J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 0 , N O . 7 , 2 0 1 7

ª 2 0 1 7 B Y T H E A M E R I C A N CO L L E G E O F C A R D I O L O G Y F O U N DA T I O N

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

I S S N 1 9 3 6 - 8 7 9 8 / $ 3 6 . 0 0

h t t p : / / d x . d o i . o r g / 1 0 . 1 0 1 6 / j . j c i n . 2 0 1 6 . 1 2 . 0 1 6

Detecting PeriproceduralMyocardial Infarction in ContemporaryPercutaneous Coronary Intervention Trials

Ernest Spitzer, MD,a,b Ton de Vries, MSC,a Rafael Cavalcante, MD, PHD,b Marieke Tuinman, MSC,a

Tessa Rademaker-Havinga, MSC,a Maaike Alkema, MSC,a Marie-Angele Morel, MSC,a Osama I. Soliman, MD, PHD,a,b

Yoshinobu Onuma, MD, PHD,a,b Gerrit-Anne van Es, PHD,c Jan G.P. Tijssen, PHD,c Eugene McFadden, MD,a,d

Patrick W. Serruys, MD, PHDe

ABSTRACT

Fro

Th

Ro

for

co

Mc

Dr

Ju

rep

Ma

OBJECTIVES This study sought to investigate the differences in detecting (e.g., triggering) periprocedural myocardial

infarction (PMI) among 3 current definitions.

BACKGROUND PMI is a frequent component of primary endpoints in coronary device trials. Identification of all

potential suspected events is critical for accurate event ascertainment. Automatic triggers based on study databases

prevent underreporting of events.

METHODS We generated automated algorithms to trigger PMI based on each definition and compared results using

data from the RESOLUTE all comers trial.

RESULTS The operationalization of current PMI definitions was achieved by defining programmable algorithms used to

interrogate the study database. From a total of 636 PMI triggers, we identified 234 for the World Health Organization

extended definition, 382 for the Third Universal definition, and 216 for the Society for Cardiovascular Angiography

and Interventions definition. Differences among the biomarkers used, different cutoff values, and in the hierarchy

among biomarkers within definitions, yielded a different number of triggers, and identified unique triggers for each

definition. Only 38 triggers were consistently identified by all definitions. Availability of ECG data, eCRF data on clinical

presentation, and the reporting of >2 post-procedural values of the same biomarker influenced considerably the number

of PMI triggers identified.

CONCLUSIONS PMI definitions are not interchangeable. The number of triggers identified and consequently the

potential number of events varies significantly, highlighting the importance of rigorous methodology when PMI is

a component of a powered endpoint. Emphasis on collection of biomarkers, ECG data, and clinical status at

baseline may improve the correct identification of PMI triggers. (J Am Coll Cardiol Intv 2017;10:658–66)

© 2017 by the American College of Cardiology Foundation.

P eriprocedural myocardial infarction (PMI) is acomponent of the composite primary endpointin coronary device trials and trials that compare

interventional with surgical revascularization

m the aCardialysis Core Laboratories and Clinical Trial Management, Rott

oraxcenter, Erasmus University Medical Center, Rotterdam, the Netherl

tterdam, the Netherlands; dDepartment of Cardiology, Cork University H

Circulatory Health, National Heart and Lung Institute, Imperial College L

mers trial was funded by Medtronic CardioVascular. Dr. Onuma is a me

Fadden has received honoraria for clinical event committee work for Abbo

. Serruys is a consultant for Abbott Laboratories, AstraZeneca Pharmaceuti

de Medical, Stentys France, Svelte Medical Systems, Inc., and Sino Medic

orted that they have no relationships relevant to the contents of this pap

nuscript received October 5, 2016; revised manuscript received Decembe

strategies. Multiple different definitions of PMI havebeen adopted, although 3 are currently preferred.Detection of all potential suspected events for adjudi-cation by an independent clinical event committee

erdam, the Netherlands; bDepartment of Cardiology,

ands; cEuropean Cardiovascular Research Institute,

ospital, Cork, Ireland; and the eInternational Centre

ondon, London, United Kingdom. The RESOLUTE all

mber of the Advisory Board of Abbott Vascular. Dr.

tt Vascular; and a travel grant fromMenarini Ireland.

cals, Biotronik, Medtronic, Volcano Europe BVBA, St.

al Sciences Technology, Inc. All other authors have

er to disclose.

r 7, 2016, accepted December 15, 2016.

