1. Peri-infarct Zone Pacing to Prevent Adverse Left Ventricular Remodeling in Patients with Large Myocardial Infarction Gregg W. Stone, MD Eugene S. Chung, Branislav Stancak, Jesper H. Svendsen, Trent M. Fischer, Fred Kueffer, Thomas Ryan, Jeroen Bax, and Angel Leon, for the Post-Myocardial Infarction Remodeling Prevention Therapy (PRomPT) Trial Investigators Results from the PRomPT Trial

2. Presenter Disclosures None

3. PRomPT: Background Patients with large MI have high rates of death and hospitalization for heart failure (HF), which markedly affects quality of life (QoL) and increases healthcare costs Patients with large MI have high rates of death and hospitalization for heart failure (HF), which markedly affects quality of life (QoL) and increases healthcare costs Delayed contractile activation in the infarct and peri- infarct regions may increase local wall stress and workload, resulting in infarct expansion, wall thinning and LV remodeling Delayed contractile activation in the infarct and peri- infarct regions may increase local wall stress and workload, resulting in infarct expansion, wall thinning and LV remodeling LV pacing in the peri-infarct zone reduces local stroke work, and may attenuate post-MI increase in LV end- diastolic volume (LVEDV), preventing ventricular remodeling and improving prognosis LV pacing in the peri-infarct zone reduces local stroke work, and may attenuate post-MI increase in LV end- diastolic volume (LVEDV), preventing ventricular remodeling and improving prognosis

4. PRomPT: Design Up to 250 patients with large first MI (CPK >3000 U/L) and QRS duration 3000 U/L) and QRS duration <120 msec within 10 days of symptom onset *Protecta XT or Consulta, with LV leads implanted in the peri-infarct zone with 2D echo guidance CRT-D* LV and RV pacing CRT-D* LV pacing only Control No implant Randomized 1:1:1 Blinded Follow-up visits and evaluations 2D echo, NYHA, MLWHF, EQ-5D, 6MWT at 1, 3, 6, 12, and 18 months Baseline 2D echo, NYHA, MLWHF, EQ-5D

5. Target (green) and acceptable (yellow) lead implantation sites Anterior infarcts Inferior and lateral infarcts A lateral (circumflex) infarct generally comprises segments 5, 6, 11, 12, and 16, while an inferior (right coronary artery) infarct typically comprises segments 3, 4, 9, 10, and 15. A LV lead in this location will be in the peri-infarct zone of either infarct. Anterior (i.e. left anterior descending) infarcts generally comprise segments 1, 2, 7, 8, 13, 14, and 17. The target location in green is in the lateral peri-infarct region, segment 12, with other acceptable regions adjacent (yellow).

6. PRomPT: Primary endpoint Change in LVEDV in the echocardiographic apical 4-chamber view from baseline to 18 months between the pooled pacing therapy groups and the control group (ΔLVEDV) Change in LVEDV in the echocardiographic apical 4-chamber view from baseline to 18 months between the pooled pacing therapy groups and the control group (ΔLVEDV) Primary analysis is performed in the as-treated cohort, including all randomized subjects, but requiring successful CRT-D implantation in the device arms Primary analysis is performed in the as-treated cohort, including all randomized subjects, but requiring successful CRT-D implantation in the device arms Analyzed by analysis of covariance (ANCOVA) with baseline LVEDV as a covariate Analyzed by analysis of covariance (ANCOVA) with baseline LVEDV as a covariate

7. PRomPT: Power analysis / sample size Assumptions Assumptions  Baseline mean LVEDV 130 mL in each group  15% greater increase in mean LVEDV from baseline to 18 months in control patients compared to the pooled pacing therapy group (e.g. 19.5 mL vs. 0 mL)  Standard deviation of ΔLVEDV from baseline to follow- up of 35 mL in the randomized groups → 156 evaluable echocardiograms would provide ~90% power at a 2-sided alpha of 0.05 to demonstrate superiority of peri-infarct pacing → 156 evaluable echocardiograms would provide ~90% power at a 2-sided alpha of 0.05 to demonstrate superiority of peri-infarct pacing

