Effects of GSK v Teva ripple after SCOTUS denies cert
Last week (May 15), the Supreme Court denied certiorari in GlaxoSmithKline v Teva Pharms USA (GSK v Teva), locking in the US Court of Appeals for the Federal Circuit’s second panel decision which held that Teva’s attempted section viii carve-out of an indication covered by a patented method of use was not ‘skinny’ enough to avoid being liable for infringement.
Although the court did not state why certiorari was denied—as is common—the order list stated that Justice Brett Kavanaugh would grant the petitions for certiorari. One possible reason that the court believes that certiorari is not ripe is that Teva’s affirmative defence of equitable estoppel, ie, if GSK’s representations to the US Food and Drug Administration (FDA) were at odds with its enforcement efforts in this case, still needs to be considered on remand. The opinions from the Federal Circuit concurring and dissenting from the denial for the petition for rehearing en banc have left copious notes for the district court to consider on remand regarding equitable estoppel.
While the district court will now have to consider the issues of equitable estoppel on remand, there are still several other effects from GSK v Teva that we see in other district court cases, at the FDA, and for biosimilars.
- GSK v Teva’s second panel decision, denial of rehearing en banc, and arguments presented to the Supreme Court
GSK sells Coreg, a drug with three approved indications, one of which (the ‘CHF indication’) is protected by a patent. Teva attempted to carve out the CHF indication with a ‘skinny’ or partial label using the proper section viii statutory pathway. Ultimately, the jury found—and the Federal Circuit agreed—that although Teva attempted to carve out the patent-protected CHF indication, Teva’s label was not a “true section viii carve-out”.
The Federal Circuit agreed that the jury’s verdict was supported by substantial evidence, namely that despite the attempted ‘skinny label’, Teva induced infringement of the method of use patent based on expert testimony, Teva’s partial label, and various press releases, including statements that Teva’s generic drug was ‘AB-rated’ to the brand drug. Judge Kimberly Prost dissented from the majority opinion, faulting them for not finding active-encouragement on the part of the alleged infringer.
The Federal Circuit then denied rehearing en banc, locking in the second panel decision. In a concurring opinion, Chief Judge Kimberly Moore explained that Teva could still rely on an equitable estoppel defence on remand, ie, that Teva cannot be held liable for infringement because Teva allegedly followed what GSK and the FDA told it to do to successfully carve out the patent-protected indication.
Despite Chief Judge Moore’s statement that the case would not have far-reaching implications, and instead was “narrow and fact dependent”, dissenting opinions were filed by Judges Prost, Timothy Dyk, and Jimmie Reyna. Each noted the very serious conflict between federal patent infringement law, and the requirements of FDA-labelling under the Hatch-Waxman Act.
At the Supreme Court, both GSK and Teva raised the same arguments they did below, along with additional public policy statements regarding the impact of the second panel decision. Interestingly, the brief for the US as amicus curiae in support of Teva’s cert. petition was signed by the US Patent and Trademark Office (USPTO), the Department of Health and Human Services, the FDA, and the solicitor general’s office, all of which supported that the second panel decision was incorrectly decided and that “no reasonable jury could have concluded that the carved-out labelling . . . was itself evidence of intent to induce infringement.”
Now that the Supreme Court has denied cert., the Federal Circuit’s second panel decision is locked in.
2. GSK’s likely arguments regarding equitable estoppel on remand
While the district court will now have to consider the issue of equitable estoppel on remand (ie, whether Teva cannot be held liable for infringement because Teva allegedly followed what GSK and the FDA told it to do to successfully carve out the patent-protected indication), GSK’s supplemental brief at the Supreme Court likely revealed GSK’s key arguments. Specifically, GSK argued that the FDA did not impose any obligation to identify sections of the labelling describing claimed methods of use until 2016, almost a decade after the events in question. And even if the FDA had required GSK to identify these sections, the FDA and Teva could not have relied on GSK’s statements because GSK submitted its Form 3542 six months after Teva launched its ‘skinny label’ generic carvedilol. GSK’s arguments on remand will likely be similar, namely that there was no “misleading conduct” and therefore no equitable estoppel.
3. Amarin v Hikma noted the fact-intensive inquiries of GSK v Teva, but that too could be reversed on appeal
Since GSK v Teva, other parties have attempted to argue that section viii carve-outs were not ‘skinny’ enough to avoid infringement. One notable case is Amarin Pharma v Hikma Pharm (2022), where Judge Andrews granted Hikma’s motion to dismiss, agreeing that Amarin had failed to state a claim for induced infringement because Hikma’s ‘skinny label’ did not instruct or encourage an infringing use.
