clinical presentation, management, and outcomes of ... · figure 3. in-hospital mortality based on...

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 . 5 , N O . 2 , 2 0 1 2

© 2 0 1 2 B Y T H E A M 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 U N D A T I O N I S S N 1 9 3 6 - 8 7 9 8 / $ 3 6 . 0 0

P U B L I S H E D B Y E L S E V I E R I N C . D O I : 1 0 . 1 0 1 6 / j . j c i n . 2 0 1 1 . 1 0 . 0 1 3

CLINICAL RESEARCH

Clinical Presentation, Management, andOutcomes of Angiographically DocumentedEarly, Late, and Very Late Stent Thrombosis

Ehrin J. Armstrong, MD,* Dmitriy N. Feldman, MD,† Tracy Y. Wang, MD, MHS, MS,‡Lisa A. Kaltenbach, MS,‡ Khung-Keong Yeo, MBBS,* S. Chiu Wong, MD,†John Spertus, MD,§ Richard E. Shaw, PHD,� Robert M. Minutello, MD,†Issam Moussa, MD,¶ Kalon K. L. Ho, MD,# Jason H. Rogers, MD,*Kendrick A. Shunk, MD, PHD**††

Sacramento and San Francisco, California; New York, New York; Durham, North Carolina;Kansas City, Missouri; Jacksonville, Florida; and Boston, Massachusetts

Objectives The aim of this study was to describe differences in treatment and in-hospital mortalityof early, late, and very late stent thrombosis (ST).

Background Early, late, and very late ST may differ in clinical presentation, management, and in-hospital outcomes.

Methods We analyzed definite (angiographically documented) ST cases identified from February2009 to June 2010 in the CathPCI Registry. We stratified events by timing of presentation: early (�1month), late (1 to 12 months), or very late (�12 months) following stent implantation. Multivariablelogistic regression modeling was performed to compare in-hospital mortality for each type of STafter adjusting for baseline comorbidities.

Results During the study period, 7,315 ST events were identified in 7,079 of 401,662 patients (1.8%)presenting with acute coronary syndromes. This ST cohort consisted of 1,391 patients with early ST(19.6%), 1,370 with late ST (19.4%), and 4,318 with very late ST (61.0%). Subjects with early ST had ahigher prevalence of black race and diabetes, whereas subjects with very late ST had a higher preva-lence of white race and a lower prevalence of prior myocardial infarction or diabetes. In-hospital mortal-ity was significantly higher in early ST (7.9%) compared with late (3.8%) and very late ST (3.6%, p � 0.001).This lower mortality for late and very late ST persisted after multivariable adjustment (odds ratio:0.53 [95% confidence interval (CI): 0.36 to 0.79] and 0.58 [95% CI: 0.43 to 0.79], respectively).

Conclusions Significant differences exist in the presentation and outcomes of early, late, and very late ST.Among patients with acute coronary syndromes who are undergoing percutaneous coronary intervention forangiographically documented ST, early ST is associated with the highest in-hospital mortality. (J Am CollCardiol Intv 2012;5:131–40) © 2012 by the American College of Cardiology Foundation

From the *Division of Cardiology, University of California, Davis Medical Center, Sacramento, California; †Division ofCardiology, New York Presbyterian Hospital, Weill Cornell Medical College, New York, New York; ‡Duke Clinical ResearchInstitute, Division of Cardiology, Duke University Medical Center, Durham, North Carolina; §Saint Luke’s Mid America HeartInstitute, Division of Cardiology, University of Missouri, Kansas City, Missouri; �Division of Cardiology, California PacificMedical Center, San Francisco, California; ¶Division of Cardiology, Mayo Clinic, Jacksonville, Florida; and the #Division ofCardiology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts; **Division of Cardiology,University of California, San Francisco, Veterans Affairs Medical Center, San Francisco, California; and the ††Division ofCardiology, University of California, San Francisco, San Francisco, California. This research was supported by the American Collegeof Cardiology Foundation’s NCDR (National Cardiovascular Data Registry). The views expressed in this manuscript represent those of
the author(s), and do not necessarily represent the official views of the NCDR or its associated professional societies identified at theNCDR website. CathPCI Registry is an initiative of the American College of Cardiology Foundation and the Society for Cardiovascular

itIpdaaotadptiaidarpooabtNeas

P2faadrcSsosiIoao1

b

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 . 5 , N O . 2 , 2 0 1 2

F E B R U A R Y 2 0 1 2 : 1 3 1 – 4 0

Armstrong et al.

Outcomes After ST

132

Stent thrombosis (ST) is an uncommon but highly morbidcomplication of percutaneous coronary intervention (PCI),with 10% to 20% mortality at 1 year (1–4). Causes of ST canbe grouped according to factors related to the stent, theprocedure, and the patient (5). Early ST (occurring within 1month of stent implantation) is more commonly due toprocedure-related complications (6). In contrast, late ST (oc-curring 1 to 12 months after stent implantation) and very late

See page 141

ST (occurring �12 months after stent implantation) may becaused by impaired re-endothelialization, stent malapposi-tion or fracture, a hypersensitivity reaction to the polymer usedin drug-eluting stents (DES), or de novo plaque rupture(7–10). Inadequate antiplatelet therapy and/or resistance toantiplatelet agents are also associated with an increased risk forST (11,12). DES have a similar rate of early and late STrelative to bare-metal stents (BMS), but DES may be associ-

ated with increased rates of verylate ST (13). These heterogeneouscauses of ST may result in differ-ent presentations and outcomesfor early, late, and very late ST.

