APPENDIX 1 Neurologic Academic Research Consortium Participants
ONLINE APPENDIX
APPENDIX 1 – Neurologic Academic Research Consortium Participants
Academic Research Organization Representatives:
Donald E. Cutlip, MD, Harvard Clinical Research Institute; Beth Israel Deaconess Medical Center, Boston, MA, US
Mitchell W. Krucoff, MD, Duke Clinical Research Institute, Durham, NC, US
Alexandra J. Lansky, MD, Yale Cardiovascular Research Group; Yale School of Medicine, New Haven, CT, US
Roxana Mehran, MD, Cardiovascular Research Foundation; Mount Sinai Medical Center, New York, NY, US
Gerrit-Anne van Es, PhD, Cardialysis, Rotterdam, NL
Physician Specialty Experts:
Interventional and Structural Cardiology
Andreas Baumbach, MD, cDepartment of Cardiology, St Bartholomew's Hospital and William Harvey Research Institute and Queen Mary University of London, London, United Kingdom;
John K. Forrest, MD, Yale School of Medicine, New Haven, CT, US
David Holmes, MD, Mayo Clinic, Rochester, MN, US
Samir Kapadia, MD, Cleveland Clinic, Cleveland, OH, US
Susheel Kodali, MD, Columbia University Medical Center / New York-Presbyterian Hospital, New York, NY, US
Axel Linke, MD, University of Leipzig, Leipzig, DE
Vivian G. Ng, MD, Yale Cardiovascular Research Group; Yale School of Medicine, New Haven, CT, US
Cody G. Pietras, Yale Cardiovascular Research Group; Yale School of Medicine, New Haven, CT, US
Jeffrey Popma, MD, Beth Israel Deaconess Medical Center, Boston MA, US
Bernard Prendergast, MD, St. Thomas’ Hospital, London, UK
Eugene McFadden, MD, Cork University Hospital, Cork, IE
Jeffrey Moses, MD, Columbia University Medical Center / New York-Presbyterian Hospital, New York, NY, US
Joachim Schofer, MD, Albertine Heart Center, Hamburg, DE
Szilard Voros, MD, Global Institute for Research and Global Genomics Group, Richmond, VA, US
Cardiac Surgery
Arie P. Kappetein, MD, PhD, Erasmus University Medical Center, Rotterdam, NL
Michael Mack, MD, The Heart Hospital Baylor Plano Research Center, Plano TX, US
Electrophysiology
Joseph Akar, MD, Yale University School of Medicine, New Haven, CT, US
Neurology and Neuroradiology
Kevin J. Abrams, MD, Baptist Hospital of Miami, Miami, FL, US
Michel Bilello, MD, PhD, Hospital of the University of Pennsylvania, Philadelphia, PA, US
Michael Dwyer, PhD, University of Buffalo, Buffalo, NY
Karen L. Furie, MD, MPH, Rhode Island Hospital, Providence, RI, US
David Greer, MD, Yale School of Medicine, New Haven, CT, US
Daryl Gress, MD, University of Nebraska Medical Center, Omaha, NE, US
Steven R. Messé, MD, Hospital of the University of Pennsylvania, Philadelphia, PA, US
Claudia Scala Moy, PhD, National Institute of Neurological Disorders and Stroke, Bethesda, MD
Nils H. Petersen, MD, PhD, Yale School of Medicine, New Haven, CT, US
H. Bart van der Worp, MD, University Medical Center, Utrecht, NL
Robert Zivadinov, MD, PhD, University at Buffalo, State University of New York
Neuropsychology
Adam M. Brickman, PhD, College of Physicians & Surgeons, Columbia University, New York, NY, US
Jeffrey N. Browndyke, PhD, Duke University Medical Center, Durham, NC, US
Ronald M. Lazar, PhD, Columbia University Medical Center / New York-Presbyterian Hospital, New York, NY, US
Biostatistics
Michael Parides, PhD, Icahn School of Medicine at Mount Sinai Group, New York, NY, US
Pathology
Renu Virmani, MD, CVPath Institute, Gaithersburg, MD, US
U.S. Food and Drug Administration:
Andrew Farb, MD, Medical Officer, Division of Cardiovascular Devices, Center for Devices and Radiological Health
John Laschinger, MD, Medical Officer, Division of Cardiovascular Devices, Center for Devices and Radiological Health
Bram Zuckerman, MD, Director, FDA Division of Cardiovascular Devices
Carlos Peña, PhD, MS, Director, FDA Division of Neurological and Physical Medicine Devices
Industry Representatives:
Boston Scientific Corporation: Kenneth M. Stein, MD
Cardiogard Medical Ltd.: Doug Post
CardioMed Device Consultants: Semih Oktay, PhD
Claret Medical, Inc.: Tony Fields, MS, MSc; James Hallums; Azin Parhizgar, PhD; Peyton Willert
Edwards Lifesciences: Faouzi Kallel, PhD
Emboline, Inc.: Amir Belson, MD; Scott Russell
Innovative Cardiovascular Solutions: R. Kevin Plemmons
Keystone Heart Ltd.: Hank Hauser; Pauliina Margolis, MD, PhD; Shuki Porath
Medtronic: Sidney A. Cohen, MD, PhD; Manish Gupta, MS, MBA; Manuela Negoita, MD; Marc Turco, MD
NeuroSave, Inc.: Thomas Kreck, MD
St. Jude Medical, Inc.: Eric Manasse, MD, MBA; Angie Roach, MSc
Transverse Medical, Inc.: Eric Goslau
W.L. Gore & Associates: Coby Larsen, PhD
Appendix II
Post-procedural Neurologic Assessment and Stroke Severity Determination
In general board-certified neurologists will be most sensitive and accurate in detecting neurologic dysfunction and diagnosing stroke, and are recommended to screen all patients in efficacy endpoint trials (Category II and III). For safety trials (Category I), a credentialed non-neurologist physician or non-physician performing a standardized assessment should have adequate sensitivity to detect changes from the baseline examination, which should prompt confirmation by an independent neurologist (1). All neurologic events in clinical trials require independent adjudication.
