Bivalirudin or unfractionated heparin in patients with acute coronary syndromes managed invasively with and without ST elevation (MATRIX): randomised controlled trial

(1)

open access

For numbered affiliations see end of article.

Correspondence to: M Valgimigli marco.valgimigli@insel.ch Additional material is published online only. To view please visit the journal online.

Cite this as: BMJ 2016;354:i4935

http://dx.doi.org/10.1136/bmj.i4935 Accepted: 2 September 2016

Bivalirudin or unfractionated heparin in patients with acute

coronary syndromes managed invasively with and without

ST elevation (MATRIX): randomised controlled trial

Sergio Leonardi,

1

_{Enrico Frigoli,}

2

_{Martina Rothenbühler,}

3,4

_{Eliano Navarese,}

5

_{Paolo Calabró,}

6

Paolo Bellotti,

7

_{Carlo Briguori,}

8

_{Marco Ferlini,}

1

_{Bernardo Cortese,}

9

_{Alessandro Lupi,}

10

_{Salvatore Lerna,}

11

Dennis Zavallonito-Parenti,

12

_{Giovanni Esposito,}

13

_{Simone Tresoldi,}

14

_{Antonio Zingarelli,}

15

Stefano Rigattieri,

16

_{Cataldo Palmieri,}

17

_{Armando Liso,}

18

_{Fabio Abate,}

19

_{Marco Zimarino,}

20

Marco Comeglio,

21

_{Gabriele Gabrielli,}

22

_{Alaide Chieffo,}

23

_{Salvatore Brugaletta,}

24

_{Ciro Mauro,}

25

Nicolas M Van Mieghem,

26

_{Dik Heg,}

3,4

_{Peter Jüni,}

27

_{Stephan Windecker,}

28

_{Marco Valgimigli}

28

for the MATRIX Investigators

ABSTRACT

ObjeCtive

To test the optimal antithrombotic regimen in patients with acute coronary syndrome.

Design

andomised controlled trial. setting

Patients with acute coronary syndrome with and without ST segment elevation in 78 centres in Italy, the Netherlands, Spain, and Sweden.

PartiCiPants

7213 patients with acute coronary syndrome and planned percutaneous coronary intervention: 4010 with ST segment elevation and 3203 without ST segment elevation. The primary study results in the overall population have been reported previously.

interventiOns

Patients were randomly assigned, in an open label fashion, to one of two regimens: bivalirudin with glycoprotein IIb/IIIa inhibitors restricted to procedural complications or heparin with or without glycoprotein IIb/IIIa inhibitors.

Main OutCOMe Measures

Primary endpoints were the occurrence of major adverse cardiovascular events, defined as death, myocardial infarction or stroke; and net adverse clinical events, defined as major bleeding or major adverse cardiovascular events, both assessed at 30 days. Analyses were performed by the principle of intention to treat.

results

Use of a glycoprotein IIb/IIIa inhibitor in patients assigned to heparin was planned at baseline in 30.7% of patients with ST segment elevation, in 10.9% without ST segment elevation, and in no patients assigned to bivalirudin. In patients with ST segment elevation, major adverse cardiovascular events occurred in 118 (5.9%) assigned to bivalirudin and 129 (6.5%) assigned to heparin (rate ratio 0.90, 95% confidence interval 0.70 to 1.16; P=0.43), whereas net adverse clinical events occurred in 139 (7.0%) patients assigned to bivalirudin and 163 (8.2%) assigned to heparin (0.84, 0.67 to 1.05; P=0.13). In patients without ST segment elevation, major adverse cardiovascular events occurred in 253 (15.9%) assigned to bivalirudin and 262 (16.4%) assigned to heparin (0.97, 0.80 to 1.17; P=0.74), whereas net adverse clinical events occurred in 262 (16.5%) patients assigned to bivalirudin and 281 (17.6%) assigned to heparin (0.93, 0.77 to 1.12; P=0.43). COnClusiOns

A bivalirudin monotherapy strategy compared with heparin with or without glycoprotein IIb/IIIa inhibitors, did not result in reduced major adverse cardiovascular events or net adverse clinical events in patients with or without ST segment elevation.

trial registratiOn

ClinicalTrials.gov NCT01433627. Introduction

In patients with acute coronary syndromes managed invasively with coronary angiography and revasculari-sation the optimal strategy for preventing coronary thrombosis and ischaemia, while limiting bleeding, is uncertain. Two commonly adopted regimens consist of unfractionated heparin, an indirect thrombin inhibitor, with or without concomitant use of a glycoprotein IIb/ IIIa inhibitor in case of percutaneous coronary inter-vention; and bivalirudin, a direct thrombin inhibitor, with the use of glycoprotein IIb/IIIa inhibitors restricted to periprocedural thrombotic complications. This latter strategy has been associated with reduced bleeding, an outcome extensively associated with mortality. Previ-ous moderately sized studies in patients with acute cor-onary syndrome provided conflicting results on the

WhAT IS AlReAdy knoWn on ThIS TopIC

The optimal antithrombotic regimen in patients with acute coronary syndrome with and without ST segment elevation is not yet defined

WhAT ThIS STudy AddS

A regimen of bivalirudin alone compared with unfractionated heparin with provisional use of glycoprotein IIb/IIIa inhibitors in patients with acute coronary syndrome with or without ST segment elevation, did not result in reduced major adverse cardiovascular events or net adverse clinical events

A strategy of bivalirudin monotherapy was associated with reduced bleeding compared with heparin with provisional use of glycoprotein IIb/IIIa inhibitors

(2)

comparative efficacy and safety of these two strategies.1-4_{We therefore designed the Minimizing} Adverse Haemorrhagic Events by Transradial Access Site and Systemic Implementation of Angiox (MATRIX) phase 3 programme in patients with acute coronary syndrome managed invasively via the radial or femoral route.5_{In this large, randomised controlled trial we} assessed whether bivalirudin is superior to unfraction-ated heparin with or without concomitant use of glyco-protein IIb/IIIa inhibitors. Bivalirudin did not significantly reduce the two coprimary endpoints of major adverse cardiovascular events and net adverse clinical events in the overall acute coronary syndrome population of 7213 patients.6

However, the use of these two strategies in clinical practice,7_{as well as the profile of efficacy on ischaemic} and thrombotic complications (eg, stent thrombosis), and safety in terms of bleeding8_{might differ depending} on the type of acute coronary syndrome at presentation— that is, with or without persistent ST segment elevation.