Prior studies have identifiedclinical predictors of ST after stentimplantation (2,14,15), but mostof these studies have limitedpower to detect differences in ratesof major adverse cardiovascularevents with treatment for early,late, or very late ST. To addressthis gap in knowledge, we identi-fied definite, angiographicallydocumented, ST in the CathPCI

registry and analyzed differences in the presentation, treatment,and in-hospital mortality of ST as a function of time sinceoriginal stent implantation.

Angiography and Interventions. Dr. Feldman has received consulting fees from Maquet,Gilead Sciences, and is on the Speaker’s Bureau for Eli Lilly, Daiichi Sankyo, AbbottVascular, and The Medicines Company. Dr. Wang has received consulting fees fromMedco and AstraZeneca; grant support from BMS-Sanofi, Schering-Plough/Merck, EliLilly, Daiichi Sankyo, Canyon Pharmaceuticals, Heartscape, Medco, AstraZeneca, andThe Medicines Company; and payment for lectures and educational conferences fromthe American College of Cardiology Foundation. Ms. Kaltenbach is an employee of theDuke Clinical Research Institute, which receives funding support from the AmericanCollege of Cardiology-NCDR. Dr. Spertus has a contract with the American College ofCardiology Foundation to serve as an Analytic Center for the NCDR. Dr. Minutello ison the Speaker’s Bureau for Merck and Possis. Dr. Rogers has received consulting feesfrom Volcano, Medtronic, Cordis, and Boston Scientific. Dr. Shunk has receivedinstitutional research support from Siemens Medical Systems, Abbott Vascular, andInfraRedx. All other authors have reported that they have no relationships relevant to thecontents of this paper to disclose.

Abbreviationsand Acronyms

BMS � bare-metal stent(s)

CI � confidence interval

DES � drug-eluting stent(s)

OR � odds ratio

PCI � percutaneouscoronary intervention

ST � stent thrombosis

STEMI � ST-segmentelevation myocardialinfarction

TIMI � Thrombolysis InMyocardial Infarction

fManuscript received August 16, 2011; revised manuscript received October 19, 2011,accepted October 28, 2011.

Methods

Study population. The CathPCI registry (version 4.3) is annitiative of the American College of Cardiology Founda-ion and the Society for Cardiovascular Angiography andnterventions. The registry receives data from over 1,000articipating hospitals across the United States, and itsatabase includes demographic factors, procedural details,nd in-hospital outcomes of patients undergoing coronaryngiography and PCI. The CathPCI registry is composedf standard data collection forms with explicit data defini-ions, uniform data entry and transmission requirements,nd data quality checks (16). The NCDR (National Car-iovascular Data Registry) has a data quality program inlace to ensure consistent and reliable data. The institu-ional review board of each hospital approved participationn the CathPCI registry. Participating sites collected data inn anonymous fashion for consecutive patients who met thenclusion criteria during the initial hospitalization. Built-inata quality checks include exclusion of outliers, edit checkst time of entry, consistency checks, quarterly reports, andandom site audits by trained abstractors. In addition,articipating facilities are afforded a wide variety of trainingptions. Details on the data collection process have previ-usly been published, and the data definitions can be foundt NCDR’s Website (17). All data elements were proposedy the CathPCI Version 4 Dataset Workgroup, reviewed byhe CathPCI Steering Committee, and approved by theCDR Science Quality and Oversight Committee. Data

lements are recorded at the time of cardiac catheterizationnd then forwarded electronically to a secure data servertored at the ACC Heart House in Washington, DC.

For this study, we examined registry patients undergoingCI procedures from February 3, 2009, through June 29,010. ST was identified using a pre-defined data elementor ST by the participating centers at the time of PCIccording to Figure 1. To ensure quality of data recordingnd to model closely the Academic Research Consortiumefinition of definite ST (18), we also required the angiog-apher to indicate that the patient presented with an acuteoronary syndrome; that a prior PCI had occurred; and thatT was present. To refine the cohort, we excluded anyubject in which the lesion was encoded as a chronic totalcclusion or if the date of implantation of the thrombosedtent was missing or unknown. Furthermore, in cases wherenitial flow was categorized as Thrombolysis In Myocardialnfarction (TIMI) flow grade 3, the case was excluded if theperator failed to indicate the presence of visible thrombust the time of intervention. Early ST was defined asccurring within 1 month of prior stent implantation, late asto 12 months, and very late as �1 year.Baseline demographics and procedural details were reported

y participating centers. Race/ethnicity was based on patient or
amily self-report, with options consisting of white, black,

o(wsai

amSpavcDI

i

tmu(twcdgcc

R

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 . 5 , N O . 2 , 2 0 1 2 Armstrong et al.