Serial assessment of neurologic change using established instruments such as the National Institutes of Health Stroke Scale (NIHSS), or of delirium using the Confusion Assessment Methods (CAM or CAM-ICU), will add consistency to study results both within and across trials. The NIHSS is a validated quasi-ordinal measure of neurologic dysfunction and stroke severity, with standards for training and proficiency in its administration (https://learn.heart.org/nihss.aspx). The exam is explicit, detailed, and focused, evaluating cranial nerve functions, proximal strength, sensation, cerebellar function, vision, and cortical functions (including language and neglect). The NIHSS is limited in that it does not assess weakness of the distal upper and lower extremities, including hand weakness, and does not identify nystagmus, vertical gaze abnormalities, or truncal ataxia. Severity measured early after stroke onset using the NIHSS is one of the strongest predictors of long-term outcome (2,3) and is reliable across neurologists, non-neurologist physicians, nurses and research coordinators (4,5). There are no data evaluating the utility of the NIHSS as a screening tool to identify acute stroke across a trial population and research on this approach is needed; however, it is a reasonable application that has been used in numerous trials. (6-9) For highest reliability, the same individual should perform all NIHSS evaluations of a given trial subject when possible;if performed by a non-neurologist and a change is noted, a neurologist should assess the patient to confirm any deficits and determine whether a stroke has occurred.
The Confusion Assessment Methods (CAM) (10) is a systematic and minimally burdensome assessment technique for identifying core features of delirium with high sensitivity and specificity, (10) and with training can be reliably administered by physician or non-physician personnel. Versions of the CAM have been developed for assessment of intubated [CAM-ICU(11)] and non-acute patients [3D-CAM(12)]. Identification of delirium should trigger neuroimaging and evaluation by a neurologist. In summary, well-trained certified non-neurologists and neurologists that are free of conflict, using established neurologic and delirium instruments will enhance the quality of unbiased clinical information, particularly for more subtle clinical changes.
Long-term Stroke Ascertainment and Disability Determination
Protocol-driven long-term stroke and disability assessment (at 1 year and ideally longer) should be contextualized for procedures in which stroke is primarily a safety concern (Category I), and for devices intended to prevent procedural or long-term stroke (Categories II and III). The modified Rankin Scale (mRS) (13) is an ordinal scale designed to measure post-stroke disability and ranges from 0 (normal without symptoms or disability) to 6 (dead). An mRS score of 0 or 1 is consistent with a return to all usual activities (non-disabling). Because stroke symptoms generally improve over time, it is helpful to identify patients having “stroke with complete recovery” i.e., without residual deficits at follow up (an mRS score of 0). For patients with baseline disability, a new “stroke with complete recovery” would represent a return of the patient to his or her baseline mRS and the absence of any ongoing symptoms related to the new stroke.
Additional disability and QoL scales may also be used, including the Barthel Index, a semi-quantitative assessment of instrumental activities of daily living, and the NeuroQol or EQ-5D, which are standardized and validated questionnaires frequently used in stroke trials (14,15). Health related quality of life (HRQoL) measurements may include not only disease-specific measures, but also generic instruments such as the EQ-5D, Medical Outcomes Study Short-Form 36 (SF-36), or the Short-Form 12 (SF-12) (16). Whereas stroke-specific measures maybe more sensitive and easier to interpret, generic measures offer comparability across different diseases and countries. Moreover, the EQ-5D provides a single index value of health, making it a suitable measure for health economic analyses. Differing mortality rates among treatments can complicate HRQoL interpretation, and analysis methodologies should be carefully considered, including complementing analysis of the entire cohort with additional analysis in only the surviving patients.
