Percutaneous coronary angioplasty versus coronary artery bypass grafting in treatment of unprotected left main stenosis Nordic–Baltic–British left main revascularisation study (NOBLE) A prospective, randomised, open-label, non-inferiority trial NOBLE Evald Høj Christiansen Timo Mäkikallio, Niels R Holm, Mitchell Lindsay, Mark S Spence, Andrejs Erglis, Ian B A Menown, Thor Trovik, Markku Eskola, Hannu Romppanen, Thomas Kellerth, Jan Ravkilde, Lisette O Jensen, Gintaras Kalinauskas, Rikard B A Linder, Markku Pentikainen, Anders Hervold, Adrian Banning, Azfar Zaman, Jamen Cotton, Erlend Eriksen, Sulev Margus, Henrik T Sørensen, Per H Nielsen, Matti Niemelä, Kari Kervinen, Jens F Lassen, Michael Maeng, Keith Oldroyd, Geoff Berg, Simon J Walsh, Colm G Hanratty, Indulis Kumsars, Peteris Stradins, Terje K Steigen, Ole Fröbert, Alastair NJ Graham, Petter C Endresen, Matthias Corbascio, Olli A Kajander, Uday Trivedi, Juha Hartikainen, Vesa Anttila, David Hildick–Smith, Leif Thuesen, and Evald H Christiansen On behalf of the NOBLE investigators

Disclosure Statement of Financial Interest Within the past 12 months, I have had a financial interest/arrangement or affiliation with the organization(s) listed below. Affiliation/Financial Relationship Company Grant/Research Support Biosensors

Background PCI is increasingly used in the treatment of left main coronary artery (LMCA) stenosis Evidence suggest that PCI provides clinical outcomes comparable to CABG in patients with less complex coronary artery disease Dedicated, adequately powered trials are required to confirm that PCI is a valid alternative

Hypothesis PCI with drug-eluting stents produce non-inferior clinical results compared with CABG in revascularization of patients with unprotected left main coronary artery stenosis

Inclusion criteria Stable angina, unstable angina, or acute coronary syndrome A significant left main lesion Visually assessed stenosis diameter >50% or fractional flow reserve ≤0.80 Located in the ostium, mid-shaft, or bifurcation No more than three additional non-complex lesions Local interventional cardiologists and cardiac surgeons determined that equivalent revascularization could be achieved with CABG or PCI

Exclusion criteria Additional non left main complex lesions Chronic total occlusions Bifurcation lesions requiring two stent techniques Calcified or tortuous vessel morphology ST-elevation infarction within 24 h Being considered too high-risk for CABG or PCI Expected survival of less than 1 year

Design A prospective, randomized, open-label, non-inferiority trial, carried out at 36 hospitals in Latvia, Estonia, Lithuania, Germany, Norway, Sweden, Finland, United Kingdom, and Denmark Enrollment: December 2008 to January 2015

Primary endpoint A composite of major adverse cardiac and cerebrovascular events (MACCE) Death from any cause Non-procedural myocardial infarction Repeat revascularization Stroke

Sample size SYNTAX trial MACCE after 2 years*(30% of PCI,23% of CABG) A HR of 1·35 was defined accordingly as the clinically acceptable non-inferiority limit not to be exceeded by the one-sided 95% CI 1- β (power) = 80% 275 events, with 1200 patients, 600 in each group needed The primary endpoint assessment was January 2015 changed to include MACCE endpoints occurring between 2 and 5 years to reach a total of 275 events September 2015 it was estimated that 275 events could not be reached within full 5 years (January 2020), and the primary endpoint assessment was changed to median 3 years *Predicted from preliminary 1-year results in the SYNTAX trial

Secondary endpoints The individual components of the primary endpoint Definite stent thrombosis Symptomatic graft occlusion Procedural myocardial infarction (post hoc) Repeat revascularization Target lesion Left main coronary artery target lesion De novo lesion

Did not receive PCI (n=13) Died before PCI (n=1) Randomized (n= 1201) Allocated to PCI (n=598) Received PCI (n=585) Did not receive PCI (n=13) Died before PCI (n=1) Patient declined PCI (n=4) PCI operator declined (n=4) LMCA lesion not significant (n=4) Allocated to CABG (n=603) Received CABG (n=570 ) Did not receive CABG (n=33) Died before CABG (n=1) Patient declined CABG (n=15) Not eligible for CABG (n=15) Cross over by mistake (n=2) Lost to follow-up (n=6) Emigration (n=1) Contact lost (n=2) Withdrawal (n=3) Lost to follow-up (n=11) Emigration (n=0) Contact lost (n=0) Withdrawal (n=11) Patients allocated to PCI in analysis (n=592) 580 received PCI 7 received CABG Patients allocated to CABG in analysis (n=592) 567 received CABG 23 received PCI

Baseline characteristics   PCI CABG P–value Age (yrs) 66·2±9·9 66·2±9·4 0·91 Gender (female) 116 (20%) 140 (24%) 0·09 BMI (kg/m2) 27·9±4·5 28·1±4·4 0·53 Diabetes type I or type II 86 (15%) 90 (15%) 0·94 Family history of IHD 321 (58%) 307 (56%) 0·45 Statin treatment 482 (82%) 464 (78%) 0·17 Hypertension 386 (65%) 389 (66%) Active smoking 108 (19%) 127 (22%) 0·18