F E B R U A R Y 2 0 1 2 : 1 3 1 – 4 0 Outcomes After ST

133

Hispanic, Asian, Native American, or other. Because theCathPCI registry only contains data related to a given PCIprocedure (i.e., there is no ability to link patients acrossdifferent interventions), procedural details of the initial stentimplantation could not be ascertained other than the initialstent type (BMS vs. DES). This study was approved by theDuke University Medical Center and was determined to meetcriteria for research not requiring informed consent.Clinical endpoints. The primary endpoint was in-hospitalmortality. Secondary endpoints during the hospital admis-sion included occurrence of a cerebrovascular event (definedas a focal neurologic deficit lasting �24 h due to an ischemicr hemorrhagic cause), a bleeding event within 72 hdefined as �3 g/dl drop in hemoglobin, transfusion ofhole blood or packed red blood cells, or an intervention or

urgical procedure to correct bleeding related to the PCI),nd lesion treatment success (defined as TIMI flow grade 3n the culprit vessel at the conclusion of the procedure).Statistical analysis. Baseline patient subgroups are presenteds counts with percentages for categorical variables andedian with interquartile range for continuous variables.

tatistical comparison of between-group differences waserformed using chi-square test for categorical variables andnalysis of variance or the Kruskal-Wallis test for continuousariables. Statistical significance was defined as p � 0.05 for allomparisons. All statistical analyses were performed by theuke Clinical Research Institute using SAS (version 9.3, SAS

nstitute, Cary, North Carolina).Multivariable logistic regression modeling with general-

Figure 1. Study Population

Records retrieved from the CathPCI registry database (version 4.3). ACS � acuST � stent thrombosis; TIMI � Thrombolysis In Myocardial Infarction.

zed estimating equations was performed to evaluate if the

iming of ST was an independent predictor of in-hospitalortality while adjusting for within hospital correlation

sing a previously validated model for mortality after PCI19). Early ST was used as the reference group. Variables inhe mortality model included: age, sex, race, presentationith ST-segment elevation myocardial infarction (STEMI),

ardiogenic shock, previous heart failure, cerebrovascularisease, peripheral vascular disease, chronic lung disease,lomerular filtration rate, dialysis, New York Heart Asso-iation functional class heart failure, highest risk segmentategory, PCI status, diabetes, and lesion bifurcation.

esults

A total of 7,315 ST cases with known timing occurred in7,079 patients at 977 hospitals during the study period(Fig. 1). Stent thromboses represented 1.8% of all PCI foracute coronary syndromes during the study period. A totalof 1,391 early, 1,370 late, and 4,318 very late ST wereidentified, representing 19.6%, 19.4%, and 60% of the studypopulation, respectively. Among subjects with very late ST,925 (21.4%) occurred 1 to 2 years after an index PCI, and3,393 (78.6%) occurred �2 years after index PCI.

Patient characteristics differed based on the timing of ST(Table 1). The prevalence of black race, diabetes, and priorheart failure was highest among subjects with early ST, andthese subjects were more likely to present with cardiogenicshock than the other ST groups were. Patients with late STwere more likely to have previous coronary artery bypass

nary syndrome(s); PCI � percutaneous coronary intervention;
te coro
graft and renal failure and to present with unstable angina


F E B R U A R Y 2 0 1 2 : 1 3 1 – 4 0

Armstrong et al.

Outcomes After ST

134

rather than acute myocardial infarction. The group present-ing with very late ST had the highest prevalence of whiterace and a lower prevalence of prior myocardial infarction,renal failure, or diabetes.

Most patients had TIMI flow grade 0 (complete throm-botic occlusion) on initial coronary angiography (Table 2);TIMI flow grade 0 was more common in early ST (82%) thanin late (61.9%) or very late ST (68.3%). The culprit lesion wasmore commonly located in the left anterior descending coro-nary artery (48%) in patients with early and late ST, whereas

Table 1. Demographic Factors and Presentation of P

Early ST(n � 1,391)

Age, yrs 61.0 (52–71)

Male 933 (67.1)

Race/ethnicity

White 1,150 (82.7)

Black 189 (13.6)

Asian 30 (2.2)

Hispanic 90 (5.6)

Body mass index, kg/m2 29.0 (25–34)

Current/recent smoker 582 (41.8)

HTN 1,193 (85.8)

Dyslipidemia 1,191 (85.6)

Family history of CAD 317 (22.8)

Prior MI 988 (71.0)

Prior HF 237 (17.0)

Prior CABG 204 (14.7)

GFR 73.9 (57–91)

Dialysis 50 (3.6)

Cerebrovascular disease 184 (13.2)

Peripheral vascular disease 204 (14.7)

Chronic lung disease 264 (19.0)

Diabetes 608 (43.7)

Noninsulin-requiring 346 (24.9)

Insulin-requiring 262 (18.8)

Type of previous stent

BMS 456 (32.8)

DES 827 (59.5)

Unknown 104 (7.5)

Symptoms on admission

Unstable angina 156 (11.2)

NSTEMI 306 (22.0)

STEMI 929 (66.8)

Pre-procedure shock 155 (11.1)

Pre-procedure IABP 8 (4.4)

Cardiac arrest 87 (6.3)

Cardiogenic shock 186 (13.4)

Values are median (interquartile range) or n (%).

BMS � bare-metal stent(s); CABG � coronary artery bypass graft; CA

filtration rate (calculated using the MDRD [Modification of Diet in Ren

intra-aortic balloon pump; MI � myocardial infarction; NSTEMI � non–

ST-segment elevation myocardial infarction.

