CNS SUMMIT 2013 ABSTRACTSNovember 14–17, 2013
Boca Raton, Florida

Innov Clin Neurosci. 2013;10(11–12 Suppl B):1–18

CNS DRUG COMPOUNDS/THERAPIES

QT Effects of Levomilnacipran ER at Therapeutic and Supratherapeutic Doses in Healthy Subjects

Presenters: Chen L, Chen C, Carrothers T, Greenberg W, Periclou A, Ghahramani P

Affiliations: All authors are from the Forest Research Institute Harborside Financial Center in Jersey City, New Jersey.

Background: The effect of levomilnacipran extended release (LVM) on QTc interval was assessed.

Methods: Healthy adults were randomized to three groups: escalating doses of LVM QD Days 1–24 of study administration (up to 120mg on Day 11; 300mg on Day 24) and placebo (PBO) Day 25; moxifloxacin (Day 1) and PBO (Days 2–25); or PBO (Days 1–24) and moxifloxacin (Day 25). On Days -2, 1, 11, 24, and 25, 24-hour electrocardiograms (ECGs) were obtained (all in supine position), and on Day -1 an ECG was obtained while exercising to provide increased heart rate at baseline. QT intervals were corrected using Fridericia (QTcF) or individual method (QTcNi) with QT/RR data from Day -1 or Day -2.

Results: Assay sensitivity was demonstrated. Upper bound of the 90-percent confidence interval (CI) for least-squares mean difference in QTc between LVM and PBO (DDQTc) was below 10msec for QTcNisupine (correction with supine data) and QTcF, but was 10.38msec for QTcNiexercise (correction with exercise data) at three and eight hours (LVM 120mg) and 10.96msec at 16 hours (LVM 300mg). Analysis of QT-RR showed the relationship differed between Day -1 and Day -2; the Day -2 data aligned more closely with relationship post-baseline. Furthermore, because of hysteresis, the number of data points that could be collected on Day -1 (exercise) was inadequate for a number of subjects. For these reasons, the exercise correction was considered inappropriate in this study.

Conclusion: At doses of 120/300mg, LVM does not prolong QTc to a clinically meaningful extent.

Financial disclosures/funding: All authors are employees of Forest Research Institute, sponsor of this study.

Keywords: QTc interval, levomilnacipran ER, safety

PharmacoMRI and Cognitive Effects of the Low-trapping NMDA Open-channel Blocker AZD6765 Compared with Ketamine in Untreated Major Depressive Disorder

Presenters: Dawson G, Deakin J, Williams S, Downey D, McKie S, Goodwin G, Rylands A, Harmer C, Craig K, Dourish C, McCarthy D, Smith M

Affiliations: Dawson G, Rylands A, Craig K, and Dourish C are from P1vital Limited in the United Kingdom; Deakin J, Williams S, Downey D, and McKie S are from the University of Manchester, United Kingdom; Goodwin G and Harmer C are from the University of Oxford, United Kingdom; McCarthy D and Smith M are from Astra Zeneca in the United States.

Background: The aim of this study was to determine the immediate effects of AZD6765 and ketamine on neural activity in the subgenual cingulate cortex (SGC) and its relationship with subsequent change in depressive symptoms and emotion processing using pharmaco- and functional magnetic resonance imaging (phMRI, fMRI).

Methods: Sixty treatment-naïve men and women aged 18 to 45 years with major depressive disorder (MDD) were randomly assigned to three groups to receive intravenously ketamine, AZD6765, or placebo during a 60-minute phMRI scan. Twenty-four hours later, behavioral and fMRI responses to emotional stimuli were recorded. Baseline and follow-up Montgomery Asberg Depression Rating Scale (MADRS) scores were recorded.

Results: Both AZD6765 and ketamine increased SGC BOLD signal responses; no decreases were seen in any brain region. The SGC responses correlated with improvement in MADRS scores 24 hours and seven days post-infusion. Following administration of AZD6765, interviewer-rated psychotic and dissociative symptoms were minimal and not statistically significant. In contrast, ketamine produced a moderate statistically significant increase in dissociative symptoms. Both drugs reduced amygdala responses to fear and sadness in the emotional faces task 24 hours post-infusion.

Conclusion: Activation of the SGC was seen following both drugs and this effect was associated with improvement in depressive symptoms 24 hours and seven days post-infusion. The results suggest that AZD6765 and ketamine both have antidepressant-like effects on emotion processing in the brain and that diminished NMDA glutamate neurotransmission in the SGC is a likely proximal mechanism.

Financial disclosures/funding: Dawson G, Deakin J, Goodwin G, Harmer C, Dourish C, McCarthy D, Smith M are employees of AstraZeneca. This study was supported by funding from AstraZeneca.

Keywords: Depression, AZD6765

PH10 may be a New Rapidly Acting, Intranasally Administered Antidepressant

Presenters: Liebowitz M, Nicolini H, Hanover R, Monti L

Affiliations: Liebowitz M is from the Medical Research Network in New York, New York; Nicolini H is from the Carracci Medical Group in Mexico City, Mexico; Hanover R is from the Westport Compass Consulting in Salt Lake City, Utah; Monti L is from Pherin Pharmaceuticals in Los Altos, California.

Background: More rapidly acting antidepressants are needed. Pherines, which are intranasally administered neurosteroids, may have rapid and potent efficacy in major psychiatric syndromes. PH10 is a pherine, or a small molecule that specifically engages peripheral chemoreceptors in the nasal passages, and triggers neural impulses in the central nervous system (CNS).

Methods: Thirty patients with MDD and a HAM-D-17 score of greater than 17 were randomized to receive intranasally low-dose PH10 (3200ng/day), high-dose PH10 (6400ng/day), or placebo for eight weeks. The primary outcome measure was the endpoint HAM-D score. Secondary outcome measures included CGI and Q-LES-Q-SF.

Results: Effect sizes (Cohen’s d) were 1.13 for the low-dose group and 0.77 for the high-dose group when compared to the placebo group at endpoint. The low-dose group also showed an effect size of 1.05 and the high-dose group 0.75 in comparison to placebo after one week of treatment. Side effects were benign.

Conclusion: PH10 may be a novel, rapidly acting, potent, and well tolerated antidepressant. Further trials are clearly indicated.

Financial disclosures/funding: Pherin Pharmaceuticals provided funding for this study to Dr. Nicolini. Drs. Liebowitz, Hanover, and Monti have stock or stock options in Pherin.

Keywords: Rapidly acting antidepressant, novel antidepressant

ALKS 5461, a Novel Opioid Modulator for the Treatment of Depression: Phase 2 SPCD Results

Presenters: Martin W, Bodkin J, Marshall R, Leigh-Pemberton R, Thase M, Memisoglu A, DeSomer M, Ehrich E, Fava M

Affiliations: Martin W, Marshall R, Leigh-Pemberton R, Memisoglu A, DeSomer M, and Ehrich E are from Alkermes, Inc. in Waltham, Massachusetts; Bodkin J is from McLean Hospital in Belmont, Massachusetts; Thase M is from the University of Pennsylvania in Philadelphia, Pennsylvania; Fava M is from the Massachusetts General Hospital in Boston, Massachusetts.

Background: Placebo response in antidepressant trials is highly variable and often substantial. The sequential parallel comparison design (SPCD) methodology has been shown to enhance signal detection with relatively smaller samples in clinical trials. Although the endogenous opioid system plays a key role in mood regulation, the contemporary use of opioids for depression is limited by abuse potential. ALKS 5461 is a co-formulation of buprenorphine (BUP), a partial mu agonist, and ALKS 33, a counteracting mu antagonist designed to yield a nonaddictive opioid modulator.