AB BR E V I A T I O N S

AND ACRONYM S

CEC = clinical event committee

CK-MB = creatine kinase-

myocardial band

ECG = electrocardiogram

eCRF = electronic case report

form

NSTEMI = non-ST-elevation

myocardial infarction

PCI = percutaneous coronary

intervention

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 0 , N O . 7 , 2 0 1 7 Spitzer et al.A P R I L 1 0 , 2 0 1 7 : 6 5 8 – 6 6 Detecting Periprocedural Myocardial Infarction

659

(CEC) is critical to ensure complete ascertainment ofendpoints in individual trials and to enable compari-sons among trials. Where biomarker values are avail-able, automatic triggering based on biomarker andother data in study databases can aid in preventingunderreporting of events. Independent adjudicationof clinical events is a critical component of the overallquality of randomized trials (1), allowing for a standard,nonbiased, adjudication of events (2). The identifica-tion of potential events (i.e., triggers) and the collectionof relevant source documents is generally entrusted toclinical research organizations (Figure 1) (3).

SEE PAGE 667PMI = periprocedural

myocardial infarction

RAC = RESOLUTE All-Comers

trial

SCAI = Society for

Cardiovascular Angiography

and Interventions

STEMI = ST-segment elevation

myocardial infarction

TUD = Third Universal

definition

WHO = World Health

Organization

Event triggers may have multiple sourcesincluding: 1) potential events reported by investiga-tors, generally through electronic case report forms(eCRF); 2) potential (unreported by the investigator)events detected by monitoring of source documents;3) eCRF-derived programmable event triggers; 4) corelaboratory-derived event triggers, such as the detec-tion of potential stent thromboses; and 5) event trig-gers identified by the CEC members during review ofsource documentation related to other events. eCRF-derived programmable event triggers are identifiedthrough algorithms with specific cutoff values forbiomarkers, clinical parameters, and findings ondiagnostic techniques. This process presupposesstructured integration of the eCRF and the CEC data-base, and the core laboratory databases when avail-able (e.g., electrocardiogram [ECG], angiography).

Three commonly used definitions of PMI areencountered in current percutaneous coronary inter-vention (PCI) trials. First, the World Health Organi-zation (WHO) extended definition, which wasdeveloped in the context of the RAC (RESOLUTE All-Comers) trial. The RAC trial originally aimed to usethe WHO historical definition (4), because of thepreference of the industry at that time aiming toensure comparability and poolability of data amongtrials (5,6). However, given the need for adjudicationof event triggers in which biomarkers were notavailable, the WHO extended definition, which is theoldest considered in this analysis, introduced a hier-archy that would allow adjudication even in theabsence of complete biomarker data. Second, theThird Universal definition (TUD) of the Joint Euro-pean Society of Cardiology/American College ofCardiology Foundation/American Heart Association/World Heart Federation Task Force (7), which hastroponin as the preferred biomarker always inconjunction with ancillary criteria, such as symptomsor ECG findings consistent with ischemia or angio-graphic complications. Finally and most recently, the

Society for Cardiovascular Angiography andInterventions (SCAI) provided recommenda-tions for the diagnosis of PMI based on crea-tine kinase-myocardial band (CK-MB) as thepreferred biomarker and with a markedlyhigher threshold based on evidence suggest-ing that this threshold, irrespective of othercriteria, is associated with clinically relevantPMI (8).

To accurately identify event triggers basedon these definitions, we aimed to operation-alize them into programmable algorithms. Forvalidation purposes, we used the study data-base of the RAC trial, an investigation spon-sored by Medtronic CardioVascular, whichcompared drug-eluting stents releasing eitherzotarolimus or everolimus (9).

METHODS

PMI definitions were derived from the orig-inal sources (5,7,8); algorithms drafted; anddiscussed in several meetings of cardiolo-gists, statisticians, and scientists with rele-

vant expertise in the topic to reach a consensus thatwould permit reliable identification of PMI triggerswhile minimizing false positives. Operationalizationof definitions was primarily based on biomarkervalues. However, when ECG findings or clinicaldetails recorded in the eCRF could help to bettercharacterize the triggers, they were also integrated.

