AstraZeneca Pharmaceuticals LP v. Teva Pharmaceuticals USA, Inc. (Fed. Cir. 2009)

Sometimes, Sins of Omission Are Not Inequitable Conduct

    By Kevin E. Noonan --

St. Thomas Aquinas recognized two types of sin:  those of omission and those of commission.  In patent law, inequitable conduct comes closest to this concept of sin, where frequently the accused behavior is the failure to submit to a patent examiner during prosecution information that is material to patentability.  However, mere failure to submit is not enough; there must be some evidence of deceptive intent.  Kingsdown Medical Consultants, Ltd. v. Hollister Inc., 863 F.2d 867 (Fed. Cir. 1988) (en banc).  These principles were reaffirmed last Friday by the Federal Circuit in AstraZeneca Pharmaceuticals LP v. Teva Pharmaceuticals USA, Inc.

AstraZeneca sued Teva and Sandoz pursuant to 35 U.S.C. § 271(e)(2)(A) for filing ANDAs for a generic version of Seroquel®, an antipsychotic agent.  Each ANDA contained a Paragraph IV certification that AstraZeneca's U.S. Patent No. 4,879,288 (the '288 patent), its Orange Book-listed patent for Seroquel®, was invalid and/or not infringed.  The patent claims quetiapine (the generic name for Seroquel®), an atypical antipsychotic having the chemical structure:

(An "atypical" antipsychotic agent is characterized by what it does not do:  it does not cause involuntary body movements such as "torsion spasms, muscle spasms . . . dystonia's of the face neck or back with protrusion of the tongue and tonic spasms of the limbs (dyskinesias).")  The District Court granted summary judgment that AstraZeneca was not guilty of inequitable conduct in obtaining the '288 patent.

The Federal Circuit agreed, in an opinion by Judge Newman, joined by Judges Rader and Prost.  The material information asserted by Teva and Sandoz to have been omitted by AstraZeneca concerned a number of quetiapine analogs, including the following:

AstraZeneca submitted several prior art references relating to these compounds, specifically a German-language publication (Compound 210786), U.S. Patent No. 3,389,139 to Schmutz ("Schmutz I"; Perlapine), U.S. Patent No. 3,539,573 to a second Schmutz reference ("Schmutz II"; Compound 24028), and U.S. Patent No. 4,308,207 to Hunzikerr (Fluperlapine).

The Examiner issued an obviousness rejection on chemical structural obviousness grounds over the Schmutz II compound in view of additional references, including U.S. Patent No. 4,097,597 to Horrom disclosing the Horrom Compound, and another species from the Schmutz II reference, termed Schmutz X by the parties.

AstraZeneca responded to the rejection with argument regarding the unpredictability of "the critical physiological property of atypicality," wherein the prior art did not provide a rationale for choosing the substitutions resulting in quetiapine.  These arguments proved unavailing, with the Examiner demanding that AstraZeneca "provide proof that the prior art compounds do not necessarily or inherently posses the characteristics of the claimed compounds."  Specifically, the Examiner required that AstraZeneca overcome his structural obviousness rejection "by a side-to-side comparison with the closest prior art compound(s)."  The Examiner required this evidence for the Horrom compound and the Schmutz X compound but none of the other compounds disclosed by AstraZeneca.

The patent applicant responded with a declaration by one of the inventors, Dr. Migler, with data for each of the asserted compounds.  In his declaration, Dr. Migler attested to the antipsychotic activity not of the Schmutz X compound but a related compound termed Schmutz B:

The reason for this substitution provided to the Examiner by AstraZeneca was that data for Schmutz X did not exist and "would be very expensive to generate right now."  The data Dr. Migler submitted with his declaration showed that the antipsychotic behavior of these compounds was "typical."  In addition, Dr. Miglar attested in his declaration that another compound, termed "Schmutz A," showed no antipsychotic behavior; this compound differed from the quetiapine compound in having a chlorine atom substitution in the left benzene ring of the heterocyclic structure and having a methyl group as a side chain.

Defendants argued at trial, and before the Federal Circuit, that what AstraZeneca did not disclose to the Patent Office was internal data showing atypical antipsychotic behavior for perlapine, fluperlapine, Compound 21076, and Compound 24028, and that withholding this data was "deliberately misleading."  The essence of Defendants' argument, according to Judge Newman's opinion, was that "AstraZeneca presented internal test data about similar compounds that were typical while omitting internal test data about similar compounds that were potentially atypical."  This argument did not convince the District Court, which ruled that AstraZeneca had disclosed the closest prior art and had addressed the arguments and provided the evidence required by the Examiner.

