Listen to this manuscript’s

audio summary by

Editor-in-Chief

Dr. Valentin Fuster on

JACC.org.

J O U R N A L O F T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y V O L . 7 3 , N O . 7 , 2 0 1 9

ª 2 0 1 9 B Y T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y F O UN DA T I O N

P U B L I S H E D B Y E L S E V I E R

ORIGINAL INVESTIGATIONS

Dual Antiplatelet Therapy DurationBased on Ischemic and Bleeding RisksAfter Coronary Stenting

Francesco Costa, MD, PHD,a,b David Van Klaveren, MSC,c,d Fausto Feres, MD, PHD,e Stefan James, MD, PHD,f

Lorenz Räber, MD, PHD,b Thomas Pilgrim, MD,b Myeong-Ki Hong, MD,g Hyo-Soo Kim, MD,h Antonio Colombo, MD,i

Philippe Gabriel Steg, MD,j,k Deepak L. Bhatt, MD, MPH,l Gregg W. Stone, MD,m,n Stephan Windecker, MD,b

Ewout W. Steyerberg, PHD,c Marco Valgimigli, MD, PHD,b for the PRECISE-DAPT Study Investigators

ABSTRACT

ISS

FrobSw

for

Tu

Sci

Un

Se

Un

Ita

FramC

Re

BACKGROUND Complex percutaneous coronary intervention (PCI) is associated with higher ischemic risk, which can be

mitigated by long-term dual antiplatelet therapy (DAPT). However, concomitant high bleeding risk (HBR) may be pre-

sent, making it unclear whether short- or long-term DAPT should be prioritized.

OBJECTIVES This study investigated the effects of ischemic (by PCI complexity) and bleeding (by PRECISE-DAPT

[PREdicting bleeding Complications in patients undergoing stent Implantation and SubsequEnt Dual AntiPlatelet

Therapy] score) risks on clinical outcomes and on the impact of DAPT duration after coronary stenting.

METHODS Complex PCI was defined as$3 stents implanted and/or$3 lesions treated, bifurcation stenting and/or stent

length >60 mm, and/or chronic total occlusion revascularization. Ischemic and bleeding outcomes in high ($25) or non-

high (<25) PRECISE-DAPT strata were evaluated based on randomly allocated duration of DAPT.

RESULTS Among 14,963 patients from 8 randomized trials, 3,118 underwent complex PCI and experienced a higher

rate of ischemic, but not bleeding, events. Long-term DAPT in non-HBR patients reduced ischemic events in both

complex (absolute risk difference: �3.86%; 95% confidence interval: �7.71 to þ0.06) and noncomplex PCI strata

(absolute risk difference: �1.14%; 95% confidence interval: �2.26 to �0.02), but not among HBR patients, regardless of

complex PCI features. The bleeding risk according to the Thrombolysis In Myocardial Infarction scale was increased by

long-term DAPT only in HBR patients, regardless of PCI complexity.

CONCLUSIONS Patients who underwent complex PCI had a higher risk of ischemic events, but benefitted from

long-term DAPT only if HBR features were not present. These data suggested that when concordant, bleeding,

more than ischemic risk, should inform decision-making on the duration of DAPT. (J Am Coll Cardiol 2019;73:741–54)

© 2019 by the American College of Cardiology Foundation.

N 0735-1097/$36.00 https://doi.org/10.1016/j.jacc.2018.11.048

m the aDepartment of Clinical and Experimental Medicine, Policlinic “G. Martino,” University of Messina, Messina, Italy;

iss Cardiovascular Center Bern, Bern University Hospital, Bern, Switzerland; cDepartment of Medical Statistics and Bioin-

matics, Leiden University Medical Center, Leiden, the Netherlands; dInstitute for Clinical Research and Health Policy Studies,

fts Medical Center, Boston, Massachusetts; eIstituto Dante Pazzanese de Cardiologia, Sao Paulo, Brazil; fDepartment of Medical

ences and Uppsala Clinical Research Center, Uppsala University, Uppsala, Sweden; gSeverance Cardiovascular Hospital, Yonsei

iversity College of Medicine, Seoul, Korea and Severance Biomedical Science Institute, Yonsei University College of Medicine,

oul, Korea; hDepartment of Internal Medicine, Seoul National University Hospital, Seoul, Korea; iInterventional Cardiology

it, San Raffaele Scientific Institute, Milan, Italy and Interventional Cardiology Unit, EMO-GVM Centro Cuore Columbus, Milan,

ly; jFACT, Assistance Publique–Hôpitaux de Paris (AP-HP), Paris, France; kUniversité Paris-Diderot, Bichat Hospital, Paris,

nce; lBrigham and Women’s Hospital Heart and Vascular Center and Harvard Medical School, Boston, Massachusetts;

olumbia University Medical Center/New York-Presbyterian Hospital, New York, New York; and the nCardiovascular

search Foundation, New York, New York. Mr. Van Klaveren and Dr. Steyerberg were supported by the National Institutes of

ABBR EV I A T I ON S

AND ACRONYMS

ACS = acute coronary

syndrome

ARD = absolute risk difference

CI = confidence interval

DAPT = dual antiplatelet

therapy

HBR = high bleeding risk

MI = myocardial infarction

PCI = percutaneous coronary

intervention

ST = stent thrombosis

TIMI = Thrombolysis In

Myocardial Infarction

TVR = target vessel

revascularization

Health (gra

Abbott Vas

received sp

and has re

Squibb, CSL

Medicines

scape Card

Cardiovasc

member of

Cleveland C

Institute; h

ACC.org; V

Institute; c

Institute (c

American C

committee

(Secretary/

NCDR-ACT

research fu

Ironwood,

royalties fr

Investigato

of Cardiolo

fees from

Medical De

Vascular N

consultant

institution

research gr

Company,

other autho

Manuscript

Costa et al. J A C C V O L . 7 3 , N O . 7 , 2 0 1 9

PRECISE-DAPT Score and Complex PCI F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4

742

T he PRECISE-DAPT (PREdictingbleeding Complications in patientsundergoing stent Implantation and

SubsequEnt Dual AntiPlatelet Therapy) scoreis a 5-item bleeding risk prediction modeldeveloped to estimate the bleeding risk in pa-tients who receive dual antiplatelet therapy(DAPT) after stent implantation (1). This scorehas been validated in 2 large independentpatient populations with acute coronarysyndrome (ACS) (2) and in a contemporaryreal-world registry (3). The categorization ofpatients based on the PRECISE-DAPT scorewas shown to be useful to inform decision-making for duration of DAPT in stentedpatients (1), which drove its endorsement byinternational guidelines (4,5). In particular,patients with a score of $25 experienced net

SEE PAGE 755

clinical harm if treated with a DAPT regimen forlonger than 3 to 6 months after stenting (1). However,this prediction tool deserves further validation toextend its application in multiple patient subgroups,including those at higher ischemic risk.

Extensive coronary artery disease necessitatingcomplex percutaneous revascularization techniquesfrequently drives practitioners to prescribe a longer

nt U01NS086294). Dr. James has received intuitional research

cular, and Boston Scientific. Dr. Pilgrim has received personal fee

eakers fees from Biotronik and Boston Scientific. Dr. Steg has rec

ceived speaking or consulting fees from Amarin, Amgen, Astra

-Behring, Daiichi-Sankyo, GlaxoSmithKline, Janssen, Lilly, Merck

Company. Dr. Bhatt has been a member of the advisory Board fo

iology, and Regado Biosciences; has been on the Board of Director

ular Patient Care; has been the Chair for the American Heart As

the Data Monitoring Committees for Baim Institute for Clinical Re

linic, Duke Clinical Research Institute, Mayo Clinic, Mount Sinai

as received honoraria from the American College of Cardiology

ice-Chair, ACC Accreditation Committee), Baim Institute for Cli

linical trial steering committee), Belvoir Publications (Editor-in-C

linical trial steering committees), HMP Communications (Editor-in

ollege of Cardiology (Guest Editor; Associate Editor), Populatio

), Slack Publications (Chief Medical Editor, Cardiology Today’s I

Treasurer), and WebMD (CME steering committees); has received

ION Registry Steering Committee (Chair), VA CART Research

nding from Abbott, Amarin, Amgen, AstraZeneca, Bristol-Myers

Ischemix, Lilly, Medtronic, PhaseBio, Pfizer, Regeneron, Roche,

om Elsevier (Editor, Cardiovascular Intervention: A Companion

r for Biotronik, Boston Scientific, St. Jude Medical (now Abbott),

gy; and has received unfunded research from FlowCo, Merck, PLx

Velomedix, Toray, Vascular Dynamics, Miracor, TherOx, Reva,

velopment Technologies; has received other fees from MedFocus

onotransfer Technologies, Cagent, Qool Therapeutics, Caliber,

on prasugrel patent litigation paid for by Lupin Pharmaceuticals

from Abbott, Amgen, Boston Scientific, Biotronik, Medtronic,

ants from The Medicines Company, Terumo, and AstraZaneca;