Methods: The SPCD methodology, with two 4-week treatment stages, was used to evaluate the safety, tolerability, and efficacy of ALKS 5461 in subjects with major depressive disorder (MDD) with inadequate response to 1 to 2 courses of a selective serotonin reuptake inhibitor (SSRI) and/or serotonin-norepinephrine reuptake inhibitor (SNRI). Treatment groups included 2mg/2mg BUP/ALKS 33, 8mg/8mg BUP/ALKS 33, and matching placebo. All subjects remained on background SSRI/SNRI therapy.

Results: One-hundred and forty-two patients were enrolled. In the combined estimates SPCD analyses, ALKS 5461 significantly reduced depressive symptoms on both the 17-Item Hamilton Depression Rating Scale (HAMD-17; p=0.006) and the Montgomery-Asberg Depression Rating Scale (MADRS; p=0.001). As predicted, drug-placebo differences on response rates, HAMD-17 scores, and MADRS scores were greater for both doses in Stage 2 versus Stage 1. The most common adverse events included nausea, vomiting, and sedation, typical of opioid therapy.

Conclusion: Opioid modulation showed important efficacy and may be a novel treatment approach for treatment-resistant MDD. The SPCD methodology reduced placebo response in the second treatment phase, enhancing the efficacy signal in Stage 2 compared to Stage 1.

Financial disclosures/funding: This study was sponsored by Alkermes, Inc. Drs. Martin, Marshall, Leigh-Pemberton, Memisoglu, DeSomer, and Ehrich are employees of Alkermes. Drs. Bodkin, Thase, and Fava are paid consultants to Alkermes.

Keywords: ALKS 5461, opioid modulator, depression

Anticipated Clinical Effect of the New Alzheimer Drug ANAVEX 2-73 on a Calibrated Quantitative Systems Pharmacology Platform

Presenters: Missling C, Maurice T, Spiros A, Roberts P, Geerts H

Affiliations: Missling C is from Anavex Life Sciences in New York, New York; Maurice T is from the University of Montpellier 2 INSERM U 710 and the Amylgen in Montpellier, France; Spiros A, Roberts P, and Geerts H are from In Silico Biosciences in Lexington, Massachusetts; Roberts P is also from OHSU in Portland, Oregon; Geerts H is also from the University of Pennsylvania in Philadelphia, Pennsylvania.

Background: ANAVEX 2-73 is a novel cholinergic compound with sigma-1 agonist properties and Na-channel modulation pharmacology in development for Alzheimer’s disease (AD). While the sigma-1 pharmacology could lead to neuroprotection, this study simulated the symptomatic dose-response effect of ANAVEX 2-73 in a well-characterized quantitative system pharmacology (QSP) model of AD.

Methods: A complex computer-based biophysically realistic cortical network model with the physiology of 12 different membrane central nervous system (CNS) targets simulating the stability of a memory trace was calibrated with clinical trials in healthy humans, schizophrenia patients, and AD conditions.

Results: The results suggested that ANAVEX 2-73 alone shows a dose-dependent improvement in ADAS-Cog at target engagement levels well above 60 percent (measured with M2 mAchR radioligand 18F-TZTP); however, augmentation with 5mg donepezil significantly improved the clinical outcome already at target engagement above 40 percent with an anticipated maximal differential effect of 3 to 4 points on the ADAS-Cog scale (12 weeks) and 1.5 to 2.5 points (26 weeks) on top of the donepezil alone response. With regard to the placebo values, this corresponds to a maximal effect of anticipated ADAS-Cog response of 6 to 7 points at 12 weeks and 4.5 to 5.5 points at 26 weeks for the combination donepezil/ANAVEX 2-73. The Na-channel pharmacology (at higher doses) can reduce the clinical outcome dependent upon the nature of the interaction with different Na-channel isoforms. Full antagonism of ANAVEX 2-73 with the presynaptic M1 mAChR lowers the maximal clinical response by 0.5 to 1 point on the ADAS-Cog scale.

Conclusion: ANAVEX 2-73 has potential in symptomatic treatment of AD, especially when combined with low-dose donepezil (5mg) at target engagement levels above 40 percent and at earlier time points. The effect will likely be clinically detectable and meaningful. However, at target engagement levels beyond 70 percent, the negative effect of Na-channel pharmacology could reduce the clinical outcome.

Financial disclosures/funding: This study was commissioned by Anavex Life Sciences to In Silico Biosciences who had full control over the study.

Keywords: Alzheimer’s disease, ANAVEX 2-73, computer simulation

Amitifadine, Gender, and Alcohol Consumption in MDD

Presenters: Pedrelli P, Freeman M, McKinney A, Bradshaw M, Hurt S, Hsu T, Fava M

Affiliations: Pedrelli P, Freeman M, and Fava M are from the Massachusetts General Hospital, Department of Psychiatry, at Harvard Medical School, Boston, Massachusetts; McKinney A and Hsu T are from Euthymics Bioscience/ Neurovance Inc., Cambridge, Massachusetts; Bradshaw M is from Global Consulting Partners in Medical Biometrics; Hurt S is from the Weill Medical College of Cornell University, Department of Psychiatry, New York, New York.

Background: Amitifadine is a novel serotonin-preferring triple-reuptake inhibitor with a ratio of 1:2:8 for the transporters of serotonin, norepinephrine, and dopamine, respectively. A previous trial demonstrated a significant effect of amitifadine 100mg in a six-week, randomized, controlled trial for major depressive disorder (MDD).

Methods: We present data on alcohol consumption in a subsequent multicenter, randomized, double-blind, placebo- and paroxetine- controlled trial of amitifadine in MDD unresponsive to one previous adequate trial of antidepressant treatment. Alcohol consumption data were collected with the MGH Compulsive Habits Questionnaire.

Results: Patients (N=342) were enrolled in the study, and treated patients in the completer arms varied from 42 to 50 compared to 131 placebo-treated patients. Among patients who consumed any alcohol, paroxetine 40mg increased drinking by 0.12 drinks/day, whereas EB-1010-100mg reduced drinking by an average of 0.12 drinks/day (p=0.0722). At baseline, 5.26 percent (3/57) of patients in the paroxetine arm consumed 1 or more drinks per day, and after six weeks treatment 8.77 percent (5/57) consumed 1 or more drinks per day. In contrast, 3.57 percent (2/54) of patients on EB-1010 100mg consumed 1 or more drinks per day at baseline and this was reduced to 1.79 percent (1/56) after six weeks treatment. Women were primarily responsible for the mean reduction compared to men (-1.75 vs 0.00).

Conclusion: Amitifadine is a triple-reuptake inhibitor with a unique profile of monoamine affinity associated with a directional difference compared with that of paroxetine in alcohol consumption during treatment for a major depressive episode, particularly among women.

Financial disclosures/funding: McKinney A and Hsu T are employees of Euthymics Bioscience Inc.

Keywords: Amitifadine, alcohol, gender

GENOMICS ANALYSIS

Toward Precision Medicine: Tute Genomics, a Cloud-based Application for Analysis of Personal Genomes

Presenters: Robison R, Wang K

Affiliations: Robison R and Wang K are with Tute Genomics in Provo, Utah.

Background: The cost of sequencing the human genome has dropped significantly in recent years, making it possible for whole genome and exome sequencing to be used in clinical settings. However, data analysis remains as the major hurdle before widespread adoption and individualized “precision medicine” can become a reality. Precision medicine involves the use of genomic information to identify disease origins, develop targeted therapies, and improve outcomes. Whole genome data can help clinicians and researchers identify disease-causing variants, quantify health risks, and personalize treatments through pharmacogenomics.

