Disappearing Polymorphs And Gastrointestinal Infringement

Results in cases brought in Canada and the United States by GlaxoSmithKline Beecham (“GSK”) against various generic companies concerning GSK’s PAXIL anti-depressant serve to highlight significant differences between pharmaceutical patent litigation in the two countries. This article considers litigation that has taken place to date in Canada and the United States, focussing on the contrasts in claim interpretation.

BACKGROUND
Paroxetine hydrochloride is used to treat depression and other disorders. A British company, Ferrosan, discovered paroxetine hydrochloride and its anti-depressant qualities in the 1970s. Ferrosan subsequently licensed its patent rights to GSK which began experimenting with the anhydrate form of paroxetine hydrochloride (the only form known to exist at that time).

The inherently unstable anhydrate form presented difficulties in the manufacturing process. GSK discovered a hemihydrate form of paroxetine hydrochloride, which proved to have significantly better handling characteristics. Tests conducted on patients confirmed the safety and efficacy of the hemihydrate in the treatment of depression. GSK obtained patents in various countries, including Canada and the United States incorporating claims to “crystalline paroxetine hydrochloride hemihydrate”.

LITIGATION AGAINST APOTEX IN CANADA
In 1996, Apotex filed for a Notice of Compliance (“NOC”) to market the anhydrate form of paroxetine hydrochloride in Canada. GSK moved in the Federal Court against Apotex pursuant to the Patented Medicines (Notice of Compliance) Regulations seeking an order prohibiting the grant of an NOC to Apotex. GSK argued that the generic anhydrate was unstable and would convert to the more stable hemihydrate form either during the tableting process or over time after exposure to heat or humidity. GSK adduced evidence of experiments showing that batches of anhydrate converted to hemihydrate over time.

In 1999, Justice McGillis of the Federal Court dismissed GSK’s application for an order of prohibition. McGillis J. held that GSK’s evidence raised “no more than a possibility of infringement”, concluding that Apotex should not be prevented from taking anhydrate tablets to market “on the basis of a potential conversion to hemihydrate at some undisclosed and imprecise time in the future”. McGillis J. did note, however, in the event that Apotex’ anhydrate tablets convert to the patented hemihydrate form “in whole or in part, it will face ‘very grave’ consequences at that point in time”. McGillis J. apparently interpreted GSK’s claim 1, for “crystalline paroxetine hydrochloride hemihydrate”, to cover any amount of hemihydrate. Her ruling was upheld on appeal.

LITIGATION AGAINST APOTEX IN THE U.S. DISTRICT COURT
GSK also commenced infringement proceedings against Apotex in the United States District Court for the Northern District of Illinois. At issue was claim 1, for “crystalline paroxetine hydrochloride hemihydrate”.

GSK argued that its claim covered the hemihydrate in any amount. GSK further argued that Apotex’ product would necessarily infringe because the anhydrate form would in all cases convert to the patented hemihydrate form. GSK then had to explain why Ferrosan’s production of the anhydrate form well before GSK’s patent filing date would not anticipate GSK’s hemihydrate patent. GSK advanced the “disappearing polymorph” theory: as Ferrosan and its successors improved the manufacturing process for paroxetine hydrochloride, the relatively unstable anhydrate morphed into the more stable hemihydrate and the general environment became seeded with crystals of the new polymorph. In the seeded environment, even as little as a single crystal of the hemihydrate was enough to convert the anhydrate, at least in part, to the hemihydrate form.

GSK argued its tests showed that Apotex’ product would inevitably convert to hemihydrate when combined with moisture, pressure and hemihydrate crystals. Lastly, GSK argued that, in any event, Apotex’ product converted to the hemihydrate form once exposed to the patient’s gastrointestinal juices and Apotex knew that this would be the result.

In March 2003, Judge Posner issued his reasons for decision, ruling that the GSK patent was valid but not infringed by Apotex’ formulation. In Judge Posner’s view, GSK’s construction would render the claim indefinite since a potential infringer would never be able to know whether its product would infringe, as even a single, undetectable crystal of the hemihydrate would be enough for infringement. The district court therefore construed GSK’s claim to only cover the hemihydrate in “any commercially significant quantity”.

Judge Posner found that Apotex’ product would likely contain only small quantities of hemihydrate produced during the formulation process, so there could be no patent infringement. Further, even if Apotex’ product converted to the patented hemihydrate form once exposed to the patient’s gastrointestinal juices, Apotex was not liable for inducing infringement since there was no evidence that it “desires or is working to achieve, this result”.

Judge Posner was prepared to find GSK’s claim invalid for indefiniteness but instead interpreted it to exclude commercially insignificant amounts of hemihydrate produced by involuntary conversion. He then held that, in the event GSK’s claim were construed to cover any amount of hemihydrate, Apotex would have an equitable defence to infringement since GSK itself was responsible for Apotex’ inability to create a non-infringing product because GSK had seeded the environment with the hemihydrate polymorph.

LITIGATION AGAINST GENPHARM IN THE FEDERAL COURT OF CANADA
In June 2003, subsequent to the release of Judge Posner’s reasons for decision, Justice Heneghan of the Federal Court of Canada heard argument in an NOC dispute between GSK and Genpharm, which also was seeking an NOC to market the anhydrate form of paroxetine hydrochloride.

GSK argued that the words “crystalline paroxetine hydrochloride hemihydrate” were plain and unambiguous and, as such, there was no need to refer to the patent disclosure to give them meaning. In response, Genpharm argued that the claim must be read in the context of the entire patent specification, and that the claim could only encompass a level of hemihydrate that could be found through detection techniques that existed as of the date that the Canadian patent was granted.

