III. Issues
[11]
In very general terms, the assessment of Apotex’s damages
involves five steps:
1.
determine the duration of the period of liability (the
Relevant Period);
2.
determine the overall size of the ramipril market during
the Relevant Period (the Ramipril Market);
3.
determine the portion of the Ramipril Market that would
have been retained by Sanofi and the portion that would have been held by
generic manufacturers during the Relevant Period (the Generic Market);
4.
determine the portion of the Generic Market that would have
been held by Apotex (Apotex’s Lost Volumes); and
5.
quantify the damages that would have been suffered by
Apotex in respect of Apotex’s Lost Volumes (Apotex’s Net Lost Profits).
[12]
In the case before me, these steps require consideration of
a number of issues where the parties are in disagreement. These issues are as
follows:
1.
What is the appropriate date for the commencement of the
Relevant Period for which loss can be claimed by a second person under s. 8 of
the Regulations:
a.
April 26, 2004, the date Health Canada’s review of Apotex’s drug submission was completed and Apotex was advised
that an NOC would not issue until the requirements of the Regulations
were met (referred to as the “patent hold” date; see Exhibit 1, Tab 2); or
b.
December 13, 2005, the date of expiry of Canadian Patent
No. 1,246,457 (the '457 Patent), which was the
subject of a Prohibition Order (the Prohibition Order or the Order) granted by
Justice Simpson in Aventis Pharma Inc v Apotex Inc, 2005 FC 1381, 281
FTR 233 [Ramipril NOC #2 (FC)];
2.
What is the appropriate date for the ending of the Relevant
Period, having regard to whether Apotex was a “second person” for purposes of
the Regulations:
a.
May 2, 2008, the date of the dismissal of the last
prohibition proceeding in Court File No. T-87-06 by Order of Prothonotary
Aalto;
b.
June 27, 2006, the date of dismissal of Court File No.
T-1499-04; or
c.
December 12, 2006, the date of Apotex’s NOC for ramipril?
3.
What would have been the size of the Ramipril Market over
the Relevant Period?
4.
What would have been the size of the Generic Market during
the Relevant Period?
5.
What would have been Apotex’s Lost Volumes during the
Relevant Period? Subsidiary to this question are the following sub-issues:
a.
In assessing Sanofi’s liability under s. 8, is Sanofi’s
liability to be assessed on the basis of a single “but for” world which
includes all potential generic manufacturers?
b.
What other generics would likely have come to market during
the Relevant Period and when? Specifically, would any or all of Teva,
Laboratoire Riva Inc. (Riva) and/or Pharmascience Inc. (Pharmascience or PMS),
or an authorized generic have launched during the Relevant Period?
c.
What portion of the Generic Market would have been captured
by Apotex during the Relevant Period? In other words, what would have been
Apotex’s Lost Volumes?
6.
Based on my finding as to Apotex’s Lost Volumes, what is
Apotex’s Lost Gross Sales, having regard to the pricing of Apo-ramipril during
the Relevant Period, considering the provincial formularies?
7.
Based on my finding as to Apotex’s Lost Volumes, what is
Apotex’s Net Lost Profits, having regard to relevant matters, including:
a.
sales returns;
b.
likely trade spend (including discounts and allowances)
that would have been paid by Apotex to pharmacists and distributors to stock
Apo‑ramipril;
c.
likely price of the active pharmaceutical ingredient (API);
and
d.
other potential adjustments?
8.
Is a second person entitled to recover under s. 8 of the Regulations
for lost sales that would have been made as a result of prescriptions that were
aimed at unapproved indications?
IV. Essential
Background
A. Statutory framework under the PM (NOC) Regulations
[13]
This action arises solely out of the operation of the PM
(NOC) Regulations. Quite simply, Apotex was kept off the market for a
period of time by the actions of Sanofi that were ultimately found to be
unsustainable. In his decision in Apotex Inc v Merck & Co, 2008 FC
1185 at paras 35-51, [2009] 3 FCR 234 [Alendronate (FC)], Justice
Hughes provides a comprehensive history and rationale of the Regulations
and s. 8, in particular. Although the decision in Alendronate (FC) was
overturned in part by the Court of Appeal in Apotex
Inc v Merck & Co, 2009 FCA 187, [2010]
2 FCR 389, rev’g 2008 FC 1185, leave to appeal to SCC
refused [2009] SCCA No 347 [Alendronate
(FCA)], Justice Hughes’s
description of the background to the PM (NOC) Regulations remains a
valuable tool. Rather than restate this history here, I commend the identified
passages to the reader.
[14]
The damages suffered by Apotex are statutory in that they
arise only because of the operation of s. 8 of the PM (NOC) Regulations.
The liability of Sanofi, in this case, is better understood if s. 8 is examined
in the context of the entire statutory scheme. I will provide a brief overview
of the statutory scheme that gives rise to Apotex’s claim. Ms. Anne Bowes, the
Director of the Office of Patented Medicines and Liaison Therapeutic Products
Directorate, Health Canada, was
helpful in explaining the operation of the applicable regulations and policies
engaged on the facts of this case.
[15]
Before a pharmaceutical company can market a prescription
drug in Canada, it must comply with the
provisions of the Food and Drug Regulations, CRC, c 870 [F&D
Regulations]. Section C.08.002 of the F&D Regulations provides,
in part that:
[16]
As provided for in s. C.08.002(1)(a) of the F&D
Regulations, anyone who wishes to sell a drug in Canada must submit, to the
Minister of Health (through Health Canada), either a new drug submission (NDS) or an abbreviated new drug
submission (ANDS). An NDS is filed by an innovative drug company, or “first
person”, seeking approval to market a new drug product. In contrast and in very
general terms, an ANDS is filed by a generic manufacturer, or “second person”,
that wishes to market a generic version of a drug that has already been
approved. The second person may rely on much of the technical, health and
safety information originally filed as part of the NDS by the first person. In
other words, it may compare its drug with, or make reference to, a brand name
drug (F&D Regulations, above at s. C.08.002.1.(1)).
[17]
An essential element of the regulatory scheme is the
“Patent Register”. The PM (NOC) Regulations allow an innovator who has
filed an NDS or a supplement to a new drug submission (SNDS) to submit a list
of the associated patents to the Minister of Health (Minister) for inclusion on
the register of patents (Patent Register or Register) (s. 4(1)). The Regulations
require that the Minister maintain a register of all listed patents (s. 3(2)).
Subsections 4(2) and (3) of the Regulations describe the eligibility
requirements for listing.
[18]
If a patent is listed on the Patent Register, s. 5 of the PM
(NOC) Regulations provides that the second person, with respect to each
patent on the Patent Register, must, in its application for an NOC:
·
state that it accepts that the NOC will not issue until the
patent expires (s. 5(1)(a)); or
·
allege that:
o
the first person is not the patentee or licensee of the
listed patent (s. 5(1)(b)(i));
o
the patent has expired (s. 5(1)(b)(ii));
o
the patent is not valid (s. 5(1)(b)(iii)); or
o
the second person will not infringe the listed patent (s.
5(1)(b)(iv)).
The second person
identifies its election on the Form V submitted with its application.
[19]
If a second person alleges that an NOC should issue in
spite of the listed patents, it must serve a notice of allegation on the first
person (Regulations, above at s. 5(3)). The first person may, within 45
days after service, apply to the Federal Court for an order prohibiting the
Minister from issuing an NOC until the expiration of a patent that is the
subject of the notice of allegation (Regulations, above at s. 6(1)).
[20]
The specific circumstances during which the Minister may
not issue the NOC are dealt with in s. 7(1) of the PM (NOC) Regulations.
Of relevance to these proceedings, the Minister may not issue an NOC to a
second person before the latest of:
·
the day on which the second person complies with the
requirements of s. 5 (s. 7(1)(b));
·
the expiration of any patent on the Register that is not
the subject of an allegation (s. 7(1)(c));
·
the expiration of 45 days after the receipt of proof of
service of a notice of allegation under paragraph 5(3)(a) in respect of any
patent on the Register (s. 7(1)(d));
·
the expiration of 24 months after the receipt of proof of
the making of any application under s. 6(1) (s. 7(1)(e)); and
·
the expiration of any patent that is the subject of an
order of prohibition pursuant to s. 6(1) (s. 7(1)(f)).
[21]
Regardless of the election made by a second person under s.
5(1) of the Regulations, Health Canada will process the application for
all health and safety considerations and will assign a drug identification
number (DIN) (F&D Regulations, above at s. C.01.014.2.(1)). However,
no NOC will be issued until the relevant patents on the Patent Register either
expire (assuming an election to await expiry) or until all such patents have
been addressed through the PM (NOC) Regulations process. The day on
which a generic drug product would have otherwise received its NOC is called
the “patent hold date”.
[22]
The 24 months referred to in s. 7(1)(e) of the Regulations
is referred to as a “statutory stay” or “automatic stay”; the Minister is
enjoined for a period of up to 24 months from issuing the NOC while the first
person pursues its rights in the Federal Court.
[23]
After hearing the application, the court may dispose of an innovator’s
prohibition application in several ways. First, if the court finds that none of
the generic’s allegations are justified, it must issue an order prohibiting the
Minister from issuing an NOC to the generic (Regulations, above at
s. 6(2)). In that case, the generic will not receive its NOC until patent
expiry (unless the decision of the Federal Court is overturned on appeal).
[24]
Alternatively, the court may dismiss the innovator’s
application in whole or in part (Regulations, above at s. 6(5)), or
the application may be withdrawn or discontinued by the first person. If an
application is dismissed, withdrawn, or discontinued, the generic will quickly
receive its NOC. Most relevant to this case, the generic will also be able to
invoke s. 8 of the Regulations. Section 8 allows a generic to bring an
action against an innovator for compensation for the period it was kept off the
market as a result of the innovator’s unsuccessful prohibition application.
[25]
The full text of s. 8 is set out below:
B. Ramipril patents
[26]
Sanofi, either as patentee or licensee, holds the rights to
a series of Canadian patents that include claims to ramipril or its uses. The
initial patent was Canadian Patent No. 1,187,087 (the '087 Patent), a
product-by-process patent for ramipril, issued May 14, 1985. The '087 Patent
was set to expire on May 14, 2002, after 17 years of patent protection. Sanofi,
in efforts to extend patent protection for ramipril, proceeded to obtain a
further series of patents and protect those patents through listings on the
Patent Register. Sanofi describes these subsequent patents and the measures it
took, through litigation and under the PM (NOC) Regulations, as “Altace
Lifecycle Management” (Exhibit 89, Tab 373 at 23). Others – including generic
manufacturers – have referred to the subsequent patents as “evergreening”.
[27]
The following chart describes those subsequent patents
involving ramipril or its uses and identifies when each patent was listed on
the Patent Register:
[28]
The '549 and '387 Patents are referred to, collectively, as the HOPE Patents after the
Heart Outcomes Prevention Evaluation study (HOPE study), discussed in more
detail below.
C. Apotex’s regulatory submissions and litigation
[29]
Between July 2003 and May 2008, Apotex was continuously
engaged in litigation under the PM (NOC) Regulations with respect to
ramipril. The chart that follows describes this history:
[30]
Even though Justice Mactavish, in Ramipril NOC #1 (FC),
dismissed Sanofi’s Notice of Application in respect of the '206 Patent, there were other patents on the Patent Register that needed to
be addressed before an NOC could be issued to Apotex. In particular, Apotex had
to clear the hurdles caused by Sanofi’s decision to commence prohibition
proceedings with respect to the '457, '089, '948, '549 and '387 Patents.
[31]
Ultimately, and as a result of the decision of the Supreme
Court of Canada in AstraZeneca Canada lnc v Canada (Minister of Health),
2006 SCC 49, [2006] 2 S.C.R. 560 [AstraZeneca (SCC)], the Minister of
Health determined that Apotex did not need to address the HOPE Patents. An NOC
was issued to Apotex on December 12, 2006.
[32]
The following day, Sanofi filed an application for judicial
review (T‑2196-06), seeking, among other things, an order quashing the
decision to issue an NOC to Apotex, an order prohibiting the issuance of an NOC
to Apotex, a declaration that the Minister had misinterpreted AstraZeneca
(SCC) and s. 5(1) of the PM (NOC) Regulations, and an interim order
pursuant to s. 18.2 of the Federal Courts Act, RSC 1985, c F-7, staying
the effect of the decision to issue the NOC. In an Interlocutory Order dated
December 29, 2006, Justice von Finckenstein granted the stay. The Order stayed the operation of the NOC and required
Apotex and the Minister to comport themselves as if the NOC had not been
issued. A stay of Justice von Finckenstein’s Order was granted on January 8,
2007 by the Court of Appeal (Sanofi-Aventis Canada Inc v
Apotex Inc, 2007 FCA 7, 54 CPR (4th) 402), thereby removing any
impediment to the operation of the NOC which had issued on December 12, 2006.
Other than the short period between Justice von Finckenstein’s Order on
December 29, 2006 and the Court of Appeal’s stay of that Order on January 8, 2007,
Apotex’s NOC has been in full force and effect since December 12, 2006.
[33]
At the end of the litigation, Sanofi was only ever
successful in one proceeding; that is, Ramipril NOC #2 (FC), where
Justice Simpson issued an Order of Prohibition to be in force until the expiry
of the '457 Patent. Apotex commenced
an appeal of that decision (Court of Appeal File No. A-494-05), which appeal
was discontinued on October 13, 2006.
[34]
To provide a complete picture, it should be noted that
Apotex was not the only company challenging the “evergreening patents”;
beginning in February 2003 and continuing up to December 2006, Pharmascience,
Riva, Teva, Cobalt Pharmaceuticals Inc. (Cobalt) and Sandoz Canada Inc. served
notices of allegation. In each and every case, except for Cobalt’s notice of
allegation in August 2006, Sanofi chose to commence prohibition applications
under the Regulations.
[35]
After its loss in Ramipril NOC #1 (FC), Sanofi
commenced an action against Apotex claiming that Apotex had infringed the '206 Patent (Court File No. T-161-07). In a decision dated June 29, 2009,
this Court dismissed the action and a companion action against Teva (then
Novopharm Inc.), in Court File No. T-1161-07, and declared the '206 Patent to be invalid (Sanofi-Aventis Canada Inc v Apotex Inc,
2009 FC 676, 350 FTR 165).
This decision was affirmed by the Court of Appeal (Sanofi-Aventis Canada Inc
v Apotex Inc, 2011 FCA 300, 426
NR 196). At the time of writing, Sanofi’s application for leave
to appeal to the Supreme Court of Canada remains pending.
[36]
This discussion of the statutory framework, the ramipril
patents and the relevant NOC proceedings forms the context for these Reasons
for Decision.
V. Relevant
Period
[37]
A critical determination for the Court is the commencement
and end dates of the Relevant Period. The parties do not agree on either the
beginning or the end date for the Relevant Period.
A. Commencement date
[38]
As set out in s. 8(1)(a) of the PM (NOC) Regulations,
a first person (Sanofi) is liable to a second person (Apotex) for any loss
suffered during the period:
[39]
In Alendronate (FC), above at paragraphs 106-116,
Justice Hughes explained that s. 8 thus gives the Court discretion to select a
more appropriate date for the beginning of the liability period, although the
presumptive period begins on the patent hold date.
[40]
Here, the parties appear to agree that “the date, as
certified by the Minister, on which a notice of compliance would have been
issued” is April 26, 2004. This patent hold date is set out in a letter dated
April 29, 2004 from Health Canada to Apotex (Exhibit 1, Tab 2). Apotex submits that this should be the date
used for the commencement of the Relevant Period. Sanofi disagrees, arguing
that December 13, 2005 is the appropriate date for the commencement of the
Relevant Period.
[41]
Sanofi’s argument is founded on the existence of a
Prohibition Order of Justice Simpson arising from her decision in Ramipril
NOC #2 (FC). The Prohibition Order prohibited the Minister from issuing an
NOC to Apotex until the expiry of the '457
Patent. Since the application upon which the Prohibition Order was based was
never withdrawn, discontinued, dismissed or reversed on appeal, Sanofi claims
that Apotex cannot allege that it has a s. 8 claim in respect of this
application. Sanofi’s principal argument is that I cannot ignore the
Prohibition Order. Based on that Order, regardless of what transpired with
respect to other notices of application, Apotex would not have been able to
come to market until December 13, 2005, when the '457 Patent expired.
[42]
I do not accept Sanofi’s arguments on this point. In light of the subsequent decision of Justice Tremblay-Lamer in Ramipril
NOC #3 (FC), the Prohibition Order of Justice Simpson had, in my view, no
effect on either the issuance of an NOC to Apotex or Apotex’s s.
8 claim. This is due to the unusual facts of this case.