8. PRomPT: Organization / quality control Principal Investigator: Gregg W. Stone, Columbia University Principal Investigator: Gregg W. Stone, Columbia University Co-principal Investigator: Angel Leon, Emory University Co-principal Investigator: Angel Leon, Emory University Steering Committee: Eugene S. Chung, Inder S. Anand, Jeroen J. Bax, Michael R. Gold, Robert Gorman, MD, Heinz Theres, James Udelson, Mark Pfeffer Steering Committee: Eugene S. Chung, Inder S. Anand, Jeroen J. Bax, Michael R. Gold, Robert Gorman, MD, Heinz Theres, James Udelson, Mark Pfeffer Clinical Events Adjudication Committee: John Herre (Chair), Michael Dickinson, Ayesha Hasan, Mark Kremers Clinical Events Adjudication Committee: John Herre (Chair), Michael Dickinson, Ayesha Hasan, Mark Kremers Echocardiography Core Laboratory: Thomas Ryan, Ohio State University Echocardiography Core Laboratory: Thomas Ryan, Ohio State University Implant Lead Adjudication Committee: Jose Dizon, Jagmeet Singh, Oussama Wazni Implant Lead Adjudication Committee: Jose Dizon, Jagmeet Singh, Oussama Wazni Data Safety and Monitoring Board: Dwight W. Reynolds (Chair), Juan Aranda, Tim Church Data Safety and Monitoring Board: Dwight W. Reynolds (Chair), Juan Aranda, Tim Church Trial and Site Management, Data Monitoring, Management and Analysis and Sponsor: Medtronic, plc Trial and Site Management, Data Monitoring, Management and Analysis and Sponsor: Medtronic, plc

9. PRomPT: Enrollment Given difficult recruitment and the goal of the present trial to inform a pivotal trial design, it was determined during enrollment, while the study leadership was still blinded, to accept 80% power and an alpha of 0.10, requiring 96 evaluable echocardiograms Given difficult recruitment and the goal of the present trial to inform a pivotal trial design, it was determined during enrollment, while the study leadership was still blinded, to accept 80% power and an alpha of 0.10, requiring 96 evaluable echocardiograms Assuming 20% of subjects would not have data available at the 18-month visit due to early study exit, death, or non-evaluable echocardiograms, a sample size of 120 as-treated patients was required (40 per group) Assuming 20% of subjects would not have data available at the 18-month visit due to early study exit, death, or non-evaluable echocardiograms, a sample size of 120 as-treated patients was required (40 per group) 9

10. 10 Between December 2010 and October 2013, 126 patients were randomized at 27 sites in Europe, the Middle East, and the United States PRomPT: Enrollment LV and RV pacing (n=41) LV pacing only (n=40) No implant (n=45) Randomized 1:1:1 18-month FU (n=38) As-treated (n=37) ITT (n=41) 18-month FU (n=36) As-treated (n=38) ITT (n=40) 18-month FU (n=36) As-treated (n=45) ITT (n=45) Successful implant (n=37) Successful implant (n=38) 1 withdrew 1 withdrawn 1 withdrew 1 withdrawn 5 withdrew 3 lost to FU 1 missed 18-mo FU 1 withdrew 2 lost to FU

11. PRomPT: Top 5 Enrolling Sites 1. East-Slovak Institute of Cardiovascular Diseases, Kosice, Slovakia (20): B. Stancak, J. Ignac, M. Jankajova, E. Komanova 2. The Lindner Research Center, Cincinnati, OH, USA (17): E. Chung, T. Waller, S. Menon, M. Gupta, K. Bailey, E. Schloss, D. Kereiakes, D. Wahl 3. Rigshospitalet, Copenhagen, Denmark (11): J. Svendsen, J. Bro-Jeppesen 4. Emory University Hospital Midtown, Atlanta, GA, USA (9): A. Leon, M. El-Chami, M. Lloyd, M. Hoskins, J. Langberg, D. DeLurgio 5. University of Pennsylvania, Philadelphia, PA, USA (7): W. Matthai, W. Groh, R. Li, J. Stern, J. Bullinga, G. Chang, C. Frankil, C. Gasperetti, C. Leng, A. Moak

