BLOCK HF Study Biventricular versus Right Ventricular Pacing in Patients with Left Ventricular Dysfunction and Atrioventricular Block – Preliminary Results Anne B. Curtis, Buffalo, NY Seth J. Worley, Lancaster, PA Philip B. Adamson, Oklahoma City, OK Eugene S. Chung, Cincinnati, Ohio Imran Niazi, Milwaukee, WI Lou Sherfesee, Minneapolis, MN Timothy S. Shinn, Ann Arbor, MI Martin St. John Sutton, Philadelphia, PA On behalf of the BLOCK HF Trial Investigators and Coordinators

Acknowledgments BLOCK HF Steering Committee Curtis AB (Principal Investigator), Adamson PB , Chung ES, St. John Sutton MG, Worley SJ Echo Core Lab St. John Sutton MG, Plappert T Adverse Events Advisory Committee Boehmer JP, Meyer TE(Chair), Smith AL, De Lurgio DB Data Monitoring Committee Steinberg JS (Chair), DeMarco T, Elkayam U, Louis TA (Statistician) Investigators Canada: Rinne C, Thibault B United States: Adamson PB, Al-Sheikh T, Andriulli J, Barber MJ, Beau S, Bell M, Borgatta L, Brodine W, Canosa R, Chung ES, Compton S, Curtis AB, Ellison K, Evonich R, Faddis M, Foreman B, Murray C, Guerrero M, Herre J, Hodgkin D, Huang D, Keim S, Kocovic D, Kusmirek SL, Lessmeier T, Levanovich P, Lobban JH, Mackall JA, Manaris A, McBride W, McKenzie J, Mela T, Merliss A, Mitrani R, Mittal S, Mounsey P, Navone A, Niazi I, Obel O, Oren J, Patel P, Patel V, Pickett A, Rao A, Rist K, Rosenblum A, Saba S, Sakaguchi S, Sandler D, Sangrigoli R, Shinn TS, Simmons T, Simonson J, Smith JE, Telfer EA, Tobias S, Tomassoni G, Worley SJ Sponsor Medtronic Inc. Clinical Trials.gov Identifier: NCT00267098 Caution: Use of CRT devices for AV block and systolic dysfunction patients without ventricular dyssynchrony is not an approved use in the United States. You may want to lead into the presentation to your fellow investigators with a summary of the August 2012 DMC meeting and what has led everyone to this meeting. BLOCK HF

Background Clinical Importance Over 1 million people world-wide and 819,000 people in the US have atrioventricular (AV) block Currently treated with standard pacemaker (i.e. right ventricular (RV pacing)) therapy Approximately 6 million in the US are currently diagnosed with heart failure (HF) and approximately 670,000 new cases confirmed each year According to AHA 2012 statistics, this costs the US approximately $20 to $56 billion annually DAVID and MOST Trial results have shown that RV pacing may have long-term deleterious effects Can biventricular (BiV) pacing prevent progression of heart failure and its clinical and economic consequences in AV block? BLOCK HF

Study Design BLOCK HF AV block necessitating pacing ELIGIBILITY CRITERIA AV block necessitating pacing Left ventricular ejection fraction (LVEF) < 50% NYHA functional class I, II or III Absence of a Class I indication for resynchronization therapy No previous pacemaker or implantable cardioverter defibrillator (ICD) Echocardiography performed at Randomization, 6, 12, 18 and 24 months Implant (CRT-P/D) Establish OMT (30-60 days) Randomize 1:1 Control: RV pacing Treatment: BiV pacing You might want to comment that a brief summary of why a Bayesian approach was chosen and how to interpret the data via Bayesian methods will be described in an upcoming slide. Double-Blind OMT=optimal medical therapy CRT-P=cardiac resynchronization therapy pacemaker CRT-D=CRT defibrillator Follow-up Every 3 months Follow-up Every 3 months BLOCK HF