[43]
As set out above, Apotex served a first notice of
allegation with respect to the '457 Patent, alleging non-infringement, in
August 2003. In response, Sanofi commenced a prohibition application in Court
File No. T-1851-03. On October 6, 2005, in Ramipril NOC #2 (FC), Justice
Simpson found that Apotex’s allegation of non-infringement was not justified
and issued the Prohibition Order, prohibiting the Minister from issuing an NOC
to Apotex until after the expiry of the '457 Patent. Apotex commenced an appeal
of Justice Simpson’s Order, but abandoned it in October 2006, following the
expiry of the '457 Patent.
[44]
In November 2003, Apotex served a second notice of
allegation with respect to the '457 Patent, this time alleging invalidity. On
December 29, 2003, Sanofi commenced a prohibition application in Court File No.
T‑2459-03. On November 4, 2005, in Ramipril NOC #3 (FC), Justice
Tremblay-Lamer dismissed Sanofi’s application, concluding that Apotex’s invalidity
allegation based on obviousness was justified. Sanofi appealed.
[45]
The '457 Patent then expired, and Apotex moved to dismiss
Sanofi’s appeal on the ground of mootness. Apotex’s arguments found favour with
the Court of Appeal which, in Aventis Pharma Inc v Apotex Inc, 2006 FCA 328, 354 NR
316 [Ramipril NOC #3 (FCA)], dismissed the appeal as moot.
Moreover, the Court of Appeal refused to exercise its discretion to hear the
appeal in any event, because Sanofi had failed to show that the decision would
have any practical effect.
[46]
In support of its assertion that Apotex could not have
entered the market prior to the expiration of the '457 Patent, Sanofi relies on
the words of the Court of Appeal in Ramipril NOC #3 (FCA), above at
paragraph 20, where the court stated, “Simpson J.’s prohibition order has
remained in effect until the expiration of the '457 patent”. Sanofi’s reliance
on this sentence, however, ignores the context of that decision. Apotex brought
its motion to dismiss Sanofi’s appeal after the expiry of the '457
Patent. In Ramipril NOC #3 (FCA), the court was not asked to rule
on whether the Prohibition Order was enforceable or of practical effect before
the expiry of the '457 Patent, because of Justice Tremblay-Lamer’s decision in Ramipril
NOC #3 (FC). When the Court of Appeal stated that the Prohibition Order
“remained in effect”, it was not expressing any opinion on the enforceability
of the Order after the decision in Ramipril NOC #3 (FC). That is the
precise question before me.
[47]
In my view, the second '457 Patent decision in Ramipril
NOC #3 (FC) effectively “unlocked” the door for Apotex to receive an NOC
vis-à-vis that particular patent. The logical result was that the first
decision was subsumed or “trumped” by the second. As of the decision of Justice
Tremblay-Lamer, Apotex had addressed the '457 Patent; the Prohibition Order of
Justice Simpson was no longer enforceable or of any practical effect.
[48]
Stated differently, although Justice Simpson’s Prohibition
Order regarding Apotex’s allegation of non-infringement of the '457 Patent was
neither a nullity nor void ab initio, as a result of Justice
Tremblay-Lamer’s subsequent finding that Apotex’s allegation of invalidity was
justified, the Prohibition Order nonetheless could not be acted upon. In
particular, it cannot be used as a basis for holding that Apotex could not have
entered the market until after the expiry of the '457 Patent.
[49]
The logic of this result is reinforced when one considers
what the outcome would have been if Apotex had served one notice of allegation
raising both its non-infringement and invalidity allegations. Had that
happened, a court would have likely found that:
·
the allegation of non-infringement was not justified (as
Justice Simpson found); and
·
the allegation of invalidity was justified (as Justice
Tremblay-Lamer concluded).
Even though Apotex would
likely have been unsuccessful on one of its allegations, Sanofi’s Application
would have been dismissed. There would have been no Prohibition Order.
[50]
There is no principled reason why the result should be any
different just because Apotex served and pursued two notices of allegation
rather than one.
[51]
Sanofi argues that Apotex, having pursued two separate
notices of allegation in respect of the '457 Patent, should live with the
result of its litigation strategy. This argument is without merit. Certainly,
it would have been more efficient to serve one notice alleging both
non-infringement and invalidity. However, this inefficiency does not mean that
the Prohibition Order remains in force and effect until the expiry of the '457
Patent. Indeed, one could argue that Sanofi’s litigation strategy in responding
to the second notice of allegation – which turned out to be without merit –
contributed to or even caused the “mess” that we are now in. Apotex’s
litigation strategy is not to blame for Sanofi’s unsuccessful challenge to the
second '457 allegation.
[52]
Both parties put forward case law that they argue supports
their respective positions. The problem, of course, is that none of the
jurisprudence directly answers the question of the enforceability of a
“prohibition order” after a subsequent application is dismissed with respect to
the same patent.
[53]
For its part, Apotex points to several cases which it says
support the proposition that a second person who has been prohibited on one
notice of allegation may receive an NOC if it succeeds on a subsequent,
discrete allegation. In this regard, Apotex places the most reliance on the
decision in Apotex Inc v Canada
(Minister of National Health and Welfare) (1997), 129 FTR 300 (TD),
aff’d (1997), 153 DLR (4th) 68 (CA), leave to appeal to SCC refused, [1997]
SCCA No 528 [Nizatidine]. While Nizatidine
is authority for the proposition that
a prohibition order “must be confined to the specific allegations advanced in
those proceedings” (Nizatidine, above at para 24), that case does not
directly address the practical effect of an earlier prohibition order with
respect to the same patent, as the first notice of allegation in Nizatidine
was based on non‑infringement due to a licence, while the second alleged
a non-infringing process. Apotex’s reliance on the decisions regarding the drug
norfloxacin, which are summarized in Apotex
Inc v Merck & Co, 2010 FC 287 at paras 2-3,
363 FTR 137 [Norfloxacin (FC)], is similarly wide of the
mark. In particular, and as Apotex acknowledged in argument, none of those
decisions explicitly considered the effect of a prohibition order on a
subsequent notice of allegation.
[54]
Sanofi relies on AB Hassle v Apotex Inc, 2008
FCA 416, 384 NR 372 [AB
Hassle]. In
that case, both the Federal Court and the Court of Appeal held that Apotex
could not use its success in a third NOC case to set aside two prohibition
orders in prior NOC cases, involving different patents. In other words, a
second person cannot “unlock” the NOC door until and unless all patents are
addressed. AB Hassle was a situation where the extant prohibition orders
were in respect of different patents for the same drug; the second person had
failed to address those different patents. That is not the same – despite
Sanofi’s arguments to the contrary – as the situation where the subsequent
dismissal is in respect of exactly the same patent that is the subject of the
prohibition order. In the case before me, Apotex “unlocked” the door by fully
addressing the '457 Patent in a subsequent proceeding; there were no
prohibition orders with respect to other patents.
[55]
In sum on this point, I conclude that the Prohibition
Order, as of the date of Ramipril NOC #3 (FC), could not have prevented
Apotex from obtaining an NOC with respect to the '457 Patent. It follows that
December 13, 2005 is not an appropriate date for the commencement of the
Relevant Period. I find that April 26, 2004, the date of the “patent hold”, is
the appropriate date to begin the liability period.
B. End date
[56]
I now turn to a discussion of the proper “end date” for the
Relevant Period. As set out in s. 8(1)(b), a first person is liable to a second
person “for any loss suffered during the period . . . ending on the date of the
withdrawal, the discontinuance, the dismissal or the reversal”. The drafters of
the Regulations may have contemplated a much simpler scenario than has
been placed before me. In the “normal” circumstances, the NOC would issue as
soon as an application for prohibition is “withdrawn or discontinued by the
first person or is dismissed by the court hearing the application”.
[57]
Here, there are five different dismissal dates relating to
five separate prohibition applications. This case also presents the very
unusual situation in which the second person received an NOC prior to the
disposition of the last prohibition proceeding.
[58]
The parties disagree on the question of the end date.
Apotex would like me to conclude that the Relevant Period ends on May 2, 2008;
Sanofi argues that June 27, 2006 is the correct end date. For the reasons
explained below, neither Apotex’s nor Sanofi’s preferred date can be accepted.
The Relevant Period must end on December 12, 2006.
(1) Apotex’s date: May 2, 2008
[59]
Apotex submits that the Relevant Period ends on May 2,
2008, that being the date of the dismissal of the last prohibition proceeding
in Ramipril NOC #6 (FC). In Apotex’s view, “[t]he plain wording of
section 8 entitles Apotex to claim its damages to this date”.
[60]
The sequence of events in the ramipril history under the
Regulations led to some unusual results. As described in Part IV.C of these
Reasons, Apotex filed its sixth and final notice of allegation with respect to
ramipril and the HOPE Patents on November 29, 2005. Sanofi commenced
prohibition proceedings on January 17, 2006 (Court File No. T-87-06). On
December 8, 2006, the Minister advised Apotex that Apotex was not required
to address the HOPE Patents. However, the Minister determined that Apotex could
not receive an NOC until Apotex had disposed of the prohibition proceeding in
T-87-06, as the Minister held that he remained bound by the 24-month stay
imposed by the Regulations (Exhibit 37, Tab 11). On December 12, 2006,
after receiving representations from counsel for both Apotex and Sanofi, the
Minister decided that Apotex was “no longer considered to be a ‘second person’”
in respect of the HOPE Patents, and that therefore s. 7 of the Regulations
was “not applicable to prohibit the issuance of the NOC”. Apotex accordingly
received an NOC for Apo-ramipril on December 12, 2006, and proceeded to launch
its product following a brief delay (described earlier in these Reasons).
However, the prohibition application in T-87-06 was not technically disposed of
until May 2, 2008, when, upon motion by Sanofi, Prothonotary Aalto dismissed
the application (Ramipril NOC #6 (FC)).
[61]
As correctly pointed out by Apotex, s. 8(1)(b) of the Regulations
requires that the liability period end on the date of the dismissal (or withdrawal,
discontinuance, or reversal) of the relevant prohibition application. In Alendronate
(FC), above at paragraphs 106-109, Justice Hughes observed that, although
s. 8(1)(a) allows the Court to choose a more appropriate date for the beginning
of the liability period, s. 8(1)(b) does not give the Court any discretion to
choose an end date other than “the date of the withdrawal, the discontinuance,
the dismissal or the reversal”. In this case, Apotex argues that the end date
is May 2, 2008, the date on which T-87-06 was dismissed. I do not agree.
[62]
In his Order (Ramipril NOC#6 (FC)), Prothonotary
Aalto concluded that the underlying prohibition application in respect of the
HOPE Patents was moot as of the date of the issuance of the NOC to Apotex. As
stated by Prothonotary Aalto, “[t]here is little doubt that this Application is
moot and became moot when the NOC was issued to Apotex”. Stated differently,
the prohibition application was effectively dismissed as of that date.
[63]
Moreover, May 2, 2008, being the date of Ramipril NOC#6
(FC), has no rational meaning within the context of the Regulations.
It is merely an arbitrary date on which Prothonotary Aalto dealt with a motion
before him. This order could just as easily have been brought on December 13,
2006 or as late as today. Nothing changes the fact that the prohibition
application became moot on December 12, 2006. Even if a motion for dismissal
had never been brought, I cannot imagine that the situation would be any
different. Surely, Sanofi’s liability does not stretch to infinity merely
because neither party thought to bring a motion in respect of a matter that had
become moot.
[64]
Thus, for the purposes of s. 8(1)(b) of the Regulations,
December 12, 2006 – and not May 2, 2008 – must be considered to be the
date of dismissal of the prohibition application.
(2) Sanofi’s date: June 27, 2006
[65]
Sanofi argues that the end date of the Relevant Period
should be June 27, 2006, on the basis that Apotex ceased to be a
second person as of that date.
[66]
The ability to claim damages under s. 8 of the Regulations
is undeniably linked to a claimant being a “second person” under the Regulations.
Subsection 8(1) states that the first person’s liability is “to the second
person”. Under s. 8(2), a “second person” may apply to the court for an order
requiring the first person to compensate “the second person for the loss
referred to in [s. 8(1)]”.
[67]
Sanofi submits that, with respect to NOC proceedings
related to the HOPE Patents (T‑87‑06), Apotex was never a
second person. Accordingly, it argues, Apotex can have no claim under s. 8(1)
related to any period involving the NOC proceedings in T-87-06. Accepting that
Apotex was a second person for all other patents on the Patent Register, Sanofi
then asserts that Apotex ceased to be a second person as of the date of the
dismissal of the final prohibition application where it was a second person;
that was on June 27, 2006, the date when the prohibition
application related to the '948
Patent was dismissed (Ramipril NOC#5 (FC)).
[68]
This argument amounts to an assertion that the HOPE Patent
NOC Proceedings were void ab initio and is founded on Sanofi’s
interpretation of the jurisprudence in AstraZeneca (SCC) and Ferring
Inc v Canada (Minister of Health),
2007 FC 300, [2008] 1 FCR 19, aff’d
2007 FCA 276 [Ferring].
[69]
Contrary to the submissions of Sanofi, neither AstraZeneca
(SCC) nor Ferring goes so far as to declare that Apotex was never a
second person or that the HOPE proceedings were void ab initio.
[70]
The question before the Supreme Court in AstraZeneca
(SCC) was whether the Regulations required a generic manufacturer to
address patents on the Patent Register that had been listed subsequent to the
drug “copied” by the generic manufacturer (in that case, Apotex). In concluding
that the later patents did not need to be addressed, Justice Binnie stated at
paragraph 39 that:
In my view, s. 5(1) of
the NOC Regulations requires a patent‑specific analysis, i.e. the
generic manufacturer is only required to address the cluster of patents listed
against submissions relevant to the NOC that gave rise to the comparator drug,
in this case the 1989 version of Losec 20.
[71]
The Supreme Court was not asked to consider, nor did it
consider, whether its decision would strip Apotex of its claim to damages under
s. 8. Nor did the Supreme Court declare that Apotex was never a second person
or that the prohibition application initiated by Astrazeneca was void ab
initio. In effect, all that the Supreme Court decided was that the Minister
could issue an NOC to Apotex.
[72]
In Ferring, Justice Hughes was faced with five
separate applications for judicial review, all of which dealt with actions
taken by the Minister following the release of AstraZeneca (SCC), above.
In addition to ruling on the five individual applications for judicial review
of the Minister’s decisions, Justice Hughes provided general remarks on the
application of AstraZeneca (SCC); in other words, he provided further
guidance on when a generic manufacturer was obligated to address a patent on
the Patent Register. In his decision, Justice Hughes framed the question in
terms of when a generic is a “second person” for the purposes of s. 5(1) of the
Regulations. For example, at paragraph 61, he states:
If section 5(1) is not
triggered, then the generic is not a “second person” and is not required to
file a notice of allegation. The NOC Regulations do not come into play.
The Supreme Court said [in AstraZeneca (SCC)], at paragraph 41 of its
Reasons:
41. However, it
is clear that AstraZeneca did not market any product pursuant to the subsequent
NOCs and that the preconditions to any obligations of Apotex under s. 5(1) were
therefore not triggered.
[Emphasis in original]
[73]
I acknowledge that Ferring appears to support
Sanofi’s view. However, I think that Ferring unnecessarily frames the
issue in AstraZeneca (SCC) (i.e. whether a generic needs to address a
subsequently listed patent) in terms of whether the generic is a “second
person”. In particular, at paragraph 26 of Ferring, Justice
Hughes states that a generic will be placed on “patent hold” until it has
either successfully dealt with the listed patents, the patents expire, or “as AstraZeneca
points out, the generic can demonstrate that it is not a ‘second person’ as
described in the Regulations and thus does not need to address the patents at
all”. Again, at paragraphs 59-60, Justice Hughes writes that:
[59] [...]
Section 5(1) of the NOC Regulations are specific in stating that a
person is only required to take steps to issue a notice of allegation to the
innovator who has listed patents (thus become a “second person”) if:
• that person has
filed for an NOC;
• that person has
compared reference or made reference to another drug;
• for the purposes of
demonstrating bioequivalence;
• and that other drug
has been marketed in Canada pursuant to an NOC; and
• there is a patent
list pertinent to that NOC.
[60] These
requirements are cumulative. Thus, if there is no comparison or reference for
the purpose of bioequivalence, section 5(1) is not triggered.
[74]
However, the Supreme Court did not frame this as an issue
of being a “second person”. Rather, Justice Binnie wrote at paragraph 39 that
“s. 5(1) of the NOC Regulations requires a patent-specific analysis,
i.e. the generic manufacturer is only required to address the cluster of
patents listed against submissions relevant to the NOC that gave rise to the
comparator drug”.
[75]
In Ferring, Justice Hughes was not asked to
consider, nor did he consider, whether his decision would strip Apotex of its
claim to damages under s. 8. Nor did he declare that Apotex was never a second
person or that some of the prohibition applications initiated by Ferring Inc.
or Sanofi, in that case, were void ab initio. In effect, all that
Justice Hughes decided was whether or not the Minister could issue an NOC in
the circumstances.