Six specific issues were identified and discussed up-front because these would affect the structure of thealgorithms. First, the interchangeability of units re-ported as multiples by the investigational sitesbecause in most trials these are not uniformlyreported. Thus, upper reference limit was inter-changeably used together with upper limit of normal.Second, all 3 definitions classify the status of thebiomarkers as stable, rising, or falling, based theoret-ically on pre-procedure biomarkers. However, whenonly 1 pre-procedure value of a specific biomarker wasavailable, we allowed comparisons between pre- andpost-PCI values to achieve this classification. Third,the presence of MI at baseline (i.e., before the pro-cedure) was defined as the presence of ST-segmentelevation myocardial infarction/non-ST-segmentelevation myocardial infarction (STEMI/NSTEMI) atbaseline according to the investigator, CK-MB, orcardiac troponin >1 upper limit of normal/upperreference limit (for all 3 definitions), and also if CK >1upper limit of normal/upper reference limit for WHOextended, ignoring the hierarchies. When both base-line biomarkers and clinical diagnosis were missing

FIGURE 1 Flow From Event Trigger Generation to Adjudication of Events

Adjudica onEvent Trigger Prepara on

Inves gator-reported event

triggers

Monitor-detected under-reported event triggers

eCRF-derived programmable event triggers

Core lab-derived event triggers

CEC-generated event triggers

Common event triggers database

Clustering of event triggers

Event triggers requiring source

documents

Inves ga onal sites

Clinical research associates

Data manager

Event triggers with complete source documenta on

CEC adjudica on

Event Trigger Sources

Adjudicated events

CEC ¼ clinical event committee; eCRF ¼ electronic case report form.

Spitzer et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 0 , N O . 7 , 2 0 1 7

Detecting Periprocedural Myocardial Infarction A P R I L 1 0 , 2 0 1 7 : 6 5 8 – 6 6

660

the most sensitive path (i.e., the one with the lowestthreshold for generating a trigger) was chosen.Fourth, ECG data were included in the algorithmsfor WHO extended and SCAI, because this wouldimprove specificity without affecting sensitivity(i.e., biomarker cutoffs change according to ECGfindings). The presence of new Q waves (or left bundlebranch block) after the procedure in patients in stablecondition at baseline, or the presence of significantchanges after the procedure as proxy, would deter-mine the path in the algorithm. When ECG data weremissing, ECG changes were presumed present. Fifth,the hierarchical use of biomarkers was implementedfor all definitions. WHO extended uses first CK, thenCK-MB, and then troponin, with 1 exception (5). TUDuses first troponin, then CK-MB. SCAI uses firstCK-MB, and then troponin. CK is not used in TUD orSCAI. This hierarchy is used regardless of whether pre-procedure information is available for the selectedbiomarker. If the pre-procedure value is not available,a worst case approach is used (value 0 is assumed).Sixth, when no post-procedure biomarkers wereavailable, patients were further classified according tothe presence of MI at baseline, and if ECG findingswere reported.

Once algorithms for all definitions were completed,they were programmed and executed in the studydataset of the RAC trial (9).

RESULTS

The algorithms used for the operationalization ofeach definition are shown in Figures 2 to 5. Theavailability in the RAC trial of parameters used in thealgorithms is presented in Online Table 1.

WHO EXTENDED DEFINITION. The WHO extendeddefinition results are found in Figure 2 and OnlineFigure 1. Out of 2,509 procedures included in thisanalysis (2,292 index and 217 staged), 234 yielded aPMI trigger; 161 were related to the index procedure,and 73 to staged procedures; 114 procedures followeda different algorithm (Figure 5, Online Figure 2)because there were no post-procedure biomarkersavailable. Overall, 916 presented with a STEMI/NSTEMI at baseline, whereas 1,479 were proceduresfor patients in stable condition. From the former, 845had >1 value of the same biomarker post-PCI, whereas71 only had 1 value. When >1 value was available, only8.0% were classified as triggers, whereas when only 1value was available, 22.5% were so classified. Patientsin stable condition were further categorized accordingto the presence of clinically relevant changes in theECG at discharge relative to baseline. A total of 98 of278 (35.3%) with ECG changes (8) or missing ECG data(270) were classified as triggers, whereas 52 of 1,201(4.3%) without ECG changes were so classified.

FIGURE 2 Algorithm for PMI Triggers Based on the World Health Organization Extended PMI Definition (When Post-Procedure Biomarkers Are Available)

Biomarker rulesGroup characteriza on

Myocardial infarc on at inclusion defined as:CK-MB >1xULN/URL or Troponin >1xULN/URL or CK >1xULN/URL, or MI (STEMI, NSTEMI) reported

by inves gator on baseline form

Are there any new ischemic changes in the ECG?(i.e. Q waves, ST-T changes, LBBB)

Without MI at inclusion

Post-procedure CK-MB>1xULN/URL,or Tn>1xULN/URL, or CK>1xULN/URL

If previous post-PCI biomarker ≤1xULN/URL, use:CK>2xULN/URL (plus Tn or CK-MB >1xULN/URL), or CK-MB

