In February, the U.S. Court of Appeals for the First Circuit held, in a direct purchaser antitrust action, that an innovative pharmaceutical company marketing an injectable drug product had “improperly listed” in FDA’s Orange Book a patent claiming a mechanism used in the drug’s delivery device. As I explain below, the ruling creates the specter of antitrust liability for steps taken in good faith to comply with a complex regulatory framework that overlaps in part with patent law. I explain below how the ruling puts biopharmaceutical innovators in a tough spot.
First, the legal framework.
Federal law requires each company that submits a new drug application to identify the patents that claim the drug or a method of using the drug (if a claim of patent infringement could reasonably be asserted against someone who made, used, or sold the drug without a license). The application cannot be approved, if the company fails to submit the required information on a patent that satisfies the listing standard. (See section 505(d)(6) of the drug statute, here.) FDA publishes the patent numbers and expiration dates in the “Orange Book,” which takes the form of a PDF and electronic database.
Federal law also requires a generic drug applicant to take a position with respect to every patent that claims the drug or a method of using the drug — effectively, every patent listed in the Orange Book. For every unexpired patent, the generic applicant has two choices, which dictate when its application can be approved. (There’s a third option for a patent claiming a method of using the drug, which isn’t relevant here.)
It can choose to wait for patent expiry, which means filing a “paragraph 3 certification.” In this scenario, FDA cannot approve its generic drug for market entry until expiry of the patent.
Or it can say that it plans to market right away, because its product doesn’t infringe the patent or because it thinks the patent invalid, which means filing a “paragraph 4 certification.” In this scenario, it must notify the innovator (and patent owner, if different). (I’ll just say “innovator,” going forward.) As far as this patent is concerned, FDA can approve the generic drug for market entry as soon as its review is complete and assuming the generic drug is otherwise approvable with one important exception. If the patent was listed before the generic drug company submitted its application, and if the innovator files a patent infringement suit within 45 days of receiving notice, then final approval of the generic application is stayed for 30 months or until a district court ruling in the generic company’s favor (whichever happens first). The paragraph 4 certification is considered an act of infringement, which creates federal court jurisdiction.
The patent listing mechanism is intended to facilitate litigation of patent issues before market entry, which both industries wanted. The generic companies wanted a way to litigate these issues before launching, for example, because doing so avoids the risk of damages (for more information, see my article on the history and political economy of the legislation). The scheme encourages generic companies to participate by offering 180-day exclusivity in the market for the first to file a (true) generic application with paragraph 4 certification, and it encourages innovators to participate by offering the 30-month stay that makes it possible for the patent to be litigated before the generic drug launches.
These rules apply to companies that file true generic applications, for exact copies of the innovator’s drug. And with one exception they also apply to companies that file a different type kind of abbreviated application known as a 505(b)(2) application. The distinction between the types of application isn’t critical here. The one exception is that companies filing 505(b)(2) applications with paragraph 4 certifications aren’t eligible for 180-day exclusivity.
Second, applying the framework to combination products in particular.
The listing standard — “any patent which claims the drug for which the applicant submitted the application or which claims a method of using such” — has proved vexing to interpret.
In 1994, FDA published its first regulation interpreting this provision, stating that it meant “drug substance (ingredient) patents, drug product (formulation and composition) patents, and method of use patents,” but not “process patents.” But there have been questions about a variety of patent types over the years, and in 2003 — responding in part to requests for elaboration — the agency revised its regulations to provide more details about what it required to be listed and what was not to be listed.
At issue here: what to do with combination products. These products combine two regulated components, such as a device and a drug. Two discrete products packaged together for use together are, together, considered a “combination product.” But the phrase also means a single finished product that comprises two regulated components — thus a drug and device produced as a single entity. Combination products thus include prefilled drug delivery devices — such as a prefilled drug syringe, an auto-injector, or an metered dose inhaler (see here).
The question is whether the statute requires companies to list patents associated with the device component of these products.
