Author + information
- Received June 5, 2018
- Revision received July 24, 2018
- Accepted July 31, 2018
- Published online November 19, 2018.
- Peter T Hu, MDa,b,
- W. Schuyler Jones, MDb,c,
- Thomas J. Glorioso, MSd,e,
- Anna E. Barón, PhDd,e,
- Gary K. Grunwald, PhDd,e,
- Stephen W. Waldo, MDd,f,
- Thomas M. Maddox, MD, MScg,
- Mladen Vidovich, MDh,i,
- Subhash Banerjee, MDj,k and
- Sunil V. Rao, MDb,c,l,∗ ()
- 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
- dDenver 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
- kVeterans Affairs North Texas Health Care System, Dallas, Texas
- lDurham Veterans Affairs Medical Center, Durham, North Carolina
- ↵∗Address for correspondence:
Dr. Sunil V. Rao, Durham VA Medical Center, 508 Fulton Street (111A), Durham, North Carolina 27705.
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 procedural 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 = 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.
- multivessel coronary artery disease
- multivessel coronary revascularization
- one-time multivessel revascularization
- percutaneous coronary intervention
- staged percutaneous coronary intervention
Multivessel coronary artery disease (MVCAD) is present in approximately two-thirds of patients who require revascularization (1). Two options exist when performing multivessel percutaneous coronary intervention (MVPCI): staged percutaneous coronary intervention (PCI) or one-time multivessel revascularization (OTMVR). In staged PCI, patients often receive revascularization of a culprit (thrombotic) artery or the most severe stenosis, followed by a planned revascularization of a nonculprit artery or arteries or less severe or complex stenosis at a later time. This is in contrast to OTMVR, in which more than 1 diseased lesion is revascularized at the time of the index procedure. Patients with MVCAD who undergo MVPCI constitute a heterogenous patient population. The decision to perform staged PCI or OTMVR may be related to many factors, including patient-specific factors (e.g., chronic kidney disease, indication for PCI such as ST-segment elevation myocardial infarction [STEMI], non–ST-segment elevation myocardial infarction [NSTEMI], unstable angina, or stable angina), physician- or hospital-specific factors, and procedural characteristics (e.g., anatomic complexity, disease burden, presence of chronic total occlusion, contrast load, radiation exposure). Each of these factors plays an important role in the decision-making process in MVPCI. For example, patients with stable angina and chronic total occlusions who undergo MVPCI are more likely to undergo staged PCI (2). Furthermore, among interventional cardiologists, renal function is the strongest factor in the decision to pursue staged PCI (3).
In patients with STEMI, several observational studies have suggested that multivessel staged PCI may be associated with lower mortality compared with OTMVR (4–7). Observational data comparing staged versus OTMVR in those with NSTEMI are lacking. Two issues hinder the use of these data to guide practice. First, the optimal timing and associated outcomes of MVPCI across the spectrum of clinical presentation remain unclear, and second, observational studies are subject to multiple biases, including unmeasured confounding. To date, there are insufficient observational data and limited randomized controlled trials to provide recommendations on the optimal timing of MVPCI (8).
Using standardized electronic health system data from the national Veterans Affairs (VA) CART-CL (Clinical Assessment, Reporting, and Tracking System for Catheterization Laboratories) Program, we sought to determine: 1) factors associated with the decision to perform staged PCI or OTMVR among patients who underwent MVPCI, including unmeasured site variation; and 2) the association between staged PCI or OTMVR and clinical outcomes (death, rehospitalization, transfusion) (9).
The VA CART Program is a national VA clinical quality initiative that was launched in 2005 and has been described previously (10). A key feature of the CART Program is standardized data capture and reporting across all VA catheterization laboratories through a clinical software application. The software is embedded in the VA electronic health record and allows providers to enter patient and procedural information (pre-procedural assessment, cardiac catheterization, and PCI) as part of routine clinical work flow. The application is integrated with the VA’s patient electronic health record, and data elements automatically populate a clinical note. These data are combined with longitudinal VA administrative data in a clinical data repository to support the quality assessment, quality improvement, and clinical research missions of the CART Program. Data elements conform to the definitions and standards of the American College of Cardiology’s National Cardiovascular Data Registry (9,11). Cardiac catheterization reports generated by the VA CART Program have been shown to demonstrate excellent data validity and completeness (10). This study was approved by the Institutional Review Board at the Denver VA Medical Center and the Durham VA Medical Center, with a waiver of the requirement to obtain informed consent.
