Professional Documents
Culture Documents
22, 2018
Peter T Hu, MD,a,b W. Schuyler Jones, MD,b,c Thomas J. Glorioso, MS,d,e Anna E. Barón, PHD,d,e
Gary K. Grunwald, PHD,d,e Stephen W. Waldo, MD,d,f Thomas M. Maddox, MD, MSC,g Mladen Vidovich, MD,h,i
Subhash Banerjee, MD,j,k Sunil V. Rao, MDb,c,l
ABSTRACT
OBJECTIVES The aim of this study was to determine predictors and outcomes associated with staged percutaneous
coronary intervention (PCI) versus one-time multivessel revascularization (OTMVR) in patients with multivessel coronary
artery disease.
BACKGROUND Prior observational studies have not evaluated predictors and outcomes of staged PCI versus OTMVR in
a heterogenous population of patients with multivessel coronary artery disease who undergo multivessel
revascularization.
METHODS Data from the Veterans Affairs (VA) CART (Clinical Assessment, Reporting, and Tracking) Program were used
to evaluate patients who underwent PCI of >2 vessels between October 1, 2007, and September 3, 2014. Associations
between individual factors and the decision to perform staged PCI were assessed. Additionally, the impact of measured
patient and procedural factors, site factors, and unmeasured site factors on the decision to perform staged PCI was
compared. Cox proportional hazards models were used to determine the association between staged PCI and mortality.
RESULTS A total of 7,599 patients at 61 sites were included. The decision to perform staged PCI was driven by pro-
cedural characteristics and unmeasured site factors. Staged PCI was associated with lower risk-adjusted mortality
compared with OTMVR (adjusted hazard ratio [HR]: 0.78; 95% confidence interval [CI]: 0.72 to 0.84; p < 0.01). This
mortality benefit was observed among the ST-segment elevation myocardial infarction (HR: 0.31; 95% CI: 0.21 to 0.47;
p < 0.01), non–ST-segment elevation myocardial infarction (HR: 0.74; 95% CI: 0.64 to 0.87; p < 0.01), unstable angina
(HR: 0.75; 95% CI: 0.64 to 0.89; p < 0.01) and stable angina (HR: 0.88; 95% CI: 0.77 to 1.00; p ¼ 0.05) groups.
CONCLUSIONS The decision to pursue staged PCI was driven by procedural characteristics and unmeasured site
variation and was associated with lower mortality compared with OTMVR. After adjustment, there was an association
between staged PCI and reduced mortality. Given the observational nature of these findings, a randomized trial
comparing the 2 is needed to guide practice. (J Am Coll Cardiol Intv 2018;11:2265–73) Published by Elsevier on behalf of
the American College of Cardiology Foundation.
From the aDepartment of Cardiology, Cleveland Clinic, Cleveland, Ohio; bDepartment of Medicine, Duke University Medical
Center, Durham, North Carolina; cDuke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina;
d
Denver Veterans Affairs Medical Center, Denver, Colorado; eColorado School of Public Health, University of Colorado Anschutz
Medical Campus, Aurora, Colorado; fSection of Cardiology, University of Colorado School of Medicine, Aurora, Colorado; gDivision
of Cardiology, Washington University, St. Louis, Missouri; hUniversity of Illinois College of Medicine, Chicago, Illinois; iJesse
Brown Veterans Affairs Medical Center, Chicago, Illinois; jUniversity of Texas Southwestern Medical Center, Dallas, Texas;
k
Veterans Affairs North Texas Health Care System, Dallas, Texas; and the lDurham Veterans Affairs Medical Center, Durham,
North Carolina. Dr. Jones has received research funding from the Agency for Healthcare Research and Quality, the American Heart
Association, AstraZeneca, Bristol-Myers Squibb, Daiichi-Sankyo, and the Patient-Centered Outcomes Research Institute. Dr. Waldo
M
ABBREVIATIONS ultivessel coronary artery disease associated outcomes of MVPCI across the spectrum of
AND ACRONYMS (MVCAD) is present in approxi- clinical presentation remain unclear, and second,
mately two-thirds of patients observational studies are subject to multiple biases,
ACS = acute coronary
syndrome(s)
who require revascularization (1). Two including unmeasured confounding. To date, there
options exist when performing multi- are insufficient observational data and limited ran-
CI = confidence interval
vessel percutaneous coronary intervention domized controlled trials to provide recommenda-
HR = hazard ratio
(MVPCI): staged percutaneous coronary tions on the optimal timing of MVPCI (8).