Study Purpose and Objectives Purpose: Biventricular pacing is superior to RV apical pacing in patients with AV block and LVEF <50% who require ventricular pacing Endpoints: Primary: Composite of: All-cause mortality, HF-related urgent care, defined as HF hospitalization requiring IV therapy, or Any unplanned visit requiring intravenous HF therapy, and Increase in left ventricular end systolic volume index (LVESVI) >15% Key Secondary: All-cause mortality, All-cause mortality/HF hospitalization, HF hospitalization 1. Describe the difference between the primary HF morbidity objective and the secondary HF hospitalization in that for the primary, the patient must have presented with HF symptoms in which IV diuretics or in which was classified as meeting the primary objective even if HF wasn’t the primary reason for the hospitalization, but later developed HF and given IV diuretics while being hospitalized. The secondary objective hospitalization was met if the patient was admitted for HF and may or may not have received IV therapies BLOCK HF

Study Success Criteria Methods Study featured: Adaptive sample size Pre-specified interim analyses Stopping rules for success, safety, futility Analysis: Intention-to-treat Bayesian survival analysis using time to first event Parameter of interest for each endpoint: BiV to RV hazard ratio (HR) Metric: Probability that HR < 1 BLOCK HF utilized a Bayesian statistical design which is an adaptive type of design in terms of sample size, which allows only enough subjects to be randomized in the trial to conclude enrollments and possibly reach a successful study outcome. Had the design not been adaptive in nature, a pre-specified sample size may have produced a negative or non-significant result if too small, or if too large, would have taken longer to enroll. (Keep in mind, that it took almost 8 years to enroll 918 patients.) The metric for passing the objective in Bayesian statistics does not make use of P-values. Instead the probability that BiV is superior to RV (represented by a Hazard Ratio less than 1) is determined through a combination of pre-trial assumptions and accumulated trial data. For this to be concluded, the probability that the hazard ratio is less than 1 must exceed the pre-specified threshold seen in the table. So for the primary objective, the probability of the hazard ratio being less than 1 must be at least 0.9775 and for the secondary objectives the probability must be at least 0.95. Endpoint Study Success Criteria Primary Endpoint (Mortality, HF Urgent Care, LVESVI) Probability of (HR <1) > 0.9775 Secondary Endpoints All-cause Mortality All-cause Mortality/HF Hospitalization HF Hospitalization Probability of (HR <1) > 0.95 BLOCK HF

918 Assessed for eligibility Study Flow Diagram Enrollment 227 Subjects not randomized: 95 Subjects for which AV conduction testing criteria not met prior to implant 14 Subject withdrawals prior to implant 51 Unsuccessful implants 67 Implanted subjects not randomized 918 Assessed for eligibility 691 Randomized 1:1 Allocation 349 Allocated to Biventricular Pacing 346 Received allocated intervention 3 Did not receive allocated intervention 342 Allocated to Right Ventricular Pacing 342 Received allocated intervention Follow-up 52 Exited/lost to follow-up 75 Deaths 13 Crossed over to Right Ventricular Pacing 3 Met primary endpoint prior to crossover 50 Exited/lost to follow-up 90 Deaths 84 Crossed over to Biventricular Pacing 50 Met primary endpoint prior to crossover Analysis Implanted subject not randomized (n =67) Implanted subject exited prior to randomization (n=21) Implanted subject died prior to randomization (n=16) Programmed to BiV pacing prior to randomization (n = 10) Randomization visit missed (n = 10) LV lead complication (n = 6) Subject did not meet inclusion/exclusion criteria (n = 4) 349 Analyzed 83 Censored for primary endpoint due to missing LVESVI data 342 Analyzed 71 Censored for primary endpoint due to missing LVESVI data BLOCK HF

Follow-Up Experience BLOCK HF BiV (N=349) RV (N=342) Average Follow-up (months) 36.3 ± 23.1 37.9 ± 23.5 Follow-up Compliance (% of visits) 94.6% 93.8% BLOCK HF

