Misleading Affidavits
Canada has legislation based on the US Hatch-Waxman Act, which allows a generic pharmaceutical company to base its application for health regulatory approval on data provided by the brand, so long as the generic can show that its product will not infringe any valid patent listed against the drug by the brand. Validity of the patent is commonly challenged in so-called Notice of Compliance (NOC) proceedings but, because the proceedings are carried out by way of affidavit, with limited discovery, a determination of invalidity is not res judicata, but simply allows the generic to obtain regulatory approval for launch. When a generic launches its product after prevailing in the NOC proceedings, it is normal for the patentee to immediately commence an infringement action in which the same questions of validity and infringement are re-litigated between the same parties.
This occurred in respect of Plavix (clopidogrel), an enantiomer of a prior art racemate. One question relating to validity of the Plavix patent is whether the separation method used by Sanofi was obvious. Litigation between Sanofi and Apotex went to the Supreme Court of Canada in the NOC proceedings, where the Court held that the separation method was not obvious, and the validity of the patent was upheld. The contrary conclusion was reached at trial in the subsequent infringement litigation, which invalidated the patent. In holding that the separation method was inventive, the Supreme Court pointed to the affidavit evidence of the inventor responsible for separation, which testified to the difficulty of achieving separation. In the infringement proceedings, the trial judge noted that there was no documentary evidence of the key series of failed experiments that established the difficulty of separation. These experiments were not recorded in the inventor’s lab notebooks, even though the lab notebooks were otherwise complete. The trial judge also pointed to a number of other discrepancies; for example, the inventor’s deposition in the US infringement proceedings was inconsistent with his affidavit evidence in the Canadian NOC proceedings in this respect. The trial judge consequently concluded that the inventor’s evidence as to the difficulty of achieving separation was “unpersuasive”.
This illustrates why the NOC proceedings are not binding on the subsequent infringement proceedings. But the discrepancy between the affidavit evidence and the evidence at trial, after discovery and cross-examination, also illustrates the inherent trade-off between speed and cost, on the one hand, and thoroughness on the other. Are misleading or false affidavits a real problem in EPO opposition proceedings? This Kat imagines the problem might be worse, as the issue in NOC proceedings are usually re-litigated, but if misleading or incomplete evidence kills a patent at the EPO, then the patentee will never have the chance to disprove this evidence in a national court. If so, what is the solution? The most obvious solution would be the sanction of contempt, but contempt is not available in the EPO, and in any event, I am not sure that would be cost-effective, given the relatively high standard for establishing contempt. (I note that there was no suggestion in the Plavix decision that the discrepancies at issue amounted to swearing a false affidavit.) Also, there is a free-rider problem, as one patentee bears the cost of a given contempt action, while the incentive to be accurate in swearing affidavits going forward benefits all future litigants. Are there other IP (or non-IP) contexts in which this problem arises and has been solved?
This occurred in respect of Plavix (clopidogrel), an enantiomer of a prior art racemate. One question relating to validity of the Plavix patent is whether the separation method used by Sanofi was obvious. Litigation between Sanofi and Apotex went to the Supreme Court of Canada in the NOC proceedings, where the Court held that the separation method was not obvious, and the validity of the patent was upheld. The contrary conclusion was reached at trial in the subsequent infringement litigation, which invalidated the patent. In holding that the separation method was inventive, the Supreme Court pointed to the affidavit evidence of the inventor responsible for separation, which testified to the difficulty of achieving separation. In the infringement proceedings, the trial judge noted that there was no documentary evidence of the key series of failed experiments that established the difficulty of separation. These experiments were not recorded in the inventor’s lab notebooks, even though the lab notebooks were otherwise complete. The trial judge also pointed to a number of other discrepancies; for example, the inventor’s deposition in the US infringement proceedings was inconsistent with his affidavit evidence in the Canadian NOC proceedings in this respect. The trial judge consequently concluded that the inventor’s evidence as to the difficulty of achieving separation was “unpersuasive”.
This illustrates why the NOC proceedings are not binding on the subsequent infringement proceedings. But the discrepancy between the affidavit evidence and the evidence at trial, after discovery and cross-examination, also illustrates the inherent trade-off between speed and cost, on the one hand, and thoroughness on the other. Are misleading or false affidavits a real problem in EPO opposition proceedings? This Kat imagines the problem might be worse, as the issue in NOC proceedings are usually re-litigated, but if misleading or incomplete evidence kills a patent at the EPO, then the patentee will never have the chance to disprove this evidence in a national court. If so, what is the solution? The most obvious solution would be the sanction of contempt, but contempt is not available in the EPO, and in any event, I am not sure that would be cost-effective, given the relatively high standard for establishing contempt. (I note that there was no suggestion in the Plavix decision that the discrepancies at issue amounted to swearing a false affidavit.) Also, there is a free-rider problem, as one patentee bears the cost of a given contempt action, while the incentive to be accurate in swearing affidavits going forward benefits all future litigants. Are there other IP (or non-IP) contexts in which this problem arises and has been solved?
