1. Seung-Jung Park, MD, PhD On behalf of the BEST investigators Professor of Medicine, University of Ulsan College of Medicine, Heart Institute, Asan Medical Center, Seoul, Korea Trial of Everolimus-Eluting Stents or Bypass Surgery for Coronary Disease (BEST Trial)

2. Background Recent studies have demonstrated that the rates of most adverse clinical outcomes in patients with multivessel coronary-artery disease are lower following CABG than with PCI. However, previous studies may have been limited by their use of first- generation drug-eluting stents.

3. BEST Trial Design DESIGN: A prospective, open-label, randomized trial OBJECTIVE: To compare PCI with everolimus-eluting stents and CABG for optimal revascularization of patients with multivessel coronary artery stenosis. PRINCIPAL INVESTIGATOR Seung-Jung Park, MD, PhD, Asan Medical Center, Seoul, Korea

4. Global Multi-Centers (N=27) CountrySiteInvestigator KoreaAsan Medical CenterSeung-Jung Park KoreaKeimyung University Dongsan Medical CenterSeung Ho Hur KoreaThe Catholic University of Korea Seoul St. Mary's HospitalKi Bae Seung KoreaGachon University Gil HospitalTae hoon Ahn KoreaGangnam Severance HospitalHyuck Moon Kwon KoreaKorea University Guro HospitalSeung Un Na KoreaKorea University Anam HospitalDo Sun Lim KoreaChonnam National University HospitalMyung-Ho Jeong KoreaKangwon National University HospitalBong-Ki Lee KoreaHanyang University Medical CenterKyoung Soo Kim KoreaKonyang University HospitalJang Ho Bae KoreaInje University Sanggye Paik HospitalByung Ok Kim KoreaWonju Christian HospitalJunghan Yoon KoreaInje University Pusan Paik HospitalTae-Hyun Yang KoreaSeverance HospitalYang-Soo Jang KoreaNational Health Insurance Corporation Ilsan HospitalJoo-Young Yang KoreaYeungnam University Medical CenterJong-Seon Park KoreaInje University Ilsan Paik HospitalSung Yun Lee KoreaPusan National University Yangsan HospitalJun Hong Kim KoreaSt.carollo HospitalJang-Hyun Cho KoreaThe Catholic University of Korea, Yeouido St. Mary's HospitalYun Seok Choi KoreaUlsan University HospitalSang-Gon Lee ChinaSir Run Run Shaw HospitalGuo Sheng Fu ChinaZhongshan HospitalJunBo Ge MalaysiaNational Heart InstituteRobaaya Zambahari MalaysiaSarawak General HospitalTiong Kiam Ong ThailandSiriraj HospitalDamras Tresukosol

5. Major Inclusion Criteria  18 years of age. Angiographically confirmed mutivessel coronary artery disease (>70%) Suitable candidates for either PCI or CABG by their treating physicians and surgeons Symptoms of angina and/or objective evidence of myocardial ischemia.

6. Major Exclusion Criteria Any contraindication to dual antiplatelet therapy Severe heart failure (NYHA III or IV) Planned surgery Previous CABG Prior PCI with DES implantation within 1 year CTO ≥2 STEMI within 72 hours Elevated cardiac enzyme Disabled stroke Other comorbidity

7. Study Procedures Everolimus-Eluting Xience Stent for all lesions Strong recommendation of IVUS-guidance Other adjunctive devices at the physician’s discretion Use of LIMA to LAD anastomosis Off- or on-pump surgery at the surgeon’s discretion DAPT at least for 1 year after PCI Standard medical treatment after PCI and CABG

8. Follow-up Clinical follow-up at 30 days and 6, 9, and 12 months, and annually thereafter, via clinic visit or telephone interview. Secondary preventive medication was strongly recommended according to clinical guideline Routine angiographic follow-up was strongly discouraged for all patients to reduce the occurrence of repeat revascularization driven by angiography alone without evidences of ischemia.

9. Primary End Point A composite of major adverse cardiac events (MACE) at the 2 years after randomization including - Death from any cause - Myocardial infarction ≤48 h: New Q waves and CK-MB > 5 times >48 h: Any CK-MB elevation and ischemic symptoms or signs - Target vessel revascularization

10. Original Power Calculation Assumed MACE rate: 12% at 2 years A noninferiority margin : 4% A one-sided type I error rate : 0.05 Power : 80% Dropout rate: 5% Assumed sample size: 1776 patients Non-inferiority Design for Primary Endpoint

11. Premature Termination of Trial The enrollment rate was slower than expected, which was thought to be a consequence of the rapid spread of measurement of fractional flow reserve in clinical practice. The data and safety monitoring board recommended stopping enrollment in October 2013 when 880 patients had been enrolled. We extended the follow-up period with a median of 4.6 years.

