Columbia University Medical Center The Cardiovascular Research Foundation State of the Art Review: Carotid Stenting, Patient Selection, and Clinical Trial Results State of the Art Review: Carotid Stenting, Patient Selection, and Clinical Trial Results William A. Gray MD Director of Endovascular Services Associate Professor of Clinical Medicine

Disclosures Consultant Consultant  Abbott Vascular  Cordis/J&J  BSC  Medtronic  Contego  Silk Road Consultant Consultant  Abbott Vascular  Cordis/J&J  BSC  Medtronic  Contego  Silk Road

Overview Large scale CEA data Large scale CEA data  Landmark  Recent validation of high risk population and differential outcomes What has the evolution in CAS data been? What has the evolution in CAS data been? What predicts stroke in CAS? What predicts stroke in CAS? What are the data for outcomes as regards: What are the data for outcomes as regards:  Experience  Devices Large scale CEA data Large scale CEA data  Landmark  Recent validation of high risk population and differential outcomes What has the evolution in CAS data been? What has the evolution in CAS data been? What predicts stroke in CAS? What predicts stroke in CAS? What are the data for outcomes as regards: What are the data for outcomes as regards:  Experience  Devices

Year of study % without stroke Surgical-ACAS Medical-ACAS Surgical-NASCET Medical-NASCET Surgical-ACST Medical-ACST 17% 6% ACST crossover Landmark CEA studies ACAS, ACST and NASCET: Comparative Results

NY State Registry Results State of the state of CEA Stroke. 2009;40:

Validation of high-risk CEA subsets. Validation of high-risk CEA subsets.

Conclusions from these recent data Although limited in terms of surveillance for events (retrospective, not neurologist evaluated): Although limited in terms of surveillance for events (retrospective, not neurologist evaluated):  Confirmatory data on the outcomes after CEA, though probably underestimates events  Clearly identifies subgroups at increased risk Including patients >80 yearsIncluding patients >80 years Although limited in terms of surveillance for events (retrospective, not neurologist evaluated): Although limited in terms of surveillance for events (retrospective, not neurologist evaluated):  Confirmatory data on the outcomes after CEA, though probably underestimates events  Clearly identifies subgroups at increased risk Including patients >80 yearsIncluding patients >80 years

Evolution in CAS Outcomes

Incidence of stroke and death in high risk carotid stent IDE trials:

ARMOUR results: 1° Endpoint 30d Results (ITT & Full Population) 30d Results by Symptoms and Age (ITT)

Predictors of Outcomes

VariableSymptomatic Pre-dilatation w/o EPD (n=143) Age ≥ 80 (n=788) Multiple stents per procedure (n=196) CAPTURE 3500: Predictors of CAS Outcomes All Patients (3307) Symptomatic (479) Asymptomatic (2517) >80 (809) <80 (2611) Odds Ratio

CAS Outcomes by age and symptoms CAPTURE 3500

CAPTURE 3500: Stroke by Location 44% 22% 18% of all strokes in CAPTURE are non-ipsilateral More non-ipsilateral strokes were minor c/w ipsilateral

Experience and Outcomes

Experience Preface (assumption: volume equals experience) Preface (assumption: volume equals experience) Trial-based outcome data supporting experience as a determinant of outcomes Trial-based outcome data supporting experience as a determinant of outcomes  Surgical  Pro-CAS  EVA-3S  SPACE  PMS registries Preface (assumption: volume equals experience) Preface (assumption: volume equals experience) Trial-based outcome data supporting experience as a determinant of outcomes Trial-based outcome data supporting experience as a determinant of outcomes  Surgical  Pro-CAS  EVA-3S  SPACE  PMS registries

Retrospective CEA Survey Demonstrates Inverse Relationship between Volume and Neurologic and Mortality Outcome Maryland state 1998 J Vasc Surg. 1998;27:25-33

