The ongoing covid-19 pandemic has demonstrated the speed at which the biopharma sector can act when facing a medical emergency. On 4 February 2020, the US Department of Health and Human Services declared a public health emergency relating to covid-19. In less than six months, the US Food and Drug Administration (FDA) granted emergency use authorisation for two repurposed drugs:
- remdesivir, which had been previously developed as an antiviral but not previously approved for any indications (May 2020); and
- corticosteroids, such as dexamethasone (July 2020).
Other well-known examples of repurposed drugs (ie, drugs approved to treat a first indication that can also be used to treat a second indication) include aspirin to deter heart attacks and thalidomide for the treatment of leprosy and multiple myeloma. Repurposing drugs offers drug developers a more rapid, cost-effective way to develop treatments in a medical emergency or to identify new treatments for diseases (see Table 1).
Table 1. Drug repurposing
|Drug repurposing||New drug development|
|Success rate||Three in 10||One in 10,000|
|Safety||Known safety||Unknown safety|
|Typical timeline||One to three years||12 to 19 years|
|Cost||Millions of dollars||Often $1 billion to $2 billion|
Certainly, where clinical trials of the new use are required, the cost will be higher and the timeline will be longer, but both are still much less than for an original approval. Nonetheless, to make the necessary investments, the industry needs some assurance that it will be able to recoup the costs along with a reasonable profit on marketing the repurposed drug.
Protecting repurposed drugs
To protect pharmaceuticals and biologics, composition-of-matter patents covering the active ingredient are the gold standard. For repurposed drugs, however, the original composition-of-matter patents may have long since expired. For example, while Bayer’s original aspirin patent expired in 1917, it was not until 1985 that aspirin’s use for the prevention of heart attacks was approved. Similarly, the original dexamethasone patent expired in the early 1980s, nearly 40 years before its use in treating covid-19 was discovered.
Sometimes a repurposed drug uses a new formulation or dosing regimen. In those cases, a patent covering those attributes may protect the repurposed drug from generic competitors. For example, in the early 1990s, Bayer introduced its patented low-dose aspirin formulation, which reduced the risk of a heart attack while minimising the risk of gastrointestinal bleeding.
However, such protection may be limited. For example, while a formulation or dosing patent prevents generic competitors from manufacturing and selling the new formulation (eg, a low-dose aspirin tablet), it will not prevent marketing the original formulation, which compounders or patients may then modify for the new use (eg, cutting in half an original higher dose aspirin tablet).
Method-of-treatment patents in the United States (Swiss-type claims and European Patent Convention 2000 composition-for-use claims in non-US jurisdictions) may also protect the use of repurposed drugs. However, they are use-limited and provide no protection for the drug itself or its earlier known uses.
In addition, it is usually only the clinician or patient who practises the patented method of treatment. By contrast, it is the generic drug manufacturer or distributor who directly infringes patent claims to compositions of matter, formulations and dosing regimens (to the extent that the dosing regimens affect drug formulation). To assert a method-of-treatment patent against a generic competitor, therefore, a patent owner must rely on one of two theories of indirect patent infringement:
- contributory infringement; and
- induced infringement.
Contributory infringement occurs when a component is sold for use in a patented product or process (35 USC Section 271(c)). Accordingly, a patent owner must first prove that someone (eg, a clinician or patient) directly infringed the patent by using the method of treatment. Next, the patent owner must prove that the generic manufacturer knew both of the patent and that its customers’ use of the component was infringement (Global-Tech Appliances Inc v SEB SA, 563 US 754 (2011)). Finally, the patent owner must prove that there is no substantial non-infringing use for the component (35 USC Section 271(c)). For repurposed drugs, this last prong may be dispositive, especially if the patent covering the method of treatment for which the drug was initially approved has expired (ie, by definition, there is a substantial non-infringing use).
Induced infringement occurs when the generic manufacturer actively induces infringement of the patent (35 USC Section 271(b)). Here, a patent owner must prove that there was direct infringement and that the generic manufacturer had knowledge of the patent (or was wilfully blind as to its existence) and had an affirmative intent that its product be used to infringe (MGM Studios Inc v Grokster Ltd, 545 US 913, 936 (2005)). This intent can be established by demonstrating that the generic manufacturer advertised an infringing use or provided instructions to customers to engage in that use. This may be a difficult burden to meet if the label only provides instructions to use the generic drug in a non-infringing manner.
