Can the claim term "substantially pure" mean two different things when the specification uses the same term to refer to both an intermediate compound and a final drug product in a specification?  The Federal Circuit recently answered the question in the affirmative, in Aventis Pharma. Inc. v. Amino Chemicals Ltd., depending of course on the context that the term is used in the claims.  The Court also focused on the claim term "substantially pure regioisomer," instead of the truncated "substantially pure" from the lower court's decision, and determined that the term required "largely but not wholly the para-regioisomer of the intermediate of the structure shown, as compared to the meta isomer."  Judge Bryson in dissent, however, found it more compelling that the patentees appeared to use the term "substantially pure" interchangeably in both the specification and the prosecution history when referring to both the intermediate and final drug products.  Therefore, he would have affirmed the lower court's claim construction of "substantially pure" to require "at least 98% purity with respect to all impurities."

The technology at issue in this case was the process for producing large quantities of fexofenadine, the active ingredient of Allegra^® and Allegra-D^® 24 hour.  The prior art process of making piperidine derivatives such as fexofenadine was inefficient because it required purification steps after the derivatives were fully formed, and this added to the both the cost and the time to obtain a pharmaceutically acceptable final product.  Dr. Thomas E. D'Ambra overcame this problem by using piperidine and cyclopropylketone ("CPK") intermediates earlier in the reaction.  The CPK intermediate could take on at least two conformations, para-CPK and meta-CPK:

but it was the para-CPK that gave rise to the biologically active piperidine derivatives.  The molecule could also adopt the ortho-CPK form, but it was rarely produced and was of little biological efficacy.  These different conformations are referred to as regioisomers.  The prior art processes used a "Friedel-Crafts" reaction, which produced a statistical admixture for the "second mixture of aromatic regioisomers" of 67% meta-isomers and 33% para-isomers.  The final products (third mixture) were then obtained by converting these intermediates, although the ratio of the regioisomers persisted.  It was therefore nearly impossible to completely separate the para-isomers to pharmaceutical purity.  Dr. D'Ambra's use of CPK intermediates made it possible to obtain "substantially pure regioisomers" of para-CPK, and consequently a subsequent "substantially pure" para-piperdine derivative end product.

Claim 1 of the patent-in-suit, U.S. Patent No. 5,750,703 ("the '703 patent"), provides (with the claim term highlighted):

1.  A process of preparing a piperidine derivative compound of the formula:

wherein

R₁ is hydrogen or hydroxy;

R₂ is hydrogen;

or R₁ and R₂ taken together form a second bond between the carbon atoms bearing R₁ and R₂;

R₃ is –COOH or –COOR₄ ;

R₄ has 1 to 6 carbon atoms;

A, B, and D are the substituents of their aromatic rings, each of which may be different or the same, and are selected from the group consisting of hydrogen, halogens, alkyl, hydroxy, alkoxy, or other substituents,

said process comprising:

providing a substantially pure regioisomer of the following formula:

converting the substantially pure regioisomer to the piperidine derivative compound with a piperidine compound of the formula:

Even a casual reading of this claim revels that the term "substantially pure" only modifies the intermediate regioisomer and not the final product.  However, the specification uses the same term in reference to the final product of the prior art process, and a divisional patent, U.S. Patent No. 5,578,610 ("the '610 patent") has claims drawn to substantially pure piperdine derivative compounds.

The procedural history of this case was complex and involved dozens of parties in twenty different lawsuits.  The defendants had filed a Drug Master File that was referenced in the Abbreviated New Drug Applications of Mylan Pharmaceuticals Inc. and Teva Pharmaceuticals USA Inc. to market generic versions of antihistamines containing fexofenadine.  Upon submission of the ANDAs, Aventis timely brought suit alleging, among other things, infringement of the '703 patent.  At the claim construction phase, the lower court construed the term "substantially pure" to mean "at least 98% purity with respect to all impurities."  As a result, Aventis stipulated that it could not prove infringement, and so the District Court entered final judgment to allow Aventis to appeal the claim construction ruling.