12. PRomPT: Baseline features Pooled pacing (n=75) Control (n=45) Age (yrs)*59 ± 11 54 ± 11 Gender (male)73.3% Body mass index (kg/m 2 )29 ± 629 ± 4 Hypertension56.0%55.6% Hyperlipidemia40.0%46.7% Diabetes mellitus20.0%22.2% Current smoking45.3%42.2% MI location Anterior76.0%77.8% Non-anterior24.0%22.2% PCI performed96.0%97.8% Symptom onset to PCI (hrs)7.6 ± 8.29.0 ± 12.7 Peak CPK (U/L)4513 ± 13165248 ± 2832 *P=0.02; otherwise there were no significant differences between groups

13. PRomPT: Implant performance Pooled pacing groups* Devices were successfully implanted in 75/76 attempts (98.7%) at 7.0 ± 2.3 days after MI Devices were successfully implanted in 75/76 attempts (98.7%) at 7.0 ± 2.3 days after MI The LV lead was successfully placed within the echocardiographic peri-infarct zone in 95.5% of patients, and in the pre-specified target zone in 62.7% of patients The LV lead was successfully placed within the echocardiographic peri-infarct zone in 95.5% of patients, and in the pre-specified target zone in 62.7% of patients The median (25%, 75%) percent LV pacing during follow-up was 98.9% (97.9%, 99.0%) The median (25%, 75%) percent LV pacing during follow-up was 98.9% (97.9%, 99.0%) 13 *There were no significant differences between single and dual lead groups

14. PRomPT: Primary endpoint - ΔLVEDV Paired echocardiographic results between the baseline and 18-month follow-up visits LVEDV (mL) nBaseline18-month Adj mean change (95% CI) Difference (95%CI) P Control34 118.6 ± 29.4 133.9 ± 38.9 15.8 (5.5, 26.2) 0.6 (-12.3, 13.5) 0.92 Pooled pacing 64 106.1 ± 27.6 122.8 ± 40.4 16.4 (8.9, 23.9)

15. PRomPT: Primary Endpoint – ΔLVEDV Paired echocardiographic results between the baseline and 18-month follow-up visits 100 Mean ∆LVEDV (mL) 0 Control N with data: 75 50 25 0 -25 -50 1361218 31 33 34 Months after randomization 29 25 31 27 28 33 27 34 32 33 34 37 44 Single-site Dual-site Control Single SiteDual Site

16. PRomPT: ΔLVEDV - Subgroup analysis PRomPT: ΔLVEDV - Subgroup analysis ΔLVEDV was not significantly different in pts in whom the LV lead was vs. was not implanted in the pre-specified target location (adjusted mean diff (95%CI) = -4.3 (-20.4, 11.8) mL, P=0.60) ΔLVEDV (Intervention – Control) Mean (95% CI) ΔLVEDV (Intervention – Control) Interaction P value Variable # Patients ALL patients980.6 (-12.3, 13.5) 0.87 Age <median (58 years) ≤median (58 years) Intervention betterControl better -40-35-30-25-20-15-10-50510152025303540 48 50 3.3 (-14.1, 20.7) 1.2 (-18, 20.4) 0.42 Gender Female Male 27 71 8.2 (-15.7, 32) -3.2 (-18.6, 12.2) 0.73 Hypertension No Yes 42 56 -2.2 (-21.2, 16.8) 2.3 (-15.2, 19.8) 0.23 Hyperlipidemia No Yes 59 39 6.2 (-10, 22.3) -9.6 (-30.1, 10.9) 0.19 Diabetes No Yes 81 17 -3.6 (-17.6, 10.4) 20.7 (-13.2, 54.7) 0.34 Infarct Location Anterior Non-anterior 78 20 3.6 (-10.6, 17.8) -12.1 (-41.4, 17.3) 0.54 Peak CPK <median (4447 U/L) ≥median (4447 U/L) 44 45 0.5 (-19.1, 20.1) 8.7 (-9, 26.4) 0.42 Baseline LVEDV <median (105.8 mL) ≥median (105.8 mL) 49 6.5 (-12.9, 25.9) -4 (-21.5, 13.4) 0.09 Baseline LVEF <median (43.1%) ≥median (43.1%) 49 9.9 (-8.2, 27.9) -11.5 (-28.6, 5.5)