MVCAD = multivessel coronary
artery disease
intervention (PCI) or one-time multivessel Using standardized electronic health system data
revascularization (OTMVR). In staged PCI, from the national Veterans Affairs (VA) CART-CL
MVPCI = multivessel
percutaneous coronary patients often receive revascularization of a (Clinical Assessment, Reporting, and Tracking Sys-
intervention culprit (thrombotic) artery or the most severe tem for Catheterization Laboratories) Program, we
NSTEMI = non–ST-segment stenosis, followed by a planned revasculari- sought to determine: 1) factors associated with the
elevation myocardial infarction
zation of a nonculprit artery or arteries or decision to perform staged PCI or OTMVR among
OR = odds ratio less severe or complex stenosis at a later patients who underwent MVPCI, including unmea-
OTMVR = one-time multivessel time. This is in contrast to OTMVR, in which sured site variation; and 2) the association between
revascularization
more than 1 diseased lesion is revascularized staged PCI or OTMVR and clinical outcomes (death,
PCI = percutaneous coronary
at the time of the index procedure. Patients rehospitalization, transfusion) (9).
intervention
with MVCAD who undergo MVPCI constitute
REM = reference effect
a heterogenous patient population. The deci-
measure METHODS
sion to perform staged PCI or OTMVR may be
STEMI = ST-segment elevation
myocardial infarction related to many factors, including patient-
STUDY SAMPLE. The VA CART Program is a national
VA = Veterans Affairs specific factors (e.g., chronic kidney disease,
VA clinical quality initiative that was launched in
indication for PCI such as ST-segment eleva-
2005 and has been described previously (10). A key
tion myocardial infarction [STEMI], non–ST-segment
feature of the CART Program is standardized data
elevation myocardial infarction [NSTEMI], unstable
capture and reporting across all VA catheterization
angina, or stable angina), physician- or hospital-
laboratories through a clinical software application.
specific factors, and procedural characteristics (e.g.,
The software is embedded in the VA electronic health
anatomic complexity, disease burden, presence of
record and allows providers to enter patient and
chronic total occlusion, contrast load, radiation expo-
procedural information (pre-procedural assessment,
sure). Each of these factors plays an important role in
cardiac catheterization, and PCI) as part of routine
the decision-making process in MVPCI. For example,
clinical work flow. The application is integrated with
patients with stable angina and chronic total occlu-
the VA’s patient electronic health record, and data
sions who undergo MVPCI are more likely to undergo
elements automatically populate a clinical note.
staged PCI (2). Furthermore, among interventional
These data are combined with longitudinal VA
cardiologists, renal function is the strongest factor
administrative data in a clinical data repository to
in the decision to pursue staged PCI (3).
support the quality assessment, quality improve-
SEE PAGE 2274 ment, and clinical research missions of the CART
Program. Data elements conform to the definitions
In patients with STEMI, several observational and standards of the American College of Cardiol-
studies have suggested that multivessel staged PCI ogy’s National Cardiovascular Data Registry (9,11).