Clinical characteristics   PCI CABG P–value Ejection fraction (% [IQR]) 60 [55;65] 60 [52;64] 0·27 Stable angina pectoris 466 (79%) 476 (81%) 0·61 Acute coronary syndrome 106 (18%) 100 (17%) 0·65 NYHA class I 244 (53%) 195 (43%) II 135 (30%) 150 (33%) III 57 (13%) 77 (17%) IV 23 (5%) 33 (7%) 0·012 EUROSCORE 2 [2:4] 0·18

Lesion characteristics   PCI CABG P–value SYNTAX score 22·5±7·5 22·4±8·0 0·74 Distal LMCA lesion 477 (81%) 482 (81%) 0·77

Treatment PCI LMCA stenting involving ostium and not the bifurcation 59 (10%) Shaft LMCA stenting only 11 (2%) LMCA bifurcation lesion stenting 508 (88%) IVUS pre–evaluation 270 (47%) IVUS post–evaluation 430 (74%) Complete revascularization 543 (94%)

Treatment CABG Off–pump technique 88 (16%) Arterial graft 532 (95%) Arterial graft to LAD 526 (93%)

Results Primary endpoint: MACCE 28·9% HR 1·48 (1·11–1·96); p=0·0066 19·1% PCI did not show non-inferiority and CABG was superior to PCI

Results All-cause mortality 11.5% HR 1·07 (0·67–1·72); p=0·77 11·6% 9.5% 9·5%

Results Non-procedural myocardial infarction HR 2·88 (1·40–5·90); p=0·004 6·9% 1·9%

Results Total repeat revascularization HR 1·50 (1·04–2·17); p=0·03 16·2% 10·4%

Results Stroke HR 2·25 (0·92–5·48); p=0·07 4·9% 1·7%

Results Secondary endpoints   PCI CABG P–value Definite ST or symptomatic graft occlusion* 3%(9) 4%(15) 0·22 Procedural myocardial infarction (post hoc)  5%(16/296) 7%(16/238) 0·52  *Kaplan-Meier 5 year estimates by intention-to-treat

Results Secondary endpoints   PCI CABG P–value Repeat revascularization 16%(71) 10%(47) 0·032 Target lesion revascularization 12%(50) 8%(36) 0·14 De novo lesion revascularization 6%(24) 3%(11) 0·018 Target LMCA revascularization 10%(41) 9%(33) 0·37 Kaplan-Meier 5 year estimates by intention-to-treat

Results SYNTAX score subgroups K-M estimates 4.9% 1.9% HR 1·88 (1·23–2·89); p=0·0031 HR 1·16 (0·76–1·78); p=0·48 HR 1·41 (0·62–3·20); p=0·41 SYNTAX score assessed by independent corelab (CERC)

Conclusions PCI did not meet non-inferiority for the primary endpoint of 5- year MACCE compared to CABG CABG was superior to PCI PCI resulted in higher rates of non-procedural myocardial infarctions Repeat revascularization was higher after PCI, primarily due to de novo lesions and non LMCA target lesion revascularization All-cause mortality was similar for PCI and CABG

Backup slides

MACCE: Death, MI, stroke HR 1.33 (0.92-1.93); p=0.13 12.7% 8.8%

MACCE: Death, MI, stroke HR 1.47 95% CI (1.06-2.05) p=0.02

Kaplan-Meier 5 year estimates by intention-to-treat K-M estimates 4.9% 1.9%

Outcomes between index procedure and 30 days of follow-up by treatment group K-M estimates 4.9% 1.9% (*assessable in 296 [50%] of 592 and 238 [40%] of 592 patients)

1-year clinical outcome by treatment group K-M estimates 4.9% 1.9%

1-year clinical outcome by treatment group K-M estimates 4.9% 1.9%

Completed follow-up

Treatment PCI LMCA stenting involving ostium and not bifurcation 59 (10·2%) Shaft LMCA stenting only 11 (1·8%) LMCA bifurcation lesion stenting 508 (87·7%) IVUS pre–evaluation 270 (46·8%) IVUS post–evaluation 430 (74·9%) Total stent length in LMCA lesion (mm) 24 [IQR 15:35] Number of treated lesions 1 312 (53·4%) 2 191 (32·5%) 3 55 (9·6%) 4 14 (2·4%) 5 3 (0·5%) Complete revascularisation 543 (94·1%)

Treatment PCI: LMCA bifurcation Angulation less than 70 degrees 110 (22·3%) Visible calcification 213 (42·5%) Severe tortuosity 55 (10·8%) Stenting of LMCA–LAD only 300 (59·4%) Stenting of LMCA–Cx only 21 (4·2%) Culotte 119 (23·9%) Crush 20 (4·0%) T–stenting 41 (8·4%) V–stenting 1 (0·2%) Other technique 4 (0·8%) Kissing balloon post inflation 277 (54·5%)

Treatment CABG Off–pump technique 88 (15·6%) Arterial graft 532 (94·5%) Arterial graft to LAD 526 (93·4%) LIMA + RIMA grafts 44 (7·9%) LIMA + venous graft 480 (85·7%) Radial artery graft 26 (4·8%) Venous grafts only 27 (5·0%) Grafts per patient 1 23 (4·1%) 2 294 (52·0%) 3 220 (39·0%) 4 25 (4·4%) 5 3 (0·6%)