Center for Advanced Study & Research on Intellectual Property


CASRIP Newsletter - Summer 2007, Volume 14, Issue 3

Federal Circuit Judges Split Their Views on Obvious-to-Try in Nonobviousness of Chemical Compounds

Pfizer, Inc. v. Apotex, Inc.,[1]
Takeda Chem. Indus., Ltd. v. Alphapharm Pty., Lty.,[2]

By Toshiko Takenaka

On May 27, 2007, Federal Circuit denied the patent owner’s request for rehearing en banc with respect to a panel decision which found the patent invalid and reversed a judgment for the patentee. Three dissenting opinions authored by Judges, Newman, Laurie and Rader highlighted different views in applying the obvious-to-try standard for chemical compounds in the post-KSR era.

In Pfizer, the patent at issue, U.S. Patent No. 4,879, 303 (“the ‘303 patent”) is directed to amlodipine besylate which is contained in the drug sold under the trademark “Norvasc”, for treating hypertension and angina. In 1982, Pfizer filed in the U.K. a patent application for certain dihydropyridine compounds and their pharmaceutically-acceptable acid addition salts. Following a US counterpart application claiming priority from the U.K. patent, a U.S. patent was issued in 1986 as U.S. Patent No. 4,572, 909 (“the ‘909 patent”). The ‘909 patent discloses that the pharmaceutically-acceptable acid addition salts of amlodipine "are those formed from acids which form non-toxic acid addition salts containing pharmaceutically acceptable anions, such as hydrochloride, hydrobromide, sulphate, phosphate or acid phosphate, acetate, maleate, fumarate, lactate, tartrate, citrate and gluconate salts," and that the preferred salt is maleate.

For commercialization of a drug, Pfizer researchers originally selected amlodipine meleate which later was found unsuitable for a direct compression tablet product due to chemical instability and stickiness of the tablet blend of amlodipine meleate. After testing the stability of different salt formulations, Pfizer researchers selected the besylate salt which showed a better stability over the meleate. In 1986, Pfizer filed a patent application in the U.K. for the amlodipine besylate, and a counterpart patent application in US in 1987. During the prosecution proceedings in the United States, all claims were rejected for obviousness over the ‘909 patent in view of two references. One disclosed that aryl sulphonic acid salts, which include besylate, are superior to the maleate of the '909 patent. The other reference provides an example of a pharmaceutical compound wherein the besylate form is specifically identified as the preferred embodiment. To overcome the rejection, Pfizer emphasized the unexpected properties of the besylate salt of amlodipine and the fact that one skilled in the art would not find it as an obvious compound. Following two continuation applications, the claims for amlodipine besylate were issued as the US ‘303 patent.

Pfizer filed suit against Apotex, alleging that Apotex’s filing of an abbreviated new drug application with FDA regarding amlodipine besylate tablets before the expiration of ‘303 patent infringe on that patent. Apotex denied infringement, challenging the validity of ‘303 patent on several grounds including obviousness. The District Court agreed with Pfizer, finding the ‘303 patent valid and infringed. It rejected Apotex’s obviousness argument because there was no expectation of success in making besylate salt of amlodipine and because of the unexpected superior property in the amlodipine besylate.

On Appeal, a Federal Circuit panel consisting of Judges Michel, Mayer and Lynn agreed with Apotex that amlodipine besylate is obvious over the ‘909 patent in view of the “Berge” reference.[d1] First, it found clear error in the district court’s nonobviousness analysis for its improper placement of burden of proof on each party. This was because the district court’s obviousness analysis relied heavily on the examiner’s determination during the prosecution. As Stated, an examiner’s interim finding of prima facie obviousness cannot render the claims of an issued patent prima facie obvious, because courts must presume a patent valid. Thus the patent challenger bears the burden of proving the patent invalid by clear and convincing evidence.

The Federal Circuit panel also found the district court’s nonobviousness ruling incorrect because Apotex showed with clear and convincing evidence that a person having ordinary skill in the art (PHOSITA) would have been motivated to combine the ‘909 patent and Berge, to product the amlodipine besylate. The court rejected Pfizer’s argument that its ‘909 patent did not suggest a motivation to make amlodipine besylate because none of the anions disclosed in the ‘909 patent had a cyclic structure like besylate. However, it found a motivation to make amlodipine besylate from the nature of the problems that the inventor sought to solve in the ‘303 patent. The court reemphasized the flexible nature of its teaching-suggestion-motivation test highlighted in Dystar.[3] As stated, a motivation or suggestion to combine the prior art references can be found explicitly or implicitly from different sources, including common knowledge, the prior art as a whole or the nature of the problem.

