Abstract
Recent guidelines for implantable cardioverter-defibrillator (ICD) use in patients with early ventricular arrhythmia (VA) after acute myocardial infarction (MI) are based on systolic function and revascularization status, yet decision to implant an ICD remains highly subjective. We aimed to determine characteristics, utilization of ICDs, and long-term outcomes of survivors of early VA (<48 h) after acute MI.
We retrospectively analyzed clinical characteristics, ICD therapies, and survival in 128 patients with early VA after acute MI from 2002–12. Patients were classified for appropriateness of ICD implantation, per 2013 Appropriate Use Criteria (AUC). In 128 early VA survivors after MI, older age, female gender, history of coronary artery bypass graft surgery (CABG) or MI, non-ST-elevation MI or ventricular tachycardia (VT) at presentation predicted worse overall survival (all P < 0.05). While left ventricular ejection fraction (LVEF) did not predict mortality (HR = 1; P = 0.86), post-MI ventricular fibrillation (VF) portended a better long-term prognosis than VT (HR = 0.37; P = 0.001). Twenty-six (20%) early VA survivors received ICD, corresponding well with AUC. Implantable cardioverter-defibrillator recipients had lower post-MI LVEF (P = 0.02) and more frequently presented with non-ST-elevation MI (P = 0.007). Over 2.4 years of median follow-up, ICD recipients had a greater mortality rate than non-ICD recipients (42 vs. 17%; P = 0.02). Appropriate and inappropriate ICD discharges were high in ICD recipients.
Early VA survivors after MI receiving ICD due to suspected non-reversible arrhythmogenic substrate have high rates of appropriate ICD therapy and mortality. Our ICD implantation practice corresponds well with the AUC. Sustained monomorphic VT and non-ST-elevation MI at presentation predict increased risk for death. Larger prospective studies are necessary to confirm our findings, such as to provide evidence for future ICD guidelines.
The 2013 Appropriate Use Criteria (AUC) offer guidance for implantable cardioverter-defibrillator (ICD) use in early ventricular arrhythmia (VA) survivors, stratifying patients based on revascularization status and systolic function.
Between 2002 and 2012, ICD implantation practices for early VA at our institution correspond well with the AUC.
Patients treated with ICD for early VA after acute myocardial infarction (MI) had high rates of appropriate ICD discharge, but still had a higher mortality rate.
Predictors for worse outcome in early VA survivors include female gender, older age, prior cardiac history, NSTEMI (vs. STEMI) as the presenting acute MI type, and monomorphic ventricular tachycardia (vs. ventricular fibrillation) as the presenting VA type, while left ventricular ejection fraction at presentation was not associated with long-term outcomes.
Future ICD guidelines should take into account cardiac comorbidities and presenting MI and VA type.
Introduction
Although the prevalence of ventricular arrhythmias (VAs) after acute myocardial infarction (MI) has steadily decreased over time due to the advent of thrombolytic therapy and percutaneous coronary intervention (PCI),1–3 ∼5% of individuals still experience VA following an MI.4,5 Post-MI VA is traditionally divided into two types: early VA occurs within 48 h of MI symptom onset, whereas late VA occurs over 48 h after MI.6,7 Early VA is often attributed to the transient ischaemic milieu and is thought to have a benign long-term prognosis if the underlying ischaemia is relieved by revascularization. Late VA, particularly sustained monomorphic ventricular tachycardia (VT), is more frequently due to a reentrant mechanism resulting from the presence of scar tissue,8,9 and the arrhythmogenic substrate is thought to be irreversible in these patients. While consensus guidelines agree that survivors of late VA after MI should be given an implantable cardioverter-defibrillator (ICD) prior to hospital discharge,7,10 the utility of ICDs for patients experiencing early VA after MI is less definite.
