In this Hatch-Waxman case, Bristol-Myers Squibb, owner of the drug entecavir (sold as Baraclude®, indicated for hepatitis B (HBV) infection), sued Teva Pharmaceuticals for patent infringement. Teva responded that the patent (the only patent at issue was US5206244, priority date 10/18/1990), which claims the chemical structure of entecavir, was obvious. The district court found in favor of Teva. This finding was affirmed by the Federal Circuit panel. En banc review was denied.

The obviousness argument was that 2’-CDG was the lead compound, and that entecavir was merely a minor modification of 2’-CDG, termed a “small conservative change.”

Fig1

 

Teva argued that 2’-CDG in turn is structurally similar to the natural nucleoside deoxyguanosine. Teva also argued that “Madhaven Compound 30,” a 2’-hydroxy analog of entecavir (and §102(b) prior art to the ‘244 patent) is the exo-methylene analog of aristeromycin. Madhaven compound 30 was said to have substantially improved antiviral properties as compared to aristeromycin.

Fig2

The panel found no clear error in the district court findings that entecavir was obvious in view of:

  • Structural similarity between entecavir and 2’-CDG;
  • Teachings of Madhavan;
  • The exo-methylene substitution is a “small conservative change”; and
  • The totality of the prior art on 2’-CDG

The panel found that a skilled artisan would have been motivated to substitute an exocyclic methylene at the 5’ position of 2’-CDG with a reasonable expectation of success of creating a compound with beneficial antiviral properties.

The most significant feature of this case is the treatment of post-filed data. BMS argued that several features of entecavir were discovered after the ‘244 patent was filed in 1990, but the district court discounted these arguments, which were all upheld in Chen’s opinion finding entecavir obvious.

Thus, BMS challenged the selection of 2’-CDG as a lead compound because it was found to be toxic in the 1990’s, after the file date of the ‘244 patent. The panel dismisses this argument, asserting that obviousness is measured at the time of the invention, citing to Velander v. Garner, 348 F.3d 1359, 1377 (Fed. Cir. 2003). Likewise, Madhavan compound 30 was found to more toxic than aristeromycin after the ‘244 patent was filed, but again Chen asserts that the reasonable expectation of success provided by Madhavan is measured as of the date of the invention, citing to Amgen v. Roche, 580 F.3d 1340, 136 (Fed. Cir. 2009).

BMS also argued that post-filed data supported its arguments that secondary considerations rebutted the prima facie case of obviousness. BMS argued that entecavir had the unexpected properties of (1) high potency against hepatitis B, (2) a larger than expected therapeutic window, and (3) a high genetic barrier to resistance. All of these features were discovered after the ‘244 patent was filed, but Chen agreed with the District Court that entecavir was expected to have some antiviral activity based on the similarity to 2’-CDG, and Chen asserted that these unexpected properties were merely differences in degree.

However, Judge Newman’s dissent forcefully argues the panel disregarded the long established practice that comparative post-filng data can be used to rebut a prima facie case of obviousness and can be used to show a new property or use unexpected in light of the prior art. 769 F.3d at 1347-1348.

BMS also argued that a new chemical entity cannot be obvious as a matter of law if unexpected properties are demonstrated. Chen dismisses this argument, holding that an unexpected result or property does not by itself support a finding of non-obviousness, 752 F.3d at 977, citing to In re Dillon, 919 F.2d  , 693, 697 (Fed. Cir. 1990). Judge Taranto responded to this argument in his dissent to the denial of en banc review, arguing that the panel took the statement from Dillon out of context. 769 F.3d at 1354. Tarranto argues that in the reasonable expectation of success prong, the panel erred by not considering “particular unexpected results” pertaining to efficacy and safety. This argument again comes down the matter of whether post-filed data is relevant to the obviousness analysis.

Tarranto also argues that the issue decided in Dillon was only what was needed to establish a finding of prima facie obviousness, not what could properly rebut a finding of obviousness. Finally, Tarranto argues that the panel should not have ruled on the doctrinal relationship between a finding of unexpected results and a finding of the prima facie case elements, because the panel concluded (even if erroneously) that there were no appreciable unexpected results.

Judge Dyk’s concurrence with the denial of en banc review agreed with the panel that the post-filed data was properly excluded.

Judge O’Malley concurred with the denial of en banc review, writing to assuage fears that this decision rewrites the test for obviousness in pharmaceutical patents, and asserting that the standard is unchanged. O’Malley argued that the post-filed data was properly considered but doesn’t save BMS. However, O’Malley also suggested that BMS raised evidence of entecavir’s unexpected properties and reasonable expectation of success (presumably its HBV activity) as “an afterthought” in its opening brief.

Comment: while this case seemingly tips the scales strongly in favor of generics, caution should be exercised in reading too much into this holding. This case appears to be the first time the Federal Circuit found a novel pharmaceutical compound invalid as obvious on a full review,[1] based on structural similarities in the prior art with a “lead compound” analysis. A number of previous compounds were all found to be not obvious, including rosuvustatin, pregabalin, olmestartan, aripiprazole, rabeprazole, and pioglitazone. So the question is, will the Federal Circuit start finding other compound patents invalid routinely?

The panel did not address the commonly applied principle that the chemical and pharmaceutical arts are inherently unpredictable (Eisai Co. Ltd., et al. v. Dr. Reddy’s Laboratories, Ltd., et al. 533 F.3d 1353, 1359 (Fed. Cir. 2008)), an omission that cuts in favor of the innovators.

Also, the split panel in the en banc review (6-4) (Prost did not participate in the decision to deny en banc review) does not give the victor here a clean win.

But for generics, this case suggests that in future arguments along these lines, generic challengers should stick to arguments that post-filed data is of limited probative value. Generic challengers should also emphasize that modifications to a lead compound are minor, and result in predictable changes in activity. Generic challengers should seek to exclude later filed data using this case as precedent, arguing that obviousness should be determined at time of invention or filing.

This case may call into question the paradigm of the lead compound analysis. It may be difficult to argue going forward that any chemical modification to a lead compound is anything but a “small conservative change.” However, this concern should be coupled with whether chemical and pharmaceutical arts are predictable. If chemical and pharmaceutical arts really are unpredictable, a small conservative change will not have a predictable effect. So a new procedure may be needed to evaluate the obviousness of chemical structures.

Case citation: Bristol-Myers Squibb Co. v. Teva Pharms. USA, Inc., 752 F.3d 967 (Fed. Cir. 2014); Judges Prost, Plager, and Chen; Opinion by Chen. En banc review denied, 769 F.3d 1339 (Fed. Cir. 10/20/2014); concurring in the denial were opinions by Dyk and O’Malley; dissenting from the denial of review were opinions by Newman and Taranto.

[1] But in Altana Pharma AG v. Teva Pharms. USA, Inc., 566 F.3d 999 (Fed. Cir. 2009), pantoprazole was held to be obvious on a motion for a preliminary injunction, i.e., a likelihood of success standard.