Heneghan J. reviewed the case law on construction of claims, including the Supreme Court of Canada’s recent decision in Whirlpool v. Camco. She ruled that “the purposive approach to construction must refer to the actual words used, identify the essential elements of the invention and consider the purpose for which the patent was granted in the first place”, and that an interpretation of a claim could not be so broad as to overstep the stated purpose of the patent.

Heneghan J. held that the claim construction sought by GSK would be “inimical” to the stated purpose of the patent, that is to improve the handling properties of a manufactured drug product that includes paroxetine hydrochloride as the active ingredient. She therefore interpreted the claim as meaning “crystalline paroxetine hydrochloride hemihydrate when it is found in a drug in sufficient quantity that it improves the handling properties of such drug during manufacture”.

Heneghan J. referred to the equitable defence of non-infringement but did not consider it necessary to rule on the issue. She also did not find the patent to be anticipated or obvious; if the public was aware of the end product but not its composition, then it was not made “available to the public”.

Heneghan J. did not cite the Supreme Court of Canada’s statement in Whirlpool that a purposive construction “is capable of expanding or limiting a literal text”. It is clear, however, that she applied this principle in order to exclude from infringement something that would otherwise fall within the literal scope of the claims.

APPEAL FROM U.S. DISTRICT CCOURT DECISION
GSK appealed Judge Posner’s decision to the Court of Appeals for the Federal Circuit. GSK argued that the trial judge erred by: (i) limiting claim 1 to cover only commercially significant amounts of the hemihydrate; (ii) holding that a construction of claim 1 that covers any amount of hemihydrate would render claim 1 indefinite; (iii) creating an equitable defence to infringement based on GSK causing the infringement; (iv) holding that the granting of injunctive relief was a matter of discretion; and (v) excluding GSK’s evidence of induced infringement.

Apotex cross-appealed, contending Judge Posner erred in granting summary judgment that GSK’s clinical trials qualified as experimental use.

The Court of Appeals affirmed Judge Posner’s factual findings that: (i) Apotex’s product will contain trace amounts of the hemihydrate based on GSK’s evidence of seeding and its disappearing polymorph theory; and (ii) Apotex’s product will not contain detectable quantities of hemihydrate.

The Court of Appeals reviewed the patent and file history and found that nothing contained in those documents defined the invention in terms of commercially significant quantities. The Court therefore construed the claim as encompassing, without limitation, the hemihydrate. The Court also held that it was improper for Judge Posner to consider the effects of alternative constructions on validity or infringement. In particular, the Court held it was irrelevant that a broad interpretation would prevent a third party from preparing the prior art anhydrate. The claim was not indefinite just because a defendant could not determine whether its product contained trace amounts of hemihydrate. In other words, the scope of claim was clear even if Apotex’s infringement was not. The Court of Appeals was therefore prepared to find that Apotex infringed based upon the Court’s construction of GSK’s claim.

The Court next turned to the issue of anticipation since Apotex argued that it only intended to practice the prior art. The Court concluded the hemihydrate likely did not come into existence until GSK’s discovery, so the paroxetine hydrochloride that had been produced by Ferrosan prior to GSK’s involvement did not anticipate GSK’s invention because it did not convert to the hemihydrate form. The Court accepted GSK’s disappearing polymorph theory as satisfactorily explaining why Ferrosan’s production of the anhydrate did not create the hemihydrate even though later production of the anhydrate did so.

The Court of Appeals declined to rule on the equitable defence of non-infringement, finding that the patent was invalid on the ground of a prior public use under 35 U.S.C. § 102(b). More particularly, the Court found that the hemihydrate was in public use more than one year prior to the U.S. filing date since GSK’s clinical trials were not an experimental use. The Court found that the invention was ready for patenting and was used by persons other than the inventor who were under no confidentiality obligation: patients and administering physicians. GSK argued that the clinical trials were an experimental use. The Court of Appeals, however, held that the experimental use had to be of the claimed invention; in the case before it, the claim did not cover the use of the hemihydrate as an antidepressant.

On the issue of gastrointestinal infringement, the Court of Appeals declined to rule but noted that a recent case (Schering v. Geneva) had held a compound claim anticipated because the prior art substance metabolized into the claimed compound upon ingestion by a patient. Recognition of the conversion process at the time of the prior art was not necessary to prove inherent anticipation. Therefore, if GSK proved contributory infringement by showing that the anhydrate metabolized into the hemihydrate upon ingestion, it may also have proven that the hemihydrate was inherent in the prior art.

CONCLUSION
The different treatment of GSK’s claim in Canada and the United States is certainly attributable to differences in the patent schemes of the two countries and the manner in which Canadian and American courts approach patent construction, contributory infringement and related issues.

However, given that Canadian NOC proceedings are not final determinations of infringement or validity, the issues considered by Judge Posner and subsequently by the Court of Appeals for the Federal Circuit may yet be influential in infringement proceedings in Canada.

This publication is made available on the Blakes™ Web site to provide comments on legal issues of significance. Comments reflect the view(s) of the individual author(s) and are not intended to provide legal advice. For specific advice or information, please contact a member of the Intellectual Property Practice Group directly. For permission to reprint articles, please contact Blakes™ Marketing Department at 416.863.2403 or lynn.spencer@blakes.com.

©2004 Blake, Cassels & Graydon LLP

Back to Top