may be associated with lower mortality compared Cardiac catheterization reports generated by the VA
with OTMVR (4–7). Observational data comparing CART Program have been shown to demonstrate
staged versus OTMVR in those with NSTEMI are excellent data validity and completeness (10). This
lacking. Two issues hinder the use of these data to study was approved by the Institutional Review
guide practice. First, the optimal timing and Board at the Denver VA Medical Center and the
receives unrelated research funding to the Denver Research Institute from Abiomed, Cardiovascular Systems Inc., and Merck
Pharmaceuticals. Dr. Maddox was the former director of the VA CART Program. Dr. Vidovich has received research funding from
Boston Scientific, Biosensors International, and Sanofi; and is a consultant for Boston Scientific and Abbott and serves on the
advisory board for Merit. Dr. Banerjee has received honoraria from Medtronic, Gore, and Cardiovascular Systems Inc., and
research funding from Boston Scientific and Merck. All other authors have reported that they have no relationships relevant to the
contents of this paper to disclose.
Manuscript received June 5, 2018; revised manuscript received July 24, 2018, accepted July 31, 2018.
JACC: CARDIOVASCULAR INTERVENTIONS VOL. 11, NO. 22, 2018 Hu et al. 2267
NOVEMBER 26, 2018:2265–73 Staged Versus One-Time Multivessel Revascularization
Durham VA Medical Center, with a waiver of the in prior inpatient, outpatient, and CART records were
requirement to obtain informed consent. used to determine medical comorbidities. Two
STUDY POPULATION. Patients included in the anal- outpatient records or 1 inpatient record with the In-
ysis were those who had MVCAD and underwent PCI to ternational Classification of Diseases-9th Revision
2 or more coronary artery distributions at a VA cathe- code of interest were needed to confirm a comorbid
terization laboratory between October 1, 2007, and condition.
September 30, 2014, and had their records collected in
PROCEDURAL CHARACTERISTICS. Information from
CART. Patients were excluded from analyses if they
either the index procedure or a combination of the
had left main coronary artery treatment, presented
index and staged procedure was used for the proce-
with cardiogenic shock, underwent PCI in the previous
dural characteristic variables. Number of stents used
60 days either inside or outside of the VA, underwent
and number of diseased vessels treated were summed
more than 1 staged procedure, or were at a VA cathe-
over both procedures if a patient underwent multiple
terization laboratory that performed no staged PCI or
PCIs. PCI status, indication, and access site (radial vs.
had performed 10 or fewer PCIs overall. If a patient
femoral) were used from the index procedure. If a
underwent more than 1 MVPCI over this time frame,
patient had multiple access sites listed for the index
only the first PCI was used. Patients with missing data
procedure, the second access site was used.
were not included in the analyses (Figure 1).
DEFINITIONS AND OUTCOMES. O T M V R . OTMVR SITE CHARACTERISTICS. Site regional status was
was defined as PCI of >2 diseased coronary artery based on U.S. census regions (West, Midwest, South,
distributions at the time of an index procedure. A and Northeast). Certain patient-level characteristics
diseased left anterior descending, left circumflex, or that may better describe the catheterization labora-
right coronary artery was defined by a segment with tory site, such as racial breakdown and urban or rural
>70% stenosis in a main branch or side branch of an status, were summarized at the site level by taking
epicardial artery. the overall proportion of patients in the cohort with
S t a g e d P C I . To be included in the staged PCI group, the characteristic of interest at the site.
a subsequent PCI of a diseased coronary artery had to STATISTICAL ANALYSIS. All data were reviewed for
be performed within 60 days of an index procedure. missing values and quality. Some patients were
The second procedure could not have treated a missing values for body mass index (1.2%) and urban
segment that had been treated during the index pro- or rural status (1.1%), and a single imputation was
cedure. The subsequent procedure could not have a performed simulating values on the basis of associa-
PCI status of emergent, urgent, or salvage. It also tions with observed data; 449 patients with other
could not have an indication of STEMI or NSTEMI. If missing values were excluded because of the diffi-
comments in CART-CL indicated that a patient should culty in correctly imputing values.
