Author + information
- Received August 10, 2015
- Revision received April 26, 2016
- Accepted May 18, 2016
- Published online August 30, 2016.
- Nadia R. Sutton, MD, MPHa,
- Milan Seth, MSa,
- Cyril Ruwende, MD, PhDb and
- Hitinder S. Gurm, MBBSa,c,∗ ()
- aDepartment of Internal Medicine, Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, Michigan
- bMichigan Heart, IHA, St. Joseph Mercy Hospital, Ypsilanti, Michigan
- cDepartment of Medicine, Section of Cardiology, Veterans Affairs Medical Center, Ann Arbor, Michigan
- ↵∗Reprint requests and correspondence:
Dr. Hitinder S. Gurm, Division of Cardiovascular Medicine, University of Michigan Cardiovascular Center, 2A381, SPC 5869, 1500 East Medical Center Drive, Ann Arbor, Michigan 48109-5869.
Background Atrial fibrillation (AF) is increasing in prevalence, and patients with a history of AF commonly undergo percutaneous coronary intervention (PCI). There is a paucity of contemporary data on the association between AF and clinical outcomes after PCI.
Objectives The study sought to evaluate the association between AF and in-hospital adverse outcomes using a large, prospective multicenter registry.
Methods Data for consecutive PCI cases from 47 hospitals performed between April 2011 and December 2014 were utilized for the analysis. Propensity-matched multivariate analysis was used to adjust for differences in baseline characteristics between patients with and without a history of AF.
Results Of 113,283 PCI cases during the study period, a history of AF was present in 13,912 patients (12%), which varied by institution (range 2.5% to 18.4%). At baseline, patients with a history of AF were older and were more likely to have comorbid congestive heart failure, cardiomyopathy, cerebrovascular disease, and chronic lung disease. Patients with a history of AF were more likely to have in-hospital complications, including in-hospital mortality (3% vs. 1%). In propensity-matched analysis, patients with a history of AF were more likely to be treated with a bare-metal stent (27% vs. 18%). In the propensity-matched model, AF remained independently associated with an increased risk of developing post-procedural bleeding (odds ratio [OR]: 1.32; 95% confidence interval [CI]: 1.15 to 1.52), heart failure (OR: 1.33; 95% CI: 1.17 to 1.52), cardiogenic shock (OR: 1.26; 95% CI: 1.08 to 1.48), and in-hospital mortality (OR: 1.41; 95% CI: 1.18 to 1.68).
Conclusions AF is common among patients undergoing PCI. AF is associated with older age, the presence of other comorbidities, and independently associated with in-hospital post-procedural heart failure, cardiogenic shock, and mortality.
Atrial fibrillation (AF) is the most common sustained arrhythmia treated in the United States (1,2). The prevalence of AF is increasing, reflective of increasing numbers of older patients and the pervasiveness of comorbid illness (2,3). Because there is overlap in risk factors for AF and coronary artery disease (CAD), patients with AF often have coexistent CAD and are treated with percutaneous coronary intervention (PCI). Patients with AF who require PCI pose a challenging clinical dilemma. In addition to the need for dual antiplatelet therapy after PCI, patients often warrant anticoagulation to lower the risk of stroke. There is a heightened level of interest in this topic, given the need to weigh the bleeding risk of these therapies against the risk of future thrombotic events.
As AF is associated with higher left atrial filling pressure and myocardial remodeling, we hypothesized that AF could be a risk factor for poor clinical outcomes in patients undergoing PCI (4,5). Despite the frequency with which this scenario is encountered in clinical practice, there are scant contemporary published reports describing the incidence and characteristics of patients with AF undergoing PCI. Furthermore, there is minimal data describing the procedural characteristics and in-hospital outcomes in this population. In this large, database-derived study using the Blue Cross Blue Shield Registry, the goal was to describe the population and to evaluate the association between AF and in-hospital adverse outcomes.
The Blue Cross Blue Shield of Michigan Cardiovascular Consortium (BMC2) is a prospective, multicenter registry that collects demographic, clinical, procedural, and in-hospital outcomes data from consecutive PCI cases at all nonfederal hospitals in the state of Michigan. Details of BMC2 data collection and audit processes have been described previously (6,7).