Terumo, St. Jude Vascular, Alvimedica, Abbott Vascular, AstraZan

rs have reported that they have no relationships relevant to the

received August 15, 2018; revised manuscript received Novembe

than average duration of DAPT (6). This is currentlysupported by 2 retrospective analyses, one from asingle large study and another from multiple small-to medium-sized combined studies, that showed anabsolute greater ischemic risk reduction in patientswho received long-term, instead of short-term,DAPT after complex intervention (7,8). Nonetheless,patients who undergo complex intervention simul-taneously carry features that significantly increasetheir bleeding hazards, such as multiple comorbid-ities, renal disease, or previous bleeding. Themutual role of ischemic and bleeding risks on out-comes and whether long- or short-term DAPT shouldbe prioritized in the setting of patients with bothhigh bleeding and ischemic risk features remainsunclear and constitutes the focus of the presentanalysis.

METHODS

STUDY DESIGN AND POPULATION. The study popu-lation, which consisted of a total of 14,963 patientstreated with percutaneous coronary intervention(PCI) and subsequent DAPT was previously described(1). In brief, patients treated with coronary stenting in

grants from The Medicines Company, AstraZeneca,

s from Biotronik, Medtronic, and Edwards; and has

eived grants from Bayer, Merck, Servier, and Sanofi;

Zeneca, Bayer, Boehringer Ingelheim, Bristol-Myers

Novartis, Pfizer, Regeneron, Sanofi, Servier, and The

r Cardax, Elsevier Practice Update Cardiology, Med-

s for Boston VA Research Institute and the Society of

sociation Quality Oversight Committee; has been a

search (formerly Harvard Clinical Research Institute),

School of Medicine, and Population Health Research

(Senior Associate Editor, Clinical Trials and News,

nical Research (formerly Harvard Clinical Research

hief, Harvard Heart Letter), Duke Clinical Research

-Chief, Journal of Invasive Cardiology), Journal of the

n Health Research Institute (clinical trial steering

ntervention), Society of Cardiovascular Patient Care

other fees from Clinical Cardiology (Deputy Editor),

and Publications Committee (Chair); has received

Squibb, Chiesi, Eisai, Ethicon, Forest Laboratories,

Sanofi, and The Medicines Company; has received

to Braunwald’s Heart Disease); has been a site co-

Svelte; has been a trustee for the American College

Pharma, and Takeda. Dr. Stone has received personal

V-wave, Matrizyme, Ablative Solutions, Neovasc,

family of funds, Guided Delivery Systems, Micardia,

Aria, and Biostar family of funds; and has been a

. Dr. Windecker has received research grants to his

St. Jude, and Terumo. Dr. Valgimigli has received

and has received personal fees from The Medicines

eca, and Correvio outside of the submitted work. All

contents of this paper to disclose.

r 7, 2018, accepted November 8, 2018.

J A C C V O L . 7 3 , N O . 7 , 2 0 1 9 Costa et al.F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4 PRECISE-DAPT Score and Complex PCI

743

an elective, urgent, or emergent setting were pooledat an individual level from 8 randomized controlledtrials (Online Figure 1) (9–16). The duration of DAPTwith aspirin and a P2Y12 inhibitor was randomlyassigned to short- (3 or 6 months) or long-term (12 or24 months) treatment in 10,081 patients in 5 of the 8included studies (11–15), whereas the duration lastedaccording to international guidelines in 2 studies(9,10), and ranged from 1 to 12 months based on pa-tient characteristics in 1 study (16) (Online Figure 1).The terminology used for short- or long-term DAPTduration throughout the paper is based on theaforementioned randomization scheme and refers tothe short- (i.e., 3 or 6 months) or long-term (i.e., 12 or24 months) duration of the assigned DAPT treatment.

All clinical trials were approved by the ethics com-mittees at each study center, and all patients providedwritten informed consent. For the purpose of thepresent analysis, we stratified the population based onquantifiable bleeding versus ischemic risks, andanalyzed outcomes in terms of both safety and efficacyendpoints in the overall cohort. The impact of DAPTduration, further stratified by ischemic and bleedingrisks, was restricted to studies that randomly allocatedpatients to a short- or long-term DAPT regimen.BLEEDING RISK ASSESSMENT. The PRECISE-DAPTscore was calculated as previously described basedon age, creatinine clearance, hemoglobin, whiteblood cell count, and previous spontaneous bleeding(1). Validation of the PRECISE-DAPT score and in-struction for its calculation are discussed elsewhere(1,4). Patients were considered at high bleeding risk(HBR) for scores $25 and non-HBR for scores <25.Sensitivity analysis that took into account the quar-tiles of bleeding risk (i.e., very low risk: score #10;low risk: score 11 to 17; moderate risk: score 18 to 24;and high risk: score $25) was performed.ISCHEMIC RISK ASSESSMENT. The predictedischemic risk after PCI was quantified using previ-ously validated and guideline-endorsed criteria of PCIcomplexity (4,7) as follows: PCI with $3 stentsimplanted, $3 lesions treated, and/or 3 coronaryvessels treated; and/or bifurcation with 2 stentsimplanted, total stent length >60 mm, and/or treat-ment of a chronic total occlusion. The presence of atleast 1 element of complexity stratified the patientinto the complex PCI group. Occurrence of >1 elementof complexity in the same patient was also accountedfor and analyzed (i.e., complex PCI score). We alsoanalyzed the impact of an alternative high ischemicrisk definition, which accounted for both the clinicaland anatomical and/or procedural features (i.e.,complex patient), by considering ACS at presentationas an additional criterion to the previously

established anatomical and/or procedural complexPCI features. Both high ischemic risk assessmentmethods were pre-specified in the study plan.

OUTCOMES. All clinical and laboratory variablesincluded in the present analysis were prospectivelycollected (1). Major and minor bleeding, according tothe Thrombolysis In Myocardial Infarction (TIMI) (12)definition (17), and the composite of myocardialinfarction (MI), definite stent thrombosis (ST) (18),stroke, or target vessel revascularization (TVR) wereappraised up to 2 years of follow-up, and were pre-specified as primary safety and efficacy endpoints,respectively. Net adverse clinical events, which wereobtained after pooling ischemic and bleeding events,as well as other ischemic and bleeding secondaryendpoints, were also explored. All clinical eventswere adjudicated by a blinded independent clinicalevents committee for each of the included studies.

STATISTICAL ANALYSIS. The patient population wasstratified according to the presence or absence of ahigh PRECISE-DAPT score and/or the presence orabsence of at least $1 complex PCI criteria. Categori-cal variables are reported as percentages, continuousvariables are reported as median and interquartilerange, and in-between group differences wereassessed using the chi-square test and Wilcoxon rank-sum test, respectively. Event rates in the 4 exploredgroups were evaluated using Kaplan-Meier estimatesat 24 months and compared with the log-rank test.The impact of a randomized DAPT duration amongthe PRECISE-DAPT bleeding risk strata was evaluatedin a large subgroup of patients (n ¼ 10,081) (OnlineFigure 1). In this analysis, events were counted afterthe landmark point of treatment divergence in the 2study arms. For instance, if 1 of the DAPT trialsenvisioned 3-month treatment versus 12-monthtreatment, events that occurred in the first3 months, when treatment was identical in the 2study arms, were censored. The absolute risk differ-ence (ARD) and its 95% confidence interval (CI) werecalculated according to Altman et al. (19). The methodfor the calculation of the risk difference, which is adifference between proportions, requires the calcu-lation of the CIs of the 2 proportions separately. Withl1 to u1 being the 95% CI of the first proportion (p1) andl2 to u2 being the 95% CI of the second proportion (p2),the 95% CI for the difference is given by:

95% CI ¼ RD�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðp1 � l1Þ2 þ ðu2 � p2Þ2

qto

RDþffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðp2 � l2Þ2 þ ðu1 � p1Þ2

q

The ARD was calculated to evaluate the differencein event incidence after a long-term DAPT duration