…More CEA Volume Data Mortality Stroke Length of stay J Vasc Surg. 1998;27:25-33

Pro-CAS: Prospective registry of CAS Prospective, multi-center German registry Prospective, multi-center German registry  25 sites/6 year enrollment (July 1999-June 2005)  5341 interventions (outcomes of first 3267 published 2004) Median # of center CAS before enrollment: 38 (0-1200)Median # of center CAS before enrollment: 38 (0-1200) Median # of center CAS SPACE enrollment: 140 (10-806)Median # of center CAS SPACE enrollment: 140 (10-806) No learning curve data availableNo learning curve data available  No defined inclusions or exclusions, or procedural methods  No angiographic core lab  Voluntary independent neurologic assessment (no stroke scales or CEC)  Primary endpoint: in-hospital death and stroke Prospective, multi-center German registry Prospective, multi-center German registry  25 sites/6 year enrollment (July 1999-June 2005)  5341 interventions (outcomes of first 3267 published 2004) Median # of center CAS before enrollment: 38 (0-1200)Median # of center CAS before enrollment: 38 (0-1200) Median # of center CAS SPACE enrollment: 140 (10-806)Median # of center CAS SPACE enrollment: 140 (10-806) No learning curve data availableNo learning curve data available  No defined inclusions or exclusions, or procedural methods  No angiographic core lab  Voluntary independent neurologic assessment (no stroke scales or CEC)  Primary endpoint: in-hospital death and stroke Stroke 2008;39; ;

Pro-CAS: Overall outcomes Median age: 70 years (32-96) Median age: 70 years (32-96) In-hospital stoke and death: 3.6% In-hospital stoke and death: 3.6% Median age: 70 years (32-96) Median age: 70 years (32-96) In-hospital stoke and death: 3.6% In-hospital stoke and death: 3.6%

Pro-CAS: Effect of experience* *Univariate analysis

Pro-CAS: effect of age and gender* *Univariate analysis

Pro-CAS: effect of symptom status* *Univariate analysis

Pro-CAS: Multivariate analysis of predictors

The trouble with (European) randomized CAS data Trial analysis in CAS using ICSS as a primer Trial analysis in CAS using ICSS as a primer

Columbia University Medical Center The Cardiovascular Research Foundation

Columbia University Medical Center The Cardiovascular Research Foundation

Columbia University Medical Center The Cardiovascular Research Foundation

Columbia University Medical Center The Cardiovascular Research Foundation

Columbia University Medical Center The Cardiovascular Research Foundation

Columbia University Medical Center The Cardiovascular Research Foundation

EVA-3S Randomized CEA vs CAS RR 2.5 (95% CI, ) P= Mas JL, et al. New Engl J Med. 2006;355:

EVA-3S: Trial design Randomized, multicenter, non-inferiority trial of standard surgical risk patients with symptomatic carotid stenosis >60% Randomized, multicenter, non-inferiority trial of standard surgical risk patients with symptomatic carotid stenosis >60%  Primary endpoint: 30-day death and stroke  Secondary endpoint: 30-day MAE plus ipsilateral stroke to 4 years 30 centers in France 30 centers in France 527 patients 527 patients Randomized, multicenter, non-inferiority trial of standard surgical risk patients with symptomatic carotid stenosis >60% Randomized, multicenter, non-inferiority trial of standard surgical risk patients with symptomatic carotid stenosis >60%  Primary endpoint: 30-day death and stroke  Secondary endpoint: 30-day MAE plus ipsilateral stroke to 4 years 30 centers in France 30 centers in France 527 patients 527 patients Lancet Neurol 2008; 7: 885–92

EVA-3S critique Slow enrollment resulted in limited investigator experience Slow enrollment resulted in limited investigator experience  1.7 CAS patients/year/site Early and/or non-standard technique resulted in unnecessary morbidity Early and/or non-standard technique resulted in unnecessary morbidity  Use of EPD not widespread or familiar Lack of use in the early phase of the trial likely responsible for 4-5 excess strokes (~20% of all strokes in the CAS arm)Lack of use in the early phase of the trial likely responsible for 4-5 excess strokes (~20% of all strokes in the CAS arm)  5% stent procedure failure requiring emergency surgery in this trial resulting in 2 strokes in the CAS group –Major pivotal trials in this country (e.g., SAPPHIRE, ARCHeR) have not reported any emergent surgical conversions  No pre-dilation in >80% of procedures (standard in US)  Significant (beyond local) anesthesia was employed in ~30% of procedures (estimated <5% in US) Slow enrollment resulted in limited investigator experience Slow enrollment resulted in limited investigator experience  1.7 CAS patients/year/site Early and/or non-standard technique resulted in unnecessary morbidity Early and/or non-standard technique resulted in unnecessary morbidity  Use of EPD not widespread or familiar Lack of use in the early phase of the trial likely responsible for 4-5 excess strokes (~20% of all strokes in the CAS arm)Lack of use in the early phase of the trial likely responsible for 4-5 excess strokes (~20% of all strokes in the CAS arm)  5% stent procedure failure requiring emergency surgery in this trial resulting in 2 strokes in the CAS group –Major pivotal trials in this country (e.g., SAPPHIRE, ARCHeR) have not reported any emergent surgical conversions  No pre-dilation in >80% of procedures (standard in US)  Significant (beyond local) anesthesia was employed in ~30% of procedures (estimated <5% in US)

