Supplementary Appendix

Transcription

1 Supplementary Appendix This appendix has been provided by the authors to give readers additional information about their work. Supplement to: Xaplanteris P, Fournier S, Pijls NHJ, et al. Five-year outcomes with guided by fractional flow reserve. N Engl J Med. DOI: /NEJMoa

2 Table of contents I. Investigators and Committee Member... 4 II. Supplementary Methods... 5 A. Patient inclusion and exclusion criteria... 5 B. End point definitions... 6 C. Data and Safety Monitoring Board stopping rules... 8 D. Sample size considerations... 9 III. Supplementary Figures Figure S1: Flow chart of the complete FAME-2 population including registry patients Figure S2: Kaplan Meier curves for the primary composite endpoint of death, myocardial infarction or urgent revascularization for the two randomized groups and the registry cohort Figure S3: Kaplan Meier curves for spontaneous myocardial infarction (A) and peri-procedural myocardial infarction (B) for the two randomized groups Figure S4: Kaplan Meier curves for the composite of death or myocardial infarction for the two randomized groups Figure S5: Kaplan Meier curves for any revascularization (A) and non-urgent revascularization (B) for the two randomized groups Figure S6: Kaplan Meier curves for the composite of death, myocardial infarction or any revascularization for the two randomized groups Figure S7: Kaplan Meier curves for the composite of cardiac death or myocardial infarction for the two randomized groups Figure S8: Landmark analyses for the primary composite endpoint (A) and its components death (B), myocardial infarction (C), and urgent revascularization (D) for the two randomized groups Figure S9: Subgroup analyses for the primary composite endpoint for the two randomized groups Figure S10: Caterpillar plot showing the variation in risk ratios of urgent revascularization across trial sites IV. Supplementary Tables Table S1: List of hospitals that participated in the 5-year follow-up and respective enrolment in the randomized groups and registry cohort

3 Table S2: Baseline clinical, angiographic, and FFR characteristics of randomized and registry patients Table S3: Baseline clinical, angiographic, and FFR characteristics of randomized patients from centers participating in the 5-year follow-up Table S4: Baseline clinical, angiographic, and FFR characteristics of randomized patients from centers not participating in the 5-year follow-up Table S5: Baseline clinical, angiographic, and FFR characteristics of patients according to site participation in the 5-year follow-up Table S6: Overview of patient flow according to site participation in the 5-year follow-up Table S7: Cardiac medications at baseline, and at 1 month, 6 months, 1 year, 2 years, 3 years and 5 years of follow-up Table S8: Causes of death by treatment group Table S9: Types of myocardial infarction and additional composite endpoints at 5-year follow-up Table S10: Subgroup analyses according to participation in the 5-year followup for the primary endpoint and its components for the two randomized groups

4 I. Investigators and Committee Members The members of the FAME II study group are as follows: Steering Committee B. De Bruyne, W.F. Fearon, N.H.J. Pijls, E. Barbato, P.A.L. Tonino, P. Jüni, U. Lonn Clinical Events Committee M.J. Kern (Chairman), E. Mahmud, M. Lim; Data Safety Monitoring Board S. Windecker (Chairman), B.J. Gersh, S.J. Pocock. Study Investigators: United States Stanford University Medical Center and Palo Alto Veterans Affairs Health Care Systems, Stanford, CA: W. Fearon, T. Brinton, D. Lee, T. Carroll, D. Daniels, J. Giacomini, J. Trammel, A. Yeung; Eastern Maine Medical Center, Bangor, ME: P. Ver Lee, G. Crespo, R. Fincke, M. McKay, M. Rowe, P. Vom Eigen, A. Wiseman; Atlanta Veterans Affairs Medical Center, Decatur, GA: K. Mavromatis, P. Block, G. Kumar; Emory University School of Medicine, Atlanta, GA: H. Samady, V. Babaliaros, J. Douglas, M. McDaniel, D. Morris, L. Sperling, S. Tanveer Rab; Tulane University Heart and Vascular Institute, New Orleans, LA: S. Arain. Canada Hopital du Sacré-Coeur de Montréal, Montreal: E. Schampaert; Le Centre Hospitalier de l'université de Montréal, Montreal: S. Mansour, F. Gobeil, A. Kokis, F. Lemire, C. Pilon. United Kingdom King s College Hospital, London: P. MacCarthy, J. Byrne, N. Melikian; Southampton University Hospital NHS Trust, Southampton: N. Curzen, A. Calver, H. Gray; Royal Victoria Hospital, Belfast: G. Manoharan; Golden Jubilee National Hospital, Glasgow: K. Oldroyd, H. Etieba, M. Lindsay; Edinburgh Heart Centre, Edinburgh: N. Uren, N. Cruden, P. Henriksen. Germany Klinikum der Universitat Munchen-Campus-Innenstadt, Munich: H. Sohn, M. Leibig; Heart Center Leipzig, Leipzig: S. Möbius-Winkler; Herzzentrum Munchen- Bogenhausen, Munich: J. Rieber, M. Deichstetter. The Netherlands Department of Cardiology, Catharina Hospital, Eindhoven, and Department of Biomedical Enginering, Eindhoven University of Technology: N.H.J. Pijls, S. de lo Fuente, P. Tonino, I. Wijnbergen, J. Willem Sels; Isala Klinieken, Zwolle: J. Dambrink; St. Antonius Ziekenhuis, Nieuwegein: B.J.W.M. Rensing, B.J.M. Mulder, A.N.van den Akker. Czech Republic Masaryk University and University Hospital, Brno: P. Kala, P. Jerabek, P. Neugebauer; Na Homolce Hospital, Bohemia: M. Mates, P. Korvicek. Sweden Karolinska Institutet at Södersjukhuset, Stockholm: N. Witt, P. Alstrom; Örebro University Hospital, Orebro: O. Fröbert. Belgium Cardiovascular Center Aalst, OLV-Clinic Aalst: B. De Bruyne, E. Barbato, J. Bartunek, G. Heyndrickx, C. Van Mieghem, W. Wijns, M. Vanderheyden, E. Wyffels.L. Vandriessche, A. Heyse Denmark Rigshospitalet University Hospital, Copenhagen: T. Engstroem, E. Jorgensen, H. Kelbaek. France Cardiovascular Hospital, Lyon: G. Rioufol, G. Finet. Hungary Hungarian Institute of Cardiology, Budapest: Z. Piroth, P. Andréka, G. Fontos, G. Tóth. Italy University Hospital of Ferrara, Ferrara: M. Valgimigli, G. Campo, C. Tumscitt. Serbia Clinical Center Kragujevac, Kragujevac: N. Jagic, V. Miloradovic, D. Nikolic. Independent adjudication of urgent and non-urgent revascularizations regarding trigger and angina severity: Giulio Stefanini, MD, Thomas Pilgrim, MD and Stefan Stortecky, MD. Events ascertained after the original close-out were adjudicated by two blinded cardiologists not involved in the study: Mariano Pellicano, MD and Giovanni Ciccarelli, MD. 4