Judge Newman's opinion noted that there was no assertion that AstraZeneca withheld any "relevant" reference.  The art AstraZeneca did cite (the Hunzinger patent, for example) disclosed atypical antipsychotics including fluperlapine, and any internal data possessed by AstraZeneca was thus cumulative.  The Federal Circuit rejected the Defendants' allegations that AstraZeneca made material misrepresentations during prosecution (such as it being "too expensive" to provide data for the Schmutz X compound) as being without evidence, nor was there evidence that, had AstraZeneca made and tested the Schmutz X compound, material evidence would have been produced, said the Court.  The Court said that the unpredictability of whether a particular analog had atypical antipsychotic properties was "undisputed," and that "the record demonstrates that structural similarity is not a predictor of whether antipsychotic behavior would be typical or atypical."  The Court also assessed whether the compounds having data regarding antipsychotic behavior that AstraZeneca did not disclose were "equally close" in chemical structure to the compounds whose data was disclosed and to quetiapine.  In its analysis, the Court found that substitution of the Schmutz B compound data for Compound 24028 data was not a material representation, based on structural grounds (and despite evidence that their antipsychotic behavior differed in ways that did not support AstraZeneca's argument).  The Court's analysis rested on its assessment that AstraZeneca submitted data for the "structurally closest" compounds, and the compounds asserted by the Examiner.  The Court opined:

A reasonable examiner would not have understood the Migler declaration as stating that no prior art product had the atypical property shown by quitiapine, for it was known that other atypical antipsychotics existed . . . .  A reasonable examiner would understand AstraZeneca's statements to refer to the closest prior art compounds, not all prior art compounds.

Thus, according to the opinion, the evidence did not support a finding that AstraZeneca misrepresented or omitted material information, despite AstraZeneca's failure to disclose all its data for several of these prior art compounds.

Similarly, the Court found that the Defendants did not establish deceptive intent by clear and convincing evidence.  Defendants' argument suffered because it suggested that, due to the "high degree of materiality" it had shown regarding the withheld information, "they therefore need a proportionally lesser showing of intent to deceive to establish the requisite threshold level of intent."  The Court was succinct in its evaluation of this argument:  "[t]hat is incorrect," it said, and "[e]vidence of mistake or negligence, even gross negligence, is not sufficient to support inequitable conduct . . .," citing Kingsdown Medical.  A court must balance evidence of materiality and deceptive intent, said the Federal Circuit, but only after threshold levels of materiality and deceptive intent have been shown by clear and convincing evidence.  The Defendants' evidence of deceptive intent was merely the existence of the undisclosed data showing that some of the prior art compounds had atypical antipsychotic behavior and AstraZeneca's failure to expressly disclose this evidence to the Patent Office.  Intent to withhold, however, is not the same as intent to deceive, according to the Court, and "[i]ntent to deceive cannot be inferred simply from the decision to withhold [information] where the reasons given for withholding are plausible," citing Dayco Products Inc. v. Total Containment, Inc.  In the absence of any evidence of bad faith, the Court found AstraZeneca "presented plausible reasons for its presentation of arguments and data during the prosecution" which did not support a finding of deceptive intent.

Regardless of the soundness of the Federal Circuit's opinion, the case illustrates how difficult making an inequitable conduct determination can be.  Even in situations where no affirmative sins of commission appear to have been committed, omissions that are sinful can be devilishly difficult to discern, and their motivations even more so.

AstraZeneca Pharmaceuticals LP v. Teva Pharmaceuticals USA, Inc. (Fed. Cir. 2009)
Panel:  Circuit Judges Newman, Rader, and Prost
Opinion by Circuit Judge Newman

Comments

[Dan Feigelson]

A welcome outcome, but hardly surprising, given that Judges Newman and Rader were on the panel. At least their positions have been consistent over the years. What do you make of Judge Prost's position here versus her position in Aventis v Amphastar?

[EG]

Kevin,

Nice article. But my question remains: what is the difference between Compound 24028 and Schmutz X? From the formulas in the case, these look to be structurally identical.

[Kevin E. Noonan]

Dear Dan:

I think the difference illustrates the problem - inequitable conduct cases will always depend heavily on the facts and the "smell test" - does it look like something underhanded was done? And, of course, the panel - I suspect Judges Newman and Rader can be very persuasive that finding inequitable conduct should be rare, and particularly rare in cases where the district court doesn't think it exists, and even more rare when the district court grants summary judgment of no inequitable conduct.

But I agree it isn't a surprise. Thanks for the comment.

[Kevin E. Noonan]

Dear EG:

There must have been a typo in the case, since you are correct, they are identical.

Thanks for noticing.

[EG]

Kevin,

Thanks for confirming that what I saw isn't a figment of my imagination. If it is a typo, then I wonder which of the formulas for Compound 24028 and Schmutz X is correct.

I've got any even scarier thought: what if the formulas shown in the case are correct, i.e., Compound 24028 and Schmutz are the same. If they are correct/the same, then Teva/Sandoz has a real beef about why AstraZeneca represented that they had no data on Schmutz X.

We'll have to wait and see if the Federal Circuit issues a correction notice.

[Mike]

The point isn't that they didn't have data on 24028. The point was that Schmutz B was more similar to quetiapine than Schmutz X and that the existing compounds were unpredictable.

That said, maybe they should have presented all compounds and argued that the effects of "simple" substitutions were not predictable.