how the skilled person would understand DE 455’s
reference to “optically pure.”
[235] Finally, Dr. Davies (for AstraZeneca) claimed that “optically pure”
by convention suggests a meaning of 90%ee or greater. However, Dr. Davies also
argued that this suggestion, without the disclosure of specific methods or
evidence of purity, is without substance. With the significant amount of
expert critique in this trial of methods for assessing purity, I am sympathetic
to Dr. Davies’ hesitation to affirm a specific degree of purity that appears to
be unsubstantiated.
[236] Whether or not other essential elements of the ‘653 patent were
disclosed in the DE 455 was also in dispute (e.g. salts, the (-)-enantiomer of
omeprazole, the properties of the omeprazole enantiomers). However, I need not
go into detail about those purported essential elements because the element
that received the most thorough analysis and which had the most evidence raised
in its respect (99.8%ee optical purity) was not disclosed.
[237] A patent that provides for an “optically pure” substance, which, in
context, means 90%ee or greater, does not necessarily infringe a patent that
specifies a purity of 99.8%ee. Consequently, the prior disclosure criterion of
novelty is not met in this case.
B.
Enablement
(1)
The Law of Enablement
[238] While I have concluded that there is no prior disclosure, I will
briefly discuss the second part of the novelty test: enablement.
[239] Enablement is directed at whether the skilled person, with reference
to the single prior art reference, can “perform or make
the invention of the second patent without undue burden” (Sanofi-Synthelabo
Plavix, at para 33). In Sanofi-Synthelabo Plavix, the Supreme Court
does not provide a definition for “undue burden” and instead relies on a
non-exhaustive list of factors that “will apply in
accordance with the evidence in each case” (at para 37). In particular,
the Supreme Court summarizes those factors (which I paraphrase) as follows:
a. the prior patent as a whole including the specification and the
claims;
b. the skilled person’s common general knowledge; and
c. the nature of the invention, meaning, that what is considered to be “undue
burden” within a particular field depends on how common potentially burdensome
activities such as trials and experiments are in that field.
[240] The Supreme Court also referred to a fourth factor: the extent to
which obvious errors and omissions do not prevent enablement if “reasonable skill and knowledge in the art could readily
correct the error or find what was omitted” (at para 37). In my view,
this is more an application of the second factor above to the assessment of
undue burden than an independent factor, though it is helpful for it to be
specifically pointed out.
(2)
The Evidence Regarding Enablement
[241] The Court was presented with ample conflicting evidence on the issue
of enablement. In particular, both parties relied on their respective
successful and failed attempts at following DE 455 to reach the purity
described in the ‘653 patent. As these attempts applied the prior patent (DE
455) and purported to follow an approach reflective of the skilled person my
analysis of both parties testing evidence in addition to expert commentary on
that evidence considers the enablement factors described above.
[242] AstraZeneca provided fact evidence from Dr. Larsson, a former
employee of AstraZeneca. He described AstraZeneca’s unsuccessful efforts in
applying the DE 455 to produce sufficient quantities of the enantiomers of
omeprazole for testing in 1993. Apotex responded with a thorough critique of
both Dr. Larssons’ limited personal involvement with the experiments and the
apparent errors in AstraZeneca’s attempted application of the DE 455 based on
the laboratory notes of Mr. Niman. Mr. Niman, who performed the experiments,
was not called as a witness.
[243] Apotex’s evidence was similarly inconclusive. Apotex relied on the
fact evidence of Dr. Taylor, an esteemed chemist who possesses skills and knowledge
well in excess of the skilled person in this case. Further, AstraZeneca
provided a satisfactory critique of how Dr. Taylor went beyond the scope of DE
455 in his experiments.
[244] On the whole, I am unconvinced that the DE 455 enables the ‘653
patent.
(3)
No Enablement of 99.8% Enantiomeric Excess
[245] Neither party advanced compelling evidence regarding enablement.
Viewed in its entirety, the evidence regarding testing is “controversial and
inconclusive.” As a consequence, I assign it
little weight (Bristol-Myers Squibb Canada Co v Apotex Inc, 2009 FC 137
at paras 132-42). Further, as both parties’ arguments with respect to
enablement rested primarily on the testing, I conclude that Apotex has failed
to meet its burden with respect to enablement.
[246] Dr. Larsson (for AstraZeneca) described many struggles in performing
the DE 455. Ultimately, these tests were unable to obtain optically pure
enantiomers of omeprazole and encountered “major degradation.” That being
said, Apotex had a strong critique of Dr. Larsson’s evidence and of the testing
he described. First, Apotex pointed out many flaws in Dr. Larsson’s evidence,
namely:
a. The experiments were conducted by Mr. Niman, who was not called to
give evidence at trial.
b. Mr. Niman’s work began in 1992, prior to Dr. Larsson joining
AstraZeneca in March 1993.
c. Mr. Niman’s work was initially supervised and instructed by Dr.
Hjalmarsson, who himself was not called to give evidence at trial.
d. Dr. Larsson did not conduct the actual experiments.
e. Dr. Larsson says he observed some of the experiments, but could not
say which ones.
f. Dr. Larsson has not discussed Dr. Niman’s work with Dr. Niman since
1993.
g. Dr. Larsson could not explain whether or why certain steps were
performed.
h. Dr. Larsson was unable to confirm whether analyses were performed at
various stages.
[247] Second, Apotex convincingly argued that Mr Niman, based on the
evidence of his laboratory notebooks, did not faithfully follow the steps in DE
455 as a skilled chemist. In particular, Apotex observed that:
a. Mr. Niman at times failed to control the pH of the reactions and
failed to properly neutralize the solution after hydrolysis exposing it to
destructive acid pHs.
b. Mr. Niman failed to carry out the chromatography as a skilled
chemist, repeatedly beginning purifications on a chromatographic column and
leaving them overnight.
c. Mr. Niman, when unable to chromatographically purify omeprazole
using one solvent system, failed to try other solvents.
[248] In light of these critiques it is difficult to give much weight to
Dr. Larsson’s evidence or to consider that evidence compelling with respect to
anticipation.
[249] Unlike AstraZeneca, Apotex encountered little difficulty when
performing the DE 455. However, AstraZeneca’s critique of Apotex’s evidence is
similarly persuasive.
[250] First, in Apotex’s favour, I accept that Dr. Taylor, who organized
the Apotex testing of the DE 455, was insulated from bias or any pre-conception
as to the ultimate objective because of how Apotex approached his retainer. In
particular, I note that Dr. Taylor:
a. was not given the ‘653 patent;
b. was not told that a desired level of purity was desirable; and
c. would have had no way of knowing whether it favoured Apotex that the
examples worked (e.g. supporting an anticipation argument) or did not work
(e.g. supporting an inutility argument).
[251] Dr. Taylor is an exceptional chemist with skills well-beyond that of
the skilled person. While it is true that solely attacking his credentials
would amount to a mere ad hominem attack (as Apotex argued), the entire context of the Apotex testing
(including Dr. Taylor’s impressive credentials) undermines its credibility.
Apotex claims that the DE 455, with its “exquisite level
of detail,” would easily enable the skilled chemist to attain compounds
as pure as the ‘653 patent. Yet they enlisted a highly esteemed chemist, who
performed several ex parte tests of the DE 455 before performing those same tests in the
presence of representatives for AstraZeneca, and who was not instructed to
perform the tests as if he was a chemist in 1993.
[252] Mere speculation about how Dr. Taylor does not perfectly align with
the skilled person is insufficient to discredit his evidence. However, when
coupled with multiple variations that he used when performing DE 455, his
evidence loses weight. In particular, Dr. Taylor’s use of unusual solvent
ratios and cooling techniques, without convincing evidence as to how the
skilled person would have approached DE 455 similarly, and viewed in light of
Dr. Taylor’s exceptional expertise, leaves me with doubts about his success with
DE 455 in 2013 being reflective of the skilled person’s success with DE 455 in
1993. As a consequence, I am not convinced that the skilled person could
perform DE 455 and achieve the purity described in the ‘653 patent without
“undue burden.”
[253] The above flaws aside (which defeat Apotex’s claim of enablement
because Dr. Taylor’s approach was not reflective of the skilled person), I
accept Apotex’s argument that Dr. Taylor ultimately achieved the degree of
purity specified in the ‘653 patent.
[254] Dr. Taylor calculated the degree of purity in his tests by two
methods: High-performance Liquid Chromatography (HPLC) and Nuclear Magnetic
Resonance (NMR). Despite its efforts, AstraZeneca’s attempt at discrediting
the findings from these methods was unpersuasive.