Patients included in the analysis were those who had MVCAD and underwent PCI to 2 or more coronary artery distributions at a VA catheterization laboratory between October 1, 2007, and September 30, 2014, and had their records collected in CART. Patients were excluded from analyses if they had left main coronary artery treatment, presented with cardiogenic shock, underwent PCI in the previous 60 days either inside or outside of the VA, underwent more than 1 staged procedure, or were at a VA catheterization laboratory that performed no staged PCI or had performed 10 or fewer PCIs overall. If a patient underwent more than 1 MVPCI over this time frame, only the first PCI was used. Patients with missing data were not included in the analyses (Figure 1).
Definitions and outcomes
OTMVR was defined as PCI of >2 diseased coronary artery distributions at the time of an index procedure. A diseased left anterior descending, left circumflex, or right coronary artery was defined by a segment with >70% stenosis in a main branch or side branch of an epicardial artery.
To be included in the staged PCI group, a subsequent PCI of a diseased coronary artery had to be performed within 60 days of an index procedure. The second procedure could not have treated a segment that had been treated during the index procedure. The subsequent procedure could not have a PCI status of emergent, urgent, or salvage. It also could not have an indication of STEMI or NSTEMI. If comments in CART-CL indicated that a patient should be scheduled for a staged procedure, the patient was included as part of the staged PCI group using an intention-to-treat approach.
The primary efficacy outcome was all-cause death through latest follow-up. Secondary outcomes included all-cause rehospitalization and rehospitalization for myocardial infarction through latest follow-up. These were time-to-event outcomes using the latest discharge date from the index or staged procedure as a baseline date. Follow-up data were available through September 30, 2015, and thus patients with no events were censored as of this date. For outcomes other than mortality, patients were censored at the date of death or the last date of follow-up. The primary safety outcome was the rate of blood transfusions within 72 h after both the index PCI and subsequent PCI if a patient had a staged procedure.
With the exception of body mass index and baseline renal function, International Classification of Diseases-9th Revision codes in prior inpatient, outpatient, and CART records were used to determine medical comorbidities. Two outpatient records or 1 inpatient record with the International Classification of Diseases-9th Revision code of interest were needed to confirm a comorbid condition.
Information from either the index procedure or a combination of the index and staged procedure was used for the procedural characteristic variables. Number of stents used and number of diseased vessels treated were summed over both procedures if a patient underwent multiple PCIs. PCI status, indication, and access site (radial vs. femoral) were used from the index procedure. If a patient had multiple access sites listed for the index procedure, the second access site was used.
Site regional status was based on U.S. census regions (West, Midwest, South, and Northeast). Certain patient-level characteristics that may better describe the catheterization laboratory site, such as racial breakdown and urban or rural status, were summarized at the site level by taking the overall proportion of patients in the cohort with the characteristic of interest at the site.
All data were reviewed for missing values and quality. Some patients were missing values for body mass index (1.2%) and urban or rural status (1.1%), and a single imputation was performed simulating values on the basis of associations with observed data; 449 patients with other missing values were excluded because of the difficulty in correctly imputing values.