5 II. Supplementary Methods A. Patient inclusion and exclusion criteria Eligible subjects include patients with stable clinical condition and one-, two- or threevessel disease at coronary angiography and amenable for. Inclusion is based on the combination of clinical data, non-invasive testing, when available, and angiographic data. Inclusion criteria: 1. Stable angina pectoris (Canadian Cardiovascular Society Class [CCS] 1, 2, 3); or angina pectoris CCS class 4 subsequently stabilized medically (minimum 7 days); or atypical or no chest pain but documented ischemia on noninvasive testing 2. At least one stenosis of at least 50% diameter reduction in at least one major native epicardial coronary artery with a diameter of at least 2.5 mm and supplying viable myocardium 3. Eligible for 4. Signed written informed consent obtained Note: a) Patients with restenosis in native coronary arteries may be included b) Patients with previous stents and restenosis may be included. c) Total occlusion may be included if this vessel supplies viable myocardium, and if recanalization is deemed likely and useful by the operator and if it is not the only lesion with positive FFR. In total chronic occlusions it is not mandated to measure FFR. The FFR value will be set at an arbitrary value of 0.5. d) Patients who sustained a STEMI or a NSTEMI more than one week ago may be included in the trial (see exclusion criteria) Exclusion criteria: 1. Patients in whom the preferred treatment is CABG 2. Patients with left main coronary artery disease requiring revascularization 3. Patients with a recent (less than 1 week) STEMI or NSTEMI 4. Prior CABG 5. Contraindication to dual antiplatelet therapy 6. LVEF < 30% 7. Severe LV hypertrophy (defined as a septal wall thickness at echocardiography of more than 13 mm) 8. Planned need for concomitant valvular or aortic surgery 9. Extremely tortuous or calcified coronary arteries precluding FFR measurements 10. A life expectancy of less than 2 years 11. Age under Pregnancy or intention to become pregnant during the course of the trial 5

6 13. Refusal or inability to sign an informed consent. Mental condition (psychiatric or organ cerebral disease) rendering the subject unable to understand the nature, scope, and possible consequences of the trial or mental retardation or language barrier such that the patient is unable to give informed consent 14. Potential for noncompliance towards the requirements in the trial protocol (especially the medical treatment) or follow-up visits 15. Participation or planned participation in another cardiovascular clinical trial before twoyear follow-up is completed B. End point definitions The primary end point of the FAME II trial was defined as a composite of the 24-month: o All cause death o Documented myocardial infarction o Unplanned hospitalization leading to urgent revascularization as adjudicated by the Clinical Events Committee (CEC). Death All patient deaths will be documented on the CRF. The Sponsor or designee must be notified of a patient's death within 24 hours after the clinical site has knowledge of the event. The principal investigator's narrative summary of the circumstances of death is required. Autopsy results, when available, should be reported to the CRO. In the primary comparison of the two treatment strategies, all deaths will be examined. Death due to specific causes will be investigated and adjudicated by the CEC. All deaths are considered cardiac unless an unequivocal non-cardiac cause can be established. Cardiac death: Any death due to immediate cardiac causes (e.g. MI, low-output failure, fatal arrhythmia). Unwitnessed death and death of unknown cause will be classified as cardiac death. Vascular death: Death due to cerebrovascular disease, pulmonary embolism, ruptured aortic aneurysm, dissecting aneurysm, or other vascular cause. Non-cardiovascular death: Any death not covered by the above definitions, including death due to infection, sepsis, pulmonary causes, accident, suicide or trauma. Myocardial infarction 6

7 A myocardial infarction (MI) will be considered an event whether it occurred within the first 24 hours after randomization or more than 24 hours after randomization. A definite diagnosis of myocardial infarction is made based on the following: a. Within 24 hours after randomization or any : I. CK-MB above 10 x 99th percentile upper reference limit (URL) determined on a single measurement, OR II. CK-MB above 5 x 99th percentile URL determined on a single measurement PLUS at least one of the following: o new pathological Q waves in at least 2 contiguous leads or new persistent non-rate related LBBB, o angiographically documented native coronary artery occlusion, o imaging evidence of new loss of viable myocardium b. More than 24 hours after randomization: I. Detection of rise and/or fall of cardiac biomarkers, CK-MB or troponin with at least one value above the 99th percentile of the URL together with evidence of myocardial ischemia with at least one of the following: o Symptoms of ischemia o ECG changes indicative of new ischemia [new ST-T changes or new left bundle branch block (LBBB)], OR II. Development of pathological Q waves ( 0.03 seconds in duration and 1 mm in depth) in 2 contiguous precordial leads or 2 adjacent limb leads) of the ECG, OR III. Imaging evidence of loss of viable myocardium or new regional wall motion abnormality For each MI adjudicated by the CEC, the type of MI will also be described as: ST-Elevation MI (STEMI) Also categorize as: o Q-wave o Non-Q-wave o Unknown (no ECG or ECG not interpretable) Non-ST-Elevation MI (NSTEMI) Also categorize as: o Q-wave o Non-Q-wave o Unknown (no ECG or ECG not interpretable) Unknown (no ECG or ECG not interpretable) Unplanned hospitalization leading to an urgent revascularization procedure This component of the primary endpoint will be present only if the patient is hospitalized unexpectedly because of persisting or increasing complaints of chest pain (with or without ST-T changes, with or without elevated biomarkers) AND a revascularization is performed 7