[76]
As I read these two decisions, the impact of AstraZeneca
(SCC) and Ferring is two-fold:
·
in respect of newly-initiated submissions for generic drug
approval under the PM (NOC) Regulations, a generic manufacturer is no
longer required to address certain patents on the Patent Register; in which
case, it will never be a second person vis-à-vis those patents; and
·
for prohibition applications commenced before the decisions
in AstraZeneca (SCC) and Ferring, and where certain patents on
the Patent Register do not now need to be addressed, the generic will
immediately receive its NOC (assuming that all other relevant patents have been
addressed), in which case, it will cease being a second person upon the
issuance of the NOC.
In no way do I interpret
AstraZeneca (SCC) and Ferring as stripping generic manufacturers
who have been kept off the market due to the actions of a brand company of
their right to claim s. 8 damages.
[77]
While Apotex raised the additional argument that the
doctrines of election and estoppel apply to prevent Sanofi from arguing that
Apotex was not a second person, I do not need to consider this argument given
my conclusion that Apotex was a second person in relation to the HOPE Patents.
[78]
I also note that Apotex was treated by the Minister as a
second person in relation to the HOPE Patents until December 12, 2006, when the
Minister decided to issue an NOC to Apotex. The Minister’s letter of December
8, 2006, in which it advised Apotex that it was not required to address the
HOPE Patents, does not contain any determination that Apotex was not a “second
person”. In the Minister’s letter of December 12, 2006 the Minister stated
simply that Apotex was “no longer considered to be a ‘second person’ in
respect of the '387 and '549 patents” [emphasis added]. This carefully-worded statement by the
Minister is, in my view, a correct interpretation of the teachings of the
Supreme Court in Astrazeneca (SCC).
[79]
Finally, rejecting June 27, 2006 as an end date is consistent
with the fact that s. 8 compensates a second person for the loss occasioned by
the operation of the statutory stay (see Alendronate (FCA), above at
para 71). In this case, Apotex did not receive an NOC until December 12, 2006.
The dismissal of the proceeding in T-1499-04 on June 27, 2006 did not allow
Apotex to enter the market at that time. Accordingly, June 27, 2006 cannot be
accepted as the end of the Relevant Period.
[80]
In sum, I am satisfied that there are no grounds to support
Sanofi’s view that June 27, 2006 is the “end date”.
(3) Alternative date: December 12, 2006
[81]
Both parties point to December 12, 2006 as an alternative
end date for the Relevant Period. Apotex notes that this is the date on which
it received an NOC for Apo-ramipril, while Sanofi submits that, if Apotex was a
second person, then its status as such terminated when it was no longer
required to address the HOPE Patents and received an NOC.
[82]
As discussed above, in my view, December 12, 2006 is the
correct end date for the Relevant Period.
C. Conclusion on Relevant Period
[83]
I find that the Relevant Period for the assessment of
Apotex’s losses is April 26, 2004 to December 12, 2006.
VI. Overall
Size of the Ramipril Market
[84]
Having determined the Relevant Period of April 26, 2004 to December
12, 2006, three major steps remain before I can begin an assessment of Apotex’s
Lost Profits:
1.
estimate the size of the total ramipril market during the
Relevant Period (i.e. the Ramipril Market);
2.
estimate the portion of the Ramipril Market that would have
been acquired by generic manufacturers during the Relevant Period (i.e. the
Generic Market); and
3.
estimate the share of the Generic Market that would have
accrued to Apotex.
[85]
The first step requires me to estimate the size of the
total ramipril market during this hypothetical period. Stated in different
terms, I must estimate the total number of capsules of
ramipril that would have been sold by all manufacturers during the Relevant
Period. In this task, I was assisted by two economists, Dr. Aidan
Hollis (produced by Apotex) and Dr. Robert Carbone (produced by Sanofi). Each
of these experts prepared forecasts to estimate the overall Ramipril Market,
the share of the market that would have been captured by the generic
manufacturers and Apotex’s share of that market. Dr. Hollis, in a simple and
effective diagram, depicted the general problem as follows:
[86]
In addition, I had the evidence of Dr. Iain Cockburn, whose
mandate was, as I see it, to do no more or less than to criticize Dr. Hollis’s
expert opinion. Dr. Cockburn made no estimates of the size of the Ramipril
Market. He had a myriad of criticisms of Dr. Hollis, referred to by Apotex in
final argument quite aptly as a “scorched earth attack”. Most of Dr. Cockburn’s
criticisms were addressed during the course of the testimony of all of the
experts.
[87]
We know what the actual sales for ALTACE were between April
26, 2004 and December 12, 2006. The key question is what impact the entry
of generic manufacturers, or “genericization”, would have had on those actual
sales. Dr. Carbone’s opinion on the size of the Ramipril Market reflects a
considerable impact of genericization, while Dr. Hollis concludes that the
“estimated effect of generic entry on ramipril sales is very modest” (Exhibit
44, vol 1 at para 37).
[88]
For each of a series of possible scenarios, Both Dr.
Carbone and Dr. Hollis began their tasks using the actual ramipril sales made
by Sanofi for the period April 26, 2004 to December 12, 2006. Both experts
used a time series forecasting model to estimate the sales that would have been
made after December 2006 in the absence of genericization. This enabled the
experts to come up with a “generic effect”. In his responding report, Dr.
Hollis explains the overall approach as follows (Exhibit 47 at para 16):
[Dr.] Carbone’s report and my report take a similar
approach in estimating the total ramipril volume during the damages period, had
a generic entry taken place in 2004. In order to do this, we both construct a
model to predict the likely sales volume of ramipril, had there been no generic
entry . . . in December 2006. We then compare these predicted values to the
actual values to try to estimate the effect of generic entry on total sales of
ramipril.
[89]
Although the overall approach of each expert was similar,
there were substantial differences in the details of their analyses.
[90]
Dr. Carbone’s methodology involves four phases:
·
Phase One: Dr. Carbone uses
market data for the period prior to the actual formulary listing of generic
ramipril to forecast the size of the ramipril market after December 12, 2006,
assuming that no generics ever entered the market.
·
Phase Two: Dr. Carbone next
subtracts the forecasted sales of ramipril after the formulary listing date
(the quantity forecasted in Phase One) from actual sales of ramipril after the
formulary listing date. He then divides this difference by the forecasted
sales. His calculation produces a series of “impact percentages” which
represent the impact of generic competition on the size of the ramipril market.
The results at this point of the analysis are presented in Table 7 of Dr.
Carbone’s report. He observes that generic competition reduced the size of the
ramipril market over time for all formulations, with the exception of the 1.25
mg strength (Exhibit 94, vol 1 at paras 65-66).
·
Phase Three: Dr. Carbone constructs
an “impact model” using a process called Bass Diffusion modelling (the Impact
Model). This technique estimates the change in ramipril sales over time based
on how demand reacts to influences on product diffusion such as advertising,
media coverage and word of mouth by customers already using the product (Exhibit
94, vol 1 at Appendix J). The purpose of the Impact
Model is to predict the (negative) linear trend in ramipril market size based
on the impact percentages calculated in Phase Two (see Exhibit 94, vol 1 at
Appendix K for the modelling results.)
·
Phase Four: Dr. Carbone subtracts
the values generated by the Impact Model from the size of the ramipril market
(without generic competition) forecasted in Phase One. The result is the total
forecasted size of the Ramipril Market over the Relevant Period.
[91]
Overall, Dr. Carbone concludes that there would be a
significant reduction in the size of the Ramipril Market during the Relevant
Period. Dr. Carbone attributes much of this reduction to the cessation of
promotion by Sanofi (Exhibit 94, vol 1 at paras 67-70). This particular factor,
however, may not be as relevant in the case of ramipril as it is in other
instances.
[92]
While I accept that an innovator will usually stop
promoting a product after its genericization, this did not happen immediately
or completely in the case of ALTACE. As acknowledged by Mr. Benoit Gravel,
Sanofi’s vice president of sales, promotion of ALTACE continued until the end
of March 2007 – some three months after genericization. In addition, Sanofi
introduced a combination formulation – ALTACE HCT – into the market in November
2006 and continued promoting that product after the genericization of ramipril.
Dr. Carbone agreed that there would be a benefit to ALTACE sales arising from
promotion of ALTACE HCT.
[93]
In addition, Dr. Carbone’s rationale does not accord with a
recently published report of the Patented Medicine Prices Review Board (PMPRB),
entitled “The Impact of Generic Entry on the Utilization of the Ingredient”,
September 2011 (the PMPRB Report) (Exhibit 48). The purpose of the study
carried out by the PMPRB was to determine whether, upon genericization, a drug
continues to be utilized to the same extent. This is exactly the question that
Dr. Carbone and Dr. Hollis addressed in their reports. The authors of the
PMPRB Report studied seven top‑selling – “blockbuster” – drugs that had
lost patent protection in the period between 2000 and 2006. The conclusion of
the PMPRB at page 25 was:
Generally, this
research shows that there is very little change in the trend in utilization
once the first generic version is launched. Typically the number of claims and
market share following generic entry continue the trend established by the brand
name under market exclusivity. In most cases the changes in utilization that
are identified cannot be directly and/or solely attributable to generic entry.(PMPRB Report, above
at 25).
[94]
During his oral testimony, Dr. Carbone expressed very
negative views of every aspect of the PMPRB Report. I give little weight to his
criticisms, most of which were seriously undermined during cross-examination.
In any event, I am not relying on the PMPRB Report as the foundation of my
decision on the Ramipril Market. Rather, the PMPRB Report simply shows that,
directionally, the conclusion of Dr. Hollis is preferable to that of Dr.
Carbone.
[95]
One of the more significant criticisms of Dr. Carbone’s
work was of his use of the proprietary “Futurcast” system to forecast the size
of the non-genericized ramipril market after December 2006 in Phase One of his
analysis. Dr. Hollis opined that, while Futurcast might be perfectly
appropriate “to prognosticate into the future”, where one is attempting simply
to predict future sales of ramipril, the software is not as useful to predict
“sales in the past”, because it fails to take into account actual information
over the Relevant Period that is pertinent to the analysis. This, says Dr.
Hollis, “constrains the utility and accuracy of Futurcast” (Exhibit 47 at para
18).
[96]
In contrast, Dr. Hollis’s approach was much simpler and did
not rely on proprietary software.
[97]
At this stage of his analysis, Dr. Hollis uses an
econometric (regression analysis) model to estimate the Ramipril Market size in
the “but for” world pre-December 2006 using nationally aggregated data from the
actual ramipril market in and after December 2006. He finds that his model
estimates track very closely the actual effects of genericization post-December
2006. Dr. Hollis also performs an “alternative” modelling analysis using
time series and provincially disaggregated data as a “check” on his preferred
method. In contrast to Dr. Carbone, Dr. Hollis also uses additional
available data – including the total volume of sales of other ACE inhibitors –
to refine his forecasts.
[98]
The simplicity of Dr. Hollis’s approach at this stage has
much to recommend it. Rather than attempting to construct a complex econometric
model that might account for the direct influence of advertising behaviour,
diffusion of product information, the introduction of alternative formulations
(such as ATLACE HCT) or other explanatory variables, Dr. Hollis uses a
relatively simple model based on the assumption that the overall Ramipril
Market can be predicted mainly by the time after generic entry. Given the
accuracy of this simple model in predicting trends in the real world ramipril
market, and absent any apparent differences in the “but for” world that would
affect total market size, it is likely unnecessary to include more explanatory
variables to create a reliable model.
[99]
One potential drawback of Dr. Hollis’s approach is that he
cannot distinguish the individual causal factors that drive changes in the size
of the Ramipril Market. This “drawback”, however, is significant only if there
is reason to believe that some factor would have operated in the “but for”
world that did not operate in the real world, or vice versa. Since the goal
here is not to explain market dynamics but to make an accurate quantitative
prediction about the “but for” world, I do not see this “drawback” as a reason
to reject Dr. Hollis’ approach.
[100]
Another criticism of Dr. Hollis’s analysis was his use of
national data. In the face of vigorous cross-examination, Dr. Hollis was clear
and consistent in defending his approach. Dr. Hollis explained that the
use of national data is appropriate where there is nothing “very different”
happening across provinces between the Relevant Period and the time period that
was being modelled after December 2006:
So if there isn’t, in
fact, a substantial change in what’s happening across the provinces, if you
expect things are going to be materially the same, there is no reason to add
extra complication by worrying about the provinces.
[101]
Nonetheless, it is significant that, although Dr. Hollis’s
estimates closely track real world trends at the aggregate national level, they
diverge to a greater degree once they are disaggregated at the provincial
level. For example, Dr. Hollis’s graphs showing the predicted and actual doses
of ramipril for Prince Edward Island and Saskatchewan demonstrate that the
“fitted values” tend to be a poorer predictor of the growth of the Ramipril
Market post-generic entry at the provincial level than they are at the national
level (Exhibit 44, vol 1 at Tab 6).
[102]
My final criticism of Dr. Hollis’s approach at this stage
concerns his explanation of his model’s “conservative” estimate of the effect
of generic entry on the Ramipril Market (Exhibit 44, vol 1 at para 38):
In this model, greater
sales of other ACE inhibitors is found to increase ramipril sales. Thus, if
there is a reduction in promotion of ramipril after generic entry that leads to
an increase in sales of other ACE inhibitors, the effect will be to increase
the predicted values, relative to actual values. And this will, by definition,
increase … the estimated impact of genericization on ramipril sales.
[103]
This explanation, in my view, is entirely speculative. Dr.
Hollis has provided no evidence to suggest that a reduction in the promotion of
ramipril, if it occurred in the real world, would cause sales of other ACE
inhibitors to increase. He offers no justification for presuming that his
estimates are conservative on this basis.
[104] Even with the above concerns in mind, I am satisfied that Dr. Hollis’s
analysis represents a sound approach to predicting the size of the Ramipril
Market in the “but for” world. I prefer his model – and hence his results – to
quantify the size of the Ramipril Market.
VII. Size of the Generic
Market
[105] Having
determined the size of the overall Ramipril Market for the Relevant Period, I
must now establish the size of the Generic Market. The notion that generics
will acquire a portion of the Ramipril Market is described as “market
penetration”. Looking at this from Sanofi’s perspective, Dr. Carbone referred
to this as market “erosion”. Stated in different terms, the issue is to
determine how ALTACE and generic versions of ramipril would have shared the
Ramipril Market.
[106]
Once again, Drs. Hollis, Carbone and Cockburn provided
expert opinions on this step of the analysis. The
different views of the parties on the size of the Generic Market appear to
relate to two significant areas: (1) modelling to account for the degree of
market penetration; and (2) timing of formulary listings. I will consider each
issue in turn.
A. Market
penetration
[107]
In predicting the size of the Generic Market, Dr. Hollis
uses the same conceptual foundation as he applied to determine the size of the
Ramipril Market. In other words, Dr. Hollis begins with the presumption that
the available observed data from the ramipril market post-generic entry is an
accurate predictor of the “but for” world, unless there is good reason to
believe that there are significant differences between these two worlds.
[108]
Although Dr. Hollis’s opinion was harshly criticized, I see
little reason to doubt either his approach or his conclusion that the
regulatory conditions would have been the same or very similar between the real
world and the “but for” world.
[109]
Dr. Carbone’s prediction of erosion in the “but for” world
differs from Dr. Hollis in one significant regard. In particular, Dr. Carbone
opines that markets with fewer generic entrants will demonstrate a slower
erosion of the brand name manufacturer’s market share. The basis of
Dr. Carbone’s opinion is a multivariate regression analysis set out at
Appendix O to his report (see Exhibit 94, vol 1 at Tab O). I agree with Apotex
that there are “methodological and logical flaws that make the[se] results
difficult to rely upon”.
[110]
A serious problem that I have with Dr. Carbone’s opinion is
his use of over 100 “Bass Diffusion” estimations. These estimations were
performed with a formula that was not provided in his reports. Indeed, as we
discovered in Dr. Carbone’s cross-examination, the Bass Diffusion formula set
out in Appendix J to his report was not the formula that he (or his assistants)
actually used to reach his predictions.
[111]
A further problem arose with the number of discrepancies
found in the coefficients contained in Appendix P when compared with the
coefficients for the same molecules in Appendices L and N. As pointed out by
Apotex:
For the 53 of 112
values where the source variables were provided, at least 8 (or 15%) of the
values did not correspond to the values in the source Appendices.
[112]
These and other problems identified by Apotex give me
reason to discount – at least to some extent – the final opinions of Dr.
Carbone on the important question of erosion during the Relevant Period.