[255] Regarding the HPLC testing, AstraZeneca argued that the calculations
may have been compromised by impurities. After each hydrolysis, the material
obtained was purple. Both parties agreed that this purple colour was an
impurity because the enantiomers of omeprazole are colourless. As a
consequence, the HPLC results were potentially unreliable. However, Apotex’s
response to this “purple impurity theory” is
compelling. More specifically, Apotex convincingly argued that this theory is
inordinately speculative and forces the Court to “postulate
the existence of a number of highly improbable facts,” namely:
a. The impurity is undetectable on NMR – Dr. Taylor’s NMR did not show
any impurity.
b. The impurity just happens to co-elute with the major enantiomer in
sufficient amount to affect the enantiomeric purity calculation.
c. The impurity just happens to co-elute at precisely the same point as
the major enantiomer such that the symmetry of the peak representing the major
enantiomer is not distorted.
d. The impurity has the same spectral properties as the major
enantiomer at two different wavelengths even though molecules that are
different colours (e.g. colourless omeprazole and the purple impurity) are
likely to have different spectra both in the colour spectrum and in the UV
region.
e. The impurity would have to generate a peak that is 20 times larger
than the peak associated with the major enantiomer in order to visually impact
the peak associated with the enantiomer in question.
[256] Regarding the NMR testing, there was a lively debate between Drs.
Jacobsen and Davies over the assignment of peaks in the calculation of the
diastereomeric ratio. In the end, however, the use of NMR as a tool for
monitoring the progress of a reaction, as opposed to calculating the purity of
the final product (as HPLC does), makes this debate largely inconsequential.
In light of the HPLC results, I conclude that Dr. Taylor was able to reach the
level of purity specified by the ‘653 patent.
[257] Finally, AstraZeneca’s last attempt at discrediting Dr. Taylor’s
approach was to critique that he went beyond the scope of DE 455 by making
salts. This critique belies two factors outlined in Sanofi-Synthelabo
Plavix, at para 37: the plain language of the DE 455 and how obvious errors
and omissions do not prevent enablement if “reasonable
skill and knowledge in the art could readily correct the error or find what was
omitted.” First, the DE 455 application states that salts of racemic
substituted benzimidazoles with bases can be made “in a
customary manner by reaction of the compounds with the appropriate hydroxides.”
In fact, Dr. Taylor made a sodium salt, and the DE 455 application specifically
mentions as base addition salts of the racemic benzimidazoles among others the
sodium, potassium, calcium, magnesium and titanium salts. Further, even if
this direction was not clear enough for the skilled person to follow, as Dr.
Jacobsen opined, the procedure described for the making of omeprazole salts in
European Patent Application 0 124 495 dated November 7, 1984 [the ‘495] could
be used, without modification, to obtain the salts of the enantiomers of
omeprazole. As Dr. Taylor independently located the ‘495 as part of the
testing he performed, reliance on its methods falls within the scope of the
diligent skilled chemist.
[258] In sum, the ‘653 patent was novel. The DE 455 did not disclose the
essential characteristic of a specified degree of high optical purity.
Moreover, Dr. Taylor’s performance of the DE 455 does not convince me that the
skilled person would have been enabled to reach the degree of purity specified
in the ‘653 patent by performing DE 455. As a consequence, Dr. Taylor’s
evidence regarding his testing was inconclusive and therefore incapable of
satisfying Apotex’s burden with respect to enablement.
VII.
Obviousness
[259]
Apotex’s final argument for the invalidity of
the ‘653 patent is that it is invalid because it was “obvious.” It is a
requirement for a valid patent under sections 2 and 28.3 of the Patent Act
that the subject of the patent not be obvious. In essence, an alleged patent
is obvious if, from the perspective of the skilled person as of the patent
claim date, no inventive step was required to get from the state of the art to
the inventive concept of the patent.
[260] The Supreme Court set out a four part test for obviousness in Sanofi-Synthelabo
Plavix, at para 67:
a. Identify the person skilled in the art and the relevant common
general knowledge;
b. Identify the inventive concept of the claim in question or, if that
cannot readily be done, construe it;
c. Identify what, if any, differences exist between the matter cited as
forming part of the "state of the art" and the inventive concept; and
d. Viewed without any knowledge of the alleged invention as claimed, do
those differences constitute steps that would have been obvious to the skilled
person or do they require a degree of invention?
[261] The first step of this test is not in dispute. The parties
substantially agreed on the characteristics of the skilled person, as set out
above. The skilled person of the ‘653 patent is a composite of:
•
An organic or medicinal chemist
•
A pharmacologist
•
A pharmaceutical formulator
•
A physician familiar with the pharmaceutical
treatment of excess gastric acid secretion and related diseases
[262] Further, the common general knowledge of that skilled person is
also, for the most part, undisputed between the parties (also set out above).
It includes:
•
Stereochemistry
•
The role of stereochemistry in drug action
•
Omeprazole, in particular
▪ its salts, and
▪ its mechanism of action
•
Esomeprazole and its salts
•
The use of salts to improve drugs
•
General resolution techniques, and
•
The motivation to purify enantiomers
[263] The remaining three steps in the Sanofi-Synthelabo Plavix four-step
obviousness test merit more elaborate consideration.
A.
Step 2: The Inventive Concept of the Claims in
Question
[264] In this case, the first disputed step of the Sanofi-Synthelabo
Plavix obviousness test is the identification of the inventive concept of
the claims in question.
(1)
The Meaning of the Inventive Concept
[265] The meaning of the “inventive concept” of a patent’s claims was the
subject of controversy in this case. The parties adopted conflicting
interpretations of the inventive concept. Whereas AstraZeneca argued for a
truncated promise (lowering its utility burden), it argued for an expansive inventive
concept (raising Apotex’s obviousness burden). Apotex, on the other hand,
argued the reverse: for a broader promise (raising AstraZeneca’s utility
burden) and a narrower inventive concept (lowering its own obviousness burden).
[266] The parties had conflicting views on the legal principles
underpinning the inventive concept as well. AstraZeneca, in its closing,
argued that the inventive concept, promise of the patent, and claims
construction, are “just one construction for all
purposes.” By contrast, Apotex argued that all three exercises are
distinct inquiries. Such a stark contrast in the basic legal framework
underlying key doctrines in patent law, between two highly sophisticated
litigants, is alarming to say the least.
[267] Courts have consistently held that the identification of the
inventive concept begins with the claims of the patent, and only looks to the
disclosure when necessary. In its leading authority on obviousness the Supreme
Court states that a court, at the second step of the obviousness inquiry, must “[i]dentify the inventive concept of the claim in question or
if that cannot readily be done, construe it” (Sanofi-Synthelabo
Plavix, at para 67).
[268] The Supreme Court goes on to say:
A bare chemical formula in a patent claim may
not be sufficient to determine its inventiveness. In such cases, I think it
must be acceptable to read the specification in the patent to determine the
inventive concept of the claims. Of course, it is not permissible to read the
specification in order to construe the claims more narrowly or widely than the
text will allow (at para 77).
[269] Thus, similar to claims construction, the identification of the
inventive concept begins with the claims, and only considers the remainder of
the patent (the disclosure) if necessary. Courts have consistently affirmed
this approach (see e.g. Abbvie Corp v Janssen Inc, 2014 FC 55 at para
123).
(2)
The Inventive Concept of the ‘653 Patent
[270] The ‘653 patent is not one that requires a consideration of the full
specification to identify the inventive concept. It does not merely claim a “bare chemical formula” as discussed in Sanofi-Synthelabo
Plavix. Rather, it claims a multitude of things (compounds, purities,
uses) none of which require a consideration of the disclosure to
understand their inventiveness. On that basis, Apotex’s interpretation of the
inventive concept, which flows unambiguously from the claims of the ‘653
patent, should be accepted over AstraZeneca’s interpretation of the inventive
concept, which unnecessarily imports aspects of the disclosure.
[271] Two inventive concepts were proposed by the parties in this case.
Apotex argued for a narrower inventive concept: “the
salts of esomeprazole, including the magnesium salt, having an ee of up to
≥99.8% ee.” By contrast, AstraZeneca argued for a broader
inventive concept: “an optically pure salt of
(-)-omeprazole as claimed, together with improved pharmacokinetic and metabolic
properties over omeprazole, and high stability to racemization in neutral and
basic pH.” Consequently, the inventive concept proposed by AstraZeneca
may be viewed as an extension of the inventive concept proposed by Apotex.
More specifically, the disagreement between the parties distils to whether or
not the inventive concept of the claims in the ‘653 includes
esomeprazole’s improved properties described in the disclosure of the
patent.
[272] There is, however, no need to look to the disclosure for improved
properties within the inventive concept of the ‘653 patent because a viable
inventive concept is present in the claims alone. Both parties focus their
analysis of the inventive concept on claims 7 and 8 of the ‘653, which read:
7. A compound according to any one of claims 1
to 6 having an optical purity of 98% or greater.