The ACCF/HRS/AHA/ASE/HFSA/SCAI/SCCT/SCMR 2013 Appropriate Use Criteria (AUC) for ICD and Cardiac Resynchronization Therapy11 offers recommendations for ICD therapy in early VA survivors after MI, classifying patients into three groups for appropriateness of ICD implantation (Table 1). Among early VA survivors in whom total revascularization is completed or not indicated, the AUC states that ICD implantation may be appropriate for those with LVEF ≤35%, while it is rarely appropriate for those with LVEF >35%. In early VA survivors who are incompletely revascularized or not amenable to revascularization ICD is appropriate if LVEF is ≤35% and maybe appropriate if LVEF >35%.11
2013 AUC recommendations for ICD implantation for secondary prevention in patients with early VAs (<48 h) after acute MI (newly diagnosed, no prior assessment of LVEF)
| Indication | Appropriate use score (1–9) | ||
|---|---|---|---|
| LVEF | |||
| ≥50% | 36–49% | ≤35% | |
| Total revascularization completed after VA | |||
| 1. Single episode early (<48 h) VF or polymorphic VT | R (2) | R (3) | M (4) |
| 2. Recurrent early VF or polymorphic VT | R (3) | R (3) | M (5) |
| 3. Early VF or polymorphic VT | M (5) | A (7) | A (8) |
| Nonsustained VT 4 days post-MI | |||
| Inducible VT/VF at electrophysiologic study at least 4 days after revascularization | |||
| No revascularization indicated (i.e. no significant coronary artery disease) | |||
| 4. Single episode early VF or polymorphic VT | R (2) | R (3) | M (4) |
| 5. Recurrent early VF or polymorphic VT | R (2) | R (3) | M (5) |
| Obstructive coronary artery disease with coronary anatomy not amenable to revascularization | |||
| 6. Early VF or polymorphic VT | M (5) | M (5) | A (7) |
| No electrophysiological study done | |||
| Indication | Appropriate use score (1–9) | ||
|---|---|---|---|
| LVEF | |||
| ≥50% | 36–49% | ≤35% | |
| Total revascularization completed after VA | |||
| 1. Single episode early (<48 h) VF or polymorphic VT | R (2) | R (3) | M (4) |
| 2. Recurrent early VF or polymorphic VT | R (3) | R (3) | M (5) |
| 3. Early VF or polymorphic VT | M (5) | A (7) | A (8) |
| Nonsustained VT 4 days post-MI | |||
| Inducible VT/VF at electrophysiologic study at least 4 days after revascularization | |||
| No revascularization indicated (i.e. no significant coronary artery disease) | |||
| 4. Single episode early VF or polymorphic VT | R (2) | R (3) | M (4) |
| 5. Recurrent early VF or polymorphic VT | R (2) | R (3) | M (5) |
| Obstructive coronary artery disease with coronary anatomy not amenable to revascularization | |||
| 6. Early VF or polymorphic VT | M (5) | M (5) | A (7) |
| No electrophysiological study done | |||
Adapted, with permission, from J Am Coll Cardiol.11
A, Appropriate; AUC, Appropriate Use Criteria; ICD, implantable cardioverter-defibrillator; LVEF, left ventricular ejection fraction; M, Maybe appropriate; MI, myocardial infarction; R, Rarely appropriate; VF, ventricular fibrillation; VT, ventricular tachycardia.