>3xULN/UNL, or Tn>3xULN/URL

If previous post-PCI biomarker >1xULN/URL, use:CK>2xULN/URL (plus Tn or CK-MB >1xULN/URL) and ≥50%

of previous level, or CK-MB> 3xULN/URL and ≥50% of previous level, or Tn>3xULN/URL and ≥50% of previous level

CK>2xULN/URL (+Tn or CK-MB >1xULN/URL) and ≥50% of baseline level, or (if CK missing) CK-MB>3xULN/URL and ≥50% of baseline

level, or (if both missing) Tn>3xULN/URL and ≥50% of baseline level

CK>2xULN/URL (plus Tn or CK-MB >1xULN/URL),or CK-MB>3xULN/UNL, or Tn>3xULN/URL

All biomarker checks are performed hierarchically following the defini on

When biomarkers are stable, falling or rising

Only when biomarkers are rising

Decision

With MI at inclusion

Is there >1 measurement of the same biomarker post-PCI (within 48 h)?

Event trigger

No event trigger

Event trigger

No event trigger

Event trigger

No event trigger

Event trigger

No event trigger

Yes

No

Yes

No

Yes

No

Yes

No

Yes

No

Yes

No

CK-MB ¼ creatine kinase-myocardial band; ECG ¼ electrocardiogram; LBBB ¼ left bundle branch block; MI ¼ myocardial infarction; NSTEMI ¼ non-ST-elevation

myocardial infarction; PCI¼ percutaneous coronary intervention; PMI¼ periprocedural myocardial infarction; STEMI¼ ST-segment elevation myocardial infarction; Tn¼troponin; ULN ¼ upper limit of normal; URL ¼ upper reference limit.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 0 , N O . 7 , 2 0 1 7 Spitzer et al.A P R I L 1 0 , 2 0 1 7 : 6 5 8 – 6 6 Detecting Periprocedural Myocardial Infarction

661

THIRD UNIVERSAL DEFINITION. The TUD results arefound in Figure 3 and Online Figure 3. Overall, 382procedures were associated with a PMI trigger; 318were related to the index procedure, and 64 to stagedprocedures; 150 procedures followed a differentalgorithm (Figure 5, Online Figure 2) because therewere no post-procedure biomarkers available; 833presented with a STEMI/NSTEMI at baseline, whereas1,526 were procedures for patients in stable condi-tion. From the former, 756 had >1 value of the samebiomarker post-PCI, whereas 77 only had 1 value.When >1 value was available, 11.8% were classified astriggers, whereas when only 1 value was available,48.1% were so classified. No subcategorizationaccording to ECG findings was performed because forthe TUD, the ECG findings do not mandate differentcutoff values for the biomarkers.

SCAI DEFINITION. The SCAI definition results arefound in Figure 4 and Online Figure 4. Overall, 216procedures were associated with a PMI trigger; 204

were related to the index procedure and 12 to stagedprocedures; 150 procedures followed a different algo-rithm (Figure 5, Online Figure 2) because there were nopost-procedure biomarkers available; 833 presentedwith a STEMI/NSTEMI at baseline, whereas 1,526 wereprocedures for patients in stable condition. From theformer, 756 had >1 value of the same biomarker post-PCI, whereas 77 only had 1 value. When >1 value wasavailable, 185 (24.5%) were classified as triggers,whereas when only 1 value was available, 9 (11.7%)were so classified. Patients in stable condition werefurther categorized according to the presence of clini-cally relevant changes in the ECG at discharge relativeto baseline; 14 of 276 (5.1%) with ECG changes (10) ormissing ECG data (266), were classified as triggers; 8 of1,250 (0.6%) without ECG changes were so classified.

COMPARISON AMONG DEFINITIONS. When all defi-nitions were combined, 636 PMI triggers were identi-fied corresponding to 26.6% of all procedures.However, only 38 triggers (6.0%) were identified based

FIGURE 3 Algorithm for PMI Triggers Based on the Third Universal PMI Definition (When Post-Procedure Biomarkers Are Available)

Biomarker rulesGroup characteriza on

Myocardial infarc on at inclusion defined as:Tn >1xULN/URL, or CK-MB >1xULN/URL, or MI (STEMI, NSTEMI) reported by inves gator on

baseline form

Without MI at inclusion Post-procedure Tn≥5xULN/URL, or (if Tn missing) CK-MB≥3xULN/URL

Rise in Tn values > 20% from the previous nadir level, or (if Tnmissing) CK-MB values > 20% from the previous nadir level

Nadir level is lowest level measured before the tested post-procedure. If screening value unavailable, 0 is assumed

Biomarkers >20% above screening value, and Tn≥5xULN/URL, or (if Tn missing) CK-MB≥3xULN/URL

All biomarker checks are performed hierarchically following the defini on

When biomarkers are stable, falling or rising

With screening values available

Decision

With MI at inclusion

Is there >1 measurement of the same biomarker post-PCI (within 48 h)?