[113] That is not to say that Dr. Hollis’s evidence is perfect. Dr. Hollis
responds, in part, to Dr. Carbone’s claim that markets with fewer generic
entrants will demonstrate a slower erosion by indicating that Dr. Carbone has
inferred a causal relationship between the number of generic entrants and the
erosion rate that is not supported by the evidence. Dr. Hollis posits that the
important explanatory variable is not the number of entrants but the size of
the overall market. He argues that larger markets can be expected to generate
faster erosion rates, which may explain the differences observed by Dr.
Carbone. In my view, Dr. Hollis’s presumed causal relationship between market
size and the erosion rate is subject to the same criticism that he directs at
Dr. Carbone. Dr. Hollis performs no statistical analysis to test his
hypothesis. He could have attempted to model this relationship in his
responding report, but he opted not to do so.
[114]
Having said that, I find Dr. Hollis’s logic that, in
effect, other determinants beyond the number of generics in the market must
explain why some drug markets support multiple generic entrants, while others
motivate only one generic entrant, to be reasonably persuasive. Some “other”
factor or factors must underlie generic manufacturers’ choices to enter a given
market or not. It follows that Dr. Carbone’s attempt to distinguish between the
“but for” world and the real world based on the number of generic entrants is
at best an incomplete explanation and, in my view, insufficient to deviate from
the baseline assumption that the size of the generic market in the real world
is a good predictor of that market in the “but for” world.
B. Formulary
listings
[115] The second substantial area of difference between the parties is the
question of formulary listing dates. Dr. Hollis carried out his own inquiries
into formulary listing dates. In contrast, Dr. Carbone relied on the
formulary listing dates provided to him by Mr. Palmer. Mr. Palmer’s assumptions
differed from Dr. Hollis’s.
[116]
As stated by Dr. Hollis in his Expert Report, “[a]n
important determinant of sales of pharmaceuticals in Canada is
listing on provincial formularies” (Exhibit 44, vol 1 at
para 62). Dr. Hollis explained why this was so (Exhibit 44 at paras 62-63 [footnotes omitted]):
62. [. . .] The reason for this is
that, generally speaking, across Canada, the cost of prescription
pharmaceuticals for seniors and the indigent is covered by the provincial drug
benefit plans once the drugs achieve formulary listing. Collectively, this
makes the provincial drug benefit plans the largest payers for prescription
pharmaceuticals, accounting for approximately 40% of expenditures in 2005. For
generic pharmaceuticals, the listing on formularies is even more critical
because, typically, formulary listings permit and, in some cases require, the
pharmacist to substitute a lower-priced generic version of the branded drug
product.
63. For ramipril, as for most other drugs, retail
sales increase markedly when listed on provincial formularies (although
significant wholesale sales can occur before then in the expectation of
formulary listings). In constructing the hypothetical sales data, therefore, it
is important to account for the likely listing dates on provincial formularies.
[. . . .]
[117]
Dr. Hollis’s analysis of this factor is again built on the
reasonable assumption that the available observed data from the ramipril market
post-generic entry is an accurate predictor of the “but for” world unless there
is good reason to believe that there are significant differences between the
two worlds. Dr. Hollis “tested” this assumption against the average listing
delays for Apotex products for the period 2004 to 2006.
[118]
Having carried out his analysis, Dr. Hollis concludes that:
(1) there is no significant difference between the average speed of formulary
listing in 2004 as compared to 2006; (2) there is no difference in the average
number of days on which the provinces approved Apotex products between 2004 and
2006; and (3) the same speed of approval for Apo-ramipril as actually happened
would have been possible in the “but for” world, since the relevant committees
approved other drugs at about the same time as they could have approved
Apo-ramipril (Exhibit 44, vol 1 at para 70).
[119]
Again, I see no reason to doubt Dr. Hollis’s conclusion
that the regulatory conditions would have been the same or very similar between
the real world and the “but for” world, leading to approximately the same times
to formulary listings.
[120]
In final argument, Sanofi disputed only one aspect of Dr.
Hollis’s formulary listing dates. Specifically, Sanofi pointed to a portion of
Mr. Palmer’s testimony where he opined that:
I do note the date
listed for BC is January 27th, and that is the correct date. That is different
from the date that Dr. Hollis had in his report, which is I believe April.
[121]
On the basis of this one statement, Sanofi suggests that
“applying Dr. Hollis’s approach starts erosion 86 days (or almost 3 months)
early in British Columbia, thus over-stating the
generic share of the market”. In reaching this conclusion, Sanofi ignores the
balance of the cross-examination of Mr. Palmer, where he acknowledged that his
revised estimates for Alberta and British Columbia were, in fact, consistent with those of Dr. Hollis.
[122]
I therefore accept, as reasonable and more probable than
not, the formulary listing dates developed and relied on by Dr. Hollis in
assessing the size of the Generic Market.
C. Conclusion on Generic Market
[123]
In conclusion on the Generic Market, I accept the analysis
of Dr. Hollis over that of Dr. Carbone and, where applicable, Mr. Palmer.
Of particular significance to the next step of the analysis (establishing
Apotex’s share of the Generic Market), I accept Dr. Hollis’s opinion that the
size of the Generic Market would not be materially impacted by the number of
generic entrants or the timing of their entry. To the extent that it is
necessary to predict formulary entries in the hypothetical world, I would apply
the estimates used by Dr. Hollis over those used by Mr. Palmer.
VIII. Apotex’s
Share of the Generic Market
[124]
The next step in the analysis is to determine Apotex’s
share of the Generic Market for ramipril. This assessment provides an estimate
of Apotex’s Lost Volumes over the Relevant Period.
[125]
I begin by observing that Sanofi does not argue that Apotex
would have been unable to produce sufficient quantities of generic ramipril to
supply whatever market share it would have acquired in the “but for” world. The
evidence before me is clear and compelling that Apotex would have had the means
to obtain sufficient quantities of API and incipient ingredients and sufficient
plant capacity to meet market demand throughout the Relevant Period. The only
caveat is that Apotex may have been required to add further machinery or
to run additional shifts. This could result in higher production costs, a
matter considered later in these Reasons.
[126]
Having found that Apotex had the physical capacity to have
supplied the entire market, the question is whether it would have done so alone
or with competitors.
[127]
Before I begin a detailed assessment of this issue, Sanofi
raises a question that must be addressed. Should Apotex’s share of the Generic
Market be assessed on the basis of one “but for” world? Once I have responded
to this question, I will move to the next steps of determining:
·
the participants in the hypothetical world and the timing
of their entries; and
·
the market share of Apotex (i.e. Apotex’s Lost Volumes).
A. Sanofi’s view of the “but for” world
[128]
Sanofi asserts that there can be only one “but for” world that
should apply to all s. 8 claims for ramipril. In Sanofi’s view, a second person
would receive an inappropriate windfall if all other generics were not
considered. Sanofi explains that, absent a single finding as to the overall
generic market in the “but for” world, the result could be “absurd
consequences”.
[129]
Sanofi attempted to illustrate the potential “absurdity” of
Apotex’s position through the use of a hypothetical example. In the example,
Sanofi assumed three generic manufacturers, with A approvable at year 0, B
approvable at year 1 and C approvable at year 2. The other assumptions were:
·
total generic market of 20 units/year;
·
NOC proceedings against each generic;
·
NOC proceeding against A dismissed at year 3;
·
all generics enter at year 3; and
·
each generic advances a s. 8 claim.
[130]
If each of A, B and C commence an action for recovery of
damages under s. 8 and if the claims are assessed as three independent “but
for” markets with no other hypothetical entrants, the results, as posited by
Sanofi, would be the following:
·
A would claim three years at 20 units/year for a total of
60;
·
B would claim two years at 20 units/year for a total of 40
units; and
·
C would claim one year at 20 units/year for a total of 20
units.
[131]
This would constitute a total generic recovery of 120
units, whereas the total generic market over that period would be only 60
units. The result would be that Sanofi’s liability would be double that which
would rationally be possible.
[132]
I do not disagree with Sanofi’s arithmetic. I also acknowledge
that, if this were to happen, the result would be, if not “absurd”, at least questionable. That said, Sanofi’s argument contains
a number of flaws.
[133]
The first issue that I take with Sanofi’s argument is that
it misrepresents Apotex’s position. Apotex is not arguing that the hypothetical
world under the Regulations must consider Apotex to be a sole-source
manufacturer with no competitors throughout the Relevant Period. Rather, as I
understand Apotex’s argument, Apotex is submitting that other entrants in the
market must be considered on a case-by-case basis.
[134]
While I agree with Sanofi that the “but for” world must
consider the inclusion of potential competitors, I do not go so far as Sanofi
asserts. In other words, I reject Sanofi’s urging that I establish one “but
for” world that will apply in this case and in any others involving the
genericization of ramipril.
[135]
The assessment of damages can and should be made on the
facts of each case. To the extent that there are common elements that impact on
the quantification of damages, these will more likely than not come forth
during the trial.
[136]
Another serious flaw in Sanofi’s argument is that the
evidence in one case may establish a different Relevant Period than in another
case. This will impact on many elements of the assessment of damages. In this
case, for example, I have determined that Apotex would have entered the market
on April 26, 2004. This finding means that different considerations will come
into play with respect to the possible entry of an authorized generic than if I
had concluded that an entry date of December 13, 2005 was more appropriate. In
the companion Teva case (Court File No. T-1161-07), I have concluded that a
different Relevant Period is established and different considerations were
relevant. Following Sanofi’s urging would accordingly require that I disregard
evidence in either Teva’s case or this one.
[137]
By their very nature, damages in this action are
hypothetical. It follows that estimates must be made and a market constructed
that will not be perfect. As I re‑write history, hypotheses must be
constructed and evaluated. Those hypotheses will necessarily change depending
on the facts of each case. I am striving to be reasonable and fair – I
cannot achieve perfection. As pointed out by Lord Shaw in Watson, Laidlaw
& Co Ld v Pott, Cassels, and Williamson (1914), 31 RPC 104 at 118 (HL):
The restoration by way
of compensation is therefore accomplished to a large extent by the exercise of
a sound imagination and the practice of the broad axe.
[138]
With respect to ramipril, Sanofi has identified only Teva,
Apotex and Riva as participants in the “but for” world. I am quite certain that
the damages in those three actions will not be greatly – if at all – in excess
of the award of damages that would be made had the three cases been joined and
one “but for” world established. Since Sanofi is the defendant in all three
cases, it is well aware of the total damages being claimed. If that amount
raised a real threat that Sanofi’s total liability would exceed the bounds of
rationality, Sanofi could urge the Court to consider an adjustment to the
compensation pursuant to s. 8(5) of the Regulations.
[139]
There may be a situation where Sanofi’s fear has some
merit. It certainly is not this case.
B. Other generics
[140]
Having determined that the particular facts before me will
inform my conclusions on this issue, the next task is to establish which
generics would have entered the market during the Relevant Period and when they
would have entered. The parties put forward evidence related to three possible
entrants into the “but for” market – Teva, Riva and an authorized generic.
[141]
Before I examine the individual market entrants, I note
some disagreement as to the burden.
[142]
While both parties agree that Apotex bears the burden of
proving its losses, they disagree on the specific issue of generic competition
in the “but for” world. For its part, Sanofi says that Apotex “must prove on a
balance of probabilities what position it would have been in ‘but for’ the
prohibition proceedings”, including that it would have been able to obtain an
NOC and enter the ramipril market in the period alleged, that no AG or other
generic would have done so until the dates alleged, and the pricing of its
product.
[143]
Apotex submits that Sanofi bears the burden of proving
“affirmative defences” and that Apotex is not required to disprove those
defences. In particular, Apotex submits that Sanofi bears the burden of proving
that any generic competitors or an AG would have entered the hypothetical market.
[144]
It is trite law that there are both legal and evidential
burdens in a trial. In Hoffmann-La Roche Ltd v Canada
(Minister of National Health and Welfare) (1996), 205 NR 331 at
para 8(3) (CA), 205 NR 331, the Court of Appeal
explained that,
[The primary] burden, known in a civil case as either the
“persuasive burden” or the “legal burden”, is the burden of establishing a case
to the civil standard of proof. By contrast, the “evidential burden” consists
of the burden of putting an issue in play and means that a party has the
responsibility to ensure that there is sufficient evidence of the existence or
nonexistence of a fact or an issue on the record to pass the threshold for that
particular fact or issue.
[145]
The legal burden does not shift over the course of a trial,
but the evidential burden does. Once there is “prima facie proof or
presumption of the truth of an allegation, which ought therefore to be found
true in the absence of further evidence”, the evidential burden shifts to a
defendant “to adduce evidence in answer to the prima facie proof” (Ontario
Equitable Life and Accident Insurance Co v Baker, [1926] S.C.R. 297 at 308-309
[Baker]). At the end of a case, the court must weigh all of the evidence
called by both parties (Baker, above).
[146]
It cannot be Apotex’s burden to call every potential
participant in the market and prove that they would not have launched a
ramipril product. In this regard, the position of Apotex is much like the
situation faced by the defendant in Rainbow Industrial Caterers Ltd v
Canadian National Railway Co, [1991] 3 S.C.R. 3, [1991] SCJ No 67. In that case, involving an action in tort, Rainbow was seeking damages
from Canadian National Railway (CN) for, inter alia, negligent
misrepresentation in respect of a catering contract. CN argued that the claimed
loss was not all attributable to CN since Rainbow would have entered into other
contracts. On the issue of burden, at page 15, the Supreme Court states that:
Once the loss
occasioned by the transaction is established, the plaintiff has discharged the
burden of proof with respect to damages. A defendant who alleges that a
plaintiff would have entered into a transaction on different terms sets up a
new issue. It is an issue that requires the court to speculate as to what
would have happened in a hypothetical situation. It is an area in which it is
usually impossible to adduce concrete evidence. In the absence of evidence to
support a finding on this issue, should the plaintiff or defendant bear the
risk of non-persuasion? Must the plaintiff negate all speculative hypotheses
about his position if the defendant had not committed a tort or must the
tortfeasor who sets up this hypothetical situation establish it?
[Emphasis added]
[147]
In holding that CN bore the burden, the Supreme Court
comments as follows, at page 16:
The appellant CN
alleged that the loss was not all attributable to the misrepresentation because
Rainbow would have entered into a different contract on other terms which would
have resulted in at least some of the loss. What the respondent would have done
had it not been for the tortious act requires a great deal of speculation, and,
on the basis of the principles which I have reviewed above, I would apply the
legal burden of proof against the appellant.
[148]
In the context of an undertaking in damages, a successful
defendant bears the onus of proving its loss (see e.g. Les Laboratoires
Servier and another v Apotex Inc and others, [2008] EWHC 2347 (Ch) (QL),
[2008] All ER (D) 79 (Oct), rev’d on other grounds [2010] EWCA Civ 279,
[2010] All ER (D) 238 (Nov)).
However, in Algonquin Mercantile Corp v Dart Industries
Canada Ltd (1987), [1988] 2 FC 305 (QL) at para 8 (CA), 79 NR 305, the
Federal Court of Appeal held that an unsuccessful plaintiff bears the burden of
proving cannibalization where that issue was raised by the plaintiff:
[T]he existence of cannibalization was a question which was
introduced into the debate as a result of plaintiff’s allegations. Accordingly,
it was for plaintiff to prove it, not for defendant to show, as the Trial Judge
said, that it “would not have occurred”.
[149]
Taking all of this into consideration, in my view, the
proper approach is the following: Once Apotex has led prima facie
evidence of its losses, the evidential burden shifts to Sanofi to adduce
evidence in response. Sanofi cannot simply allege that other generics would
have entered the market without leading evidence in support of such assertions.
[150]
In this case, Apotex does not, at least initially, bear the
burden of disproving the hypothetical sales of third party generics. However,
Apotex, at all times, bears the legal burden of proving its losses, and the
evidence adduced by Apotex must ultimately be weighed against any evidence
adduced by Sanofi establishing sales by third party generics. To the extent
that Sanofi succeeds in discharging its evidential burden by proving third
party sales, Apotex must address that evidence in order to discharge its legal
burden.
(1) Teva
[151]
Sanofi produced evidence and asserts that Teva would have
been part of the Generic Market.
[152]
There are two aspects to the question of whether Teva would
have commenced sales in the hypothetical world. The first step is to examine
the regulatory context to determine whether there were regulatory impediments
to Teva’s entry. The analysis of this question will lead to a determination of
the possible entry date for Teva’s launch of a generic version of ramipril. As
a second step, I must address the practical considerations that would have
likely arisen as of the hypothetical entry date. Such matters as plant capacity,
access to API and motivation are relevant at this second step.
[153]
Neither Apotex nor Sanofi questions Teva’s physical
capacity to manufacture ramipril during the Relevant Period. Thus, the focus of
my examination is on the first step of the analysis. In other words, the
determinative question is: From a regulatory and legal perspective, would Teva
have been able to come to market and, if so, when?