All PCI cases in the BMC2 registry performed between April 2011 and December 2014 were used for the analysis. Baseline characteristics between patients with a history of AF and patients without AF were compared using Pearson chi-square tests for categorical variables and Student t tests for continuous variables. Continuous variables are summarized using mean ± SD. Propensity score matching and multivariate regression models were used to adjust for differences in baseline characteristics. A propensity score was estimated using logistic regression, and patients with a history of AF were matched on a 1:1 basis using a greedy algorithm without replacement to patients without AF (8). Matched pairs were required to have propensity scores within a caliper of 0.25 SD, and were exact matched on the clinical presentation category: ST-segment elevation myocardial infarction (STEMI), non–ST-segment elevation myocardial infarction (NSTEMI), unstable angina (UA), and stable angina. Covariates included in the propensity score model are listed in Online Table 1. Patients missing covariate data were excluded from the matching exercise and thus from the matched cohort. The association between AF and outcomes was assessed univariately using Fisher exact tests and through multivariable logistic regression models fitted to the propensity-matched cohort, adjusting for variables included in the propensity score.
The primary outcome of interest was in-hospital mortality. Secondary outcomes were in-hospital post-procedural myocardial infarction, cardiogenic shock, heart failure, stroke, new requirement for dialysis, acute kidney injury, vascular complications requiring treatment, need for blood transfusion, and bleeding events within 72 h of PCI (9). In-hospital death was defined as death from both cardiac and noncardiac causes during the admission in which PCI was performed. A composite of pseudoaneurysm, arteriovenous fistula, femoral neuropathy, retroperitoneal hematoma, access site hematoma requiring transfusion or associated with prolonged hospital stay, drop in hemoglobin >3.0 g/dl, or any access site complication requiring surgical repair comprised vascular complications. Bleeding events were defined as a suspected or confirmed bleeding event observed and documented in the medical record that was associated with any of the following: 1) hemoglobin drop ≥3 g/dl; 2) transfusion of whole blood or packed red blood cells; 3) procedural intervention or surgery at the bleeding site to stop bleeding (e.g., surgical closure or exploration of the arteriotomy site, balloon angioplasty to seal an arterial tear, endoscopy with cautery for a gastrointestinal bleed).
For the primary outcome of in-hospital mortality, a subgroup analysis was performed evaluating the association of mortality and AF within subgroups of the matched patient cohort defined by a number of baseline characteristics, including age, sex, presentation category, history of heart failure, and pre-procedural left ventricular ejection fraction. These analyses were conducted using multivariate logistic regression models, adjusting for baseline covariates fitted within each subgroup. In addition, the extent to which subgroup definitions modified the effect of AF on the risk of mortality was assessed statistically through likelihood ratio tests assessing the inclusion of the subgroup by AF interaction terms in the regression model.
Data on oral anticoagulant therapy is available in the BMC2 registry for medications administered during the hospitalization. Except in rare cases, patients were not taking oral anticoagulation at the time of PCI. In our study population, the absolute numbers of inpatients receiving oral anticoagulation was small (data not shown) and could be under-representative of the use of oral anticoagulation prior to admission, as oral anticoagulants are often stopped in anticipation of PCI. Anticoagulation is typically not recommended until 48 h after PCI, and many patients are discharged prior to that time point. Therefore, patients receiving oral anticoagulation in the hospital are likely not representative of the broader population, and data on oral anticoagulant use is therefore not reported here.
Between April 2011 and December 2014, there were 113,283 PCIs performed at 47 hospitals in Michigan. Among those, 13,912 (12.3%) patients had a history of AF. Among the 47 hospitals, there was a wide range (2.5% to 18.4%) in the proportion of patients with a history of AF undergoing PCI (Figure 1). Among the presentation categories, 8.0% of STEMI, 13.9% of NSTEMI, 12.4% of UA, and 12.0% of stable angina patients had a history of AF. UA was the most frequent indication for PCI in patients with and without a history of AF. Baseline demographic and clinical features are shown in Table 1. Compared with patients without AF, patients with a history of AF were more likely to be older and to have prior congestive heart failure, cardiomyopathy, cerebrovascular disease, prior CAD, hypertension, chronic lung disease, diabetes, and chronic kidney disease. Compared with patients with a history of AF, patients without AF were more likely to smoke. Patients with a history of AF were more likely to present with NSTEMI and less likely to present with STEMI, compared with patients without AF. Compared with patients without AF, patients with a history of AF were more likely to present in cardiogenic shock or with cardiac arrest, and to be in cardiogenic shock at the start of PCI.
Procedural characteristics for patients with and without a history of AF are shown in Table 2. At baseline, patients with a history of AF were more likely to receive a bare-metal stent (BMS) or angioplasty alone and less likely to receive a drug-eluting stent (DES), compared with patients without AF. Data for calculation of CHADS2 score were available for a subset of patients undergoing PCI after January 2013 (10). Use of angioplasty alone correlated positively with CHADS2 score in patients with a history of AF, but there was no trend in the percent use of DES or BMS with increasing CHAD2 scores (Online Table 2). Compared with patients without AF, patients with a history of AF were more likely to require intra-aortic balloon pump placement or other mechanical support. PCI was categorized more often as urgent and less often as emergent or elective in patients with a history of AF, compared with patients without AF.