TABLE 1 Baseline Characteristics in the Four Explored Groups by Bleeding Risk and Intervention Complexity

p Value

PRECISE-DAPT<25 (non-HBR)Noncomplex PCI

(n ¼ 8,982)

PRECISE-DAPT<25 (non-HBR)Complex PCI(n ¼ 2,273)

PRECISE-DAPT$25 (HBR)

Noncomplex PCI(n ¼ 2,863)

PRECISE-DAPT$25 (HBR)Complex PCI(n ¼ 845)

PRECISE-DAPT<25 (non-HBR)Noncomplex PCIvs. Complex PCI

PRECISE-DAPT$25 (HBR)

Noncomplex PCIvs. Complex PCI

Age, yrs 61 (54–68) 62 (55.0–68.6) 76.4 (71.8–81.0) 77.2 (72.9–82.0) 0.04 0.008

Women (vs. men) 25.8 23.6 42.6 40.6 0.03 0.30

Hypertension 69.4 68.7 79.0 80.9 0.49 0.23

Dyslipidemia 62.2 65.7 56.0 57.3 0.002 0.50

Current smoking 29.0 31.1 12.7 12.1 0.06 0.61

Diabetes 26.7 28.5 29.4 34.6 0.06 0.003

Insulin-dependent 5.2 5.6 4.9 7.6 0.41 0.002

PVD 6.1 8.1 17.3 21.5 0.02 0.03

Previous MI 19.2 19.5 20.7 23.0 0.76 0.15

Previous PCI 15.0 17.0 17.7 18.1 0.02 0.78

Previous CABG 5.1 5.9 8.4 7.6 0.13 0.43

Previous stroke 2.7 4.0 5.4 5.7 0.003 0.77

Previous bleeding 0.0 0.0 5.3 5.4 1.00 0.91

Weight, kg 75 (66–85) 75 (66–85) 68.4 (60–78) 70 (61–80) 0.94 0.03

Creatinine clearance, ml/min 89.3 (76.0–106.6) 89.4 (76.7–103.9) 52.3 (40.5–62.2) 51.9 (41.1–62.7) 0.99 0.41

White blood cell count, 103 U/ml 7.7 (6.2–10.1) 7.8 (6.4–9.7) 8.2 (6.7–10.6) 8.0 (6.7–10.4) 0.42 0.39

Hemoglobin, g/dl 14.0 (13.1–15) 14.2 (13.2–15.1) 12.7 (11.4–14.1) 12.4 (11.1–13.7) 0.05 0.03

LV ejection fraction, % 55.0 (50.0–61) 55.0 (45.0–64.4) 55.0 (45.0–60.0) 54.0 (40.0–60.4) <0.001 0.94

LV ejection fraction <35% 6.7 8.9 13.2 15.7 0.001 0.08

Clinical presentation

SCAD 45.7 48.4 40.1 35.0 <0.0001 <0.0001

UA 23.2 20.1 23.5 21.4

NSTEMI 11.3 13.8 18.3 28.2

STEMI 19.8 17.7 18.1 15.5

Single-vessel disease 64.9 38.2 48.4 23.8 <0.0001 <0.0001

No. of vessels treated

1 87.0 60.8 81.3 41.4 <0.0001 <0.0001

2 12.8 31.7 18.5 44.9

3 0.2 7.5 0.2 13.7

>1 lesion treated 18.0 50.5 20.6 67.4

Overall stent length, mm 24.0 (18.0–33.0) 58.0 (36.0–78.0) 24.0 (18.0–33.0) 63.0 (43.0–81.0) <0.0001 <0.0001

Type of stent implanted

BMS 11.5 8.5 19.4 16.3 <0.0001 0.004

First-generation DES 7.5 8.5 9.4 6.8

Second-generation DES 81.0 83.0 71.2 76.9

Therapy at discharge

Aspirin 99.1 99.1 97.3 97.0 0.92 0.70

Clopidogrel 87.6 78.0 90.2 83.5 <0.0001 <0.0001

Prasugrel 8.1 12.1 3.2 4.4 <0.0001 0.10

Ticagrelor 2.8 8.5 2.3 8.5 <0.0001 <0.0001

Statin 90.8 90.7 84.6 87.4 0.95 0.05

Beta-blocker 74.4 76.9 71.9 73.9 0.02 0.27

ACE/ARB 65.8 70.8 66.0 66.5 <0.0001 0.80

PPI 34.2 29.6 51.3 51.3 0.004 0.99

Values are n (range) or %.

ACE/ARB ¼ angiotensin-converting enzyme inhibitor or angiotensin-II receptor blocker; BMS ¼ bare-metal stent; CABG ¼ coronary artery bypass grafting; DAPT ¼ dual-antiplatelet therapy; DES ¼ drug-eluting stent; HBR ¼ high bleeding risk; LV ¼ left ventricular; MI ¼ myocardial infarction; NSTEMI ¼ non�ST-segment elevated myocardial infarction; PCI ¼ percutaneous coronary intervention; PPI ¼ protonpump inhibitor; PRECISE-DAPT ¼ PREdicting bleeding Complications in patients undergoing stent Implantation and SubsequEnt Dual AntiPlatelet Therapy; PVD ¼ peripheral vascular disease; SCAD ¼ stablecoronary artery disease; STEMI ¼ ST-segment elevated myocardial infarction; UA ¼ unstable angina.

Costa et al. J A C C V O L . 7 3 , N O . 7 , 2 0 1 9

PRECISE-DAPT Score and Complex PCI F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4

744

versus short-term DAPT duration in each stratum forprimary and secondary ischemic and bleeding end-points. We tested the null hypothesis of a homoge-neous ARD (long-term DAPT treatment vs. short-term

DAPT treatment) within complex PCI patients versusnoncomplex PCI patients. A p value <0.05 was usedto detect significant heterogeneity across subgroups.ARD for bleeding and ischemia after a long treatment

TABLE 2 Ischemic and Bleeding Events According to Bleeding (PRECISE-DAPT) and Intervention Complexity Risk Profiles

p Value

PRECISE-DAPT<25 (non-HBR)Noncomplex PCI

PRECISE-DAPT<25 (non-HBR)Complex PCI

PRECISE-DAPT$25 (HBR)

Noncomplex PCI

PRECISE-DAPT$25 (HBR)Complex PCI

PRECISE-DAPT<25 (non-HBR) vs.PRECISE-DAPT$25 (HBR)

Noncomplex PCIvs. Complex PCI

MI, definite ST, stroke or TVR 7.6 13.9 12.5 20.5 <0.0001 <0.0001

MI, definite ST, stroke 3.5 6.8 8.6 13.3 <0.0001 <0.0001

Definite ST 0.7 1.6 1.3 0.8 0.03 0.009

Death for all causes 2.4 2.8 12.2 14.7 <0.0001 0.008

Cardiovascular death 1.4 1.7 8.3 9.9 <0.0001 0.01

Noncardiovascular death 0.9 1.1 4.3 5.2 <0.0001 0.22

TIMI major or minor 1.4 1.8 4.8 5.8 <0.0001 0.06

TIMI major 0.7 1.1 2.5 2.9 <0.0001 0.08

Fatal bleeding 0.07 0.3 0.8 1.3 <0.0001 0.05

Values are %. 24 months Kaplan-Meier estimates are presented. Log-rank test was used for comparison.

HBR ¼ high bleeding risk; MI ¼ myocardial infarction; ST ¼ stent thrombosis; TIMI ¼ Thrombolysis In Myocardial Infarction; TVR ¼ target vessel revascularization.

J A C C V O L . 7 3 , N O . 7 , 2 0 1 9 Costa et al.F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4 PRECISE-DAPT Score and Complex PCI

745

and their CIs were plotted for the 4 groups ofPRECISE-DAPT high and/or non-high versus complexand/or noncomplex PCI (19). All analyses were per-formed with R package version 3.3.2 (R Foundation,Vienna, Austria).