Methods: We developed a cloud-based application called Tute Genomics to process, store, annotate, and interpret personal genomes using the following approach: 1) annotation, bringing in outside data sources, such as ClinVar and HGMD; 2) filtering, using default or customized pipelines for variant filtering and stepwise reduction; 3) probabilistic modeling, using a machine-learning algorithm to score all variants in a genome by their likelihood of having functional consequences; and 4) storage, an ultrafast data-querying system for realtime retrieval and organization of user files and analyses.

Results: Users can upload genome variants files via a secure web interface, and all major file formats are supported (i.e., VCF, ASM). Tute can also perform case-control and family-based analyses on a collection of personal genomes, to help identify and prioritize genomic variants and biomarkers related to any given phenotype or genetic condition.

Conclusion: Tute Genomics is opening a new door for precision medicine by enabling researchers and clinicians to utilize human genome data for scientific discovery and individualized treatment.

Financial disclosures/funding: None were reported.

Keywords: Genomics, cloud-based application, data analysis

PATIENT ASSESSMENT

METHODS/TOOLS

Analysis of Serum Biomarker Expression in Adolescent and Adult MDD Patients and Nondepressed Subjects

Presenters: Bilello J, Thurmond L, Smith K, Renshaw P, Kondo D

Affiliation: Bilello J and Thurmond L are from Ridge Diagnostics, Research Triangle Park, North Carolina; Smith K is from the Ridge Diagnostics, Research Triangle Park, San Diego, CA; Renshaw P and Kondo D are from the University of Utah Brain Institute in Salt Lake City, Utah.

Background: An estimated 25 percent of young adults will experience a major depressive episode (MDE) by age 24, the highest incidence rate of any adult age-group. However, the percentage of youth with an MDE receiving treatment for depression declined from 40 percent in 2004 to 38 percent in 2011. Since this group is both underdiagnosed and undertreated, we investigated whether biomarker expression and analysis could be a clinically useful aid to early diagnosis and treatment.

Methods: Male and female subjects 13 to 20 years in age were eligible for inclusion in this study. Thirty-seven subjects (19 with major depressive disorder [MDD] and 18 nondepressed) were recruited, and evaluated. A data set of 68 adult MDD patients and 86 adult non-MDD control subjects was used for comparison. Immunoassays were used to quantify the serum levels of 10 biomarkers (Papakostas et al, Molecular Psychiatry 2013).

Results: Only two proteins, myeloperoxidase (p=0.002) and epidermal growth factor (p=0.08), showed significant differences between nondepressed adolescent subjects and those with MDD. Unpredictably, serum cortisol levels were 17.1±5.8 and 14.7±6.7µg/dL for normal subjects and MDD patients, respectively. Serum GABA levels in MDD patients (783±344µg/mL) were slightly higher than normal subjects (630±341µg/mL) but not statistically significant (p=0.19). While no individual biomarker could be used to identify MDD patients, algorithms (e.g., MDD Score, SVM) were explored.

Conclusion: We found significant differences in both BMI and biomarker patterns between this group of youthful subjects and representative adult populations. Nonetheless, preliminary analysis of the data suggests we can develop a panel and algorithm specifically for adolescents.

Financial disclosures/funding: Bilello J, Thurmond L, and Smith K are employees and stockholders of Ridge Diagnostics, Inc. Renshaw P is a stockholder in Ridge. The study was supported in part by NIMH grant 1R43MH90817-1.

Keywords: Depression, adolescents, biomarkers

Key-Lock Principle in Pharmaco-EEG: A Tool for Identifying a Possible Drug Effect

Presenters: Dorffner G, Anderer P, Saletu B

Affiliations: Dorffner G and Saletu B are from the Medical University of Vienna, Austria; Dorfnner G and Anderer P are from The Siesta Group in Vienna, Austria.

Background: This presentation provides an overview of insights into the so-called “key-lock effect” in pharmaco-electroencephalography (pharmaco-EEG). The key-lock effect—first coined by Saletu et al—originally referred to the observation that many central nervous system (CNS)-active drugs show changes in the pharmaco-electroencephalography (EEG) that go in the opposite direction of changes observed in patients of certain psychiatric or neurological disorders, as compared to normal subjects. This observation does not only help in proving whether a drug is likely to counteract a disorder’s effects on the brain, but can also serve as an early tool to identify a compound’s probable target, as well as to rule out such targets.

Results: Our overview of study results focuses on psychiatric disorders and includes examples from our own research as well as that from other groups. We demonstrate that, based on the key-lock principle, characteristic patterns of EEG changes can often be seen in early phase trials that help in the decision-making process about where a compound can go and where it cannot. All commonly used paradigms of pharmaco-EEG— including spectral analysis of vigilance-controlled EEG and event-related potentials—play a role in this process and can thus be of use.

Conclusion: We conclude that, while pharmaco-EEG at times might look like an old and thus somewhat “outdated” method, it should be considered as an extremely valuable biomarker in drug development that serves its purpose more than ever, nicely complementing more modern techniques like brain imaging.

Financial disclosures/funding: None were reported.

Keywords: Key-Lock Principle, pharmaco-EEG, drug effect, pharmaco-electroencephalography, electroencephalography, EEG, biomarker, drug development, CNS

Development and Optimization of the P1vital® Oxford Emotional Test Battery (ETB) for Use in Translational Efficacy Studies in Healthy Volunteers and in Patient Trials for Depression

Presenters: Dawson G, Dourish C, Goodwin G, Harmer C, Kingslake J

Affiliations: Dawson G, Dourish C, and Kingslake J are from the Forest Research Institute Harborside Financial Center in Jersey City, New Jersey; Goodwin G and Harmer C are from the University of Oxford Department of Psychiatry in the United Kingdom.

Background: The worldwide market for antidepressants is estimated at $45 billion, but available drugs have limited efficacy, a slow onset of action, and a significant side effect burden. Many compounds have entered clinical development to address these problems but with little success. A particular limitation is that due to their slow onset of action and large placebo effects, antidepressants need to be administered for 6 to 8 weeks in large-scale trials to demonstrate efficacy.

Results: P1vital’s University of Oxford collaborators were the first to demonstrate that antidepressants modulate negative emotional bias (NEB) in both healthy volunteers and patients with depression. They observed significant shifts of bias away from negative and toward positive stimuli, usually within seven days of initiating antidepressant treatment. The effects are specific to antidepressants and occur prior to any measurable mood changes that may be apparent several weeks later. These findings prompted development of the P1vital® Oxford Emotional Test Battery (ETB), a suite of validated tasks specifically designed and optimized to be sensitive to early changes in NEB.

Conclusion: To ensure the successful deployment of the ETB in a wide range of research and clinical trial settings, P1vital has established and validated a robust system that meets the high regulatory and safety standards for clinical studies in both Europe and the United States. The ETB is currently being used in early development studies both in healthy volunteers and in patients by pharmaceutical companies that require an early indication of antidepressant efficacy before investing in the expensive large scale clinical trials.

Financial disclosures/funding: The authors are shareholders in P1vital LTD.

Keywords: ETB, P1vital® Oxford Emotional Test Battery, translational efficacy studies, depression

Profiling Cognitive Dysfunction Across Psychiatric, Developmental, and Degenerative Disorders Using the Cambridge Neuropsychological Test Automated Battery (CANTAB)

Presenters: Hermans L, Rock P, Housden C, Saxby B, Riedel W

Affiliations: Hermans L, Rock P, Housden C, and Riedel W are from the Cambridge Cognition Ltd. in Cambridge, United Kingdom; Rock P is also from the University of Oxford Department of Psychiatry in the United Kingdom; Housden C is also from the University of Cambridge Department of Psychiatry in the United Kingdom; Saxby B is from the Cambridge Cognition LLC in Chicago, Illinois, and the Newcastle University Institute for Ageing and Health in the United Kingdom; Ridel W is also from the Maastricht University Faculty of Psychology and Neuroscience in The Netherlands.

