FDA's Thinking On New Psychiatric Endpoints: What Does Brintellix CRL Portend For Nuplazid?

FDA deviated from its advisory committee's recommendation on the first cognitive function claim for an antidepressant when it issued a complete response letter for Brintellix. The decision could signal trouble for Acadia's Nuplazid for Parkinson's psychosis, which also relies on a new psychiatric endpoint.

FDA's complete response action on a cognitive function claim for Takeda Pharmaceutical Co. Ltd. and H. Lundbeck AS's antidepressant Brintellix could bode poorly for Acadia Pharmaceuticals Inc.'s Nuplazid, which is in a similar position of trying to gain approval based on a novel psychiatric endpoint.

Both drugs have received advisory committee support, but the March 29 complete response letter for Brintellix is a reminder that FDA does not always follow the advice of its advisory committees. In both cases, FDA's psychiatry drug review team raised concerns about the use of new clinical endpoints that the agency has not found adequate to identify and quantify the cognitive and psychiatric effects of drugs.

The CRL was announced the same day that FDA's Psychopharmacologic Drugs Advisory committee voted 12-2 that Nuplazid's benefits outweighed risks in treatment of Parkinson's psychosis. The same committee had voted 8-2 in support of the Brintellix cognitive function claim in major depressive disorder in February, raising the prospect that the agency is taking a stricter line on the adequacy of clinical trial endpoints than the researchers, clinicians and advocates who sit on the panel.

Psychiatry became a static area of drug development after the dramatic rise of serotonin reuptake inhibitors and atypical antipsychotics. Applications reviewed by FDA's Division of Psychiatry Products and related Division of Neurology Products consistently rank at the bottom of measures of innovation (Also see "Once More, With Statistics: FDA Defends Division Productivity Differences" - Pink Sheet, 31 Aug, 2015.).

Now, though, FDA is being asked to evaluate new endpoints to support new types of claims. Takeda and Lundbeck aim to make Brintellix (vortioxetine) the first drug to include data specifically on cognitive function in patients with major depression in its labeling. Acadia developed Nuplazid (pimavanserin) as the first product for Parkinson's disease psychosis, earning an FDA breakthrough therapy designation in the process.

FDA's briefing material and presentations to the Psychopharmacologic Drugs Advisory Committee on both applications show the division's struggle with its regulatory imperative to approve claims only if there is evidence of an effect on clinically meaningful endpoints when the endpoints themselves are inadequate efforts to quantify effects in domains, like cognition and delusion, that have a large subjective component.

"Clinical meaningfulness is not always easy to assess in psychiatric drug trials," Division of Psychiatry Products Director Mitchell Mathis observed in briefing memo on Nuplazid. "Clinical meaningfulness is difficult to quantify but it is, in the most general sense, an effect produced by the drug that matters to the patient or to the treating clinician."

Takeda and Lundbeck are not providing much clarity on the CRL or their plans for resubmission, but their statement noted that "the companies were pleased that FDA recognized the importance of cognitive dysfunction in MDD and view it as a legitimate target for drug development."

Weighing The Scales

The Brintellix cognitive function trials, FOCUS and CONNECT, used the Digital Symbol Substitution Test (DSST), a test that involves the substitution of simple digits for symbols, in their primary endpoints. The DSST is well established as a cognitive assessment, but FDA questioned the instrument's utility in major depressive disorder because the "range of cognitive abilities that can be affected by MDD is variable and includes domains beyond those that are clearly measured by the DSST" (Also see "Brintellix Review To Focus On Cognitive Function Measurement" - Pink Sheet, 1 Feb, 2016.).

The advisory committee, however, largely agreed that while the DSST was an incomplete measure of cognitive function, it did serve as a proxy for a broader battery of neurocognitive tests (Also see "Takeda's Brintellix Edges Closer To Cognitive Effects Labeling Claim" - Pink Sheet, 3 Feb, 2016.).

The pivotal efficacy trial supporting the Nuplazid application, Trial 020, is the first to use a new scale developed out of an established schizophrenia measure, the Scale for the Assessment of Positive Symptoms (SAPS). Acadia designed the SAPS-PD scale by factor analysis of previous pimavanserin Parkinson's psychosis trials that had failed to show an effect on the SAPS instrument, Mathis noted: "The questions on the SAPS that showed the most favorable change in the failed trials were compiled into a 9-item scale that measured the domains of hallucinations and delusions."

Using the new scale, Trial 020 was "strongly statistically positive," Mathis commented. But "there has been a fair amount of consideration among the review team members about how to characterize the clinical meaningfulness of the statistically significant change seen in this trial," he said.

The review decision sought a consultation from FDA's Clinical Outcomes Assessment team, which identified limitations to the SAPS-PD, like a potential overweighting of auditory hallucinations over the visual hallucinations more common in Parkinson's patients, but concluded that "we do not view them as critical flaws that would preclude the use of the SAPS-PD as a clinical outcome assessment to assess clinical benefit for regulatory use."

The COA team did, however, disagree with the sponsor's definition of clinical benefit as a 3-point improvement in SAP-PD score. The threshold should be at least 5-7 points, the COA consult said (Also see "Committee Will Sift Through 'Noise' Associated With Nuplazid Efficacy" - Pink Sheet, 28 Mar, 2016.).

"While not a regulatory requirement, in the spirit of optimizing measurement for future clinical trials, we recommend further instrument development work be done including: investigation of whether the SAPS-PD is missing key psychosis symptoms such as illusions in Parkinson disease and confirmation of the adequacy of SAPS-PD using additional patient, caregiver or clinical expert input," the COA consult stated. "The goal of this additional research is to confirm that the most important and relevant features are being assessed in a way that optimizes accuracy, reliability and ability to detect clinically meaningful change."

Weighing Medical Need

For the clinicians on the advisory committee, patient need often seemed to outweigh technical concerns about the validity of the scale used in clinical trials. Thomas Grieger, a staff psychiatrist at the Maryland Department of Health and Mental Hygiene, commented that "if you go from having a symptom every day to having a symptom twice a week, that's a huge improvement." Nuplazid "really fills a niche that nothing else is approved for that's easy to use" (Also see "Acadia's Nuplazid May Need Difficult Post-Market Trial, Panel Says" - Pink Sheet, 29 Mar, 2016.).

Medical need was probably less of a factor in FDA's decision about Brintellix, which is already on the market in the well-served therapeutic area of depression.

For Nuplazid, safety, including mortality, poses another challenge that FDA must weigh in its decision-making. The committee discussed a post-marketing study and the drug's safety profile (see related story, (Also see "Nuplazid's Black Box: Warning May Depend On Antipsychotic Classification" - Pink Sheet, 4 Apr, 2016.)).

"FDA has not approved an antipsychotic drug with this safety signal for use in the agitated, psychotic or demented elderly populations," psychiatry clinical reviewer Paul Andreason told the Nuplazid committee. "However, previously when these applications came to us, the drugs that were under review were already on the market."

"If we follow our usual logic and didn't approve pimavanserin for the treatment of Parkinson's disease psychosis, it couldn't be used off-label," he emphasized. "It wouldn't be available. So that is our dilemma from a regulatory point of view."

The user fee deadline for Nuplazid is May 1.