Friday, March 28, 2014

Start and End Dates for Section 8 Liability Period

Apotex Inc v Sanofi-Aventis Canada Inc [Apotex s 8 Liability Appeal] 2014 FCA 68 Sharlow JA, Pelletier JA concurring, Mainville JA dissenting, var’g 2012 FC 553 Snider J ( blog blog);
Teva Canada Ltd v Sanofi-Aventis Canada Inc [Teva s 8 Liability Appeal] 2014 FCA 67 Sharlow JA, Dawson JA concurring, Mainville JA dissenting, var’g 2012 FC 552 Snider J (blog blog)
            ramipril / ALTACE

Under the patent linkage set provided for in the PM(NOC) Regulations, a generic that would otherwise be in a position to receive an NOC as a result of an ANDS will be placed on “patent hold” until it has addressed all the patents listed against the drug in question by the innovator / patentee whose drug was the reference product for the ANDS. The generic can respond with a Notice of Allegation, alleging invalidity or non-infringement, and if the patentee responds, a statutory stay is triggered under s 7, preventing the generic from receiving its NOC until the patents have expired or the generic has prevailed on its allegations in NOC proceedings. If the generic is successful, s 8(1) makes the patentee liable to the generic for losses suffered from having been wrongly (in hindsight) held off the market by the statutory stay. The patent linkage system is analogous to an automatic interlocutory injunction, and s 8 is analogous to the undertaking in damages which is usually required for a plaintiff to obtain such an injunction.

The Apotex and Teva s 8 Liability Appeals raise some unusual issues regarding the start and end dates for the section 8 liability period (which I have previously called the compensable period). While the facts were unusual, this means that the FCA had the opportunity to clarify some principles which are not as apparent in more routine situations.

Start Date: Teva
S 8(1) provides that the liability presumptively begins on the patent hold date, “(ii) unless the court concludes that a date other than the certified date is more appropriate.” In Teva, the main issue arose because the patent hold date was two years before the beginning of the statutory stay, raising the concern that if the patent hold date was used as the start date, the generic would be compensated for a period when off the market for reasons unrelated to the statutory stay.

The patent hold date was 14 October 2003 [36], but Teva had elected to await the expiry of the one of the listed patents, the ‘457 patent, which expired on December 13, 2005. Consequently Teva did not serve its NOA until September 12, 2005, thereby triggering the statutory stay on 31 October when Sanofi replied with its Notice of Application [40fc], just over two years after the patent hold date. Snider J held that as a matter of law, “the liability period cannot predate the statutory stay” [60fc]. This meant that the liability period could not start on the patent hold date, and on the facts, she concluded that “a more appropriate date” was the expiry of the ‘457 patent [75fc].

On appeal, Mainville J speaking for the unanimous court on this point [142], held that Snider J was wrong to hold that the liability period could not precede the statutory stay as a matter of law [79]. However, in the circumstances the FCA agreed with Snider J that the expiry of the ‘457 patent was more appropriate than the patent hold date [80], on the basis that “Teva took a business decision not to challenge the ‘457 Patent under either the NOC Regulations or the Patent Act, and it must now live with that business decision” [85].

Thus it seems that in principle at least, under the FCA rule a generic which was simply dilatory in serving its NOA could get a start date prior to the statutory stay, while that would not have been permitted by Snider J. This raises a point made by Snider J, which is that ordinary principles of causation would prevent recovery prior to the start of the stay, because it was Teva’s delay rather than Sanofi’s conduct which caused the loss [58fc]. Mainville J responded that “Though ordinary principles of the law of damage have an important role to play under section 8 of the NOC Regulations, the liability set out under that section is purely statutory. As a result, the clear language of the Regulations (such as that set out in para. 8(1)(a)) must in all cases prevail over general principles” [78]. I don’t find this compelling. The language in para 8(1)(a), specifying the patent hold date, is indeed clear, but it is clearly only a prima facie date. Given the general discretion that is clearly allowed by subpara (ii), I would have thought that the very fundamental principle that a party should not be liable for a loss that it did not cause, would have full rein. In my view, while Snider J may have been wrong to hold that the start date could never predate the statutory stay, her causation argument is powerful. In any event, the causation argument is not dead, since Mainville J did not say that causation could never be a consideration under subpara (ii), and the FCA’s holding on facts respected the causation requirement.