A mixed logistic regression model with a random intercept for site was estimated using the glmmML package in R version 3.2.5 (R Foundation for Statistical Computing, Vienna, Austria). Associations of individual measured patient and site factors with staged PCI versus OTMVR, adjusted for other patient, procedure, and site factors, were examined using this model. We used reference effect measure (REM) methods to explore the extent to which unmeasured site variation contributed to the decision to use staged PCI relative to sets of measured factors (12). A REM value is an odds ratio (OR) comparing a patient at a given risk percentile with a median-risk patient. For variation between sites not accounted for by measured patient or site characteristics (unmeasured site variation), this is based on the estimated normal distribution of the site random effects. For variation due to measured demographics, comorbidities, procedure, or site characteristics, REM is based on the corresponding empirical distributions of patient risk scores. For each distribution, the OR for staged PCI for a low-risk (2.5th percentile) and a high-risk (97.5th percentile) patient relative to a median-risk patient formed 95% REM ranges. Wider ranges indicated greater differences in “risk” or probability of having a staged procedure on the basis of the factors of interest and thus suggested a larger contribution to the procedural decision.
A propensity score approach using inverse probability of treatment weighting assessed the association between staged PCI and efficacy outcomes balancing for differences in patient risk. A hybrid method using both parametric and nonparametric models was used to produce propensity scores and weights (13). Variables in the propensity score models included patient demographics (age, sex, race, urban vs. rural), comorbidities (body mass index, congestive heart failure, chronic kidney disease, chronic obstructive pulmonary disease, diabetes, hyperlipidemia, hypertension, peripheral artery disease, prior myocardial infarction, prior coronary artery bypass graft surgery, prior PCI, prior stroke or transient ischemic attack, tobacco use), procedural characteristics (use of at least 1 bare-metal stent, number of diseased vessels, number of stents, indication, status), site characteristics (proportion of urban patients, proportion of white patients, coronary artery bypass graft surgery on-site), and time trend. A Cox proportional hazards model using these weights and a frailty term for site was then estimated using the coxme package, and we report the hazard ratio (HR) for staged PCI versus OTMVR. Diagnostics from the propensity-weighted cohort are provided in Online Table 1 and Online Figure 1. Sensitivity analyses incorporating unmeasured site variation into the propensity score also produced similar results (14). Additionally, sensitivity analyses for unmeasured confounding examined changes in the association between OTMVR versus staged PCI and mortality with the inclusion of a hypothetical unmeasured confounder with assumed prevalence in each group and strength of association with the outcome (15) (Online Table 2).
Because of the small number of events in the safety outcome, an unadjusted chi-square test compared the proportion of patients with blood transfusions by staged PCI versus OTMVR. Results were restricted to the 7,578 of 7,599 patients who survived to the 72-h cutoff after their last procedure. Finally, subgroup analyses were performed using the previously described propensity score weighting approach among patients presenting with acute coronary syndrome (ACS), non-ACS, STEMI, NSTEMI, unstable angina, and stable angina. All analyses were performed by the VA CART analytic center at the Denver VA Medical Center using R version 3.2.5.
A total of 10,498 patients underwent MVPCI at 67 sites between October 1, 2007, and September 30, 2014. After exclusion criteria were applied, the final study sample included 7,599 patients at 61 sites (Figure 1). Of the patients in the analysis cohort, 2,743 (36.1%) underwent staged PCI, while 4,856 (63.9%) underwent OTMVR.
A summary of baseline patient and procedure characteristics grouped by staged PCI versus OTMVR is shown in Table 1. Including both the index and subsequent staged procedures, patients with MVCAD who underwent staged PCI had higher rates of treated 3-vessel disease (12.3% vs. 4.4%; p < 0.01), STEMI presentation (8.1% vs. 3.0%; p < 0.01), and a PCI status of emergent or salvage (8.0% vs. 3.4%; p < 0.01) compared with patients with OTMVR. The mean total number of stents placed during index PCI and subsequent PCI for those who underwent staged procedures (3.2) was greater than that for patients who underwent OTMVR (2.6) (p < 0.01). Baseline mean glomerular filtration rate did not differ between those who underwent staged PCI versus OTMVR (mean glomerular filtration rate 74.7 vs. 75.2 ml/min/1.73 m2; p = 0.38). Compared with OTMVR, patients undergoing staged PCI had higher rates of atherectomy use (14.4% vs. 8.3%; p < 0.01), treated chronic total occlusions (16.4% vs. 8.8%; p < 0.01) and treated calcified lesions (39.4% vs. 31.3%; p < 0.01) during either the index or staged procedure.