17. PRomPT: Secondary echo endpoints Paired echocardiographic results between the baseline and 18-month follow-up visits nBaseline18-month Adj mean change DifferenceP LVESV (mL)(95% CI) Control34 67.7 ± 24.9 81.5 ± 32.5 14.0 (5.4, 22.7) -2.8 (-13.5, 7.9) 0.61 Pooled pacing 64 61.8 ± 21.2 73.1 ± 33.1 11.2 (5.0, 17.5) LVEF (%) Control34 43.8 ± 9.1 40.5 ± 10.1 -3.0 (-5.8, -0.2) 2.3 (-1.2, 5.7) 0.20 Pooled pacing 64 42.5 ± 8.9 41.9 ± 10.0 -0.8 (-2.8, 1.3)

18. PRomPT: Secondary endpoint – 6MWT Paired 6 minute walk test results between the 1-month and 18-month follow-up visits 6MWT (meters) nBaseline18-month Adj mean change (95% CI) Difference (95%CI) P Control28 399.6 ± 155.0 410.3 ± 151.8 15.6 (-32.1, 63.2) 22.0 (-35.3, 79.3) 0.45 Pooled pacing 63 384.2 ± 134.8 424.0 ± 145.6 37.6 (5.8, 69.3)

19. PRomPT: Secondary endpoints - QoL Paired Minnesota Living with Heart Failure and EQ-5D results between the baseline and 18-month follow-up visits nBaseline18-month Adj mean change DifferenceP MLWHF score(95% CI) Control32 29.6 ± 29.7 26.8 ± 25.4 -0.1 (-8.5, 8.3) 0.5 (-9.6, 10.7) 0.92 Pooled pacing 69 24.6 ± 24.5 26.2 ± 24.2 0.4 (-5.3, 6.1) EQ-5D score Control33 0.8 ± 0.2 0.1 (0.0, 0.1) -0.0 (-0.1, 0.1) 0.99 Pooled pacing 69 0.8 ± 0.2 0.1 (0.0, 0.1)

20. PRomPT: Secondary endpoints - NYHA Paired NYHA classification between the baseline and 18- month follow-up visits NYHA class Control groupPooled pacing groups Proportion P=0.67

21. PRomPT: Death or HF Hospitalization Pooled Pacing vs. Control HR 0.79 (0.34, 1.86) P = 0.59 Pooled Pacing (n=75) Control (n=45) 82.6% Number at Risk: 75 45 Pooled Pacing Control 1.0 Freedom from Death or HF Hospitalization 0.8 0.6 0.4 0.2 0 67 38 50 66 35 100 Days after Randomization 63 34 150 62 32 200 61 32 250 61 32 300 61 31 350 61 31 400 60 30 450 59 30 500 58 25 Month 18 0.0 78.3%

22. PRomPT: Adverse events during FU Pooled pacing (n=75) Control (n=45) HR (95% CI) P value Death 3 (4.0%) 2 (4.7%) 0.83 0.83 (0.14, 4.98)0.84 - Cardiac death - Cardiac death 3 (4.0%) 1 (2.4%) 1.64 1.64 (0.17, 15.82)0.66 - Sudden death - Sudden death 1 (1.4%) 1 (2.4%) 0.56 0.56 (0.03, 9.03)0.68 - Non-cardiac death - Non-cardiac death 0 (0.0%) 1 (2.3%) -0.19 Hospitalization, all 29 (39.9%) 18 (44.7%) 0.81 0.81 (0.45, 1.46)0.48 - For heart failure - For heart failure 11 (15.0%) 8 (19.6%) 0.75 0.75 (0.30, 1.87)0.54 Death or hosp 31 (41.5%) 19 (46.3%) 0.82 0.82 (0.46, 1.46)0.50 Death or HF hosp 13 (17.4%) 9 (21.7%) 0.79 0.79 (0.34, 1.86)0.59 Data are time to event Kaplan-Meier estimates, compared by log rank

23. PRomPT: Limitations Sample size – reduced to 120 as-treated patients Sample size – reduced to 120 as-treated patients  Nonetheless, the largest trial of its type to date  With 95% confidence we can exclude a >12.3 mL reduction in ΔLVEDV with successful peri-infarct pacing The study was not powered to assess most secondary endpoints, or compare outcomes between the individual pacing groups Cannot exclude benefit in subgroups that the trial was under-powered to detect Cannot exclude benefit in subgroups that the trial was under-powered to detect

24. In the present multicenter, randomized trial, peri-infarct pacing did not prevent LV remodeling or improve functional or clinical outcomes during 18 months of follow-up in patients with large first MI 24 PRomPT: Conclusions