We present the results of prespecified, randomised comparisons of bivalirudin with heparin in two MATRIX subpopulations with acute coronary syndrome: 4010 patients presenting with persistent ST elevation and 3203 patients presenting without persistent ST segment elevation for whom percutaneous coronary interven-tion was planned at the time of coronary angiography. Methods

study design

MATRIX was a programme of three independent ran-domised controlled trials in an all comers population with acute coronary syndrome, with or without per-sistent ST segment elevation. The first trial, MATRIX Access Site, compared transradial access with trans-femoral access in 8404 patients with acute coronary syndrome, 4010 (47.7%) with ST segment elevation and 4394 (52.3%) without ST segment elevation.9_The sec-ond trial, MATRIX Antithrombin, was a randomised comparison of two antithrombotic strategies: bivaliru-din with use of glycoprotein IIb/IIIa inhibitors restricted to angiographic complications (eg, giant thrombus) compared with unfractionated heparin with use of gly-coprotein IIb/IIIa inhibitors left to the discretion of the investigator.6_{These strategies were compared in the} overall population with ST segment elevation (n=4010) and in patients without ST segment elevation for whom percutaneous coronary intervention was planned (n=3203, 72.9% of the overall population without ST segment elevation). The third trial, MATRIX Treatment Duration, was a randomised comparison within patients assigned to bivalirudin, comparing extended bivalirudin administration after percutaneous coronary intervention with short term administration during per-cutaneous coronary intervention only.6

Participants

Patients with ST segment elevation were eligible: 1) if they experienced more than 20 minutes of ischaemic symptoms with ST segment elevation of ≥1 mm in two or more contiguous electrocardiogram leads, or with a

new left bundle branch block, or 2) in case of ST seg-ment depression of ≥1 mm in two or more of leads V1-V3 with a positive terminal T wave if presented within 12 hours of symptom onset, or 3) if there was evidence of continuing ischaemia or previous fibrinolytic treatment between 12 and 24 hours after symptom onset. We use the term acute coronary syndrome with ST segment ele-vation instead of ST eleele-vation myocardial infarction to recognise the possibility of an aborted myocardial infarction in patients presenting with acute coronary syndrome and persistent ST segment elevation.10

Patients without ST segment elevation were eligible if they had a history consistent with new or worsening cardiac ischaemia, occurring at rest or with minimal activity within seven days before randomisation, and fulfilled at least two high risk criteria among the follow-ing: age 60 years or older, increase in cardiac biomark-ers, or electrocardiographic changes consistent with cardiac ischaemia; consideration as a possible candi-date for percutaneous coronary intervention after com-pletion of coronary angiography.

Key exclusion criteria (for patients both with and without ST segment elevation) were treatment with low molecular weight heparins within the past six hours; treatment with any glycoprotein IIb/IIIa inhibitor in the previous three days; contraindications to angiography, including but not limited to severe peripheral vascular disease; and presumed life expectancy of less than 30 days. Detailed inclusion and exclusion criteria have been published elsewhere and are listed in the appendix.12_{All patients gave written informed consent.} randomisation

Using a computer generated random sequence, we ran-domised patients in a 1:1 ratio to receive the bivalirudin strategy or unfractionated heparin strategy, with a ran-dom block size stratified by type of acute coronary syn-drome—ie, with ST segment elevation versus without ST segment elevation—intended or ongoing use of P2Y12 inhibitor (clopidogrel versus ticagrelor or prasugrel), and study site.

Owing to the high likelihood of patients with ST seg-ment elevation receiving percutaneous coronary inter-vention, we simultaneously randomised all patients with ST segment elevation to the comparisons of access site, antithrombin, and treatment duration (one step randomisation process). To account for the patients without ST segment elevation who would not undergo percutaneous coronary intervention after angiography and to minimise the risk that this might unbalance the distribution of patients for the tested drug options, we also allowed randomisation to the antithrombin com-parison (either bivalirudin or heparin strategy) after coronary angiography and the decision had been taken to proceed to percutaneous coronary intervention. Hence, a two step randomisation process (randomisa-tion to access site before angiography followed by ran-domisation to antithrombin, including ranran-domisation to treatment duration, only in those undergoing percu-taneous coronary intervention) was selected in most of the patients without ST segment elevation. Additionally,

(3)

randomisation of patients without ST segment eleva-tion for which a two step randomisaeleva-tion was selected by the investigator was stratified by the actual treatment strategy (ie, coronary angiography only, not followed by percutaneous coronary intervention; or coronary angi-ography followed by planned percutaneous coronary intervention).

Procedures

We administered all interventions in an open label fash-ion. Bivalirudin was given according to the product labelling, with a bolus of 0.75 mg/kg body weight fol-lowed immediately by an infusion of 1.75 mg/kg body weight hourly until completion of percutaneous coro-nary intervention. Bivalirudin was then stopped at the end of percutaneous coronary intervention, or pro-longed in accordance with the subsequent random assignment. In patients allocated to prolonged treat-ment, bivalirudin could be administered for up to four hours at the full dose or at a reduced dose of 0.25/mg/kg body weight hourly for at least six hours, at the discre-tion of the treating doctors. Heparin was dosed at 70-100 units/kg body weight in patients not receiving glycoprotein IIb/IIIa inhibitors and at 50-70 units/kg body weight in patients receiving glycoprotein IIb/IIIa inhibitors. Subsequent heparin dosing based on acti-vated clotting time was again left to the discretion of the investigator. A glycoprotein IIb/IIIa inhibitor could be administered before percutaneous coronary interven-tion in patients in the heparin group based on investiga-tor’s judgment, but in the bivalirudin group it was restricted only to patients with thrombotic complica-tions at the time of percutaneous coronary intervention, including no reflow or giant thrombus. Use of other drugs was allowed as per guidelines. Clinical follow-up was performed at 30 days.

Outcomes

The MATRIX Antithrombin trial had two coprimary out-comes, identical for patients with and without ST seg-ment elevation: major adverse cardiovascular events and net adverse clinical events, both assessed at 30 days. Considering the extensive association of bleeding with mortality and that both experimental treatments tested in MATRIX (radial access and bivalirudin mono-therapy) were expected to reduce bleeding, we included net adverse clinical events as the coprimary endpoint.

We defined major adverse cardiovascular events as the composite of all cause mortality, myocardial infarc-tion, or stroke; and net adverse clinical events as the composite of major bleeding unrelated to coronary artery bypass graft (Bleeding Academic Research Con-sortium type 3 or 5, also see supplementary file),11_or major adverse cardiovascular events. Secondary comes included each component of the composite out-comes, cardiovascular mortality, and stent thrombosis. Bleeding was also assessed and adjudicated on the basis of the Thrombolysis In Myocardial Infarction (TIMI) and Global Utilization of Streptokinase and Tis-sue Plasminogen Activator for Occluded Coronary Arteries (GUSTO) scales.12

An independent clinical events committee, blinded to randomised treatment allocation, adjudicated all suspected events according to prespecified definitions. The appendix details the definitions for endpoints adju-dicated by the clinical events committee.

statistical analysis

The overall MATRIX Antithrombin trial was designed as a superiority study on two coprimary outcomes: major adverse cardiovascular events and net adverse clinical events. We expected that the incidence of 30 day major adverse cardiovascular events (death, myocardial infarction, or stroke) on an intention to treat basis to be 4.2% in the bivalirudin group and 6.0% in the unfrac-tionated heparin group (relative risk 0.70). With an α error set at 2.5% to correct for the two coprimary end-points, we considered that a sample size of 6800 ran-domised patients or 3400 patients in each group would provide 85% power. For 30 day net adverse clinical events (death, myocardial infarction, stroke, or major bleeding), we expected an incidence on an intention to treat basis of 6.3% in the bivalirudin group and 9.0% in the unfractionated heparin group (relative risk 0.70). We determined that we would need 6800 randomised patients or 3400 patients in each group to provide 95% power with an α error set at 2.5% to correct for the two coprimary endpoints. Otherwise, we made no adjust-ments for multiple comparisons when analysing sec-ondary endpoints. We analysed the data according to the intention to treat principle. Outcomes were assessed separately for patients with and without ST segment elevation as time to first event using the Mantel-Cox method to estimate unadjusted rate ratios and corre-sponding 95% confidence intervals.13_{The proportional} hazard assumption was verified on the basis of Schoen-feld residuals. We used a two sided log-rank test, and survival curves were constructed using Kaplan-Meier estimates. We performed stratified analyses according to prespecified subgroups (centre’s annual volume of percutaneous coronary intervention, age, sex, body mass index, intended start or continuation of prasugrel or ticagrelor, diabetes, estimated glomerular filtration rate, history of peripheral vascular disease, previous heparin, and randomisation to access site) and esti-mated possible effect modifications using interaction terms or tests for trend across ordered groups. We chose not to adjust for centre because the number of observa-tions in each centre varies substantially (range 1-794). However, to account for the within group correlation, we applied a shared frailty model using Cox regression and observed that hazard ratios were similar to our orig-inal Cox regression models where we did not use a shared frailty model. All analyses were performed using the statistical packages Stata 14.1 and R 3.3.0.

