"The Novartis case should have been treated in the same way as other cases where a patentee secured a broad claim construction to prove infringement, only to have the patent invalidated because the claims as construed were not enabled as to their full scope."
In Novartis Pharms. Corp. v. Torrent Pharma Inc., ___F. 4 ___, 2025 U. S. App. LEXIS 486 (Fed. Cir., Jan. 10, 2025), the U. S. Court of Appeals for the Federal Circuit shut down efforts to market a generic version of the heart disease drug ENTRESTO which is sold under a patent which claims a "combination" of two drugs, valsartan and sacubitril. Because the proposed generic drug accused of infringement was a "complex" of bonded valsartan and sacubitril, the patentee Novartis needed to urge and had successfully urged that "combination" be construed to include both bonded complexes and non-bonded (i.e. separate) components. Absent such construction, the accused product would not have infringed.
But in addressing an enablement defense, in which the subject patent had to be enabled as to the full scope of the patent claims as construed, the Federal Circuit effectively held that the patent did not claim bonded valsartan-sacubitril complexes, such that enablement as to the bonded combination -- or the lack thereof -- was a moot point.
This was problematic at best. It contravened Supreme Court precedent by allowing the patentee to extend its monopoly to drug products for which the patentee did not provide the required quid pro quo of an enabling disclosure. And the decision seems to have been driven by a factual distraction and case law which, in all likelihood, is no longer viable.
The patent asserted by Novartis in the case was U.S. Patent No. 8,101,659. For present purposes, the claims of the '659 Patent called for a pharmaceutical combination comprising valsartan and sacubitril "administered in combination" in about a 1:1 ratio. The accused generic product, like Novartis's ENTRESTO, comprised a complex of non-covalently bonded valsartan and sacubitril.
During claim construction, the defendant MSN urged that the claims be limited to the two active agents as two separate components. Inasmuch as such a limiting construction would have led to a finding of non-infringement, Novartis argued that "combination" be given its plain and ordinary meaning, i.e. including bonded complexes. The District Court agreed and thereafter conducted a bench trial
In addressing enablement after the bench trial, the district correctly recognized that enablement is judged as of the patent's priority date (here, 2002), such that a later-existing state of the art may not properly be considered in an enablement analysis. The Federal Circuit's summary of the District Court's ruling on enablement was -- in its entirety -- as follows (with citations omitted):
"[B]ecause complexes of valsartan and sacubitril were unknown in the art in 20022, the [district] court determined that they need not have been enabled in the '659 patent....The [district] court further found that MSN had failed to establish that pharmaceutical complexes, more generally, were known or were nascent as of the 2002 priority date."
In a unanimous precedential opinion, the Federal Circuit affirmed the District Court's decision rejecting the defense of lack of enablement. The case also involved defenses of lack of written description and obviousness, which are beyond the scope of this article.
The Court of Appeals began its analysis by stating that a specification must "only" enable the claimed invention. It then noted that the claimed invention "as construed by the district court" is a composition in which valsartan and sacubitril are administered "in combination." It then stated that "the patent does not claim as its invention valsartan-sacubitril complexes.
It is equally well settled that an enabling disclosure must support the full scope of the claim as construed. Just two years ago, the Supreme Court reaffirmed that:
"If a patent claims an entire class of...compositions of matter, the patent's specification must enable a person skilled in the art to make and use the entire class. In other words, the specification must enable the full scope of the invention as defined by its claims. The more one claims, the more one must enable."
Amgen Inc. v. Sanofi, 588 U.S. 594, 609 (2023).
The Federal Circuit concluded its enablement discussion in Novartis by stating that was affirming on the issue of enablement "because [i] the '659 patent does not expressly claim complexes, and because [ii] the parties do not otherwise dispute that the '659 patent enables that which it does claim...." (Emphasis added). As to part [i], whether or not the '659 patent "expressly" claims complexes is neither here nor there. It does in fact claim complexes because Novartis sought and obtained a broad construction which does include complexes.
As for part [ii], it is difficult to discern from the appellate opinion exactly what the parties argued. If the court was suggesting here that the '659 patent does enable some compositions within the scope of "combinations" as construed -- viz., separate components -- that too is neither here nor there. It simply doesn't begin to address the question whether the full scope of the claims as construed are enabled.