Neurologic Event Adjudication for Clinical Trials
The complexity of neurologic endpoint assessment requires central adjudication by a Clinical Events Committee (CEC) for consistency. A particular challenge for cardiovascular trials is the expertise necessary for evaluation of stroke signs, symptoms, disability, and imaging in the diagnosis and classification of stroke and CNS infarction. We therefore recommend the involvement of a stroke neurologist in all phases of clinical trial planning and execution, including event adjudication by the CEC, with neuroradiology expertise available when required. Neuropsychological data should be interpreted by a central core laboratory, and the results provided to the CEC. In randomized trials, all assessments (neurologic, imaging, and cognitive) should be performed by persons unaware of treatment assignment whenever possible.
Magnetic Resonance Imaging for the Detection and Quantification of CNS Infarction
MRI is the imaging modality of choice for detection and quantification of brain ischemia related to cardiovascular procedures (8,17,18). While cerebral computed tomography (CT) has high specificity, it has relatively poor sensitivity for detecting brain ischemia, particularly in the acute peri-procedural phase of injury (0.39/1.0 for CT vs 0.99/0.92 for MRI) (19), and is likely to miss the numerous but often small lesions common after cardiovascular interventions. Therefore, MRI is recommended as the default imaging modality for cardiovascular clinical trials, even if a CT was obtained. Use of 3.0 Tesla MRI scanners is recommended given their high sensitivity in detecting small lesions that may be missed on 1.5 T systems. Within an individual trial, every effort should be made to ensure consistency with respect to MRI field strength to avoid bias resulting from variability of lesion detection and quantitative measurements. Deviations should be limited to explicit contraindications (e.g., new pacemaker), in which case scan parameters should be matched as closely as possible.
Pre-Procedure and Late Follow-up MRI Assessments in Clinical Trials
A baseline DWI scan can increase specificity by identifying pre-existing acute lesions, and sensitivity by facilitating subtraction imaging. However, the benefits of a baseline image must be weighed on a trial-by-trial basis against practical concerns, including additional cost and increased drop out from subsequent MRI assessments due to patients’ reluctance to undergo repeat examinations. The prevalence of acute lesions detected on baseline studies can be low (20); however, a baseline MRI may better isolate true procedure-related insults and should be considered in populations with a significant background risk of embolism, or when an invasive diagnostic procedure is required prior to index intervention. Ideally in procedural trials, a time delay of at least 10-14 days should be imposed between any invasive diagnostic examination and the primary post-procedure MRI assessment. These recommendations do not preclude immediate brain imaging with MRI in symptomatic patients.
Systematic follow-up MRI can be considered to assess the ongoing new ischemic burden in a study population, and to evaluate the conversion of acute DWI lesions to chronic lesions (acknowledging the caveats above). If follow-up MRI is intended, it should be performed approximately 30 days (or longer depending on the trial goals) after the index procedure (21), and include DWI (to detect ongoing acute ischemic burden) and T2-FLAIR and T2 sequences (to identify chronic lesion burden and assess evolution of the acute lesions). For comparative purposes, a follow-up MRI should be paired with a baseline or post-procedural study to avoid inclusion of pre-existing age-related T2 lesions.
MRI Outcomes and Reporting
Advantages of MRI endpoints in efficacy (Category II and III) trials include: 1) DWI lesions accurately identify and quantify ischemic neural tissue; 2) the temporal specificity of DWI can reliably attribute CNS infarction to a particular procedure or device; and 3) DWI sensitivity to CNS infarction allows clinical trials to be sufficiently powered with achievable sample sizes. The clinical and cognitive impact of DWI lesions have yet to be fully elucidated; therefore, a primary measure cannot be chosen based on prognostic value. However, as the best measure of extent of CNS injury, total acute DWI ischemic lesion volume is recommended as a primary imaging endpoint in clinical trials designed to assess procedural neuroprotection. Other recommended measures are summarized in Table 4. MRI measures should also be reported by vascular territory and correction for covariates such as total brain volume should be considered, particularly when investigating correlations with cognitive and other outcomes. While the selected DWI measures primarily apply to post-procedure assessment, similar outcomes can be assessed at later time points using T2-visible lesions.
Limitations of MRI
The benefits of MRI must be weighed against numerous practical considerations. MRI is costly and often difficult to schedule in an acute clinical setting. In particular, the use of multiple MRI time points (baseline, acute and potentially late) poses substantial logistical challenges. Pacemakers and other implantable devices may not be MR-compatible (or may limit the field strength to 1.5 T), and some patients may not tolerate investigation in the post-operative setting; either factor can cause dropout or out-of-window acquisition that may bias MRI results, presenting a challenge in clinical trial interpretation, where methods and utility of data imputation are unclear. Therefore every effort should be made to minimize loss to post-procedure MRI follow-up (17,18).
Role of Transcranial Doppler in Cardiovascular Clinical Trials
Use of transcranial Doppler ultrasound (TCD) in prospective clinical trials can provide mechanistic insight into the cause, timing and location of cerebral microembolization. TCD allows real-time detection of cerebral microembolic signals (MES) and is the only technology available to provide data on embolic risk during various procedural steps; as such, it can help identify the likely source of procedural embolism. One important application of this technology might be early in the clinical evaluation of a device in order to measure embolic load and allow comparison of device designs and technical approaches.