Factors associated with staged PCI
Table 2 shows adjusted associations between individual factors and staged PCI versus OTMVR. With respect to baseline patient characteristics, the only comorbidity associated with staged PCI was chronic kidney disease (adjusted OR: 1.31; 95% confidence interval [CI]: 1.14 to 1.52; p < 0.01). Procedural characteristics associated with staged PCI included 3-vessel disease treated versus 2-vessel disease (adjusted OR: 2.17; 95% CI: 1.77 to 2.68; p < 0.01) and STEMI presentation (adjusted OR: 2.48; 95% CI: 1.74 to 3.52; p < 0.01).
In Figure 2, we observe a wide 95% REM range for unmeasured site variation (OR: 0.20 to 5.08), which was comparable in width with the 95% REM range for procedural characteristics (OR: 0.44 to 6.17) and exceeded demographics (OR: 0.79 to 1.11), comorbidities (OR: 0.69 to 1.39), and measured site characteristics (OR: 0.44 to 2.14). On the basis of unmeasured site variation, a patient at a low-“risk” site (2.5th percentile) has odds of a staged procedure 0.20 times lower compared with a patient a median-“risk” site, while a patient at a high-“risk” site (97.5th percentile) has odds 5.08 times greater. The wide ranges show large differences in the likelihood of receiving staged PCI due to unmeasured site variation and procedural characteristics, indicating they are driving the procedural decision.
Association between staged PCI and efficacy outcomes
The median time of follow-up was 43.7 months for the primary outcome of all-cause death. In unadjusted analysis, staged PCI was associated with a lower risk for death compared with OTMVR (HR: 0.76; 95% CI: 0.68 to 0.85; p < 0.01). After adjustment including applying the propensity scores weights, staged PCI was still associated with a lower risk for death (HR: 0.78; 95% CI: 0.72 to 0.84; p < 0.01). For the secondary outcomes of all-cause rehospitalization and rehospitalization for myocardial infarction after propensity adjustment, there were no statistically significant differences (Table 3). Sensitivity analyses (Online Table 2) show the effect of an unmeasured confounder with varying prevalence and its association with death.
Association between staged PCI and transfusion
The proportion of patients receiving blood transfusion within 72 h of both index PCI and subsequent PCI if a patient underwent a staged procedure was not significantly different between the staged PCI (1.4%) and OTMVR (1.3%) groups (p = 0.96).
Similar HRs for mortality were observed when comparing staged PCI with OTMVR among patients who presented with ACS (adjusted HR: 0.73; 95% CI: 0.66 to 0.81; p < 0.01). This significant relationship was consistent in those who presented with STEMI (HR: 0.31; 95% CI: 0.21 to 0.47; p < 0.01), NSTEMI (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), or stable angina (HR: 0.88; 95% CI: 0.77 to 1.00; p = 0.048) (Table 4).
In this contemporary study of patients with MVCAD undergoing staged PCI or OTMVR, we report several key findings: 1) although some patient and procedure factors are associated with staged PCI, site variation due to unmeasured factors had a substantial impact on the decision to perform staged PCI; 2) after adjustment for differences in patient and procedure characteristics, staged PCI was associated with lower risk for death compared with OTMVR overall and across all subgroups; and 3) overall the rates of bleeding with staged PCI and OTMVR were low and not significantly different between groups. However, given the influence of unmeasured site variation and potential unmeasured confounding on these relationships, further exploration of which factors (e.g., patient preference, physician preference) is warranted, and a prospective randomized trial is needed to guide clinical practice in this population.
