ISCTM 2010 Autumn Conference – Presentation Abstracts
Session 1: MATRICS Update and Beyond
The MATRICS (Measurement and Treatment Research to Improve Cognition in Schizophrenia) program was developed by NIMH with the goal of facilitating the development of pharmacological agents for treatment the cognitive impairments in schizophrenia. The developers of the request for proposals – Steve Hyman, Wayne Fenton, and Ellen Stover – were concerned that industry was reluctant to address new treatment targets where there was uncertainty about important issues in drug development including: the best methods for measuring treatment outcome; the optimal design for clinical trials; the most promising molecular targets; and the likelihood that the FDA would be willing to approve a drug for this indication. This brief talk will provide an overview of the products delivered by MATRICS during its 2 years of funding as well as the principles that guided the consensus building processes.
The MATRICS Approach to Measuring Cognition: Foundational Issues and Critical Decisions
M. Green
The NIMH MATRICS Initiative was launched to overcome obstacles to drug development for cognitive impairment in schizophrenia. Major obstacles included the lack of an accepted definition of cognition for schizophrenia and the lack of a consensus assessment battery that could serve as an endpoint for clinical trials. The mandate of the MATRICS Initiative was to create a path for drug approval in this area; hence, the focus was on development of methods for Phase III trials.
For the development of MATRICS Consensus Cognitive Battery (MCCB), the MATRICS Initiative followed a carefully planned method that involved committees and a series of large public meetings. Out of these discussions arose key decisions that influenced the subsequent process and products. These decisions included: 1) A focus on measurement at the level of cognitive domains versus global composite scores; 2) Neutrality on the relative importance of computerization of the battery; 3) Inclusion of the relatively new domain of social cognition in the MCCB; 4) Neutrality on the inclusion of tests from clinical neuropsychology versus experimental cognitive psychology; and 5) Clear emphasis on psychometrics and tolerability/practicality of tests included in the final consensus battery. One additional factor is that there was uncertainty in how to apply the criterion of sensitivity to treatment effects given the lack of a potent drug in this area. This presentation will discuss briefly each of the key decisions and their implications for MATRICS and follow up activities.
Real-world Application of the MATRICS Consensus Cognitive Battery (MCCB) in Multisite Clinical Trials
R. Keefe
Based upon responses from over 90 experts on cognition, clinical trials and schizophrenia, the MATRICS Neurocognition Committee set guidelines for the most important characteristics of a cognitive battery for clinical trials in schizophrenia. These included test-retest reliability, small practice effects, correlations with measures of functional capacity, practicality and tolerability. In initial validation studies completed during the MATRICS contract with NIMH, the MCCB clearly demonstrated these characteristics. It has been an empirical question whether the MCCB would demonstrate these favorable characteristics when administered in the context of the type of large multi-site industry trial for which it was designed. Further, little is known about the size of the placebo effect that can be expected for cognitive enhancement trials in schizophrenia, and whether international versions of the MCCB in non-English languages are able to collect data comparable to the English version used in the United States. Methods: This presentation will discuss five recent unpublished schizophrenia cognition trials that have used the MCCB as a primary outcome measure and followed the MATRICS study design guidelines. The total N from these studies is 1135 subjects with schizophrenia. One trial included 215 stable outpatients from 22 US sites, and the MCCB was assessed at screening, baseline and weeks 4, 8 and 10. A second trial included 323 patients from 29 US sites. This trial used the MCCB to assess cognition at screening, baseline, week 6, month 3 and month 6. Two NIMH TURNS studies have each assessed 60 patients with the MCCB as the primary outcome measure, and a final industry study with MCCB assessments at baseline (N=477), Week 8 (N=452), Week 16 (N=339) and Week 24 (N=367) will be described. Two international trials using non-English versions of the MCCB are underway, and data from these studies will be briefly presented as well. Results: In the 215-patient trial, the composite scores had excellent test-retest reliability (ICC = 0.88) and sensitivity to impairment at screening (T-score = 25.1+/-SD=11.6 for patients compared to 50.0+/-10.0 for controls) and baseline (T=26.7+/-12.1). The Pearson correlation with the UPSA-2 scores was 0.56 (P<.001). In the placebo group, the effect size of the improvement on the MCCB composite score from baseline to week 8, encompassing three assessments, was small (d=0.2). The reliability between assessments in the placebo group was high (all ICC’s >.90). The MCCB had a very low rate of missing data with 99.7% of measures non-missing and none of the 215 patients missing a composite score. The results from the 323-patient trial confirmed these findings, with excellent test-retest reliability of the MCCB composite score (ICC=0.88), low missing data rate (99.8% complete), large correlation with the UPSA-B composite score (r=.61, df=304, P<.001), and a small practice effect (z=0.18). Discussion: The ultimate criterion for the success of a test battery is its repeatable sensitivity to true treatment effects, and no test battery has convincingly demonstrated this criterion. However, the results from five separate trials suggest that the MCCB has proven reliability, practicality, and tolerability for multisite clinical trials, and the practice effect is small, allowing ample room for potential treatment effects.