be scheduled for a staged procedure, the patient was A mixed logistic regression model with a random
included as part of the staged PCI group using an intercept for site was estimated using the glmmML
intention-to-treat approach. package in R version 3.2.5 (R Foundation for Statistical
O u t c o m e s . The primary efficacy outcome was all- Computing, Vienna, Austria). Associations of individ-
cause death through latest follow-up. Secondary ual measured patient and site factors with staged PCI
outcomes included all-cause rehospitalization and versus OTMVR, adjusted for other patient, procedure,
rehospitalization for myocardial infarction through and site factors, were examined using this model. We
latest follow-up. These were time-to-event outcomes used reference effect measure (REM) methods to
using the latest discharge date from the index or explore the extent to which unmeasured site variation
staged procedure as a baseline date. Follow-up data contributed to the decision to use staged PCI relative
were available through September 30, 2015, and thus to sets of measured factors (12). A REM value is an odds
patients with no events were censored as of this date. ratio (OR) comparing a patient at a given risk percen-
For outcomes other than mortality, patients were tile with a median-risk patient. For variation between
censored at the date of death or the last date of follow- sites not accounted for by measured patient or site
up. The primary safety outcome was the rate of blood characteristics (unmeasured site variation), this is
transfusions within 72 h after both the index PCI and based on the estimated normal distribution of the site
subsequent PCI if a patient had a staged procedure. random effects. For variation due to measured de-
PATIENT COMORBIDITIES. With the exception of mographics, comorbidities, procedure, or site charac-
body mass index and baseline renal function, Inter- teristics, REM is based on the corresponding empirical
national Classification of Diseases-9th Revision codes distributions of patient risk scores. For each
2268 Hu et al. JACC: CARDIOVASCULAR INTERVENTIONS VOL. 11, NO. 22, 2018
subsequent staged procedures, patients with MVCAD Age (yrs) 65.3 9.0 65.2 8.9 65.3 9.1 0.49
who underwent staged PCI had higher rates of Sex (male) 0.987 (7,502) 0.989 (2,713) 0.986 (4,789) 0.34
Race (white) 0.848 (6,441) 0.859 (2,356) 0.841 (4,085) 0.04
treated 3-vessel disease (12.3% vs. 4.4%; p < 0.01),
Hispanic 0.044 (338) 0.044 (120) 0.045 (218) 0.86
STEMI presentation (8.1% vs. 3.0%; p < 0.01), and a
Urban 0.581 (4,415) 0.575 (1,578) 0.584 (2,837) 0.46
PCI status of emergent or salvage (8.0% vs. 3.4%;
Comorbidities
p < 0.01) compared with patients with OTMVR. The
BMI (kg/m2) 30.3 5.6 30.2 5.5 30.3 5.7 0.76
mean total number of stents placed during index PCI
CHF 0.231 (1,752) 0.212 (582) 0.241 (1,170) <0.01
and subsequent PCI for those who underwent staged
CKD 0.202 (1,538) 0.209 (572) 0.199 (966) 0.33
procedures (3.2) was greater than that for patients COPD 0.208 (1,583) 0.199 (546) 0.214 (1,037) 0.14
who underwent OTMVR (2.6) (p < 0.01). Baseline Diabetes 0.501 (3,809) 0.486 (1,334) 0.51 (2,475) 0.05
mean glomerular filtration rate did not differ GFR (ml/min/1.73 m2) 75.0 25.4 74.7 25.1 75.2 25.6 0.38
between those who underwent staged PCI versus Hyperlipidemia 0.898 (6,822) 0.883 (2,422) 0.906 (4,400) <0.01
OTMVR (mean glomerular filtration rate 74.7 vs. Hypertension 0.902 (6,857) 0.886 (2,431) 0.911 (4,426) <0.01
75.2 ml/min/1.73 m 2; p ¼ 0.38). Compared with PAD 0.194 (1,473) 0.184 (504) 0.2 (969) 0.10
OTMVR, patients undergoing staged PCI had higher Prior CABG 0.134 (1,020) 0.131 (360) 0.136 (660) 0.59
rates of atherectomy use (14.4% vs. 8.3%; p < 0.01), Prior MI 0.343 (2,609) 0.309 (847) 0.363 (1,762) <0.01
treated chronic total occlusions (16.4% vs. 8.8%; Prior PCI 0.357 (2,715) 0.331 (908) 0.372 (1,807) <0.01
p < 0.01) and treated calcified lesions (39.4% vs. Prior stroke/TIA 0.085 (649) 0.076 (209) 0.091 (440) 0.03
31.3%; p < 0.01) during either the index or staged Tobacco 0.635 (4,823) 0.601 (1,649) 0.654 (3,174) <0.01
Procedural characteristics
procedure.