12. Patient Flow 4654 patients were screened Between July 2008 and September 2013 PCI (n=438) Treated PCI: 413 Treated CABG: 9 Treated medically: 6 CABG (n=442) Treated CABG: 382 Treated PCI: 51 Treated medically: 9 1725 patients were eligible 880 patients consented and enrolled 1 Year FU (N=438) 3 Year FU (N=369) 5 Year FU (N=172) 1 Year FU (N=442) 3 Year FU (N=373) 5 Year FU (N=171) 1 patient2 patients Lost to Follow-up At 5 year

13. Statistical Analysis Kaplan-Meier method to estimate survivals with comparison using log-rank test. Noninferiority test using the Z-test with 95% CI of difference in the 2-year MACE rate. Survival analyses for longer-term outcomes using all available follow-up data as an exploratory analyses. Subgroups analysis using the Cox regression model with tests for interaction. Primary analysis in intention-to-treat principle

14. Baseline Clinical Characteristics

16. AspirinThienopyridine StatinBeta blocker Medication at Follow-Up CABG PCI % % % %

17. Procedural Characteristics* PCI 464 Total stents number 3.4 ± 1.4 Total stent length, mm 85.3 ± 38.2 Mean stent diameter, mm 3.1 ± 0.3 IVUS guidance 333 (71.8) Complete revascularization 236 (50.9) † CABG 401 Total no. of grafted vessels3.1 ± 0.9 Total no. of arterial grafts2.1 ± 1.1 Total no. of vein grafts1.0 ± 0.8 Left internal mammary artery graft398 (99.3) Off-pump surgery258 (64.3) Complete revascularization274/383 (71.5) † * Data were summarized according to the as-treated analysis † P<0.05 between PCI and CABG group

18. Noninferiority Test for Primary End Point Prespecified non-inferiority margin: 4% Upper 1-sided 95% CI -2012345678910 Non-inferiority P=0.32 Absolute Risk Difference 3.1% points 95% CI -0.8-6.9 2-year MACE rate, PCI: 11.0% CABG: 7.9% Difference (percentage point) of 2-year MACE rate (PCI – CABG)

19. Long-Term Follow-up

20. Primary End Point 012345 0 5 10 15 20 25 30 17.0% No. at Risk PCI CABG 402 415 362 377 305 326 438 442 242 262 126 145 CABG PCI 11.7% Cumulative Incidence, % Log-rank P=0.043 Years Since Randomization Event rates were derived from Kaplan-Meier estimates

21. Death, MI or Stroke 012345 0 5 10 15 20 25 30 No. at Risk PCI CABG 413 419 373 381 318 329 438 442 Log-rank P=0.26 255 263 133 144 Years Since Randomization 13.4%13.4% 10.2% Cumulative Incidence, % Event rates were derived from Kaplan-Meier estimates CABG PCI

22. Death 012345 No. at Risk PCI CABG 426 433 387 397 333 346 438 442 Log-Rank P=0.30 268 278 146 154 7.5% 5.5% Years Since Randomization 5 10 15 20 25 30 0 Cumulative Incidence, % Event rates were derived from Kaplan-Meier estimates Cardiac Death: HR 1.15 (0.58-2.25), P=0.69 Non-Cardiac Death: HR 1.87 (0.69-5.05), P=0.21 CABG PCI

23. Myocardial Infarction 012345 0 No. at Risk PCI CABG 419 422 382 386 325 335 438 442 Log-Rank P=0.11 261 271 140 151 5.5% 2.8% Years Since Randomization 5 10 15 20 25 30 Cumulative Incidence, % Event rates were derived from Kaplan-Meier estimates CABG PCI

24. Land Mark Analysis of MI Event rates were derived from Kaplan-Meier estimates 0 0 3 6 9 12 15 12345 Years Since Randomization 30days Cumulative Incidence, % CABG PCI >30 days: HR, 5.04 (95% CI, 1.46-17.4), P=0.004 4.2% 0.8% 0-30days: HR, 0.67 (95% CI, 0.24-1.88), P=0.44

25. Land Mark Analysis of Death Event rates were derived from Kaplan-Meier estimates 0 0 3 6 9 12 15 12345 Years Since Randomization 30days 0-30days: HR, 0.51 (95% CI, 0.09-2.76), P=0.42 >30 days: HR, 1.53 (95% CI, 0.84-2.78), P=0.16 7.0% 4.6% Cumulative Incidence, % CABG PCI