RESULTS

Among 14,963 patients who underwent PCI in thepooled study cohort, 11,845 (79.2%) underwentnoncomplex PCI, of whom 8,982 (60.0%) were not atHBR and 2,863 (19.1%) of whom were at HBR. ComplexPCI was performed in 3,118 (20.8%) patients, of whom2,273 (15.2%) were not at HBR and 845 (5.6%) of whomwere at HBR. Among those who underwent complexPCI, 1,668 (53.5%) had a single element that qualifiedfor complexity and 1,450 (46.5%) had $2 elements(Online Figure 2). Patients with complex PCI wereolder and more commonly had a history of diabetesmellitus and peripheral vascular disease, reducedejection fraction, and were more commonly treatedwith second-generation drug-eluting stents, as well asticagrelor or prasugrel at discharge (Table 1). The dis-tribution of the PRECISE-DAPT score in the complexPCI population, further stratified according to eachcomplex PCI criterion is presented in Online Figure 3.The median PRECISE-DAPT score was slightly lower inpatients with complex PCI features compared withthose without complex PCI features (16.2 score pointsvs. 17.0 score points; p¼0.001), but the score increasedmodestly with the increasing number of elements ofcomplexity (noncomplex: 16.2 points vs. 1 element:16.0 points vs. $2 elements: 18.2 points; p < 0.0001).

CLINICAL EVENTS BASED ON ISCHEMIC AND

BLEEDING RISKS. The occurrence of the primarycomposite endpoint of MI, definite ST, stroke, or TVR

was higher among patients in the complex PCI groupcompared with the noncomplex PCI group, both withand without HBR features (Table 2, Figure 1A). Thepresence of >1 element of complexity was associatedwith greater risk of the primary endpoint (OnlineTable 1). HBR was also associated with a greater rateof ischemic events (Table 2, Figure 1A). Major or minorbleeding events were roughly 3-fold higher in theHBR groups, both with (5.8% vs. 1.8%; p <0.001) andwithout (4.8% vs. 1.4%; p < 0.001) complex PCI fea-tures. Fulfillment of complex PCI criteria alone wasnot associated with a significant increase in bleedingrisk for TIMI major or TIMI major or minor bleeding(Table 2, Figure 1B). Results were consistent whenconsidering the 4 quartiles of bleeding risk (OnlineFigure 4) and when the analysis was limited to thefirst 12 months of follow-up (Online Table 2). Multi-variable models including the PRECISE-DAPT scorevariables and the complex PCI criteria are presentedin the Online Appendix (Online Tables 3 and 4).

IMPACT OF RANDOMIZED DAPT DURATION ACCORDING

TO ISCHEMIC AND BLEEDING RISKS. Kaplan-Meiercurves for the composite endpoint of MI, definite ST,stroke or TVR, and TIMI major or minor bleeding inthe high and nonhigh bleeding and ischemic riskgroups, based on randomized long-term DAPT treat-ment versus short-term DAPT treatment are providedin Figures 2 and 3.

A long-term DAPT duration was associated with areduction of ischemic events in patients without HBRin both the complex and noncomplex PCI strata (ARDfor long-term DAPT vs. short-term DAPT: noncomplexPCI: �1.14%; 95% CI �2.26% to �0.02% vs. complexPCI: �3.86%; 95% CI: �7.71% to þ0.06%; pint ¼ 0.19)(Central Illustration, Table 3). In contrast, no benefit of

FIGURE 1 Ischemic and Bleeding Events

0 3 6 9 12Time (Months)

15 18 21 24

Non-HBR / Non-Complex PCINon-HBR / Complex PCIHBR / Non-Complex PCIHBR / Complex PCI

8,981

0.00

0.05

0.10Cu

mul

ativ

e In

cide

nce

of M

I, De

f ST,

Str

oke

or T

VR

0.15

0.20

8,645 8,446 8,314 7,571 4,8892,273 2,105 2,050 2,000 1,738 1,1562,863 2,605 2,495 2,409 2,191 1,139845 737 697 658 584 321

4,5931,1051,066

312

3,9681,017968287

2,032582381134

vs. p < 0.001p < 0.001vs.p < 0.001vs.p < 0.001vs.

A

0 3 6 9 12Time (Months)

15 18 21 24

Non-HBR / Non-Complex PCINon-HBR / Complex PCIHBR / Non-Complex PCIHBR / Complex PCI

8,981

0.00

0.05

0.10

Cum

ulat

ive

Inci

denc

e of

TIM

I Maj

or o

r Min

or B

leed

ing

0.15

0.20

8,805 8,711 8,653 7,949 5,2122,273 2,200 2,181 2,166 1,926 1,3072,863 2,663 2,592 2,528 2,325 1,250845 782 757 733 668 378

4,9211,2641,182368

4,2901,1811,089344

2,156646392162

vs. p < 0.001p < 0.001vs.p = 0.50vs.p = 0.11vs.

Non-HBR / Non-Complex PCI Non-HBR / Complex PCIHBR / Non-Complex PCI HBR / Complex PCI

B

Continued on the next page

Costa et al. J A C C V O L . 7 3 , N O . 7 , 2 0 1 9

PRECISE-DAPT Score and Complex PCI F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4

746

J A C C V O L . 7 3 , N O . 7 , 2 0 1 9 Costa et al.F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4 PRECISE-DAPT Score and Complex PCI

747

long-term treatment with DAPT was observed amongHBR patients, regardless of complex PCI features(ARD for long-term DAPT vs. short-term DAPT:noncomplex PCI: þ1.45%; 95% CI: �1.84% to þ4.72%vs. complex PCI: þ1.30%; 95% CI: �6.99% to þ9.57%;pint ¼ 0.97) (Central Illustration, Table 3).

Long-term DAPT was associated with an excess ofbleeding in patients with HBR regardless of PCIcomplexity, with a significant increase of TIMI majoror minor bleeding among noncomplex PCI-treatedpatients (ARD for long-term vs. short-termDAPT: þ2.61%; 95% CI: þ0.89% to þ4.31%) and a nu-merical, but not statistically significant, increase inthe complex PCI stratum (ARD for long-term DAPT vs.short-term DAPT: þ3.04%; 95% CI: �2.97% to þ8.82%;pint ¼ 0.89) (Central Illustration, Table 3). In contrast,among patients without HBR, long-term DAPT was notassociated with higher bleeding liability, regardless ofPCI complexity (ARD for long-term DAPT vs. short-term DAPT: noncomplex PCI: þ0.12%; 95%CI: �0.25% to þ0.50% vs. complex PCI: þ0.28%; 95%CI: �0.46% to þ1.26%; pint ¼ 0.73) (Central Illustration,Table 3).

Results remained consistent when both ischemicand bleeding events were evaluated in the compositenet adverse clinical event endpoint (CentralIllustration, Table 3). The net clinical benefit fromlong-term DAPT among non-HBR patients wasnumerically higher among complex PCI patients, butthe interaction term did not reach statistical signifi-cance (ARD for long-term DAPT vs. short-term DAPT:noncomplex PCI: �0.91%; 95% CI: �2.07% to þ0.25%vs. complex PCI: �4.05%; 95% CI: �7.96% to �0.07%;pint ¼ 0.14). Results in terms of ischemic, bleeding, orcombined ischemic and bleeding endpoints werecorroborated when only ACS patients were analyzed(Online Table 5) or when the number of complex PCIcriteria were further accounted for (Online Figure 5).Further stratification according to the 4 bleeding riskquartiles showed consistent ischemic benefits in pa-tients at very low, low, or moderate bleeding risks,but not for those with a PRECISE DAPT score $25(Online Table 6).

When the study population was stratified accord-ing to either ACS presentation and/or procedural

FIGURE 1 Continued

(A) Ischemic and (B) bleeding events over time for patients with and/or

percutaneous coronary intervention (PCI). Kaplan-Meier estimates for (A

stroke or target vessel revascularization (TVR), and (B) Thrombolysis In

24-month follow-up are presented. The log-rank test was used for comp

among patients with complex PCI or HBR and highest in the group in wh

major or minor bleeding was higher among patients with HBR but not a

patient complexity (i.e., ACS at presentation with orwithout complex PCI: Complex Patient criteria), sig-nificant interactions were noted with respect toischemic and net adverse events for treatment dura-tion and patient complexity among non-HBR in-dividuals, but not among those with HBR (Figure 4,Online Table 7).

DISCUSSION

The main findings of the present analysis are asfollows:

1. In univariate analysis, patients who underwentcomplex PCI compared with noncomplex PCI hadhigher ischemic and mortality risks, which werefurther increased in those with >1 element of thecomplexity criteria. The fulfillment of complex PCIcriteria did not significantly affect bleeding risk.Patients with HBR had a higher incidence of bothbleeding and ischemic events, including mortality,compared with non-HBR patients.