Access (radial) 0.141 (1,072) 0.126 (345) 0.15 (727) <0.01
FACTORS ASSOCIATED WITH STAGED PCI. Stent type (BMS) 0.151 (1,144) 0.15 (412) 0.151 (732) 0.98
Table 2 shows adjusted associations between indi- Number of stents 2.8 1.3 3.2 1.5 2.6 1.1 <0.01
vidual factors and staged PCI versus OTMVR. With Treated vessels <0.01
respect to baseline patient characteristics, the only 2 0.927 (7,047) 0.877 (2,405) 0.956 (4,642)
comorbidity associated with staged PCI was chronic 3 0.073 (552) 0.123 (338) 0.044 (214)
kidney disease (adjusted OR: 1.31; 95% confidence Indication <0.01
interval [CI]: 1.14 to 1.52; p < 0.01). Procedural char- STEMI 0.048 (367) 0.081 (222) 0.03 (145)
acteristics associated with staged PCI included NSTEMI 0.213 (1,618) 0.213 (585) 0.213 (1,033)
3-vessel disease treated versus 2-vessel disease Stable angina/chest 0.418 (3,180) 0.409 (1,123) 0.424 (2,057)
pain/asymptomatic
(adjusted OR: 2.17; 95% CI: 1.77 to 2.68; p < 0.01) and Unstable angina/ACS 0.262 (1,991) 0.241 (661) 0.274 (1,330)
STEMI presentation (adjusted OR: 2.48; 95% CI: 1.74 Other/valve disease 0.058 (443) 0.055 (152) 0.06 (291)
to 3.52; p < 0.01). CTO treated 0.116 (878) 0.164 (450) 0.088 (428) <0.01
In Figure 2, we observe a wide 95% REM range for Calcified treated 0.342 (2,600) 0.394 (1,080) 0.313 (1,520) <0.01
unmeasured site variation (OR: 0.20 to 5.08), which Atherectomy 0.105 (797) 0.144 (395) 0.083 (402) <0.01
was comparable in width with the 95% REM range for Distal protection <0.001 (6) 0.001 (3) 0.001 (3) 0.78
procedural characteristics (OR: 0.44 to 6.17) and Embolic protection 0.001 (8) 0.001 (2) 0.001 (6) 0.78
ities (OR: 0.69 to 1.39), and measured site character- Elective 0.642 (4,878) 0.632 (1,733) 0.648 (3,145)
Urgent 0.307 (2,336) 0.288 (791) 0.318 (1,545)
istics (OR: 0.44 to 2.14). On the basis of unmeasured
Emergent/salvage 0.051 (385) 0.08 (219) 0.034 (166)
site variation, a patient at a low-“risk” site (2.5th
percentile) has odds of a staged procedure 0.20 times Values are mean SD or proportion (n).
