The standard of review at the Federal Circuit is important. Even though it often appears like cases are re-litigated at the Appeals Court, it is actually very difficult (or at least should be) to overturn a lower court's factual determinations. This is especially true for a finding that a patent was not shown to be invalid, because the burden at the trial court is clear and convincing evidence. Moreover, the standard of review for any factual determination made by the lower court is "clear error." Such a high standard might have saved Pfizer's Lyrica® patents from being invalidated earlier today in Pfizer Inc. v. Teva Pharmaceuticals USA, Inc.
Lyrica® is a drug for treating seizures and certain types of pain. The present case was brought pursuant to the Hatch-Waxman scheme of 35 U.S.C. § 271(e)(2) after several companies filed ANDAs to market a generic version of this drug. These companies included Teva, Sun Pharma Global, Inc, Mylan Pharmaceuticals, and about nine other companies. Plaintiffs, which included Pfizer, CP Pharmaceuticals International C.V., Warner-Lambert Company LLC, and Northwestern University, asserted two patents, U.S. Patent No. 6,197,819 ("'819 patent") and U.S. Reissue Patent No. RE41,920 ("RE '920 patent"). Ultimately, only claim 2 of the '819 patent was at issue on appeal. This claim reads:
2. 4-amino-3-(2-methylpropyl) butanoic acid, or a pharmaceutically acceptable salt thereof.
This compound is also known as 3-isobutylGABA. The active ingredient in Lyrica® is the S-enantiomer of this compound, which is specifically claimed in claim 1 of the same patent. Based on the lower court claim construction of claim 2, the Appellants stipulated to infringement. In addition, the trial court found that the Appellants had not established with clear and convincing evidence that claim 2 was invalid as non-enabled, not described, or obvious. As their moniker would suggest, the Appellants appealed these determinations.
Claim Construction and Infringement
The first issue was the construction of the claim term "4-amino-3-(2-methylpropyl) butanoic acid." Of course, claim construction is a matter of law, which the Federal Circuit reviews de novo, thanks to Cybor Corp. The lower court had construed this term to cover 3-isobutylGABA generally, and therefore it would cover racemic mixtures, pure enantiomers, or any ratio in between. The Appellants alleged that the term should be limited to the racemic (50:50) mixture, based on statements made in the specification and the prosecution history. They were, of course, advocating such a construction because their proposed products contained non-racemic mixtures, and therefore would have avoided infringement.
Nevertheless, the Federal Circuit agreed with the lower court that the plain and ordinary meaning of the claim term does not include such a narrowing limitation. In fact, the Court noted that the use of the terms "racemate," "racemic," the prefix (R,S) by the inventors in the application indicated that they had no such intention to so limit the claims. "[T]he patentee's use of the prefix in the tables demonstrates that it knew how to specify racemic 3-isobutylGABA as distinguished from the compound generally and chose not to do so in claim 2." Slip Opinion at 9-10. The Appellants pointed to the fact that only racemic mixtures were discussed in the "results" section of the specification. However, "[a]bsent a clear disavowal or lexicographic definition in the specification or the prosecution history, the reporting of test results limited to a racemate does not warrant importing a racemic limitation into claim 2." Because Appellants stipulated to infringement under this construction, the Court affirmed that finding.
Even though enablement is a question of law, and therefore reviewed without deference, the underlying factual inquiries are reviewed for clear error. The District Court had held that a person of ordinary skill in the art could have made the claimed 3-isobutylGABA compound with no more than routine experimentation based on the disclosure in the '692 application's specification. The '692 application was filed in 1990, and if it did not enable claim 2, intervening art before the '819 patents filing date may have invalidated the claim. The Appellants had argued that, because the 3-isobutylGABA claim term was construed to cover all enantiomeric ratios, the patent disclosure must teach "every conceivable mixture," including 100% pure enantiomers. The trial court had found that there was no dispute that the co-inventors were the first to create and claim the "3-isobutylGABA" chemical compound. In addition, the court had found that the specification stated that "enantiomers may be prepared or isolated by methods already well known in the art," and that there was sufficient support for this statement based on "the prior art, the prosecution history, and witnesses' trial testimonies." Id. at 12.