Data on intraprocedural medications and in-hospital antiplatelet therapy is shown in Table 3. At baseline, patients with a history of AF were more likely to receive heparin alone and less likely to receive a glycoprotein IIb/IIIa inhibitor in addition to heparin. Compared with patients without AF, patients with a history of AF were slightly more likely to receive bivalirudin. Those with a history of AF more frequently received clopidogrel and less frequently received prasugrel or ticagrelor, compared with those without AF.
In-hospital outcomes for patients undergoing PCI are shown in Table 4. Patients with a history of AF were more likely to have post-procedural cardiogenic shock, heart failure, stroke, acute kidney injury, dialysis initiation, vascular complications, blood transfusion, bleeding complications, and in-hospital mortality.
A total of 13,498 patients with a history of AF were successfully matched to similar patients without AF. In the matched cohort, the absolute standardized difference for all variables was ≤2.7, indicating that the cohort was well matched (Table 1, Online Figure 1). Differences in choice of stent remained after propensity matching; patients with a history of AF were more likely to receive a BMS or angioplasty alone, and less likely to receive a DES compared with the matched cohort (Table 2). After matching, compared with patients without AF, patients with a history of AF were more likely to have had radial arterial access and less likely to have had femoral arterial access.
In-hospital adverse outcomes after propensity matching are shown in Table 4 and in the Central Illustration. After propensity matching, compared with patients without AF, patients with a history of AF were more likely to have post-procedural bleeding, a need for blood transfusion, heart failure, cardiogenic shock, and in-hospital mortality. There were no differences in the incidence of post-procedural stroke, acute kidney injury, dialysis initiation, or vascular complications in the matched cohorts. Outcome analysis using a conditional logistic regression model to account for the paired nature of the matched cohort yielded results that were consistent with the results of the multivariable logistic regression model (data not shown).
Subgroup analysis revealed that the association between AF and risk-adjusted mortality was consistent across subgroups (Figure 2). No subgroup by AF interactions (including age) was associated with a statistically significant improvement in model fit. The standardized differences of baseline characteristics of patients with and without a history of AF, by subgroup, are shown in Online Figure 2.
The key finding in this study is that AF is strongly associated with in-hospital adverse outcomes, even after extensive adjustment for other relevant factors. A history of AF was present in 12% of patients undergoing PCI in this study, highlighting how commonly this scenario occurs in contemporary clinical practice, although the prevalence of AF in patients undergoing PCI varied considerably by institution. A history of AF was associated with the presence of other medical conditions, and with the presence of cardiac arrest and cardiogenic shock on presentation. Those with a history of AF were found to have a higher rate of post-procedural bleeding, need for transfusion, heart failure, cardiogenic shock, and mortality. These findings indicate that a history of AF is a marker of patients who are vulnerable to serious in-hospital complications after PCI.
Consistent with prior reports, we found that patients with a history of AF were older and had more comorbid illness (11–16). After propensity matching, AF remained an independent risk factor for poor clinical outcomes. Patients with a history of AF had a higher rate of in-hospital mortality. This was consistent across subgroups. The association between AF and mortality has been described, with the bulk of the literature relating to AF in the setting of acute myocardial infarction. Most prior studies have found an association between AF and mortality, with variations in populations studied, including all patients with acute coronary syndrome (ACS) (11,17), STEMI (14,15,18–25), NSTEMI (19,23–26), and UA (27), and all presentations for PCI (12,16). To our knowledge, among recent studies in which risk adjustment for other factors known to influence poor outcomes was performed, there is only 1 other report of the association of AF with short-term mortality in the setting of STEMI (14). Here we report a 1% risk of in-hospital mortality in patients without AF, compared with a 3% risk of mortality in patients with a history of AF, and 2% versus 3% mortality, respectively, after propensity matching. It is noteworthy that the exaggerated risk of mortality among patients with a history of AF was noted across the spectrum of clinical presentation, and should be taken into account when considering risk estimation and benchmarking.