8 within the same hospitalization. This event should be clearly distinguished from the crossover revascularization procedure (secondary endpoint) which is performed on nonurgent basis. A complete and detailed narrative of this event by the investigator will be needed. Cerebrovascular event (stroke) A focal neurological deficit of central origin lasting more than 72 hours and results in irreversible brain damage or permanent body impairment. Type and severity of symptoms is dependent on the location and extent of brain tissue whose circulation has been involved. Strokes will be further classified as ischemic or hemorrhagic based on imaging studies. When blood flow to the brain is interrupted because of rupture of a vessel causing bleeding into or around the brain, it is considered hemorrhagic. When a vessel that supplies the brain is blocked, the event is considered ischemic. Reversible ischaemic neurological deficits (RINDs) and transient ischemic attacks (TIAs) events will be submitted to the CEC for final adjudication. Revascularization procedure Every subsequent revascularization procedure and its indication will be reported and documented in the appropriate CRF. Stent thrombosis Stent thromboses will be classified by the CEC according to the Academic Research Consortium definition as definite, probable, or possible and as early (0 to 30 days), late (31 to 360 days), or very late (>360 days). - Definite stent thrombosis requires the presence of an acute coronary syndrome with angiographic or autopsy evidence of thrombus or occlusion. - Probable stent thrombosis includes unexplained deaths within 30 days after the procedure or acute myocardial infarction involving the target- vessel territory without angiographic confirmation. - Possible stent thrombosis includes all unexplained deaths occurring at least 30 days after the procedure. Intervening target lesion revascularization is defined as any repeated percutaneous revascularization of the stented segment, including the 5-mm proximal and distal margins, that preceded stent thrombosis. C. Data and Safety Monitoring Board stopping rules The Data and Safety Monitoring Board (DSMB) charter specifies that the recommendation of stopping the study would be based on safety concerns as evidenced by statistical and clinical judgment to the rate of serious adverse events. There were no pre-specified formal statistical stopping rules for the study. An independent third party CRO was responsible for the data analysis and the CRO statistician provided the interface with DSMB members. The DSMB received twice-monthly reports of enrollment and events, including events as reported by the site and as adjudicated by the CEC and had full access to the data. The DSMB formally met to review study data by treatment group in April 2011, August 2011 and November The DSMB communicated its recommendation to stop the trial to the sponsor and the steering committee on November After careful consideration, the sponsor and the steering group decided to follow the recommendation 8

9 in early January 2012 and the sponsor notified the investigators per on January 15, 2012 that patient recruitment would be stopped immediately. D. Sample size considerations The trial was powered to determine whether with medical therapy was superior to the medical therapy alone with respect to the primary end point at 24 months. On the basis of findings from previous studies and using binomial proportions, we estimated that the cumulative incidence of the primary end point at 24 months would be 12.6% in the group and 18.0% in the medical-therapy group, corresponding to a relative risk reduction with of 30%, and that with 816 patients in each group, the study would have more than 84% power to detect this relative risk reduction at a two-sided type I error rate of However, patient recruitment was halted on January 15, 2012, after the randomization of 54% of patients of the initially planned sample size. 9

10 III. Supplementary Figures Figure S1: Flow chart of the complete FAME-2 population including registry patients. Underwent FFR (n=1220)* Randomized (n=888) Patients with FFR >0.80 in all lesions included in registry (n=332)** Allocated to +MT (n=447) Received allocated intervention (n=435) Did not receive allocated intervention (n=12) Treated with balloon angioplasty (n=3) Underwent CABG rather than (n=4) Received OMT, planned for staged procedure (n=3) Received OMT, unsuccessful (n=1) Received OMT, FFR >0.8 (n=1) Allocated to MT alone (n=441) Received allocated intervention (n=439) Did not receive allocated intervention (n=2) Erroneously received DES (n=2) Randomly selected to undergo follow-up (n=166) Received OMT alone (n=165) Received DES (n=1) Follow up information available (n=417) Followed up and alive (n=394) Deceased (n=23) Follow up information unavailable (n=30) Withdrew (n=10) Lost to follow up (n=20) Follow up information available (n=406) Followed up and alive (n=383) Deceased (n=23) Follow up information unavailable (n=35) Withdrew (n=8) Lost to follow up (n=27) Follow up information available (n=148) Followed up and alive (n=141) Deceased (n=7) Follow up information unavailable (n=18) Withdrew (n=2) Lost to follow up (n=16) Analyzed on clinical endpoints (n=447) Censored at time of loss to follow up or withdrawal (n=30) Analyzed on clinical endpoints (n=441) Censored at time of loss to follow up or withdrawal (n=35) Analyzed on clinical endpoints (n=166) Censored at time of loss to follow up or withdrawal (n=18) *1220 patients were screened, consented to undergo FFR and to be included in the randomized trial or registry; no information is available on patients who refused participation. **Note that 6 patients had total occlusions supplying a kinetic myocardium and were therefore not considered for ; 1 patient had 2 FFR negative lesions and was therefore included in the registry, however a subsequently detected total occlusion was eventually treated with DES. 10

11 Figure S2: Kaplan Meier curves for the primary composite endpoint of death, myocardial infarction or urgent revascularization for the two randomized groups and the registry cohort. Cumulative incidence (%) No. at risk Registry vs. : HR 0.46 (95% CI ) P<0.001 vs. Registry: HR 0.88 (95% CI ) P=0.57 vs. Registry: HR 1.91 (95% CI ) P= Years after randomization Registry Cumulative incidence of the primary endpoint (death, myocardial infarction or urgent revascularization) in the two randomized groups and the registry cohort. 11

12 Figure S3: Kaplan Meier curves for spontaneous myocardial infarction (A) and peri-procedural myocardial infarction (B) for the two randomized groups. A 100 vs. : HR 0.62 (95% CI ) Cumulative incidence (%) Cumulative incidence (%) Years after randomization No. at risk Years after randomization B 100 vs. : HR 0.77 (95% CI ) Cumulative incidence (%) Cumulative incidence (%) Years after randomization No. at risk Years after randomization Cumulative incidence of spontaneous myocardial infarction (A) and peri-procedural myocardial infarction (B) in the two randomized groups. Hazard ratios below 1.00 denote a lower incidence of events in the group than in the medical therapy group. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 12