Patient involvement

No patients were involved in setting the research ques-tion or the outcome measures, nor were they involved in developing plans for recruitment, design, or implemen-tation of the study. No patients were asked to advise on interpretation or writing up of results. We plan to

(4)

disseminate the results of the research to all the scien-tific community, including study participants. Results

Patients

Between 11 October 2011 and 7 November 2014 a total of 7213 patients were randomised in the MATRIX Anti-thrombin trial in 78 centres in Italy, the Netherlands, Spain, and Sweden. Of these, 4010 (56%) qualified for acute coronary syndrome with ST segment elevation at presentation (randomised after a median of 3.1 hours from symptom onset) and 3203 (44%) for acute coronary syndrome without ST segment elevation (randomised after a median of 36.4 hours from symptom onset, fig 1 ). Of the 4010 patients with ST segment elevation, an inves-tigator reported primary percutaneous coronary inter-vention as the indication for angiography in 3647—that is, 91%. Of the remaining 363 patients with ST segment elevation (9.0%), 190 did not receive percutaneous coro-nary intervention and 173 had percutaneous corocoro-nary intervention for an indication different from primary per-cutaneous coronary intervention. Overall, 3442 patients (95.3%) in the bivalirudin group and 3474 (96.4%) in the heparin group actually received the allocated treatment. At 30 days, complete follow-up information was avail-able for 7198 (99.8%) patients (see supplementary figures S1 and S2). Baseline characteristics were similar within qualifying groups with acute coronary syndrome, with burden of risk factors for atherothrombosis and comor-bidities (including hypertension, diabetes, hypercholes-terolaemia, previous myocardial infarction) higher in patients who qualified for the group without ST segment elevation than those who qualified for the group with ST segment elevation (table 1). Use of a glycoprotein IIb/IIIa inhibitor in patients assigned to heparin was planned at the time of randomisation in 30.7% of patients with ST segment elevation and 10.9% without ST segment eleva-tion. In accordance with the protocol, no patient was assigned to bivalirudin for whom a glycoprotein IIb/IIIa inhibitor was planned at the time of randomisation (see supplementary appendix). Bailout use of a glycoprotein IIb/IIIa inhibitor was not significantly different between randomised groups in patients without ST segment ele-vation and higher in the bivalirudin group (6.0%) com-pared with heparin (4.3%) in the population with ST segment elevation (P=0.014). Administration of an oral P2Y12 inhibitor (clopidogrel, ticagrelor, or prasugrel) before coronary angiography was similar between patients in the bivalirudin group and heparin group, in the populations both with and without ST segment ele-vation (see supplementary table S1). The appendix lists the other procedural characteristics.

Clinical outcomes

In patients with ST segment elevation, major adverse cardiovascular events occurred in 118 (5.9%) assigned to bivalirudin and 129 (6.5%) assigned to heparin (rate ratio 0.90, 95% confidence interval 0.70 to 1.16, P=0.43), whereas net adverse clinical events occurred in 139 (7.0%) assigned to bivalirudin and 163 (8.2%) assigned to heparin (0.84, 0.67 to 1.05, P=0.13, fig 2 and table 2).

Femoral (n=4207 patients) Radial

(n=4197 patients)

Patients without ST segment elevation managed invasively (n=4394)

Access site randomisation

(n=8404 patients) Antithrombin randomisation(n=7213 patients) Patients with ST segment elevation

with planned percutaneous coronary intervention (n=4010)

Patients with ST segment elevation with planned percutaneous coronary intervention (n=4010) Patients without ST segment elevation

without planned percutaneous coronary intervention (n=1191; 27.1%)

Patients without ST segment elevation with planned percutaneous coronary intervention (n=3203; 72.9%) Coronary angiography R 1:1 Heparin with or without glycoprotein IIb/IIIa inhibitors (n=3603 patients) Bivalirudin (n=3610 patients) R 1:1 R 1:1

Bivalirudin long infusion Bivalirudin short infusion

Fig 1 | Flow of study participants

table 1 | baseline characteristics of participants with acute coronary syndrome with and without st segment elevation. values are numbers (percentages) unless stated otherwise Characteristics

With st segment elevation Without st segment elevation bivalirudin

(n=2012) unfractionated heparin (n=1998) bivalirudin (n=1598) unfractionated heparin (n=1605) Mean (SD) age (years) 63.9 (12.2) 63.9 (12.0) 67.4 (11.2) 67.3 (11.4) Age ≥75 years 445 (22.1) 423 (21.2) 461 (28.8) 481 (30.0) Male sex 1549 (77.0) 1544 (77.3) 1182 (74.0) 1220 (76.0) Mean (SD) weight (kg) 77.8 (13.5) 77.4 (14.0) 77.7 (14.0) 77.4 (13.7) Mean (SD) body mass index 27.1 (4.1) 26.9 (4.1) 27.3 (4.3) 27.2 (4.1) Diabetes mellitus 363 (18.0) 355 (17.8) 452 (28.3) 431 (26.9) Insulin dependent 73 (3.6) 79 (4.0) 123 (7.7) 108 (6.7) Current smoker 820 (40.8) 815 (40.8) 487 (30.5) 487 (30.3) Hypercholesterolaemia 783 (38.9) 785 (39.3) 813 (50.9) 773 (48.2) Hypertension 1150 (57.2) 1084 (54.3) 1114 (69.7) 1138 (70.9) Family history of CAD 561 (27.9) 563 (28.2) 430 (26.9) 429 (26.7) Previous MI 188 (9.3) 196 (9.8) 342 (21.4) 305 (19.0) Previous PCI 200 (9.9) 186 (9.3) 336 (21.0) 318 (19.8)

Previous CABG 29 (1.4) 23 (1.2) 98 (6.1) 72 (4.5)

Previous TIA or stroke 88 (4.4) 84 (4.2) 93 (5.8) 101 (6.3) Peripheral vascular disease 106 (5.3) 123 (6.2) 190 (11.9) 161 (10.0)

COPD 100 (5.0) 85 (4.3) 116 (7.3) 135 (8.4)

History of renal failure 21 (1.0) 24 (1.2) 27 (1.7) 23 (1.4)

Dialysis 3 (0.1) 0 (0.0) 2 (0.1) 2 (0.1) Clinical presentation Cardiac arrest 67 (3.3) 72 (3.6) 13 (0.8) 11 (0.7) Killip class: I 1814 (90.2) 1778 (89.0) 1461 (91.4) 1462 (91.1) II 129 (6.4) 157 (7.9) 95 (5.9) 107 (6.7) III 38 (1.9) 29 (1.5) 38 (2.4) 35 (2.2) IV 31 (1.5) 34 (1.7) 4 (0.3) 1 (0.1)