We found that patients with a history of AF were more likely to present in cardiogenic shock or with cardiac arrest. In line with prior reports, patients with a history of AF were more likely to experience post-procedural heart failure and cardiogenic shock (14,17,18,22,23). As with mortality, only 1 other recent report concluded that AF was associated with post-procedural cardiogenic shock or cardiac arrest after risk adjustment in the setting of STEMI (14). Even though patients had undergone PCI, those with AF may have had a greater burden of nonrevascularized coronary artery disease, which could have contributed to this finding. It has been theorized that AF is a clinical manifestation of a fibrotic atrial cardiomyopathy that can occur independently of age and comorbidities (28,29). Whether AF drives this structural process or is a consequence is unclear. The data presented herein are consistent with the theory that AF represents underlying cardiomyopathy, as a history of AF was found to be an independent risk factor for poor clinical outcomes related to hemodynamic alterations and pump function.
At baseline, there was a higher risk of bleeding and stroke in patients with a history of AF, similar to prior reports (11–13,16,20,22). Following propensity matching, we did not observe an increased risk of stroke, but an increased risk of bleeding remained. Except for rare circumstances, oral anticoagulation is held pre- and post-procedurally. Most patients with AF resume anticoagulation 48 h post-procedure, often after discharge. These data suggest that in-hospital bleeding complications may be related to periprocedural anticoagulation and antiplatelet agents. We found that clopidogrel was used more frequently and prasugrel and ticagrelor less frequently in patients with AF relative to the comparator group. Patients with a history of AF were less likely to receive eptifibatide intraprocedurally, were more likely to receive heparin alone, and had similar rates bivalirudin utilization. This suggests that clinicians are attempting to mitigate a presumed elevated risk of bleeding in this population. Despite this, patients with a history of AF were still more likely to have a bleeding complication or require a blood transfusion. The absence of a detectable difference in the incidence of stroke and post-procedural myocardial infarction between cohorts differs from prior reports describing an increased risk of these poor outcomes in patients with AF (14,15). This inconsistency could be due to improvements in treatment practices over time or differences in time points examined (14–16).
We report here that patients with a history of AF were far more likely to receive a BMS or angioplasty alone, and were less likely to receive a DES. This finding likely stems from the need for oral anticoagulation in many patients with AF. Selecting a BMS affords the possibility of shortening the duration of dual antiplatelet therapy plus oral anticoagulation (triple therapy) if a bleeding event occurs. In general, current guidelines suggest that the benefit of DES with regard to restenosis be weighed against the bleeding risk associated with triple therapy, and that DES be avoided in patients who have a higher bleeding risk and, when possible, in patients who will require oral anticoagulation long term (30–33).
A prior study spanning 2004 to 2006 also found increased use of BMS in patients with AF, although the absolute percent of patients receiving BMS has declined substantially in all patients in the interval (14). We are unable to discern from these data the exact reasons for increased BMS use in patients with a history of AF in this study. Additional explanations for this observation are concern for bleeding on triple therapy, medication compliance, a history of bleeding, or other unmeasured confounders.
These data are important for several reasons. First, this report defines the prevalence of AF in patients undergoing PCI in contemporary clinical practice, in itself a useful observation. We found that the prevalence of a history of AF in patients undergoing PCI was 12.3%, which is higher than in recent reports (5% to 11%) (11,12,14–16,19). It is possible that this reflects the increased prevalence of AF in the general population, as well as the increased prevalence of obesity, a known risk factor for AF in the PCI population (34). Furthermore, we demonstrate a wide range in the proportion of patients with a history of AF undergoing PCI by institution. These data strongly suggest that AF should be taken into account for individual patients, as well as when grading institutions on quality and the outcomes of patients undergoing PCI.
We identified that patients with a history of AF undergoing PCI are more susceptible to developing post-procedural complications of bleeding, heart failure, cardiogenic shock, and mortality, and acknowledgment of this potential may be valuable clinically. The noted association is an important reminder that the presence of AF marks underlying pathology. The association of AF with mortality after PCI was consistent across a spectrum of presentation types. Resources could be targeted to further evaluate the factors responsible for this association and to lower morbidity and mortality in this high-risk subgroup.
This represents the largest contemporary report on the association of AF with clinical outcomes after PCI. The BMC2 database was sufficient to report on >25,000 propensity-matched cases. Patients were not excluded on the basis of eligibility for clinical trials; therefore, this is truly reflective of patients treated in clinical practice. In this study, AF could have occurred at any time prior to PCI, not only periprocedurally. As a result, these data capture a broader population of patients with AF. This is of relevance because there are implications for stent choice and antiplatelet agents in patients with a history of AF, as well as in patients with new-onset AF at the time of PCI. Other recent studies have been limited by size, limitation to DES, a single presentation type, or lack of a comparator group (12,14–16,35). The BMC2 was an ideal database to use to overcome these limitations, with detailed information on baseline clinical variables, presentation, procedural variables, and clinical outcomes.