13 Figure S4: Kaplan Meier curves for the composite of death or myocardial infarction for the two randomized groups. 100 vs. : HR 0.72 (95% CI ) Cumulative incidence (%) Cumulative incidence (%) Years after randomization No. at risk Years after randomization Cumulative incidence of death or myocardial infarction in the two randomized groups. Hazard ratios below 1.00 denote a lower incidence of events in the group than in the medical therapy group. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 13

14 Figure S5: Kaplan Meier curves for any revascularization (A) and non-urgent revascularization (B) for the two randomized groups. A 100 vs. : HR 0.19 (95% CI ) Cumulative incidence (%) No. at risk Years after randomization B 100 vs. : HR 0.18 (95% CI ) Cumulative incidence (%) No. at risk Years after randomization Cumulative incidence of any revascularization (A) and non-urgent revascularization (B) in the two randomized groups. Hazard ratios below 1.00 denote a lower incidence of events in the group than in the medical therapy group. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 14

15 Figure S6: Kaplan Meier curves for the composite of death, myocardial infarction or any revascularization for the two randomized groups. 100 vs. : HR 0.28 (95% CI ) Cumulative incidence (%) No. at risk Years after randomization Cumulative incidence of death, myocardial infarction or any revascularization in the two randomized groups. Hazard ratios below 1.00 denote a lower incidence of events in the group than in the medical therapy group. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 15

16 Figure S7: Kaplan Meier curves for the composite of cardiac death or myocardial infarction for the two randomized groups. 100 vs. : HR 0.70 (95% CI ) Cumulative incidence (%) Cumulative incidence (%) Years after randomization No. at risk Years after randomization Cumulative incidence of cardiac death or myocardial infarction in the two randomized groups. Hazard ratios below 1.00 denote a lower incidence of events in the group than in the medical therapy group. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 16

17 Figure S8: Landmark analyses for the primary composite endpoint (A) and its components death (B), myocardial infarction (C) and urgent revascularization (D) for the two randomized groups. A Cumulative incidence (%) to 7 days: HR, 2.49 (95% CI, ) 8 days to 3 yrs: HR, 0.34 (95% CI, ) 3 to 5 yrs: HR, 0.60 (95% CI, ) Years after randomization B Cumulative incidence (%) to 7 days: no events 8 days to 3 yrs: HR, 0.79 (95% CI, ) 3 to 5 yrs: HR, 1.35 (95% CI, ) Years after randomization 17

18 C Cumulative incidence (%) to 7 days: HR, 9.01 (95% CI, ) 8 days to 3 yrs: HR, 0.56 (95% CI, ) 3 to 5 yrs: HR, 0.39 (95% CI, ) Years after randomization D Cumulative incidence (%) to 7 days: HR, 0.74 (95% CI, ) 8 days to 3 yrs: HR, 0.24 (95% CI, ) 3 to 5 yrs: HR, 0.29 (95% CI, ) Years after randomization Cumulative incidences of the primary endpoint (A), death (B), myocardial infarction (C) and urgent revascularization (D) in the two randomized groups, stratified on the basis of landmark points at 7 days and 3 years after randomization (vertical dashed lines). Hazard ratios for versus medical therapy were calculated separately for events that occurred within 7 days, between day 8 and 3 years and between 3 and 5 years of follow-up. Patients who experienced an event during the first seven days are not included in the subsequent periods, and patients who experienced an event prior to 3 years are not included in the period between 3 and 5 years. The insets show the data for days 0 to 7 on an expanded 18

19 y axis. For the primary endpoint, p-values were p<0.001 for the interaction between the first and second period, and p=0.13 for the interaction between the second and third period. P-values were p=0.002 and p=0.48 for myocardial infarction, and p=0.14 and p=0.74 for urgent revascularization for the interaction between the first and second time period, and for the interaction between the second and third period, respectively. No deaths had occurred during the first period, and the p-value for interaction between the second and third period was p=0.38. Hazard ratios below 1.00 denote a lower incidence of events in the group than in the medical therapy group. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 19

20 Figure S9: Subgroup analyses for the primary composite endpoint for the two randomized groups. MT Hazard ratio (95% CI) P value for interaction Age, years 0.81 >60 41 (14.5) 76 (27.2) 0.48 ( ) (12.7) 43 (26.5) 0.44 ( ) Gender 0.70 Male 50 (14.0) 94 (27.8) 0.45 ( ) Female 12 (13.2) 25 (24.3) 0.52 ( ) Diabetes 0.78 Yes 23 (18.7) 40 (34.2) 0.49 ( ) No 39 (12.0) 79 (24.4) 0.45 ( ) History of stroke/tia 0.53 Yes 5 (15.2) 11 (39.3) 0.33 ( ) No 57 (13.8) 108 (26.2) 0.48 ( ) History of MI 0.60 Yes 30 (18.3) 53 (32.1) 0.51 ( ) No 32 (11.3) 66 (23.9) 0.43 ( ) History of 0.49 Yes 15 (18.8) 22 (28.9) 0.57 ( ) No 47 (12.8) 97 (26.6) 0.43 ( ) LVEF % 14 (16.9) 20 (35.7) 0.38 ( ) >50% 48 (13.2) 99 (25.7) 0.47 ( ) FFR 0.15 < (14.3) 66 (32.2) 0.37 ( ) (13.5) 53 (22.5) 0.58 ( ) Diameter stenosis % 51 (14.7) 92 (29.0) 0.45 ( ) <70% 11 (10.9) 27 (21.8) 0.46 ( ) Subgroup analyses for the primary endpoint. Presented are numbers (percentages) of primary endpoint events in the two randomized groups and hazard ratios with corresponding 95% confidence intervals from Mantel-Cox comparisons stratified according to different baseline characteristics. Interaction P values are from tests of heterogeneity between strata. Hazard ratios below one denote lower risks of the primary endpoint after compared to medical therapy alone. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 20