8. A compound according to any one of claims 1
to 6 having an optical purity of 99.8% or greater.
(emphasis added)
[273] AstraZeneca argues that these are merely “bare compound claims” that
necessitate a consultation of the disclosure to ascertain their inventive
concept. I disagree. Claims 7 and 8 refer not merely to a “bare” compound,
but its specific degree of optical purity. Notably, that degree of optical
purity, and whether or not an inventive step was necessary to achieve it, was
the primary focus of much of the expert testimony in this case. While it is
true that the inventive concept could be expanded by referring to the
properties of these compounds discussed in the disclosure, that would be an improper
approach. The Supreme Court in Sanofi-Synthelabo Plavix permits a
consideration of the disclosure for instances like “bare compound claims”
because they “may not be sufficient to determine”
the inventive concept, not because it would be possible to expand on a viable
inventive concept that has already been identified (at para 77; emphasis
added). Indeed, the Supreme Court notes that “it is not
permissible to read the specification in order to construe the claims more
narrowly or widely than the text will allow” (at para 77). Such an
improper interpretation is precisely what AstraZeneca advances in this case.
The ‘653 patent never claims the improved properties AstraZeneca seeks
to add to the inventive concept. Further, an inventive concept may be
identified in the claims alone with respect to novel compounds with a
previously unattainable degree of purity. As such, the inventive concept,
which is rooted in the claims of the patent, is found in claims 7 and 8 alone.
[274] Both claims identify a compound (esomeprazole) with a specific trait
(high purity). The inventive concept is therefore the compound and its
specific high degree of purity that was obtained. More specifically, the
inventive concept is the compound with the highest extent of purity claimed (claim
8); or, as Apotex argued: “the salts of esomeprazole,
including the magnesium salt, having an ee of up to ≥99.8% ee.”
[275] I need not look to other claims in the ‘653 patent because the
“claims in question” in this case (i.e. the claims to which the pleadings and
experts were directed with respect to obviousness) were only claims 7 and 8.
B.
Step 3: Differences Between the State of the Art
and the Inventive Concept
[276]
This takes us to the third step of the Sanofi-Synthelabo
Plavix obviousness test. Before discussing the differences between the
prior art and inventive concept, the time frame for prior art must be
established. Section 28.3 sets out the relevant dates for prior art depending
on its source. In this case, Apotex did not advance any evidence of prior art
originating directly or indirectly from the applicant (AstraZeneca). As a
consequence, the only prior art that is relevant must have been “made available
to the public” before the claim date by any person other than AstraZeneca (Patent
Act, s 28.3(b)).
[277] Apotex, given its views on anticipation, argues that there is no
difference between the state of the art and the inventive concept because, in
its view, the ‘653 patent was anticipated by the DE 455 German patent
application. I have already explained, above, how the DE 455 does not
anticipate the ‘653 because it lacks an essential element of the ‘653: a
specified optical purity of 99.8%ee.
[278] In the alternative, Apotex argues, with respect to obviousness, that
the only differences between the state of the art and the inventive concept are
the following:
a. a quantitative (e.g. 99.8%ee in the ‘653) rather than qualitative
(e.g. “optically pure” in DE 455) description of purity and
b. the listing of specific base addition salts (e.g. magnesium in the
‘653) rather than the general description of base addition salts (e.g. “salts
with bases” in DE 455).
[279] AstraZeneca does not dispute these differences between the state of
the art and the inventive concept. They are quite plainly differences between
the DE 455 and the ‘653. Instead, AstraZeneca adds three further differences:
a. the isolated (-) enantiomer of omeprazole,
b. in alkaline salt form, and
c. its properties.
[280] I will assess each of the proposed differences in turn, ultimately
concluding that all of the proposed differences between the state of the art
and the inventive concept are valid except the enantiomers in alkaline salt
form or their advantageous “properties.” I note, though, that compiling a list
of differences between the state of the art and an impugned patent does not
protect it from invalidity because those differences could have been obvious to
overcome.
[281] Attaining the specified optical purity of 99.8%ee is a difference
between the state of the art and the inventive concept. Though DE 455
describes its compounds as “optically pure” I have already described above how,
when read in context, that means approximately 90% pure. The only other piece
of prior art referred to on this point is the Erlandsson paper which describes
a method achieving 91.2%ee optical purity. As a consequence, the degree of
optical purity attained and specified in the ‘653 is a difference between the
state of the art and the inventive concept of the ‘653.
[282] The listing of specific base addition salts (such as magnesium) is
also a difference between the state of the art and the inventive concept. The
only piece of prior art referred to on this point is the DE 455 which more
generally refers to “salts with bases” as opposed to enumerating specific salts
as the ‘653 does in its claims.
[283] The isolated (-) enantiomer of omeprazole similarly differs from the
state of the art. The DE 455 only exemplifies the (+) enantiomer of
omeprazole. Moreover, Erlandsson only provides for a partial separation and
isolation of the omeprazole enantiomers.
[284] Finally, the two remaining differences between the state of the art
and the inventive concept proposed by AstraZeneca must be dismissed. Neither
the enantiomer in alkaline salt form nor its improved properties, such as
improved pharmacokinetic properties, is part of the inventive concept claimed
in the ‘653. Rather, they are both discussed in the disclosure outside of the
relevant claims (7 and 8) and do not form part of the inventive concept.
Accordingly, they cannot be a “difference” between the state of the art and the
inventive concept of the ‘653.
[285] In conclusion, the differences between the state of the art and the
inventive concept include the following:
a. A specified optical purity of 99.8%ee;
b. The listing of specific base addition salts; and
c. The isolated (-) enantiomer of omeprazole.
C.
Step 4: Inventive Step from State of the Art to
the Inventive Concept
[286] Finally, the fourth step of the Sanofi-Synthelabo Plavix obviousness
test considers whether overcoming the above differences (that is, getting from
the state of the art to the inventive concept) requires a “degree of invention”
or, in other words, an “inventive step.” I will first outline the proper
approach to the obviousness inquiry, and in particular, the “obvious to try”
inquiry, and then follow that approach in reaching the conclusion that the ‘653
patent was not obvious because an inventive step was required to achieve its
high optical purity.
(1)
The Law Governing Obviousness
(a)
The Merged Obviousness and Obvious to Try Inquiries
[287] As previously described, the general obviousness inquiry was
provided in Sanofi-Synthelabo Plavix. However, at the fourth step of
that general obviousness inquiry, an overlapping inquiry – the “obvious to try”
inquiry – may arise (see Sanofi-Synthelabo Plavix, at para 67). The
obviousness and obvious to try inquiries should not be conceived of as discrete
inquiries. Rather, the four-step obviousness test always governs the analysis,
with the notion of whether or not a particular experiment was “obvious to try”
comprising “only one factor to assist in the obviousness
inquiry” (Sanofi-Synthelabo Plavix, at para 64). In that sense,
they are overlapping inquiries, wherein the general obviousness test is always
applied and the obvious to try inquiry occasionally supplements the fourth step
of the general obviousness test, depending on whether or not the area of
science engaged by the patent in question is sufficiently experimental.
[288] More specifically, the obvious to try approach may be applied at the
fourth step of the general obviousness test in “areas of
endeavour where advances are often won by experimentation” (Sanofi-Synthelabo
Plavix, at para 68). In particular, the Supreme Court in Sanofi-Synthelabo
Plavix notes that “some inventions in the
pharmaceutical industry might warrant an ‘obvious to try’ test since there may
be many chemically similar structures that can elicit different biological
responses and offer the potential for significant therapeutic advances” (at
para 68). The ‘653 patent, which relates to optically pure enantiomers with
promised therapeutic advances (esomeprazole) in contrast with its predecessor,
a previously discovered and similarly structured racemate (omeprazole),
satisfies this criterion and ought to be evaluated pursuant to the obvious to
try approach. That being said, obvious to try is still merely a factor to be
considered. Accordingly, the factors outlined in the jurisprudence with
respect to both the general obviousness inquiry and the obvious to try inquiry
are relevant in this case.
(b)
Obvious to Try and Obviousness Factors
[289] I begin with factors related to the obvious to try inquiry. In Sanofi-Synthelabo
Plavix, the Supreme Court outlines a non-exhaustive list of primary and
secondary factors which may be considered when evaluating whether or not an
invention was obvious to try (at paras 69-71). The primary factors may be
summarized as follows:
a. Whether what was being tried obviously ought to work and
whether there are a finite number of identified predictable solutions known to
the skilled person.
b. The extent, nature, and amount of effort required to achieve
the invention and whether the experimentation amounts to trials that are
routine or trials that are prolonged and arduous.
c. Whether there is a motive provided in the prior art to find
the solution the patent addresses.