FDA considered this in the 2003 rulemaking. The final regulation is 21 C.F.R. § 314.53, but the agency’s explanation of the regulation in the Federal Register — which has the formal status of an “advisory opinion” — is just as important.
The agency decided that “patents claiming a package or container must not be submitted.” Packaging and containers are “distinct from the drug product.”
Several commenters also argued that patents claiming devices that are “integral” to the drug product or require approval should be listed. FDA offered what it labeled as a “response.” The agency didn’t write that these patents “should” be listed, or that they “should not” be listed. Instead it said that a “drug product” is the drug in its “finished dosage form” — meaning the form administered to patients. And, it added, the current list of “dosage forms for approved products” — which appears in an appendix to the Orange Book — includes “aerosols, capsules, metered sprays, gels, and pre-filled drug delivery systems.” Elsewhere it wrote that a patent claiming the finished dosage form “must be submitted for listing.”
Now, the litigation and First Circuit ruling.
Sanofi-Aventis holds the approved marketing application for Lantus (insulin glargine recombinant), a long-acting human insulin analog used in treating diabetes. At first the company sold Lantus in multiple dose vials and in cartridges for use with a (separate) insulin delivery device. In 2007, however, FDA approved a supplemental application for sale of Lantus in a single-patient-use prefilled injector pen.
Sanofi has listed several patents in connection with Lantus. In connection with the prefilled pen, the company listed U.S. Patent No. 8,556,864 (drive mechanisms suitable for use in drug delivery devices), which issued in October 2013 and expires in March 2024. The parties agree that the ’864 patent claims the drive mechanism used in the Lantus pens, and FDA would not have approved the prefilled pens without a showing that the pen (including the drive mechanism) ensures patients safely receive accurate doses. But — and this turned out to be critical in the end — the patent doesn’t mention insulin glargine. Nevertheless, according to the agency, an insulin injector pen is a prefilled drug delivery system. And this makes it a dosage form. And patents claiming dosage forms must be listed.
In 2013, Eli Lilly submitted a 505(b)(2) application for a copy of Lantus, which it planned to market as Basaglar. It included a paragraph 4 certification to the ’864 patent and to various other patents as well. Sanofi brought suit. The case settled on the morning trial was scheduled to begin, with Lilly agreeing to pay for a license to launch in December 2016, seven years before patent expiry.
The plaintiffs in this antitrust litigation are drug wholesalers. They claim, among other things, that Sanofi improperly listed the ’864 patent. (As far as I can tell, Lilly didn’t raise the issue.) The district court dismissed their first amended complaint, pointing out that FDA has interpreted “drug products” to include “prefilled drug delivery systems” and that patents claiming drug products must be listed. The plaintiffs amended their complaint, but the district court dismissed again on largely the same grounds. Under a “reasonable interpretation” of the agency’s regulations, Sanofi had to submit the patent for listing. So, it couldn’t have been improper conduct to list the patent.
The First Circuit’s ruling came as a shock. In a unanimous decision, Judge Kayatta wrote that Sanofi had improperly listed the patent. He reasoned as follows. First, the statute and regulations call for listing of patents that claim the drug, and the patent doesn’t even mention the drug. Second, in 2003 FDA didn’t adopt the proposal that devices “integral” to the product should be listed. Instead, the agency said that companies should list patents that claim the finished dosage form. And this patent doesn’t, the court wrote; it claims a device that can be combined with other components to produce the finished dosage form.
Finally, the implications.
The innovative pharmaceutical industry has asked FDA repeatedly since 2003 — at least four times, including in citizen petitions — to clarify whether patents directed to drug delivery systems are supposed to be listed, if they don’t recite the drug’s active ingredient or formulation. The agency never answered these requests.
Although FDA’s failure to respond has been frustrating, it is my understanding that most companies — consulting with patent and regulatory counsel — have concluded these patents should be listed and that, in fact, they list them, and FDA publishes them. I have always thought this was the best reading of what FDA wrote in 2003. At the very least, it is a reasonable reading of what FDA wrote. It is deeply concerning that the First Circuit now purports to answer this question for the agency — no, these patents do not satisfy the listing standard — in litigation to which FDA was not a party and could not explain its interpretation of the statute or its expectations.