lower compared with a patient a median-“risk” site, ACS ¼ acute coronary syndrome(s); BMI ¼ body mass index; BMS ¼ bare-metal stent(s); CABG ¼ coronary
artery bypass graft; CHF ¼ congestive heart failure; CKD ¼ chronic kidney disease; COPD ¼ chronic obstructive
while a patient at a high-“risk” site (97.5th percentile) pulmonary disease; CTO ¼ chronic total occlusion; GFR ¼ glomerular filtration rate; MI ¼ myocardial infarction;
has odds 5.08 times greater. The wide ranges show NSTEMI ¼ non–ST-segment elevation myocardial infarction; OTMVR ¼ one-time multivessel revascularization;
PAD ¼ peripheral artery disease; PCI ¼ percutaneous coronary intervention; STEMI ¼ ST-segment elevation
large differences in the likelihood of receiving staged myocardial infarction; TIA ¼ transient ischemic attack.
PCI due to unmeasured site variation and procedural
2270 Hu et al. JACC: CARDIOVASCULAR INTERVENTIONS VOL. 11, NO. 22, 2018
with a lower risk for death (HR: 0.78; 95% CI: 0.72 to
T A B L E 2 Predictors of Staged Percutaneous Coronary
Intervention Versus One-Time Multivessel Revascularization
0.84; p < 0.01). For the secondary outcomes of all-
cause rehospitalization and rehospitalization for
OR (95% CI) p Value
myocardial infarction after propensity adjustment,
Demographics
there were no statistically significant differences
Urban 0.91 (0.81–1.03) 0.13
(Table 3). Sensitivity analyses (Online Table 2) show
Hispanic 0.91 (0.69–1.21) 0.52
the effect of an unmeasured confounder with varying
Age (10 yrs) 0.98 (0.92–1.05) 0.63
prevalence and its association with death.
Race (white) 1.13 (0.96–1.33) 0.14
Sex (male) 1.25 (0.76–2.06) 0.38 ASSOCIATION BETWEEN STAGED PCI AND TRANS-
Comorbidities FUSION. The proportion of patients receiving blood
Prior stroke/TIA 0.87 (0.71–1.06) 0.15 transfusion within 72 h of both index PCI and subse-
CHF 0.87 (0.76–1.01) 0.06
quent PCI if a patient underwent a staged procedure
Prior MI 0.88 (0.77–1.00) 0.04
was not significantly different between the staged PCI
Hypertension 0.88 (0.73–1.06) 0.19
(1.4%) and OTMVR (1.3%) groups (p ¼ 0.96).
Diabetes 0.90 (0.80–1.01) 0.07
Tobacco use 0.91 (0.81–1.02) 0.10
SUBGROUP ANALYSIS. Similar HRs for mortality
to an increased risk for in-stent thrombosis with PCI PCI on the basis of the group of factors, suggesting that they contribute more to the
procedural decision. The width of the 95% reference effect measure (REM) range for
of nonculprit vessels, higher contrast load leading to
unmeasured site variation (odds ratio [OR]: 0.20 to 5.08) exceeded that of demographics
contrast-induced nephropathy, which has been (OR: 0.79 to 1.11), comorbidities (OR: 0.69 to 1.39), and site characteristics (OR: 0.44 to
associated with increased mortality, or greater risk for 2.14) and was comparable in width to procedural characteristics (OR: 0.44 to 6.17).
procedural complications with increased procedure
times and/or intervening on nonculprit vessels
without optimal evaluation (e.g., percentage diam- toward insignificance in the stable angina group from
eter stenosis, fractional flow reverse, or intravascular the STEMI, NSTEMI, and unstable angina groups is
ultrasound) (23). In those without STEMI, we likely attributed to a lesser degree of acuity and less
observed a positive yet weaker association with lower of an inflammatory state as well as procedural related
mortality in the NSTEMI (HR: 0.74; 95% CI: 0.64 to complications. This is the first analysis we are aware
0.87; p < 0.01), unstable angina (HR: 0.75; 95% CI: of to find this association across subgroups.