Background: CANTAB is a suite of touchscreen, computerized, cognitive tests that has been used in clinical trials across a range of therapeutics areas. Although the specific test battery composition is customized according to each study hypothesis, there are commonalities in the core cognitive domains of attention, memory, and executive function. We examined data from commonly used CANTAB tests to compare the magnitude of deficits between schizophrenia, attention deficit hyperactivity disorder (ADHD), and Alzheimer’s disease (AD) in patients entering clinical trials.

Methods: Pooled baseline data were extracted from our clinical trials database. Fourteen studies met the criteria, contributing n=2,746 cases for the analyses: 554 schizophrenia patients; 985 AD; 783 ADHD; and 424 healthy control subjects. Cohen’s d effect sizes (ES) reflecting the performance of patients relative to healthy controls were calculated for each CANTAB test measuring sustained attention, processing speed, visual memory, episodic memory, and executive function.

Results: Distinct cognitive profiles were evident across disorders: As expected, AD patients showed greater deficits to episodic and visual memory, whereas patients with schizophrenia showed impairment across all domains. Both schizophrenia and ADHD had larger deficits to executive function (ES=-2.27 and -1.79) than other domains, highlighting the importance of assessing more complex higher-order tasks.

Conclusion: It is feasible to use a common battery of tests to assess the core domains of cognition across patient groups, facilitating a dimensional approach to measuring and remediating cognitive impairments across diagnostic categories.

Financial disclosures/funding: LH, PLR, CH, BKS, and WJR are employees of Cambridge Cognition, provider of the CANTAB computerized cognitive tests to the pharmaceutical industry and academia.

Keywords: Cognition, effect size

Functional Brain Imaging for Differential Diagnosis and Prediction of Treatment Efficacy in Neurological and Psychiatric Disease

Presenters: Lewine J

Affiliations: Lewine J is from the MIND Research Network, Albuquerque, New Mexico.

Background: Functional brain imaging technologies, including functional MRI and magnetoencephalography (MEG), continue to emerge as valuable tools in the differential diagnosis of neurological and psychiatric disorders. In some cases, these methods also provide valid intermediate markers of disease progression and treatment, with data collected at baseline predictive of subsequent treatment response.

Methods: As part of our effort to advance pharmaco-imaging, we have acquired a large database of information from control subjects, including more than 1,000 MRI/fMRI scans and more than 200 MEG/EEG evaluations. Data are also available from a number of clinical populations including subjects with Alzheimer’s dementia, epilepsy, autism, schizophrenia, depression, traumatic brain injury, and multiple sclerosis. Using a variety of data analysis methods, including independent component analysis, it is possible to reveal hidden patterns of biomarkers that correlate with disease state and severity.

Results: In each of these areas, we have identified imaging-based biomarkers associated with disease state and severity. In several conditions, including autism, depression, and traumatic brain injury, we have identified markers that track disease progression and treatment response, with some baseline data predictive of final clinical outcome.

Conclusion: Multimodal imaging methods can provide novel biomarkers and intermediate treatment end-points for clinical trials.

Financial disclosures/funding: None were reported.

Keywords: Functional brain imaging, differential diagnosis, treatment efficacy, neurology, psychiatry, MRI, EEG, MEG, biomarkers

Head To Head Comparison Of 11C-Pib and 18F-AZD4694 (NAV4694) for Beta-Amyloid Imaging in Ageing and Dementia

Presenters: Reininger C, Pejoska S, Mulligan R, Jones G, Chan J, Svensson S, Cselényi Z, Masters C, Villemagne V, Rowe C

Affiliations: Reininger C is from Navidea Biopharmaceuticals in Andover, Massachusetts; Pejoska S, Mulligan R, Jones G, Chan J, Villemagne V, and Rowe C are from the Austin Health Department of Nuclear Medicine and Centre for PET in Heidelberg, Victoria, Australia; Villemagne V and Rowe C are from the University of Melbourne Department of Medicine in Parkville, Victoria, Australia; Svenson S, Cselényi Z, Villemagne V, and Rowe C are from AstraZeneca R&D in Södertälje, Sweden; Masters C and Villemagne V are from The Mental Health Research Institute, University of Melbourne in Parkville, Victoria, Australia.

Background: 11C-PiB (PiB) is the benchmark radiotracer for imaging of beta-amyloid (Ab) plaque in Alzheimer’s disease (AD). Unfortunately, the 20-minute radioactive decay half-life of carbon-11 (11C) limits the use of PiB to centers with an on-site cyclotron and 11C radiochemistry expertise, making the access to PiB PET restricted and with costs prohibitive for routine clinical use. Fluorine-18 labeled Ab tracers, subsequently developed for clinical use, show higher nonspecific white matter binding and, in some cases, lower cortical binding in AD that could lead to less accurate interpretation of scans. Preliminary studies with the “second generation” fluorine-18 labeled amyloid tracer, which has a close structural resemblance to PiB, 18F-NAV4694 showed a robust separation between AD patients and healthy age-matched control (HC) subjects and less white matter binding than reported with other 18F-labelled amyloid tracers. We compared the cortical and white matter binding of a new F-18 labeled Ab tracer, 18F-NAV4694 to PiB in the same subjects.

Methods: Forty-five participants underwent PET imaging with PiB and AZD4694 (25 healthy elderly [HC], 10 mild cognitive impairment [MCI], 7 probable AD, 3 probable frontotemporal dementia). Images were co-registered so that region of interest placement was identical on both scans and standard uptake value ratios (SUVR) using the cerebellar cortex as reference region were calculated between 40- and 70-min post-injection for both tracers.

Results: NAV4694 showed reversible binding kinetics very similar to PiB reaching apparent steady state 50-min post injection. Both radiotracers showed a similar dynamic range of neocortical SUVR (1.1–3.3 and 1.0–3.2 SUVR for PiB and NAV4694, respectively) and identical, low, nonspecific white matter binding with frontal cortex to white matter ratios of 0.7±0.2 and 1.3±0.2 for both radiotracers in HC and AD, respectively. There was an excellent linear correlation between PiB and NAV4694 neocortical SUVR (slope of 0.95, r=0.99, p<0.0001).

Conclusion: Our results demonstrate that NAV4694 18F-NAV4694 displays nearly identical imaging characteristics to 11C-PiB. NAV4694 provides images of similar appearance to PiB without the limitation of the short 11C radioactive decay half-life. The striking similarity with PiB suggests that the results from longitudinal studies that are clarifying the relationship between Ab accumulation and cognitive decline and asserting the value of Ab imaging as a predictor of cognitive decline and progression to clinical Alzheimer’s disease can be directly translated to NAV4694. The high cortical binding in AD and low nonspecific white matter binding also suggests that NAV4694 images may be more easily and reliably read in clinical practice than other F-18 labeled PET tracers for brain amyloid.

Financial disclosures/funding: None were reported.

Keywords: Alzheimer’s disease, amyloid imaging, Ab, positron emission tomography Rating

Scales Regional Baseline Symptom Patterns in Major Depressive Disorder Trials

Presenters: Busner J, Montgomery S

Affiliation: Busner J is from the Penn State College of Medicine Department of Psychiatry and Bracket Global in Wayne, Pennsylvania; Montgomery S is from the Imperial College in London, United Kingdom.