21 Figure S10: Caterpillar plot showing the variation in risk ratios of urgent revascularization across trial sites. RR (95% CI) 0.08 (0.00, 1.36) 0.12 (0.01, 2.01) 0.14 (0.02, 0.99) 0.18 (0.04, 0.74) 0.19 (0.05, 0.82) 0.22 (0.07, 0.71) 0.29 (0.09, 0.97) 0.30 (0.03, 3.49) 0.32 (0.04, 2.90) 0.33 (0.02, 6.37) 0.33 (0.01, 7.65) 0.45 (0.05, 4.16) 0.49 (0.18, 1.31) 0.50 (0.05, 4.90) 0.50 (0.05, 4.94) 0.56 (0.06, 5.14) 0.65 (0.12, 3.63) 0.91 (0.07, 12.69) 1.00 (0.10, 9.61) 1.25 (0.11, 14.34) 2.57 (0.13, 52.12) 3.43 (0.16, 71.36) Risk Ratio (RR) Caterpillar plot showing the variation in risk ratios of urgent revascularization across trial sites in the 22 sites with patients who had experienced an urgent revascularisation. The heterogeneity statistic I 2 for the heterogeneity between sites was 0.0% on a scale from 0 to 100%, with a p for heterogeneity between sites of 0.93, suggesting that the variation between sites was not over and above of what would be expected by chance alone. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 21

22 IV. Supplementary Tables Table S1: List of hospitals that participated in the 5-year follow-up and respective enrolment in the randomized groups and registry cohort. Hospital Patients enrolled (n) Gottsegen Hungarian Institute of Cardiology 115 Clinical Center Kragujevac 114 Heart Center Leipzig- University Leipzig 109 Catharina Hospital 78 Södersjukhuset 77 Cardiovascular Center Aalst 76 Cardiovascular Hospital - Hospices Civils de Lyon 76 Masaryk University and University Hospital Brno 65 Stanford University and Palo Alto VA 45 Golden Jubilee National Hospital 37 Rigshospitalet University Hospital 35 Atlanta VA Medical Center 28 Örebro University Hospital 25 Klinikum der Universitat München 14 Isala Klinieken 11 Emory University 10 Na Homolce Hospital 6 Tulane University 6 University Hospital of Ferrara 4 22

23 Table S2: Baseline clinical, angiographic, and FFR characteristics of randomized and registry patients. Patient characteristics Randomized trial Medical therapy Registry Number of patients (n = 447) (n = 441) (n = 166) Baseline characteristics P value comparing randomized trial with registry Age, years 63.5± ± ± Age> (63.1) 279 (63.3) 103 (62.0) 0.78 Male 356 (79.6) 338 (76.6) 113 (68.1) BMI 28.3± ± ± Family history of CAD 216 (48.4) 207 (46.9) 76 (45.8) 0.65 Current smoker 89 (19.9) 90 (20.4) 35 (21.1) 0.79 Hypertension 347 (77.6) 343 (77.8) 136 (81.9) 0.23 Hypercholesterolemia 330 (73.8) 348 (78.9) 118 (71.1) 0.15 Diabetes mellitus 123 (27.5) 117 (26.5) 42 (25.3) 0.65 Insulin requiring diabetes 39 (8.7) 39 (8.8) 10 (6.0) 0.24 Renal insufficiency (creatinine >2.0 mg/dl) 8 (1.8) 12 (2.7) 7 (4.2) 0.14 Peripheral vascular disease 43 (9.6) 47 (10.7) 8 (4.8) History of stroke/tia 33 (7.4) 28 (6.3) 10 (6.0) 0.69 History of MI 164 (36.7) 165 (37.4) 60 (36.1) 0.82 History of in target vessel 80 (17.9) 76 (17.2) 34 (20.5) 0.37 Angina Class 0.64 No angina/ Asymptomatic 53 (11.9) 46 (10.5) 17 (10.2) 23

24 CCS class I 82 (18.3) 98 (22.3) 42 (25.3) CCS class II 204 (45.6) 197 (44.8) 74 (44.6) CCS class III 80 (17.9) 65 (14.8) 23 (13.9) CCS class IV 28 (6.3) 34 (7.7) 10 (6.0) Silent Ischemia 73 (16.3) 73 (16.6) 27 (16.3) 0.96 Left ventricular ejection fraction<50% 83 (18.6) 56 (12.7) 27 (16.3) 0.84 Angiographic findings Significant lesions per patient 1.9± ± ±0.6 <0.001 Vessels with significant lesions per patient < vessel 251 (56.2) 261 (59.2) 136 (81.9) 2 vessels 156 (34.9) 146 (33.1) 26 (15.7) 3 vessels 40 (8.9) 34 (7.7) 4 (2.4) Significant lesion in proximal or middle LAD 291 (65.1) 276 (62.6) 74 (44.6) <0.001 FFR findings Significant lesions per patient 1.5± ± ±0.2 <0.001 Vessels with significant lesions per patient vessel 331 (74.0) 343 (77.8) 5 (3.0) 2 vessels 102 (22.8) 85 (19.3) 0 (0.0) 3 vessels 14 (3.1) 13 (2.9) 0 (0.0) Significant lesion in proximal or middle LAD 279 (62.4) 263 (59.6) 1 (0.6) <0.001 Lesion characteristics Number of lesions (n = 890) (n = 815) (n = 241) Angiographic findings 24

25 Lesions with stenosis of >50% of diameter 837 (94.0) 764 (93.7) 219 (90.9) 0.13 Stenosis <0.001 <50% of diameter 53 (6.0) 51 (6.3) 22 (9.1) 50 69% of diameter 317 (35.6) 332 (40.7) 177 (73.4) 70 90% of diameter 383 (43.0) 330 (40.5) 37 (15.4) >90% of diameter 101 (11.3) 80 (9.8) 0 (0.0) Total occlusion 36 (4.0) 22 (2.7) 5 (2.1) FFR findings Lesions with FFR (76.3) 625 (76.7) 5 (2.1) <0.001 Mean FFR in lesions with FFR ± ± ± Presented are means±sd or numbers (percentages) in the two randomly assigned groups and in the registry. P values compare the combined groups that underwent randomization and the registry and are from Student s t test, two-sided chisquared or Fisher s exact test for patient characteristics and from mixed-effects regression models to account for the correlation of lesions within patients for lesion characteristics. Differences to previously published characteristics are due to reclassifications caused by additional data cleaning and validation. 25