0.64 to 0.89; p < 0.01), and stable angina (HR: 0.88; To assess whether a second procedure in selected
95% CI: 0.77 to 1.00; p ¼ 0.05) groups. This trend patients would be associated with an increase in
2272 Hu et al. JACC: CARDIOVASCULAR INTERVENTIONS VOL. 11, NO. 22, 2018
PERSPECTIVES
WHAT IS KNOWN? In patients with STEMI, several OTMVR. We found that the decision to pursue staged PCI
observational studies have suggested that multivessel was driven by specific patient characteristics, procedural
staged PCI may be associated with lower mortality characteristics, and unmeasured site variation, and staged
compared with OTMVR. PCI was associated with a significant mortality benefit.
WHAT IS NEW? This study incorporates a heterogenous WHAT IS NEXT? A randomized controlled trial evalu-
patient population with STEMI, NSTEMI, unstable angina, ating the efficacy and optimal timing of staged PCI is
and stable angina who underwent either staged PCI or needed to guide clinical practice.
REFERENCES
1. Pocock SJ, Henderson RA, Rickards AF, et al. guideline for the management of ST-elevation 18. Engstrøm T, Kelbæk H, Helqvist S, et al. Com-
Meta-analysis of randomised trials comparing myocardial infarction: a report of the American plete revascularisation versus treatment of the
coronary angioplasty with bypass surgery. Lancet College of Cardiology/American Heart Association culprit lesion only in patients with ST-segment
1995;346:1184–9. Task Force on Clinical Practice Guidelines and the elevation myocardial infarction and multivessel dis-
2. Brilakis ES, Banerjee S, Karmpaliotis D, et al. Society for Cardiovascular Angiography and In- ease (DANAMI-3-PRIMULTI): an open-label, rando-
Procedural outcomes of chronic total occlusion terventions. Circulation 2016;133:1135–47. mised controlled trial. Lancet 2015;386:665–71.
percutaneous coronary intervention: a report from 9. Box TL, McDonell M, Helfrich CD, et al. Stra- 19. Smits PC, Boxma-de Klerk BM. Fractional flow
the NCDR (National Cardiovascular Data Registry). tegies from a nationwide health information reserve-guided multivessel angioplasty in myocar-
J Am Coll Cardiol Intv 2015;8:245–53. technology implementation: the VA CART story. dial infarction. N Engl J Med 2017;377:397–8.
3. Dangas GD, George JC, Weintraub W. Popma JJ J Gen Intern Med 2010;25 Suppl 1:72–6.
20. Thiele H, Akin I, Sandri M, et al. PCI strategies in
Timing of staged percutaneous coronary inter-
10. Byrd JB, Vigen R, Plomondon ME, et al. Data patients with acute myocardial infarction and
vention in multivessel coronary artery disease.
quality of an electronic health record tool to cardiogenic shock. N Engl J Med 2017;377:2419–32.
J Am Coll Cardiol Intv 2010;3:1096–9.
support VA cardiac catheterization laboratory
21. Li Z, Zhou Y, Xu Q, Chen X. Staged versus one-
4. Russo JJ, Wells GA, Chong AY, et al. Safety and quality improvement: the VA Clinical Assessment,
time complete revascularization with percuta-
efficacy of staged percutaneous coronary inter- Reporting, and Tracking System for Cath Labs
neous coronary intervention in STEMI patients
vention during index admission for ST-elevation (CART) Program. Am Heart J 2013;165:434–40.
with multivessel disease: a systematic review and
myocardial infarction with multivessel coronary
11. Brindis RG, Fitzgerald S, Anderson HV, Shaw RE, meta-analysis. PLoS ONE 2017;12:e0169406.
disease (insights from the University of Ottawa
Weintraub WS, Williams JF. The American College
Heart Institute STEMI Registry). Am J Cardiol 22. Vlaar PJ, Mahmoud KD, Holmes DR Jr., et al.
