More Debate on the Compensable Period under Section 8

Apotex Inc v Sanofi-Aventis Canada Inc / ramipril (s 8) 2012 FC 553 Snider J

Yesterday’s post discussed the question of the start date of the compensable period for a s 8 claim in Teva / raimpril (s 8). Today’s post discusses the parallel question in Apotex / ramipril (s 8), in which both the start date and the end date were in dispute. Note that of the four dates to be determined in the two cases, only the end date in Teva was uncontroversial. Nor were any of the disputes easy to resolve; all raised significant legal questions relating interpretation of the Regulations. While the facts of both Teva / ramipril (s 8) and Apotex / ramipril (s 8) were unusual in certain respects, NOC litigation is very complex, and no doubt many cases are unusual in one way or another. That the first two cases to arrive at an assessment of s 8 damages have raised difficult issues of interpretation suggests that future litigation will continue to reveal uncertainty in the application of s 8 of the NOC Regulations. This is not to impugn the drafting; anticipating the future fact patterns is no easier for legislative drafters than it is for patent drafters.

The start date under s 8(1)(a)(i) of the NOC Regulations is presumptively the “patent hold” date, the date on which Health Canada’s review of the drug submission is completed and the generic is advised that an NOC will not issue until the listed patents are addressed. Section 8(1)(a)(ii) gives the court the discretion to select a more appropriate date.

In this case the patent hold date was April 26, 2004. The start date issue arose because prior to the patent hold date Apotex had challenged one of Sanofi’s patents, the ‘457 patent, in NOC proceedings, alleging non-infringement. Sanofi prevailed, and on October 6, 2005, Simpson J 2005 FC 1381 issued a prohibition order preventing the Minister from issuing an NOC to Apotex until the expiry of the ‘457 patent, which occurred on December 13, 2005. A few months after serving the first NOA, Apotex served a second NOA, this time alleging invalidity. Apotex was successful in this proceeding at first instance [44]. Apotex and Sanofi appealed the respective decisions, but the ‘457 patent expired before the appeals were heard. Apotex abandoned its appeal [43], and Sanofi’s appeal was dismissed as moot [45].

Sanofi argued that the start date for the compensable period should be the date of the expiry of the ‘457 patent, as Apotex could not have entered the market because of Simpson J’s prohibition order, which was never reversed on appeal. Snider J rejected this argument, holding that “the first decision was subsumed or ‘trumped’ by the second” [47]. Snider J noted that if Apotex had alleged both non-infringement and invalidity in a single NOA, the application would likely have been dismissed on the invalidity ground, and “[t]here is no principled reason why the result should be any different just because Apotex served and pursued two notices of allegation rather than one” [50]. I am not sure that this is persuasive. When a generic serves a notice of allegation, the innovator has to decide whether to respond; if it does not, s 8 liability will not be triggered. In principle, an innovator might feel very confident of its case regarding infringement, but less confident of its case on validity, and so choose to respond to one but not the other. If Sanofi had not responded to the second NOA, would it still be right to ignore the order of prohibition in determining the start date? In a sense a shortened compensation period would reflect Apotex’s own choice to allege only non-infringement in the first NOA. With that said, Snider J’s larger point is that we now know that Apotex should not have been kept out of the market on the basis of the ‘457 patent, and the compensable period should not be reduced because of an “invalid” patent. Presumptively, the compensable period runs from the patent hold date, and an order based on an “invalid” patent is not sufficient to justify a different date. It is also significant that Sanofi will only be liable for damages for the period of the statutory stay, which it would normally expect. This is in contrast to the situation in Teva / ramipril (s 8), where the start date argued for by Teva would have extended the liability period by a potentially arbitrary amount.

In Apotex / rampril (s 8) the end date of the compensable period was also in issue. Section 8(1)(b) provides that the end date is the “the date of the withdrawal, the discontinuance, the dismissal or the reversal” of the order of prohibition. The court has no discretion to vary this, but what exactly is that date? Apotex argued that the date should be May 2, 2008, which was the date of the dismissal of the last prohibition proceeding [59]. However, because of some unusual procedural history, the Minister actually issued the NOC on December 12, 2006, before the dismissal of the last proceeding [60]. The last proceeding was dismissed for mootness on May 2, 2008 [62]. Snider J was of the view that the appropriate date was the date on which the proceeding actually became moot, that is, the date of the issuance of the NOC, and not the date on which it was formally declared moot [63]. May 2, 2008 “is merely an arbitrary date on which Prothonotary Aalto dealt with a motion before him. . . . 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” [63]. This logic is compelling. The only difficulty is that s 8 itself does not list “mootness” as one of the various ways in which a proceeding can be terminated.