Limitations of this study include that this is an analysis of observational, nonrandomized data. This report includes data from 1 state, and it is unknown to what degree geographic variations in treatment patterns may influence the results. The data is drawn from hospitals participating in a quality-improvement initiative, and it is unclear if the findings would apply to hospitals not participating in such an initiative. Although there was extensive adjustment for risk factors, residual measured or unmeasured confounding factors may exist. The specific reasons for stent choice and choice of antiplatelet agent were not available. We do not report on the association of oral anticoagulant use with clinical outcomes because patients were rarely administered these medications periprocedurally. We are unable to correlate with long-term clinical adverse outcomes.
A history of AF is common among patients presenting for PCI, and is associated with comorbid illnesses. The prevalence of a history of AF in patients undergoing PCI varies substantially by institution. Patients with a history of AF were more likely to receive a BMS or angioplasty alone, and less likely to receive a DES. We found that a history of AF is associated with a higher risk of in-hospital complications after PCI, including bleeding, heart failure, cardiogenic shock, and mortality. These findings highlight the importance of AF as an independent risk factor for poor clinical outcomes after PCI. AF can be viewed as a harbinger of in-hospital complications, including mortality, after PCI.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: A history of AF in patients undergoing PCI is associated with an increased risk of in-hospital adverse outcomes, including bleeding, post-procedural heart failure, cardiogenic shock, and mortality.
TRANSLATIONAL OUTLOOK: Prospective clinical studies are needed to evaluate the impact of prophylactic therapies on post-procedural morbidity and mortality in patients with a history of AF undergoing PCI.
For supplemental figures and tables, please see the online version of this article.
Executive and clinical support for the Blue Cross Blue Shield of Michigan Cardiovascular Consortium is provided by Blue Cross and Blue Shield of Michigan (BCBSM), and by Blue Care Network under the aegis of BCBSM’s Value Partnerships program. Although Blue Cross and Blue Shield of Michigan (BCBSM) and BMC2 work collaboratively, the opinions, beliefs and viewpoints expressed by the authors do not necessarily reflect the opinions, beliefs, and viewpoints of BCBSM or any of its employees. Dr. Gurm has received research funding from Blue Cross Blue Shield of Michigan and the National Institutes of Health; and has served as a consultant to Osprey Medical. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- Blue Cross Blue Shield of Michigan Cardiovascular Consortium
- bare-metal stent
- coronary artery disease
- drug-eluting stent
- non–ST-segment elevation myocardial infarction
- percutaneous coronary intervention
- ST-segment elevation myocardial infarction
- unstable angina
- Received August 10, 2015.
- Revision received April 26, 2016.
- Accepted May 18, 2016.
- American College of Cardiology Foundation
- Miyasaka Y.,
- Barnes M.E.,
- Gersh B.J.,
- et al.
- Hirsh B.J.,
- Copeland-Halperin R.S.,
- Halperin J.L.
- Moscucci M.,
- Rogers E.K.,
- Montoye C.,
- et al.
- NCDR CathPCI Registry v4
- Lopes R.D.,
- Pieper K.S.,
- Horton J.R.,
- et al.
- Lopes R.D.,
- Elliott L.E.,
- White H.D.,
- et al.
- Rene A.G.,
- Généreux P.,
- Ezekowitz M.,
- et al.
- Kinjo K.,
- Sato H.,
- Ohnishi Y.,
- et al.,
- Osaka Acute Coronary Insufficiency Study (OACIS) Group
- Pizzetti F.,
- Turazza F.M.,
- Franzosi M.G.,
- et al.,
- GISSI-3 Investigators
- Pedersen O.D.,
- Bagger H.,
- Køber L.,
- et al.
- Crenshaw B.S.,
- Ward S.R.,
- Granger C.B.,
- et al.,
- GUSTO-I Trial Investigators
- Rathore S.S.,
- Berger A.K.,
- Weinfurt K.P.,
- et al.
- Al-Khatib S.M.,
- Pieper K.S.,
- Lee K.L.,
- et al.
- Calkins H.,
- Kuck K.H.,
- Cappato R.,
- et al.
- O'Gara P.T.,
- Kushner F.G.,
- Ascheim D.D.,
- et al.
- Authors/Task Force Members,
- Steg P.G.,
- James S.K.,
- et al.
- European Heart Rhythm Association,
- European Association for Cardio-Thoracic Surgery,
- Camm A.J.,
- et al.,
- ESC Committee for Practice Guidelines
- Lip G.Y.H.,
- Huber K.,
- Andreotti F.,
- et al.
- Buschur M.E.,
- Smith D.,
- Share D.,
- et al.
- Ruiz-Nodar J.M.,
- Marin F.,
- Hurtado J.A.,
- et al.