of Cardiology-National Cardiovascular Data Regis-
2015;116:1157–62. Culprit vessel only versus multivessel and staged
try (ACC-NCDR): building a national clinical data
percutaneous coronary intervention for multi-
5. Kornowski R, Mehran R, Dangas G, et al., for the repository. J Am Coll Cardiol 2001;37:2240–5.
vessel disease in patients presenting with
HORIZONS-AMI Trial Investigators. Prognostic
12. Glorioso TJ, Grunwald GK, Ho PM, Maddox TM. ST-segment elevation myocardial infarction: a
impact of staged versus “one-time” multivessel
Reference effect measures for quantifying, pairwise and network meta-analysis. J Am Coll
percutaneous intervention in acute myocardial
comparing and visualizing variation from random and Cardiol 2011;58:692–703.
infarction: analysis from the HORIZONS-AMI
fixed effects in non-normal multilevel models, with
(Harmonizing Outcomes With Revascularization 23. Giacoppo D, Madhavan MV, Baber U, et al.
applications to site variation in medical procedure use
and Stents in Acute Myocardial Infarction) trial. Impact of contrast-induced acute kidney injury
and outcomes. BMC Med Res Methodol 2018;18:74.
J Am Coll Cardiol 2011;58:704–11. after percutaneous coronary intervention on
13. Zhu Y, Ghosh D, Mitra N, Mukherjee B. A data- short- and long-term outcomes: pooled anal-
6. Iqbal MB, Nadra IJ, Ding L, et al. Culprit vessel
adaptive strategy for inverse weighted estimation ysis from the HORIZONS-AMI and ACUITY trials.
versus multivessel versus in-hospital staged
of causal effects. Health Serv Outcomes Res Circ Cardiovasc Interv 2015;8:e002475.
intervention for patients with ST-segment eleva-
Methodol 2014;14:69–91.
tion myocardial infarction and multivessel disease: 24. Kong JA, Chou ET, Minutello RM, Wong SC,
stratified analyses in high-risk patient groups and 14. Li F, Zaslavsky AM, Landrum MB. Propensity Hong MK. Safety of single versus multi-vessel
anatomic subsets of nonculprit disease. J Am Coll score weighting with multilevel data. Stat Med angioplasty for patients with acute myocardial
Cardiol Intv 2017;10:11–23. 2013;32:3373–87. infarction and multi-vessel coronary artery dis-
ease: report from the New York State Angioplasty
7. Manari A, Varani E, Guastaroba P, et al. Long- 15. Lin DY, Psaty BM, Kronmal RA. Assessing the
Registry. Coron Artery Dis 2006;17:71–5.
term outcome in patients with ST segment sensitivity of regression results to unmeasured
elevation myocardial infarction and multivessel confounders in observational studies. Biometrics
disease treated with culprit-only, immediate, or 1998;54:948–63. KEY WORDS multivessel coronary artery
staged multivessel percutaneous revascularization 16. Wald DS, Morris JK, Wald NJ, et al. Random- disease, multivessel coronary
strategies: insights from the REAL registry. Cath- ized trial of preventive angioplasty in myocardial revascularization, one-time multivessel
eter Cardiovasc Interv 2014;84:912–22. infarction. N Engl J Med 2013;369:1115–23. revascularization, percutaneous coronary
intervention, staged percutaneous coronary
8. Levine GN, Bates ER, Blankenship JC, et al. 2015 17. Gershlick AH, Khan JN, Kelly DJ, et al. Ran-
intervention
ACC/AHA/SCAI focused update on primary percu- domized trial of complete versus lesion-only
taneous coronary intervention for patients with ST- revascularization in patients undergoing primary
elevation myocardial infarction: an update of the percutaneous coronary intervention for STEMI and A PP END IX For supplemental tables and a
2011 ACCF/AHA/SCAI guideline for percutaneous multivessel disease: the CvLPRIT trial. J Am Coll figure, please see the online version of this
coronary intervention and the 2013 ACCF/AHA Cardiol 2015;65:963–72. paper.