[154]
Sanofi subpoenaed Mr. Barry Fishman, the president and
chief executive officer of Teva, to speak to Teva’s potential participation in
the hypothetical market. As confirmed by Mr. Fishman, the following steps
were taken by Teva to come to market:
·
Teva filed its ANDS on December 24, 2001 (Exhibit 126, Tab
3);
·
Teva’s certified “patent hold” date, limited to 2.5, 5 and
10 mg ramipril capsules, was October 14, 2003, as of which date, Teva had
agreed to wait until the expiry of the '457
Patent on December 13, 2005 (Exhibit 126, Tabs 6, 8);
·
on September 12, 2005 ('206 Patent) and September 14, 2005 ('089, '948, '549 and '387 Patents), Teva served notices of allegation;
·
on October 31, 2005 (Court File No. T-1965-05, with respect
to '206 Patent) and November 2,
2005 (Court File No. T-1979-05, with respect to '089, '948, '549 and '387 Patents), Sanofi
commenced prohibition applications;
·
on September 25, 2006, the Federal Court dismissed
T-1965-05 as “an abuse of process” (Sanofi-Aventis Canada Inc v
Novopharm Limited, 2006 FC 1135, 306
FTR 56);
·
on December 8, 2006, the Minister advised that Teva was
required to address the '089 and '948 Patents but not the '549
or '387 Patents;
·
on December 15, 2006, Teva withdrew the portion of its
notice of allegation relating to the '549
and '387 Patents;
·
on April 27, 2007, the Federal Court of Appeal dismissed
T-1979-05 as an abuse of process (Sanofi-Aventis Canada Inc v
Novopharm Limited, 2007 FCA 167, rev’g 2006 FC 1547);
and
·
on May 2, 2007, Teva received its NOC.
[155]
One very relevant fact related to Teva is that, as part of
its ANDS, Teva certified that it would await expiry of the '457 Patent. Moreover, when it finally served its notices of allegation on
Sanofi in September 2005, it did not address the '457 Patent. In other words, the record demonstrates that Teva was not
seeking regulatory approval to enter the market before December 13, 2005,
when the '457 Patent expired. Thus,
Teva’s earliest possible entry date in this case is December 13, 2005.
[156]
Apotex submits that Teva would not have been able to come
to market until the end of October 2007. In the actual world, Teva effectively
“unlocked” the regulatory door by following on the footsteps of Apotex. Apotex
explains that the “but for” world should be constructed on the basis that
Sanofi had not commenced prohibition proceedings against Apotex. For that
starting point, Apotex points to the Federal Court of Appeal decision in Norfloxacin
(FCA), above at paragraph 75, where the Court of Appeal noted that,
[I]t must be
remembered that the Federal Court had to assess Apotex’s damages on the basis
of a hypothetical question: what would have happened had Merck not brought an
application for prohibition?
[157]
Applied to this case, this means that I must assess the s.
8 damages on the basis that Sanofi had not commenced T-1742-03, in respect of
Apotex’s notice of allegation for the '206
Patent. Apotex correctly observes that, if Sanofi had not commenced T-1742-03,
there would have been no decision and order of Justice Mactavish in Ramipril
NOC #1 (FC). In the real world, Teva was able to leverage Apotex’s success
in T-1742-03 to have T-1965-05 dismissed as an abuse of process. If there had
been no decision in Ramipril NOC #1 (FC), Teva would not have been able
to have T-1965-05 dismissed as an abuse of process. Teva would have had to
pursue its arguments, in T-1965-05, that its allegations of invalidity of the '206 Patent were justified. As acknowledged by Mr. Fishman, Teva, in those
circumstances, would only have expected to receive its NOC towards the end of
the 24-month stay period in T-1965-05; that is, towards the end of October
2007.
[158]
Conceptually, the logic of Apotex’s argument is
inescapable. It is not only Teva’s reluctance to address the '457 Patent, but its delay in serving a notice of allegation with respect
to the '206 Patent, that creates an
impediment to its market entry into the “but for” world.
[159]
The only problem that I can see with Apotex’s argument is
that it has ignored the NOC proceedings of Riva. Riva served its notice of
allegation with respect to the '206, '457 and '089 Patents on June 10, 2004,
with Sanofi commencing T-1384-04 on July 23, 2004. Without Apotex’s challenge
to the '206 Patent in T-1742-03, the
NOC proceedings in T-1384-04 would have likely unfolded in the ordinary course
with the result that the Court would have dismissed Sanofi’s prohibition
application no later than July 2006. I am assuming that the result of the Riva
NOC proceedings would have been that Riva’s allegation of invalidity of the '206 Patent would have been justified – a reasonable assumption given what
happened in the real world. Teva would then have been able to leverage Riva’s
success to obtain a dismissal of T-1965-05 as an abuse of process within a few
days of the dismissal of T-1384-04. As in the real world, a success by Riva
would likely have set into motion a sequence of events that would have resulted
in Teva receiving an NOC very shortly thereafter, in spite of the continued
existence of other use patents.
[160] In sum, for purposes of this action, I will assume that Teva would have
been able to come to market on or about August 1, 2006.
(2) Riva
[161]
Some scenarios discussed during the trial involved the
possible participation of Riva in the “but for” world. As with the possible
participation of Teva in the hypothetical market, there are two aspects to
Riva’s entry during the Relevant Period. The first step is to examine the
regulatory context to determine whether there were regulatory impediments to
Riva’s entry. The second step is to address the practical considerations that
would have likely arisen as of the hypothetical entry date. Such matters as
plant capacity, access to API and motivation are relevant at this second step.
[162]
The evidence before me establishes that Riva could not have
received an NOC for its ramipril during the Relevant Period. Thus, even if Riva
could have made arrangements to market ramipril in some or all parts of Canada, it could not have come to market in the Relevant Period
due to a regulatory impediment.
[163]
Riva’s path to regulatory approval of its generic ramipril
in the real world was as follows:
·
on June 8, 2004, Riva submitted its ANDS to Health Canada for ramipril capsules, cross referencing Pharmascience’s
ANDS (Exhibit 107, Tabs 74 and 66 at 3);
·
on June 10, 2004, Riva served a notice of allegation with
respect to the '206, '457 and '089 Patents;
·
Riva’s “patent hold” date was June 18, 2004;
·
on July 23, 2004, Sanofi commenced a prohibition
application (Court File No. T-1384-04) in response to Riva’s notice of
allegation regarding the '206, '457 and '089 Patents;
·
on September 8, 2004, Riva served a notice of allegation
with respect to the '948 Patent;
·
on October 22, 2004, Sanofi commenced a prohibition
application (Court File No. T-1888-04) regarding the '948 Patent;
·
on December 5, 2006, Riva served a notice of allegation
with respect to the '549 and '387 Patents;
·
on January 19, 2007, Sanofi commenced a prohibition
application (Court File No. T-127-07) regarding the '549 and '387 Patents;
·
on May 17, 2007, the Court dismissed T-1384-04 and
T-1888-04 (Sanofi-Aventis Inc v Laboratoire Riva Inc, 2007 FC 532, 315
FTR 59);
·
on March 4, 2008, the Court dismissed T-127-07 (Sanofi-Aventis
Canada Inc v Laboratoire Riva Inc, 2008 FC 291, 331
FTR 259); and
·
Riva received its NOC on March 14, 2008.
[164]
A serious problem arose for Riva from its decision to
cross-reference the submission of Pharmascience for pms-ramipril.
[165]
As confirmed by Ms. Bowes, Health Canada’s policy and its advice to Riva was that it would not receive an NOC for
ramipril in advance of Pharmascience obtaining its NOC. Riva was informed of
this regulatory hurdle in a letter from Health Canada dated April 24, 2007 (Exhibit 107, Tab 66 at 4) which advised Riva as
follows:
[W]e would note that,
since Riva’s submission has been cross-referenced with another submission, the
NOC will not be issued to Riva until the NOC is issued for the cross-referenced
submission for pms-ramipril, in accordance with Health Canada’s Policy
entitled “Filing of Supplemental New Drug Submissions, Supplemental Abbreviated
New Drug Submissions, Notifiable Changes and Cross-Referenced Submissions”.
[166]
On May 24, 2007, Riva brought an application for judicial
review of Heath Canada’s
decision (Exhibit 107, Tab 67; Court File No. T-896-07). Health Canada subsequently reversed its position and, in a letter dated
June 21, 2007 (Exhibit 107, Tab 69), Riva was informed as follows:
[. . .]
Health Canada is no longer of the view that Riva cannot receive a notice
of compliance until such time as the Pharmascience submission to which Riva’s
product is ‘cross-referenced’ is itself approved. As a result, should Riva
ultimately be successful in the prohibition proceedings ongoing in T-127-07,
and otherwise meet all of its obligations under the Patented Medicines
(Notice of Compliance) Regulations, it will be eligible to receive a notice
of compliance, regardless of whether the Pharmascience submission has fully
complied with the NOC Regulations and received a notice of compliance.
[167]
Riva accordingly discontinued its application for judicial
review. However, the fact is that, separate and apart from any notice of
allegation proceedings under the Regulations, Riva could not have
entered the ramipril market before Health Canada changed its position on Riva’s cross-referenced ANDS. The earliest that
Riva could have obtained an NOC is June 21, 2007, after the end of the Relevant
Period in this trial. Neither Sanofi nor Apotex presented evidence or argument
that Riva could have entered the market during the Relevant Period.
[168] Given the facts before me in this case, I conclude that Riva would not
have been a participant in the Generic Market during the Relevant Period.
(3) Authorized
generic
[169]
The final potential market entrant is an authorized generic
manufacturer. In this trial, Sanofi submits that it would have launched an
authorized generic, or AG, simultaneously with – or shortly after – Apotex’s
launch of Apo-ramipril in the “but for” world.
[170]
As described by a number of witnesses, the term “authorized
generic” refers to a drug that is manufactured by an innovative drug company –
in this case, Sanofi – but sold by a generic company under the generic’s name.
While the composition of the authorized generic product is identical to the
innovator’s product, it has a separate DIN and NOC. Authorized generics obtain
regulatory approval by submitting an administrative NDS instead of an ANDS. No
bioequivalence study is required, as the innovator manufactures the authorized
generic product. As a result, AGs quickly obtain approval. The generic company
simply files an administrative NDS referencing and relying upon the innovator’s
submission, and the innovator provides a letter of access that allows the
authorized generic to cross-reference its submissions.
[171] The introduction of an AG allows an innovator to participate in both the
brand and generic markets, as the innovator effectively sells two distinct, but
identical products. The brand company can thus recoup some of the market that
has been lost to generics. It is obvious that a brand company will not
introduce a generic until and unless there is an unauthorized or “true” generic
manufacturer coming on to the market. Otherwise, the AG would only have the
result of cannibalizing sales of the brand drug.
[172] When ramipril was finally genericized in late 2006, the first market
entrant was ratiopharm inc. (ratiopharm) operating as an AG of Sanofi.
Anticipating a generic entrant, Sanofi had entered into an agreement with
ratiopharm allowing it to rely on Sanofi’s regulatory filings to obtain an NOC
ahead of the pack. Would Sanofi have launched an authorized generic during the
Relevant Period?
[173] Apotex
argues against the inclusion of an AG in the hypothetical world. Its key
arguments, in brief, are that:
1.
section 8 of the Regulations should be interpreted
as precluding the presence of an AG;
2.
Sanofi has not established that it would have introduced an
AG;
3.
if an AG had been launched, it would not have come to
market until four months after Apotex’s launch; and
(a) Do the
Regulations preclude an AG?
[174] Apotex
asserts that the predominant purpose of an AG is “to truncate section 8 rights
of a second person”. Thus, Apotex submits that:
[T]he Court should not permit the intent of the section to
be defeated in this manner, particularly as it creates a windfall for the
originator who has been found not to be entitled to its monopoly.
[175] It is
unarguably the case that the inclusion of an AG in the “but for” world results
in a lower award of damages to the second person. However, I am not persuaded
that s. 8 precludes the consideration of an AG in assessing the second person’s
losses.
[176] As
pointed out by Sanofi, the Regulations, as a whole, contemplate the
existence of AGs. This can be seen, for example, at s. 7(3), pursuant to which
a generic manufacturer can obtain an NOC with the consent of the first person.
An AG is a manufacturer entering with the consent of the first person.
[177] Generic
drug companies have raised the allegation of inequities caused by AGs in the
past. The Regulatory Impact Analysis Statement
(RIAS) that accompanied the 2006 amendments contains the following remarks (Regulatory Impact Analysis Statement, (2006) C Gaz II, 1503 at 1525
[emphasis added]):
As a final note, certain generic drug companies also argued
very forcefully that the Government should incorporate measures in these
amendments to address what they perceive as diminishing market incentives in
their industry. More specifically, they contend that innovators are
increasingly entering into licencing arrangements with willing generic
companies (so-called “authorized generics”) in order to pre-empt genuine
generic competitors and retain market share past patent expiry. This practice,
which is also said to be prevalent in the US, is currently being studied by the US Federal Trade
Commission. While the Government is of the view that there is insufficient
information on the impact of this practice on market dynamics in the industry
to support regulatory action at this time, it will be examining this
practice more closely in response to these concerns.
[178]
At that time, the Governor in Council was aware that there
was an issue surrounding AGs and chose not to make amendments to s. 8 to
exclude consideration of AGs in a claim under s. 8. In the absence of clear
statutory language, I cannot simply, as urged by Apotex, exclude the AG from
the s. 8 assessment.
[179]
Section 8 damages compensate a second person for losses it
suffered as a result of the automatic stay (Alendronate (FCA), above at
para 71). Excluding an AG where the evidence demonstrates that an AG would have
been present would artificially increase Apotex’s compensation under s. 8. This
is because, in such a situation, there would have been no impediment to the
launch of an AG by Sanofi. It follows that, by excluding the AG from the “but
for” world, Apotex’s damages would exceed the revenues it would have earned had
Sanofi not brought a prohibition application.
[180]
In brief, I share the concerns expressed by Apotex.
Nevertheless, I can see no legitimate means to exclude the existence of an AG
(where demonstrable on the facts) from the “but for” market. The Governor in
Council would have to make that decision.
(b) Would Sanofi have decided to launch an AG?
[181]
The next question is whether it was more likely than not
that Sanofi would have decided to launch an AG in the “but for” world. Apotex’s
expert, Dr. Hollis, opined that there was a 60% probability that Sanofi would
have used an AG in a scenario where Apotex entered on April 26, 2004, Teva
entered on December 13, 2005, and ratiopharm entered on January 26, 2005
(Exhibit 44, vol 1 at para 150). However, in final argument, Apotex asserted
that the probability was only 25%.
[182]
There are a number of factors that lead me to conclude that
it is more likely than not that Sanofi would have decided to launch an AG as
soon as possible after the launch of Apo-ramipril on April 26, 2004.
[183]
Mr. Gravel provided very credible testimony about Sanofi’s
approach to AGs. He acknowledged that Sanofi does not launch AGs for all of its
products upon genericization. Mr. Gravel explained that Sanofi considers a
number of factors before deciding whether to launch an authorized generic. [Redacted]
[184]
One of the most important factors in determining whether
Sanofi would have decided to launch an AG is the importance of ALTACE to
Sanofi. Mr. Gravel testified that, following the publication of the HOPE study,
ALTACE sales “increased significantly year-over-year and became the leading
product in Canada”. He also stated that at one
point, ALTACE was Sanofi’s largest product.
[185]
A second factor weighing heavily in favour of an AG launch
in 2005 is the real world action of Sanofi in authorizing ratiopharm to market
ramipril in 2006.
[186]
Apotex points to a number of instances of “large products”
where Sanofi did not launch an AG. On cross-examination, Mr. Gravel
acknowledged that Sanofi did not launch an AG upon the genericization of many
drugs. The key consideration for Sanofi is obviously financial. A further
important consideration, which also affects the financial viability of an AG,
is the business arrangements under which the drug is sold. Thus, for example,
the partnership with Bristol-Myers Squibb for PLAVIX provides a reasonable
explanation for not introducing an AG for that drug. The failure to introduce
an AG for a drug being sold in a partnership arrangement or where the economics
do not support its introduction does not lead me to conclude that Sanofi would
have come to the same decision with respect to ALTACE.
[187]
The evidence clearly establishes that Sanofi had
contemplated the possibility of generic entry and the launch of an AG for
ramipril since at least 1999. This is because the '087 Patent was due to expire
in May 2002. Mr. Leprince testified that Sanofi considered the same options in
response to the expiry of the '087 Patent as it considered with respect to many
patents, and that the “usual process” involved allowing a company called
Altimed to market an AG two or three months prior to patent expiry. In that
regard, Mr. Leprince spoke to a document which contemplated generic entry in
mid-2002. While the document Mr. Leprince spoke to did not actually model
the introduction of an AG, it is clear that Sanofi considered that possibility.