[290] Additionally, the Court notes “[a]nother
important factor” (arguably, a secondary factor), that being “the actual course of conduct which culminated in the making of
the invention,” in so far as it relates to how the skilled person would
have acted in light of the prior art – the proper focus of the obviousness
inquiry (at para 70).
[291] Next, the factors related to the general obviousness inquiry. In Janssen-Ortho
Inc v Novopharm Ltd, 2007 FCA 217 [Novopharm Levofloxacin], the
Federal Court of Appeal outlines a non-exhaustive list of principal and
secondary factors which may be considered when evaluating obviousness more
generally (at para 25). In particular, the court emphasizes that “this list is a useful tool, but no more” and that it
should not be “slavishly followed” (at para 27).
As a consequence, my analysis will be directed at those factors (whether
outlined in Novopharm Levofloxacin or not) which, based on the patent in
question and evidence presented, are relevant to the question of obviousness in
this case. Not surprisingly, there is significant overlap between the factors
discussed by the Supreme Court with respect to the obvious to try inquiry and
the factors discussed by the Federal Court of Appeal with respect to the
general obviousness inquiry since they are overlapping inquiries.
[292] The primary factors in the general obviousness inquiry from Novopharm
Levofloxacin may be summarized as follows. First, there are three primary
factors which I have already addressed above in the earlier stages of the Sanofi-Synthelabo
Plavix obviousness test:
a.
The invention
b.
The skilled
person, and
c.
The skilled
person’s common general knowledge
[293] Second, there are additional primary factors that are analyzed in
light of the above three factors, namely:
a. The climate in
the relevant field at the time the alleged invention was made.
b. The motivation in existence at the time of the alleged
invention to solve a recognized problem.
c. The time and effort involved in the invention
[294] Notably, and not surprisingly, these factors align significantly
with the obvious to try factors. Finally, there are also secondary factors to
consider, namely:
a. The commercial success of the invention, and
b. Meritorious awards received by the
invention.
(c)
The Threshold to be Met to Satisfy the Obvious
to Try and Obviousness Inquiries
[295] The above factors are considered with a view to the ultimate legal
question of the obviousness of the impugned patent.
[296] For an invention to be obvious to try, “there
must be evidence to convince a judge on a balance of probabilities that it was
more or less self-evident to try to obtain the invention. Mere possibility
that something might turn up is not enough” (Sanofi-Synthelabo Plavix,
at para 66).
[297] More broadly, for an invention to be obvious, it must require an
“inventive step” to get from the state of the art to the inventive concept,
meaning, similarly, that it must have been “obvious” or “very plain” to
discover the invention given the state of the art (Sanofi-Synthelabo Plavix¸
at para 65).
(2)
Application of the Law Governing Obviousness to
the ‘653 Patent
[298] As described above, there are three proposed differences between the
state of the art and the inventive concept that must be addressed, namely:
a. A specified optical purity of 99.8%ee;
b. The listing of specific base addition salts; and
c. The isolated (-) enantiomer of omeprazole.
[299] The second and third differences are easily characterized as obvious
extensions from the state of the art. As a consequence, I will briefly address
the non-inventive step that would have brought the skilled person from the
state of the art to those differences before analyzing the inventive step
required to reach a specified optical purity of 99.8%ee in greater detail.
(a)
The Non-Inventive Step from the State of the Art
to Specific Salts and the (-) Enantiomer of Omeprazole
[300] Both the listing of specific salts and the isolation of the (-)
enantiomer of omeprazole were obvious extensions on the state of the art, and
in particular, on the DE 455.
[301] First, the listing of specific base addition salts was an obvious
extension on the state of the art. The DE 455 generally describes and claims
“salts with bases” which AstraZeneca contrasts with the ‘653’s specific
enumeration of base addition salts. This difference, however, is superficial.
The skilled person would have understood the DE 455’s reference to “salts with
bases” as referring to pharmaceutically acceptable salts that can be formed
from the addition of a base. No inventive ingenuity would have been required
to know that these would include salts formed by the addition of metal cations,
such as sodium, potassium, calcium, and magnesium, in an appropriate
hydroxide. The DE 455 notes that basic salts of the racemic substrates can be
formed with these metallic cations in a “customary manner” and that the skilled
person would recognize that these same cations could also be used to form salts
of the resulting enantiomers. While AstraZeneca’s experts (Drs. Davies and
Armstrong) emphasized that the specific salts enumerated in the ‘653 were not
exemplified in the DE 455, this came across in their testimony as a minor
technical difference easily overcome by the skilled person. Further, as Dr.
Jacobsen explained:
[P]reparation of salts is advantageous for
purification for storage, for handling of compounds, and it's the norm in
pharmaceuticals. In fact, omeprazole, which was a marketed drug at that time,
was sold as a salt. As a salt, they called, with base. That simply means that
the salt is generated by adding a base and generating a deprotonated form of
omeprazole as a salt. That salt has to have a cation, a plus charge, to go
with it. Those are listed as sodium, potassium, calcium, magnesium, types of
pharmaceutically acceptable salts. This is very, very standard language in any
pharmaceutical patent or description and a skilled chemist would understand
what's referred to by those salts. It's simply sort of a simple matter of
considering the readily accessible salts of the compound in question. (Jacobsen Chief, Trial Transcript, Vol 2, p 322 at line 26 – p 323
at line 15)
[302] With that understanding, which Dr. Jacobsen explained credibly,
there is no reason to doubt that the listing of specific salts would be an
obvious extension of “salts with bases.” Indeed, the experts for AstraZeneca
did little more than point to the absence of an enumerated list in the DE 455
(a difference between the state of the art and the inventive concept) as
opposed to describing the ingenuity required in listing specific salts (whether
it takes an inventive step to make up for that difference). When viewed
in light of the compelling evidence raised by Apotex on this point,
AstraZeneca’s evidence was insufficient to rebuild its claim that listing
specific salts is inventive.
[303] Further, as chemistry is an experimental field, this conclusion is
bolstered by how identifying specific salts in light of the state of the art
is, at a minimum, obvious to try. The procedure described for the making of
omeprazole salts in the ‘495 could be used, without modification, to obtain the
salts of the enantiomers of omeprazole, and the skilled person would know that
the same method could be employed to form salts of the enantiomers of
omeprazole with bases and could readily determine which of the salts formed the
best crystalline solid. As a consequence, no inventive step would be required
to go from the state of the art (salts with bases) to the inventive concept in
the ‘653 (enumerated salts).
[304] The (-) enantiomer of omeprazole was also an obvious extension on
the state of the art. The DE 455 patent application exemplifies the
preparation of the (+) enantiomer in its examples 5 and 6. Further, the
skilled person would understand that this example also teaches how to obtain
the (-) enantiomer using the enantiomer of the reagent specified. The
obviousness experts from both sides (Drs. Jacobsen and Davies) essentially
agreed on this point.
[305] Consequently, getting from the state of the art to either the
listing of specific salts or the isolation of the (-) enantiomer would not have
required an inventive step. While they are differences between the state of
the art and the inventive concept of the ‘653, they would have been overcome by
routine non-inventive experimentation.
(b)
The Inventive Step from the State of the Art to
a Specified Optical Purity of 99.8%ee
[306] In contrast with the differences described above, an inventive step
would have been required for the skilled person to get from the state of the
art to the optical purity specified in the ‘653. This non-obvious difference,
in light of it being the most disputed and controversial difference on the record,
merits a more elaborate analysis of the factors pertaining to the obviousness
and obvious to try inquiries. I will discuss every factor relevant in this
particular case in turn.
(i)
Motivation to Separate the Enantiomers of
Omeprazole
[307]
The motivation to separate the enantiomers of
omeprazole is central to the question of obviousness in this case. Absent such
a motivation, it is hard to believe that the innovation found in the ‘653,
which is predicated on such a separation, was obvious.
[308] The most compelling evidence on motivation came from Dr. Armstrong.
Before his testimony there was minimal engagement between the experts – they
simply disagreed about whether or not the skilled person would have been
motivated to separate the enantiomers of omeprazole before the filing date.
However, after Dr. Armstrong’s testimony, it became clear that the differences
in opinion amongst the experts with respect to omeprazole could be explained by
the varying specificity of their analysis. Put differently, Dr. Armstrong explained
how, while there was a motivation generally to separate racemic drugs,
there was no motivation specifically to separate omeprazole because of
its unique attributes, namely, a wide therapeutic profile and no toxicity.