EVA-3S critique Limited investigator experience and number of trained sites/operators Limited investigator experience and number of trained sites/operators  Experienced operators defined by 12 lifetime CAS procedures or 5 CAS procedure if 35 supra-aortic procedure These operators were deemed experienced and allowed to tutor the non-experiencedThese operators were deemed experienced and allowed to tutor the non-experienced  No centralized training qualification process (local proctors pronounced the operators qualified)  Approximately 2/3 of sites were under tutelage at the beginning of their randomized participation. Limited investigator experience and number of trained sites/operators Limited investigator experience and number of trained sites/operators  Experienced operators defined by 12 lifetime CAS procedures or 5 CAS procedure if 35 supra-aortic procedure These operators were deemed experienced and allowed to tutor the non-experiencedThese operators were deemed experienced and allowed to tutor the non-experienced  No centralized training qualification process (local proctors pronounced the operators qualified)  Approximately 2/3 of sites were under tutelage at the beginning of their randomized participation.

EVA 3S Prototypical low operator experience multi- center trial Prototypical low operator experience multi- center trial Outcomes for CAS in EVA-3S for symptomatic standard risk patients are higher than the contemporary cohorts Outcomes for CAS in EVA-3S for symptomatic standard risk patients are higher than the contemporary cohorts Prototypical low operator experience multi- center trial Prototypical low operator experience multi- center trial Outcomes for CAS in EVA-3S for symptomatic standard risk patients are higher than the contemporary cohorts Outcomes for CAS in EVA-3S for symptomatic standard risk patients are higher than the contemporary cohorts

SPACE: an incomplete study Randomized, multicenter non-inferiority study of CEA vs. CAS in standard surgical risk symptomatic patients with 70% carotid stenosis Randomized, multicenter non-inferiority study of CEA vs. CAS in standard surgical risk symptomatic patients with 70% carotid stenosis  Primary endpoint 30-day ipsilateral stroke and death  Only 27% EPD use  Stopped before non-inferiority was proven ($$$)  Pre-specified secondary analyses include: AgeAge SexSex Type of eventType of event Side of interventionSide of intervention Degree of stenosisDegree of stenosis High-grade contralateral stenosisHigh-grade contralateral stenosis Randomized, multicenter non-inferiority study of CEA vs. CAS in standard surgical risk symptomatic patients with 70% carotid stenosis Randomized, multicenter non-inferiority study of CEA vs. CAS in standard surgical risk symptomatic patients with 70% carotid stenosis  Primary endpoint 30-day ipsilateral stroke and death  Only 27% EPD use  Stopped before non-inferiority was proven ($$$)  Pre-specified secondary analyses include: AgeAge SexSex Type of eventType of event Side of interventionSide of intervention Degree of stenosisDegree of stenosis High-grade contralateral stenosisHigh-grade contralateral stenosis

SPACE Randomized CEA vs. CAS symptomatic patients SPACE collaborators. Lancet 2006;368: Abs diff: 0.51, 90%CI , P=0.09 (non-inferiority)

SPACE: Effect of volumes on outcome Neuroradiology, Sept 2008

SPACE: sub-group analysis (PP)

SPACE: Hazard ratio for 30-day MAE plus ipsilateral stroke to 2 years

The evolution of CAS in symptomatic patients: EVA-3S, SPACE, ICSS vs. the world AHA guideline limit

Long-term CAS efficacy for stroke prevention

EVA-3S: 4-year outcomes Any ipsilateral strokeAny strokeAny stroke or death

SPACE: K-M plots of 2-year outcomes Ipsilateral stroke and vascular death 30-day stroke/death plus ipsilateral stroke to 2 years

SPACE and Pro-CAS: conclusions Experience determines outcomes in CAS Experience determines outcomes in CAS Age related gradient for outcomes in CAS, favoring the younger patient (even over surgery) Age related gradient for outcomes in CAS, favoring the younger patient (even over surgery) Long-term efficacy of CAS similar to CEA Long-term efficacy of CAS similar to CEA Experience determines outcomes in CAS Experience determines outcomes in CAS Age related gradient for outcomes in CAS, favoring the younger patient (even over surgery) Age related gradient for outcomes in CAS, favoring the younger patient (even over surgery) Long-term efficacy of CAS similar to CEA Long-term efficacy of CAS similar to CEA

Post-market US registries

CAS after device approval: Increasingly safe in the “real world” * Hierarchical Events – Includes only the most serious event for each patient and includes only each patient’s first occurrence of each event C: n=573; E # : n=204 C2 #: n=197C: n= 3574; E: n=1917; C2: n=1788 All stroke/death* #Small symptomatic cohorts preclude any comparisons ARCHeR 8.3%