[188]
Although the issuance of the '206 Patent in 2001 averted the threat of generic entry for a time, Mr.
Leprince testified that that threat resurfaced in 2003, when Pharmascience
filed a notice of allegation, followed quickly by Apotex. Sanofi also points
out that Dr. Sherman’s patent application for a ramipril formulation came to
its attention on April 16, 2003. This application would have raised the threat
of an attack on ALTACE by Apotex.
[189]
Mr. Gravel similarly testified that Sanofi was considering
the possibility of generic entry into the ramipril market when he became
involved with ALTACE at the end of 1999 and beginning of 2000.
[190] In view of the evidence before me, I am persuaded that it is more likely
than not that Sanofi would have determined that it would launch an AG during
the Relevant Period to respond to the generic entry by Apotex and Teva.
(c) When
would the AG have entered the market?
[191]
The final question is to determine when Sanofi and a
generic manufacturer would have completed all of the steps necessary to
commence sales of the AG.
[192] Sanofi
does not agree that it would have been caught off guard by Apotex’s entry into
the ramipril market. As noted above, Sanofi points to a patent application by
Dr. Sherman for a ramipril formulation as a “warning sign” that, together with
“market intelligence”, would have had Sanofi on notice that Apotex was about to
launch a generic version of ramipril.
[193] Given
Sanofi’s view of the strength of the '206 Patent and its aggressive litigation
strategy, I do not believe that the patent application, in and of itself, would
have caused Sanofi to enter into the complex negotiations and preparations for
an AG launch. As for “intelligence”, no witness for Sanofi described any
particular knowledge that would have been of assistance. More directly, when
asked for an example of a situation where an employee of Sanofi found out about
a generic having submitted an ANDS for a Sanofi product, Mr. Gravel responded
that he did not recall any.
[194] Beyond
that, we know from many witnesses – including Ms. Bowes on behalf of Health Canada –
that Apotex’s ANDS and all of its subsequent regulatory activity are kept
confidential. In the “but for” world, Apotex would not have served notices of
allegation on Sanofi.
[195] I am
satisfied that it is more likely than not that the launch by Apotex would have
been a surprise in the “but for” world.
[196] In
the actual world, in December 2006, Sanofi launched an AG – ratio-ramipril –
almost simultaneously with Apotex’s launch. However, in that case, Sanofi was
well aware of the entry of multiple generics in or around December 2006. [Redacted] In a surprise launch, Sanofi would have had to begin from
a “standing start” to introduce an AG. Apotex says this would have taken four
months; Sanofi asserts that an AG would have been ready to launch in 44 days.
[197] There
are basically two stages to launching an AG: (a) select a drug company and
conclude negotiations for an AG agreement; and (b) obtain the required
regulatory approvals and physically prepare to launch. These steps are
well-established in a number of documents presented at trial. [Redacted]
[198] [Redacted]
[199] The
problem is that the Business Review and other such planning documents are
prepared in the abstract and may not reflect what would actually happen. Thus,
the real world 2006 experience of Sanofi with its launch of an AG for ramipril
is more helpful in determining what likely would have happened in 2004. In that
case, Sanofi had some – but not much – warning. [Redacted]
Ratiopharm received its NOC on December 13, 2006 and had AG
product on the market within a few days of receiving its NOC. [Redacted] It appears that Sanofi pulled out all stops to launch the
AG [Redacted] and likely could have
accomplished a similar timeline in 2004, once a letter of intent was in place.
[200] The
only time that needs to be added to this [Redacted]
estimate is the time required to negotiate a letter of
intent and AG agreement. Sanofi would have had to select and then negotiate
with a generic manufacturer. Apotex submits that it would have taken two to
four months for Sanofi to internally reach a decision on how to proceed, to
find a partner and to come to an agreement. In his report (Exhibit 26, Appendix
12), Mr. Derek Rostant, Apotex’s accounting expert, estimates a timeline of
over seven months from the identification of an AG partner to the final signing
of a formal agreement. I think that this timeframe is unreasonably long. First,
contrary to the assertion of Apotex, I do not believe that, in 2004, Sanofi
would have had to obtain approval from Sanofi France,
Sanofi Germany and Sanofi’s North American head office.
Moreover, with the pressure of time, many of the steps to the agreement could
have been compressed.
[201] On
the other hand, I find Sanofi’s estimate of seven days to conclude a letter of
intent to be overly optimistic. Sanofi has not, it appears, factored in the
selection of an appropriate generic manufacturer. However, I accept that the
motivation to launch the AG would be a powerful driver of the selection and
negotiations. In circumstances such as a surprise launch, it is common sense to
assume that negotiations could be concluded quickly. In my estimation, Sanofi
could have selected a generic manufacturer (likely ratiopharm) and concluded an
AG agreement within one month.
[202] With
a month to finalize an AG agreement and two months to complete the steps to
launch, I find that it is more likely than not that Sanofi would have been able
to launch its AG version of ramipril by July 26, 2004 – three months after the
beginning of the Relevant Period.
(4) Conclusion
on other generics in the “but for” world
[203]
With respect to other generic entrants, I conclude that the
Generic Market during the Relevant Period would more likely than not have
included an AG, as of July 26, 2004, and Teva, as of August 1, 2006.
C. Apotex’s share of the Generic Market
(1) Apotex’s percentage share
[204]
My next task is to determine how these three market
entrants would have shared the Generic Market over the Relevant Period and,
specifically, to estimate Apotex’s Lost Volumes over the Relevant Period.
[205]
Dr. Carbone acknowledged that the order of generic entry
“appears to matter with respect to market share distribution” (Exhibit 94, vol
1 at para 98). However, in the absence of information on rebates, he felt that
it was not reasonably possible “to apply any rule other than to assume an even
allocation of market share in the case of a simultaneous market entry”. In a
scenario where there were two or three participants entering at different
times, Dr. Carbone opined as follows (Exhibit 94, vol 1 at 53, fn 38):
If only one generic is
present, the next one entering would tend to capture 50% of the total volume as
time elapses. If a third generic manufacturer subsequently enters the
marketplace, it would capture 50% of the volume previously allocated to the
second entrant. Henceforth, if three generic companies compete the first generic
manufacturer eventually captures, as time elapses, 50% of the total volume, the
second and third entrant 25% each.
[206]
It appears that an application of Dr. Carbone’s “rule”
would result in Apotex capturing 100% of the market from April 26, 2004 to July
26, 2004. Upon entry of the AG, this market share would drop to 50%. When Teva
entered on August 1, 2006, Teva would capture its market share from the AG,
leaving Apotex with its 50% share.
[207]
The biggest problem that I have with Dr. Carbone’s “rule”
is that it is completely arbitrary. It does not differentiate between a second
generic that enters within six months and a second generic that enters after
two years; in both cases, Dr. Carbone would conclude that the second generic
would capture 50% of the generic market. Dr. Carbone also fails to explain why
a third generic, under his model, reduces only the second generic’s share of
the market and not the first entrant’s.
[208]
In his report, Dr. Hollis acknowledges that the task of
allocating the Generic Market is “complex”, given that the timing of entry in
the “but for” scenarios is different from that experienced in the actual world
(Exhibit 44, vol 1 at para 76). I agree that this is a difficult assignment.
Whereas the estimates of the size of the Ramipril Market and the Generic Market
were reasonably based on the total ramipril market and total generic market in
the real world, there are no comparables in the real world to assist us in
assigning a hypothetical market share to Apotex in the “but for” world.
[209]
Nevertheless, Dr. Hollis attempts the impossible! He uses a
series of econometric models to estimate Apotex’s relative share of the Generic
Market using data from drug markets other than ramipril with two or more
generic entrants. As in Step 1, I find that Dr. Hollis’s econometric approach
at this stage is generally reliable, subject to any caveats I note below. A
serious problem, however, is that Dr. Hollis was not asked to model the
scenario that I have found for this case; nor am I certain that his model could
accommodate such a scenario.
[210]
In its overall design, Dr. Hollis’s model deviates from
what might be considered a more conventional approach to estimating market
share. Rather than estimating a single time-series model that includes time
factors as independent explanatory variables, Dr. Hollis opts to estimate a
separate model corresponding to each time quarter in his dataset. Which model
is selected will ultimately depend on the number of generic entrants and the
timing of entry that defines the “but for” world. The downside of this approach
is that fewer observations are available to estimate each model. All else being
equal, an econometric model with more observations yields more accurate
predictions compared to a model that includes fewer observations.
[211]
That being said, Dr. Hollis’s approach appears to be
reasonably robust for four reasons:
·
First, the models by and large yield statistically
significant estimates for each of the independent variables – meaning that the
individual coefficients estimated are considered to be reasonably reliable
using standard statistical criteria.
·
Second, the R-squared values (which represent the models’
“goodness of fit”), are close to or above 50% in most cases, meaning that the
independent variables account for approximately half of the observed variation
in the dependent variable (i.e. Generic Market share).
·
Third, the models yield reasonably stable coefficient
predictions for a given independent variable across all models. For example, as
one moves across the rows in Exhibit AH-17 (Exhibit 44, vol 2 at Tab 17), the
coefficient estimates for a given variable tend not to fluctuate dramatically.
·
Fourth, within a given model, the timing of entry variables
collectively estimate the kind of linear trend that one would expect to see
based on Dr. Hollis’s theory. For example, as one moves down the rows of
coefficient estimates in the Quarter 9 model in Exhibit AH-17 (-0.123, ‑0.116,
-0.198, -0.210, -0.215, etc.), the impact of late entry on market share
generally becomes increasingly negative as the lateness of the entry increases.
This trend supports the hypothesis that, all else being equal, later generic
entrants capture a relatively smaller proportion of the Generic Market.
[212]
Because of the way that Dr. Hollis has constructed his
model, he is not able to account for the interaction between the number of
generic entrants and the timing of generic entry. For example, consider two
generic manufacturers, A and B. Assume that A enters the market in Quarter 1
and B enters the market in Quarter 4. We might expect that the effect of
B’s “late entry” would differ depending on the number of generic manufacturers
already competing in the market. However, Dr. Hollis’s model is designed in
such a way that we cannot isolate this interactive effect.
[213]
Dr. Hollis prefers predictions from the real world ramipril
market post-December 2006 over his model estimates whenever the “but for” world
is likely to mirror real world experience. If Dr. Hollis observes that Apotex
faced a “more competitive” ramipril market in the real world compared to a
given “but for” scenario, he chooses to apply Apotex’s actual market share
rather than the market share predicted by his model (Exhibit 44, vol 1 at para
88). While this appears to be a reasonable approach in theory, Dr. Hollis fails
to define in more detail what he means by a “more competitive” ramipril market.
If by “more competitive” he is referring simply to cases where a greater number
of generic manufacturers are operating in the market, this criterion would fail
to account for the significant effect of entry timing on competition predicted
by his own model.
[214]
As a result, I am not convinced that Dr. Hollis’s model is
particularly helpful for my purposes.
[215]
As any judge in my position will say, the assessment of
damages can never be exact. The allocation of market share amongst the generic
entrants appears to be too complex to estimate with any accuracy. Nevertheless,
I must do my best to come up with Apotex’s market share in the Generic Market,
recognizing that perfection is impossible.
[216]
A very helpful piece of evidence before me on this
particular question was an internal Sanofi market analysis report, to which Mr.
Gravel testified (Exhibit 89, vol 1 at Tab 297). This report analyzed
prospective market shares post-genericization and was based on a careful review
of IMS CompuScript data in respect of a number of drugs. The most helpful
findings of the report were that:
·
[Redacted]
·
[Redacted]
[217]
The report does not address the situation where the AG lags
the first entrant by three months, as in the case before me. In that
circumstance, I would expect that the AG’s share of the Generic Market would
not achieve the [Redacted] rate seen with simultaneous entry. I would
estimate a market share of about 30% when averaged over two years.
[218]
In this case, there are three phases to the Generic Market:
·
Period 1: April 26, 2004 to July
26, 2004, when Apotex is alone in the market;
·
Period 2: July 26, 2004 to
August 1, 2006, when Apotex and an AG are the only participants; and
·
Period 3: August 1, 2006 to
December 12, 2006, when Apotex, Teva and an AG are sharing the Generic Market.
[219]
In my view, a reasonable assumption of Apotex’s market
share over the course of the Relevant Period, with the market participants and
entry timing that I have identified, would be as follows:
·
Period 1: 100%
·
Period 2: 70%
·
Period 3: 50%
[220]
In general terms, Apotex’s Lost Volumes during the Relevant
Period are calculated by multiplying the volumes estimated as the Generic
Market by the percentage of market share allocable to Apotex. My conclusions as
to the market shares, unfortunately, do not match any of the scenarios modelled
by the experts. Moreover, the task is somewhat complicated by the different
percentages in each of three periods. I have not carried out these calculations
but expect that Mr. Rostant and Mr. Ross Hamilton, Sanofi’s accounting expert,
would be very able to do so.
(2) Pipeline adjustment
[221]
One issue that arises with respect to the calculation of
Apotex’s Lost Volumes is what is referred to as a “pipeline adjustment” or
“channel stuffing”. Before a sale of a drug product by a pharmacist can occur,
a company must sell the product to distributors – either a retailer directly or
a wholesaler who then sells the drug to a pharmacy. In general, sales of
specific drugs are tracked through IMS EUTRx data. IMS data records sales made
at the pharmacy level and does not include the sales made “ex factory” to
supply the product pipeline.
[222]
In determining Lost Volumes, however, we need to account
for the time lag between Apotex making a sale and a pharmacist dispensing the
Apo-ramipril capsules. Each of Mr. Hamilton, Mr. Rostant and Dr. Hollis
agreed that it was appropriate to make such an adjustment, where one is working
with IMS data. This adjustment results in an addition to the forecasted Lost Volumes.
[223]
Mr. Hamilton used Dr. Carbone’s forecasts for his
calculations. Since Dr. Carbone’s forecast did not account for the pipeline
adjustment, Mr. Hamilton applied an inventory adjustment of an additional two
months of capsule sales (Exhibit 120, vol 1, Schedule 2.4 at fn 5). This
adjustment was based on Mr. Hamilton’s comparison of Apotex’s invoiced sales
over a 24‑month period in 2007-2008 with IMS data over the same period
(Exhibit 118, vol 1 at paras 35-38).
[224] Mr. Rostant made an adjustment to account for what he referred to as
“channel stuffing”. Mr. Rostant’s calculation was made by determining
incremental sales as a percentage of IMS data over the period January 2007 to
April 2007 (Exhibit 26 at 32, Appendix 13). Based on his analysis of only four
months of sales from the IMS data, Mr. Rostant’s adjustment was roughly equal
to 2.4 months of additional sales. I agree with Mr. Hamilton that this
adjustment is too high (Exhibit 119 at paras 66-68).
[225] Dr. Hollis undertook the most rigorous analysis of this problem. He
calculated the average invoiced amount of Apo-ramipril sales for the two
four-month periods following Apotex’s entry into the ramipril market: January
to April 2007 and May to August 2007. He then calculated Apo-ramipril sales for
the same two periods using IMS data. He then compared the ratio of the average
of the first four months IMS data to the actual invoice data to the ratio for
the second period. He completed this exercise for each dosage strength. His
results showed that the ratios (coefficients) when using IMS data were lower.
Dr. Hollis next used these coefficients to adjust hypothetical sales volumes in
the first four months of the Relevant Period (Exhibit 44, vol 1 at paras
96-99). Dr. Hollis’s final projections of lost sales incorporate this into his
overall forecasts of lost sales. Dr. Carbone criticizes Dr.
Hollis’s adjustment at paragraph 52 of his responding statement on the basis
that it “fails to account for the subsequent draw down of the load-in
inventories as wholesalers balance demand and supply as time elapses”. Dr.
Carbone says that Dr. Hollis erroneously assumes that wholesalers would
continue to maintain a level of inventory similar to the load-in values (see
Ex. 95). I agree with this criticism.
[226]
I agree that an inventory adjustment is appropriate. On
balance, I prefer Mr. Hamilton’s simple but effective method for calculating
the appropriate amount of the adjustment. Accordingly, I would direct that a
pipeline adjustment, computed in accordance with Mr. Hamilton’s methodology
be applied. That would result in an addition to the Lost Volumes of an
additional two months of Apo-ramipril capsule sales for each capsule dosage
strength.
IX. Apotex’s Lost Gross Sales
[227]
The next step is a calculation of revenues attributable to
Apotex’s lost gross sales. This calculation of Apotex’s gross lost sales
revenues is based on the number of Apo‑ramipril capsules that Apotex
would have sold during the period (the adjusted Lost Volumes determined in Part
VIII.C(2) of these Reasons) and the prices at which Apotex would have sold the
capsules (Lost Gross Sales). Simplistically stated, the product of Lost Volumes
and the selling price equals the estimated lost gross sales. This step requires
an analysis of pricing throughout the Relevant Period. The parties disagree
with respect to the pricing of Apo-ramipril.