Baseline Demographics CRT-P CRT-D BiV (N=243) RV (N=241) BiV (N=106) RV (N=101) % Male 75% 70% 82% 80% Age, years 74 ± 10 74 ± 11 72 ± 9 71 ± 10 LVEF, % 43 ± 7 33 ± 8 Heart Rate, beats/min 69 ± 23 69 ± 24 68 ± 17 69 ± 17 QRS Duration, ms 125 ± 33 125 ± 31 123 ± 30 119 ± 30 NYHA I NYHA II NYHA III 14% 58% 27% 20% 52% 28% 10% 63% 26% 16% 57% Left Bundle Branch Block 35% 31% Ischemic Heart Disease 39% 38% 1st Degree AV Block 2nd Degree AV Block 3rd Degree AV Block 17% 33% 49% 15% 29% 56% 40% 32% ACE Inhibitor/ARB at Randomization 71% 74% 83% 88% Beta Blocker at Randomization 78% 92% Diuretics at Randomization 64% 66% 72% DMC recommended trial be continued at least through October 15; Data in presentation is data through September 24, 2012 (data cleaned and frozen on Oct 15). The medications are bolded showing consistency across both groups at Randomization versus at Baseline. BLOCK HF

All Randomized Subjects Primary Endpoint Results: Mortality/HF Urgent Care/LVESVI Key Take-Away: There was demonstrated to be at least a 10% relative reduction in risk (95% upper bound for HR is 0.9) and an estimated 26% relative reduction (estimated HR is 0.74) in mortality/HF urgent care risk and significant increase in left ventricular volume in the BiV arm versus the RV arm. This finding holds up in the CRT-P subjects alone as well. In the CRT-D subjects, the HR estimate was comparable to that of the CRT-P subjects, as were the BiV and RV event rates themselves (see next slide). Cohort Estimated HR (95% CI) Probability HR < 1 Threshold All Randomized Subjects 0.74 (0.60, 0.90) 0.9978 0.9775 CRT-P Only CRT-D Only 0.73 (0.58, 0.91) 0.75 (0.57, 1.02) BLOCK HF

Clinical Components of Primary Endpoint: Mortality/HF Urgent Care Visits Graphs show the composite primary objective without the echo endpoint for both types of devices. Even without LVESVi, we passed. These results demonstrate at least a 19% relative reduction in mortality/HF urgent care (upper 95% confidence bound was 0.81), and an estimated 38% relative reduction in risk of events (estimate for HR was 0.62). In both device groups the confidence bounds fell below 1, showing evidence of a significant reduction. Cohort Estimated HR (95% CI) Probability HR < 1 Threshold All Randomized Subjects 0.73 (0.57, 0.92) 0.997 N/A CRT-P Only CRT-D Only 0.73 (0.56, 0.94) 0.73 (0.53, 1.02) BLOCK HF

Secondary Endpoint: Mortality/HF Hospitalization Graph shows secondary endpoint of mortality and the HF hospitalization endpoint in combination. However, the difference between this endpoint and the primary endpoint is that a HF hospitalization is a Cardiovascular-related hospitalization in which the primary reason for admission is a heart failure-related adverse event as previously defined (for the primary endpoint, the hospitalization need not have HF as the primary reason for admission, as long as the patient presents with signs or symptoms consistent with HF and receives IV therapy for HF during the hospitalization). IV therapy was not a requirement for this secondary endpoint. The probability that HR <1 was 0.9802, exceeding the pre-specified threshold of 0.95. Thus, the objective was met and demonstrates a significant relative reduction in risk of mortality/HF hospitalization. The estimated relative reduction was 22% (estimated HR was 0.78). Note: Medtronic can provide a specific example for you. Cohort Estimated HR (95% CI) Probability HR < 1 Threshold All Randomized Subjects 0.78 (0.61, 0.99) 0.9802 0.95 CRT-P Only CRT-D Only 0.77 (0.58, 1.00) 0.80 (0.58, 1.13) BLOCK HF