Previous lytic treatment 96 (4.8) 101 (5.1) 1 (0.1) 0 (0.0) Mean (SD) SBP (mm Hg) 136.2 (26.8) 136.0 (27.1) 141.5 (24.6) 140.9 (24.0) Mean (SD) heart rate (beats/min) 77.0 (17.3) 77.1 (17.2) 75.2 (16.3) 74.1 (15.3) Mean (SD) LVEF (%) 48.7 (9.5) 49.1 (9.5) 52.8 (9.1) 53.2 (9.1) Mean (SD) eGFR (mL/min) 84.4 (25.5) 85.2 (25.8) 82.6 (24.5) 83.4 (25.6) eGFR <60 307 (15.4) 301 (15.3) 282 (17.7) 287 (17.9)

eGFR<30 20 (1.0) 25 (1.3) 18 (1.1) 15 (0.9)

CAD=coronary artery disease; MI=myocardial infarction; PCI=percutaneous coronary intervention; CABG=coronary artery bypass graft; TIA=transient ischaemic attack; COPD=chronic obstructive pulmonary disease; SBP=systolic blood pressure; LVEF=left ventricular ejection fraction; eGFR=estimated glomerular filtration rate.

(5)

In patients without ST segment elevation, major adverse cardiovascular events occurred in 253 (15.9%) assigned to bivalirudin and 262 (16.4%) assigned to hep-arin (0.97, 0.80 to 1.17, P=0.72), and net adverse clinical events occurred in 262 (16.5%) assigned to bivalirudin and 281 (17.6%) assigned to heparin (0.93, 0.77 to 1.12, P=0.43, fig 2 and table 2).

Days since randomisation No at risk

No at risk

Coprimary composite endpoint of all cause mortality, myocardial infarction, or stroke

Cumulative incidence (%) 0 6 9 12 15 18 3 0 5 10 15 20 25 30 1998 2012 1605 1598

With ST segment elevation

Unfractionated heparin Bivalirudin

Without ST segment elevation

Unfractionated heparin Bivalirudin 1911 1928 1368 1363 1889 1905 1354 1353 1875 1899 1349 1349 1866 1894 1345 1346 1859 1888 1341 1343 1841 1869 1324 1335 Unfractionated heparin With ST segment elevation Without ST segment elevation Bivalirudin

Days since randomisation Coprimary composite endpoint of all cause mortality, myocardial infarction, stroke, or BARC type 3 or 5 bleeding*

Cumulative incidence (%) 0 6 9 12 15 18 P=0.68 P=0.50 3 0 5 10 15 20 25 30 1998 2012 1605 1598

Unfractionated heparin Bivalirudin 1881 1912 1359 1356 1854 1885 1338 1345 1843 1879 1331 1340 1833 1874 1326 1337 1825 1869 1322 1334 1808 1850 1305 1326

Fig 2 | Coprimary composite outcomes in patients with acute coronary syndrome with or without st segment elevation. *Major bleeding unrelated to coronary artery bypass graft tab le 2 | Pr es pec ified clinic al out come s in patients w ith ac ut e cor on ar y sy ndr ome w ith and w ithout s t segment elev ation. Per cent ag es ar e cu mu lati ve inc idenc e es tim at es Pr es pe cifi ed out co m es W ith s t s eg m en t e le va tion W ith out s t s eg m en t e lev at io n P f or in te rac tio n bi va lir udin (n =20 12 ) unf ra ct io na te d he par in (n =1 99 8) ra te r at io ( 95 % C i) P v al ue bi va lir udin (n =1 59 8) unf rac tio na te d he par in (n =1 605 ) ra te r at io ( 95 % C i) P v al ue Co pr imar y co m po si te : M aj or a dve rs e c ar dio va sc ul ar e ve nt s 11 8 ( 5. 9) 12 9 ( 6. 5) 0. 90 ( 0. 70 t o 1 .16 ) 0. 43 25 3 ( 15 .9 ) 26 2 ( 16 .4 ) 0. 97 ( 0. 80 t o 1 .17 ) 0. 72 0. 68 Ne t a dv er se cli ni ca l e ven ts 13 9 ( 7.0 ) 16 3 ( 8. 2) 0. 84 (0. 67 to 1. 05) 0.1 3 26 2 ( 16 .5 ) 28 1 ( 17. 6) 0. 93 (0. 77 to 1. 12) 0. 43 0. 50 Al l c au se m or ta lit y 42 (2 .1) 61 (3 .1) 0. 68 ( 0. 46 t o 1 .0 1) 0.0 53 17 (1 .1) 22 (1. 4) 0. 77 ( 0. 41 t o 1 .4 6) 0. 43 0. 73 M yo ca rdi al inf ar ct io n 73 (3. 7) 58 (3 .0 ) 1. 24 ( 0. 88 t o 1 .7 6) 0. 21 23 4 ( 14 .7 ) 24 5 ( 15 .3 ) 0. 95 ( 0. 78 t o 1 .16 ) 0. 64 0.1 9 St ro ke 5 ( 0. 3) 14 (0 .7 ) 0. 35 (0. 13 to 0. 97 ) 0. 035 8 ( 0. 5) 2 ( 0.1 ) 4.0 4 (0 .8 6 t o 1 9.0 4) 0.0 57 0. 00 52 BA RC t yp e 3 o r 5 b le ed in g* 33 (1. 7) 54 (2 .8) 0. 60 (0. 39 to 0. 92) 0. 019 16 (1 .0 ) 34 (2 .1) 0. 47 (0. 26 to 0. 85) 0. 012 0. 52 De fin ite s te nt th ro m bo si s 26 (1. 3) 14 (0 .7 ) 1.8 4 ( 0. 96 to 3 .5 2) 0.0 63 10 (0 .6 ) 7 (0 .4) 1. 44 ( 0. 55 t o 3 .7 7) 0. 46 0. 68 Ac ut e s te nt th ro mb osi s 17 (0 .8 ) 9 ( 0. 5) 1.8 7 ( 0. 83 t o 4 .2 0) 0.1 3 3 ( 0. 2) 4 ( 0. 3) 0. 75 ( 0. 17 t o 3 .3 8) 0. 71 0. 29 Su ba cu te s te nt th ro mb osi s 9 ( 0. 5) 5 ( 0. 3) 1. 78 ( 0. 59 t o 5 .3 1) 0. 30 7 (0 .4) 3 ( 0. 2) 2. 35 ( 0. 61 t o 9 .12 ) 0. 20 0. 75 De fin ite o r pr ob able s te nt th ro mb osi s 32 (1. 6) 20 (1. 0) 1. 58 ( 0. 90 t o 2 .7 7) 0.1 1 13 (0 .8 ) 15 (1 .0 ) 0. 87 ( 0. 41 t o 1 .8 3) 0. 71 0. 21 Ac ut e s te nt th ro mb osi s 19 (0 .9 ) 10 (0. 5) 1.8 8 ( 0. 87 t o 4 .0 6) 0.1 0 3 ( 0. 2) 6 (0 .4) 0. 50 ( 0. 13 t o 2 .0 1) 0. 32 0.0 92 Su ba cu te s te nt th ro mb osi s 13 (0 .7 ) 10 (0. 5) 1. 28 ( 0. 56 t o 2 .9 3) 0. 55 10 (0 .6 ) 9 (0 .6 ) 1.1 2 ( 0. 45 t o 2 .7 5) 0. 81 0. 82 BA RC =B le ed in g A ca de m ic R es ea rc h C on so rt iu m . *M aj or b le ed in g u nr el at ed t o c or on ar y a rt er y b yp as s g ra ft.