[228]
Pricing of drugs is set by provincial formularies and can
differ province to province. In general, private pricing follows the public
prices set out in the formularies (Exhibit 113, vol 1 at para 37). For generic
drugs, pricing is set by each formulary as a percentage of the brand price. In
general terms, pricing will be dependent on two variables: The number of
generic manufacturers on the market; and, the regulatory pricing in place under
the different provincial formularies.
[229]
Ontario is the most significant
player in the pricing of generic drugs. In his report, Mr. Rostant set out his
understanding of the pricing for the Ontario Drug Benefit (ODB) formulary
(Exhibit 26 at 17). Between April 26, 2004 and December 31, 2005, the
first generic would enter at 70% of the brand price. Upon second and subsequent
generic entries, the price would have dropped to 63% of the brand price.
[230]
In June 2006, Ontario enacted Bill 102, Transparent Drug
System for Patients Act, 2nd Sess, 38th Leg, Ontario,
2006 [Bill 102]. Under this legislation, which came into force in late 2006,
the reimbursement level for a generic could not exceed 50% of the price of the
brand product. One question that came up was the effective date for the
lowering of existing prices pursuant to Bill 102. Mr. Fraser, speaking from his
position of knowledge as director of drug program services with the Ontario
Public Drug Programs, within Ontario’s Ministry of Health, testified that there was a transition period
between October 2006 and January 2007. Mr. Fahner’s testimony was that Apotex
only reduced its actual invoice prices to 50% as of January 1, 2007. As that
date is after the end of the Relevant Period on December 12, 2006, Bill 102 is
not relevant to the assessment of Apotex’s losses.
[231]
I accept that, in Ontario, the price of Apo-ramipril would have been at a 70% level until July 26,
2004 (while Apotex was the only generic on the market); and 63% between July
26, 2004 and December 12, 2006, the end date of the Relevant Period.
[232]
The experts appear to agree that prices in Canada during the Relevant Period tended to be uniform across all
provinces with the ODB formulary being the driver. For the period between April
2004 and December 2006, Mr. Rostant carried out his calculations for
multi-generic scenarios based on an across-the-board price of 63%, except for Alberta where he used a price of 67.5% (see Exhibit 27, Appendix
1). Mr. Palmer used a price of 68% for British Columbia but 63% for Alberta.
[233]
For the other Canadian provinces, then, I accept, as
reasonable and rationally supported by the evidence of both Mr. Palmer and Mr.
Rostant, an average price of 65% of the ALTACE listed prices for the period
from July 26, 2004 to December 12, 2006.
[234]
The remaining period – April 26, 2004 to July 26, 2004 – is
that during which Apo-ramipril was the only generic brand on the market. For a
single generic entrant, Mr. Rostant used a 75% price for Alberta and a 70% price for the rest of Canada. Mr. Palmer consistently used 70% across all provinces (Exhibit 113,
vol 2 at Appendix H). In my view, given the short period of time and lower
volumes that would have been sold in this period due to ramp-up, I am satisfied
that a 70% price overall is not unreasonable.
[235]
Apotex submitted that it would have been able to negotiate
a higher price for Apo‑ramipril while it was the sole generic. In
general, I accept that there are many instances, across the country, where a
sole-source generic has been able to achieve a higher price for its product. I
am also aware, however, that higher prices are not normally permitted
and entail negotiations with provincial regulatory authorities. In the case
before me I do not believe that the three months of being sole source would
have permitted Apotex to obtain a price higher than the regulated formulary
listings.
[236]
As a result, I conclude that Apotex’s Lost Gross Sales
should be calculated on the basis that the price for Apo-ramipril during the
Relevant Period, expressed as a percentage of the ALTACE listed price, would
have been:
·
70% from April 26, 2004 to July 26, 2004; and
·
65% from July 26, 2004 to December 12, 2006.
I expect that Mr.
Hamilton and Mr. Rostant will be able to perform the necessary calculations.
X. Apotex’s Net Lost Profits
[237]
Having determined the basis upon which Apotex’s Lost Gross
Sales should be calculated, I must address the deductions from this quantum to
come up with Apotex’s Net Lost Profits.
[238]
Mr. Hamilton applied an “incremental cost approach” to
estimate Apotex’s Net Lost Profits. Simply stated, Mr. Hamilton used a two-step
methodology: He estimated Apotex’s sales revenues for the Relevant Period and
deducted incremental expenses that Apotex would have
incurred in earning the
lost sales revenues (Exhibit 118 at para 24). Schematically, this methodology
utilizes the following steps:
Estimate of Apotex’s
Lost Volumes (A)
Estimate of lost sales
revenues (A x weighted average cost of ramipril capsules = B)
Less: Rebates and
allowances (C)
Less: early payment
discounts (D)
Estimate of
lost net sales revenue (B – (C + D) = E)
Less: cost
of sales (F)
Estimate of lost gross
profit (E – F = G)
Less: sales
commissions (H)
Less: freight and
distribution expense (I)
Estimate of lost
incremental profit (G – (H + I) = J)
Pre-judgment interest
(K)
TOTAL = J + K
[239]
In his second report, Mr. Rostant described his methodology
and Mr. Hamilton’s as a “Contribution Margin Approach” (Exhibit 27 at 4). Mr.
Hamilton also opined that his incremental cost approach and Mr. Rostant’s
methodology were the same and that he did not note any “material differences”
in the two approaches (Exhibit 119 at para 18).
[240] As I
understand the submissions made in final argument, there are three substantial
areas of disagreement:
·
sales returns;
·
trade spend; and
·
cost of API.
A. Sales
returns
[241]
Mr. Rostant included in his calculations an amount of 0.37%
as “sales returns”. In his Report, he described how, when he analyzed actual
sales for Apo-ramipril for the period post‑genericization, he observed
that the average sales returns as a percentage of gross sales was 0.37% (Sales
Return Rate) (Exhibit 26 at 19). From total gross sales he deducted an amount
equal to the product of the Sales Return Rate and the total gross Apo-ramipril
sales.
[242]
Mr. Hamilton did not make any such deduction because “the
sales projections provided by Dr. Carbone are based on IMS script sales (i.e.
prescriptions provided to patients) which are net of sales returns” (Exhibit
119 at para 30).
[243]
If sales projections – i.e. Apotex’s Lost Volumes – are
based on “IMS script sales”, then I agree with Mr. Hamilton. The problem is
that I am not able to make that determination at this time, given the
uncertainty with respect to the calculation of Apotex’s Lost Volumes.
B. Trade spend
[244]
One of the most significant costs to Apotex would be the
allowance or “trade spend” paid to pharmacists or distributors to “entice” them
to carry Apo-ramipril as a product. For the purposes of these Reasons, trade
spend includes allowances provided to pharmacists and the distribution
allowance paid to wholesalers but excludes prompt payment discounts. These are
expenses that must be deducted from Apotex’s Lost Gross Sales as a cost of
sale. The parties are in disagreement as to the appropriate level of this
allowance. The higher the trade spend, the lower the profit that Apotex would
have earned and the lower the award of damages.
[245]
One point that is not in dispute is that a manufacturer’s
trade spend is much higher when it is selling a product into a multi-generic
market. The reason is simple; in a multi-generic market, a company faces stiff
competition to convince pharmacies and wholesalers to stock its version of a
generic product. Once the product becomes part of a multi-generic market,
allowances rise dramatically.
[246]
In the case before me, Apotex would have been on the market
by itself for the short period of April 26, 2004 to July 26, 2004; competing
with one other generic for the period between July 26, 2004 and July 31, 2006;
and competing with two generics from August 1, 2006. What would have been the
levels of allowances during that period of time?
[247]
Both Mr. Rostant and Mr. Hamilton gave their opinions with
respect to the likely levels of trade spend. Mr. Rostant provided an analysis
based on Apotex’s experience with trade spend (a) for Apo-ramipril beginning in
December 2006; and (b) for other Apotex products. The range of allowances that
he provided was as follows:
·
Apotex as an exclusive manufacturer: [Redacted];
·
Apotex with one other generic competitor: [Redacted];
and
·
Apotex with more than one competitor: [Redacted].
[248]
Mr. Hamilton was asked to assume a sole-source rate of [Redacted]
and a multi-source rate of [Redacted]. He assessed the reasonableness of
the rebates and allowances assumptions by reviewing Apotex’s actual experience
with Apo-ramipril since its launch.
[249]
During the course of their testimony, Mr. Hamilton and Mr.
Rostant came very close to an agreement that [Redacted] was a reasonable
rate for a scenario with multiple participants.
[250]
Initially, there was some disagreement between the two
experts as to the appropriate sole‑source rate. Mr. Rostant admitted that
“gremlins” had resulted in an erroneous sole‑source rate of [Redacted].
During his oral testimony, he agreed that the correct number was [Redacted].
Thus, the range that he found for other products, where Apotex was an exclusive
supplier, should have been [Redacted].
[251]
As acknowledged by a number of witnesses (including Dr.
Sherman, Mr. Woloschuk and Mr. Fishman), allowance rates have increased
over time. Thus, by using data that post dates December 2006, Mr. Rostant may
have actually overstated the likely trade spend during the Relevant Period. On
the other hand, as pointed out by Mr. Hamilton, legislative initiatives in
Ontario and Quebec in late 2007 limit allowances, thereby dampening the effect
of increasing trade spend rates (Exhibit 118, vol 1 at para 60).
[252]
Apotex, in final argument, submitted that any discount
allowance rate for the Relevant Period “requires a downward adjustment from any
use of 2007 allowance rates”. Apotex argues that the rates should be: [Redacted]
for sole-source ramipril; [Redacted] for dual source (with an AG);
and [Redacted] when Apotex was competing with multiple entrants. These
rates are unsupportable on the evidence, except anecdotally and – perhaps –
wishfully. The careful assessments by Mr. Rostant and Mr. Hamilton are to be
preferred.
[253]
Sanofi argues that Mr. Rostant’s dual-source rate of [Redacted]
is too low. I acknowledge that Mr. Rostant’s analysis was based on a small
number of data points. However, given that the only competitor during the
dual-source period would be an AG, I do not believe that [Redacted] is
an unreasonable rate.
[254]
In sum, I am satisfied that the trade spend rates, inclusive
of all allowances and discounts except for early payment discounts, should be
applied as follows:
·
[Redacted], for the period when
Apotex would have been alone on the market;
·
[Redacted], for the period when
Apotex and an AG would have competed; and
·
[Redacted], for the period when
Apotex, Teva and an AG would have been on the market.
C. Cost of API
[255]
The calculation of Apotex’s damages must take into account
the cost of materials used to produce Apo-ramipril in the Relevant Period.
Stated simply, the higher the cost of materials, the lower the profit that
Apotex would have earned and the lower the award of damages. One area of
disagreement between the parties is the cost at which Apotex could have
acquired the API for ramipril during the Relevant Period.
[256]
Mr. Rostant observed that, during the period December 2006
to April 2008, Apotex purchased API from a third party supplier and from Apotex
Pharmachem Inc. (Pharmachem), a related company (Exhibit 26 at 26-27). Prices
charged by Pharmachem ([Redacted]) were significantly higher than third
party purchases ([Redacted]). Accordingly, Mr. Rostant did not use the
non-arm’s length purchase price. Rather, he used the API prices from Tektrade
Ltd., the third party supplier; that is, [Redacted].
[257]
For his calculations, Mr. Hamilton’s estimates of API price
were based on Apotex’s “weighted average cost of ramipril for 2007 and 2008,
which were [Redacted] and [Redacted] respectively” (Exhibit 118,
vol 1 at para 84). Mr. Hamilton also noted that a portion of these actual
purchases were made through a related company and commented as follows (para
84):
If I had assumed that
during the Delay Periods that Apotex would have purchased all of its ramipril
from Tektrade (Aarti Industries) and therefore, would have paid [Redacted],
my estimate of Apotex’s incremental cost of sales would have been reduced, and
Apotex’s lost incremental profits would be
increased . . .
[258]
I prefer Mr. Hamilton’s assessment on the basis of actual
costs of ramipril API to Apotex in 2007 and 2008; this price reflects a “blend”
of third party and Pharmachem purchases. As stated by Mr. Hamilton in his
responding report (Exhibit 119 at para 52):
I do not agree that
Apotex Pharmachem’s invoiced costs of ramipril (ranges from [Redacted])
should be excluded in determining Apotex’s costs in the delay period. These
costs are as a result of Apotex’s actual decisions in acquiring ramipril upon
entering the market in December 2006. As such, I have assumed that Apotex would
have acted in a similar fashion if they entered the market earlier in April
2004.
[259]
I agree with this observation. This is also supported by
the evidence of Dr. Sherman that, in 2004, the company would have bought its
API from the predecessor to Pharmachem. In all likelihood, Apotex would have
taken the same business decision in 2004 that it took in 2007 to buy a portion
of its API at a premium from a related company. Accordingly, the price used by
Mr. Hamilton in his calculations is supportable on the evidence and preferable
to the API pricing used by Mr. Rostant. Specifically, I accept Mr. Hamilton’s
assessment of ramipril API cost as [Redacted] for 2003 to 2007.
D. Other
potential costs or adjustments
[260]
Sanofi, in final argument, raises two other areas which
might require adjustments and Apotex raises one topic.
(1) Medichem
[261]
The first area relates to a non-arm’s length company
referred to as Medichem. As described by Mr. Fahner, Medichem is “a company
that Apotex utilizes in respect to payment of its allowances to customers”.
Medichem apparently is paid a service fee of a flat 0.75% of all allowances.
Sanofi submits that this amount should be added to Apotex’s cost of sales. In
my view, this expense would be captured in the overall trade spend rate. No
further adjustment is necessary.
(2) Plant
capacity
[262]
The second possible additional cost of sales could be the
added costs of increasing plant capacity to accommodate the Apotex Lost
Volumes. As observed by Mr. Hamilton during his testimony, there may be
additional costs associated with increased production:
My point was simply
that if there is a capacity pinch, where they were getting close to running at
full capacity, what Mr. Fahner suggested is he could, you know, reduce
inventory carrying time or he could run a six-day shift -- a shift on a sixth
day, or buy an additional piece of equipment.
So my only point is,
that being the case, I accept that, and we just need to factor that into our
costing. So if there is a sixth day sometimes on the weekends, you know, there
will be a shift premium you have to pay. If you are buying another piece of, I
think, encapsulation equipment, you know -- and I think Mr. Fahner said about
half a million, then I think, you know, we just need to factor that into the
costing.
I don’t mean to
suggest that they wouldn’t produce it, simply that in order to produce it,
they’re going to have to make some sort of investment. We need to factor that
into the costing. It is not currently in the costing.
[263]
Mr. Rostant, during his testimony agreed that these matters
would have to be accounted for.
[264] It seems to me that Sanofi is correct in this assertion. At this stage, I
have no evidence before me that would enable me to determine that any such
costs would be incurred during the Relevant Period. Accordingly, I would direct
the parties to determine whether, over the Relevant Period, there would be a
need for added encapsulating machinery or other added costs (such as shift
premiums) – not already accounted for – to produce the Lost Volumes.
(3) Subsequent
ramp-up
[265]
Apotex claims that it should be entitled to recover an
amount that it refers to as a second “ramp-up” or “ramp-up damages”. Sanofi
submits that Apotex is not entitled to any such recovery.
[266]
In general terms, as I understand it, the term “ramp-up”
refers to the period of time that it takes a drug manufacturer after initial
approval of its drug to reach its final level of sales. It takes some time to
negotiate agreements with pharmacies and distributors, to get formulary
listings and to physically get product to drug stores. In the hypothetical
world, Apotex would have experienced a ramp-up period for which it does not
seek compensation. However, Apotex does seek compensation in respect of
its “real world” or “duplicate” ramp-up which it argues was only incurred because
of Sanofi’s actions.
[267]
Mr. Rostant described this “ramp up” during the “Subsequent
Loss Period” (i.e. after December 12, 2006) as follows (Exhibit 26 at 33):
When Apotex launched Apo-Ramipril in
December 2006, there was a “ramp up” period before it earned profits on a fully
functional basis (“Ramp Up Period”). After receiving its NOC, Apotex commenced
the marketing and sale of Apo-Ramipril, including obtaining formulary listings.
Had Apotex commenced sale of Apo-Ramipril at the start of the Initial Loss
Period, it would have only experienced the “ramp up” at that earlier date, such
that, in the period in December 2006 and following, it would have made its
sales on a fully functional basis.
. . . . [t]he lost profit associated with
the ramp up period in the Subsequent Loss Period is the difference between what
Apotex would have sold had it ramped up in the Initial Loss Period and what it
sold in the Subsequent Loss Period when it ramped up.