26 Table S3: Baseline clinical, angiographic, and FFR characteristics of randomized patients from centers participating in the 5-year follow-up. Patient characteristics Randomized trial Medical therapy Number of patients (n = 389) (n = 395) Baseline characteristics Age, years 63.9± ±9.3 Age> (64.3) 257 (65.1) Male 297 (76.3) 318 (80.5) BMI 28.2± ±4.3 Family history of CAD 178 (45.8) 190 (48.2) Current smoker 80 (20.6) 78 (19.7) Hypertension 301 (77.4) 309 (78.2) Hypercholesterolemia 300 (77.1) 282 (71.4) Diabetes mellitus 102 (26.2) 112 (28.4) Insulin requiring diabetes 34 (8.7) 34 (8.6) Renal insufficiency (creatinine >2.0 mg/dl) 10 (2.6) 8 (2.0) Peripheral vascular disease 41 (10.5) 38 (9.6) History of stroke/tia 26 (6.7) 31 (7.8) History of MI 144 (37.0) 151 (38.2) History of in target vessel 64 (16.5) 72 (18.2) Angina Class No angina/ Asymptomatic 41 (10.6) 49 (12.4) CCS class I 76 (19.6) 73 (18.5) CCS class II 179 (46.1) 172 (43.5) CCS class III 58 (14.9) 73 (18.5) CCS class IV 34 (8.8) 28 (7.1) Silent Ischemia 65 (16.7) 67 (17.0) Left ventricular ejection fraction<50% 49 (12.6) 78 (19.7) Angiographic findings Significant lesions per patient 1.7± ±1.1 Vessels with significant lesions per patient 26

27 1 vessel 253 (65.0) 223 (56.5) 2 vessels 109 (28.0) 135 (34.2) 3 vessels 27 (6.9) 37 (9.4) Significant lesion in proximal or middle LAD 249 (64.0) 255 (64.6) FFR findings Significant lesions per patient 1.4± ±0.8 Vessels with significant lesions per patient 1 vessel 320 (82.3) 291 (73.7) 2 vessels 62 (15.9) 91 (23.0) 3 vessels 7 (1.8) 13 (3.3) Significant lesion in proximal or middle LAD 237 (60.9) 243 (61.5) Lesion characteristics Number of lesions (n = 726) (n = 788) Angiographic findings Lesions with stenosis of >50% of diameter 677 (93.3) 739 (93.8) Stenosis <50% of diameter 49 (6.7) 49 (6.2) 50 69% of diameter 300 (41.3) 274 (34.8) 70 90% of diameter 291 (40.1) 345 (43.8) >90% of diameter 64 (8.8) 86 (10.9) Total occlusion 22 (3.0) 34 (4.3) FFR findings Lesions with FFR (77.0) 598 (75.9) Mean FFR in lesions with FFR ± ±0.1 Presented are means±sd or numbers (percentages) in the two randomly assigned groups. P values are from Student s t test, two-sided chi-squared or Fisher s exact test for patient characteristics and from mixed-effects regression models to account for the correlation of lesions within patients for lesion characteristics. 27

28 Table S4: Baseline clinical, angiographic, and FFR characteristics of randomized patients from centers not participating in the 5-year follow-up. Patient characteristics Randomized trial Medical therapy Number of patients (n = 52) (n = 52) Baseline characteristics Age, years 63.6± ±9.6 Age>60 29 (55.8) 25 (48.1) Male 41 (78.8) 38 (73.1) BMI 30.0± ±4.2 Family history of CAD 29 (55.8) 26 (50.0) Current smoker 10 (19.2) 11 (21.2) Hypertension 42 (80.8) 38 (73.1) Hypercholesterolemia 48 (92.3) 48 (92.3) Diabetes mellitus 15 (28.8) 11 (21.2) Insulin requiring diabetes 5 (9.6) 5 (9.6) Renal insufficiency (creatinine >2.0 mg/dl) 2 (3.8) 0 (0.0) Peripheral vascular disease 6 (11.5) 5 (9.6) History of stroke/tia 2 (3.8) 2 (3.8) History of MI 21 (40.4) 13 (25.0) History of in target vessel 12 (23.1) 8 (15.4) Angina Class No angina/ Asymptomatic 5 (9.6) 4 (7.7) CCS class I 22 (42.3) 9 (17.3) CCS class II 18 (34.6) 32 (61.5) CCS class III 7 (13.5) 7 (13.5) CCS class IV 0 (0.0) 0 (0.0) Silent Ischemia 8 (15.4) 6 (11.5) Left ventricular ejection fraction<50% 7 (13.5) 5 (9.6) Angiographic findings Significant lesions per patient 1.7± ±0.9 Vessels with significant lesions per patient 28

29 1 vessel 29 (55.8) 28 (53.8) 2 vessels 22 (42.3) 21 (40.4) 3 vessels 1 (1.9) 3 (5.8) Significant lesion in proximal or middle LAD 27 (51.9) 36 (69.2) FFR findings Significant lesions per patient 1.3± ±0.8 Vessels with significant lesions per patient 1 vessel 42 (80.8) 40 (76.9) 2 vessels 10 (19.2) 11 (21.2) 3 vessels 0 (0.0) 1 (1.9) Significant lesion in proximal or middle LAD 26 (50.0) 36 (69.2) Lesion characteristics Number of lesions (n = 89) (n = 102) Angiographic findings Lesions with stenosis of >50% of diameter 87 (97.8) 98 (96.1) Stenosis <50% of diameter 2 (2.2) 4 (3.9) 50 69% of diameter 32 (36.0) 43 (42.2) 70 90% of diameter 39 (43.8) 38 (37.3) >90% of diameter 16 (18.0) 15 (14.7) Total occlusion 0 (0.0) 2 (2.0) FFR findings Lesions with FFR (74.2) 81 (79.4) Mean FFR in lesions with FFR ± ±0.1 Presented are means±sd or numbers (percentages) in the two randomly assigned groups. P values are from Student s t test, two-sided chi-squared or Fisher s exact test for patient characteristics and from mixed-effects regression models to account for the correlation of lesions within patients for lesion characteristics. 29

