Kenneth W. MahaffeyDaniel M. WojdylaStefan K. JamesHugo A. KatusSteen HustedGabriel StegChristopher P. CannonRichard C. Becker

Claes HeldNardev KhurmiDebra MontgomeryAnders HimmelmannRobert A. HarringtonLars Wallentin

for the PLATO investigators

Characterization of the Myocardial Infarction Endpoints, Impact of Event Adjudication, and Ticagrelor Effects in the Platelet Inhibition and Patient Outcomes (PLATO) Trial

Background

• Clinical Events Committees (CEC) are common for endpoint adjudication

• In PLATO, a CEC prospectively defined, systematically identified and adjudicated all events

• In this report, we explore:

– Types of MI events reported by site investigators and CEC

– The concordance between the site investigators and CEC

– The treatment effects observed

Worldwide Country ParticipationArgentinAustraliaAustriaBelgiumBrazil BulgariaCanada

GreeceHong KongHungaryIndiaIndonesiaIsraelItaly

Portugal RomaniaRussiaSingaporeSlovakiaSpainSweden

SwitzerlandSouth AfricaSouth KoreaTaiwanThailandTurkeyUkraine

United KingdomUnited States

ChinaCzech RepublicDenmarkFinlandFranceGeorgiaGermany

MalaysiaMexicoThe NetherlandsNew ZealandNorwayPhilippinesPoland

Event Adjudication Process

Suspected events identified from:1. eCRF data (biomarkers; procedures; etc.)2. Monitors during site visits3. Endpoint Office personnel4. Independent Clinical Adjudication

Committee

1. Notification via Endpoint Office2. Source documents from site

when necessary3. Review of all clinical data DisagreeAgree

Phase 2 —Committee

reviewDONE

Phase 1 — MD Reviews

MI DefinitionsAbbreviated

Non-procedural MI:

CK-MB > ULN or Troponin > 99th % AND one of the following:1. Ischemic Symptoms2. ECG changes indicative of ischemia3. New Q-waves

PCI:

CK-MB≥ 3x ULN with 24 hours from a normal or decreasing level or new Q-waves

CABG:

CK-MB≥ 10x ULN or CK-MB≥5x ULN and new Q-waves

Silent MI: New Q-waves without symptoms

Events reported by PIN = 2705

Cardiac Ischemic Event FormSTEMI/NSTEMI* diagnosis

N = 1198

Cardiac Ischemic Event Form‘Other’ diagnoses**

N = 1507

Adjudicated as MI by CEC

N = 862

Not adjudicated as MI by CEC

N = 336

Adjudicated as MI by CEC

N = 187

Not adjudicated as MI by CEC

N = 1320

Not reported on Cardiac Ischemic

Event formN = 251

MI events adjudicated by CEC

N = 1300

* Include events with final diagnosis “Other” with text suggesting MI

**Include events with final diagnosis “Unstable Angina”, “Stable Angina”, and “Other” with text not suggesting MI

Event Level AnalysisCEC

Yes No

Investigator Reported

Yes 862 336

No 438** —

Comparisons: CEC and Site Investigator

** Includes 187 events reported in CIE form but with final diagnosis not STEMI or NSTEMI and 251 events that were not reported in CIE form.

Patient Level AnalysisCEC

Yes No

Investigator Reported

Yes 762 272

No 385 17205

Using the first event for patient with multiple MI events reported or adjudicated

Patient Level Analysis– Overall,

96% agreement– Of those reported by

site investigator,

74% agreement– Of those reported by

CEC

66% agreement

60%

18%

7%

15%

Non-procedural

PCI related

CABG related

Stent throm-bosis related

85%10%

3%1%

Non-proceduralPCI

related

CABG related

Other

CEC MIs Investigator Reported MIs

DisagreementsCEC and Site Investigator

Total Ticagrelor Clopidogrel0

100

200

300

400

500

438

215 223

Reported by CEC but not Site Investigator

Reported by Site Investigator but not CEC

Total Ticagrelor Clopidogrel0

100

200

300

400

500

336

184152

TypeTotal # Events

# EventsTicagrelor

# EventsClopidogrel

HR (95% CI)Tic. vs Clop.

Any 1107 (100%) 508 599 0.840 (0.746 – 0.945)

Non Silent 1097 (99%) 504 593 0.842 (0.748 – 0.948)

Silent 11 (1.0%) 5 6 0.829 (0.253 – 2.716)Non-Procedure Related 652 (59%) 303 349 0.861 (0.738 – 1.005)

Procedure Related 304 (27.5%) 144 160 0.895 (0.715 – 1.122)

PCI Related 223 (20%) 99 124 0.794 (0.609 – 1.034)

CABG Related 82 (7%) 45 37 1.212 (0.784 – 1.872)Associated with Stent Thrombosis 172 (15.5%) 69 103 0.666 (0.491 – 0.903)

Treatment Effect by Type of MI CEC MIs

The first event of each type counted.On patient with one silent and on non-silent MI

Ticagrelor ClopidogrelHR (95% CI) (Tic vs. Clo)

CEC

All MI 508 599 0.840 (0.746–0.945)

STEMI 117 159 0.731 (0.576–0.928)

NSTEMI 356 404 0.875 (0.758–1.008)

Not evaluable 45 57 0.785 (0.531–1.160)

Q-wave 42 46 0.909 (0.598–1.381)

Non Q-wave 333 375 0.881 (0.761–1.022)

Not evaluable 160 208 0.763 (0.621–0.938)

Site Investigator

All MI 459 516 0.883 (0.779–1.001)

STEMI 156 196 0.791 (0.641–0.976)

NSTEMI 287 307 0.929 (0.791–1.091)

Other 29 31 0.931 (0.561–1.545)

Treatment Effect: CEC and Site Investigator

0 60 120 180 240 300 3600%

1%

2%

3%

4%

5%

Days since randomization

Even

t rat

e

Non-proceduralHR 0.86 (0.74–1.01)

PCI relatedHR 0.79 (0.61–1.03)

CABG relatedHR 1.21 (0.78–1.87)

Treatment Effect by MI TypeCEC MIs

Ticagrelor Clopidogrel

Limitations

• This was pre-specified analysis from PLATO but the trial was not powered to evaluate treatment effect in MI alone or in subtypes of MI

• Reporting of MI by site investigators used different format than the CEC adjudication forms so direct event level comparisons were difficult

• We did not contact sites to determine reasons for why the CEC and the site investigators disagreed on individual MI events

Summary

• In PLATO, ticagrelor significantly reduced MI compared with clopidogrel

• The CEC procedures identified more MI endpoints than reported by the Site Investigators particularly procedural related MIs

• A consistent treatment effect was observed across most MI subtypes and for events reported by Site Investigators or CEC

• Comparison of the CEC and Site Investigator showed:– CEC identified 438 events (215 ticagrelor; 223 clopidogrel) not

reported by the investigators– CEC disagreed with Site Investigator reporting for 336 events

(184 ticagrelor; 152 clopidogrel)

• Understanding CEC procedures, MI definitions and treatment effects across MI subtypes is important in interpreting trial results