[268]
According to the calculations of Mr. Rostant, the lost
profits suffered by Apotex during the subsequent ramp-up were $9,205,121. Mr.
Hamilton calculated this amount as $7,211,327 (Exhibit 119, Schedule 9).
[269]
Although the value of the second or duplicate ramp-up
period is obviously a loss to Apotex, it is a loss occurring after the Relevant
Period. The scope of a claim under s. 8 of the PM (NOC) Regulations was
addressed by the Court of Appeal in Alendronate (FCA), above. In
that case, Apotex had pleaded that, under s. 8 of the Regulations, it
was entitled to damages in respect of “lost sales and permanent market share”
(see Alendronate (FC), above at para 118). Most
relevant to the question before me, the Court of Appeal held that s. 8 does not
include damages for “future losses”, such as decreased market share due to delayed
entry into the generic market. It is worthwhile repeating the determinative
portion of the decision, at paragraphs 99 to 102:
[99] According to the analysis of the Federal Court
Judge, the losses claimed by Apotex were caused during the period since
that is when Apotex was prevented from occupying the market and obtaining the
market share which, based on its claim, it would otherwise have had. No one
takes issue with this reasoning. The question is whether the decrease in sales
which occurs in future years as a result of this decreased market share comes
within section 8. The Federal Court Judge, by allowing the claim for losses
beyond “May 26, 2005” to proceed, answered this question in the affirmative.
[100] When regard is had to the broad grant of authority
conferred by subsection 55.2(4) of the Patent Act, it seems clear that
the measure of the compensation which can be awarded under the PM(NOC)
Regulations is a matter within the discretion of the Governor-in-Council.
It is also clear that in keeping with the purpose of the PM(NOC)
Regulations and the balance which the Patent Act seeks to achieve, a
range of compensation was open to the Governor-in-Council in the exercise of
this discretion.
[101] In this case, we have the advantage of knowing
that in 1998 the Governor-in-Council focused on this very issue, and chose to
limit the measure of the losses which can be compensated by way of damages to
those suffered during the period. No issue of principle flows from this.
The Governor-in-Council could have extended the measure of the losses to
include those caused during the period, regardless of when they are suffered.
However, it did not do that.
[102] The Governor-in-Council’s clearly expressed intent
must be given effect to. This excludes compensation for losses occurring in
future years since such losses cannot be said to have been suffered during the
period. It follows, for instance, that Apotex’s entitlement to damages for lost
sales resulting from the alleged decrease in its market share must be confined
to sales that can be shown to have been lost within the period. In order to be
compensated, the losses must be shown to have been incurred during the period.
I therefore conclude that the appeal should be allowed on this limited point.
[Emphasis in original]
[270]
Apotex argues that the decision of the Court of Appeal in Alendronate
(FCA) did not extend to a claim for subsequent ramp-up. I do not agree. The
holding of the Court of Appeal is directly applicable to this type of loss.
Apotex is claiming for a loss that may have been caused during the Relevant
Period but that was not incurred during that time. The claimed loss – however
named – falls squarely within the exceptions set out in Alendronate (FCA)
and, unfortunately, is not recoverable.
[271]
If I am wrong in my application of Alendronate (FCA)
to the facts before me, I have made no attempt to reconcile the differences
between Mr. Rostant’s $9,205,121 and Mr. Hamilton’s $7,211,327.
XI. Unapproved
Indications
[272]
One final issue to deal with relates to the possibility of
an adjustment to Apotex’s Net Lost Profits to reflect the sale of Apo-ramipril
for unapproved indications.
[273]
Sanofi, in its pleadings, claims that:
·
Apotex was not in a position to market or sell Apo-ramipril
. . . for any use other than for the treatment of hypertension (Amended
Statement of Defence at para 23); and
·
Apotex’s ramipril product is only approved for a limited
indication (Amended Statement of Defence at para 36).
[274]
From the evidence and arguments presented before me in this
trial, it is clear that the factual context of Sanofi’s argument is
specifically the “HOPE indications”, which are discussed in greater detail
below. By Order dated November 25, 2011, this Court dismissed an appeal from an
Order of Prothonotary Aalto, in which he denied a motion by Sanofi to amend its
pleadings to include specific reference to the HOPE indications. In that Order,
I stated that Sanofi was not precluded from presenting its legal argument that
s. 8 does not contemplate recovery of damages in respect of lost sales of a
generic product for an unapproved indication.
[275]
Sanofi submits that the “loss” referred to in s. 8 of the PM
(NOC) Regulations does not contemplate recovery by a second person for
sales attributable to an unapproved indication or use. Thus, Sanofi argues, the
damages awarded to Apotex should include a “downward significant adjustment to
Apotex’s recoverable loss” to reflect sales of Apo-ramipril that would have
been attributable to the HOPE indications.
[276]
Sanofi put forward Dr. Peter Lin, a physician, to describe
the HOPE study and its effects on practising physicians. Dr. Lin provided great
assistance in understanding the HOPE indications, drugs useful in the treatment
and prevention of cardiovascular events and the prescribing practices of
physicians.
[277]
As described by Dr. Lin and other witnesses, the HOPE study
was a Canadian-led study, apparently undertaken with the involvement of
Sanofi’s predecessor company, Hoechst. The study assessed the role of ramipril
in patients who were at high risk for cardiovascular events but who did not
have left ventricular dysfunction or heart failure (Exhibit 122 at Tab 4: The
Heart Outcomes Prevention Evaluation Study Investigators, “Effects of an
Angiotensin-Converting-Enzyme Inhibitor, Ramipril, on Cardiovascular Events in
High-Risk Patients” (January 20, 2000) 342:3 NEJM 145 at 145 [NEJM]). The
investigators found that ramipril was “beneficial in a broad range of patients
without evidence of left ventricular systolic dysfunction or heart failure who
are at high risk for cardiovascular events” (NEJM, above at 150). In
particular, the investigators reported that “[t]reatment with ramipril reduced
the rates of death, myocardial infarction, stroke, coronary revascularization,
cardiac arrest, and heart failure as well as the risk of complications related
to diabetes and of diabetes itself” (NEJM, above at 150). Thus, the term “HOPE
indications” has come to be associated with the patient profiles from the HOPE
study where vascular protection was demonstrated.
[278]
The results of the HOPE study were first presented in
August 1999 at the European Society of Cardiology meeting in Barcelona, and later reported in the January 20,
2000 edition of the New England Journal of Medicine (Exhibit 122 at fn 3, Tab
4). Sales of ALTACE immediately increased quite dramatically (Exhibit 82 at
1932).
[279]
ALTACE was approved for the HOPE indications on February
13, 2001. However, as shown in the table at paragraph 27 of these Reasons,
Sanofi did not protect its claim to the use of ALTACE for the HOPE indications
until 2005, when the two HOPE Patents were granted and Sanofi obtained two
listings on the Patent Register. By that time, the rate of increase of sales of
ALTACE had started to slow down.
[280]
One of the last steps in a drug approval process is the
finalization of the product monograph. The product monograph, in part, sets out
the uses or indications for which the drug is intended. The NOC issues with
reference to the product monograph. From time to time, negotiations take place
between Health Canada and
a generic manufacturer as to what the approved indications will be. Apotex,
having initially included the HOPE indications in its product monograph,
removed those indications on December 14, 2006. In our hypothetical world, it
is therefore likely that Apotex’s final product monograph, as of April 26,
2004, would not have included reference to anything other than hypertension. In
other words, as of April 26, 2004, Apotex would likely have launched
Apo-ramipril with no reference to the use of ramipril to treat proteinuaria ('948 Patent) or the HOPE indications ('549
and '387 Patents). As we know,
Sanofi had listed these patents on the Patent Register.
[281]
In spite of the “limited approval”, it is more likely than
not that some sales of ramipril during the Relevant Period would have related
to the HOPE indications. The question is whether Apotex can recover for those
sales.
[282]
Sanofi submits that Apotex is not entitled to recover in
respect of ramipril sales during the Relevant Period that would have been made
to address the HOPE indications. Since Apotex would have had no entitlement to
make sales of Apo-ramipril for the HOPE indications during the Relevant Period,
it therefore cannot now claim a loss attributable to those sales.
[283]
While Sanofi’s argument has logical appeal, it is not
supported by the facts (or – in my view – the law) which demonstrate that sales
of generic drugs for unapproved or “off-label” prescriptions can and do,
legally, take place. There are a number of arguments that run counter to
Sanofi’s submission:
·
the fact that generic manufacturers do not promote drug
products for specific indications;
·
the fact that off-label prescribing and substitution take
place and that product monographs are not relevant to physicians;
·
the fact that, in the real world, Sanofi did not oppose the
listing of Apo-ramipril as fully interchangeable with ALTACE; and
·
the availability to Sanofi of an action for patent
infringement with respect to the HOPE Patents.
[284]
I will discuss each of these.
[285]
First, I observe that generic products are not promoted for
specific uses, but are instead sold as drug products. Dr. Sherman’s testimony
in this regard is clear and credible:
Q. Who
do you promote your products to, sir?
A. Only
to pharmacists.
Q. And --
A. Just
for the filling of prescriptions, not for the writing [of] prescriptions.
Q. In
promoting to pharmacists, what use, if any, do you make of monographs?
A. We don't.
Q. Okay. Does
Apotex market its products to patients?
A. No.
Q. Does Apotex
sell its products directly to patients?
A. No.
[286]
Sanofi places much emphasis on the various versions of
product monographs submitted by Apotex, some of which included reference to the
HOPE indications. It is true that the originally-filed product monograph for Apo-ramipril
included unapproved indications. As explained by Dr. Sherman, this is because
Health Canada’s policy is that generic
monographs should mirror as closely as possible the monograph of the brand
product. Sanofi’s argument could have some traction if the product monograph
formed the basis on which ramipril was marketed or sold. This is not the
situation.
[287]
It appears that approved indications in product monographs
are not determinative in the promoting, prescribing or selling of drugs. Dr.
Lin, during cross-examination, acknowledged that he was not aware if any of the
Apo-ramipril monographs were distributed to doctors, pharmacists or patients.
[288]
Dr. Lin testified that doctors started prescribing ALTACE
for the HOPE indications immediately after the presentation of the HOPE study
results in August 1999. This was some 18 months before Sanofi received its NOC
with respect to the HOPE indications. This was “off-label” prescribing, or the
prescription of a product for a use that is not set out in a product monograph.
This practice was common, particularly with a known drug such as ALTACE. As
stated by Dr. Lin:
Q. And
you say that off-label prescribing is a common and accepted practice?
A. When
there is a trial that is ahead of the indication. So, in other words, we
follow what the trials are saying. If there is benefit for people, then often
times we would prescribe that medication to protect those people, especially
because the medication is already there.
[289]
There appears to be nothing “illegal” about off-label
prescribing. Nor does Sanofi plead illegality.
[290]
It seems that a significant number of physicians would have
prescribed generic ramipril for the HOPE indications during the Relevant
Period, even if those indications were not included in the generic’s product
monograph. While I did not hear evidence from any front-line pharmacist, I am
prepared to accept that the more usual practice would be for the pharmacist to
substitute, even where ALTACE is written on the prescription. As Dr. Lin
pointed out, this result would have likely occurred in any event as a result of
mandatory generic substitution (Exhibit 124 at 2889).
[291]
I have no evidence that, at the time when Apo-ramipril was
launched in late 2006, Sanofi opposed the listing of Apotex’s product as fully interchangeable;
or that Sanofi required Apotex (or any other generic entrant) to obtain a
limited listing for its product.
[292]
It is, therefore, more likely than not that Apotex would
have been able to make sales for the HOPE indications during the Relevant
Period, without objection. It follows that any sales made during the Relevant
Period which were solely related to the HOPE indications are still lost sales
that Apotex would have made in the absence of Sanofi’s statutory stay, and
losses for which Apotex is entitled to recover under s. 8.
[293]
If Sanofi believes that Apotex is infringing or inducing
infringement of the HOPE Patents, then Sanofi has a cause of action under the Patent
Act, RSC 1985, c P-4. In that regard, I note that, since Apotex and other
generics began selling generic ramipril, Sanofi – who is no stranger to
litigation – has not brought an action against any of the manufacturers for
infringement of the HOPE Patents.
[294]
Even if Sanofi is correct, and s. 8 prevents a second
person from recovering for sales for an unapproved use, there is insufficient
evidence to merit a reduction of Apotex’s damages on the facts of this case.
[295] I conclude that Apotex is not precluded from recovering losses associated
with the HOPE indications. That is not to say that a second person may always
recover for unapproved indications. Another s. 8 claim may provide a clear
defence in the pleadings and a different set of facts that would warrant a
different finding or a downward adjustment to the second person’s damages pursuant
to s. 8(5) of the Regulations. But, not in this case.
XII. Conclusions
[296]
In concluding, I would like to make a general comment. As
noted at the beginning of these Reasons, right before the trial of this matter,
I heard a companion case in Court File No. T‑1161‑07 – Teva
Canada Limited v Sanofi-Aventis Canada Inc and Sanofi-Aventis
Deutschland GmbH. There are obviously many similarities between the two
cases. However, each case proceeded separately, with a different record. I wish
to assure the parties and the readers that my decision in each case was made
completely on the basis of the arguments made and the records before me in the
applicable case.
[297]
Having addressed all of the issues before me, I am quite
disappointed that I cannot finalize a quantum of damages. However, I am hopeful
that Sanofi and Apotex, with the capable assistance of their lawyers and
experts can quickly agree on a final amount to be paid by Sanofi to Apotex
based on these Reasons for Judgment.
[298]
In summary, the key findings that I have made, on the basis
of the record before me, are as follows:
1.
The Relevant Period for the determination of Apotex’s Net
Lost Profits commences on April 26, 2004 and ends on December 12, 2006.
2.
The Ramipril Market during the Relevant Period should be
quantified in accordance with the calculations of Dr. Hollis.
3.
The Generic Market during the Relevant Period should be
quantified in accordance with the calculations of Dr. Hollis.
4.
Apotex’s Lost Volumes should be calculated on the basis
that Apotex would have entered the market as of April 26, 2004, an authorized
generic would have entered on July 26, 2004 and Teva would have entered on
August 1, 2006, with Apotex holding the following shares of the Generic Market:
a.
April 26, 2004 to July 26, 2004 (Period 1) – 100%;
b.
July 26, 2004 to August 1, 2006 (Period 2) – 70%; and
c.
August 1, 2006 to December 12, 2006 (Period 3) – 50%.
5.
A pipeline adjustment, computed in accordance with Mr.
Hamilton’s methodology should be applied, resulting in an addition to the Lost
Volumes of an additional two months of Apo-ramipril capsule sales for each
capsule dosage strength.
6.
With respect to the calculation of Apotex’s Net Lost
Profits:
·
pricing of Apo-ramipril during the Relevant Period,
expressed as a percentage of the ALTACE listed price, would have been:
o
70% during Period 1; and
o
65% during Periods 2 and 3;
·
if necessary, an amount for “sales returns” should be
accounted for;
·
allowances (trade spend) in the following percentages,
including distribution allowances, sales discounts, credit card discounts and
cost of free goods, should be included as a cost of sales at the rates of:
o
[Redacted] for Period 1;
o
[Redacted] for Period 2; and
o
[Redacted] for Period 3;
·
a price of [Redacted] for API should be applied;
·
no adjustment should be made in respect of Medichem;
·
the parties are to direct their minds to the Apotex Lost
Volumes to determine whether, over the Relevant Period, there would be a need
for added encapsulating machinery or other added costs (such as shift premiums)
– not already accounted for – to produce the Lost Volumes and, if so, to
account for these costs;
·
no adjustment is to be made with respect to the duplicate
ramp-up period; and
·
no adjustment should be made in respect of unapproved
indications.
7.
Pre-judgment and post-judgment interest is payable on the
award of damages as follows:
·
pre-judgment interest, not compounded, calculated
separately for each year since April 26, 2004 at the average annual bank rate
established by the Bank of Canada at the minimum rate at which the Bank of
Canada makes short-term advances to the banks listed in Schedule 1 of the Bank
Act, SC 1991, c 46; and
·
post-judgment interest, not compounded, at the rate of 5%
established by the Interest Act, RSC 1985, c I-15, s. 4.
[299]
In addition, there is the question of costs. I would hope
that the parties can agree on costs. In the event that the parties cannot agree
on the amount of costs by June 15, 2012, they may make submissions to this
Court, such submissions not to exceed ten pages. The parties will have a
further 15 days to make reply submissions, if they choose, not to exceed five
pages.
[300]
I wish to express my gratitude to counsel for their
diligence, competence and professionalism throughout the pre-trial matters and
the trial. You brought me solutions not problems! Thank you.
Email this Content