- Regulators urged sponsors to consult with them early in the development process to get feedback on study design (Farchione, Peña).
- Assessments used in a clinical trial will need to demonstrate outcomes that are clinically meaningful to patients (Farchione).
- The regulatory pathway for neurostimulation and neurodiagnostic devices includes evaluation of risks and benefits (Peña).
- In evaluating a potential treatment for cognitive dysfunction in depression, drug regulators will likely require evidence that drug effects are not pseudo-specific, and may also require co-primary endpoints (Farchione, Laughren).
NOTE: These points were made by the individual speakers identified above; they are not intended to reflect a consensus among workshop participants.
Regulatory agencies have one central goal: to ensure the safety and effectiveness of a variety of medical products, including both drugs and devices. With regard to products targeting cognitive impairment in depression, regulatory approval of any product has yet to be achieved. According to Thomas Laughren, the time is right for the pharmaceutical industry to address cognitive impairment in depression. He said the field seems ready to coalesce around the notion that cognitive impairment in depression represents a critical unmet medical need and a legitimate target for treatment development, as described in earlier chapters. From the
perspective of regulatory agencies, he presented several important questions that the field and individual sponsors will need to address to move forward:
- What domains of cognitive impairment should be targeted?
- What assessments are optimal?
- What populations should be studied?
- What study designs are optimal?
- Assuming a successful drug development program, what claims can be supported?
- What is the clinical significance of any demonstrated drug effect?
While the job of sponsors is to put their best case forward to regulatory agencies and back it up with data, the job of the regulators is to evaluate the evidence, identify data gaps, and ensure that the claims made by the sponsor are supported by the data, said Tiffany Farchione. The burden is on the sponsor to make the case for the target population and possible enrichment strategies; the scientific rationale for targeting specific domains and using adjunctive versus monotherapy; the trial design (e.g., active comparator versus placebo controlled); the type of assessments that will provide the best measure of efficacy; how to quantify improvement; and what constitutes a clinical meaningful change. Farchione urged sponsors to consult with the agency as early as the study design stage to get feedback on aspects of a trial that will come under regulatory scrutiny.
In the United States, FDA also handles regulation of neurologic medical devices, such as the neurostimulation devices discussed earlier as well as neurodiagnostic devices, through the Office of Device Evaluation at the Center for Devices and Radiological Health (CDRH). Carlos Peña, director of neurological and physical medicine devices, said the regulatory path for devices includes evaluating risks and benefits. Medical devices are classified into three classes, with level of risk and regulatory control increasing from levels I to III. For neurodiagnostics, the diagnostic capability may be one of several factors that determine the level of regulatory oversight. Neurotherapeutics require attention to some of the same issues that are required for drugs: validated outcome assessments, clearly defined parameters for what constitutes a clinically meaningful
change, and a strong focus on patient needs. Peña pointed to a number of Guidance Documents available on the FDA website that delineate the path toward regulatory approval and, like Farchione, he urged sponsors to consult with the agency at the earliest stages of development.
The approach to regulatory approval in Europe largely mirrors that in the United States, albeit with different regulations, according to Maria Isaac, senior scientific officer at the EMA. As in the United States, the European depression guidelines do not identify cognition as a primary therapeutic target in depression. Isaac noted that the EMA works closely with FDA, exchanging views and sharing expertise in order to optimize and facilitate global development. Sponsors can seek parallel FDA–EMA qualification advice and hold joint discussions with the two agencies; however, each agency will issue separate responses to sponsors’ questions.
With regard to defining the target population, several workshop participants identified that the challenge for sponsors is to sort out the specific population that should be targeted for treatment development programs and decide whether to target cognition broadly or whether to home in on a particular domain that may be impaired in a specific subgroup. As described earlier, many cognitive domains may be impaired in people with depression, and different subgroups may have different impairments that may require different treatments. According to Laughren, regulatory agencies have already accepted targeting specific domains or subgroups of other DSM-defined syndromes, so he does not see this as an insurmountable hurdle.