[309] First, it is undisputed that there was little to no motivation to
investigate the difference in PPI activity between the enantiomers of
omeprazole. This is because of the mechanism of action of omeprazole through
which its enantiomers are converted into an achiral sulfenamide – the inhibitory
active agent in the parietal cell. As explained earlier, chirality is the
primary attribute distinguishing the enantiomers. Consequently, the transfer
into an achiral sulfenamide eliminates that difference, leaving the
enantiomers with presumably equal effect.
[310] Second, there was no motivation to investigate differences in
toxicity between the enantiomers. Omeprazole was very safe, with an “unusually
high” therapeutic index (meaning that there was an inordinately high spread
between safe and unsafe doses when administering omeprazole). A safe racemic
drug with little to no negative side effects provides minimal motivation to
investigate the enantiomers to identify whether negligible amounts of toxicity
originate with a particular enantiomer.
[311] Third, there was little to no motivation to investigate different
enantiomers in the interest of improving on the properties of omeprazole. Dr.
Wainer (for Apotex) spoke at length about the “possibility” that an enantiomer
“could” have improved properties over its racemate – a contention that
AstraZeneca does not dispute. However, researchers have limited time and
resources. It is unrealistic to assume that the skilled person would pursue
every possible inquiry until it was proven unfruitful (even if information is
the “lifeblood” of the chemist, as Dr. Jacobsen described). Instead, the
skilled person, when faced with a scientific problem, would weigh costs and
benefits of various options in charting a reasonable course of research. In
this regard, Dr. Davies views were most on point:
Chemistry is certainly an experimental science,
but there are millions of experiments out there to do. You have to pick ones
that you think are reasonably going to – there’s a reasonable course of action.
If you don’t have any expectation it’s going to work, why would you go down
that track? (Davies Chief, Trial Transcript, Vol 10, p 1746 at lines 17-23)
[312] To hold otherwise would render the motivation criteria not only
permanently resolved in all cases, but further, permanently resolved against
the interests of the innovator. Virtually any scientific endeavour “could”
yield helpful results. The focus when discussing the factor of motivation in
the context of obviousness is not about whether there is a theoretical
motivation that can be speculated about, but rather, whether or not a specific
motivation would have actually existed with respect to the patent in question.
As Justice Rothstein held in Sanofi-Synthelabo Plavix, at para 90:
It is well known that the pharmaceutical
industry is intensely competitive. Market participants are continuously in
search of new and improved medications and want to reach the market with them
as soon as possible. So demand for an effective and non-toxic product to
inhibit platelet aggregation might be assumed to exist. However, nothing in the
‘875 patent or common general knowledge provided a specific motivation
for the skilled person to pursue the ‘777 invention. The prior patent was a
genus patent, and selection might be expected. However, the prior patent did
not differentiate between the efficacy and the toxicity of any of the compounds
it covered. This suggests that what to select or omit was not then self-evident
to the person skilled in the art (emphasis added).
[313] In this case, no such specific motivation to separate the
enantiomers of omeprazole existed. The state of the art with respect to
omeprazole would have informed the skilled person that it would have been “very
difficult to improve on omeprazole” because of its characteristics: a wide therapeutic
index, known safety and efficacy, and clinically negligible differences in
metabolism between individuals. Apotex’s attempt at describing the motivation
as a desire to know everything about omeprazole and its enantiomers (potential
toxicity, potential properties) is unpersuasive without a scientific problem to
be addressed. It is too easy for a generic to argue that a motivation stems
from a desire to comprehensively record all potential information about a
compound when there is nothing motivating the recording of that
information.
[314] Further, the absence of studies on other similar racemic drugs
bolsters the conclusion that there was minimal motivation to separate the
enantiomers of omeprazole. There is no evidence of any racemic drug that as of
May 1993 was considered safe and efficacious with a wide therapeutic index,
where one of the enantiomers was developed to a clinical drug because of
differences in the affinities of the enantiomers for drug metabolizing enzymes.
[315] In sum, there was little to no motivation to investigate the
enantiomers of omeprazole. Apotex convincingly argued that the skilled person
would have a general interest in enantiomers. Indeed, many other
pharmaceutical discoveries discussed in this case related to racemic drugs and their
enantiomers. However, AstraZeneca specifically distinguished omeprazole
from that general interest given omeprazole’s unique attributes. It was
a widely successful “blockbuster” drug whose enantiomers presented a low
expectation of improvement on the racemate because of its wide therapeutic
window and lack of toxicity. This absent motivation strongly militates in
favour of the inventiveness required in attaining a single enantiomer with an
optical purity of 99.8%ee.
(ii)
Climate in the Field
[316] In Novopharm Levofloxacin, this factor was described as the “attitudes, trends, prejudices and expectations” in the
field (at para 25). In a way, this complements the analysis of motivation
above because a field whose climate deters against researching a particular question
limits the motivation to research that question. In this case, the climate
surrounding enantiomeric separations, for the most part, deterred against
researching the resolution of omeprazole.
[317] The attitude that omeprazole was the “gold standard” of PPIs
contributed to a trend of investigating analogues to omeprazole rather
than improving on it directly. This further undermines the obviousness of
separating its enantiomers and investing energy in isolating them with an
optical purity reaching 99.8%ee.
[318] The evidence showed that many large pharmaceutical corporations,
rather than looking into the enantiomers of omeprazole, were instead
investigating analogues (i.e. new molecules with structural similarities to
omeprazole). Most notably, with respect to Sepracor, a company who
specifically adopted the business model of patenting single enantiomeric forms
of racemic drugs, there was no suggestion that it had ever applied to patent
esomeprazole or expressed an interest in doing so. This absence of interest
from Sepracor, of all companies, in omeprazole, casts significant doubt on
Apotex’s assertion that there was a specific motivation to investigate the
enantiomers of omeprazole, or that the climate in the field would have
encouraged such an investigation.
[319] The “climate in the field” factor is also engaged by the purported expectation
of racemisation for esomeprazole and the prejudice that this
allegedly caused against further consideration of separating the enantiomers of
omeprazole. While I ultimately find that the “fear of racemisation” advanced
by AstraZeneca is unpersuasive, the climate in the field still leans in favour
of the non-obviousness of the ‘653 because of the trends discussed above with
respect to investigating analogues. In other words, the motivation to not
investigate the enantiomers of omeprazole still stands despite the lack of a
fear of racemization. A fear of razemization is merely a deterrent
against researching esomeprazole. Eliminating that deterrent against
researching esomeprazole (removing a stick) is not equivalent with creating an
incentive to perform such research (adding a carrot).
[320] The purported fear of racemisation advanced by AstraZeneca is
unpersuasive and therefore would not have deterred the skilled person from
investigating the enantiomers of omeprazole. The basis for a fear of
racemisation was the Brändström paper (Brändström A, “Chemical
reactions of omeprazole and omeprazole analogues. III Protolytic behaviour of
compounds in the omeprazole system” Acta Chem Scand. (1989)
43:569); a paper not addressing the racemisation of omeprazole, but rather,
which provides indirect support for a speculative possibility of
racemisation. From that pillar, the experts from AstraZeneca built up a
purported fear of racemisation which would paralyze a skilled person
considering the investigation of esomeprazole. Unfortunately for AstraZeneca,
that pillar rested on a weak foundation of science and reason.
[321] First, Apotex’s experts’ disbelief in a fear of racemisation facing
the skilled chemist is more credible because its experts more closely emulated
the perspective of the skilled person. Drs. Jacobsen and Danheiser (for
Apotex) had mandates that allowed them to opine on the state of the art, in the
words of the Supreme Court of Canada, “viewed without any
knowledge of the alleged invention as claimed” (Sanofi-Synthelabo
Plavix, at para 67), while both Drs. Davies and Armstrong (for
AstraZeneca) did not. More specifically, the Apotex experts
were “blinded” from the 653 for their initial reports addressing whether and
how the enantiomers of omeprazole could be obtained. By contrast, Dr. Davies
has given extensive evidence in prior esomeprazole litigation, while Dr.
Armstrong, to a lesser extent, has also addressed the ‘653 patent in a prior
case.
[322] Admittedly, I have accepted conclusions put forward by the
AstraZeneca experts on other points above, such as motivation, despite their
not being blinded to the 653. That is because, put simply, the explanations
advanced by the experts for AstraZeneca, on those points, were more
persuasive. However, the blinding of the Apotex experts is particularly
relevant to the issue of a fear of racemisation because that fear is predicated
on a single paper that the skilled chemist would likely not have considered
particularly relevant or interpreted as the AstraZeneca experts urge. The
Apotex experts, who emulated a literature search that would have been performed
by the skilled chemist investigating the separation of omeprazole (a literature
search made particularly convincing given their lack of knowledge with respect
to the ‘653), failed to find any basis for a fear of racemisation. Their
approach to the search was rigorous and credible. Indeed, Dr. Danheiser took
emulating the skilled person so seriously that he performed a manual search
of archived chemical abstracts to literally follow the precise steps
that would have been pursued by the skilled person as of the relevant date.