MAE defined as any death, MI or stroke SAPPHIRE WW: 30 day major adverse events Asymptomatic vs. Symptomatic P= P=NS P=0.0005P= P= % AHA Asx 6% AHA Sx

CAPTURE: Asymptomatic Patients <80 Years No stroke/death at 43% (61/143) of sites 30 day stroke/death distribution by site N=2764 3%

CAPTURE 2: Asymptomatic <80 Patients N= day stroke/death distribution by site No stroke or death in 81% (134/166) of sites

Outcome Improvements with Increasing Experience in CAS % of sites with 0% 30-day MAE

The Influence of Experience: PMS Outcomes *Hierarchical events – Includes only the most serious event for each patient and includes only each patient’s first occurrence of each event. Level I CAPTURE: n=210 EXACT: n=267 CAPTURE 2: n=83 Level 2 CAPTURE: n=1879 EXACT: n=776 CAPTURE 2: n=1026 Level 3 CAPTURE: n=735 EXACT: n=482 CAPTURE 2: n=318 Asymptomatic patients <80 years old

The Influence of Experience: PMS Outcomes * Hierarchical events – Includes only the most serious event for each patient and includes only each patient’s first occurrence of each event. Level I CAPTURE: n=52 EXACT: n=34 CAPTURE 2: n=18 Level 2 CAPTURE: n= 282 EXACT: n=87 CAPTURE 2: n=105 Level 3 CAPTURE: n=93 EXACT: n=57 CAPTURE 2: n=32 Symptomatic patients <80 years old

EXACT and CAPTURE 2 30-day Composite Endpoint of Death and Stroke Hierarchical- Includes only the most serious event for each patient and includes only each patient first occurrence of each event. * EXACT (N=2145)* CAPTURE 2 (N=4175) Combined (N=6320) EXACT (N=213) CAPTURE 2 (N=548) Combined (N=761) EXACT (N=1931) CAPTURE 2 (N=3627) Combined (N=5558)

EXACT/CAPTURE 2 (combined): 30-day major adverse events symptomatic patients <80 years Hierarchical- Includes only the most serious event for each patient and includes only each patient first occurrence of each event. Symptomatic patients N= Death/StrokeDeath/Major StrokeDeath Stroke Minor (3.1%) Stroke Major (1.4%) (%) Subjects 6%AHA guideline

Hierarchical- Includes only the most serious event for each patient and includes only each patient first occurrence of each event. N= Death/StrokeDeath/Major StrokeDeath Stroke Minor (1.8%) Stroke Major (0.6%) (%) Subjects 3%AHA guideline EXACT/CAPTURE 2 (combined): 30-day major adverse events asymptomatic patients <80 years

Devices and Outcomes

CAS outcomes: No differences by device 30 day Outcomes Hierarchical Events – Includes only the most serious event for each patient and includes only each patient’s first occurrence of each event Clinical Studies are not directly comparable by methodology presented. -Data from respective studies are presented for educational purposes EXACT: n=1454 CAPTURE 2: n=1372 EXACT: n=463 CAPTURE 2: n=416 <80 years >80 years 3%

Future Data Sets CREST: February 2010 CREST: February 2010 ACT I: Lead-in phase ACT I: Lead-in phase CREST: February 2010 CREST: February 2010 ACT I: Lead-in phase ACT I: Lead-in phase Event30 days, N=118 Death, Stroke, and MI*1.7% All Stroke and Death*1.7% Major Stroke and Death*0.0% Death0.0% All Stroke1.7% Major Stroke0.0% Minor Stroke1.7% MI0.0%

Conclusions A relationship of experience to outcomes in CAS appears to be present based on indirect evidence: A relationship of experience to outcomes in CAS appears to be present based on indirect evidence:  Generally improving outcomes for the field  Operator disparity …and direct evidence …and direct evidence  EVA-3S vs. “the world”  IDE and PMS registries No convincing evidence of device influence No convincing evidence of device influence Future trials will deliver more prospective, and likely confirmatory, data Future trials will deliver more prospective, and likely confirmatory, data A relationship of experience to outcomes in CAS appears to be present based on indirect evidence: A relationship of experience to outcomes in CAS appears to be present based on indirect evidence:  Generally improving outcomes for the field  Operator disparity …and direct evidence …and direct evidence  EVA-3S vs. “the world”  IDE and PMS registries No convincing evidence of device influence No convincing evidence of device influence Future trials will deliver more prospective, and likely confirmatory, data Future trials will deliver more prospective, and likely confirmatory, data