2. Patients who did not fulfill HBR criteria had aconsistent benefit from long-term DAPT comparedwith that of short-term DAPT, with no apparenttradeoff in bleeding. The absolute magnitude ofthe ischemic risk benefit offered by a long-termDAPT regimen trended greater in patients whounderwent complex PCI compared with noncom-plex PCI, but the interaction was not significant(pint ¼ 0.14). Such a difference became statisticallysignificant when ACS was also factored in as anadditional element of ischemic risk (pint < 0.001).

3. HBR patients did not derive ischemic or mortalitybenefit from long-term DAPT, regardless of thecomplexity of the intervention they underwent orthe acute presentation at the time of PCI. Theyexperienced an excess of bleeding eventscompared with a short duration of treatment.

DAPT reduces coronary ischemic events after PCI,and its action is 2-fold: 1) prevention of MI and STarising from previously stented segments; and 2) pre-vention of MI in nonculprit segments (20,21). Stent-related ischemic events mainly occur due to stentunderexpansion during the early phase after

without high bleeding risk (HBR) and with and/or without complex

) myocardial infarction (MI), definite stent thrombosis (def ST),

Myocardial Infarction (TIMI) major or minor bleeding during the

arison. (A) The occurrence of MI, def ST, stroke, or TVR was higher

ich both characteristics were present. (B) The occurrence of TIMI

mong those with complex PCI alone.

FIGURE 2 Ischemic Events Among Patients Randomized to Long-Term or Short-Term Treatment With DAPT With and/or Without HBR and PCI Complexity

0 3 6 9 12 15 18 21 24

0.00

0.05

0.10

Cum

ulat

ive

Inci

denc

e of

MI,

Def S

T, S

trok

e or

TVR

0.15

Low Risk of Bleeding - Non-Complex PCI

ARD: –1.14%95% CI: –2.26 to –0.02

Long DAPTShort DAPT

3,255 3,117 3,082 3,037 2,644 1,6583,268 3,129 3,072 1,4833,017 2,699

1,5211,623

1,2101,185

425395

Long DAPT Short DAPT

0 3 6 9 12 15 18 21 24

0.00

0.05

0.10

0.15

Low Risk of Bleeding - Complex PCI

ARD: –3.86%95% CI: –7.71 to +0.06

689 628 619 603 473 232661 602 594 234582 464

209251

175190

4852

0 3 6 9 12 15 18 21 24

0.00

0.05

0.10

Cum

ulat

ive

Inci

denc

e of

MI,

Def S

T, S

trok

e or

TVR

0.15

High Risk of Bleeding - Non-Complex PCI

ARD: +1.45%95% CI: –1.84 to +4.72

Time (Months)Long DAPTShort DAPT

887 814 803 783 682 418881 776 767 339745 655

385373

338293

6758

Time (Months)0 3 6 9 12 15 18 21 24

0.00

0.05

0.10

0.15

High Risk of Bleeding - Complex PCI

ARD: +1.30%95% CI: –6.99 to +9.57

219 183 179 169 138 66221 187 183 77174 149

6682

5865

119

The 24-month Kaplan-Meier estimates of MI, def ST, stroke, or TVR for patients randomly assigned to long-term (12 to 24 months) or short-term (3 to 6 months) dual

antiplatelet therapy (DAPT) in the 4 explored groups with and/or without HBR and with and/or without complex PCI. Long-term DAPT was associated with a reduction

of MI, def ST, stroke, or TVR in patients without HBR, noncomplex PCI, and with a borderline reduction in patients in the complex PCI stratum (upper panels). In

contrast, long-term DAPT provided no benefit among patients with HBR (lower panels), regardless of complex PCI features. ARD ¼ absolute risk difference;

CI ¼ confidence interval; other abbreviations as in Figure 1.

Costa et al. J A C C V O L . 7 3 , N O . 7 , 2 0 1 9

PRECISE-DAPT Score and Complex PCI F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4

748

FIGURE 3 Bleeding Events Among Patients Randomized to Long-Term or Short-Term Treatment With DAPT With and/or Without HBR and PCI Complexity

0 3 6 9 12 15 18 21 24

0.00

0.05

0.10

Cum

ulat

ive

Inci

denc

e of

TIM

I Maj

or o

r Min

or B

leed

ing

Low Risk of Bleeding - Non-Complex PCI0.15

ARD: +0.12%95% CI: –0.25 to +0.50

Long DAPTShort DAPT

3,255 3,187 3,164 3,145 2,764 1,7673,268 3,217 3,183 1,6123,161 2,859

1,6301,753

1,3101,309

442421

Long DAPT Short DAPT

0 3 6 9 12 15 18 21 24

0.00

0.05

0.10

Cum

ulat

ive

Inci

denc

e of

TIM

I Maj

or o

r Min

or B

leed

ing 0.15

High Risk of Bleeding - Non-Complex PCI

ARD: +2.61%95% CI: +0.89 to +4.31

Time (Months)Long DAPTShort DAPT

887 845 833 816 720 459881 812 806 395793 710

427424

384350

7059

0 3 6 9 12 15 18 21 24

0.00

0.05

0.10

0.15

Low Risk of Bleeding - Complex PCI

ARD: +0.28%95% CI: –0.46 to +1.26

689 665 660 657 533 282661 639 635 280632 523

259299

223237

5360

Time (Months)0 3 6 9 12 15 18 21 24

0.00

0.05

0.10

0.15

High Risk of Bleeding - Complex PCI

ARD: +3.04%95% CI: –2.97 to +8.82

219 197 194 189 160 86221 207 205 98200 174

84104

7584

1212

The 24-month Kaplan-Meier estimates of TIMI major or minor bleeding for patients randomly assigned to long-term (12 to 24 months) or short-term (3 to 6 months)

DAPT in the 4 explored groups with and/or without HBR and with and/or without complex PCI. Long-term DAPT was not associated with an increase in TIMI major or

minor bleeding in patients without HBR (upper panels), regardless of PCI complexity. In contrast, long-term DAPT was associated with an increase of TIMI major or

minor bleeding among patients with HBR and noncomplex PCI features (lower panels). Abbreviations as in Figures 1 and 2.

J A C C V O L . 7 3 , N O . 7 , 2 0 1 9 Costa et al.F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4 PRECISE-DAPT Score and Complex PCI

749

CENTRAL ILLUSTRATION PRECISE-DAPT Score and Complex Percutaneous Coronary Intervention

Risk Difference ForLong vs. Short DAPT

+0.12% p = 0.53

pint

0.73

Blee

ding

*Is

chem

ia†

NACE

‡

0.89

0.19

0.97

0.14

0.83

+0.28% p = 0.57

+2.61% p = 0.003+3.04% p = 0.30

–4% –3%Long DAPT Better(% Reduction of Events

with Long DAPT)

Short DAPT Better(% Increase of Events

with Long DAPT)

–2% –1% 0 +1% +2% +3% +4%

+1.45% p = 0.39

–1.14% p = 0.04–3.86% p = 0.05

–0.91% p = 0.12

Non-Complex PCI Complex PCI

–4.05% p = 0.04PRECISE-DAPT <25

(Non-High Bleeding Risk)

PRECISE-DAPT <25(Non-High Bleeding Risk)

PRECISE-DAPT ≥25(High Bleeding Risk)

PRECISE-DAPT ≥25(High Bleeding Risk)

PRECISE-DAPT <25(Non-High Bleeding Risk)

PRECISE-DAPT ≥25(High Bleeding Risk)

+1.30% p = 0.76

+2.81% p = 0.11+1.68% p = 0.73

Costa, F. et al. J Am Coll Cardiol. 2019;73(7):741–54.

Risk difference for long-term versus short-term dual antiplatelet therapy (DAPT) duration for *bleeding, †ischemic, and ‡net adverse clinical events in patients

stratified according to the PRECISE-DAPT score and with and/or without complex percutaneous coronary intervention (PCI). *Bleeding endpoint defined according to

the Thrombolysis In Myocardial Infarction (TIMI) major and/or minor bleeding definition. †Ischemic endpoint defined according to the composite of myocardial

infarction, definite stent thrombosis, stroke, or target vessel revascularization. ‡Net adverse clinical events defined according to the composite of myocardial

infarction, definite stent thrombosis, stroke, target vessel revascularization, or TIMI major and/or minor bleeding. Complex PCI is defined as the presence of a PCI

with $3 stents implanted, $3 lesions treated, 3 coronary vessels treated, bifurcation with 2 stents implanted, total stent length >60 mm, and/or treatment of a

chronic total occlusion.