(6)

There was no heterogeneity of treatment effect on mortality according to qualifying acute coronary syndrome at baseline (P=0.73), with bivalirudin associated with fewer cases of all cause death (0.68, 0.46 to 1.01, P=0.053) and cardiovascular death (0.67, 0.45 to 1.00, P=0.048) compared with heparin in patients with ST segment elevation and directionally similar effect in patients without ST segment eleva-tion (0.77, 0.41 to 1.46, P=0.43, and 0.76, 0.40 to 1.47, P=0.42, respectively, table 2 , fig 3, and supplemen-tary figure 3A).

Bivalirudin was associated with lower rates of BARC 3 or 5 bleeding, both in patients with ST segment elevation (0.60, 0.39 to 0.92; P=0.019) and in patients without ST seg-ment elevation (0.47, 0.26 to 0.85; P=0.011) with similar treatment effect using the GUSTO (moderate or severe) and the TIMI (major or minor) classifications of bleeding (table 2 , fig 3, and supplementary figure 3D and appendix).

Acute stent thrombosis (definite and definite or prob-able) were not statistically different between the groups with ST segment elevation (1.87, 0.83 to 4.20, P=0.13, and 1.88, 0.87 to 4.06, P=0.10, respectively) and without Days since randomisation

No at risk

All cause mortality

Cumulative incidence (%) 0 2 3 4 1 0 5 10 15 20 25 30 1998 2012 1605 1598

Unfractionated heparin Bivalirudin 1955 1985 1595 1590 1945 1973 1588 1584 1936 1967 1586 1581 1931 1965 1582 1579 1924 1961 1578 1575 1907 1944 1561 1567 Unfractionated heparin With ST segment elevation Without ST segment elevation Bivalirudin

Days since randomisation Stroke Cumulative incidence (%) 0 0.4 0.6 0.8 1.0 0.2 0 5 10 15 20 25 30 1998 2012 1605 1598

Days since randomisation No at risk No at risk Myocardial infarction Cumulative incidence (%) 0 6 9 12 15 18 3 0 5 10 15 20 25 30 1998 2012 1605 1598

Days since randomisation BARC type 3 or 5 bleeding*

Cumulative incidence (%) 0 1 2 3 P=0.73 P=0.005 P=0.19 P=0.52 0 5 10 15 20 25 30 1998 2012 1605 1598

Fig 3 | Components of coprimary composite in patients with acute coronary syndrome with or without st segment elevation. *Major and unrelated to coronary artery bypass graft

(7)

ST segment elevation (0.75, 0.17 to 3.37, P=0.71, and 0.50, 0.13 to 2.01, P=0.32, respectively), with the P value for interaction with qualifying acute coronary syndrome of 0.29 for acute definite stent thrombosis and 0.09 for acute definite or probable stent thrombosis.

Rates of myocardial infarction were similar between bivalirudin and heparin in patients with and without ST segment elevation (1.24, 0.88 to 1.76, and 0.95, 0.78 to 1.16, respectively, P for interaction 0.19; fig 3). Nota-bly, 30 day event rates differed markedly between the two populations: 3.3% in patients with ST segment ele-vation and 15% in those without ST segment eleele-vation. Randomised treatment effect on stroke was direction-ally opposite in patients with and without ST elevation (P for interaction 0.0052), with bivalirudin associated with a lower risk of stroke compared with heparin in patients with ST segment elevation (0.35, 0.13 to 0.97, P=0.036) and a higher risk of stroke in patients without ST segment elevation (4.02, 0.85 to 18.95, P=0.057). additional analyses

Figures 4-7 list the randomised treatment effect on major adverse cardiovascular events and net adverse

clinical events as well as all cause mortality and bleed-ing in patients with and without ST segment elevation according to prespecified subgroups (also see the sup-plementary appendix). In patients with ST segment ele-vation, the randomised treatment effect appeared consistent across most subgroups, with the exception of patients with an estimated glomerular filtration rate of less than 60 mL/min, with previous use of unfraction-ated heparin and treunfraction-ated in centres with a high volume of percutaneous coronary intervention, where the bivalirudin strategy appeared more effective and safe than the heparin strategy. In patients with an estimated glomerular filtration rate of less than 60 mL/min, bivali-rudin appeared to be associated with a reduced risk of death and major bleeding. Treatment effects seemed consistent also in patients who qualified for the group without ST segment elevation, with a weak nominally significant interaction for intended or ongoing use of prasugrel or ticagrelor compared with clopidogrel. discussion

The main finding of this analysis—the largest ran-domised comparison of bivalirudin with heparin in Centre's annual volume of PCI

Low (247-544) Intermediate (548-991) High (1000-1950) Age (years) ≥75 <75 Sex Women Men BMI ≥25 <25

Intended start or continuation of prasugrel or ticagrelor

Yes No

Diabetes

Yes No

Estimated glomerular filtration rate (mL/min)

≥60 <60

History of peripheral vascular disease

Yes No

Previous UFH

Yes No

Randomisation to access site

Femoral access Radial access 0.03* 0.57 0.76 0.07 0.55 0.12 0.003 0.93 0.002 0.92 0.25 0.39 0.08 0.32 0.75 0.50 0.62 0.09 0.32 0.83 0.33 0.08 0.89 0.31 0.004 0.74 0.54 0.004 0.22 0.53 0.63 0.25 0.5 1 2 4

Study P value for trend

or interaction Rate ratio (95% CI) P value 1.28 (0.84 to 1.96) 0.82 (0.52 to 1.29) 0.67 (0.43 to 1.05) 0.82 (0.55 to 1.21) 0.95 (0.68 to 1.32) 0.85 (0.53 to 1.35) 0.93 (0.69 to 1.25) 0.76 (0.55 to 1.04) 1.24 (0.81 to 1.89) 0.96 (0.70 to 1.34) 0.82 (0.55 to 1.22) 0.65 (0.40 to 1.06) 1.02 (0.76 to 1.37) 1.17 (0.86 to 1.60) 0.50 (0.31-0.81) 0.89 (0.44 to 1.81) 0.92 (0.70 to 1.20) 0.53 (0.35 to 0.82) 1.22 (0.89 to 1.68) 0.89 (0.63 to 1.27) 0.92 (0.64 to 1.31) Rate ratio (95% CI) 50/682 35/715 33/615 48/445 70/1567 34/463 84/1549 70/1344 48/668 72/1361 46/651 29/366 89/1646 87/1687 26/307 14/106 104/1906 32/944 86/1068 60/1010 58/1002 Bivalirudin 39/680 43/726 47/592 55/423 74/1575 39/454 90/1544 89/1314 40/684 74/1350 55/648 42/357 87/1641 74/1675 48/301 18/123 111/1875 60/957 69/1041 66/999 63/999 UFH No of events/total