Background: The baseline level of depressive symptoms, as measured by gold standard scales, such as the Montgomery Asberg Depression Rating Scale (MADRS) or the Hamilton Depression Rating Scale (HAMD), plays an important role in allowing efficacy demonstration of antidepressants in clinical trials. Symptoms with low endorsement detract from total baseline scores and prevent demonstration of drug-related improvement. In an attempt to identify items with low endorsement and to learn if this varied by global region, we examined MADRS baseline patient ratings from multiple industry-sponsored MDD studies across multiple regions of the world.

Methods: A total of 2,224 baseline clinician-administered MADRS individual item ratings were examined from patients in multiple industry-sponsored MDD trials across 15 countries. The MADRS consists of 10 items, each rated on a 0–6 scale, with higher numbers reflecting greater severity. All raters had been trained and certified in MADRS conventions and administration prior to seeing patients in the trials. For analysis, the 15 countries were combined into four regions: United States (n=1,624), Asia (n=184), Eastern Europe (n=197), and Western Europe (n=219). Baseline MADRS items were examined by region for overall severity using multivariate analysis of variance (MANOVA), and for proportions of low endorsement (operationalized as scores of 0 or 1) using chi-square (c2). Effect sizes were examined using the eta-squared (h2) and Cramer’s V statistics, respectively.

Results: Mean scores by item and region mean total scores and individual item scores for 9 of the 10 items differed significantly by region (Item 7, “lassitude” was the exception); although statistically significant, the effect sizes (h2) were small for the total score and individual items, with the exception of Item 10 (“suicidal thoughts”). Post-hoc regional comparisons (Tukey-Kramer) for these significant but small effect sized findings (Items 1, 2, 3, 4, 5, 6, 8, 9, and Total) were scattered; for individual items, no clear pattern by region emerged, and the numeric differences in means were less than 1 point for individual items and 3 points for total score. Item 10, “suicidal thoughts,” showed a large effect size of region (h2=0.17): as shown by Tukey-Kramer post-hoc testing, Eastern Europe scores (mean=1.96) were significantly higher, and Western Europe scores significantly lower (mean=0.04) than United States and Asia scores (means=0.89 and 1.0, respectively), which did not significantly differ from each other. Collapsed across region, the proportion of baseline item scores of 0 or 1 was low (approximately 5% or less) for all items except “reduced appetite” (47.3%) and “suicidal thoughts” (71.9%). When examined by region, statistically significant differences were found for Items 4 (“reduced sleep”), 5 (“reduced appetite”), 6 (“concentration difficulties”), 7 (“lassitude”), 9 (“pessimistic thoughts”), and 10 (“suicidal thoughts”) [c2s(3) 16.2, 50.7, 15.3, 11.1, 33.0, 407.0, respectively, p’s from 0.01 to 0.0001]. The effect sizes and numerical differences by region were small for all but Items 10 (“suicidal thoughts”) and 5 (“reduced appetite”). For “suicidal thoughts,” 100 percent of scores in Asia and Western Europe were 0 or 1; in the United States, the number was 71 percent, and in Eastern Europe, in marked contrast to other regions, the number was 21.3 percent. For “reduced appetite,” 51.5 percent of United States scores were 0 or 1, in contrast to Asia (44%), Eastern Europe (31%), and Western Europe (33%).

Conclusion: For most items of the MADRS, regional differences were small, suggesting that patients with similar levels of symptom expression are entering these multinational trials. Overall, symptom ratings of 0 or 1 at baseline were infrequent, adding support to the scale’s utility in allowing demonstration of drug-related improvement. “Reduced appetite” and “suicidal thoughts” were the items with the largest regional differences and the items with the highest proportions of low endorsement. Of interest, these two items were the least frequently endorsed of all MADRS items almost 20 years ago in a United States sample of hospitalized, depressed patients. As actively suicidal patients are disallowed from most MDD protocols (and from all of the protocols included in this analysis), skewed scoring toward the less severe end is expected. “Suicidal thoughts” endorsement was higher in Eastern Europe, suggesting greater severity and some potential for trial improvement on this item. Lower endorsement of “reduced appetite” may reflect the enrollment of less melancholic patients; the symptom was endorsed at lower levels in the United States than in other regions.

Financial disclosures/funding: Dr. Busner is a full-time employee of Bracket; Dr. Montgomery is consultant, speaker and/or advisory board member to multiple pharmaceutical companies (list available upon request).

Keywords: Major depressive disorder, rating scale, MADRS, low endorsement item

Suicidal Behavior Reported During Study Participation: Predictive Relationships with Suicidal Ideation/Behavior Reports Provided at Screening/ Baseline

Presenters: Greist J

Affiliations: Greist J is from the Healthcare Technology Systems, Madison, Wisconsin.

Background: Fortunately, potentially lethal suicidal behavior is infrequent. Unfortunately, it may also be under-recognized and under-reported. The association between prospectively reported suicidal behaviors during study participation with lifetime suicidal ideation/behaviors is a topic of ongoing research and development. Several suicidal ideation and behavior (SIB) assessments are available or under development. This presentation will review electronic Columbia Suicide Severity Rating Scale (eC-SSRS) data obtained from 35 clinical studies, for both psychiatric and nonpsychiatric indications, conducted over a three-year period.

Methods: Lifetime assessments of SIB obtained at screening/baseline visits from 8,837 participants show robust short-term predictive value of risk for prospective suicidal behavior during subsequent study participation. The most severe lifetime ideation, as well as the type and number of prior lifetime suicidal behaviors, significantly increased the likelihood of prospectively reporting suicidal behavior during study participation. Relative to participants reporting no prior suicidal ideation, odds ratios (ORs) for reporting suicidal behavior during study participation range from 5.0 to 18.7 in psychiatric patients and 2.5 to 78.6 in nonpsychiatric patients, depending on the severity of lifetime ideation reported.

Results: Ideational intensity, in terms of frequency, duration, controllability, deterrents, and reasons was also associated with prospective risk. Lifetime suicidal behaviors (preparatory behavior and/or actual, interrupted, or aborted attempts) increased ORs of reporting the same behavior from 7.8 to 18.1, depending on the specific behavior reported; reporting multiple lifetime behaviors had a multiplicative effect, increasing ORs of prospective reports of behavior from 5.1 to 14.6 compared to participants reporting no lifetime behaviors.

Conclusion: Greater self-report candor for reporting SIB was evident in the six-fold greater disclosure of suicidal behaviors assessed by eC-SSRS in comparison with clinician-assessed C-SSRS.

Financial disclosures/funding: None were reported.

Keywords: Suicidal behavior, suicide ideation, rating scale, Columbia Suicide Severity Rating Scale, electronic Columbia Suicide Severity Rating Scale, eC-SSRS, C-SSRS

Validation and Normalization of the Russian Version of the Positive and Negative Syndrome Scale (PANSS-RU) in Schizophrenia: Preliminary Findings

Presenters: Opler M, Ivanova E, Reznik A, Khan A

Affiliations: Opler M is from the New York University School of Medicine in New York, New York; Opler M, Khan A, and Ivanova E are from ProPhase LLC in New York, New York; Ivanova E is also from the Health and Development Foundation in Moscow, Russia; Reznik A is from Moscow Regional Psychiatric Hospital #5 in Moscow, Russia; Khan A is from the Nathan S. Kline Institute for Psychiatric Research in Orangeburg, New York, and the Manhattan Psychiatric Center, Wards Island, New York.

Background: The PANSS is utilized in clinical trials for assessment of psychopathology. In order to determine the status of a patient, comparisons within a general population are necessary. If several groups of patients are to be compared, gender, age, comorbid diagnoses, and other factors affecting clinical status are not identically distributed. Our Phase 1 aim was to establish the initial psychometric properties of the PANSS-Ru. Our Phase 2 aim was to establish normed-reference data for the PANSS-Ru.