To date, there are no completed randomized trials in patients with MVCAD comparing staged versus complete revascularization. Prior studies comparing single-vessel PCI versus MVPCI have been conducted largely in patients presenting with STEMI who have MVCAD. The results of the PRAMI, CvLPRIT, and DANAMI-3-PRIMULTI randomized controlled trials found significant benefit with complete revascularization compared with culprit artery PCI alone (16–18). The results of these trials changed treatment recommendations for complete revascularization at the time of primary PCI in patients with STEMI without hemodynamic compromise from a Class IIIC to a Class IIB recommendation (8). The subsequent COMPARE ACUTE trial again showed improved overall outcomes in patients with STEMI who underwent complete revascularization versus infarct artery–only revascularization in the acute setting. This was driven by a decreased rate of future revascularization (19). The question of staged PCI after culprit artery revascularization in patients with STEMI is being evaluated in the ongoing COMPLETE trial. Notably PRAMI, CvLPRIT, DANAMI-3-PRIMULTI, and COMPARE ACUTE excluded patients with STEMI with cardiogenic shock. The recently published CULPRIT-SHOCK trial showed worse outcomes in patients with STEMI with cardiogenic shock who underwent culprit artery–only and non–culprit artery–only revascularization, suggesting that there is still a role for culprit artery–only revascularization in a select group of patients (20). In patients with STEMI, the COMPLETE trial will likely confirm a role for staged multivessel revascularization following culprit artery revascularization. However, the optimal timing of non–culprit artery revascularization in patients with STEMI remains unknown without a direct comparison between staged PCI and OTMVR.
An aspect of our analysis that is interesting is the finding that staged PCI is associated with lower long-term mortality versus OTMVR. In patient with STEMIs in particular, staged PCI had a strong association with mortality (HR: 0.31; 95% CI: 0.21 to 0.47; p < 0.01). The mortality benefit we observed is consistent with prior meta-analyses (21,22). The increased mortality in patients with STEMI who undergo culprit and non–culprit artery revascularization in the acute period may be attributed to a proinflammatory state leading to an increased risk for in-stent thrombosis with PCI of nonculprit vessels, higher contrast load leading to contrast-induced nephropathy, which has been associated with increased mortality, or greater risk for procedural complications with increased procedure times and/or intervening on nonculprit vessels without optimal evaluation (e.g., percentage diameter stenosis, fractional flow reverse, or intravascular ultrasound) (23). In those without STEMI, we observed a positive yet weaker association with lower mortality in the NSTEMI (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 = 0.05) groups. This trend toward insignificance in the stable angina group from the STEMI, NSTEMI, and unstable angina groups is likely attributed to a lesser degree of acuity and less of an inflammatory state as well as procedural related complications. This is the first analysis we are aware of to find this association across subgroups.
To assess whether a second procedure in selected patients would be associated with an increase in adverse procedural complications such as bleeding, we examined the association between staged PCI and blood transfusions. We found that there was no significant difference in post-PCI blood transfusion between the staged PCI and OTMVR groups. This is consistent with prior results in patients with MVCAD who underwent revascularization, suggesting that the decision to stage PCI is not associated with an adverse safety profile (4,17,24).
Despite the robustness of our results, they are observational in nature. Unlike prior studies, we documented the degree to which unmeasured site variation influenced the decision to pursue staged PCI using the REM approach. Compared with demographics and comorbidities, procedural characteristics and unmeasured site variation contributed more to the decision for staged PCI versus OTMVR as shown by the wide 95% REM ranges for these factors. These data indicate that local practice patterns may play a large role in this process. Factors that may also contribute to unmeasured site variation include physician preference, technical skill, differences in degree of illness at patient presentation across sites, and differences in systems-based practices. This substantial variation in clinical practice highlights the need for implementation of guideline directed care.
In addition to the significant impact of unmeasured site variation, we found that a composite of procedural factors affected the decision to perform staged PCI. Furthermore, individual patient factors, such as chronic kidney disease, and procedural factors, such as presentation with STEMI and 3-vessel coronary disease, were associated with the use of staged PCI. In the absence of clinical guidelines on the optimal timing of MVPCI, the decision to pursue staged PCI varies widely among cardiologists. This was observed in a survey performed by the American College of Cardiology (3). There is general agreement among cardiologists that renal function, lesion complexity, and presence of ACS are important factors in the decision to pursue staging (3).