26. Land Mark Analysis of Death and MI Event rates were derived from Kaplan-Meier estimates 0 0 3 6 9 12 15 12345 Years Since Randomization 30days 0-30days: HR, 0.62 (95% CI, 0.26-1.49), P=0.28 >30 days: HR, 1.69 (95% CI, 0.982.91), P=0.054 9.6% 5.5% Cumulative Incidence, % CABG PCI

27. Stroke 012345 0 No. at Risk PCI CABG 421 427 383 389 326 338 438 442 Log-Rank P=0.72 262 271 140 152 Years Since Randomization 3.3% 2.9% 5 10 15 20 25 30 Cumulative Incidence, % Event rates were derived from Kaplan-Meier estimates CABG PCI

28. Any Repeat Revascularization 012345 No. at Risk PCI CABG 393 414 335 365 257 286 438 442 Log Rank P=0.003 164 189 13.4% 6.6% 80 87 Years Since Randomzation 5 10 15 20 25 30 Cumulative Incidence, % 0 Event rates were derived from Kaplan-Meier estimates CABG PCI

29. Target Lesion Revascularization Event rates were derived from Kaplan-Meier estimates 012345 0 No. at Risk PCI CABG 408 424 365 386 310 334 438 442 247 267 130 147 Years Since Randomization 5 10 15 20 25 30 6.1% 4.5% Log Rank P=0.19 Cumulative Incidence, % CABG PCI

30. New Lesion Revascularization Event rates were derived from Kaplan-Meier estimates 012345 0 No. at Risk PCI CABG 416 427 370 389 317 337 438 442 254 270 138 149 Years Since Randomization 5 10 15 20 25 30 Cumulative Incidence, % Log Rank P=0.013 6.5% 2.4% CABG PCI

31. Death, MI, Stroke or Any RR (SYNTAX Primary Endpoint) 012345 0 5 10 15 20 25 30 21.7% 14.6% No. at Risk PCI CABG 389 409 341 368 288 317 438 442 229 250 117 137 Cumulative Incidence, % Log Rank P=0.01 Event rates were derived from Kaplan-Meier estimates CABG PCI

32. Definite or Probable Stent Thrombosis Event rates were derived from Kaplan-Meier estimates 012345 0 1 2 3 4 5 Years Since Randomzation 2.0% Cumulative Incidence, % 4 Definite ST 3 Probable ST

33. Subgroups for Primary Endpoint 0.1110 Subgroup Primary Outcome PCI CABG n / total n. (%) Overall 67/438 (15.3) 47/442 (10.6) Age ≥ 65 yr 41/229 (17.9) 30/252 (11.9) <65 yr 26/209 (12.4) 17/190 (8.9) Sex Male 45/304 (14.8) 34/325 (10.5) Female 22/134 (16.4) 13/117 (11.1) Diabetes Yes 34/177 (19.2) 17/186 (9.1) No 33/261 (12.6) 30/256 (11.7) ACS Yes 40/228 (17.5) 33/238 (13.9) No 27/210 (12.9) 14/204 (6.9) Ejection fraction ≤ 40% 7/17 (41.2) 4/17 (23.5) >40% 60/421 (14.3) 43/425 (10.1) Vascular extent 3VD 56/330 (17.0) 42/349 (12.0) 2VD 11/108 (10.2) 5/93 (5.4) SYNTAX score Score ≥ 33 13/66 (19.7) 10/79 (12.7) Score 23-32 30/187 (16.0) 14/177 (7.9) Score ≤ 22 24/185 (13.0) 23/186 (12.4) EuroSCORE ≥ 6 12/51 (23.5) 11/59 (18.6) <6 55/387 (14.2) 36/383 (9.4) Hazard Ratio (95% CI) P value for Interaction 1.47 (1.01-2.13) - 0.90 1.51 (0.95-2.42) 1.43 (0.77-2.63) 0.88 1.43 (0.92-2.24) 1.53 (0.77-3.05) 0.06 2.24 (1.25-4.00) 1.07 (0.65-1.76) 0.35 1.30 (0.82-2.06) 1.89 (0.99-3.60) 0.65 1.79 (0.51-6.21) 1.43 (0.97-2.12) 0.65 1.45 (0.97-2.17) 1.89 (0.66-5.43) 0.25 1.59 (0.70-3.62) 2.14 (1.13-4.03) 1.04 (0.59-1.84) 0.65 1.25 (0.55-2.84) 1.55 (1.02-2.35) PCI betterCABG better