Methods: Phase 1 included 40 patients with schizophrenia and other psychotic disorders. Phase 2 included 375 individuals (n=250 patients; n=125 controls). Responses were assessed for internal consistency, stability, reliability, discriminative validity, and construct validity. Fifth percentile norms are presented as step functions. Data were compared to United States norms.

Results: Forty subjects were enrolled for Phase 1. Alpha coefficient of 0.88 verified good internal consistency. Test-retest comparisons verified time stability (r=0.67–0.92). Correlation between subscale and the total scores ranged from 0.76 to 0.86, compared to >0.83 for United States norm studies. Internal consistencies were alpha>0.745. Mean subscale/total score were equivalent to United States population norms within 13 percent. However, there was a difference of greater than 5 norm-based points for mean general psychopathology subscale. Norms had a sensitivity of 85 percent and specificity of 89 percent.

Conclusion: Preliminary results show that different dimensions of symptom presentations in the Russian population may help to improve symptom-specific treatments and will also provide comparison data for a Russian population. Further normative studies are warranted in other populations.

Financial disclosures/funding: None were reported.

Keywords: Positive and Negative Syndrome Scale, validation, normalization, rating scale

Patient Input Toward the Final Development of the Rosenberg-Hassman Mood Scale (RHMS), a Patient-reported Outcome (PRO) for Major Depressive Episodes

Presenters: Rosenberg L, Wesnes K, Hassman H, Krefetz D, George S

Affiliations: Rosenberg L and George S are from he Center for Emotional Fitness in Cherry Hill, New Jersey; Wesnes K is from Bracket in Goring-on-Thames, United Kingdom; Hassman H and Krefetz D are from the CRI Lifetree in Mount Laurel, New Jersey.

Background: The Rosenberg-Hassman Mood Scale (RHMS) is being developed according to the United States Food and Drug Administration (FDA) Guidance for Patient Reported Outcomes (PRO) to measure depression levels for patients with major depressive disorder (MDD). Prior patient input has helped to generate new items among the 228 word/phrase synonyms in the 26 question RHMS and helped to determine the preferred response options for our scale. We are now seeking to obtain more patient input from our 2013.2 beta version of the computerized RHMS in preparation for finalizing the RHMS.

Methods: After signing informed consent, 50 patients diagnosed with either bipolar disorder (most recent episode depressed) or MDD took our computerized RHMS scale 1 to 3 times over a six-week interval. Subsequently we obtained feedback from these patients individually using a prepared questionnaire and in a group setting using an open-ended agenda.
Results: As per the FDA Guidance on PROs, feedback was both open-ended and in response to a questionnaire. The data are currently being accumulated and analyzed to prepare the final version of the RHMS.

Conclusion: Patient input toward the final development of the RHMS according to the FDA Guidance for PROs has been obtained and is being analyzed. A final version of the RHMS, a self-rated depression scale, will be presented.

Financial disclosures/funding: There are no conflicts for any of the authors. This is self-funded research.

Keywords: Depression, rating scale, Rosenberg-Hassman Mood Scale, RHMS

RATER TRAINING/RATER

ASSESSMENT

Review of Remote Videotaped SCID Interviews in Schizophrenia Trial: Lessons Learned

Presenters: Bromley T, Troyano N, Ormont M

Affiliation: All from ePharmaSolutions

Background: An alarming number of clinical trials in the field of central nervous system medicine fail to detect a signal. Increased surveillance of site ratings including reviews of diagnostic interviews has been done to improve data quality and reduce error variance in several large, global trials. Several serious administrative issues were found when reviewing Structured Clinical Interview for DSM-IV (SCID) interviews with highly experience SCID raters.

Methods: Prior to being approved as a rater in the study, raters were asked to submit their educational background as well as their experience with schizophrenia and the SCID or similar type diagnostic interview (e.g., Mini International Neuropsychiatric Interview [MINI]). Only raters with a minimum of a master’s degree with at least two years’ experience with the schizophrenic population and more than 20 SCID interviews in the past two years were approved to rate. Raters were asked to videotape the screening visits. Videotapes of the SCID administration were reviewed for adherence to scale guidelines and quality of the interview.

Results: More than a third of all SCID videos had at least one error, including failure to administer scale in prescribed fashion, failure to have administrative guideline manual open, failure to ask structured interview questions, and failure to explore topics such as previous drug use and other exclusionary behaviors.

Conclusion: Despite prior experience with scales and study population and additional training on SCID prior to administration, serious errors were found in the administration of the instrument.

Financial disclosures/funding: None were reported.

Keywords: Remote video monitoring, improving data quality

What PANSS Items are Hardest to Rate? Observations from Global Studies

Presenters: Daniel D, Dries J

Affiliations: Both authors are from Bracket Global, Wayne, Pennsylvania.

Background: The Positive and Negative Syndrome Scale (PANSS) is a complex 30-item instrument for evaluation of multiple domains of psychopathology in schizophrenia. In order to inform future training, we ranked the 30 PANSS items with respect to the degree of dissonance in scoring between site raters and same-language local experts utilizing data from 10 global schizophrenia trials.

Methods: Prior to study initiation, raters were trained at investigators meetings by highly interactive procedures, including slide presentations, rating of videotaped patient interviews, and, in some cases, interview and rating of live actors trained to portray schizophrenia symptoms. Site PANSS interviews were recorded and uploaded for evaluation by an independent same-language local expert. The external reviewers provided feedback on an ongoing basis to the site and sponsor on diagnostic and scoring accuracy and interview quality. The current results are based on data from ongoing and recently completed studies. Additional data will be reported in the future. The number of site rater versus external rater comparisons in the current analysis ranged from n=2,217 to n=2533 for the 30 PANSS items, respectively.

Results: The five PANSS items with the highest percentage of significant mismatches (?2 anchor points) between the site and external raters were G13 disturbance of volition (14.9%), N6 lack of spontaneity and flow of conversation (13.6%), G15 preoccupation (13.3%), N7 stereotyped thinking (12.5%), and N1 bunted affect (12.3%). The five items with the highest percentage of exact matches were G14 poor impulse control (69.7%), P7 hostility (69.3%), G3 guilt feelings (67.5%), P3 hallucinatory behavior (66.4%), and G8 uncooperativeness (64.7%).

Conclusion: The five PANSS items with the highest discordance are scored entirely based on objective observations of behavior and/or speech patterns during the interview. In contrast, the scoring of the items with the highest concordance all include data acquired by verbal report from the patient or informant. The results are consistent with the notion that, in the field, inter-rater reliability may be harder to achieve on PANSS items that rely purely on objective observation. If so, inter-rater reliability on such items might be enhanced by focused training. These results are preliminary and additional data is under collection.

Financial disclosures/funding: Statistical analyses were funded by Bracket Global, LLC.

Keywords: PANSS, schizophrenia, inter-rater reliability, rater training

STUDY PROTOCOL/TRIAL

METHODOLOGY

Analysis of Family History Subgroups on Drug Placebo Separation and Placebo Response in a Positive Pivotal MDD Study

Presenters: DeBonis D, Sachs G

Affiliations: All authors are from the Bracket Global, Wayne, Pennsylvania.

Background: Family history is an important predictor and risk factor for depression and may influence placebo response. Levomilnacipran ER (LVM; 1S, 2R-milnacipran) is a potent and selective serotonin and norepinephrine reuptake inhibitor approved in July 2013 for treatment of MDD in adults; NCT00969709 was an eight-week randomized (1:1), placebo-controlled, double-blind study (n=429 received treatment and had MADRS after baseline).