First, although we controlled for known confounders and used propensity matching to ensure that the staged PCI and OTMVR groups were similar, the analysis is observational and subject to unmeasured confounding. Second, it is possible that patients who underwent index PCI with plans for staged PCI may not have returned for the subsequent procedure. We accounted for this by using an intention-to-treat analysis; however, if a planned approach was not mentioned in CART-CL, some patients may have been misclassified as OTMVR or excluded entirely if only a single vessel was treated, leading to survivor bias. Similarly, it is possible that some patients who were to undergo planned OTMVR subsequently underwent staged PCI because of procedural difficulties during the index case and were not included as part of the OTMVR group. Third, our study sample was largely male, given the demographic makeup of the veteran patient population. Similarly, the results of our study may not be generalizable to a nonveteran population. Finally, we were unable to assess the degree of angiographic complexity in our study sample, and those details may account for some of the unmeasured confounding in the propensity to undergo staged PCI versus OTMVR.
In this analysis of patients with MVCAD undergoing PCI, the decision to perform staged PCI was driven by specific patient characteristics but more by procedural characteristics and unmeasured site variation, and staged PCI was associated with a significant mortality benefit compared with OTMVR.
WHAT IS KNOWN? In patients with STEMI, several observational studies have suggested that multivessel staged PCI may be associated with lower mortality compared with OTMVR.
WHAT IS NEW? This study incorporates a heterogenous patient population with STEMI, NSTEMI, unstable angina, and stable angina who underwent either staged PCI or OTMVR. We found that the decision to pursue staged PCI was driven by specific patient characteristics, procedural characteristics, and unmeasured site variation, and staged PCI was associated with a significant mortality benefit.
WHAT IS NEXT? A randomized controlled trial evaluating the efficacy and optimal timing of staged PCI is needed to guide clinical practice.
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 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.
- Abbreviations and Acronyms
- acute coronary syndrome(s)
- confidence interval
- hazard ratio
- multivessel coronary artery disease
- multivessel percutaneous coronary intervention
- non–ST-segment elevation myocardial infarction
- odds ratio
- one-time multivessel revascularization
- percutaneous coronary intervention
- reference effect measure
- ST-segment elevation myocardial infarction
- Veterans Affairs
- Received June 5, 2018.
- Revision received July 24, 2018.
- Accepted July 31, 2018.
- Brilakis E.S.,
- Banerjee S.,
- Karmpaliotis D.,
- et al.
- Dangas G.D.,
- George J.C.,
- Weintraub W.
- Russo J.J.,
- Wells G.A.,
- Chong A.Y.,
- et al.
- Kornowski R.,
- Mehran R.,
- Dangas G.,
- et al.,
- for the HORIZONS-AMI Trial Investigators
- Iqbal M.B.,
- Nadra I.J.,
- Ding L.,
- et al.
- Manari A.,
- Varani E.,
- Guastaroba P.,
- et al.
- Levine G.N.,
- Bates E.R.,
- Blankenship J.C.,
- et al.
- Byrd J.B.,
- Vigen R.,
- Plomondon M.E.,
- et al.
- Brindis R.G.,
- Fitzgerald S.,
- Anderson H.V.,
- Shaw R.E.,
- Weintraub W.S.,
- Williams J.F.
- Glorioso T.J.,
- Grunwald G.K.,
- Ho P.M.,
- Maddox T.M.
- Zhu Y.,
- Ghosh D.,
- Mitra N.,
- Mukherjee B.
- Gershlick A.H.,
- Khan J.N.,
- Kelly D.J.,
- et al.
- Engstrøm T.,
- Kelbæk H.,
- Helqvist S.,
- et al.
- Smits P.C.,
- Boxma-de Klerk B.M.
- Thiele H.,
- Akin I.,
- Sandri M.,
- et al.
- Li Z.,
- Zhou Y.,
- Xu Q.,
- Chen X.
- Vlaar P.J.,
- Mahmoud K.D.,
- Holmes D.R. Jr..,
- et al.
- Giacoppo D.,
- Madhavan M.V.,
- Baber U.,
- et al.
- Kong J.A.,
- Chou E.T.,
- Minutello R.M.,
- Wong S.C.,
- Hong M.K.