Assessment of Functional Outcomes
D. Velligan
For a medication to receive an indication for improving cognition in schizophrenia, regulatory agencies require evidence of improvement in both cognition and functional outcome. The Validation of Intermediate Measures (VIM) study was developed to assess the reliability, validity and utility of a number of intermediate measures that have face validity for assessing functional outcome in schizophrenia. Because longer-term functional outcomes such as employment, or changes in marital status are not likely to be improved during the course of a typical clinical trial, the study focused on intermediate measures of functional capacity or everyday functioning that are thought be more amenable to change over this time period. Conclusions of the VIM study indicated that the UCSD Performance-based Assessment (UPSA II), the UPSA Brief and the Brief Version of the Test of Adaptive Behavior in Schizophrenia (TABS) had the best psychometric properties of the scales compared, and were reasonable choices for use as co-primary outcome measures in medication trials examining the efficacy of agents designed to improve cognitive functioning in schizophrenia. However, because a large number of efficacy studies of novel compounds are now conducted as multi-site international trials, the cultural applicability of intermediate measures of functional outcome was examined in a companion study called the Cross-Cultural validation of Intermediate Measures (CIM) Study. In the CIM study, clinical investigators experienced in conducting medication trials in schizophrenia rated each of the subscales of the intermediate measures on the extent to which the subscale would be culturally applicable in their countries, and across genders, income levels, rural versus urban residence, and minority status. Results of the CIM study suggested that with one exception, all subscales of all the intermediate assessments of functional outcome were identified as likely to have problems in their adaptation to non-U.S. cultures. India, China and Mexico presented the greatest challenges in adaptation. A new study is proposed to investigate whether for international trials it may be most appropriate to use customized measures that differ from country to country or to use a new functional battery composed of the subscales rated in the CIM study as likely to be most adaptable across multiple countries and cultural contexts with minimal alteration.
Challenges in Drug Trials for Cognitive Impairment in Schizophrenia
M. Egan
The first wave of drug trials targeting CIAS have concluded and the results are largely negative. Reasons for failure may include issues related to dose, duration of treatment, use of concomitant medication, lack of target engagement, and pursing the wrong target. Several other possibilities are worth considering.
First, design and conduct of the clinical trials could play a role. Current trials are unable to distinguish failed trials (trial insensitive to meaningful change) from negative trials (drug doesn’t work). Demonstration of a practice effect could ameliorate this to some degree. Cognitive remediation (CR) has been suggested but including this substantially increases costs and patient burden Furthermore, CR has not been necessary to demonstrate efficacy in other diseases (e.g. cholinesterase inhibitors in AD and stimulants in ADHD).
Second, the biology of cognition and schizophrenia could require other strategies. Many drug targets may impact only 1-2 specific cognitive domains. If drugs improve, for example, only working memory, the development path would require a specific, related functional measure. This suggests that additional work relating specific domains to function might be needed. A piecemeal approach to cognition may not be ideal but it could be a more plausible alternative from a biological perspective.
Third, given the apparent heterogeneity in schizophrenia, study designs that address this should be considered. Randomized withdrawal, where only clear responders are randomized to active drug vs placebo, may have the potential to preselect a responsive subgroup.
Finally, once proof of concept is achieved, multinational trials would be needed. While much has been done to facilitate this, concerns remain over lack of norms for many languages and uncertainty on the cultural adaptability of functional capacity measures.
Given the many challenges to drug development, continued public/private collaboration may be necessary to demonstrate the viability of clinical trials in CIAS.