30 Table S5: Baseline clinical, angiographic, and FFR characteristics of patients according to site participation in the 5-year follow-up. Site participated in 5-year follow-up Site did not participate in 5-year follow-up P value Patient characteristics Number of patients (n = 931) (n = 123) Baseline characteristics Age, years 63.7± ± Age> (64.0) 68 (55.3) Male 715 (76.8) 92 (74.8) 0.62 BMI 28.1± ± Family history of CAD 435 (46.8) 64 (52.0) 0.27 Current smoker 192 (20.6) 22 (17.9) 0.48 Hypertension 730 (78.4) 96 (78.0) 0.93 Hypercholesterolemia 683 (73.4) 113 (91.9) <0.001 Diabetes mellitus 252 (27.1) 30 (24.4) 0.53 Insulin requiring diabetes 78 (8.4) 10 (8.1) 0.93 Renal insufficiency (Creatinine > 2.0 mg/dl) 25 (2.7) 2 (1.6) 0.49 Peripheral vascular disease 87 (9.3) 11 (8.9) 0.89 History of stroke/tia 64 (6.9) 7 (5.7) 0.62 History of MI 351 (37.7) 38 (30.9) 0.14 History of in target vessel 168 (18.0) 22 (17.9)

31 Angina Class No angina/ Asymptomatic 106 (11.4) 10 (8.1) CCS class I 184 (19.8) 38 (30.9) CCS class II 418 (44.9) 57 (46.3) CCS class III 150 (16.1) 18 (14.6) CCS class IV 72 (7.7) 0 (0.0) Silent Ischemia 158 (17.0) 15 (12.2) 0.18 Left ventricular ejection fraction<50% 152 (16.3) 14 (11.4) 0.16 Angiographic findings Significant lesions per patient 1.7± ± Vessels with significant lesions per patient vessel 621 (66.7) 73 (59.3) 2 vessels 255 (27.4) 45 (36.6) 3 vessels 53 (5.7) 5 (4.1) Significant lesion in proximal or middle LAD 567 (60.9) 74 (60.2) 0.88 FFR findings Significant lesions per patient 1.2± ± Vessels with significant lesions per patient vessel 642 (69.0) 82 (66.7) 2 vessels 137 (14.7) 21 (17.1) 3 vessels 10 (1.1) 1 (0.8) Significant lesion in proximal or middle LAD 481 (51.7) 62 (50.4) 0.79 Lesion characteristics 31

32 Number of lesions (n = 1730) (n = 216) Angiographic findings Lesions with stenosis of >50% of diameter 1601 (93.1) 209 (96.8) Stenosis <50% of diameter 119 (6.9) 7 (3.2) 50 69% of diameter 723 (42.0) 93 (43.1) 70 90% of diameter 667 (38.8) 83 (38.4) >90% of diameter 150 (8.7) 31 (14.4) Total occlusion 61 (3.5) 2 (0.9) FFR findings Lesions with FFR (67.2) 147 (68.1) 0.85 Mean FFR in lesions with FFR ± ± Presented are means±sd or numbers (percentages) for patients from sites that participated in the 5-year follow-up and from sites that did not participate in the 5-year follow-up. P values compare these two groups and are from Student s t test, twosided chi-squared or Fisher s exact test for patient characteristics and from mixed-effects regression models to account for the correlation of lesions within patients for lesion characteristics. 32

33 Table S6: Overview of patient flow according to site participation in the 5-year follow-up. Randomized trial Medical therapy Registry Site participated in 5-year follow-up Follow up information available through 5 years 371 (93.9%) 362 (93.1%) 133 (90.5%) Followed up and alive 350 (88.6%) 341 (87.7%) 127 (86.4%) Deceased 21 (5.3%) 21 (5.4%) 6 (4.1%) Follow up information unavailable through 5 years 24 (6.1%) 27 (6.9%) 14 (9.5%) Withdrew consent 9 (2.3%) 6 (1.5%) 1 (0.7%) Lost to follow up 15 (3.8%) 21 (5.4%) 13 (8.8%) Site did not participate in 5-year follow-up Follow up information available through 3 years 46 (88.5%) 44 (84.6%) 15 (79.0%) Followed up and alive 44 (84.6%) 42 (80.8%) 14 (73.7%) Deceased 2 (3.9%) 2 (3.9%) 1 (5.3%) Follow up information unavailable through 3 years 6 (11.5%) 8 (15.4%) 4 (21.1%) Withdrew consent 1 (1.9%) 2 (3.9%) 1 (5.3%) Lost to follow up 5 (9.6%) 6 (11.5%) 3 (15.8%) 33

34 Table S7: Cardiac medications at baseline, and at 1 month, 6 months, 1 year, 2 years, 3 years and 5 years of followup. Randomized trial Medical therapy n (%) n (%) P value comparing randomized groups n (%) Registry P value comparing trial with registry Baseline n = 447 n = 441 n = 166 Aspirin 390 (87%) 396 (90%) (87%) 0.51 P2Y12 inhibitors 220 (49%) 198 (45%) (43%) 0.40 Statin 370 (83%) 361 (82%) (77%) 0.13 Beta blocker 338 (76%) 344 (78%) (73%) 0.37 ACE-inhibitors/AT-II antagonist 308 (69%) 309 (70%) (77%) 0.08 Calcium channel blocker 104 (23%) 101 (23%) (30%) Month n = 442 n = 434 n = 162 Aspirin 437 (99%) 417 (96%) (92%) P2Y12 inhibitors 432 (98%) 194 (45%) < (35%) <0.001 Statin 422 (95%) 417 (96%) (95%) 0.68 Beta blocker 376 (85%) 392 (90%) (80%) ACE-inhibitors/AT-II antagonist 355 (80%) 354 (82%) (79%) 0.59 Calcium channel blocker 107 (24%) 143 (33%) (34%) Months n = 441 n = 435 n = 163 Aspirin 435 (99%) 412 (95%) (93%) P2Y12 inhibitors 425 (96%) 211 (49%) < (31%) <0.001 Statin 425 (96%) 414 (95%) (94%) 0.41 Beta blocker 358 (81%) 367 (84%) (79%) 0.22 ACE-inhibitors/AT-II antagonist 358 (81%) 360 (83%) (80%) 0.66 Calcium channel blocker 114 (26%) 141 (32%) (38%) Year n = 437 n = 429 n = 160 Aspirin 421 (96%) 407 (95%) (89%) P2Y12 inhibitors 294 (67%) 189 (44%) < (26%) <