The Federal Circuit did not believe that this determination was clearly erroneous, and therefore it affirmed that claim 2 was not invalid for lack of enablement. In response to the Appellants' contention that the specification provides no guidance as to how to purify the two enantiomers, the Federal Circuit stated that "the inventors were not required to provide a detailed recipe for preparing every conceivable permutation of the compound they invented to be entitled to a claim covering that compound." Id. The Court pointed out that "[w]here a claim has been construed to cover a chemical compound, the specification is not deficient merely because it does not disclose how to prepare a particular form or mixture -- among hundreds of possible permutations -- of that compound." Id. In other words, the Federal Circuit appears to be saying that to satisfy the enablement requirement, one need not enable the entire scope of genus claim. This is at least true if the methods of generating the full scope are well known in the art.
Of all of the Appellants, only Sun pursued invalidity of claim 2 for failing the written description requirement. This determination is a question of fact, and therefore is reviewed for clear error. Sun's principal argument was that the '692 application discloses nothing that would suggest that the inventors could isolate the individual enantiomers. As such, they could not have been in possession of the entire scope of claim 2. In fact, later applications provided support for the proposition that the inventors were not able to separate the enantiomers as of the filing date of the '692 application. Nevertheless, the Federal Circuit found that Sun's argument was "surprisingly similar" to the other parties' position on enablement. Important in the Court's decision was that claim 2 was a genus claim that encompassed all mixtures of the enantiomers. "But written description does not require inventors, at the time of their application for a patent, to reduce to practice and be in physical possession of every species . . . of a genus . . . claim." Id. at 14. It probably did not help Sun's cause that the lower court found that it presented a "dearth of evidence at trial . . . ." Id. at 13. As such, the Federal Circuit affirmed the lower court's finding that the '692 application contained an adequate written description for claim 2.
Finally, the Federal Circuit affirmed the lower court's determination that the claimed invention was not obvious. Such a determination is legal in nature, but it is based on underlying facts. The Appellants' argued on appeal that the lower court committed clear error for failing to find the following: (1) that three references, U.S. Patent No. 4,322,440 ("Fish"), U.S. Patent No. 5,051,448 ("Shashoua"), and a 1962 article ("Colonge"), taught that isopropylGABA may have anticonvulsant activity; (2) that a skilled artisan would have expected 3-isobutylGABA would have had the same activity as isopropylGABA because of structural similarity; and (3) 3-alkylGABA, a prior art compound, provided the motivation to try other alkyl substituents at the 3-position. However, the Court suggested that the Appellants' real failing was that "the evidence presented at trial was too sparse." Id. at 17.
The determination that a new chemical compound is obvious is a two-step process. First, the court must determine whether a particular "lead" compound would have been selected by a skilled artisan. Second, it must be established that one skilled in the art would have been motivated to modify the lead compound to derive the claimed invention. Unfortunately for Appellants, the record below contained "scant evidence" to establish either gabapentin or 3-isopropylGABA as lead compounds. Even though Appellants may have established that both of these compounds were being tested for anticonvulsant effects, Appellants presented "no testimony establishing its being tested prior to the discovery of 3-isobutylGABA." Id. at 18. With regard to gabapentin, the Court found that "there was no evidence in the record of motivation for a skilled artisan to modify gabapentin for further anticonvulsant research." Id. The evidence related to 3-isopropylGABA was described as "just as meager." Id. Even though Fish, Shashoua, and Colonge may have disclosed the compound, the lower court found that the record lacked "any explanation for 'what structures were important for anticonvulsant activity[.]'" Id. As such, there was simply no reason to select 3-isopropylGABA for further modification.
There was also a lack of evidence to establish the second step. Appellants alleged that the three prior art references would have taught modifying the 3-position with "lower alkyl substitutes." However, even though these references disclosed "trillions" of compounds, they did not call out alkyl groups, much less isobutyl substitutions. "Indeed, a vague suggestion in the prior art pointing to a broad class of compounds, without any teaching particularly identifying isobutyl among the millions of potential compounds, is not a teaching of 'specific molecular modifications' required by [Federal Circuit] precedent." Id. at 20. The lower court had found that Pfizer credibly established that drug discovery in the anticonvulsant field was "complicated," "highly unpredictable," and "largely conducted through trial and error." Id. As such, there would have been no reasonable expectation of success to teach any substitution at the 3-position.
The important take away lesson from this case is to focus on the standard of review at the Federal Circuit, especially if you are challenging a finding that a patent is not invalid. It is also essential that get all of your evidence in at trial. This is because if you need to prove that the lower court's determination was clearly erroneous, you had better have presented sufficient evidence at trial to establish that fact.
Pfizer Inc. v. Teva Pharmaceuticals USA, Inc. (Fed. Cir. 2014)
Panel: Chief Judge Rader and Circuit Judges Prost and Moore
Opinion by Circuit Judge Prost