Event trigger

No event trigger

Event trigger

No event trigger

Event trigger

No event trigger

Yes

No

Yes

No

Yes

No

Yes

No

Tn≥5xULN/URL, or (if Tn missing) CK-MB≥3xULN/URL

Without screening value available:

Event trigger

No event trigger

Yes

No

Abbreviations as in Figure 2.

Spitzer et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 0 , N O . 7 , 2 0 1 7

Detecting Periprocedural Myocardial Infarction A P R I L 1 0 , 2 0 1 7 : 6 5 8 – 6 6

662

on all definitions. A Venn diagram showing the distri-bution of triggers is presented in Figure 6A. A total of 91(14.3%), 240 (37.7%), and 147 (23.1%) triggers wereuniquely identified using the WHO extended, TUD,and SCAI definitions, respectively. Concordanceamong 2 definitions was observed in 89 (WHOextended and TUD, 14.0%), 15 (TUD and SCAI, 2.4%),and 16 (WHO extended and SCAI, 2.5%) event triggers.

CALCULATIONS BY MODIFYING THE ALGORITHM.

To further assess the relevance of different parame-ters for triggering, we repeated the analysis excludingclinical presentation (Figure 6B), ECG data (Figure 6C),or both (Figure 6D) from the algorithms. We observeda consistent reduction in the number of triggersobtained with all 3 definitions when clinical presen-tation data were available, and when ECG data wereavailable. When both were excluded from the algo-rithm, we observed an almost 3-fold increase in thenumber of triggers according to the WHO extendeddefinition (from 234 to 626 triggers), and an overall40% increase for all combined definitions. Whenclinical presentation was excluded, 15% of triggers

were detected consistently by all definitions; 7.6%when ECG data were excluded, and, 11.7% when bothwere excluded (Figures 6B to 6D).

CALCULATIONS USING THE PEAK VALUE ONLY.

When using only the peak post-procedural value of abiomarker, instead of all post-procedural valuesavailable, the number of triggers increased from636 to 867. With the SCAI definition using only 1biomarker (214 with peak value vs. 216 with all valuesavailable) there was no material effect on the numberof triggers. There were few cases only identified whenusing peak values (5) or all values (3). However, basedon the WHO extended and TUD, we observed an in-crease in the number of triggers from 234 to 470, andfrom 382 to 782, respectively.

CALCULATIONS WHEN ALL DATA ARE AVAILABLE.

The number of triggers calculated for index proced-ures without missing data (Figure 6G) using allalgorithms was 4 percentage points lower than forindex procedures with any missing data (Figure 6F)(from 27% to 23%).

FIGURE 4 Algorithm for PMI Triggers Based on the SCAI PMI Definition (When Post-Procedure Biomarkers Are Available)

Biomarker rulesGroup characteriza on

Myocardial infarc on at inclusion defined as:CK-MB >1xULN/URL, or Tn >1xULN/URL,

or MI (STEMI, NSTEMI) reported by inves gator on baseline form

Are there any new ischemic changes in the ECG?(i.e. Q waves, ST-T changes, LBBB)

Without MI at inclusion

Post-procedure CK-MB≥5xULN/URL, or (if CK-MB missing) Tn≥35xULN/URL

New CK-MB eleva on by an absolute increase of ≥10xULN/URL, or (if CK-MB missing) Tn≥70xULN/URL from the previous nadir level

Nadir level is lowest level measured before the tested post-procedure. If screening unavailable value 0 is assumed.

CK-MB eleva on by an absolute increase of ≥10xULN/URL, or (if CK-MB missing) Tn≥70xULN/URL from the baseline level CK-MB

Post-procedure CK-MB ≥10xULN/URL, or (if CK-MB missing) Tn≥70xULN/URL

All biomarker checks are performed hierarchically following the defini on

When biomarkers are stable, falling or rising

Only when biomarkers are rising

Decision

With MI at inclusion

Is there >1 measurement of the same biomarker post-PCI (within 48 h)?