Assessing Cognition in Serious Mental Illness: NIMH Perspectives
B. Cuthbert
The purpose of this talk will be to review past and future NIMH perspectives regarding the assessment of cognition in serious mental illness. The MATRICS project is seen as an invaluable first step not only for its substantive role in providing endpoints in clinical trials, but also for its role as a vehicle to address the issues that arise in such an innovative project — such as establishing co-primary measures, forging procedures to deal with language and cultural translation, and integrating various aspects of cognition. The CNTRICS process represents the next generation of assessments, and illustrates the process required to translate basic paradigms in cognitive neuroscience to practical clinical use. For both MATRICS and CNTRICS, a critical need will be demonstrate not only satisfactory psychometric properties, but also sensitivity to the effects of relevant compounds. Over the next several years, NIMH anticipates broadening the scope of these assessments beyond syndromal schizophrenia. As exemplified by the new Research Domain Critieria (RDoC) project and the NIMH Advisory Council Intervention Workgroup Report, NIMH will move toward an emphasis on circuit-based mechanisms that may cut across traditional disorder categories. It is hoped that this approach will foster new treatment development by providing targets with greater validity, and contribute to more effective personalized medicine.
Bridging the Gaps Between Animal Models, Human Neuroscience and Schizophrenia Clinical Trials: Rational Enhancement of Signal Detection
J. Sweeney
Cognitive deficits are a major cause of functional disability in schizophrenia. Developing interventions to reduce these deficits face the well-recognized challenges of a limited knowledge of the neurobiology of schizophrenia and of the neurochemical basis of its associated cognitive impairments. They also face more tractable challenges associated with selecting optimal outcome measures for trials at various stages along the drug discovery/registration pathway. Global cognitive impairment assessed with standardized neuropsychological tests (e.g., MATRICS battery) is linked to functional disability and its assessment is readily implemented across sites for Phase III trials. However, this approach has significant limitations for POC studies: 1) it provides intercorrelated measures of global deficit rather than specific measures of cognitive abilities likely to be influenced by a drug targeting a specific neurochemical mechanism; 2) its measures have proven remarkably resistant to change, not only from acute drug effects but even the presence of acute psychosis; 3) it fails to directly measure drug effect on the target organ – the functional brain systems that support specific aspects of cognition targeted by a drug; and 4) it has only modest linkage to preclinical platforms. These limitations may contribute to the high rate of negative trials across a range of targeted mechanisms.
A new strategy may be needed for POC studies. Rather than use “off-the-shelf” neuropsychological tests alone as outcome measures (be they computerized or otherwise), other types of outcomes should be considered that: 1) have a tight linkage to the neurochemical and functional brain systems specifically targeted by the drug’s mechanism of action, and the specific cognitive operations they subserve (global cognitive improvement is not the place to start); 2) have demonstrated sensitivity to effects of target drug in preclinical models (it is time for translational strategies); 3) have high sensitivity to rapid-onset drug effects (neuropsychological tests may lack sensitivity needed for small sample POC studies); and 4) assess brain as well as cognitive/behavioral outcomes (study the target organ as in other fields of medicine).
To transition to this type of POC strategy, the limited human cognitive neuroscience capabilities in industry typically will require tight industry-academic partnerships. With early involvement in drug discovery, these partnerships can integrate and parallelize preclinical and clinical programs. This strategy can enhance confidence in early go-no go decisions, provide useful ancillary measures for later phase trials, and potentially identify patients most likely to benefit from a study drug. This may require a scaling back of hopes for the quick discovery of a blockbuster drug, but an incremental rational drug discovery process may in the long run provide the greatest hope for developing treatments for the debilitating neurocognitive deficits associated with psychotic disorders.
Cognitive Remediation as a Platform for Clinical Trials: Issues for Trial Design
A. Reichenberg
NEWMEDS – Novel Methods leading to New Medications in Depression and Schizophrenia (http://www.newmeds-europe.com/) initiative is an academic-industry partnership which consists of 19 participating institutions based in nine different EU member states as well as in Iceland, Switzerland and Israel. Nearly all major biopharmaceutical companies including AstraZeneca, Eli Lilly, Janssen Pharmaceutica, Lundbeck A/S, Novartis, Orion, Pfizer, Roche and Servier are participating in the project. Academic institutions involved are: King’s College London, Karolinska Institutet (Stockholm), The University of Cambridge, Central Institute of Mental Health (Mannheim), CSIC (IIBB, Barcelona), the University of Manchester and the Bar Ilan University (Israel). Additionally, two pharmaceutical small and medium-sized enterprises, deCODE (Reykjavik) and Psynova (Cambridge) take part.
This talk will present an overview of NEWMEWDS and its scientific rational. We will discuss how the MATRICS initiative and the MCCB have shaped projects in NEWMEDS, and present work focusing on: (A) the development of cognitive measures ideal for international trials and (B) the design of trials for testing cognitive enhancing drugs in schizophrenia.