[323] In contrast, AstraZeneca did not endeavour to similarly situate its
witnesses. Further, AstraZeneca’s experts commented on publications provided
to them by counsel, even where they appeared in journals the experts themselves
had never heard of or read. This approach casts significant doubt on the fear
of racemisation proposed by AstraZeneca. While the ‘653 patent does describe
the stability against racemisation of the claimed salts as being “surprising,”
AstraZeneca’s approach leaves the impression that its experts, who consulted
the ‘653 in advance of providing their reports, went about to justify that
proposition rather than test its veracity.
[324] Second, Apotex’s experts’ disbelief in a fear of racemisation facing
the skilled chemist is more credible because their interpretation of the
Brändström paper – the foundation of the fear of racemisation – is more
reasonable. In their initial reports, the Apotex experts say little to nothing
about Brändström because, in their view, the skilled person seeking to resolve
omeprazole would concentrate primarily on the DE 455 and Erlandsson – two
pieces of prior art with clear and direct relevance to the problem before
them. By contrast, the AstraZeneca experts took a single passage from the
Brändström paper, dissociated from the paper’s thesis, and stretched that
passage as far as possible to justify a fear of racemisation. To be clear, my
issue with the fear of racemisation described by AstraZeneca is not that
racemisation is scientifically impossible, but rather, that the weight assigned
to a fear of racemization was inordinate – a fear more analogous to a paper
tiger than a lion in the path of the skilled person (AstraZeneca Canada Inc
v Teva Canada Ltd, 2013 FC 245 at para 56).
[325] To understand why AstraZeneca’s emphasis on the Brändström-based
fear of racemisation was inordinate demands a brief outline of the logic and
science behind the argument. Brändström does not stand for the proposition
that omeprazole will inevitably racemize in all circumstances and that its
resolution is therefore a fruitless endeavour. Rather, in a single passage, it
indicates that deprotonation of a methylene group is possible at a pKA of the
methylene protons that is “too high to be of interest.” From this, the
AstraZeneca experts extrapolated about how that deprotonation could lead
to racemization. But Brändström provides no experimental detail as to the
conditions in which this racemization occurred or whether those conditions are
ones that they should worry about in the formation of a salt. Further,
Brändström does not indicate that the deprotonation caused a chemical reaction
which would lead to racemization or degradation.
[326] The paralyzing fear of racemization advanced by AstraZeneca, based
on a single reference that was not found by the credible literature searches
performed by Apotex’s blinded experts, is unconvincing. At most, such an
extrapolation would have alerted the skilled person to a possible speed bump,
rather than an insurmountable barrier. To permit such extraneous concerns to
preclude the skilled person from performing routine testing in the
investigation of science would preclude the ability of the skilled person to
adequately emulate the predicted conduct of researchers in the patent law
context. Every scientific question has some footnote in an obscure publication
suggesting possible struggles when researching a question. To allow innovators
to stretch those footnotes into guaranteed ingenuity on the part of their
patents would set too high a standard for obviousness. Extrapolation is
certainly a part of distilling and interpreting complex scientific literature,
but extrapolating from minor speculative harms into blanket prohibitions on
areas of research is simply not reflective of the reasonable conduct of
scientific researchers or the skilled person.
[327] Further, Apotex convincingly argued that the fear of racemisation
flowing from Brändström was not only inordinately weighted, but ill-founded. Deprotonation
of the methylene proton is extremely unlikely to occur or would occur to an
infinitesimally small degree because of the high pKA of the protons on the
methylene group which would be very difficult to abstract. Additionally,
extensive literature on chiral sulfoxides (the family to which omeprazole
belongs) states that deprotonation does not result in racemization.
AstraZeneca claimed that omeprazole was unique and would not behave like other
chiral sulfoxides, but the mechanisms for racemisation advanced by its experts
were thoroughly debunked under cross-examination as, at best, speculative
concerns falling below an expectation that would dissuade the skilled person
from investigating the omeprazole enantiomers.
[328] In sum, the climate in the field also favours the non-obviousness of
the ‘653. Though the fear of racemisation was misplaced, the climate in the
field, which favoured research into analogues of omeprazole rather than its
separation into enantiomers, supports the ingenuity in attaining a uniquely
high optical purity of esomeprazole.
(iii)
Effort Required
[329]
With respect to the obvious to try analysis,
this factor was described as follows in Sanofi-Synthelabo Plavix, at
para 69: “What is the extent, nature and amount of effort
required to achieve the invention? Are routine trials carried out or is the
experimentation prolonged and arduous, such that the trials would not be
considered routine?” This factor as well favours the non-obviousness of
the ‘653 patent.
[330] Apotex sought to demonstrate the minimal effort required to achieve
the invention in the ‘653 patent through multiple routes. To be clear, the
obviousness inquiry considers whether or not the “state of the art,” not a
single piece of art, can reach the invention without an inventive step. That
said, three principal approaches were advanced by Apotex: (1) DE 455, (2) the
Erlandsson paper, and (3) commercially available columns. None of these
approaches, either independently or in conjunction, would have obviously led to
the degree of purity attained in the ‘653.
[331] With respect to the DE 455, I cannot accept Apotex’s contention that
attaining the ‘653 patent’s degree of purity was obvious to the skilled person
when Apotex’s purported re-enactment of the skilled person’s resolution of
esomeprazole in 1993 consisted of two highly esteemed chemists in 2013, never
instructed to emulate the skilled chemist, who only performed a trial in the presence
of opposing counsel after performing ex parte trials first. If a
99.8%ee level of purity really was obvious to the skilled person then an
individual much closer in aptitude and experience to that skilled person ought
to have reached it through routine trials, not careful calibration of the DE
455 by leading academics whose approach to the DE 455 could only be scrutinized
once polished from multiple practice runs.
[332] With respect to Erlandsson, given Apotex’s view that the skilled
person is naturally motivated to achieve the highest purity possible,
Erlandsson’s achievement of 91.2%ee purity is presumably from an optimized
process seeking a maximal purity. This view is bolstered by Dr. Armstrong’s
explanation of how the efficiency of Erlandsson’s resolution, which was likely
optimized for, engages the same factors as those related to achieving maximal
purity. As a consequence, Apotex’s assertion that Erlandsson could simply be
improved upon is unconvincing. Further, Apotex’s view that techniques from Erlandsson,
such as peak-shaving and recycling, could simply be repeated until any desired
purity was achieved are unconvincing. As Dr. Armstrong observed:
If you have broad peaks […] if you run it
through the system again, they get even broader which goes against separation
and you're in danger of the recycled portion overlapping with the other side of
the other peak. That is the danger. (Armstrong Chief, Trial Transcript, Vol
19, p 2988 at lines 9-14)
[333] Thus, it is incorrect to suggest that any material may simply be
purified to 100% purity (or 99.8%ee, which is not much less) through the
reiteration of these processes.
[334] Finally, the testing with respect to commercially available columns
was particularly unconvincing. Apotex relied on the testing results of Dr.
Conor Scully to demonstrate the ability for commercially available columns to
achieve the purity attained by the ‘653 patent. However, Dr. Scully’s use of a
1999 column to emulate a 1993 column with no evidence that they would operate
with equivalent proficiency (but rather, some evidence that advances over those
6 years likely would have occurred) is unconvincing.
[335] In particular, Dr. Okamoto, Apotex’s primary witness who testified
regarding the similarities between the 1993 and 1999 columns, conceded that he
had little to no actual knowledge of whether any changes may have been made
during that six year span. He also conceded that Daicel (the column
manufacturer) had problems with its columns in the early 1990s that it may have
sought to remedy, a fact which Dr. Armstrong also confirmed. Additionally,
Apotex made little effort to ensure that Dr. Scully’s testing reflected the
testing that would have been performed by the skilled person. Dr. Scully was
not instructed to perform the testing as such a person. Even further, Dr.
Scully was told which limited solvents to purchase, to only use those solvents,
and even which solvent ratios to start with. In total, Apotex was incapable of
demonstrating that Dr. Scully’s testing and the columns he used were materially
analogous to the skilled person and the columns that person would have used as
of May 1993. As a consequence, Apotex’s testing with 1999 columns provides
minimal support to its claim that commercially available 1993 columns would
have obviously led the skilled person to the degree of purity attained in the
‘653 patent.
[336] In sum, having reviewed all of the key pieces of prior art forming
the state of the art and the various testing results applying that state of the
art to the task of resolving omeprazole, the extent, nature and amount of
effort required to achieve the invention in the ‘653 patent did not consist
merely of routine trials. Rather, the skilled person would have been
confronted with multiple avenues, none of which would self-evidently lead to
the significant purity achieved in the ‘653 patent. As a consequence, this
factor also favours the non-obviousness of the ‘653 patent.