Methods: The primary outcome measure was site-based clinician ratings (MADRSSBR). A computer-administered diagnostic assessment for depression was also used to collect information on family history and MADRS (MADRSCOMP). Drug-placebo separation was analyzed for three groups: 1) FHx++: First- or second-degree relative diagnosed and treated for a mood disorder (n=243); 2) FHx+: Group 1 and first- or second-degree relative with an AXIS I disorder (n = 300); and 3) FHx-: No (n=82) or unknown (n=42) family history.

Results: The treatment effect favored LVM, with a between-group difference of 3.1 (P=0.005) for MADRSSBR (primary outcome) and 2.8 (P=0.014) for MADRSCOMP overall, and ~3.0 to 3.8 points (P<0.05) in the FHx++ and FHx+ groups. In the FHx- group, the between-group difference was ~1.1 points in favor of placebo (not significant). Results were similar for both MADRSSBR and MADRSCOMP.

Conclusion: The results from this large, multicenter, United States study support previous findings that family history may be an important factor in placebo response and treatment effect. Further analysis may clarify which correlates of the FHx(-) patient status (e.g., diagnostic validity) may be associated with high placebo response.

Financial disclosures/funding: DeBonis D and Sachs G are employees of Bracket. Forest Laboratories, the study sponsor, has reviewed this abstract.

Keywords: Family history, drug placebo separation, placebo response, MDD, major depressive disorder, drug response

Are Large Numbers of Investigative Sites Associated with Symptom Improvement on Placebo in Antipsychotic Randomized, Controlled Trials? A Meta-analytic Review

Presenters: Litman R, Szymialis S, Khan A

Affiliations: Litman R and Szymialis S are from CBH Health, LLC, in Rockville, Maryland; Khan A is from the Northwest Clinical Research Centerin Bellevue, Washington.

Background: Improvement on placebo treatment in recent atypical antipsychotic randomized, controlled trials (RCTs) in schizophrenia has been attributed to excessive variability in study design factors, including number and nature of investigative sites. We hypothesized that studies utilizing large numbers of sites would lead to greater improvement on placebo treatment.

Methods: We abstracted data on site number and response to placebo from recently published antipsychotic trials for schizophrenia by searching MEDLINE, Clinicaltrials.gov, and the Cochrane Database. Only trials utilizing the PANSS were analyzed. Response to placebo treatment was defined as the change in PANSS ÷ PANSS baseline score for the placebo arm, and was correlated (Pearson’s R, 2-tailed) with the number of investigative sites for each RCT.

Results: We found 56 RCTs for atypical antipsychotics in schizophrenia. Thirty were included for analysis and 26 were rejected for not meeting criteria. Improvement on placebo increased with larger numbers of sites (r=-0.38, p<0.05), and the number of sites utilized per RCT increased over time (r=0.62, p<0.01).

Conclusion: These data support an association between the recent increase in placebo response and the increase in numbers of sites in antipsychotic RCTs. Further research regarding variability due to site and other study design elements is warranted.

Financial disclosures/funding: None were reported.

Keywords: Placebo response, antipsychotic, schizophrenia, trial design, trial methodology

Pediatric Psychopharmacology Clinical Trials: Development, Issues and Strategy

Presenters: Lowy M, Douglas C, Travers J, Sorrentino M, Burch D

Affiliations: All Authors are employees of PPD, Wayne, Pennsylvania.

Background: This abstract reviews the development and implementation of clinical trials in pediatric psychopharmacology over the past 20 years with a focus on specific childhood diagnoses, such as attention deficit hyperactivity disorder, depression, anxiety, schizophrenia, and bipolar disorder studies for pediatric exclusivity. In the future, focus is likely to shift to other disorders occurring in children and adolescents, such as autism and rare diseases.

Methods: Standard literature approaches (PubMed, regulatory agency’s webpage) were used to identify key publications and position papers related to pediatric psychopharmacology trials.
Results: Current investigations of compounds for the treatment of pediatric disorders in children and adolescent patients require careful attention to diagnosis, patient selection, and selected symptoms, which may be ameliorated by medication. Clinical trial design is crucially dependent on choosing the appropriate scales with the necessary psychometric properties for measuring the anticipated drug effect. Operational procedures pertinent to these studies, including informed consent, pediatric data safety monitoring boards, and administration of rating scales to children and parents, need to be established and validated. Specific examples of drugs/diseases that have been studied will be provided.

Conclusion: A strategy for global regulatory approval needs to be developed for any new pediatric indication, which requires discussion with the regulatory authorities and the development of support in the academic community. Success factors in obtaining six-month patent extensions for agreed upon clinical trial data in children will also be required.

Financial disclosures/funding: All authors are employees of PPD.

Keywords: Pediatric psychopharmacology, pediatric clinical trials.

Investigation of Participant Perceptions of Recording Interviews During Clinical Trials: Surveillance Findings from Russian Sites

Presenters: Opler M, Ivanova E, Reznik A, Khan A

Affiliations: Opler M is from the New York University School of Medicine in New York, New York; Opler M, Khan A, and Ivanova E, are from ProPhase LLC in New York, New York; Ivanova E is also from the Health and Development Foundation in Moscow, Russia; Reznik A is from Moscow Regional Psychiatric Hospital #5 in Moscow, Russia; Khan A is from the Nathan S. Kline Institute for Psychiatric Research in Orangeburg, New York, and the Manhattan Psychiatric Center, Wards Island, New York.

Background: Many schizophrenia trials now include video, audio, or external rater surveillance measures; however, little research is done examining participant’s perspective. Recorded interviews and scoring may influence the study outcomes and how patients and raters perform during the interview. Our aim is to determine if patients with schizophrenia have different perceptions on being recorded during psychiatric clinical interviews compared to subjects without schizophrenia (controls).

Methods: Fifty participants in the Russian Federation were enrolled (schizophrenia=16, controls=34). An audio recording system was used for the Positive and Negative Syndrome Scale (PANSS). Participants were asked to complete a survey on their perceptions and experience of being recorded.

Results: In the control group, 83.35 percent indicated they were willing to be audio recorded, and 93.75 percent of the patients with schizophrenia indicated such. When asked if they would complete a video recorded interview, 44.12 percent of the controls and 25.00 percent of the patients with schizophrenia indicated they would (Fisher’s=61.234 (p<0.001). Only 38.24 percent of the controls and 25.00 percent of the patients with schizophrenia indicated they were comfortable being recorded (Fisher’s=14.238 (p=0.658). There was a significant difference between controls (47.06%) and patients with schizophrenia (25.00%) in their comfort level to complete the interview via webcam with remote interviewers (Fisher’s=52.236 (p<0.001). When asked if they felt their responses would be different if they were not recorded, 5.88 percent of the controls and 6.25 percent of the patients with schizophrenia answered yes (Fisher’s=10.365 (p<0.875).

Conclusion: It is important that sources of variability are minimized. Although audio recordings encourage inter-rater reliability, it is important to examine the patient perceptions of the audio recording process and whether these perceptions influence patient response and rapport with the interviewers.

Financial disclosures/funding: None were reported.

Keywords: Positive And Negative Syndrome Scale, participant perceptions, audio recording

Centrally Monitoring In-study Ratings for an Alzheimer’s Disease Trial Conducted in Asia

Presenters: Shen Q, Shen S, Bertzos K, Gaur R, Harrison J, Shen J

Affiliations: Shen Q and Shen S are from Pfizer Inc., New York, New York; Bertzos K and GaurR are from inVentiv Health Clinical, Princeton, New Jersey; Harrison J is from Metis Cognition Ltd. and Imperial College Department of Medicine, London, United Kingdom; Shen J is from HengRui, Shanghai City, China.