Fig 4 | stratified analysis of coprimary composite of all cause mortality, myocardial infarction, or stroke in patients with acute coronary syndrome with st segment elevation. uFH=unfractionated heparin; PCi=percutaneous coronary intervention; bMi=body mass index

(8)

patients with acute coronary syndrome with ST segment elevation and one of the largest in patients with acute coronary syndrome without ST segment elevation treated with modern antiplatelet agents—is that a regi-men of bivalirudin monotherapy, compared with unfractionated heparin with provisional glycoprotein IIb/IIIa inhibitors, did not result in reduced major adverse cardiovascular events or net adverse clinical events. These results are consistent with the findings from the overall population with acute coronary syndrome.6

efficacy and safety of bivalirudin versus heparin strategies according to type of acute coronary syndrome

A strategy of bivalirudin monotherapy was associated with reduced bleeding compared with heparin with provisional use of glycoprotein IIb/IIIa inhibitors.2-48 This reduction in bleeding was consistent across bleeding classifications (including fatal events), not clearly related to randomised access site (which was allocated in MATRIX), and of similar magnitude in patients with acute coronary syndrome with and

without ST segment elevation. These observations suggest that bivalirudin, if intended as a strategy to minimise bleeding, may be complementary to other strategies (eg, radial access) and could be applied with few restrictions to the general population with acute coronary syndrome, as also suggested by the associa-tion of bivalirudin with a reducassocia-tion in death and bleeding in patients with a low estimated glomerular filtration rate. In secondary analyses we observed that bivalirudin was associated with a reduction in all cause mortality as a result of lower cardiovascular dis-ease mortality, a difference that in this analysis was similar between patients with and without ST segment elevation (P for interaction 0.72).

The excess risk of acute stent thrombosis associated with bivalirudin is well documented.238_{We did not} observe significant differences in the rate of acute stent thrombosis between randomised strategies in the patients with and without ST segment elevation. While the risk of acute stent thrombosis with bivalirudin in patients with ST segment elevation has been well docu-mented (and is directionally similar in MATRIX), we observed no increased risk in patients without ST

0.004* 0.37 0.64 0.24 0.58 0.27 0.001 0.62 0.02 0.35 0.34 0.57 0.002 0.08 0.50 0.22 0.30 0.06 0.99 0.41 0.16 0.08 0.46 0.55 0.0004 0.33 0.26 0.004 0.81 0.08 0.72 0.25 0.5 1 2 4 Rate ratio (95% CI) 1.21 (0.82 to 1.79) 0.89 (0.59 to 1.34) 0.54 (0.36 to 0.81) 0.73 (0.51 to 1.04) 0.90 (0.67 to 1.22) 0.77 (0.50 to 1.18) 0.87 (0.66 to 1.14) 0.76 (0.57 to 1.01) 1.00 (0.69 to 1.46) 0.88 (0.66 to 1.19) 0.77 (0.54 to 1.11) 0.67 (0.43 to 1.05) 0.90 (0.69 to 1.18) 1.09 (0.82 to 1.45) 0.46 (0.30-0.72) 0.73 (0.39 to 1.37) 0.87 (0.68 to 1.11) 0.57 (0.38 to 0.84) 1.04 (0.78 to 1.38) 0.76 (0.55 to 1.04) 0.94 (0.67 to 1.32) Rate ratio (95% CI) 57/682 43/715 39/615 57/445 82/1567 39/463 100/1549 83/1344 56/668 85/1361 54/651 35/366 104/1646 103/1687 31/307 17/106 122/1906 40/944 99/1068 71/1010 68/1002 Bivalirudin 47/680 49/726 67/592 72/423 91/1575 49/454 114/1544 106/1314 57/684 95/1350 68/648 49/357 114/1641 94/1675 61/301 26/123 137/1875 70/957 93/1041 91/999 72/999 UFH No of events/total Centre's annual volume of PCI

Yes No

Diabetes

Yes No

≥60 <60

Yes No

Previous UFH

Yes No

Femoral access Radial access

or interaction P value

Fig 5 | stratified analysis of coprimary composite of all cause mortality, myocardial infarction, stroke, or barC type 3 or 5 bleeding (major and unrelated to coronary artery bypass graft) in patients with acute coronary syndrome with st segment elevation. uFH=unfractionated heparin; PCi=percutaneous coronary intervention; bMi=body mass index

(9)

segment elevation, especially if treated with potent P2Y12 inhibitors.

There was no clear effect of randomised treatment on myocardial infarction according to qualifying acute coronary syndrome. The myocardial infarction compo-nent was far more common (>4 times) in patients without ST segment elevation compared with patients with ST segment elevation, reflecting the contribution of proce-dural related myocardial infarction events in patients without ST segment elevation and a planned percutane-ous coronary intervention. So, even when procedural myocardial infarctions are not accounted for (ie, patients with ST segment elevation), a strategy of bivalirudin monotherapy did not have a measurable effect on either coprimary endpoints, indicating a complex, competitive balance between thrombosis and bleeding, which may explain the overall neutral results. Unexpectedly, a nom-inally significant interaction effect between randomised antithrombotic strategies compared with qualifying acute coronary syndrome on stroke was detected. No such effect has previously been reported, and since it currently defies, to our knowledge, any plausible biological expla-nation, we concur it might represent a spurious finding.

Finally, subgroup analysis indicates that a strategy of bivalirudin monotherapy with glycoprotein IIb/IIIa inhibitors restricted to angiographic complications might be useful in patients with ST segment elevation pretreated with heparin and with renal dysfunction. A benefit of bivalirudin in patients who have received heparin has been consistently reported.14-16_This effect, perhaps related to an insufficient level of anti-coagulation provided by bivalirudin alone, suggests that the common practice of an immediate heparin bolus in patients with ST segment elevation followed by bivalirudin seems to be safe. The effect of bivaliru-din in patients with renal dysfunction is less consis-tent,1718_{but suggests that in the acute setting (where} data on renal function are often not available) bivali-rudin has a favourable safety profile. Finally, a benefit of bivalirudin compared with a heparin strategy in patients without ST segment elevation (but not with ST segment elevation) treated with novel P2Y12 inhibi-tors might reflect a longer time required for these agents to reach their desired effect, a possible consid-eration for the selection of an antithrombotic regimen in this population.19 0.11* 0.17 0.20 0.12 0.04 0.38 0.89 0.62 0.51 0.95 0.20 0.86 0.23 0.36 0.29 0.39 0.31 0.24 0.31 0.07 0.33 0.55 0.40 0.80 0.77 0.53 0.82 0.46 0.93 0.82 0.77 0.25 0.5 1 2 4 Rate ratio (95% CI) 0.82 (0.60 to 1.11) 1.03 (0.77 to 1.36) 1.40 (0.80 to 2.44) 1.16 (0.84 to 1.60) 0.88 (0.69 to 1.12) 1.17 (0.81 to 1.69) 0.89 (0.71 to 1.11) 0.87 (0.69 to 1.10) 1.18 (0.85 to 1.64) 0.75 (0.55 to 1.03) 1.13 (0.88 to 1.44) 1.11 (0.79 to 1.56) 0.90 (0.72 to 1.14) 0.97 (0.78 to 1.21) 0.94 (0.62 to 1.42) 0.85 (0.52 to 1.41) 0.98 (0.79 to 1.20) 0.82 (0.49 to 1.38) 0.99 (0.81 to 1.22) 0.97 (0.75 to 1.26) 0.96 (0.72 to 1.27) Rate ratio (95% CI) 104/543 117/703 32/352 98/461 155/1137 81/416 172/1182 157/1095 96/503 86/663 167/935 85/458 168/1140 199/1311 54/282 38/190 215/1408 33/222 220/1376 138/802 115/796 Bivalirudin 123/546 116/713 23/346 90/481 172/1124 65/385 197/1220 177/1091 85/514 109/656 153/949 74/436 188/1169 204/1312 58/287 37/161 225/1444 40/227 222/1378 142/805 120/800 UFH No of events/total Centre's annual volume of PCI