(iv)
Actual Course of Conduct Leading to Invention
[337]
With respect to the obvious to try analysis,
this factor was described as follows in Sanofi-Synthelabo Plavix, at
para 70:
Another important factor may arise from
considering the actual course of conduct which culminated in the making of the
invention. It is true that obviousness is largely concerned with how a skilled
worker would have acted in the light of the prior art. But this is no reason to
exclude evidence of the history of the invention, particularly where the
knowledge of those involved in finding the invention is no lower than what
would be expected of the skilled person.
[338] With that in mind, Sverker von Unge (one of the named inventor’s of
the ‘653) and his two year long “informal side project” of separating the
enantiomers of omeprazole is instructive. The background to his discovery
provides further support for the non-obviousness of the ‘653 patent.
[339] First, it is refreshing to note that von Unge is much closer to the
skilled person than the highly esteemed experts in this case, and in that
sense, provides an additional lens on the struggles in isolating the enantiomers
of omeprazole.
[340] Second, von Unge’s course of conduct leading to the ‘653 patent not
only displays the non-obviousness of its invention, but additionally,
exemplifies the extent of a lacking motivation amongst pharmaceutical companies
to investigate the resolution of omeprazole (discussed earlier). The “Omeprazole Successor Project,” AstraZeneca’s research
initiative focussed on omeprazole analogues (rather than omeprazole
enantiomers) was the primary focus of AstraZeneca’s research and development aimed
at following up on omeprazole. Though von Unge was a member of the successor
project, AstraZeneca’s policy enabling researchers to spend 10-20% of their
time on personal projects enabled von Unge to pursue his informal side project
of resolving omeprazole sporadically over the course of two years. Even after
von Unge made the highly pure alkaline salts of omeprazole enantiomers,
AstraZeneca did not investigate the properties of the enantiomers readily
despite their enhanced knowledge of omeprazole.
[341] I note that von Unge’s work, which occurred from 1989-1991, precedes
DE 455 and certain chiral chromatographic options, pieces of art that inform
the skilled person in this case who exists as of May 1993. Still, von Unge’s
pursuit of resolving omeprazole over the course of two years, while sporadic,
at the very least casts doubt on Apotex’s assertion that such a resolution
reaching the same extent of purity as the ‘653 patent would have been obvious
to the skilled person using techniques despite those that came to be known
following his research. Further, von Unge’s experimentation, which involved
the manipulation of various analogues and chiral auxiliaries with varying
connectivity, suggests that his work was not a routine experiment
self-evidently resulting in high optical purity, but rather, multiple
experiments involving the exercise of discretion and indicative of inventive
ingenuity.
(v)
Commercial/Clinical Success
[342] The Federal Court of Appeal in Novopharm Levofloxacin states
that secondary factors – like commercial success – “may
be relevant but generally bear less weight because they relate to facts arising
after the date of the alleged invention” (at para 25). More
specifically, with respect to commercial success, the Court of Appeal observes:
Was the subject of the invention quickly and anxiously received
by relevant consumers? This may reflect a fact that many persons were motivated
to fill the commercial market, which may suggest inventive ingenuity. However,
it may also reflect things other than inventive ingenuity such as marketing
skills, market power and features other than the invention (at para 25).
[343] There is no doubt that Nexium was a commercially successful drug.
Indeed, that commercial success presumably underlay, in part, Apotex’s decision
to enter the market. However, evidence of clinical success, like that of
commercial success, may be attributable to factors extrinsic to the drug
itself. Whether clinical or commercial, the success of Nexium, and the degree
to which doctors may or may not prescribe it, provides limited insight into the
core question of the degree of inventive ingenuity required to invent it: Apotex
Inc v Bayer AG, 2007 FCA 243 at para 47. The Court heard evidence from two
expert witnesses in respect of the clinical success of esomeprazole: Dr. Vakil
and Dr. Howden. As will become clear, the evidence does not lean decidedly one
way or the other, such that evidence of clinical success informs the
obviousness inquiry.
[344] Dr. Nimish Vakil is a clinical professor of medicine at the University of Wisconsin, in Madison, Wisconsin. He is a Fellow of the American College of Gastroenterologists, has published extensively, lectured globally, and sits on the
editorial boards of several academic journals. He is the author of the chapter
on peptic ulcers in a leading medical textbook on gastroenterology. He has
large clinical practice dealing with more complex gastro oesophageal reflux
disease (GERD) cases. He was qualified as an expert in the field of
gastroenterology, with particular expertise in the treatment of GERD, including
the pathophysiology, diagnosis, and management of GERD and the use of PPIs in
clinical practice.
[345] Dr. Vakil testified that for patients with mild to moderate GERD
symptoms, the original PPI’s worked and continue to work well. According to
Dr. Vakil, the majority of patients do not benefit from Nexium more than other
PPIs (e.g. pantoprazole). Dr. Vakil’s opinion, however, was that there was
extensive support in the academic literature for the fact that Nexium provides
more effective treatment than other PPIs in patients with more severe cases of
erosive esophagitis (EE). EE is not present in all patients with GERD, rather
it is a complication of GERD. Erosions of the oesophagus are graded based on
an international classification system called the Los Angeles Classification,
which has four grades, A through D, to categorize the degree of severity.
[346] I recognize, from the evidence of Dr. Vakil, that esomeprazole has
been successful, based on his own clinical experience, as a therapy for Grades
C and D of EE. Accepting that in his experience it has an advantage over
omeprazole, the difference is marginal both in scope and effect and, for the
reasons that follow, does not inform the analysis of the obviousness issue.
[347] Dr. Howden is a Professor of Medicine in the Gastroenterology
Division of the Department of Medicine at Northwestern University in Chicago, Illinois, and an attending physician at Northwestern Memorial Hospital. He is a certified specialist in internal medicine and gastroenterology by the
American Board of Internal Medicine, and is equally accredited in the UK. He is, amongst others, a Fellow of the American Gastroenterological Association and
the British Society of Gastroenterology. Dr. Howden is also the associate
editor of the American Journal of Gastroenterology, sits on the editorial board
of six other relevant journals, and has refereed articles for the New England
Journal of Medicine, the Annals of Internal Medicine, JAMA, and the Lancet.
[348] Beginning with his doctoral thesis “Clinical
Pharmacological Studies with Omeprazole: A Novel Inhibitor of Gastric Acid
Secretion” in 1985, Dr. Howden’s career has focused on the treatment and
management of GERD, including consultation on the design and conduct of clinical
trials. He has considerable expertise in the use of all PPIs, including
Nexium, and is familiar with the clinical trials of esomeprazole that led to
its approval by the Food and Drug Administration (FDA).
[349] Dr. Howden was quick to acknowledge, in both his expert report and
oral testimony, many points of Dr. Vakil’s evidence with which he agreed. He
disagreed, however, with Dr. Vakil’s opinion that there is “substantial support in the academic literature for the fact
that Nexium provides more effective treatment than other PPIs in patients at
the severe end of the spectrum.” Dr. Howden’s opinion is that the
literature, on a proper reading, does not demonstrate “more effective” treatment by esomeprazole, but only modest
statistically significant differences, and that there is little evidence that
these differences equate to clinical advantages.
[350] I prefer the evidence of Dr. Howden, which reflects a more targeted
and, for the purposes of the issues before the Court, relevant reading of the
literature. I also prefer Dr. Howden’s opinion as it is consistent with the
prevailing guidance of specialized medical associations that have examined the
question of clinical effectiveness of PPIs.
[351] In sum, I am not satisfied that the evidence of clinical success in
the use of esomeprazole in the treatment of GERD assists AstraZeneca in
responding to the question of whether the ‘653 patent was obvious. There are
several reasons for this. As will be seen there are limitations in the
clinical studies comparing esomeprazole to other PPIs such that, to the extent
that conclusions can be drawn, they are limited and equivocal.