2013 AUC recommendations for ICD implantation for secondary prevention in patients with early VAs (<48 h) after acute MI (newly diagnosed, no prior assessment of LVEF)
| Indication | Appropriate use score (1–9) | ||
|---|---|---|---|
| LVEF | |||
| ≥50% | 36–49% | ≤35% | |
| Total revascularization completed after VA | |||
| 1. Single episode early (<48 h) VF or polymorphic VT | R (2) | R (3) | M (4) |
| 2. Recurrent early VF or polymorphic VT | R (3) | R (3) | M (5) |
| 3. Early VF or polymorphic VT | M (5) | A (7) | A (8) |
| Nonsustained VT 4 days post-MI | |||
| Inducible VT/VF at electrophysiologic study at least 4 days after revascularization | |||
| No revascularization indicated (i.e. no significant coronary artery disease) | |||
| 4. Single episode early VF or polymorphic VT | R (2) | R (3) | M (4) |
| 5. Recurrent early VF or polymorphic VT | R (2) | R (3) | M (5) |
| Obstructive coronary artery disease with coronary anatomy not amenable to revascularization | |||
| 6. Early VF or polymorphic VT | M (5) | M (5) | A (7) |
| No electrophysiological study done | |||
| Indication | Appropriate use score (1–9) | ||
|---|---|---|---|
| LVEF | |||
| ≥50% | 36–49% | ≤35% | |
| Total revascularization completed after VA | |||
| 1. Single episode early (<48 h) VF or polymorphic VT | R (2) | R (3) | M (4) |
| 2. Recurrent early VF or polymorphic VT | R (3) | R (3) | M (5) |
| 3. Early VF or polymorphic VT | M (5) | A (7) | A (8) |
| Nonsustained VT 4 days post-MI | |||
| Inducible VT/VF at electrophysiologic study at least 4 days after revascularization | |||
| No revascularization indicated (i.e. no significant coronary artery disease) | |||
| 4. Single episode early VF or polymorphic VT | R (2) | R (3) | M (4) |
| 5. Recurrent early VF or polymorphic VT | R (2) | R (3) | M (5) |
| Obstructive coronary artery disease with coronary anatomy not amenable to revascularization | |||
| 6. Early VF or polymorphic VT | M (5) | M (5) | A (7) |
| No electrophysiological study done | |||
Adapted, with permission, from J Am Coll Cardiol.11
A, Appropriate; AUC, Appropriate Use Criteria; ICD, implantable cardioverter-defibrillator; LVEF, left ventricular ejection fraction; M, Maybe appropriate; MI, myocardial infarction; R, Rarely appropriate; VF, ventricular fibrillation; VT, ventricular tachycardia.
To determine the clinical characteristics of patients with early VA after MI and to compare the outcomes in ICD recipients and non-recipients, we conducted a retrospective cohort study of early VA survivors at our institution. Additionally, we aimed to assess how closely ICD implantation practices at our institution have coincided with recently published AUC.
Methods
Patient population
All patients referred to the Mayo Clinic in Rochester, Minnesota for management of acute MI between January 2002 and January 2012 were screened for inclusion in this study. Initial screening of electronic medical records identified all patients with VA (using the search terms ‘VF’, ‘ventricle’ or ‘ventricular fibrillation’, ‘VT’, ‘ventricle’ or ‘ventricular tachycardia’, ‘ventricle’ or ‘ventricular arrest’, and ‘ventricle’ or ‘ventricular flutter’) documented in electronic clinical notes (including admission notes, progress notes, and discharge summaries) during the same hospitalization as acute MI as identified by International Classification of Diseases (ninth revision) codes 410 or 411.1. Of note, the search term ‘cardiac arrest’ was not used with initial screening since we were not including those with cardiac arrest due to asystole or pulseless electrical activity. A trained physician then reviewed all patient charts to confirm the presence of early VA in the same hospitalization for acute MI. Eligible individuals experienced one or more episodes of early VA (duration >30 s) after MI and survived until hospital discharge. For this study, myocardial infarction included ST-elevation and non-ST-elevation MI, and was diagnosed in the presence of two of the following three criteria: (i) typical chest pain lasting at least 20 min; (ii) elevation of creatine kinase (the MB fraction) at least two times normal or Troponin T level exceeding 0.01 ng/mL (99th percentile cutoff) in at least one blood sample, when drawn serially in the setting of strong clinical suspicion and (iii) a new Q-wave on electrocardiogram. Patients with stable, low-grade biomarker elevation after VA without evidence of a culprit lesion on angiography were excluded, since biomarker elevation was attributed primarily to the VA itself and cardiac resuscitation rather than secondary to acute MI. Ventricular arrhythmia included ventricular fibrillation (VF) as well as sustained monomorphic or polymorphic VT with rate over 100 beats per minute, lasting at least 30 s or requiring treatment with cardioversion/defibrillation due to haemodynamic instability. Early VA was defined as VA occurring within 48 h of MI symptom onset.