ST of the right coronary artery was more common (38%)

among subjects with very late ST. Aspiration thrombectomywas performed in 29% to 34% of patients. More than one-halfof patients had a new stent placed during treatment for ST.The rate of new stent placement was lowest in early ST(51.2%) and highest in very late ST (69.9%). A new DES wasplaced in 47% of patients with very late ST, and 23% ofpatients had a BMS placed (Fig. 2). Subjects with ST of apreviously implanted DES were more likely to have a newDES placed as treatment for the ST event. Subjects with ST ofa previously implanted BMS were more likely to have a BMS

s With ST

Late STn � 1,370)

Very Late ST(n � 4,318) p Value

60.0 (50–70) 61.0 (53–69) 0.02

980 (71.5) 3,325 (77.0) �0.001

�0.001

,159 (84.6) 3,900 (90.3) �0.001

158 (11.5) 294 (6.8) �0.001

28 (2.0) 62 (1.4) 0.10

96 (7.0) 169 (3.9) �0.001

28.4 (25–33) 28.6 (25–33) 0.17

611 (44.6) 1,798 (41.6) 0.15

,199 (87.5) 3,648 (84.5) 0.02

,208 (88.2) 3,861 (89.4) 0.001

359 (26.2) 1,028 (23.8) 0.09

989 (72.2) 2,920 (67.6) 0.002

226 (16.5) 461 (10.7) �0.001

329 (24.0) 576 (13.3) �0.001

70.1 (54–90) 72.6 (58–90) 0.012

57 (4.2) 58 (1.3) �0.001

195 (14.2) 462 (10.7) 0.001

226 (16.5) 505 (11.7) �0.001

252 (18.4) 659 (15.3) 0.001

513 (37.5) 1,207 (28.0) �0.001

272 (19.9) 821 (19.0)

241 (17.6) 386 (8.9)

�0.001

421 (30.7) 738 (17.1)

785 (57.3) 2,374 (54.5)

161 (11.8) 1,194 (27.7)

�0.001

295 (21.5) 462 (10.7)

362 (26.4) 954 (22.1)

713 (52.0) 2,902 (67.2)

111 (8.1) 356 (8.2) 0.003

5 (4.2) 13 (3.3) 0.97

69 (5.0) 267 (6.2) 0.26

142 (10.4) 402 (9.3) �0.001

ronary artery disease; DES � drug-eluting stent(s); GFR � glomerular

se] study equation); HF � heart failure; HTN � hypertension; IABP �

ent elevation myocardial infarction; ST � stent thrombosis; STEMI �

atient

(

1

1

1

D � co

al Disea

ST-segm

placed as treatment for the ST event if the event was early or

smTa



135

late. Among subjects with very late ST, the rate of new DESplacement was similar irrespective of the type of stent associ-ated with the ST (47% vs. 48% for previously implanted DESand BMS, respectively).

Approximately 10% of patients required mechanical supportduring the procedure. Heparin was the most frequent anti-thrombotic agent, followed by bivalirudin. Glycoprotein IIb/IIIa inhibitors and prasugrel were used with increased fre-quency in early ST compared with late or very late ST. Successof lesion treatment was slightly lower in early ST (91.6%) than

Table 2. Intraprocedural Characteristics of Patients

Angiographic characteristics

Pre-Procedure TIMI flow grade

0

1–2

3

Culprit lesion location

LAD

RCA

Circumflex

Bypass graft

Left main

Lesion length, mm

Bifurcation lesion

Procedural characteristics

Thrombectomy performed

Additional stent placed

DES

BMS

No stent placed

IABP or other mechanical support placed during procedu

Dissection

Coronary perforation

Post-procedure TIMI flow grade

3

0–2

Intraprocedural medications

Anticoagulants

Unfractionated heparin

Bivalirudin

LMWH

Fondaparinux

Aspirin

Glycoprotein IIb/IIIa inhibitor

Thienopyridines

Clopidogrel

Prasugrel

Ticlopidine

Values are n (%).

LAD � left anterior descending artery; LMWH � low molecular-wei

Infarction; other abbreviations as in Table 1.

late (93.9%) or very late ST (94.6%, p � 0.026). l

Patients with early ST had a greater likelihood of stroke(1%) and major bleeding events (5.4%) than patients with lateor very late ST did (p � 0.001 for stroke; p � 0.003 forbleeding) (Table 3). Unadjusted in-hospital mortality washigher for early ST (7.9%) than for late (3.8%) or very late ST(3.6%, p � 0.001). Mortality was higher for early ST across thepectrum of acute coronary syndromes, with the highestortality in patients with STEMI due to early ST (Fig. 3).hese mortality trends were similar for patients with ST fromDES or BMS (Fig. 4). The lower mortality for late and very

T

Early ST(n � 1,391)

Late ST(n � 1,370)

Very Late ST(n � 4,318) p Value

�0.001

1,141 (82.0) 848 (61.9) 2,948 (68.3)

150 (10.8) 330 (24.1) 872 (20.2)

96 (6.9) 191 (13.9) 489 (11.3)

�0.001

666 (48.0) 541 (39.6) 1,728 (40.2)

371 (26.8) 433 (31.7) 1,635 (38.0)

282 (20.3) 217 (15.9) 610 (14.2)

51 (3.7) 160 (11.7) 310 (7.2)

17 (1.2) 17 (1.2) 21 (0.5)