Yes No

Diabetes

Yes No

≥60 <60

Yes No

Previous UFH

Yes No

Fig 6 | stratified analysis of coprimary composite of all cause mortality, myocardial infarction, or stroke in patients with acute coronary syndrome without st segment elevation. uFH=unfractionated heparin; PCi=percutaneous coronary intervention; bMi=body mass index

(10)

limitations of this study

We acknowledge some limitations in the current analysis. Although the reported subgroups were prespecified in the protocol, and separate statistical analyses on these sub-groups were planned, we did not adjust for multiple com-parisons, increasing the risk of type 1 errors. Also, the results for the coprimary endpoints were all negative. There-fore, differential effects across subgroups of the randomised treatments, if any, must be considered hypothesis generat-ing. Also, the study was not powered for type of acute coro-nary syndrome at baseline. Accordingly, our ability to detect smaller treatment differences was limited. Finally, the study compared two antithrombotic strategies (with the heparin arm intended to reflect use of glycoprotein IIb/IIIa inhibi-tors in clinical practice) rather than two drugs. The different use of glycoprotein IIb/IIIa inhibitors in the two study arms has thus challenged our ability to understand the efficacy and safety of the anticoagulants drugs alone.

Conclusions

A strategy of bivalirudin monotherapy compared with heparin with or without glycoprotein IIb/IIIa inhibitors did not result in reduced major adverse cardiovascular

events or net adverse clinical events in patients with acute coronary syndrome with or without ST segment elevation. The choice of the anticoagulant regimen thus requires a careful balance between the expected risk of bleeding and thrombotic complications with the expected benefits for each patient. Further research is needed to better understand optimal patients, settings, and combinations of antithrombotic treatments, partic-ularly in those with a high risk of early thrombotic com-plications or a high risk of bleeding, or both.

autHOr aFFiliatiOns

1_{Fondazione IRCCS Policlinico San Matteo, Pavia, Italy}

2_{EUSTRATEGY Association, Forli’, Italy}

3_{CTU Bern, University of Bern, Switzerland}

4_{Institute of Social and Preventive Medicine (ISPM), University of}

Bern, Switzerland

5_{Policlinico Multimedica IRCSS, University of Milan, Milan, Italy}

6_{Division of Cardiology, Department of Cardiothoracic Sciences,}

Second University of Naples, Naples, Italy

7_{Department of Cardiology, ASL3 Ospedale Villa Scassi, Genoa, Italy}

8_{Clinica Mediterranea, Napoli, Italy}

9_{Ospedale Fatebene Fratelli, Milano, Italy}

10_{University Hospital Maggiore della Carità, Novara, Italy}

0.12* 0.15 0.20 0.15 0.04 0.26 0.97 0.53 0.54 0.86 0.13 0.89 0.32 0.49 0.15 0.53 0.16 0.14 0.48 0.03 0.49 0.63 0.19 0.48 0.73 0.35 0.56 0.40 0.61 0.65 0.50 0.25 0.5 1 2 4 Rate ratio (95% CI) 0.79 (0.58 to 1.07) 0.98 (0.75 to 1.29) 1.30 (0.77 to 2.20) 1.11 (0.82 to 1.52) 0.84 (0.66 to 1.06) 1.12 (0.79 to 1.59) 0.85 (0.68 to 1.06) 0.84 (0.67 to 1.06) 1.12 (0.82 to 1.54) 0.72 (0.53 to 0.98) 1.09 (0.86 to 1.38) 1.09 (0.78 to 1.51) 0.86 (0.69 to 1.08) 0.93 (0.75 to 1.15) 0.93 (0.63 to 1.37) 0.80 (0.50 to 1.28) 0.94 (0.77 to 1.15) 0.80 (0.48 to 1.34) 0.95 (0.78 to 1.16) 0.94 (0.73 to 1.22) 0.91 (0.69 to 1.20) Rate ratio (95% CI) 106/543 121/703 35/352 105/461 157/1137 86/416 176/1182 163/1095 99/503 88/663 174/935 90/458 172/1140 201/1311 61/282 42/190 220/1408 34/222 228/1376 142/802 120/796 Bivalirudin 129/546 125/713 27/346 100/481 181/1124 72/385 209/1220 189/1091 92/514 116/656 165/949 80/436 201/1169 215/1312 66/287 43/161 238/1444 42/227 239/1378 150/805 131/800 UFH No of events/total Centre's annual volume of PCI

Yes No

Diabetes

Yes No

≥60 <60

Yes No

Previous UFH

Yes No

Fig 7 | all cause mortality, myocardial infarction, stroke, or barC type 3 or 5 bleeding (major and unrelated to coronary artery bypass graft) in patients with acute coronary syndrome without st segment elevation. uFH=unfractionated heparin; PCi=percutaneous coronary intervention; bMi=body mass index

(11)

11_{Ospedale Sirai–Carbonia, Carbonia, Italy}

12_{Humanitas Research Hospital, IRCCS, Rozzano, Italy}

13_{Division of Cardiology-Department of Advanced Biomedical}

Sciences, Federico II University of Naples

14_{A O Ospedale di Desio, Lombardia, Italy}

15_{IRCCS San Martino-IST, Genoa, Italy}

16_{Interventional Cardiology Sandro Pertini Hospital Rome, Italy}

17_{Ospedale Pasquinucci, Massa, Italy}

18_{Citta’ di Lecce Hospital, Lecce, Italy}

19_{Ospedale Giovanni Paolo II, Sciacca, Italy}

20_{Università degli Studi G d’Annunzio Chieti e Pescara, Chieti, Italy}

21_{Ospedale San Jacopo, Pistoia, Italy}

22_{Azienda Ospedali Riuniti-Presidio GM Lancisi, Ancona, Italy}

23_{Ospedale San Raffaele IRCCS, Milano, Italy}

24_{Hospital Clinic, Cardiovascular Institute, IDIBAPS, Barcelona, Spain}

25_{AORN Cardarelli, Napoli, Italy}

26_{Erasmus MC, Rotterdam, Netherlands}

27_{Applied Health Research Centre, Li Ka Shing Knowledge Institute}

of St Michael’s Hospital, and Department of Medicine, University of Toronto, Canada

28_{Swiss Cardiovascular Centre Bern, Bern University Hospital,}

CH-3010 Bern, Switzerland

We thank the MATRIX patients and investigators who volunteered to help others.

Contributors: SL wrote the first draft of the manuscript and subsequent revision, and interpreted it in collaboration with the other authors. SL, EF, PC, PB, CB, MF, BC, AL, SI, DZ-P, GE, ST, AZ, SR CP, AL, FA, MZ, MC, GG, AC, SB, CM, NMvM, SW, and MV acquired data. MR, EN, DH, and PJ performed the analysis and interpreted it in collaboration with the other authors. MV conceived and designed the study and obtained funding. SL and MV had full access to the final data and had final responsibility for the content of the report and the decision to submit for publication. All authors critically revised the paper for important intellectual content and approved the final version. SL and MV are the guarantors.

