Korlym: A review of the PTAB’s institution decision for Teva’s PGR of the ‘214 patent.

Zachary Silbersher

We previously blogged about Teva’s ($TEVA) petition for post-grant review (PGR) of Corcept Therapeutics’ ($CORT) ‘214 patent.  Since then, on November 20, 2019, the PTAB granted institution of Teva’s petition.  The proceeding will now be litigated for another year before a final decision.  On a recent earnings call, Corcept stated that it believes the standard for institution is “pretty low,” and not necessarily dispositive that Teva will prevail.  Now that we have the institution decision, what can we glean from it, and how likely is that Corcept’s patent survives?

The crux of this dispute can be summarized as follows:  The ‘214 patent discloses a method of lowering the dosage of mifepristone from 900 to 1200 mg/day down to 600 mg/day when co-administered with a strong CYP3A inhibitor, such as ketoconazole.  Corcept argues that it discovered the optimal dosage of mifepristone that could be safely and effectively co-administered with ketoconazole, namely 600 mg.  According to Corcept, this was surprising because prior warnings in the literature cautioned against administering more than 300 mg of mifepristone a day when co-administered with ketoconazole.  Before the ‘214 patent, Corcept claims it was unpredictable whether any co-administered dosage above 300 mg would be safe and effective.

By contrast, Teva argues that figuring out the optimal dosage where mifepristone could be safely and efficaciously co-administered with ketoconazole was simply a matter of routine optimization.  The only reason that optimal dosage was previously “unpredictable” was because no tests had yet been performed to identify it.  According to Teva’s expert, Dr. Greenblatt,

“a POSA would have expected that co-administration of strong CYP3A inhibitors and mifepristone—at some dose—would be safe and effective to treat Cushing’s syndrome and other symptoms associated with elevated cortisol levels, and it would be a matter of routine experimentation to determine precisely how much to adjust the dosage of mifepristone when coadministered with a strong CYP3A inhibitor to achieve the optimum balance of safety and therapeutic efficacy.”

(See IPR2019-00048, Paper No. 19 at 16).

Institution decisions are not final decisions.  There are numerous instances of patent owners losing at the institution stage, but winning at the final decision.  This is true even for pharmaceutical patents.  Technically, the standard for granting institution is whether the petition has proven that at least one of the challenged claims is more likely than not to be unpatentable.  35 U.S.C. § 324(a) (2012). 

Given that, just because the PTAB grants institution does not mean that it will necessarily invalidate the patent.  Indeed, handicapping the outcome of an IPR or PGR after the petition has been instituted often turns on whether the decision to institute turned on a fact question or a legal question.  Institution decisions that turn on fact questions can theoretically still go either way at the final decision.  By contrast, IPRs that are instituted primarily on legal questions typically present a higher obstacle for the patent owner to reverse before the final decision.

The PTAB typically construes issues of fact in favor of the petitioner upon deciding whether to institute an IPR.  That means if the question of patentability appears to turn on a difference of scientific opinion between the experts, then it is not uncommon for the PTAB to grant institution.  Yet, after further discovery and the opportunity for each side to cross-examine the experts, the PTAB has in certain instances found that the petitioner has not met its burden, and the patent is not invalidated.

On the other hand, sometimes institution decisions turn less on factual questions, and more on legal questions.  For instance, a legal question may include a claim construction dispute between the parties.  If the patent owner unsuccessfully argues for a narrow construction before institution, which would require an element not contained in the petitioner’s cited prior art, then there is not necessarily a lot a patent owner can do to reverse that holding after institution.

Here, the PTAB’s institution decision acknowledged that much of its analysis turns on how a person of skill would interpret the warning in Korlym®’s label against administering more than 300 mg of mifepristone to patients concomitantly taking a strong CYP3A inhibitor.  Corcept argues that this would have discouraged any scientist from exploring whether a higher dosage could be safely and effectively administered with ketoconazole. 

Yet, the PTAB was not persuaded by that reasoning because the label expressly indicates that the warning itself is not based upon clinical data.  Rather, it is based upon the absence of any study into this particular issue.  Indeed, that is why the FDA expressly instructed Corcept to perform a drug-drug interaction study between mifepristone and ketoconazole.  Thus, given this evidence, the PTAB found the warning on Korlym®’s label was a pre-cautionary measure based on the absence of data rather than a conclusive measure based upon existing studies.

Resolution of this issue is technically a fact question.  The PTAB acknowledges that Corcept has yet to put in its own expert testimony to support its interpretation of Korlym® label’s warning regarding co-administration of mifepristone with ketoconazole.  Yet, though it is fact question, we can see what type of evidence Corcept would have to summon.  Namely, it would need credible and persuasive expert opinion that, despite instructions from the FDA to conduct a DDI study between mifepristone and ketoconazole to optimize the dosage, no reasonable scientist would have done that.  That will be a tough needle to thread, and whether Corcept can manage it remains to be seen.

Corcept’s other argument is that, in the absence of prior teaching, no scientist could have known in advance that 600 mg/day of mifepristone would have been the optimal dosage.  To prove a patent is invalid for obviousness does not require showing the precise invention existing in the prior art.  But it does require showing that, based upon the existing prior art, there was a reasonable expectation of success at coming up with the claimed invention.  On this basis, Corcept has some footing. 

But that argument quickly comes up against another relatively cornerstone principle of obviousness, which is routine optimization.  You cannot get a patent simply because you were the first to run an optimization study that was itself obvious to do.  Whether the ‘214 patent was simply the result of routine optimization is technically a legal question.  Indeed, PTAB acknowledged the tension between these two legal issues in its institution decision.  The PTAB states:

“We acknowledge that the art does not allow one to quantify with precision the exact highest dose of mifepristone that would be safe and effective when mifepristone is coadministered with a strong CYP3A inhibitor, like ketoconazole. . . . However, the inability to predict exact appropriate dosing absent experimentation does not, by itself, render the claimed dosage non-obvious. . . [I]t cannot be the case that any new dosage of co-administered mifepristone is patentable simply because one must test to quantify the extent of a recognized drug-drug interaction, especially when the prior art expressly teaches to conduct such testing.”

(See IPR2019-00048, Paper No. 19 at 20-21) (citations omitted).

 To the extent this is a legal question, the PTAB’s institution decision appears to provide an indication on the PTAB’s view on that question.  That will likely be difficult to turn around during the remainder of the proceeding.   

The PTAB, however, did indicate certain underlying fact questions that Corcept could use to uphold the patent’s validity.  Specifically, Teva presented considerable evidence showing that the 300 mg/day mifepristone dosage for coadministration with ketoconazole was understood not to be optimal.  Indeed, that is presumably why the FDA instructed Corcept to conduct the DDI study. 

To the extent that Corcept can summon evidence and expert opinion to undercut Teva’s evidence on this issue, that could give Corcept enough footing to save the ‘214 patent.  Yet, Corcept has already cited much of this evidence in its preliminary response.  The PTAB’s institution decision indicates it was not persuasive because, much like the Korlym® label, it acknowledges that there is an absence of clinical data. 

There are some other fact issues over which Corcept can make some hay.  The next big milestone will be Corcept’s response to the petition, which is currently due on February 13, 2020.  That will be where Corcept marshals all the evidence it can in favor of keeping the ‘214 patent alive, and that will also be where Corcept reveals whatever admissions it managed to extract from Teva’s expert.