Pfizer argued that the Berge reference practically taught away from combining with the ‘909 patent, due to the fact that it showed that the besylate was rarely used in the pharmaceutical industry. This argument was also rejected by the panel which held that PHOSITA would have been indeed motivated to make the amlodipine besylate in view of its known acid strength, solubility and other known chemical characters discussed in other publications. The panel found a reasonable expectation of success because a Pfizer researcher is able to select seven alternative anions including besylate as soon as the tablet processing problem is revealed. It noted that a showing of some degree of unpredictability in the art is not sufficient to establish nonobviousness.

Pfizer was also out of luck with the obvious-to-try argument. The panel rejected Pfizer’s argument that amlodipine besylate would be at most obvious-to-try because all parameters would need to be varied to seek a successful result. It pointed out that the only parameter to be varied is the anion with which to make the amlodipine acid addition salt, and that prior art motivates PHOSITA to try 53 pharmaceutically acceptable anions. The panel was not persuaded by Pfizer’s policy argument that application of the obvious-to-try analysis disadvantageously impacts research activities in the pharmaceutical industry.

Instead, it emphasized that it did not find amlodipine besylate obvious for the manner in which the invention was made. Instead, the panel’s obviousness conclusion was based on the fact that PHOSITA would have had a reasonable expectation of success in combining amlodipine with besylate salt. Describing the process for selecting anions analogous to the optimization of a range or other variable within the claims that flows from the normal desire of PHOSITA, the panel held obvious with respect to the optimization of the acid addition salt formulation for an active pharmaceutical ingredient if the formulation does not result in any therapeutic effectiveness of the ingredient and the prior art heavily suggests the particular anions used to form the salt.

Failing to prevent Apotex from establishing a prima facie case of obviousness, Pfizer argued non-obviousness relying on the evidence of secondary considerations. The panel found it errorneous that the district court compared amlodipine besylate with the amlodipine maleate to establish an unexpected result. The panel reasoned that the amlodipine maleate is not the closest prior art compound to amlodipine besylate. Further, any superior result in the patented compound over the closest prior art compound should be unexpected by PHOSITA. Nevertheless, Pfizer’s evidence fails to include what PHOSITA would have expected at the time of invention. The panel struck down the district court’s finding of unexpected result with respect to the superior physicochemical characteristics in amlodipine besylate. The evidence showed that amlodipine maleate works as good as amlodipine besylate as the active ingredient in the drug for its intended purpose. Pfizer’s routine testing to select an optimal salt for commercialization is universal and within common sense. Finally, the panel noted that the strong showing of obviousness in this case does not overcome by the evidence of secondary considerations, even if Pfizer were assumed to show the unexpected superior result of amlodipine besylate.

Pfizer’s petition for rehearing en banc was referred to the panel that heard the appeal and then referred to other judges. The court voted not to hear the case en banc. Judges Newman, Lourie and Rader dissented on whether they would rehear the case en banc. The three judges criticized the panel’s application of the obvious-to-try standard as being in direct conflict with the court’s precedent. As argued, the panel erred when deciding to find the amlodipine besylate obvious despite its unpredictable properties of amlodipine besylate from the properties of a besylate salt of a different compound. They also found it an error to refuse to give weight on the evidence of unexpected results because amlodipine besylate produces superior properties for sticking and instability problems of the prior art, without any trade-offs in salt properties which are expected by PHOSITA.

Judge Lourie wrote a separate dissenting opinion and found the panel in error for failing to defer to fact-findings by the district court that were not clearly erroneous regarding the unexpected properties of amlodipine besylate. As stated in his opinion, the panel also substituted its own finding with that of the district court with respect to a reasonable expectation of success in making amlodipine besylate from the prior art. He also found that the panel improperly placed greater therapeutic value over the “physical properties”. It is not necessary for a compound to be patentable to possess improved therapeutic properties. As emphasized in his opinion, “physical properties” such as stickiness and stability in this case are as important as therapeutic properties for commercializing chemical compounds. Judge Rader’s opinion focused on the inappropriateness of applying the obvious-to-try standard in unpredictable pharmaceutical inventions.