Start Date: Apotex
In Apotex the patent hold date was April 26, 2004 [40fc]. The problem arose because Apotex had split its NOAs respecting the ‘457 patent (a practice which is no longer permitted). Apotex first served an NOA alleging non-infringement. It was unsuccessful, and a prohibition order was granted by Simpson J [43fc]. Apotex then served an NOA alleging invalidity. That argument was successful, and Tremblay-Lamer J dismissed Sanofi’s application for a prohibition order [44fc]. The first prohibition order was never reversed; it expired of its own accord on the expiry of the ‘457 patent in December 13 2005. Sanofi argued that the prohibition order could not be ignored, and Apotex would not have been able to come to market until the expiry of the ‘457 patent, and 13 December 2005 was therefore the appropriate start date for the liability period.

Snider J and the FCA disagreed. Mainville J, speaking for the FCA on this point, held in effect that both proceedings relating to the ‘457 patent should be considered as a whole, and the net effect was that the ‘457 patent was not an impediment of the issuance of an NOC to Apotex [78]: “Liability under section 8 of the NOC Regulations should not be made to be dependent on the largely irrelevant question of whether Simpson J. or Tremblay-Lamer J. was first to issue a decision” [79].

Perhaps the most interesting aspect of the decision is that the Mainville J accepted the principle of Virgin Atlantic Airway Ltd v Zodiac Seats UK Ltd [2013] UKSC 46, which, in broad terms, is that a party should not be made liable for infringement of an invalid patent. The problem arises in Europe because of the possibility of infringement litigation in the UK running in parallel with, or before, an opposition based on invalidity in the EPO. In Unilin [2007] EWCA Civ 364 the EWCA had held that in the interest of certainty and finality, a judgment for the patentee ordering an enquiry as to damages could not be set aside even if the patent was subsequently declared invalid by the EPO. Sanofi relied on Unilin in this case, but as Mainville J pointed out [84], that case was specifically overruled in Virgin Airways.

The precise principle from Virgin will rarely be directly applicable in Canada because we do not have the same system of parallel jurisdiction, but our own version of the problem has arisen in the question of whether a generic can be entitled to s 8 damages for a patent which later turns out to be valid. The FCA has recently held that while the s 8 action won’t be set aside, damages payable under s 8 can be considered damages in a subsequent infringement action: see here. This is broadly consistent with Virgin, and also with the FCA’s holding in this case. That is, it seems to be an emerging general principle that considerations of fairness – that a party should not be held liable for a wrong that turns out not to have been wrong after all – trumps considerations of finality. Exactly how far this principle extends remains to be seen: note that in all the UK cases, including Virgin, the enquiry as to damages had not concluded, and it is not clear the same result would follow if there had been a final judgment for a liquidated sum: [Virgin 36].

End Date: Apotex
Section 8(1)(b) of the NOC Regulations provides that the liability period ends on “the date of the withdrawal, the discontinuance, the dismissal or the reversal,” of “an” application made under subsection 6(1). In Apotex, there were multiple patents listed by Sanofi against ramipril, and five different dismissal dates relating to five separate prohibition applications involving Sanofi and Apotex [38], and the question was which of these was “an” application. Fortunately, only two need concern us. The second to last, relating to the ‘948 patent, was dismissed on June 27, 2006. This date was favoured by Sanofi. The last, relating to the so-called HOPE patents, was dismissed on May 2, 2008. This date was favoured by Apotex. The difficulty arose because the HOPE patents were not issued and listed until 2005, long after Apotex had been placed on patent hold. In November of 2006, the SCC released AstraZeneca 2006 SCC 49, which held that the generic did not have to address after-listed patents. Consequently, the Minister issued the NOC to Apotex on December 12, 2006, after the ‘948 proceeding was dismissed, but before the HOPE proceeding was dismissed. Snider J held that under these very unusual circumstances, the appropriate end date was the date on which Apotex received its NOC, that is December 12, 2006, and the FCA agreed, essentially on the basis that “once AstraZeneca was issued, all pending prohibition proceedings respecting the HOPE patents became moot, and the ‘withdrawal, the discontinuance, the dismissal or the reversal’ of those proceedings under the meaning of paragraph 8(1)(b) of the NOC Regulations was concomitant with the issuance of the NOC to Apotex” [97]. Matters would have been simpler if the Regulations specified that the end date was when the NOC was issued, rather than when the last proceeding was withdrawn. Since the start date is presumptively when the NOC would have issued, it seems reasonable that the end date should be, presumptively at least, when the NOC actually issued. No doubt there is some reason which doesn’t occur to me know why it was drafted as it was. In any event, these facts are very unusual, so this issue may not arise again.

End Date: Teva
While this is all interesting, I admit to being relieved that in Teva the parties agreed on the correct end date: [35], [36fc].

No comments:

Post a Comment