Skinny labels and generic competition
Further compounding the difficulties in protecting repurposed drugs is the Drug Price Competition and Patent Term Restoration Act 1984 (known as ‘the Hatch-Waxman Act’). The act allows generic drug manufacturers to obtain FDA approval for non-patented indications by carving out from its label any approved use that is still patented. This practice is known as ‘skinny labelling’.
Generic manufacturers and patient advocacy groups argue that skinny labelling allows generic competition before it would otherwise have been possible, thereby driving down drug prices. They also argue that skinny labelling limits the efforts of branded manufacturers to maintain market exclusivity through evergreening and patent thickets.
By contrast, branded manufacturers argue that skinny labels have a chilling effect on innovation by making it more difficult to recoup the enormous investments for innovating a new drug, expanding its indications, conducting clinical trials and navigating the approval process. They also argue that clinicians often violate the skinny label by writing prescriptions for off-label uses of the generic drug, which infringe the existing use patents that were carved around.
Skinny labels also de-incentivise the development of repositioned drugs by allowing the repurposed use to be carved out of the label. The generic manufacturer then relies on the market to use the drug off-label in the repurposed method of treatment. Because the skinny label recites only non-patented methods, a generic manufacturer can avoid liability for contributory infringement because the generic product has a substantial non-infringing use in the non-patented method. Similarly, proving induced infringement based on a skinny label, which only provides instructions to use the generic drug in a non-infringing manner, may be difficult. A recent case, however, highlights that in some circumstances a skinny label may not avoid patent infringement.
In GlaxoSmithKline LLC v Teva Pharmaceuticals USA Inc (976 F.3d 1347 (Fed Cir 2020)), a divided panel of the Court of Appeals for the Federal Circuit held that Teva’s marketing efforts were sufficient, despite a skinny label, to constitute induced infringement. The FDA had approved GlaxoSmithKline’s (GSK) Coreg (carvedilol) for three indications:
- congestive heart failure (CHF); and
- left ventricular dysfunction following myocardial infarction (LVD-MI).
GSK’s method of treating the CHF patent had an expiry date well after the expiry date of its composition-of-matter patent for carvedilol.
Before either patent had expired, Teva filed an abbreviated new drug application (ANDA) for its generic carvedilol. Teva’s ANDA contained a Paragraph III certification that it would not launch the generic carvedilol until the expiry of the composition-of-matter patent (which also claimed the method of treating hypertension). Teva also certified under Paragraph IV that the CHF patent was “invalid, unenforceable, or not infringed”. On expiry of the composition-of-matter patent, Teva launched its generic carvedilol under a skinny label for use only for hypertension and LVD-MI.
GSK then sued Teva for inducing infringement of its CHF patent. At trial, Teva argued, inter alia, that because of its skinny label it did not induce infringement. GSK countered that Teva’s skinny label induced infringement because it included the LVD-MI indication that at least partially overlaps with the CHF indication and, therefore, did not completely carve out CHF. GSK also argued that Teva’s marketing materials, including press releases from 2004 and 2007, that were still on Teva’s website in 2015, indicated that its generic carvedilol was an AB Rated generic of Coreg Tablets, thereby informing clinicians that Teva’s generic carvedilol was “therapeutically equivalent” to GSK’s Coreg Tablets.
Notwithstanding the skinny label, the jury found that Teva had induced infringement of the CHF patent. The District Court, however, reversed, holding that the jury’s verdict of induced infringement was not supported by substantial evidence. The court reasoned that an essential element of induced infringement was proof that Teva’s actions caused direct infringement of the CHF patent. The court also noted that many sources of information, including publications by the American Heart Association and the American College of Cardiology, the Coreg label and GSK’s marketing materials, had already informed clinicians about the various uses of carvedilol, including for treating CHF. On those bases, the court concluded that “[a] reasonable factfinder could only have found that these alternative, non-Teva factors were what caused doctors to prescribe generic carvedilol for an infringing use” (GlaxoSmithKline, 976 F.3d at 1351).
GSK appealed to the Federal Circuit. Several amici filed briefs in support of the parties, most addressing the “skinny label issue” and asserting that an infringement finding despite the skinny label would have far-reaching, unintended consequences. In a brief in support of GSK, the Biotechnology Innovation Organisation (BIO) argued, inter alia, that the District Court’s causation standard was inconsistent with tort law because it required the inducement to result from Teva’s instructions “as opposed to other factors”. Indeed, BIO analogised Teva’s marketing strategy of making clinicians aware of the availability and suitability of its generic to a store-brand product being placed on the shelf next to a branded product, and argued that common experience demonstrates that no communication is necessary from the store to induce a consumer to purchase its product and to use it in the manner marketed by the branded company. BIO, therefore, asserted that a jury of reasonable people could infer that a generic manufacturer’s marketing choices were the cause of the direct infringement.