Event trigger

No event trigger

Event trigger

No event trigger

Event trigger

No event trigger

Event trigger

No event trigger

Yes

No

Yes

No

Yes

No

Yes

No

Yes

No

Yes

No

Abbreviations as in Figure 2.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 0 , N O . 7 , 2 0 1 7 Spitzer et al.A P R I L 1 0 , 2 0 1 7 : 6 5 8 – 6 6 Detecting Periprocedural Myocardial Infarction

663

DISCUSSION

The main findings of the present investigationregarding triggering PMI based on 3 current defini-tions are as follows:

1. Operationalization of PMI definitions (WHOextended, TUD, and SCAI) for event triggering wasaccomplished by the development of 3 definition-specific algorithms;

2. Different PMI definitions (WHO extended, TUD,and SCAI) yielded markedly different numbers ofPMI triggers and, not uncommonly, were mutuallyexclusive;

3. Differences among the number of triggers relatedto the different definitions are explained bythe biomarkers used, the hierarchies in biomarkerselection, and the cutoff values;

4. The careful collection of pre-procedural bio-markers (ideally 2 for the TUD), at least 2 sets ofpost-procedural biomarkers, the documentation ofnew ischemic ECG changes at discharge relative to

baseline (or relative to post-index for staged pro-cedures), and recording of clinical presentation atbaseline (i.e., STEMI/NSTEMI vs. others) improvesidentification of PMI triggers, potentially reducingadjudication costs.

Randomized trials use event rates to compare thesafety and efficacy of 2 or more treatment strategies,devices, or pharmacological treatments. Conse-quently, the meticulous detection of events is crucialto guarantee a fair comparison (10). In large multi-center trials this endeavor is undertaken by a jointcollaboration among the investigational sites throughaccurate and comprehensive reporting of events,clinical research associates acting as site monitorswith the responsibility of reviewing charts of patientsenrolled and prevent underreporting of events, corelaboratory-derived event triggers (i.e., angiographicor ECG findings), and a programmed review of thedatabase searching for abnormal values (1,11). Multi-ple event triggers that may result in the adjudicationof an endpoint need to be identified and the relevant

FIGURE 5 Algorithm for PMI Triggers Based on All Definitions (When Post-Procedure Biomarkers Are Not Available)

SCAI ¼ Society for Cardiovascular Angiography and Interventions; WHO ¼ World Health Organization; other abbreviations as in Figures 1 and 2.

Spitzer et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 0 , N O . 7 , 2 0 1 7

Detecting Periprocedural Myocardial Infarction A P R I L 1 0 , 2 0 1 7 : 6 5 8 – 6 6

664

source documents collected before presentation tothe CEC (Figure 1).

Exhaustive detection of triggers involves a step-wise process that can be customized for a trial. Themain source of triggers is those reported by theinvestigator. Although these triggers are clinicallyaccepted and represent the expert opinion of thetreating physician or local investigator, noise may beintroduced because of variable compliance withprotocol-mandated definitions. A local or remotemonitor who reviews source documents to verifycompleteness of reporting should perform an initialvalidation. Relevant considerations for themonitoringactivities include the independence of the monitorsand the percentage of source documents that aremonitored. Additional elements of validation includethe eCRF (highly dependent on the appropriateness ofthe questions and the completeness of data), the corelaboratories (for biomarker assessment, ECG inter-pretation, or angiographic analysis), and the in-depthreview of source documents by the adjudicators, whomay identify additional, unreported events.

eCRF-derived event triggers are defined by rules,such as cutoff values of troponin for PMI, or cutoffvalues of hemoglobin for anemia (3). These rules needto be as inclusive as possible so as not to misspotential events (false negatives) while minimizingthe inclusion of nonevents (false positives). Thedevelopment of trigger rules involves the strictimplementation of the definitions delineated in the

investigational protocols and should properly balancea trade-off between ensuring quality while mini-mizing cost. Once an event trigger is generated, itadds cost to the trial (source document collection,digitalization, dossier preparation, and adjudication,among others); thus, event triggers are of significantrelevance for investigators and sponsors whiledesigning a trial.

PMI definitions have changed over time. Thecontinuous progress in cardiovascular medicine hasbrought a shift from nonspecific enzymes, such as CKand lactate dehydrogenase, toward more specificbiomarkers, such as cardiac CK-MB, and ultimatelycardiac troponins. These changes have had a signifi-cant impact in definitions. For instance, CK is nolonger used in the SCAI or TUD. Moreover, whereastroponin is the preferred biomarker for TUD, SCAIstrongly recommends the use of CK-MB as thepreferred biomarker based on robust historical data.