Cognitive Assessment for Schizophrenia Clinical Trials in the Post-MATRICS Era
R. Bilder
The MATRICS project achieved its goals of providing consensus about cognitive domains impaired in schizophrenia, and a test battery (MCCB) to measure these. There remain critical challenges and opportunities to develop assessment methods that will ultimately yield improved measurement of relevant functions and inform rational treatment development. It should be recognized that cognitive domains defined by consensus, while offering an important way forward, remain under-specified with respect to: (1) pathophysiological hypothesis about schizophrenia; (2) putative mechanisms of treatments; and (3) specific relations to functional outcomes. Novel strategies for construct definition will be needed that go beyond dimension reduction within the cognitive variable space. On one hand we need to incorporate external biological validation. On the other hand we need to better identify the mediating role of cognitive functions as these impact other instrumental functions and quality of life. Investment in more elaborate Phase II programs is prudent to help surface validity issues that might enhance the odds of success in Phase III. These strategies include instrumenting POC studies to maximize translation of preclinical findings, and deploying “MATRICS-plus” designs in later Phase II to assure domain-specific signal is not missed. The “-plus” measures would ideally demonstrate convergent validity with respect to functional constructs identified as important in the MATRICS process, help refine these constructs psychometrically, and provide additional validation with respect to biological processes affected by schizophrenia or treatment mechanisms putatively underlying cognitive benefits. Instrument development would benefit from open-access content-banks that permit widespread dissemination of tools for basic investigations, and application of methods from modern psychometric theory to advance novel construct definition while maintaining back-compatibility with established measures. Links to real world outcomes can be studied further by leveraging this same infrastructure in broader community networks. Together these strategies can help provide a path for treatment development that puts cognitive assessment in appropriate context considering both brain biology and our shared public health mission.
Building Beyond a Cognitive Battery
S. Koslow
BRAINnet Foundation was established to facilitate the cure of human brain disease. It is a 501(c) 3 US based Tax Exempt research Foundation. The BRAINnet Foundations operating principals are: (1) Free and open electronic sharing of all data; (2) Open research network of collaborating members; (3) Continued expansion of the global database of human brain function in health and disease across the life span with its global membership; (4) Using the same proven standardized protocols and measurements techniques in all studies allowing for data pooling across disorders, sites and studies; (5) Integrates data across the genome, Brainmarkers*, and clinical measures; (6) Open sharing of data without data contribution; (7) Self organizing, cooperative collaboration and openness within the global scientific community to analyze and publish data; (7) Successful competing for grant support with collaborators to carry our large clinical studies of human brain disorders. Currently it has as part of its global consortium of 243 researchers from 13 countries. The database contains clinical, cognitive, brain and gene data acquired on the same subjects with same standardized methods and protocols, with datasets from healthy and multiple clinical brain disorders. Currently 252 peer-reviewed papers reporting on the data have been published. This presentation will provide additional information on the establishment of the Foundation, details of its operation and standardized methods and protocols as well as participation and membership opportunities.
Exemplar Applications and Results From BRAINnet: What Are The Implications for Understanding, Modeling, Diagnosing and Treating Brain Disorders?
S. Silverstein
Following Steve Koslow’s presentation on the development, goals, products and growth of BRAINnet, this talk will review examples of integrative research that has emerged from the initiative. The BRAINnet project has now produced over 250 papers, including studies of healthy people aged 6-100, normal aging, schizophrenia, depression, attention deficit-hyperactivity disorder (ADHD), PTSD, eating disorders, mild cognitive impairment (MCI), and Alzheimer’s Disease. For example, in schizophrenia, relationships between cognition, synchrony, symptoms, grey matter loss, and level of functioning, including in first episode patients, have helped clarify how the illness progresses over time, as well as the links between phenomena at multiple levels of analysis. In depression, data have demonstrated an interaction between BDNF allele type and early life stress that predicts loss of grey matter, psychophysiological reactivity, symptoms, and cognitive deficits. In ADHD, aspects of cognitive and emotion processing deficits have been linked to specific psychophysiological abnormalities. In anorexia nervosa, cognitive and electrophysiological markers have demonstrated differential responsiveness to treatment. Finally, relationships between degree of cognitive decline and specific electrophysiogical indices have differentiated Alzheimer’s disease from MCI and normal aging. These and other data demonstrate the power of an integrative approach grounded in a large and growing database. Specifically, integration of genetic, psychophysiological, cognitive, life history, and personality data can increase our ability to model the etiology and development of psychopathology. Also, the large normative database, and common testing platform for studies of multiple disorders can assist in identifying common vs. unique mechanisms and pathways across disorders. Consistent with the NIMH Research Domain Criteria (RDoC) initiative, this can lead to more effective ways of conceptualizing, classifying, diagnosing, and treating mental illness. Importantly, treatment effects can be examined at multiple levels, in an integrated fashion, with clear reference to age- and gender-matched population norms. Finally, the large (N>8000) and growing database can serve as the basis for a personalized approach to treatment and prognosis.