34. End points PCI (N=438) CABG (N=442) Hazard ratio (95% CI) P-value Primary End Points: MACE 67 (15.3)47 (10.6)1.47 (1.01-2.13)0.043 Secondary End Points Death 29 (6.6)22 (5.0)1.34 (0.77-2.34)0.30 Myocardial Infarction 21 (4.8)12 (2.7)1.76 (0.87-3.58)0.11 Spontaneous MI 19 (4.3)7 (1.6)2.75 (1.16-6.54)0.017 Stroke 11 (2.5)13 (2.9)0.86 (0.39-1.93)0.72 Death, Myocardial Infarction, or stroke 52 (11.9)42 (9.5)1.26 (0.84-1.89)0.26 Any Repeat Revascularization 48 (11.0)24 (5.4)2.09 (1.28-3.41)0.003 Target Lesion Revascularization 25 (5.7)17 (3.8)1.51 (0.82-2.80)0.19 New Lesion Revascularization 24 (5.5)10 (2.3)2.47 (1.18-5.17)0.013 Death, MI, Stroke, or Any RR 87 (19.9)59 (13.3)1.54 (1.11-2.14)0.01 Bleeding TIMI Major Bleeding‡ 30 (6.8)132 (29.9)0.20 (0.14-0.30)<0.001 Fatal Bleeding 3 (0.7)7 (1.6)0.44 (0.11-1.68)0.21 Percentages are crude rates throughout the available follow-up period Summary

35. Conclusion The BEST trial failed to show that PCI with everolimus-eluting stents was noninferior to CABG with respective to the primary end point of death, myocardial infarction, or target vessel revascularization at 2 years. At longer-term follow-up (median 4.6 years), PCI was associated with a significant increase in the incidence of the primary end point compared with CABG.

36. This article was published on March 16, 2015, at NEJM.org.

37. Reasons for Screening Failure for Enrollment

38. Myocardial Infarction Percentages are crude rates throughout the available follow-up period Crude Incidence, % P=0.48P=0.02 CABG PCI Q wave plus CK-MB>5 Times Ischemic Sx or Sign plus CK-MB elevation

39. Complete vs. Incomplete PCICABGHR (95% CI)P value Complete Revascularization N=215N=295 Death7.0%4.4%1.50 (0.71-3.15)0.29 Myocardial infarction2.3%3.1%0.75 (0.25-2.24)0.60 Spontaneous myocardial infarction1.9%1.4%1.34 (0.34-5.37)0.68 Death, MI, or Stroke11.6%9.5%1.18 (0.69-2.02)0.55 Any repeat revascularization6.5%3.4%1.89 (0.84-4.25)0.13 Target lesion revascularization4.7%2.7%1.67 (0.66-4.23)0.28 New lesion revascularization2.8%1.0%2.71 (0.68-10.8)0.16 Death, MI, or TVR13.5%9.5%1.37 (0.81-2.30)0.24 MACCE16.7%12.2%1.34 (0.84-2.13)0.22 Incomplete Revascularization N=215N=122 Death6.5%5.7%1.22 (0.49-3.02)0.68 Myocardial infarction7.4%1.6%4.85 (1.11-21.1)0.036 Spontaneous myocardial infarction7.0%1.6%4.68 (1.07-20.5)0.041 Death, MI, or Stroke12.6%9.0%1.52 (0.75-3.07)0.24 Any Repeat Revascularization15.8%10.7%1.58 (0.83-3.00)0.16 Target lesion revascularization7.0%7.4%0.96 (0.42-2.20)0.92 New lesion revascularization8.4%4.9%1.83 (0.73-4.62)0.20 Death, MI, or TVR17.7%13.1%1.45 (0.81-2.61)0.21 MACC23.7%16.4%1.59 (0.94-2.66)0.08

40. As Treated Analysis

41. Noninferiority Test for Primary End Point Prespecified non-inferiority margin: 4% Upper 1-sided 95% CI -2012345678910 Non-inferiority P=0.44 Absolute Risk Difference 3.7% points 95% CI -0.2-7.6 2-year MACE rate PCI: 11.2% CABG: 7.5% Difference (percentage point) of 2-year MACE rate (PCI – CABG)

42. Primary End Point of MACE Event rates were derived from Kaplan-Meier estimates 012345 0 5 10 15 20 25 30 No. at Risk PCI CABG 425 378 385 343 328 295 464 401 267 230 148 125 Years Since Randomization Cumulative Incidence, % 17.2%17.2% 11.0% Log-rank P=0.02 CABG PCI