Our results demonstrate that the TUD providesthe highest number of event triggers given thepreferential use of troponin, and the lower cutoffsas compared with SCAI. Nonetheless, the lack ofsystematic collection in the eCRF of ancillarycriteria is problematic for consistent event adjudi-cation. The WHO extended definition, althoughscientifically interesting for the present analysis, mayprogressively be abandoned and is mostly used toallow data comparability. The SCAI definition resultsin a lower number of triggers. The rationale proposed

FIGURE 6 Distribution of Periprocedural Myocardial Infarction Triggers Among Definitions and Variations Secondary to Data Availability

147

3816 15

89 24091

Using all data

100

1046 15

113 26693

Excluding Presenta on

146

6218 15

157 148267

Excluding ECG data

99

1499 16

184 149284

Excluding BothA CB DTotal: 636/2509 (25%) Total: 697/2509 (28%) Total: 813/2509 (32%) Total: 890/2509 (35%)

4

1942 14

195 37979

Using peak biomarker onlyETotal: 867/2509 (35%)

6.4

1.20.7 0.7

2.4 9.72.8

Using all data (%)

6.9

0.80.5 0.8

1.8 9.92.4

With complete data (%)

3.2

3.21.5 0

5.3 8.65.6

With incomplete data (%)F HGTotal: 545/2292 (24%) Total: 452/1953 (23%) Total: 93/339 (27%)

SCAI defini on Third Universal defini on WHO-extended defini on All data Index procedures excluding staged procedures

Abbreviations as in Figures 2 and 5.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 0 , N O . 7 , 2 0 1 7 Spitzer et al.A P R I L 1 0 , 2 0 1 7 : 6 5 8 – 6 6 Detecting Periprocedural Myocardial Infarction

665

for this definition is that it identifies patients withclinically relevant PMI and is therefore more appro-priate as a component of a composite primaryendpoint (12). It is noteworthy that TUD may bepreferred for trials in which a higher rate of events isneeded because of statistical power and whenmechanistic comparisons are aimed. These includestent versus stent comparisons (e.g., different strutthickness or stent design) and comparisons amongdrugs administered before PCI. The SCAI definitionmay be preferred for large randomized controlledtrials because of its simplicity and less costs involved,and in strategy trials comparing PCI with coronaryartery bypass graft. However, the relative merits ofthe Universal MI definition and the SCAI definition ofPMI are not addressed in the current investigation.

The value of a database that collects specificitems captured in the definitions of PMI has beenhighlighted by our results. When developing the eCRFinvestigators should include the clinical presentation;presence of prolonged chest pain (>20 min) during theprocedure; information regarding the baseline andpre-discharge ECGs, and more specifically the

presence of ischemic ST-segment changes, new path-ological Q waves, or new left bundle branch block; andthe presence of angiographic periprocedural compli-cations defined as slow-flow or no-flow at the end ofthe procedure (Thrombolysis In Myocardial Infarctionflowgrade<3), occlusion of amajor coronary artery or aside branch, or distal embolization (13). Imagingdemonstration of new loss of viable myocardium ornew regional wall motion abnormality in patients instable condition undergoing PCI should also beincluded. These elements would be investigator-reported with evident limitations but would provide,with appropriate verification in place, a more rationalbasis for comparison among trials. Ideally, electrocar-diographic and angiographic assessment should beperformed by a central core laboratory and the resultsprovided to the CEC (6,14).

Detecting PMI outside the setting of a randomizedcontrolled trial creates additional challenges. Clinicalcenters may have preferences among definitions,which could even be the case for individual physi-cians within 1 single institution. Consequently, largeprospective registries need standardization of

PERSPECTIVES

WHAT IS KNOWN? PMI is a frequent component of

primary endpoints in coronary device trials.

WHAT IS NEW? Accurate detection of PMI requires

careful interpretation of the definitions, because

those may largely affect event rates. Contempora-

neous definitions of PMI may be mutually exclusive

for detecting potential events.

WHAT IS NEXT? Emphasis on collection of bio-

markers, ECG data, and clinical status at baseline may

improve the correct identification of PMI triggers.

Spitzer et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 0 , N O . 7 , 2 0 1 7

Detecting Periprocedural Myocardial Infarction A P R I L 1 0 , 2 0 1 7 : 6 5 8 – 6 6

666

definitions for voluntary reporting. This has provenuseful in the National Cardiovascular DatabaseRegistry CathPCI Registry where the impact of PMIdefinitions on observed rate of events has beenrecently explored (15).

STUDY LIMITATIONS

Although we were able to show important differencesamong PMI triggers according to the 3 different defi-nitions, the sensitivity and specificity of the triggerscan only be tested if triggering and adjudication areperformed following all 3 definitions. It remainsimportant to elucidate the impact of using differentdefinitions in a single trial, because not only does thenumber of triggers change, but also the individualprocedures potentially associated with a PMI.