Costa et al. J A C C V O L . 7 3 , N O . 7 , 2 0 1 9

PRECISE-DAPT Score and Complex PCI F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4

750

implantation, and by restenosis or neoatherosclerosisin a later phase; these events appear to be criticallydependent on the degree of P2Y12 pathway residualactivation (22). The introduction of current generationdrug-eluting stents greatly reduced the risk of early,late, and very late ST (23,24), and warranted theinvestigation of a short-term DAPT for 6, 3 (12,14), oreven only 1 month after stenting (16,25). Withcontemporary devices, ST is rare and is responsible fora minority of recurrent ischemic events; in the DAPTtrial, among patients treated with second-generationDESs, ST was responsible for only 15% of recurrentischemic events in the placebo arm (26). Hence, long-term DAPT achieves its benefit by mainly preventingischemic events in the untreated coronary

vasculature, thereby reducing the global risk ofnonstent-relatedMIwith themarginal effects of strokeand venous thromboembolism (27). However, thebenefit provided by the institution and continuation ofDAPT is counterbalanced by an increase in majorbleeding risk (20) that accrues over time and affectsmorbidity, mortality, quality of life, and costs (28,29).This sets the rationale for an individualized approachin the decision-making for DAPT duration, taking intoaccount the anticipated risks and benefits (4,5,30).

Multiple studies investigated individual patient orprocedural features that enhanced, diminished, ordid not affect the anticipated risk and/or benefittradeoff of a long-term DAPT regimen (31). PCIcomplexity, defined according to 6 anatomical and/or

TABLE 3 Clinical Events According to the PRECISE-DAPT Score and the PCI Complexity Among Patients Randomized to Short-Term Versus

Long-Term DAPT Duration

PRECISE-DAPT <25 (non-HBR) PRECISE-DAPT $25 (HBR)

Short DAPT Long DAPT ARD p Value pint Short DAPT Long DAPT ARD p Value pint

MI, definite ST, stroke or TVR

Noncomplex PCI 4.52 3.38 �1.14 (�2.26 to �0.02) 0.04 0.19 6.76 8.21 þ1.45 (�1.84 to þ4.72) 0.39 0.97

Complex PCI 10.18 6.32 �3.86 (�7.71 to þ0.06) 0.05 10.3 11.6 þ1.30 (�6.99 to þ9.57) 0.76

Death, MI, definite ST, stroke or TVR

Noncomplex PCI 5.33 4.37 �0.96 (�2.21 to þ0.30) 0.13 0.38 10.66 11.05 þ0.39 (�3.46 to þ4.24) 0.84 0.75

Complex PCI 11.26 8.33 �2.93 (�7.21 to þ1.30) 0.17 15.9 18.0 þ2.14 (�8.05 to þ12.48) 0.68

Definite ST

Noncomplex PCI 0.25 0.15 �0.10 (�0.34 to þ0.14) 0.43 0.18 0.41 0.67 þ0.26 (�0.46 to þ1.11) 0.51 NA

Complex PCI 0.67 0.00 �0.67 (�1.33 to þ0.25) NA 0.00 0.00 0.00 (0.00 to 0.00) NA

Death for all causes

Noncomplex PCI 0.98 1.28 þ0.30 (�0.36 to þ0.96) 0.37 0.54 6.12 4.67 �1.45 (�4.15 to þ1.24) 0.29 0.93

Complex PCI 1.36 2.29 þ0.93 (�0.99 to þ2.86) 0.34 10.01 8.89 �1.12 (�8.55 to þ6.29) 0.77

TIMI major or minor

Noncomplex PCI 0.37 0.49 þ0.12 (�0.25 to þ0.50) 0.53 0.73 1.03 3.64 þ2.61 (þ0.89 to þ4.31) 0.003 0.89

Complex PCI 0.32 0.60 þ0.28 (�0.46 to þ1.26) 0.57 3.30 6.34 þ3.04 (�2.97 to þ8.82) 0.30

MI, definite ST, stroke TVR or TIMImajor/minor bleeding

Noncomplex PCI 4.73 3.82 �0.91 (�2.07 to þ0.25) 0.12 0.14 7.65 10.46 þ2.81 (�0.72 to þ6.33) 0.11 0.83

Complex PCI 10.58 6.54 �4.05 (�7.96 to �0.07) 0.04 12.52 14.20 þ1.68 (�7.85 to þ11.15) 0.73

Values are % and ARD (95% confidence intervals). Twenty-four month Kaplan-Meier estimates are presented for long-term versus short-term DAPT treatment. The ARD and its 95% confidence interval arecalculated according to Altman et al. (19).

ARD ¼ absolute risk difference; other abbreviations as in Tables 1 and 2.

J A C C V O L . 7 3 , N O . 7 , 2 0 1 9 Costa et al.F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4 PRECISE-DAPT Score and Complex PCI

751

procedural criteria, was recently proposed as amarker of ischemic risk and a driver for DAPT dura-tion selection (7). By applying this definition, patientswho fulfilled at least 1 of those criteria had a higherrisk of recurrent ischemic events. Most importantly,the investigators showed that patients who under-went complex PCI had a significant 44% reduction ofmajor adverse coronary events with a $12-monthDAPT course compared with a 3-month to 6-monthcourse, a difference that was not observed in thenoncomplex PCI subgroup. However, an excess ofbleeding events was consistently noted in both com-plex and noncomplex PCI patients treated with long-term DAPT (7). Therefore, it was speculated thatcomplex PCI might be a marker of more extensivecoronary disease, justifying, at least in part, the po-tential benefit offered by a longer DAPT regimen inthis patient category. Similarly, in a subanalysis of theI-LOVE-IT 2 (Evaluate Safety and Effectiveness of theTivoli DES and the Firebird DES for Treatment ofCoronary Revascularization) trial, patients who wererandomized to 6 or 12 months of DAPT were furtherstratified according to the residual SYNTAX (SynergyBetween Percutaneous Coronary Intervention WithTaxus and Cardiac Surgery) score after the indexintervention (32). A higher residual SYNTAX scorewas associated with an increased risk of recurrentischemic events and with a greater potential benefit

from the treatment with long-term DAPT versusshort-term DAPT (32). In a consistent manner, ananalysis from the PRODIGY (Prolonging DualAntiplatelet Treatment After Grading Stent-InducedIntimal Hyperplasia) trial showed that patients withleft main or proximal left anterior descending coro-nary artery disease had an increased risk of adverseischemic events and derived a 55% reduction of def-inite, probable, or possible ST after 24 months ofDAPT compared with 6 months of DAPT (33).

However, none of these previous studies assessedthe mutual and possibly competing role of highischemic risk and bleeding risk features, and whetherwhich one should preferentially affect the decision-making about DAPT duration in patients withconcomitantly high ischemic and bleeding status.

Our present analysis showed that, among patientswithout HBR features, the presence of complex PCImight further increase the absolute ischemic riskreduction achieved by long-term treatment, whichidentified this specific population as one with thehighest achievable benefit from a longer DAPT treat-ment. No signal of increased bleeding liability wasnoted in non-HBR patients despite long-term DAPT.

In contrast, HBR features portended an elevatedrisk of both ischemic and bleeding events, and long-term DAPT regimen did not mitigate the former butonly increased the latter, regardless of the baseline

FIGURE 4 PRECISE-DAPT Score and Complex Patients

Non-Complex Patient Complex Patient

Risk Difference for Long vs. Short DAPT

+0.25% p = 0.16

pint

0.49

Blee

ding

*Is

chem

ia†

NACE

‡

0.22

< 0.001

0.68

< 0.001

0.32

+0.01% p = 0.98

+3.95% p = 0.004+1.78% p = 0.12

Long DAPT Better(% Reduction of Events

with Long DAPT)

Short DAPT Better(% Increase of Events

with Long DAPT)

–4%–5% –3% –2% –1% 0 +5%+1% +2% +3% +4%

+2.28% p = 0.22

+0.43% p = 0.51–3.50% p < 0.001

+0.70% p = 0.31–3.45% p < 0.001

PRECISE-DAPT <25(Non-High Bleeding Risk)

PRECISE-DAPT <25(Non-High Bleeding Risk)

PRECISE-DAPT ≥25(High Bleeding Risk)

PRECISE-DAPT ≥25(High Bleeding Risk)

PRECISE-DAPT <25(Non-High Bleeding Risk)

PRECISE-DAPT ≥25(High Bleeding Risk)

+1.06% p = 0.64

+4.58% p = 0.03+1.37% p = 0.56

Risk difference for longversus shortDAPT duration for *bleeding, †ischemic, and ‡net adverse clinical events in patients stratified according to the PRECISE-DAPT

score and with/without complex patient criteria. *Bleeding endpoint defined according to the TIMImajor/minor bleeding definition. †Ischemic endpoint defined

according to the composite of myocardial infarction, definite stent thrombosis, stroke or target vessel revascularization. ‡Net adverse clinical events defined

according to the composite of myocardial infarction, definite stent thrombosis, stroke, target vessel revascularization or TIMI major/minor bleeding. Complex

patient is defined as the presence of at least one element among the following: acute coronary syndrome at presentation, PCI with$3 stents implanted,$3

lesions treated, 3 coronary vessels treated, bifurcation with 2 stents implanted, total stent length>60mm, treatment of a chronic total occlusion. DAPT¼ dual

antiplatelet therapy; NACE ¼ net adverse clinical events; other abbreviations as in Figures 1 and 2.