18.4 (11.2) 20.3 (12.8) 21.3 (11.9) �0.001

240 (17.3) 204 (14.9) 603 (14.0) 0.010

447 (32.1) 401 (29.3) 1,475 (34.2) 0.003

�0.001

374 (26.9) 482 (35.2) 2,031 (47.0)

338 (24.3) 284 (20.7) 988 (22.9)

679 (48.8) 604 (44.1) 1,299 (30.1)

186 (13.4) 119 (8.7) 403 (9.3) �0.001

28 (2.0) 22 (1.6) 39 (0.9) 0.002

6 (0.4) 8 (0.6) 11 (0.3) 0.172

0.003

1,274 (91.6) 1,286 (93.9) 4,085 (94.6)

96 (6.9) 71 (5.2) 198 (4.6)

989 (71.1) 943 (69.0) 3,180 (73.9) 0.001

416 (29.9) 493 (36.0) 1,354 (31.4) 0.001

169 (12.2) 154 (11.3) 499 (11.6) 0.736

15 (1.1) 19 (1.4) 57 (1.3) 0.737

1,224 (90.1) 1,219 (90.0) 3,901 (90.9) 0.536

1,023 (73.9) 876 (64.0) 2,949 (68.7) �0.001

1,102 (80.6) 1,099 (81.4) 3,355 (79.2) 0.1609

175 (12.6) 127 (9.3) 329 (7.6) �0.001

11 (0.8) 14 (1.0) 24 (0.6) 0.174

arin; RCA � right coronary artery; TIMI � Thrombolysis In Myocardial

With S

re

ght hep

ate ST persisted after multivariable adjustment, compared


F E B R U A R Y 2 0 1 2 : 1 3 1 – 4 0

Armstrong et al.

Outcomes After ST

136

Figure 2. Treatment Patterns of Early, Late, and Very Late ST

(A) Early stent thrombosis (ST) was more frequently treated with balloon angioplasty alone (POBA), whereas late and very late ST were more frequently treatedwith placement of a new drug-eluting stent (DES) (p � 0.001). (B) Among subjects with ST of a previously implanted DES, a new DES was more frequentlyplaced than a bare-metal stent (BMS), regardless of ST timing (p � 0.001). (C) Among subjects with ST of a previously implanted BMS, a new BMS was more fre-
quently placed for treatment of early ST, but DES placement was more likely among subjects with very late ST (p � 0.001).



137

with early ST, with an adjusted odds ratio (OR) for in-hospitalmortality from late ST of 0.53 (95% confidence interval (CI):0.35 to 0.79) and 0.58 (95% CI: 0.43 to 0.79) for very late ST(Table 4). These trends were unchanged and remained statis-tically significant if the analysis was limited only to patientspresenting with STEMI or non-STEMI.

Discussion

This analysis represents the largest reported series on thepresentation and outcomes of ST in a contemporary PCIsetting. Because of the likelihood of disparate underlyingetiologies of early, late, and very late ST, we hypothesizedthat significant differences would exist among these impor-tant clinical subgroups. The large sample size of the Cath-PCI registry enabled us to make comparisons among groupsthat have not been possible with smaller reported case series.Three major findings emerged from this study. First, therewere significant differences in the baseline characteristicsand presentation of patients undergoing PCI for treatmentof early, late, and very late ST, supporting the notion ofdiffering etiologies. Second, most ST cases presented more

Table 3. In-Hospital Outcomes of Patients With ST

Early ST Late ST Very Late ST p Value

Death 110 (7.9) 52 (3.8) 155 (3.6) �0.001

Stroke 14 (1.0) 3 (0.2) 11 (0.3) �0.001

Bleeding event within 72 h 74 (5.4) 43 (3.2) 154 (3.6) 0.003

Values are n (%).

ST � stent thrombosis.

Figure 3. In-Hospital Mortality Based on Presentation With Unstable Angin

The p values are: 0.06 for unstable angina; 0.038 for non-STEMI; and �0.001 for STE

than 1 year after stent implantation. Third, ST presentingwithin 1 month of stent implantation was associated with asignificantly higher risk of in-hospital mortality than SToccurring more than 1 month after initial treatment.Presentation of early, late, and very late ST. We observedsignificant baseline differences among patients with ST,depending on the time since stent implantation. Thesedifferences in baseline clinical characteristics, as a functionof ST timing, lend support to the hypotheses that differingunderlying mechanisms cause early, late, and very late ST.In the recent RESTART (Registry of Stent Thrombosis ofSirolimus-Eluting Stents), late ST, compared with very lateST, was associated with presence of heart failure, hemodi-alysis, diabetes, and low body mass index (2). Our confir-mation of a higher prevalence of renal failure amongsubjects undergoing PCI for late ST suggests that renaldysfunction may mediate late ST, possibly due to increasedrestenosis with superimposed thrombosis (20,21). Otherstudies have also found a lower prevalence of diabetesamong patients with very late ST compared with early ST,supporting the hypothesis that patients with diabetes mayhave an altered acute inflammatory response to stent place-ment (2,22). A recent study also found an associationbetween black race and development of ST irrespective ofmedication compliance or socioeconomic status (23).

Because we had access only to in-hospital data at the timeof ST, we were unable to investigate further any possibleassociation between race and development of ST.