35 Statin 417 (95%) 405 (94%) (94%) 0.56 Beta blocker 349 (80%) 349 (81%) (76%) 0.24 ACE-inhibitors/AT-II antagonist 347 (79%) 362 (84%) (79%) 0.38 Calcium channel blocker 119 (27%) 138 (32%) (40%) Years n = 427 n = 424 n = 158 Aspirin 397 (93%) 393 (93%) (88%) 0.05 P2Y12 inhibitors 86 (20%) 106 (25%) (13%) Statin 394 (92%) 392 (92%) (91%) 0.42 Beta blocker 330 (77%) 340 (80%) (75%) 0.35 ACE-inhibitors/AT-II antagonist 344 (81%) 348 (82%) (80%) 0.83 Calcium channel blocker 130 (30%) 135 (32%) (37%) Years n = 422 n = 413 n = 156 Aspirin 383 (91%) 380 (92%) (82%) P2Y12 inhibitors 57 (14%) 73 (18%) (12%) 0.33 Statin 380 (90%) 381 (92%) (89%) 0.29 Beta blocker 308 (73%) 330 (80%) (75%) 0.68 ACE-inhibitors/AT-II antagonist 335 (80%) 342 (83%) (76%) 0.12 Calcium channel blocker 126 (30%) 140 (34%) (41%) Years n = 350 n = 341 n = 127 Aspirin 298 (84%) 302 (89%) (81%) 0.10 P2Y12 inhibitors 45 (13%) 57 (17%) (12%) 0.41 Statin 311 (88%) 307 (91%) (84%) 0.13 Beta blocker 259 (74%) 261 (77%) (72%) 0.44 ACE-inhibitors/AT-II antagonist 262 (74%) 267 (80%) (69%) 0.06 Calcium channel blocker 98 (28%) 106 (31%) (39%) 0.04 : percutaneous coronary intervention, ACE: angiotensin converting enzyme, AT-II: angiotensin II. Presented are numbers (percentages) of patients taking cardiac medications at different time points in the different groups. P values are from two-sided chi-squared or Fisher s exact test, as appropriate. 35

36 Table S8: Causes of death by treatment group. Randomized trial Medical therapy Registry (n = 447) (n = 441) (n = 166) Any cause Cardiac death Heart failure Myocardial infarction 5 2 Unknown cause of death (classified as cardiac death) 2 2 Vascular death 3 4 Stroke 2 1 Pulmonary embolism 1 1 Acute mesenteric ischemia 1 Lower limb ischemia 1 Cancer Pancreatic cancer 2 Gastric cancer 1 Rectal cancer 1 Urinary bladder cancer 1 Lung cancer 4 Brain cancer 1 Prostate cancer 1 36

37 Adrenal gland cancer 1 Acute myeloid leukemia 1 Pharyngeal cancer 1 Unknown primary origin 1 Road accident 1 Acute pancreatitis 1 Pneumonia Bowel perforation 1 37

38 Table S9: Types of myocardial infarction and additional composite endpoints at 5-year follow-up. Medical therapy Hazard ratio (95% CI) Registry n=447 n=441 n=166 ST elevation myocardial infarction 7 (1.6) 7 (1.6) 0.98 ( ) 4 (2.4) Non ST elevation myocardial infarction 30 (6.7) 45 (10.2) 0.64 ( ) 11 (6.6) Unknown type of myocardial infarction 1 (0.2) 3 (0.7) 0.33 ( ) 2 (1.2) Spontaneous myocardial infarction 29 (6.5) 45 (10.2) 0.62 ( ) 13 (7.8) Periprocedural myocardial infarction 7 (1.6) 9 (2.0) 0.77 ( ) 1 (0.6) Cardiac death or myocardial infarction 43 (9.6) 59 (13.4) 0.70 ( ) 16 (9.6) Death, myocardial infarction or any revascularization 91 (20.4) 239 (54.2) 0.28 ( ) 39 (23.5) : percutaneous coronary intervention. Hazard ratios and 95% confidence intervals were calculated with the log-rank test for the group compared with the medical therapy group. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 38

39 Table S10: Subgroup analyses according to participation in the 5-year follow-up for the primary endpoint and its components for the two randomized groups. P value for Medical therapy Hazard ratio (95% CI) interaction Primary endpoint n=447 n=441 Center did not participate at the 5-year follow-up 7 (13.5) 8 (15.4) 0.69 ( ) 0.40 Center participated at the 5-year follow-up 55 (13.9) 111 (28.5) 0.44 ( ) All-cause mortality Center did not participate at the 5-year follow-up 2 (3.8) 2 (3.8) 0.49 ( ) 0.57 Center participated at the 5-year follow-up 21 (5.3) 21 (5.4) 0.99 ( ) Myocardial infarction Center did not participate at the 5-year follow-up 4 (7.7) 4 (7.7) 0.88 ( ) 0.63 Center participated at the 5-year follow-up 32 (8.1) 49 (12.6) 0.63 ( ) Urgent revascularization Center did not participate at the 5-year follow-up 4 (7.7) 6 (11.5) 0.44 ( ) 0.39 Center participated at the 5-year follow-up 24 (6.1) 87 (22.4) 0.24 ( ) Subgroup analyses for the primary endpoint and its components. Presented are numbers (percentages) of events in the two randomized groups and hazard ratios with corresponding 95% confidence intervals from Mantel-Cox comparisons stratified according to center participation at the 5-year follow-up. Interaction P values are from tests of heterogeneity between strata. Hazard ratios below one denote lower risks of events after compared to medical therapy alone. The 95% confidence intervals for secondary endpoints were not adjusted for multiple testing and any inferences drawn from the intervals as reported may not be reproducible. 39