Costa et al. J A C C V O L . 7 3 , N O . 7 , 2 0 1 9

PRECISE-DAPT Score and Complex PCI F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4

752

ischemic risk. There were multiple possible explana-tions for these novel and clinically important findings.

The occurrence of bleeding in HBR patients mighttrigger ischemic events by the abrupt cessation of allantiplatelet agents, by transfusion and subsequentinflammation, or by nonadherence tomedical therapy.

Alternatively, HBR features might be associatedwith an inherent higher ischemic risk, which was notor only minimally modifiable by antiplatelet therapyduration and/or intensity modification. In a previousmediation analysis from the ADAPT-DES (Plateletreactivity and clinical outcomes after coronary arteryimplantation of drug-eluting stents) study, HBR fea-tures, such as anemia and chronic kidney disease,were associated with an excess of ischemic risk,which was not mediated or only marginally mediatedby residual platelet reactivity (34). Hence, theimplementation of an intensified and prolongedDAPT regimen might be ineffective in preventing and

limiting that risk. Direct effects of HBR features onthrombosis were previously proposed. The anemicstate stimulates the bone marrow to release immatureplatelets that are hyperreactive, less responsive toantiplatelet agents, and prone to be more thrombo-genic (35); similarly, the presence of uremic solutes inpatients with terminal chronic kidney disease wereassociated with a higher thrombotic milieu (36).

Against previous evidence that ACS at presentationis a major treatment modifier with respect to DAPTduration (31,37), we also analyzed the acuity of pre-sentation (i.e., ACS vs. non-ACS) as an additional highischemic risk feature, with or without complex PCIcriteria. We consistently observed no evidence thatHBR patients, even if they underwent complexintervention due to ACS, derived benefit from long-term DAPT.

Our findings should also be placed in context of arecent analysis by Yeh et al. (8), which explored the

PERSPECTIVES

COMPETENCY IN PATIENT CARE AND PROCEDURAL

SKILLS: Prolonging DAPT is associated with a reduction in

ischemic events in patients undergoing complex PCI who are not

at HBR, but it is not beneficial in those at HBR, regardless of PCI

complexity. Hence, the risk of bleeding should be considered

more strongly than ischemic risk in determining the duration of

DAPT after PCI.

TRANSLATIONAL OUTLOOK: Prospective studies are needed

to verify whether patients at HBR derive net benefit from a

shorter duration of DAPT after stenting, regardless of procedural

complexity.

J A C C V O L . 7 3 , N O . 7 , 2 0 1 9 Costa et al.F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4 PRECISE-DAPT Score and Complex PCI

753

impact of lesion complexity in terms of the unpro-tected left main artery, with >2 lesions per vessel, alesion length of $30 mm, bifurcation lesion with aside branch of $2.5 mm, vein bypass graft (segment oranastomosis), or thrombus-containing lesion, in theDAPT trial dataset (8). The investigators observedthat the benefits and risks of extending DAPT beyond12 months were consistent, regardless of concomitantlesion complexity criteria, and patients with higherDAPT scores (i.e., a decision-making tool for DAPTduration based on the calculation of the net benefitfrom 12-month DAPT vs. 30-month DAPT) derived thegreatest absolute benefit from prolonged treatmentwith thienopyridines, regardless of anatomicalcomplexity (8).

The MASTER-DAPT (Management of High BleedingRisk Patients Post Bioresorbable Polymer CoatedStent Implantation with an Abbreviated Versus Pro-longed DAPT Regimen) study is currently assessingwhether the presence of at least 1 HBR feature,including a PRECISE DAPT score of $25, justifies ashorter than average DAPT duration both in patientswith or without complex PCI (NCT03023020).

STUDY LIMITATIONS. First, this was a post hocretrospective analysis, and as such, it was hypothesis-generating, which requires further prospective vali-dation. Second, as a pooled analysis of randomizedclinical trials, this study was subject to the originallimitations of the included trials. However, the highquality of the studies, which had consistent externaladjudication of clinical events, and their patient-leveldesigns ensure the robustness of our dataset. Third,elements for complex PCI criteria were prospectivelycollected in each of the included study by in-vestigators and were not reviewed by a central corelaboratory. This was of minor importance for some ofthe complex PCI features (e.g., number of stentsimplanted, number of lesions treated, or overall stentlength); however, a more granular description of thebifurcation stenting technique (e.g., culotte stentingvs. minicrush) or chronic total occlusion characteris-tics would have been desirable. Fourth, due to theretrospective design of the analysis, we could notexplore the value of other potentially important ele-ments of complexity not available in the data set.However, the main aim of the analysis was to

benchmark the PRECISE-DAPT score on a solid andwidely accepted definition of complex PCI. Fifth, thestudy power was limited to explore heterogeneity forrare events or smaller subgroups. Similarly, each in-dividual element of intervention complexity mightact as treatment modifier DAPT duration at differentdegrees. Finally, patients on oral anticoagulationwere not included in the study.

CONCLUSIONS

Patients who undergo complex PCI are at furtherincreased risk of ischemic events. However, such pa-tients do not appear to derive any additional benefitfrom a long course of DAPT if HBR features (accordingto the PRECISE-DAPT score) are also present, whichultimately presents an unfavorable net clinicaloutcome. Our analysis suggests that when concor-dant, bleeding, more than ischemic risk, shouldinform decision-making on DAPT duration. Ongoingprospective studies are investigating whether HBRpatients derive greater net clinical benefit from ashort-term DAPT regimen compared with a long-termDAPT regimen, regardless of PCI complexity.

ADDRESS FOR CORRESPONDENCE: Dr. Marco Valgi-migli, Swiss Cardiovascular Center Bern, Bern UniversityHospital, CH-3010, Bern, Switzerland. E-mail: marco.valgimigli@insel.ch. Twitter: @vlgmrc.

RE F E RENCE S

1. Costa F, van Klaveren D, James S, et al. Deri-vation and validation of the predicting bleedingcomplications in patients undergoing stent im-plantation and subsequent dual antiplatelet ther-apy (PRECISE-DAPT) score: a pooled analysis of

individual-patient datasets from clinical trials.Lancet 2017;389:1025–34.2. Wallentin L, Becker RC, Budaj A, et al. Ticagrelorversus clopidogrel in patients with acute coronarysyndromes. N Engl J Med 2009;361:1045–57.

3. Koskinas KC, Raber L, Zanchin T, et al.Clinical impact of gastrointestinal bleedingin patients undergoing percutaneous coro-nary interventions. Circ Cardiovasc Interv2015;8.

Costa et al. J A C C V O L . 7 3 , N O . 7 , 2 0 1 9

PRECISE-DAPT Score and Complex PCI F E B R U A R Y 2 6 , 2 0 1 9 : 7 4 1 – 5 4

754

4. Valgimigli M, Bueno H, Byrne RA, et al. 2017ESC focused update on dual antiplatelet therapy incoronary artery disease developed in collaborationwith EACTS: The Task Force for dual antiplatelettherapy in coronary artery disease of the EuropeanSociety of Cardiology (ESC) and of the EuropeanAssociation for Cardio-Thoracic Surgery (EACTS).Eur Heart J 2018;39:213–60.

5. Mehta SR, Bainey KR, Cantor WJ, et al. 2018Canadian Cardiovascular Society (CCS)/CanadianAssociation of Interventional Cardiology (CAIC)focused update of the guidelines for the use ofantiplatelet therapy. Can J Cardiol 2018;34:214–33.

