The Supreme Court has ruled on the Merck/Integra case that I wrote about here last month, and reversed the most recent lower court ruling. And I'm glad to see it. This all turns on a 1984 change in the patent law, called 271(e)(1) for short, which says, broadly, that it's not an act of patent infringement to make or use a patented invention during its patent term "solely for uses reasonably related to the development and submission of information under a Federal law which regulates the manufacture, use, or sale of drugs. . ."

Justice Scalia wrote the unanimous opinion, a PDF of which which is available from this list of recent decisions. It covers the issues thoroughly and concisely, and if you're really into this stuff you'll want to read the original. But here's a summary, if you're not up for 17 pages of opinion (which isn't as bad as it sounds, considering the Supreme Court's traditional paper margins):

The Court found it apparent that 271(3) was designed to exempt all uses of patented inventions that were reasonably related to the development and submission of any data under the Federal Food, Drug and Cosmetic Act (the FDCA, which established the FDA.) Said Scalia, "There is simply no room in the statute for excluding certain information from the exemption. . ."

Integra argued that "the only preclinical data the FDA is interested in is that which pertains to the safety of the drug in humans," and that broader studies on mechanism of action, PK, etc. weren't meant to be exempt under 271(e). But Scalia noted that the FDA requires summaries of all these studies when an Investigational New Drug application is filed, and that these are necessary to assess the whole risk/benefit question of whether a human clinical trial should be allowed.

Integra's counterargument to that is that Merck's experiments are disqualified from the exemption, because they weren't conducted under the FDA's Good Laboratory Practices (GLP) protocols, and thus weren't intended for regulatory use. But the Court cited the law as showing that GLP studies are only required for safety assessments, and that earlier work (on mechanism, efficacy, PK, etc.) doesn't have to be run under GLP. (And they want to see non-GLP safety studies, too, if you have them, along with an explanation for why they weren't run under the protocols.)

The Court of Appeals for the Federal Circuit, when they ruled for Integra, found that the Merck/Scripps experiments "did not supply information for the FDA, but instead identified the best drug candidate to subject to further clinical testing. . .The FDA has not interest in the hunt for drugs that may or may not later undergo clinical testing for FDA approval. . .Thus, the Scripps work sponsored by Merck was not solely for uses reasonably related to clinical testing for the FDA."

The Court rejected this line of argument, which, as Scalia writes:

". . .disregards the reality that, even at late stages in the development of a new drug, scientific testing is a process of trial and error. In the vast majority of cases, neither the drugmaker nor its scientists have any way or knowing whether an initially promising candidate will prove successful over a battery of experiments. That is the reason they conduct the experiments. . . We decline to read the "reasonable relation" requirement so narrowly as to render 271(e)(1)'s stated protection of activities leading to FDA approval for all drugs illusory.

. . .the use of a patented compound in experiments that are not themselves included in a "submission of information" to the FDA does not, standing alone, render the use infringing."

Scalia and the court are completely right here, as far as I'm concerned. Drug companies constantly make each other's patented compounds for comparisons against their own, and if 271(e)(1) was interpreted as the CAFC had it, we'd be constantly second-guessing ourselves about whether we'd infringed or not. It would slow down the development of new drugs, without a doubt, and in some cases it would bring preclinical programs to an immediate halt while licensing issues were thrashed out. Which they might never be - why should a company give a competitor a license to try to beat its patented compound? Better to go tell them to grit their teeth and wait for the patent to expire.

But that's not going to happen. We're back to the "research exemption" as we've understood it, and that's a good thing.