Secondary Objective Results: HF Hospitalization and Mortality If we have two many slides in the main body, maybe we can consolidate the mortality & HF hospitalization objective results into one slide. For Hospitalization alone: The probability that HR < 1 was 0.9938, exceeding the pre-specified threshold of 0.95. Thus, since there is a greater than 95% chance of the HR being less than 1, the objective is met. Graphs show that HF hospitalizations are reduced in the BiV arm when looking at all subjects and is the driver in the combined mortality/morbidity objective. The estimated relative reductio was 30% (estimated HR was 0.7), and the HR estimates were comparable between the two device groups (0.69 vs. 0.73). For Mortality alone: The probability that the HR <1 was found to be 0.8588, which is below the pre-specified cutoff of 0.95. Thus, since there was not a 95% or greater chance of the HR being less than 1, the objective was not met. Possible Rationale for not passing Mortality alone: Whether this was impacted by the imbalance in crossovers (predominately from RV to BiV) between the arms is not known. The curves do not separate until after 48 months in the CRT-P group, while in the CRT-D group there was some very modest separation earlier. Cohort HF Hospitalization Mortality Threshold Estimated HR (95% CI) Probability HR < 1 All Randomized Subjects 0.70 (0.52, 0.93) 0.9922 0.83 (0.61, 1.14) 0.8588 0.95 BLOCK HF

Number of Adverse Events (N, % of Subjects) Note: Table below includes pre-randomization AE’s Number of Adverse Events (N, % of Subjects) CRT-P CRT-D BiV (N=243) RV (N=241) BiV (N=106) RV (N=101) Procedure-related Complications 50 (41, 17%) 32 (25, 10%) 18 (18, 17%) 21 (16, 16%) System-related Complications* Generator-related LV lead-related 41 (36, 15%) 10 (10, 4%) 16(15, 6%) 37 (31, 13%) 11 (11, 5%) 13 (13, 5%) 48 (40, 38%) 32 (32, 30%) 6 (6, 6%) 34 (24, 24%) 13 (12, 12%) 10 (9, 9%) Many of the generator-related complications were changeouts (54 of the 66 generator-related compls), and most of these 66 comps occurred at least 2.5 years after the initial implant attempt. 20 of the 23 generator-related complications for subjects randomized to RV occurred post-randomization, and 19 of those 20 occurred more than 1 year post-randomization. 40 of the 42 generator-related complications for subjects randomized to BiV occurred post-randomization, and 39 of those 40 occurred more than 11 months post-randomization. 14 of the 23 LV lead-related comps among subjects randomized to RV occurred prior to randomization. 8 of the 22 LV lead-related comps among subjects randomized to BiV occurred prior to randomization. * Subcategories for other system-related complications such as RA or RV-related complications not included BLOCK HF

Strengths and Limitations Prospective, randomized, double-blind control design Largest, longest follow-up trial to date First to show difference in outcomes in AV block and LV systolic dysfunction patients with BiV vs. RV pacing LIMItations: Long enrollment duration Censoring due to missing LVESVI in primary objective Crossover imbalance between arms: 24.6% crossed over from RV to BiV 4.6% crossed over from BiV to RV BLOCK HF

Conclusions In patients with AV block and LV systolic dysfunction (LVEF < 50%), BiV pacing compared to RV pacing leads to a significant 26% reduction in the combined endpoint of mortality, heart-failure related urgent care, and increase in LVESVI. Furthermore, there is a 27% relative risk reduction in the composite endpoint of heart-failure urgent care and all-cause mortality BLOCK HF

Back-up Slides

Post-randomization Cross-overs Device Group RV to BiV BiV to RV CRT-P 60 (24.9%) 9 (3.7%) CRT-D 24 (23.8%) 7 (6.6%) Total 84 (24.6%) 16 (4.6%) * Percentages reflect percentage of subjects randomized subjects BLOCK HF