In October 2020, the Federal Circuit in a divided opinion reversed the District Court and reinstated the jury verdict. The majority held that substantial evidence, including Teva’s 2007 press release, marketing materials; and catalogues, indicating that Teva’s generic carvedilol was “an AB Rated generic of Coreg® Tablets” supported the jury’s verdict.
Relying on testimony that clinicians are “completely reliant” on information that generic manufacturers provide, the majority held that Teva’s highlighting of the AB rating would have caused a clinician to believe that:
- the drugs were “therapeutically interchangeable”;
- Teva’s generic “was a complete replacement” for Coreg, including for treating CHF; and
- clinicians prescribing Coreg had an expectation that pharmacists would dispense the generic unless the prescription explicitly stated ‘dispense as written’.
The majority, citing testimony from a Teva executive that, despite the skinny label, Teva expected pharmacists to dispense the generic for off-label use in CHF, also held that “when the provider of an identical product knows of and markets the same product for intended direct infringing activity, the criteria of induced infringement are met” (GlaxoSmithKline, 976 F.3d at 1355).
Chief Judge Sharon Prost, who was formerly Senator Hatch’s chief counsel on the Senate Judiciary Committee, dissented vigorously. She argued that there was no proof that Teva’s skinny label alone induced infringement of the CHF patent and that the majority’s decision “nullifies Congress’s statutory provision for skinny labels – creating liability for inducement where there should be none” (GlaxoSmithKline, 976 F.3d at 1358 (Prost, dissenting)). In the context of Teva’s press releases, the dissent noted that passively maintaining them on Teva’s website did not constitute inducement. Rather, inducement requires an affirmative act, and the only affirmative acts associated with the press releases occurred before the CHF patent was granted. Prost also argued that the third-party press releases, catalogues and other materials could not induce infringement because they were silent with respect to the authorised indications of Teva’s generic carvedilol. Moreover, Prost said that there was insufficient evidence that doctors even relied on Teva’s marketing materials when making treatment decisions (ie, there was no causation). The dissent also highlighted GSK’s substantial Coreg profits and the comparative sale price of the branded and generic versions of carvedilol. As to those economic arguments, the majority responded that any policy concerns about whether GSK had made enough money from its sale of Coreg to deny it the right to enforce its CHF patent was a matter for Congress, not the Federal Circuit. The majority also criticised the dissent for not applying the required substantial evidence standard of review.
Teva petitioned for a rehearing, which the panel granted. Oral argument was had on 23 February 2021 solely on the issue of whether there was substantial evidence to support the jury’s verdict of induced infringement, notwithstanding the skinny label. The panel’s questions focused on the specific facts and evidence, not the broader legal concepts. GSK argued that this was not a skinny label case and attempted to redirect the panel’s attention to its argument that Teva’s label was a “partial label”, not a “skinny” one, because the inclusion of LVD-MI meant that Teva failed to completely carve out the CHF indication. While it is difficult to predict how a Federal Circuit panel will rule, the oral argument suggests a more narrowly tailored decision but likely reaching the same result.
The covid-19 pandemic has demonstrated how repurposing drugs can provide a nimbler mechanism to develop treatments for emergent and other diseases. By their nature, repurposed drugs are most often protected with method-of-treatment patent claims. This requires the branded company to protect its market by suing generic manufactures only for indirect infringement. Skinny labelling helps generic competitors avoid liability for contributory infringement and minimises their risk for induced infringement. Assuming that the GSK panel maintains, in modified form, its original decision, it may provide a roadmap for the innovators of repurposed drugs to protect their new uses. It may also provide generic manufacturers a path for minimising the risk of induced infringement; for example, by ensuring that they completely carve out patent indications or by listing the approved indications in their marketing materials to avoid any jury inference that those marketing materials would induce a clinician or patient to practise a still-patented method of use. To the extent that the majority’s decision is focused on the substantial evidence standard of review, it may be narrowly applied, such that the concerns about ‘the sky is falling’ impact on skinny labels that Teva, the dissent and several amici have raised may not come to fruition.