The overall rate of presentation with STEMI in ourcohort was similar to previous reports, in which STEMI hasrepresented 70% to 80% of cases of ST. The higher

n-STEMI, or STEMI
a, No
MI. STEMI � ST-segment elevation myocardial infarction.

varfiiipirsaCv(

.7 for


F E B R U A R Y 2 0 1 2 : 1 3 1 – 4 0

Armstrong et al.

Outcomes After ST

138

prevalence of complete thrombotic occlusion in the culpritvessel among patients with early ST may reflect increasedpredisposition to thrombosis or increased prevalence ofmechanical complications related to stent implantation. Ourfinding that subjects with very late ST were more likely tohave TIMI flow grade 2 or 3 at the time of angiography isconsistent with recent reports (2).

Somewhat surprisingly, we found that only one-half ofpatients were treated with aspiration thrombectomy at thetime of ST, and that almost one-half of patients had a newstent implanted at the time of ST. Although data onoutcomes after ST are limited, the high thrombus burden ofpatients with ST would suggest that aspiration thrombec-tomy might improve short- and possibly long-term out-comes. Additionally, implantation of a new stent at the timeof ST has been associated with adverse long-term outcomes.The recent large registry of Japanese patients with ST ofsirolimus-eluting stents had slightly higher rates of aspira-tion thrombectomy, but patients in that registry were also

Figure 4. In-Hospital Mortality From ST Due to a DES or BMS

There was no difference in mortality for ST due to a BMS versus a DES (p � 0

Table 4. Risk-Adjusted In-Hospital Mortality for Early

Category Unadjusted OR

Early ST 7.9% Reference

Late ST 3.8% 0.530

Very late ST 3.6% 0.580

Variables in the model—age, sex, race, STEMI, cardiogenic shock, prev

highest risk segment category, PCI status, diabetes, and bifurcation le

CHF � congestive heart failure; CI � confidence interval; CVD � ca

PCI � percutaneous coronary intervention; PVD � peripheral vascular disease

treated with new stent implantation 36% of the time.Further studies of ST with long-term follow-up are neededto identify management strategies at the time of ST thatimprove long-term outcomes.High prevalence of very late ST. In this study, we found thatery late ST had a higher prevalence than early or late STmong subjects undergoing PCI, with most events occur-ing more than 2 years after initial stent implantation. Ourndings do not suggest that very late ST has a higher

ncidence than early or late ST does. The CathPCI registrys a cross-sectional design, such that the denominator ofatients at risk of ST is uncertain. Previous studies, includ-ng a Dutch registry of BMS and DES and a Japaneseegistry limited to sirolimus-eluting DES, were prospectivetudies that were able to identify the incidence of ST among

stable population with a defined length of follow-up.onsistent with this design, those studies suggested that

ery late ST represents only 13% to 30% of all cases of ST2,14). In comparison, this study identified all-comers with

early ST, 0.8 for late ST, and 0.9 for very late ST). Abbreviations as in Figure 2.

, and Very Late ST

Adjusted OR (95% CI)

ower (95% CI) Upper (95% CI) p Value

0.356 0.790 0.002

0.429 0.785 �0.001

F, CVD, PVD, chronic lung disease, GFR, dialysis, NYHA functional class,

cular disease; NYHA � New York Heart Association; OR � odds ratio;

, Late

L

ious CH

sion.

rdiovas

; other abbreviations as in Table 1.

at9dtNiffrbeowsmpisefmtsum



139

ST, including those with DES or BMS. These differencesin methodology may account for the high prevalence of verylate ST: early and late ST by definition consist of limitedtime frames, whereas very late ST comprises the continuedexposure to a stent from 1 year to many years after initialimplantation. In the CathPCI registry, therefore, the totalnumber of patient-years with exposure to risk of very lateST is greater than the patient-year exposure to early or lateST, so it is not unexpected that very late events wouldpredominate with this study design. Our findings of a highpopulation prevalence of very late ST are also similar to thatof a recent multicenter observational study of angiographi-cally confirmed ST (24).

Nonetheless, our results emphasize the high populationprevalence and public health importance of preventing verylate ST. Recent studies have suggested that very late SToccurs with a constant incidence of 0.5% per year for BMSand up to 1.9% per year for DES (22,25). Our findings alsoraise concerns that very late ST remains a continued riskyears after PCI; that a significant number of ST occur �2years after stent implantation; and that further studies arewarranted to address the utility of continuing antiplatelettherapy beyond 1 year (26).High mortality of early ST. The in-hospital mortality riskssociated with early ST is high and significantly greaterhan for patients with late or very late ST. The observed.15% in-hospital mortality from STEMI from early ST isramatically higher than the reported 5.5% mortality fromhe overall population of contemporaneous patients in theCDR registry presenting with STEMI (27). This finding

s consistent with previous studies, suggesting that STEMIrom ST is associated with higher mortality than STEMIrom a de novo lesion (3). This increased mortality mayeflect a pre-disposition to thrombosis, higher thrombusurden, or complications of the initial PCI (28). Notably,arly ST but not late or very late ST is associated with highn-treatment platelet reactivity, suggesting that patientsith early ST may have a pre-disposition to thrombosis and

ubsequent adverse outcomes (29,30). The increased use ofechanical support, glycoprotein IIb/IIIa inhibitors, and

rasugrel among patients with early ST suggest that more-ntensive therapies were administered to this critically illubgroup. Despite this, the increased risk of mortality in thearly ST subgroup remained after multivariable adjustmentor baseline factors using a previously validated model forortality after myocardial infarction. Given the lack of

herapies specific to ST, and the high risk of mortality,troke, and bleeding that are associated with it, this studynderscores the importance of continued efforts to identifyodifiable predictors of this dreaded complication.