In full effect, the Novartis case should have been treated in the same way as other cases where a patentee secured a broad claim construction to prove infringement, only to have the patent invalidated because the claims as construed were not enabled as to their full scope. For example, in Sitrick v. Dreamworks LLC, 516 F. 3d 993 (Fed. Cir. 2008), the patent at issue involved technology which was potentially applicable to both movies and video games. Because the defendant was only accused of infringement as to movies, the patentee sought and obtained a claim construction which encompassed both movies and video games. Even assuming that the claims were enabled as to video games, the claims were invalid if they were not enabled for movies -- and they were not.
A critical element of the District Court's and the Federal Circuit's analysis was this fact: years after the 2002 critical date for the '659 patent, Novartis obtained other patents which expressly disclosed non-covalently bound valsartan and sacubitril salts. What is difficult to unpack is how and why this fact figured in the Federal Circuit's enablement analysis
By way of background, the statute, 35 U.S.C. §112(a), states that it is the specification of a patent that must provide a disclosure that is enabling to one of ordinary skill in the art. But a patent applicant need not include in the specification that which was generally and reasonably available to those skilled in the art as of the critical date. In re Howarth, 654 F.2d 103 (C.C.P.A. 1981). Thus, Novartis's post-2002 patents could not have been relied upon by Novartis to supplement the '659 specification because their content was not generally known before 2002.
Novartis appears to have urged that the post-2002 patents somehow established that the '659 patent was excused from having to enable valsartan - sacubitril complexes. Part of that strategy seems to have been that because the later patents expressly disclosed and claimed complexes, the '659 patent necessarily could not and did not claim complexes. Immediately after stating that the '659 patent "does not claim as its invention valsartan -sacubitril complexes," the Federal Circuit stated, as if by way of proof, that "Indeed, Novartis obtained separate, later patents to such complexes." In fact, what Novartis obtained after 2002 were claims directed to a species of the genus broadly -- indeed, over-broadly -- claimed in the '659 patent.
Novartis relied on three cases from the Federal Circuit and its predecessor supposedly supporting the proposition that the mere fact that an invention can be improved upon does not establish the absence of an enabling disclosure and even excuses an absence of enablement. But of these cases were decided decades before Amgen conclusively held that all patent claims must be enabled as to their full scope.
The viewpoint of these outdated cases is well captured in Chiron Corp. v. Genentech, Inc., 363 F. 3d 1247 (Fed. Cir. 2004), where the court stated that "a patent document cannot enable technology that arises after the date of application. The law does not expect an applicant to disclose knowledge invented or developed after the filing date. Such disclosure would be impossible." Id. at 1254. To which the post-Amgen response might be something like: the law does not expect, or allow, inventors to seek claims covering that which they have not invented, for which they are incapable of providing sufficient enablement.
The analysis by the District Court and the Federal Circuit in Novartis would seem to support this anomaly. When the '659 patent issued in 2002, with a specification devoid of bonded complexes, the claims may well have been invalid for want of enablement. But as soon as the later patents came into the picture, the claims were retroactively deemed enabled, and the failure of the '659 specification to disclose bonded complexes became an excused absence.
Accordingly, in this observer's view, the post-2002 patents were a red herring. They neither prove nor disprove enablement of the '659 patent.
So where does this leave us? It must be kept in mind that MSN at all times had the burden of proving by clear and convincing evidence that that there was no enabling disclosure of valsartan -sacubitril complexes as of 2002. Thus, MSN was necessarily required to prove a negative.
But the record seems to have the equivalent of such proof. The District Court flatly stated, "It is not disputed that the specification neither discloses nor suggests a complex of valsartan and sacubitril." It also found as a fact that "complexes of valsartan and sacubitril were unknown in the art in 2002." Moreover, an effort to establish that pharmaceutical complexes, more generally, were known or were nascent technology as of 2002, had failed. (The effort had been made by MSN in an effort to address Novartis's pre-Amgen decisions.)
It thus seems hard to escape the view the '659 patent claims as broadly construed per Novartis's request were required to enable valsartan-sacubitril complexes, and they did not.