Background: One challenge conducting Alzheimer’s disease (AD) clinical trials in Asian countries is maintaining quality control of assessment scale data. The current analyses implemented a central monitoring methodology in an AD trial conducted in Asia with the aim to timely identify and rectify in-study rating errors on the Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-Cog), Neuropsychological Test Battery (NTB), and Disability Assessment for Dementia (DAD).

Methods: Three-hundred and two subjects across 24 sites in China, Hong Kong, Korea, Singapore, and Taiwan participated in a nonintervention, longitudinal, validation study with subjects with mild-moderate AD. ADAS-Cog, NTB, and DAD source documents were collected from the first three consecutive subject visits and centrally reviewed to identify whether site raters (76) administered and scored the scales accurately. When rating errors were identified, raters were provided feedback/retraining.

Results: We identified 2,443 rating errors and unusual rating patterns across the following scales: NTB (65%), ADAS-Cog (30%), and DAD (5%). More than half (52%) of the errors were recording errors (inadequate documentation), 28 percent were scoring errors, 15 percent were test administration errors, three percent were unusual change scores, and two percent were protocol deviations. Over time and following intervention, there was a 30-percent decrease in rating errors.

Conclusion: By in-study monitoring of data, rating errors were identified and raters received feedback, which resulted in decreased errors over time. Greater attention to use of in-study central monitoring programs and enhanced rater training may assist in managing potential data-related issues in AD trials conducted in Asia and lead to collection of reliable study data.

Financial disclosures/funding: This study was sponsored by Pfizer Inc. and Janssen Alzheimer Immunotherapy Research and Development, LLC. No conflicts of interest were reported.

Keywords: Alzheimer’s disease, rating scales, data monitoring

Using a Subject Registry to Create a Duplicate-Free Corridor for Conducting Clinical Trials

Presenters: Shiovitz T, Wilcox C, Gevorgyan L, Mehra V, Mangano T

Affiliations: Shiovitz T and Mangano T are from the CTSdatabase, LLC in Beverly Hills, California; Shiovitz T, Gevorgyan L, and Mangano T are from the Sherman Oaks California Neuroscience Research in Sherman Oaks, California; Wilcox C is from the Encino Pharmacology Research Institute in Newport Beach, Los Alamitos; Mehra V is from the Artemis Institute for Clinical Research in San Diego, California.

Background: Duplicate subjects are an increasingly recognized problem in central nervous system (CNS) studies. Within a single pharmaceutical company, duplicates entered into studies may represent up to five percent of subjects. When prescreening subjects are tracked across pharmaceutical companies, duplicate rates are significantly higher. These duplicate subjects may increase placebo response, may not take study medication, and almost certainly contribute to failed studies.

Methods: Twenty-six Southern California CNS sites joined together to obtain IRB approval of a Subject Database Authorization and enter partial identifiers of subjects who presented for clinical trial prescreening into the CTSdatabase subject registry between October 2011 and October 2013. Sites were notified immediately when Virtually certain matches (<10–7 likelihood of matching by chance) or probable matches ( Results: Two hundred and fifty virtually certain matches representing 500 potential duplicate subjects were found among 5,100 prescreens entered (13%). Four hundred matches, which occurred at the same site, were excluded from the data set. The number and percent of matched subjects increased substantially as the number of participating sites increased. Bringing together competing local sites for a common cause meant overcoming resistance and mistrust among several sites and required prompt access to and cooperation from the matching sites when duplicates were found. Many duplicate subjects changed their personal identifiers and/or traveled to sites that were 25, 50, or even 100 miles apart.

Conclusion: While use of a subject registry is best integrated into study protocols to ensure site participation and minimize the need for sites to communicate with each other in the case of matches, Southern California has made great strides in providing a duplicate-free corridor for conducting CNS studies. This model may be applied to other metropolitan and suburban areas with large numbers of duplicate subjects, with the goal of markedly reducing the number of inappropriate subjects entering studies, allowing more CNS studies to succeed.

Financial disclosures/funding: None were reported.

Keywords: Subject registry, dual enrollment, clinical trials, CNS trials, drug trials, patient enrollment, patient screening, data monitoring, data quality control

Impact of Implementing a National Research Subject Database to Prevent Dual Enrollment in Early and Late Phase CNS Trials

Presenters: Weingard K, Efros M

Affiliations: Weingard K is from Verified Clinical Trials, LLC in Garden City, New York; Efros M is from Verified Clinical Trials, LLC and AccuMed Research Associates in Garden City, New York.

Background: Enrollment in central nervous system (CNS) clinical trials can present considerable challenges to research sites, pharmaceutical companies, and contract research organizations (CROs). Medical inaccuracies, failure to admit to simultaneous participation in more than one clinical trial, and jumping from one trial to another without allowing sufficient time to lapse between treatments compromise the health of the subjects, data quality, and the outcome of the trial. Dual enrollment is a serious problem that can be costly to the research site, pharmaceutical company, and/or manufacturer and harmful to patients.

Results: Using three different criteria, VCT has performed 69,150 verifications in 2012. A significant amount of subjects were found to be problematic. Four percent of the subjects attempted dual enrollment across all phases of clinical research and all disease states. A much higher incidence was seen in early phase trials where stipends were higher and in certain disease states, such as healthy volunteers, CNS, psych, pain, and other subjective conditions. Seven percent attempted enrollment during their lockout period across all phases of clinical research with similar statistics across early and late phase trials. Five percent attempted to screen while actively screening at another clinical trial center. While this practice is not prohibited in some early phase trials, by tracking the dual screening activity, VCT was able to alert those CNS research centers—allowing them to be proactive with qualified alternates to meet their enrollment and dosing numbers. In most late-phase trials within the VCT system, dual screening is prohibited.

Financial disclosures/funding: None were reported.

Keywords: Dual enrollment, clinical trials, CNS trials, drug trials, patient enrollment, patient screening

High Fidelity: What is the Real Impact of a Data Monitoring Program on Data Quality?

Presenters: Yavorsky C, DiClemente G, Wolanski W, Burger F

Affiliations: All authors are from Cronos CCS, Lambertville, New Jersey.

Background: Risk-based data-monitoring is widely used in clinical trials to help manage risk implicit in subjectively derived outcome measures. Few studies to date have critically examined the impact of such programs. In this study, our intention was to determine how well the method works and then to estimate the potential impact of nonintervention on statistical power.

Methods: The present study investigated a sample of subjects enrolled in a completedschizophrenia trial using the PANSS as the primary outcome measure. The data was processed daily and, if risks to data quality were detected, contact was initiated. Feedback was provided as necessary when problems in scale use were identified. The sensitivity and specificity were computed to determine the proportion of data correctly identified as problematic. A forward analysis estimating the impact of nonintervention with raters contributing poor quality data was conducted.

Results: The first analysis allowed for summarization of the overall efficacy of the method while identifying those raters whose error would have adversely impacted trial outcomes was used to re-estimate sample size. We theorized a nonintervention scenario with patients continuing to be assessed incorrectly and subtracting these. Using this estimated reduction in sample size we recalculated power based on original parameters and found a reduction from 0.90 to 0.78.

Conclusion: Risk-based data-monitoring can detect error within reasonable estimates and be addressed in-study. This analysis aimed to outline what the error rate in a typical study as well as potential impact of reduced sample size if we assume data contributed by problematic raters was allowed to persist.

Financial disclosures/funding: All authors are employees of Cronos CCS and report no additional conflicts of interest.

Keywords: Clinical trials, data monitoring, data quality, CNS trials