In reversing the lower court's decision, the Federal Circuit highlighted that claims "must be construed in light of the appropriate context in which the claim term is used" (citing Toro Co. v. White Consol. Indus., Inc., 199 F.3d 1295 (Fed. Cir. 1999)).  In fact, the Court pointed out that the same claim term can have different constructions within the claims and the specification.  However, for the patent-in-suit, the term "substantially pure" was only used in the claims to refer to the CPK intermediate.  The Court apparently discounted the use of the term in the claims of the related '610 case, in reference to the final piperidine derivative end products.  Nevertheless, what was important was how one skilled in the art would understand that term in the context of the claims, and such a skilled artisan would not have expected that an intermediate would have the same purity as the final end product.  After all, the improvement was to provide a piperidine derivative end product of higher regioisomeric purity with less extensive purification, and requiring a consistent construction of this term ignored the distinct contexts in which the terms are used.  As a result, the Court held that the proper construction of "substantially pure" required different interpretations depending on whether it was modifying the intermediate or the end product.

In coming to this conclusion, the Court had to address the District Court's reliance on the intrinsic record that Judge Bryson found persuasive in dissent.  First, as Judge Bryson pointed out, the specification stated:

Although the second mixture of regioisomers [an intermediate] and the third mixture of regioisomers [the final piperidine derivative product] can be analyzed by HPLC experiments, a practical separation to obtain gram quantities of substantially pure regioisomers has not been achieved.

Each mixture (including the first [also an intermediate]), would be expected to contain 33% of the para isomer and 67% of the meta isomer.  Since these components are inseparable, it has not been possible to obtain either of the regioisomers in each mixture in substantially pure form.

'703 patent at col. 4 ll. 16-24.  This was at least part of the evidence that the patentee supposedly used the terms interchangeably.  However, as was pointed out, this use was only in reference to the prior art -- the patentee did not refer to his own piperidine derivative end products as substantially pure in the specification.

More difficult to reconcile were comments made by the patentee during an interference proceeding in the related '610 patent case.  In a filing in that case, the patentee wrote "[w]hen read in light of the specification, one skilled in the art would have understood that the phrase 'substantially pure,' as used in claims 1-17 of [the '610 patent], to mean that the subject compound has a pharmaceutical grade purity and is in a form purer than that attained by the prior art."  Most of claims 1-17 recite a substantially pure piperidine derivative end product.  However, claim 12 of the '610 patent recites "a piperidine derivative compound produced by a process comprising: providing a substantially pure regioisomer . . . ."  Therefore, different claims from claims 1-17 of the '610 patent contain the "substantially pure" claim term to refer to both the intermediate and the end product.  Nevertheless, the Federal Circuit's opinion found this to be of little help.  It pointed out that the focus of the interference proceeding was interpreting claims in reference to the piperidine end product, and so the comments were only in regards to the "subject compound" of the interference -- the end product.  And, since the Court had already determined that the "substantially pure" should have different interpretations, it found no justification for applying the definition provided for this term from the interference.

Finally, having found the lower court's construction incorrect, the Federal Circuit performed its own claim construction.  It noted that the term "substantially" is often not amenable to numerical boundaries.  It was clear from the '703 patent that the regioisomeric purity should be greater than 67% (because this was the purity in the prior art process), but it did not need to be as high as 98%.  So, the Court borrowed a construction of the term from Ecolab, Inc. v. Envirochem, Inc., 264 F.3d 1358 (Fed. Cir. 2001), in which "largely but not wholly" was used as a flexible approach.  Therefore, the Court adopted the definition of "largely but not wholly the para regioisomer of the intermediate of the structure shown, as compared to the meta isomer."  The case was reversed and remanded to determine the other issues in the case consistent with this claim construction.

MBHB represented Aventis in the above appeal.  To the extent that this case summary contains any opinions, the opinions would be of Dr. Williams and not Aventis or MBHB.

Aventis Pharmaceuticals Inc. v. Amino Chemicals Ltd. (Fed. Cir. 2013)
Panel: Circuit Judges Newman, Bryson, and Reyna
Opinion by Circuit Judge Reyna; dissenting opinion by Circuit Judge Bryson