In that case, Amarin accused Hikma of inducing infringement of three patents reciting methods of using icosapent ethyl (sold under the brand name Vascepa) for the reduction of cardiovascular risk. Vascepa has two indications: (1) severe hypertriglyceridemia (the “SH indication”) and (2) cardiovascular risk reduction (the ‘CV indication’). Hikma received FDA approval to sell its generic icosapent ethyl under a skinny label covering just the SH indication, and carving out the CV indication.
Judge Andrews’ decision quoted the statement from GSK v Teva that GSK is a “narrow, case-specific review”, and then distinguished the facts in GSK from those before him. He found that the label did not contain any instruction to use the drug for CV risk reduction. He next found that Hikma’s press releases and advertisements failed to promote an infringing act, even if such press releases may have demonstrated Hikma’s intent, because they did not point to a patented use. Thus, Judge Andrews held that Amarin had not pleaded a claim for induced infringement.
However, Judge Andrews separately held that an insurer, Health Net, could be potentially liable for induced infringement based on allegations that it provided Hikma’s generic capsules a favourable formulary position and in view of a prior authorisation form that continued to list the patented indication for the generic capsules. Health Net has since settled the case.
Amarin has filed an appeal to the Federal Circuit regarding its claims against Hikma. Thus, the Federal Circuit will again have the opportunity to wade into the issue of ‘skinny labels’.
4. FDA has recognised “concern” with GSK v Teva and requested legislative amendments to allay this concern
In the FDA’s fiscal year 2024 legislative proposal, the FDA stated that it was “concerned” that GSK v Teva “imperils an important statutory marketing pathway that allows earlier generic drug market entry for conditions of use of a drug not protected by a patent” and that this decision “could significantly impact the timely availability of generic drugs”.
Thus, the FDA proposed that the Hatch-Waxman Amendments be amended to “create a safe harbour from patent infringement for . . . generic drug applicants and 505(b)(2) applicants who market a drug with ‘skinny labeling’” by excluding labelling from the evidence that can be used to support a claim of patent infringement, and clarifying that statements “regarding therapeutic equivalence” cannot be used as evidence to support an infringement claim.
If Congress chooses to accept the FDA’s proposal, then legislative amendments may prevent skinny labels from being as perilous as they now appear to be.
5. Biosimilars intending to carve out licensed conditions of use must be aware of section viii carve-outs, despite being under a different statute
Biosimilars can be approved “for fewer than all of the reference product’s licensed conditions of use.” (Biosimilars and Interchangeable Biosimilars: Licensure for Fewer Than All Conditions of Use for Which the Reference Product Has Been Licensed—Draft Guidance [February 2020]). Thus, biosimilars can have ‘skinny labels’, even though Hatch-Waxman’s section viii statement approach does not apply to biosimilars. More specifically, the FDA cannot approve of a biosimilar if it has an indication protected by the applicable orphan drug exclusivity. In addition, it is incumbent on the biosimilar to determine if it wants to have its label list a licensed condition of use that is protected by one or more patents.
Taking Humira as an example, these biosimilar carve-outs are not just hypothetical, with one report stating that the approved biosimilars of Humira have carve-outs for both regulatory exclusivities and patent-protected conditions of use, despite all approved biosimilars being under settlement agreements with Abbvie (Alexander Egilman et al, Frequency of Approval and Marketing of Biosimilars With a Skinny Label and Associated Medicare Savings JAMA Int. Med., Vol. 183 No. 1 [Jan 2023])
Thus, GSK v Teva may be applied by analogy for biosimilars seeking labels with fewer than all licensed conditions of use, and innovator-side biologic companies may then sue the biosimilar for inadequately carving out patented licensed conditions of use. It’s also possible that we may see some litigation against the FDA for approving biosimilars that inadequately carve out licensed conditions of use protected by orphan drug exclusivity.
This is made more difficult for interchangeable biosimilars, which “must include information sufficient to show that the proposed interchangeable product ‘can be expected to produce the same clinical result as the reference product in any given patient’” and where the FDA expects that the interchangeable biosimilar “will submit data and information to support a showing that the proposed interchangeable product can be expected to produce the same clinical result as the reference product in all of the reference product’s licensed conditions of use”.
Thus, an interchangeable biosimilar, simply by being designated as “interchangeable”, may not be able to adequately carve out protected licensed conditions of use, which will likely be cited—as ‘AB-rated’ generic carvedilol was cited in GSK v Teva—as evidence showing that there was inadequate carving out.
Thus, despite the Supreme Court’s denial of cert, the issues from GSK v Teva continue to exist and have far broader effects than the Federal Circuit may have considered.
Chad Landmon is a partner at Axinn, Veltrop & Harkrider. He can be contacted at clandmon@axinn.com
Ross Blau is an associate at Axinn, Veltrop & Harkrider. He can be contacted at rblau@axinn.com.
Gulrukh Haroon is an associate at Axinn, Veltrop & Harkrider.