The decision is also fundamentally hostile to pharmaceutical innovators. The Hatch-Waxman scheme — statute, regulations, guidance, and precedent — is complex, and figuring out how it applies in any particular situation can be tricky. There are other unresolved listing issues, which companies and their counsel work through in good faith. The lesson here seems to be that an innovator trying to navigate uncertainty about the listing requirements does so at its peril.
On the one hand, failing to list a patent that satisfies the criteria has serious consequences. Listing is not voluntary; the statute requires it. The company must declare (“under penalty of perjury”) that its patent submission is “accurate and complete.” The patent submission form reminds the company that “a willfully and knowingly false statement is a criminal offense” under 18 U.S.C. § 1001. (And at least in theory, FDA would reject an application that lacked the required patent information, though I don’t know if this has ever happened or if it would happen.) Most importantly, failing to list a patent means there is no paragraph 4 certification, and thus no artificial act of infringement, and no opportunity to enforce the patent before generic market launch. Failing to list also passes up the benefit of the 30-month stay. And there may be concern that failure to list a patent is some sort of admission against the innovator’s interests in litigation.
On the other hand, listing a patent that doesn’t satisfy the criteria attracts antitrust scrutiny, presumably because the listing places an administrative burden on generic applicants and might trigger a stay of approval. And this case shows that a hostile court may disagree with the company’s reading of the statute, regulations, Federal Register, and precedent. Even if a company adopts what appears to many to be a reasonable interpretation of the patent listing requirements, a court might interpret the listing requirements on its own — without the benefits of FDA’s views — and force the company into expensive and time-consuming antitrust litigation. Indeed, two bloggers recently praised the decision, recommending that generic companies “examine the Orange Book listings,” as they may contain a “rich vein” for antitrust claims.
To be sure, as the court wrote, Sanofi can try to show, on remand, that its submission was “the result of a reasonable, good-faith attempt to comply with the Hatch-Waxman scheme.” This would provide a defense to any liability under the Sherman Act for “antitrust injury caused by” the submission. But the burden has shifted to the company, and much of the language in the court’s opinion suggests that this will be an uphill battle. (E.g., “The statute and regulations clearly require that only patents that claim the drug for which the NDA is submitted should be listed in the Orange Book. The ’864 patent … does not fit the bill.”)
A postscript from the administrative law side of the table.
Consider a counterfactual.
As a preliminary matter, recall that FDA’s regulations require companies to list patents on drug products. These regulations also state that the phrase “drug product” refers to a drug in its “finished dosage form.” FDA has said, for years, that a patent claiming a finished dosage form “must be submitted for listing.” Finally, it has listed prefilled drug delivery systems as a type of “dosage form” in the Orange Book.
Now suppose that FDA had responded to the industry requests for clarification and stated definitively — given what I just wrote — that a patent claiming any component of a prefilled syringe must be listed? In my view, this would be a defensible position for the agency to have taken, given the statute, the regulations, and what it has written to date.
What would have happened if this hypothetical FDA decision had gone to the First Circuit for review, in a totally different kind of lawsuit? Would the First Circuit really conclude that the agency’s interpretation of the statute was unreasonable or impermissible (Chevron)? Would it really conclude that the agency’s interpretation of its regulation was “plainly erroneous or inconsistent with the regulation” (Auer)? And if we think that the courts would (or should) defer to FDA in this hypothetical case, how can Sanofi’s decision possibly have been unreasonable?
In the end, the First Circuit’s ruling contains a troubling lesson for pharmaceutical innovators. When navigating uncertainty about a patent’s status under the patent listing requirements, even if it seems reasonable to conclude that FDA would require listing and even if the agency won’t answer the question, listing the patent in good faith creates a serious risk of facing antitrust litigation. The alternative, equally unappealing, is to relinquish the opportunity to enforce the patent before generic market entry, which conflicts with the purpose and design of the Hatch-Waxman Amendments and undermines the value of the patent.