Study limitations. The results of this study should be inter-preted in the context of several potential limitations. First,because the CathPCI registry does not link patients across
cases, we could not ascertain the details of the initial PCI
associated with the initial stent implantation, such as stentlength, the number of stents implanted, vessel sizes, com-plications of the initial procedure, or use of intravascularultrasound. However, these procedural risk factors for SThave been well described in previous studies (2,4). Second,the data collected in the CathPCI registry does not includecompliance to medications. Therefore, we cannot ascertainthe prevalence of medication noncompliance or timing ofantiplatelet discontinuation before ST. Whereas antiplateletmedication adherence is an important risk factor for ST, theemphasis of our study was on management and outcomes ofST, rather than on predictors of developing ST. Third, theidentification of ST was based on operator identificationand report without an angiographic core laboratory. Over-or underreporting of ST may have occurred; however, wedid employ multiple strategies to ensure that the cases hada clinical presentation consistent with ST. Fourth, this studywas only able to identify cases of ST for which patientsunderwent PCI and, therefore, does not include patientswith probable or possible ST as defined by the AcademicResearch Consortium criteria, or ST patients managedwithout PCI. This limitation, along with a “survivor bias”(e.g., ST resulting in pre-hospital death was not captured)may have resulted in underestimation of the true prevalenceof ST in the entire population of stented patients. Alterna-tively, there could be a selection bias due to differentproportions of out-of-hospital deaths related to the timingof ST. Finally, the outcomes of this study are limited toin-hospital events. As with other observational studies, theassociation of early ST with increased mortality does notnecessarily confer causality.

Conclusions

Early, late, and very late ST have significant differences inpresentation that may reflect different underlying causes ofST as a function of timing after the initial stent implanta-tion. Most patients with early ST were treated with balloonangioplasty, whereas patients with very late ST were morefrequently treated with DES. Early ST, compared with lateor very late ST, is associated with higher in-hospitalmortality after multivariable adjustment for baseline clinicalcharacteristics.

Reprint requests and correspondence: Dr. Ehrin J. Armstrong,Division of Cardiovascular Medicine, University of California, Davis,4860 Y Street, Suite 2820, Sacramento, California 95817. E-mail:[email protected].

REFERENCES

1. van Werkum JW, Heestermans AA, de Korte FI, et al. Long-termclinical outcome after a first angiographically confirmed coronary stent
thrombosis: an analysis of 431 cases. Circulation 2009;119:828–34.
mailto:[email protected]


F E B R U A R Y 2 0 1 2 : 1 3 1 – 4 0

Armstrong et al.

Outcomes After ST

140

2. Kimura T, Morimoto T, Kozuma K, et al., for the RESTARTInvestigators. Comparisons of baseline demographics, clinical presenta-tion, and long-term outcome among patients with early, late, and very latestent thrombosis of sirolimus-eluting stents: observations from the Regis-try of Stent Thrombosis for Review and reevaluation (RESTART).Circulation 2010;122:52–61.

3. Chechi T, Vecchio S, Vittori G, et al. ST-segment elevation myocar-dial infarction due to early and late stent thrombosis a new group ofhigh-risk patients. J Am Coll Cardiol 2008;51:2396–402.

4. de la Torre-Hernandez JM, Alfonso F, Hernandez F, et al., for theESTROFA Study Group. Drug-eluting stent thrombosis: results fromthe multicenter Spanish registry ESTROFA (estudio ESpanol sobreTROmbosis de stents FArmacoactivos). J Am Coll Cardiol 2008;51:986–90.

5. Holmes DR Jr., Kereiakes DJ, Garg S, et al. Stent thrombosis. J AmColl Cardiol 2010;56:1357–65.

6. Fujii K, Carlier SG, Mintz GS, et al. Stent underexpansion andresidual reference segment stenosis are related to stent thrombosis aftersirolimus-eluting stent implantation: an intravascular ultrasound study.J Am Coll Cardiol 2005;45:995–8.

7. Joner M, Finn AV, Farb A, et al. Pathology of drug-eluting stents inhumans: delayed healing and late thrombotic risk. J Am Coll Cardiol2006;48:193–202.

8. Cook S, Wenaweser P, Togni M, et al. Incomplete stent appositionand very late stent thrombosis after drug-eluting stent implantation.Circulation 2007;115:2426–34.

9. Cook S, Ladich E, Nakazawa G, et al. Correlation of intravascularultrasound findings with histopathological analysis of thrombus aspi-rates in patients with very late drug-eluting stent thrombosis. Circula-tion 2009;120:391–9.

10. Takano M, Yamamoto M, Inami S, et al. Appearance of lipid-ladenintima and neovascularization after implantation of bare-metal stentsextended late-phase observation by intracoronary optical coherencetomography. J Am Coll Cardiol 2009;55:26–32.

11. Airoldi F, Colombo A, Morici N, et al. Incidence and predictors ofdrug-eluting stent thrombosis during and after discontinuation ofthienopyridine treatment. Circulation 2007;116:745–54.