[352] Clinicians are restricted to the use of approved formulations of
drugs in their trials. They can only administer, and hence, evaluate, the
efficacy of the dosages approved by the FDA in the United States or Health Canada in Canada. As a result, different dosages of drugs are compared, which necessarily
undermines the lessons to be drawn from the clinical trial. In other words, it
is not an apples to apples comparison. See for example:
a. Esomeprazole Versus Other Proton Pump Inhibitors in Erosive
Esophagitis: A Meta-Analysis of Randomized Clinical Trials, Clinical Gastroenterology and Hepatology 2006;4:1452-1458 (Tab 19
Expert Statement of Nimish Vakil, MD, July 21, 2013);
b. The new proton pump inhibitor esomeprazole is effective as a
maintenance therapy in GERD patients with healed erosive oesophagitis: a
6-month, randomized, double-blinded, placebo-controlled study of efficacy and
safety, Aliment Pharmacol Ther 2001: 15: 927-935
(Tab 24 Expert Statement of Nimish Vakil, MD, July 21, 2013) which contrasts
results for different doses of esomeprazole;
c. Esomeprazole 20 mg and lansoprazole 15 mg in maintaining healed
reflux oesophagitis: Metropole study results,
Aliment Pharmacol Ther 2003: 17: 333-341 (Tab 25 Expert Statement of Nimish
Vakil, MD, July 21, 2013) which compares two different drugs in different
doses; or
d. Effect of Esomeprazole 40 mg vs Omeprazole 40 mg on 24-Hour
Intragastric pH in Patients with Symptoms of Gastroesophageal Reflux Disease, Digestive Diseases and Science, Vo. 47, No. 5 (May 2002), pp
954-958 (© 2002) (Tab 12 Expert Statement of Nimish Vakil, MD, July 21, 2013)
where the authors noted that esomeprazole 40mg “demonstrates
improved acid inhibition over omeprazole 20 mg.”
[353] I turn now to my second reservation.
[354] While some of the studies indicate a statistical difference in
treatment success rates, there is also uncertainty as to whether this
statistical difference translates into a clinically meaningful difference. I
note, for example, that in Esomeprazole Compared with Omeprazole in Reflux
Oesophagitis: Aliment Pharmacol Ther 2000: 14: 1249-1258 (Tab 14 Expert
Statement of Nimish Vakil, MD, July 21, 2013) the authors conclude at page
1253:
This greater efficacy for esomeprazole 40 mg
vs. omeprazole 20 mg was seen consistently when adjusting for baseline
oesophagitis grade and remained statistically significant after 8 weeks’
treatment based on crude rates […]. At week 4, the difference between the
esomeprazole 20 mg vs. omeprazole 20 mg group was not statistically significant
(P = 0.09). There were no differences in healing at week 8 between the
three groups with respect to age or gender.
[355] Importantly, both Dr. Vakil and Dr. Howden agreed that there was a
difference between statistically significant results and clinically significant
results.
[356] One of the articles relied on was a metadata analysis of 10 randomly
selected clinical trials, conducted over 8 weeks by Gralnek et al: Clinical
Gastroenterology and Hepatology 2006;4:1452-1458 (Tab 19 Expert Statement of
Nimish Vakil, MD, July 21, 2013). The authors’ conclusion is consistent with
Dr. Howden’s observation that the link between statistically significant and
clinically significant results in patients is tenuous:
In summary, we found that as compared with
alternative PPIs, esomeprazole provides a statistically significant but only
modest degree of improved effectiveness in the healing of EE, and this appears
to be largely limited to those individuals with more severe erosive disease (LA
grades C and D). In addition, we found no evidence of what we believe would be
considered a clinically meaningful improvement in symptom relief with
esomeprazole compared with alternative PPIs, although clinical meaningfulness
is subjective, is determined by each individual practitioner, and might vary
widely.
[357] There are only two studies which compare omeprazole against
esomeprazole in equal dosages. In A Multicenter, Randomized, Double-Blind,
8-Week Comparative Trial of Low-Dose Esomeprazole (20mg) and Standard-Dose
Omeprazole (20mg) in Patients with Erosive Esophagitis, Lightdale et al, Dig
Dis Sci (2006) 51:852-857 (Tab 16 Expert Statement of Nimish Vakil, MD, July
21, 2013) the authors observed 1176 patients over 8 weeks treated daily with 20
mg of esomeprazole and 20 mg of omeprazole. They concluded that cumulative
healing rates at 8 weeks for esomeprazole were “similarly high” for omeprazole
and that “the two treatments were comparable for other
secondary measures and had similar tolerability profiles.”
[358] The authors continue, at pages 855 and 856 and conclude:
In this study, 20 mg esomeprazole had a higher
healing rate than 20 mg omeprazole at 8 weeks, but the difference was not
significant. Similar healing rates were achieved at weeks 4 and 8 with
low-dose esomeprazole (20 mg) and standard-dose omeprazole (20 mg) in the
entire study population and when patients were classified according to baseline
severity of EE. Patients in both treatment groups also had similar control of
heartburn at week 4.
[…]
The lack of a statistically significant
difference in EE healing between 20 mg esomeprazole and 20 mg omeprazole in
this study differs from the results of other studies [16, 17].
[…]
In this study, however, 20 mg omeprazole failed
to show a significant advantage over 20 mg omeprazole in short-term EE healing
rates. In conclusion, the findings of this study show that both 20 mg
esomeprazole and 20 mg omeprazole once daily are effective for the healing of
EE and for heart-burn resolution in patients with GERD. Both PPIs are well
tolerated.
[359] I conclude that there is insufficient evidence of the clinical
success of esomeprazole to alter my analysis in respect of obviousness.
[360] As a final note, independent medical associations and one
non-governmental organization have also considered the question of omeprazole v
esomeprazole. Their conclusions reflect the ambivalence of the clinical
experience.
[361] In March 2007 the Canadian Agency for Drugs and Technologies in
Health (CADT), issued an Optimal Therapy Report on PPIs. The mandate of the
CADT is to “help health care decision makers, patients,
health care professionals, health systems leaders and policy makers make well
informed decisions.” After expressing reservations about the relatively
small number of randomized control trials, and the “poor” quality rating of the
existing trials, the authors continued:
There are no clinically important differences
among standard-doses of PPIs (omeprazole 20 mg, lansoprazole 30 mg, pantoprazole
40 mg, rabeprazole 20 mg, esomeprazole 20 mg) in the treatment of symptomatic
GERD, ENRD and esophagitis. (Exhibit 205, p 6)
[362] Further, The Canadian Consensus Conference on the management of
gastro-oesophageal reflux disease in adults – 2004 [the Consensus
Conference], as published in the Canadian Journal of Gastroenterology (2005),
Volume 19, No 1 observed that:
24 h intragastric pH studies suggest greater
suppression of gastric acidity with esomeprazole 40mg compared with
lansoprazole 30 mg, although these differences do not necessarily lead to a
difference in esophageal acid exposure. There also appears to be a
dose-response effect for some PPIs (108, 109). Corresponding to the pH
results, the results of meta-analyses suggest that standard dose omeprazole,
lansoprazole, pantoprazole and rabeprazole are equivalent to each other with
respect to healing esophagitis.
[…]
These differences do not necessarily lead to a
difference in oesophageal acid exposure.
[…]
While these meta-analyses and the largest
available randomized controlled studies suggest that esomeprazole 40 mg
produces somewhat higher four- and eight-week healing rates than standard dose
omeprazole, lansoprazole or pantoprazole, particularly in more severe (Los
Angeles grades C and D) erosive esophagitis, overall differences in healing
proportions at eight weeks are small, ranging from just over 3% to just over
6%. Furthermore, although the differences are statistically significant, their
clinical relevance is debated and the results have not been replicated
consistently in other studies.
[363] The Consensus Conference made no recommendation with respect to the
choice of PPI for initial or long-term therapy, noting that factors such as
cost affect the choice of PPI.
[364] To conclude, there is competing evidence as to whether, on a
clinically significant basis, esomeprazole has been established as superior.
There is statistical evidence that it is more effective in respect of more
severe cases of EE, but not sufficient to warrant conclusions in the specialized
professional groups responsible for the treatment of GERD. No findings can be
therefore drawn in respect of clinical success such that they inform the
obviousness inquiry.
D.
Conclusion on Obviousness
[365] The ‘653 patent was not obvious, nor were the specific methods
ultimately employed to reach its high purity obvious to try.
[366] The high degree of purity attained by the ‘653 patent could not be
achieved through routine trials informed by the state of the art as of May
1993. Certainly, methods known to the skilled person could have resulted in
potentially fruitful discoveries regarding the resolution of omeprazole, but
that resolution, and more importantly, an optical purity of 99.8%ee, would not
have been self-evidently achieved. Put differently, an inventive step would
have been required to reach the ‘653 from the state of the art. As a
consequence, its invention was not obvious.
VIII.
Conclusion
[367] The ‘653 patent, though it was novel and non-obvious, is invalid
because it lacks utility. It promised more than it could provide. Its
ingenuity and novelty could not overcome the absence of a demonstration or
sound prediction of an improved therapeutic profile such as a lower degree of
interindividual variation, which the ‘653 promised to provide. Accordingly, despite
its many strengths, for this fatal flaw, the ‘653 patent for Nexium is invalid.
Email this Content