Funding: The MATRIX programme was sponsored by the Italian Society of Invasive Cardiology (GISE), a non-profit organisation, and received grant support from The Medicines Company and TERUMO. The Medicines Company provided bivalirudin for the study. The sponsor had no role in study design, data collection, data monitoring, analysis, interpretation, or writing of the report.

Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare no support from any organisation for the submitted work. They declare the following financial relations: SL reports honorariums for advisory boards from The Medicines Company during the conduct of the study outside the submitted work; MF received honorariums for advisory boards or as speaker at scientific congresses or as consultant from Astra Zeneca, Eli Lilly, and the Medicines Company, outside the submitted work; BC reports lecture fees from Abbott, AB Medica, Hexacath, Astra, Movi, and Concept Medicals and research grants from AB Medica, Aachen Resonance, and Movi, all outside the submitted work; MZ received honorariums as speaker at scientific congresses from Astra Zeneca and Pfizer, outside the submitted work; SR reports personal fees for advisory board participation from AstraZeneca; PJ has received research grants to the institution from Astra Zeneca, Biotronik, Biosensors International, Eli Lilly, and the Medicines Company, and serves as unpaid member of the steering group of trials funded by Astra Zeneca, Biotronik, Biosensors, St Jude Medical, and the Medicines Company; MV reports grants from the Medicines Company and Terumo during the conduct of the study. No other relationships or activities that could appear to have influenced the submitted work are reported. Ethical approval: This study was approved by the ethics committee from each participating site.

Data sharing: No additional data available.

Transparency: The manuscript’s guarantors (SL and MV) affirm that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 3.0) license, which permits others to distribute, remix, adapt, build upon this work

non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/3.0/. 1 Shahzad A, Kemp I, Mars C, et al. HEAT-PPCI trial investigators.

Unfractionated heparin versus bivalirudin in primary percutaneous coronary intervention (HEAT-PPCI): an open-label, single centre, randomised controlled trial. Lancet 2014;384:1849-58. doi:10.1016/S0140-6736(14)60924-7.

2 Stone GW, Witzenbichler B, Guagliumi G, et al. HORIZONS-AMI Trial Investigators. Bivalirudin during primary PCI in acute myocardial infarction. N Engl J Med 2008;358:2218-30. doi:10.1056/NEJMoa0708191. 3 Kastrati A, Neumann FJ, Schulz S, et al. ISAR-REACT 4 Trial Investigators.

Abciximab and heparin versus bivalirudin for non-ST-elevation myocardial infarction. N Engl J Med 2011;365:1980-9. doi:10.1056/NEJMoa1109596. 4 Stone GW, McLaurin BT, Cox DA, et al. ACUITY Investigators.

Bivalirudin for patients with acute coronary syndromes. N Engl J Med 2006;355:2203-16. doi:10.1056/NEJMoa062437.

5 Valgimigli M. MATRIX investigators. Design and rationale for the Minimizing Adverse haemorrhagic events by TRansradial access site and systemic Implementation of angioX program. Am Heart J 2014;168:838-45.e6. doi:10.1016/j.ahj.2014.08.013. 6 Valgimigli M, Frigoli E, Leonardi S, et al. MATRIX Investigators.

Bivalirudin or Unfractionated Heparin in Acute Coronary Syndromes. N Engl J Med 2015;373:997-1009. doi:10.1056/NEJMoa1507854. 7 De Luca L, Leonardi S, Cavallini C, et al. EYESHOT Investigators.

Contemporary antithrombotic strategies in patients with acute coronary syndrome admitted to cardiac care units in Italy: The EYESHOT Study. Eur Heart J Acute Cardiovasc Care 2015;4:441-52. doi:10.1177/2048872614560505.

8 Cavender MA, Sabatine MS. Bivalirudin versus heparin in patients planned for percutaneous coronary intervention: a meta-analysis of randomised controlled trials. Lancet 2014;384:599-606. doi:10.1016/ S0140-6736(14)61216-2.

9 Valgimigli M, Gagnor A, Calabró P, et al. MATRIX Investigators. Radial versus femoral access in patients with acute coronary syndromes undergoing invasive management: a randomised multicentre trial. Lancet 2015;385:2465-76. doi:10.1016/S0140-6736(15)60292-6. 10 Nikus K, Pahlm O, Wagner G, et al. Electrocardiographic classification of

acute coronary syndromes: a review by a committee of the International Society for Holter and Non-Invasive Electrocardiology. J Electrocardiol 2010;43:91-103. doi:10.1016/j.jelectrocard.2009.07.009. 11 Mehran R, Rao SV, Bhatt DL, et al. Standardized bleeding definitions

for cardiovascular clinical trials: a consensus report from the Bleeding Academic Research Consortium. Circulation 2011;123:2736-47. doi:10.1161/CIRCULATIONAHA.110.009449.

12 GUSTO Investigators. An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. The GUSTO investigators. N Engl J Med 1993;329:673-82. doi:10.1056/ NEJM199309023291001.

13 Clayton D, Hills M. Statistical models in epidemiology.Oxford University Press, 1993.

14 Dangas GD, Mehran R, Nikolsky E, et al. HORIZONS-AMI Trial Investigators. Effect of switching antithrombin agents for primary angioplasty in acute myocardial infarction: the HORIZONS-SWITCH analysis. J Am Coll Cardiol 2011;57:2309-16. doi:10.1016/j. jacc.2011.01.038.

15 Koutouzis M, Lagerqvist B, James S, et al. Unfractionated heparin administration in patients treated with bivalirudin during primary percutaneous coronary intervention is associated lower mortality and target lesion thrombosis: a report from the Swedish Coronary Angiography and Angioplasty Registry (SCAAR). Heart 2011;97:1484-8. doi:10.1136/hrt.2011.224709.

16 White HD, Chew DP, Hoekstra JW, et al. Safety and efficacy of switching from either unfractionated heparin or enoxaparin to bivalirudin in patients with non-ST-segment elevation acute coronary syndromes managed with an invasive strategy: results from the ACUITY (Acute Catheterization and Urgent Intervention Triage strategY) trial. J Am Coll Cardiol 2008;51:1734-41. doi:10.1016/j.jacc.2007.12.052. 17 Saltzman AJ, Stone GW, Claessen BE, et al. Long-term impact of

chronic kidney disease in patients with ST-segment elevation myocardial infarction treated with primary percutaneous coronary intervention: the HORIZONS-AMI (Harmonizing Outcomes With Revascularization and Stents in Acute Myocardial Infarction) trial. JACC Cardiovasc Interv 2011;4:1011-9. doi:10.1016/j.jcin.2011.06.012. 18 Chew DP, Bhatt DL, Kimball W, et al. Bivalirudin provides increasing

benefit with decreasing renal function: a meta-analysis of randomized trials. Am J Cardiol 2003;92:919-23. doi:10.1016/ S0002-9149(03)00970-6.

19 Franchi F, Angiolillo DJ. Novel antiplatelet agents in acute coronary syndrome. Nat Rev Cardiol 2015;12:30-47. doi:10.1038/ nrcardio.2014.156.