6. Valgimigli M, Costa F, Byrne R, Haude M,Baumbach A, Windecker S. Dual antiplatelettherapy duration after coronary stenting in clinicalpractice: results of an EAPCI survey. Euro-Intervention 2015;11:68–74.

7. Giustino G, Chieffo A, Palmerini T, et al. Efficacyand safety of dual antiplatelet therapy aftercomplex PCI. J Am Coll Cardiol 2016;68:1851–64.

8. Yeh RW, Kereiakes DJ, Steg PG, et al. Lesioncomplexity and outcomes of extended dual anti-platelet therapy after percutaneous coronaryintervention. J Am Coll Cardiol 2017;70:2213–23.

9. Pilgrim T, Heg D, Roffi M, et al. Ultrathin strutbiodegradable polymer sirolimus-eluting stentversus durable polymer everolimus-eluting stentfor percutaneous coronary revascularisation(BIOSCIENCE): a randomised, single-blind, non-inferiority trial. Lancet 2014;384:2111–22.

10. Raber L, Kelbaek H, Ostojic M, et al. Effect ofbiolimus-eluting stents with biodegradable poly-mer vs bare-metal stents on cardiovascular eventsamong patients with acute myocardial infarction:the COMFORTABLE AMI randomized trial. JAMA2012;308:777–87.

11. Gwon HC, Hahn JY, Park KW, et al. Six-monthversus 12-month dual antiplatelet therapy afterimplantation of drug-eluting stents: the Efficacyof Xience/Promus Versus Cypher to Reduce LateLoss After Stenting (EXCELLENT) randomized,multicenter study. Circulation 2012;125:505–13.

12. Feres F, Costa RA, Abizaid A, et al. Three vstwelve months of dual antiplatelet therapy afterzotarolimus-eluting stents: the OPTIMIZE ran-domized trial. JAMA 2013;310:2510–22.

13. Valgimigli M, Campo G, Monti M, et al. Short-versus long-term duration of dual-antiplatelettherapy after coronary stenting: a randomizedmulticenter trial. Circulation 2012;125:2015–26.

14. Kim BK, Hong MK, Shin DH, et al. A newstrategy for discontinuation of dual antiplatelettherapy: the RESET Trial (REal Safety and Efficacyof 3-month dual antiplatelet Therapy followingEndeavor zotarolimus-eluting stent implantation).J Am Coll Cardiol 2012;60:1340–8.

15. Colombo A, Chieffo A, Frasheri A, et al. Sec-ond-generation drug-eluting stent implantationfollowed by 6- versus 12-month dual antiplatelettherapy: the SECURITY randomized clinical trial.J Am Coll Cardiol 2014;64:2086–97.

16. Valgimigli M, Patialiakas A, Thury A, et al.Zotarolimus-eluting versus bare-metal stents inuncertain drug-eluting stent candidates. J Am CollCardiol 2015;65:805–15.

17. Chesebro JH, Knatterud G, Roberts R, et al.Thrombolysis in Myocardial Infarction (TIMI) trial,phase I: a comparison between intravenous tissueplasminogen activator and intravenous streptoki-nase. Clinical findings through hospital discharge.Circulation 1987;76:142–54.

18. Cutlip DE, Windecker S, Mehran R, et al. Clin-ical end points in coronary stent trials: a case forstandardized definitions. Circulation 2007;115:2344–51.

19. Altman D, Machin D, Bryant T, Gardner M.Statistics with Confidence: Confidence Intervalsand Statistical Guidelines. 2nd edition. London:BMJ Books, 2000.

20. Navarese EP, Andreotti F, Schulze V, et al.Optimal duration of dual antiplatelet therapy afterpercutaneous coronary intervention with drugeluting stents: meta-analysis of randomisedcontrolled trials. BMJ 2015;350:h1618.

21. Mauri L, Kereiakes DJ, Yeh RW, et al. Twelve or30 months of dual antiplatelet therapy after drug-eluting stents. N Engl J Med 2014;371:2155–66.

22. Adriaenssens T, Joner M, Godschalk TC, et al.Optical coherence tomography findings in patientswith coronary stent thrombosis: a report of thePRESTIGE consortium (Prevention of Late StentThrombosis by an Interdisciplinary Global Euro-pean Effort). Circulation 2017;136:1007–21.

23. Wijns W, Steg PG, Mauri L, et al. Endeavourzotarolimus-eluting stent reduces stent throm-bosis and improves clinical outcomes comparedwith cypher sirolimus-eluting stent: 4-year resultsof the PROTECT randomized trial. Eur Heart J2014;35:2812–20.

24. Stone GW, Rizvi A, Newman W, et al. Ever-olimus-eluting versus paclitaxel-eluting stents incoronary artery disease. N Engl J Med 2010;362:1663–74.

25. Urban P, Meredith IT, Abizaid A, et al. Poly-mer-free drug-coated coronary stents in patientsat high bleeding risk. N Engl J Med 2015;373:2038–47.

26. Hermiller JB, Krucoff MW, Kereiakes DJ, et al.Benefits and risks of extended dual antiplatelettherapy after everolimus-eluting stents. J Am CollCardiol Intv 2016;9:138–47.

27. Cavallari I, Morrow DA, Creager MA, et al.Frequency, predictors, and impact of combinedantiplatelet therapy on venous thromboembolismin patients with symptomatic atherosclerosis. Cir-culation 2018;137:684–92.

28. Amin AP, Bachuwar A, Reid KJ, et al. Nuisancebleeding with prolonged dual antiplatelet therapyafter acute myocardial infarction and its impact onhealth status. J Am Coll Cardiol 2013;61:2130–8.

29. Valgimigli M, Costa F, Lokhnygina Y, et al.Trade-off of myocardial infarction vs. bleedingtypes on mortality after acute coronary syndrome:

lessons from the Thrombin Receptor Antagonistfor Clinical Event Reduction in Acute CoronarySyndrome (TRACER) randomized trial. Eur Heart J2017;38:804–10.

30. Levine GN, Bates ER, Bittl JA, et al. 2016 ACC/AHA guideline focused update on duration of dualantiplatelet therapy in patients with coronary ar-tery disease: A Report of the American College ofCardiology/American Heart Association Task Forceon Clinical Practice Guidelines. J Am Coll Cardiol2016;68:1082–115.

31. Costa F, Vranckx P, Leonardi S, et al.Impact of clinical presentation on ischaemicand bleeding outcomes in patients receiving6- or 24-month duration of dual-antiplatelettherapy after stent implantation: a pre-specified analysis from the PRODIGY (Prolong-ing Dual-Antiplatelet Treatment After GradingStent-Induced Intimal Hyperplasia) trial. EurHeart J 2015;36:1242–51.

32. Qiu M, Li Y, Li J, et al. Impact of six versus12 months of dual antiplatelet therapy in patientswith drug-eluting stent implantation after riskstratification with the residual SYNTAX score:results from a secondary analysis of the I-LOVE-IT 2 trial. Catheter Cardiovasc Interv 2017;89:565–73.

33. Costa F, Adamo M, Ariotti S, et al. Left main orproximal left anterior descending coronary arterydisease location identifies high-risk patientsderiving potentially greater benefit from pro-longed dual antiplatelet therapy duration. Euro-Intervention 2016;11:e1222–30.

34. Redfors B, Ben-Yehuda O, Lin SH, et al.Quantifying ischemic risk after percutaneous cor-onary intervention attributable to high plateletreactivity on clopidogrel (from the Assessment ofDual Antiplatelet Therapy with Drug-ElutingStents Study). Am J Cardiol 2017;120:917–23.

35. Ibrahim H, Schutt RC, Hannawi B, DeLao T,Barker CM, Kleiman NS. Association of immatureplatelets with adverse cardiovascular outcomes.J Am Coll Cardiol 2014;64:2122–9.

36. Chitalia VC, Shivanna S, Martorell J, et al.Uremic serum and solutes increase post-vascularinterventional thrombotic risk through alteredstability of smooth muscle cell tissue factor. Cir-culation 2013;127:365–76.

37. Udell JA, Bonaca MP, Collet JP, et al. Long-term dual antiplatelet therapy for secondary pre-vention of cardiovascular events in the subgroupof patients with previous myocardial infarction: acollaborative meta-analysis of randomized trials.Eur Heart J 2016;37:390–9.

KEY WORDS bleeding, complex PCI,coronary stenting, dual antiplatelet therapy,PRECISE-DAPT score

APPENDIX For an expanded Method sectionas well as supplemental tables and figures,please see the online version of this paper.