Interestingly, in Pfizer none of the three dissenting judges mention the ruling on the obvious-to-try standard in KSR.[4] As noted by the Supreme Court in the KSR case, when examining a combination of elements disclosed in different references in a predictable mechanical art, an adjustable pedal mechanism, a combination is obvious if there is a design need or market pressure to solve a problem and there are a finite number of identified, predictable solutions. Although the panel decision was issued before KSR was published, the panel’s application of the obvious-to-try standard shows a striking similarity with that of KSR in identifying the problems to be solved, sticking and stability problems, and emphasizing the limited number of anions which the inventors needed to try for solving the problems. However, dissenting judges tend to distinguish KSR from this case because of unpredictable nature of properties in pharmaceutical compounds.

Judge Michel who authored the panel decision acknowledged the merit of policy concerns indicated by pharmaceutical industry, that many innovations in this industry resulted from trial and error procedures and the obvious-to-try standard adversely impact pharmaceutical research. Nevertheless, He still believes that the panel’s obviousness finding in this case is appropriate, not because of the routine test used by the inventors, but because of the reasonable expectation of success of PHOSITA in making amlodipine besylate.

In KSR, the Supreme Court distinguished a simple substitution of one known element for another from a more complicated substitution and modification. Judge Michel tends to view amlodipine besylate as the former type of substitution as describing the optimization of addition salt is within the ordinary skill of PHOSITA. The panel’s analysis proves serious concerns by patent applicants and patentees if KSR would make it too easy for courts and USPTO to classify inventions into the former type and allow prima facie obvious as common sense without an analysis of the reason for combining multiple references. However, the KSR Court classified the invention into the first type, but analyzed a reason for PHOSITA to combine the references to make the invention. KSR should not be interpreted to eliminate a requirement of a reason for combination for a simple substitution.

The panel’s rejection of teaching away argument also reinforces such a concern because, both the no disclosure of besylate salt and the statement that the use of besylate is rare, would not convince the panel that the prior art would have discouraged PHOSITA from combining the ‘909 patent and Berge. As supported by the Supreme Court in KSR, the panel’s rejection of teaching away argument indicates a future trend to negate a teaching away argument with respect to chemical compound inventions unless the prior art does not show any benefit for adopting a substitute formation selected by the inventor.

An even more alarming signal for the pharmaceutical industry is the limited role given by the panel with respect to the evidence of unexpected results to rebut a prima facie obvious case. The panel emphasized the strength of prima facie obvious in this case and made it clear that such cases cannot be rebutted by the evidence of an unexpected result although, as Judge Rader points out, three separate district courts upheld the validity of the ‘303 patent. Liberating the obvious-to-try patentability negation with the weakening force of teaching away argument and an unexpected result rebuttal, may present a high hurdle for patenting variations of chemical compounds.

However, a more recent decision authored by Judge Lourie, a Pfizer dissenting judge, cited KSR and set the pharmaceutical industry at ease by reviving the pre-KSR obvious-to-try standard and confirming that KSR does not change the nonobviousness hurdle for chemical compounds. In Takeda Chemical, Judge Laurie noted that KSR acknowledged a requirement of reason for a combination invention, which is consistent with the pre-KSR Federal Circuit precedent in determining nonobviousness of chemical compounds. Such precedent requires a showing of structural similarity between the prior art and claimed invention and an adequate support in the prior art for the change from the prior art to establish a prima facie obvious case.[5]

Judge Lourie agreed with the district court that the patent directed to a drug treating type 2 diabetes. He distinguished Takeda Chemical from Pfizer in that a large number of compounds present possible solutions for the problem to be solved by the invention art, and thus there was no reasonable expectation of success to make the claimed compounds. Further, the court found that the process of modifying the prior art compounds is not routine at the time of invention. Nothing in the prior art suggests the non-toxicity property in the claimed compounds by the modification.

In short, there is a significant difference among federal circuit judges in interpreting KSR case with respect to the extent of impact on the nonobviousness standard. These two cases clearly show a difficulty to reach a consensus even if the judges decide to hear a case en banc to minimize the gap.

Top of Page


Top of Page

Last updated 4/27/2012