With regard to assessments, arguments can be made for both subjective and objective measures, said Farchione. While the agency has not endorsed or rejected any specific cognitive assessments for MDD, they have indicated that patient-reported outcomes are important to define a change that is meaningful to patients. Farchione said assessment issues would most likely be taken up by FDA’s Study Endpoints and Labeling Development (SEALD) Team, which has the psychometric and statistical expertise to evaluate the appropriateness of a measure for a specific trial.
Farchione said that although FDA operated for many years under the general assumption that cognitive dysfunction in depression was pseudo-specific, that is, if depression improved cognition would improve as well, the opinion of FDA regarding approval of drugs for depression is evolving. New data have now paved the way for FDA to consider cognition a legitimate target for treatment in depression.
Laughren, however, said he still views pseudo-specificity as a primary regulatory challenge. If a company wants to target a specific subgroup or symptom, regulators will likely ask for a demonstration that the drug works only or better in that subgroup or on that symptom. Unless data indicate that the drug is unique in some way to the specific subgroup or symptom, the agency may be unwilling to accept a narrow focus. Laughren used the example of schizophrenia. Over the past decade, accumulating data have led to acceptance across the field that cognitive impairment is a well-established aspect of schizophrenia. To address the pseudo-specificity concern, he said investigators also had to make a strong case that even when positive symptoms of psychosis are successfully treated, many patients continue to have prominent cognitive impairment, and that cognitive impairment has a different time course than other symptoms. These data have led regulatory agencies to endorse cognitive impairment in schizophrenia as a legitimate target. Regulators have also endorsed other narrow targets for drug development such as agitation in schizophrenia and bipolar disease, indicating that they are willing to accept a narrow target if the data support it. To overcome regulators’ concerns about pseudo-specificity in trials of agents targeting cognitive dysfunction in depression, Laughren suggested that different clinical trial designs, such as adjunctive, acute-phase, and switching designs (see Chapter 4), may also provide the necessary data.
Laughren opined that several aspects of depression, including cognitive impairment, might be considered by regulatory agencies as targets for intervention. Other potential narrower targets in depression might be irritability, fatigue, and apathy. What is needed, he said, is a better understanding of the biological underpinnings of the target. To address pseudo-specificity, he said, investigators need to show that even when mood and other symptoms of depression are successfully treated, cognitive impairment persists and interferes with the ability to function. This might be achieved by demonstrating a residual phase of depression or by
making the case for subtypes of depression in which certain symptoms predominate, resulting in poorer treatment response.
Another regulatory issue that emerged with regard to cognitive impairment in schizophrenia was the need for a co-primary endpoint. The basis for this concern, said Laughren, is that even if the primary endpoint shows a benefit on a fairly abstract cognitive measure, such as word recall, the clinical relevance to a patient may be unclear. For both schizophrenia and Alzheimer’s disease, regulators want to see improvement on a functional or global measure as well. Whether a similar requirement would be made for cognitive impairment in depression has yet to be determined. The question is further complicated by the possibility that a sponsor might have reason to believe a drug works on a particular domain of cognition, and therefore could make the case of using a single measure as the primary endpoint. Another complication is that an antidepressant drug potentially could worsen certain aspects of cognition, such as speed of processing, while improving other aspects such as negative bias. These issues make the choice of endpoints extremely complex, said Laughren.
Diego Pizzagalli asked whether regulators might consider a neurophysiologic change as a primary outcome with a functional improvement as the co-primary. Laughren broadened the question to ask, at what point will our understanding of behavior at a biological level allow the abandonment of artificial DSM categories and move instead to look at actual domains of function? Moreover, he suggested that if investigators were able to identify the specific brain circuits involved in a cognitive domain such as working memory, and if data showed similar impairments by a biological marker such as fMRI across different diseases such as depression and schizophrenia, it might be possible to get a broad claim for a treatment that improved that measure. However, he acknowledged that the field is nowhere near that at this point. Farchione added that even in the absence of that clear neurobiological understanding, a clinically meaningful change in a domain such as working memory, along with a functional improvement across two or three different disorders, could
provide a legitimate claim for a more general indication. The key, said Farchione, is having sufficient data to prove that an improvement is generalizable across multiple disease conditions.