CONCLUSIONS

PMI definitions are not interchangeable. The identifiednumber of triggers and consequently the potentialnumber of events vary significantly, prompting carefulconsideration when PMI is a component of a poweredendpoint. Rigorous collection of biomarkers, ECGdata,

and clinical status at baseline may improve the correctidentification of PMI triggers.

ADDRESS FOR CORRESPONDENCE: Dr. Patrick W.Serruys, International Centre for Circulatory Health,National Heart and Lung Institute, Imperial CollegeLondon, 59 NorthWharf Road, LondonW2 1LA, UnitedKingdom. E-mail: patrick.w.j.c.serruys@gmail.com.

RE F E RENCE S

1. Vranckx P, McFadden E, Mehran R, Cutlip DE.Clinical event committees in coronary stent trials:insights and recommendations based on experi-ence in an unselected study population. Euro-Intervention 2012;8:368–74.

2. Mahaffey KW, Harrington RA, Akkerhuis M,et al. Disagreements between central clinicalevents committee and site investigator assess-ments of myocardial infarction endpoints in aninternational clinical trial: review of the PURSUITstudy. Curr Control Trials Cardiovasc Med 2001;2:187–94.

3. Mahaffey KW, Harrington RA, Akkerhuis M,et al. Systematic adjudication of myocardialinfarction end-points in an international clinicaltrial. Curr Control Trials Cardiovasc Med 2001;2:180–6.

4. Nomenclature and criteria for diagnosis ofischemic heart disease. Report of the Joint Inter-national Society and Federation of Cardiology/World Health Organization task force on stan-dardization of clinical nomenclature. Circulation1979;59:607–9.

5. Vranckx P, Cutlip DE, Mehran R, et al.Myocardial infarction adjudication in contempo-rary all-comer stent trials: balancing sensitivityand specificity. Addendum to the historical MIdefinitions used in stent studies. Euro-Intervention 2010;5:871–4.

6. Cavalcante R, Serruys PW. Periproceduralmyocardial infarction in stent trials: how universalis the third universal definition? EuroIntervention2016;12:813–7.

7. Thygesen K, Alpert JS, Jaffe AS, et al. ThirdUniversal Definition of myocardial infarction. Cir-culation 2012;126:2020–35.

8. Moussa ID, Klein LW, Shah B, et al. Consider-ation of a new definition of clinically relevantmyocardial infarction after coronary revasculari-zation: an expert consensus document from theSociety for Cardiovascular Angiography and In-terventions (SCAI). J Am Coll Cardiol 2013;62:1563–70.

9. Serruys PW, Silber S, Garg S, et al. Comparisonof zotarolimus-eluting and everolimus-elutingcoronary stents. N Engl J Med 2010;363:136–46.

10. Vranckx P, McFadden E, Cutlip DE, et al.Clinical endpoint adjudication in a contemporaryall-comers coronary stent investigation: method-ology and external validation. Contemp Clin Trials2013;34:53–9.

11. Mahaffey KW, Roe MT, Dyke CK, et al. Mis-reporting of myocardial infarction end points:results of adjudication by a central clinical eventscommittee in the PARAGON-B trial. SecondPlateletIIb/IIIa Antagonist for the Reduction of Acute Cor-onary Syndrome Events in a Global OrganizationNetwork Trial. Am Heart J 2002;143:242–8.

12. Park DW, Kim YH, Yun SC, et al. Frequency,causes, predictors, and clinical significance of peri-procedural myocardial infarction following percu-taneous coronary intervention. Eur Heart J 2013;34:1662–9.

13. Ishibashi Y, Muramatsu T, Nakatani S, et al.Incidence and potential mechanism(s) of post-procedural rise of cardiac biomarker in patientswith coronary artery narrowing after implantationof an everolimus-eluting bioresorbable vascularscaffold or everolimus-eluting metallic stent. J AmColl Cardiol Intv 2015;8:1053–63.

14. Cavalcante R, Sotomi Y, Onuma Y. Polymer-free drug-coated coronary stents. N Engl J Med2016;374:892.

15. Baker NC, Lipinski MJ, Escarcega RO, et al.Definitions of periprocedural myocardial infarctionas surrogates for catheterization laboratory qual-ity or clinical trial end points. Am J Cardiol 2014;113:1326–30.

KEY WORDS clinical trial, percutaneouscoronary intervention, periproceduralmyocardial infarction

APPENDIX For supplemental tables andfigures, please see the online version of thisarticle.