Session 3: DSM5: Implications for Clincial Trials
Implications of DSM-5 Mood Disorders for Future Treatment Efficacy Trials
J. Fawcett
Jan Fawcett, M.D. is the chair of the Mood Disorders Work Group. He will review key changes in the diagnostic criteria in mood disorders and their rationale. He will address the inclusion of ratings of severity of suicidality and anxiety dimensions across all mood disorders. He will discuss the ways in which these changes might affect inclusion and exclusion criteria for future clinical trials. He will also address how results from current studies may be applied in the DSM-5 context.
Schizophrenia and Other Psychotic Disorders
J. Bustillo
Systematic application of potential validators (eg: neuroimaging, genetics, postmortem) has fallen short of a substantial understanding of the Schizophrenia syndrome as defined by DSM-IV. With the development of DSM-V, there is an opportunity to update or substantially modify the nosology of Schizophrenia and other Psychotic disorders, to better accommodate clinical and research needs. Joining Schizophrenia and Bipolar disorder into a general psychotic syndrome was considered based commonalities in some neuroimaging and genetic findings. Incorporating Schizotypal Personality disorder as part of a Schizophrenia continuum is supported by a larger body of research, but will depend on future determinations regarding the structure of the Personality disorders chapter. The following issues have been identified for possible changes:
1- Eliminate the Schizophrenia subtypes. These are rarely stable in patients, have minimal therapeutic implications and research has not validated them. Symptom dimensions within psychosis (ie: reality distortion, disorganization, negative, mood and cognitive) have been replicated in many datasets. They offer a more structured venue for clinicians to routinely assess important symptoms to target with more specific interventions. The perceived utility and reliability of these dimensions will be examined in the DSM-V field trials.
2- De-emphasize the link between Catatonia and Schizophrenia. Will review the data that supports the creation of a new category for the Catatonic syndrome versus a more conservative approach of adding a qualifier to the main disorders (Schizophrenia, Mood disorder and General Medical Condition).
3- Schizoaffective disorder. There is a clinical need to accurately capture the prominent mood symptoms that occur in schizophrenia for optimal therapeutics. Also, the proper delineation of the border between mood and psychotic disorders may accelerate the discovery of biomarkers. The DSM-IV Schizoaffective category is unreliable but widely used in clinical settings and researchers are generally averse to study it. The DSM-V field trials will examine the reliability of: a) further operationalizing Criterion C from “…mood episodes are present for a substantial portion of the total duration of the active and residual periods of the illness”, to “…for a majority of the total duration…”; and b) the elimination of Criterion B, which requires a temporal dissociation between the presence of mood and psychotic symptoms (this Criterion is a source of discrepancy with ICD-10).
4- Psychosis Risk Syndrome. The functional outcome in Schizophrenia remains poor even in compliant, non-substance abusing patients, so the need for very early identification and treatment is great. Several diagnostic schemes for defining mostly adolescents with emergent, mild to moderate delusions, hallucinations and/or disorganized speech have been tested in the last decade. The DSM-V field trial will examine the reliability of a set of criteria based on the NAPLS study, to differentiate such health seeking adolescents from those with schizophrenia, bipolar and depressive disorders. Reliable definition of this syndrome may facilitate the development of fundamentally new therapeutic strategies with the hope of altering the course early in the illness.
Regulatory Perspectives from FDA and EMEA
M. Mathis
Mitchell Mathis, M.D. is the Deputy Director of the Division of Psychiatry Products in the FDA. He will comment on the relationship of current and future diagnostic categories on both overall drug development programs and labeling. Karl Broich, M.D. Deputy Head of Federal Institute for Drugs and Medical Devices will provide perspectives from the European EMEA vantage point.
R. Hirschfeld
Robert Hirschfeld, M.D., co-chair of the session has extensive research experience in depression research and in the relationship of diagnostic criteria to clinical assessment and treatment. He will initiate the discussion with his perspective on implication for clinical trials and treatment and facilitate comments from the audience