12. Mega JL, Simon T, Collet JP, et al. Reduced-function CYP2C19genotype and risk of adverse clinical outcomes among patients treatedwith clopidogrel predominantly for PCI: a meta-analysis. JAMA2010;304:1821–30.

13. Mauri L, Hsieh WH, Massaro JM, Ho KK, D’Agostino R, Cutlip DE.Stent thrombosis in randomized clinical trials of drug-eluting stents.N Engl J Med 2007;356:1020–9.

14. van Werkum JW, Heestermans AA, Zomer AC, et al. Predictors ofcoronary stent thrombosis: the Dutch Stent thrombosis registry. J AmColl Cardiol 2009;53:1399–409.

15. Iakovou I, Schmidt T, Bonizzoni E, et al. Incidence, predictors, andoutcome of thrombosis after successful implantation of drug-elutingstents. JAMA 2005;293:2126–30.

16. Brindis RG, Fitzgerald S, Anderson HV, Shaw RE, Weintraub WS,Williams JF. The American College of Cardiology–National Cardio-vascular Data Registry (ACC-NCDR): building a national clinical datarepository. J Am Coll Cardiol 2001;37:2240–5.

17. National Cardiovascular Data Registry. Data Elements and Defini-
tions, Technology Downloads, and Risk Adjustment [web page].
Available at: http://www.ncdr.com/WebNCDR/ELEMENTS.ASPX.Accessed January 10, 2011.

18. Cutlip DE, Windecker S, Mehran R, et al., for the Academic ResearchConsortium. Clinical end points in coronary stent trials: a case forstandardized definitions. Circulation 2007;115:2344–51.

19. Peterson ED, Dai D, DeLong ER, et al., for the NCDR RegistryParticipants. Contemporary mortality risk prediction for percutaneouscoronary intervention: results from 588,398 procedures in the nationalcardiovascular data registry. J Am Coll Cardiol 2010;55:1923–32.

20. Ikari Y, Kotani J, Kozuma K, et al., for the J-PMS Study Group.Assessment of sirolimus-eluting coronary stent implantation withaspirin plus low dose ticlopidine administration: one year results fromCYPHER Stent Japan Post-Marketing Surveillance registry (J-PMS).Circ J 2009;73:1038–44.

21. Zhu ZB, Zhang RY, Zhang Q, et al. Moderate-severe renal insuffi-ciency is a risk factor for sirolimus-eluting stent thrombosis. The RIFTstudy. Cardiology 2009;112:191–9.

22. Daemen J, Wenaweser P, Tsuchida K, et al. Early and late coronarystent thrombosis of sirolimus-eluting and paclitaxel-eluting stents inroutine clinical practice: data from a large two-institutional cohortstudy. Lancet 2007;369:667–78.

23. Collins SD, Torguson R, Gaglia MA Jr., et al. Does black ethnicityinfluence the development of stent thrombosis in the drug-eluting stentera? Circulation 2010;122:1085–90.

24. Yeo KK, Armstrong EJ, Bennett WE, et al. Contemporary clinicalcharacteristics, treatment, and outcomes of angiographically confirmedcoronary stent thrombosis: results from a multicenter California regis-try. Catheter Cardiovasc Interv 2011 May 11 [E-pub ahead of print],doi:10.1002/ccd.23011.

25. Brodie B, Pokharel Y, Fleishman N, et al. Very late stent thrombosisafter primary percutaneous coronary intervention with bare-metal anddrug-eluting stents for ST-segment elevation myocardial infarction: a15-year single-center experience. J Am Coll Cardiol Intv 2011;4:30–8.

26. Mauri L, Kereiakes DJ, Normand SL, et al. Rationale and design of thedual antiplatelet therapy study, a prospective, multicenter, randomized,double-blind trial to assess the effectiveness and safety of 12 versus 30months of dual antiplatelet therapy in subjects undergoing percutane-ous coronary intervention with either drug-eluting stent or bare metalstent placement for the treatment of coronary artery lesions. AmHeart J 2010;160:1035–41.

27. Roe MT, Messenger JC, Weintraub WS, et al. Treatments, trends, andoutcomes of acute myocardial infarction and percutaneous coronaryintervention. J Am Coll Cardiol 2010;56:254–63.

28. Sianos G, Papafaklis MI, Daemen J, et al. Angiographic stent throm-bosis after routine use of drug-eluting stents in ST-segment elevationmyocardial infarction: the importance of thrombus burden. J Am CollCardiol 2007;50:573–83.

29. Bouman HJ, van Werkum JW, Breet NJ, ten Cate H, Hackeng CM,ten Berg JM. A case-control study on platelet reactivity in patients withcoronary stent thrombosis. J Thromb Haemost 2011;9:909–16.

30. Geisler T, Zürn C, Simonenko R, et al. Early but not late stentthrombosis is influenced by residual platelet aggregation in patientsundergoing coronary interventions. Eur Heart J 2010;31:59–66.

Key Words: acute coronary syndrome(s) � cardiovascular
outcomes � stent thrombosis.
http://www.ncdr.com/WebNCDR/ELEMENTS.ASPX

clinical presentation, management, and outcomes of ... · figure 3. in-hospital mortality based on...

Documents