Patients who developed early VA in the emergency room, prior to transport to Mayo Clinic, or during transport were included if the event was witnessed by a healthcare professional and the rhythm was clearly documented. Individuals experiencing VA after MI not attributable to classical acute coronary artery thrombosis (e.g. reversible electrolyte imbalance, demand ischaemia, iatrogenic MI, spontaneous coronary artery dissection, stress cardiomyopathy, vasospasm, etc.) and those with prior ICD implantation or who would have qualified for ICD for primary/secondary prevention (i.e. pre-MI EF < 35% or history of VA) were excluded from analysis. As one of the primary aims of this study was to examine the long-term outcomes in individuals experiencing early VA after MI, we excluded individuals who died during hospitalization or did not have at least one follow-up visit recorded.
Prior to study initiation, ethical approval was granted by Mayo Clinic's Institutional Review Board.
Variables of interest
Electronic medical records were reviewed for all eligible individuals. Presenting clinical characteristics, prior medical history and baseline demographic data were abstracted. Myocardial infarction clinical features, including LVEF and coronary artery involvement, were abstracted from echocardiogram and coronary angiogram reports. The following data was also collected from our institution's PCI database: coronary vessel intervened upon, method of revascularization, severity of coronary artery disease (single or multiple vessel involvement), door-to-balloon time, stents placed, PCI success (defined as <20% residual stenosis in at least one treated lesion with no in-hospital death, ST-elevation MI, or coronary artery bypass graft surgery, CABG), in-hospital medication administration, presentation with cardiogenic shock, and intra-aortic balloon pump use. Eligibility for ICD implementation was post-coded as ‘appropriate’, ‘maybe appropriate’, or ‘rarely appropriate’ based on the 2013 AUC recommendations.
Discharge records were examined for each patient and medication prescriptions were recorded. Data regarding cardiac health was abstracted from each subsequent clinical follow-up visit. For individuals with an ICD, all ICD therapies including shocks or anti-tachycardia pacing were identified by reviewing clinical notes, ICD reports, and ICD tracings, where available. Mortality was obtained from clinical records and confirmed for all individuals using the National Social Security Registry. When available, the cause of death was collected from death certificates and autopsy reports.
Statistical analysis
Statistical analysis was performed using SAS version 9.3 (SAS Institute). For continuous measures, the results are presented as mean (SD) or median (IQR) for skewed distributions. Categorical measures are presented as percentages. To compare individuals who received ICDs and those that did not, one-way analysis of variance and χ2 tests were used, as appropriate. Logistic regression was then used to create a model to predict ICD implantation in this study population.
Kaplan–Meier survival curves were created and pairwise differences between those that received ICDs and those that did not were tested using the log-rank test. Univariate Cox proportional hazards models were used to assess potential predictors of survival. All risk factors with a P-value <0.05 were tested as potential predictors of death in a stepwise process to derive the final multivariate model. Relative risks are expressed as hazard ratios (HRs) with 95% confidence intervals. For all comparisons, a P-value of <0.05 was considered statistically significant.
Results
Between January 2002 and January 2012, 925 patients with International Classification of Diseases (ninth revision) codes for acute MI had ‘VA’ documented in clinical charts during the same hospitalization at Mayo Clinic, Rochester. Of these, 128 (73% male) patients experienced VA within 48 h of MI symptom onset, met the inclusion/exclusion criteria, and were included in this study. Figure 1 diagrams how the patients were selected. Of 128 study patient cohort, 92 (71.9%) presented initially with MI symptoms to outside facilities and were transferred to our tertiary care referral center for further management, while the remaining 36 (28.1%) were admitted to Saint Mary's Hospital at Mayo Clinic, Rochester, through the Emergency Department.
Selection of patients in the final study cohort.
Baseline characteristics
Of 128 early VA survivors, 26 (20%) received an ICD prior to discharge. As shown in Table 2, ICD recipients were more likely to have a history of prior MI, CABG, and PCI than those not treated with ICD (P = 0.001, P = 0.008, and P < 0.001, respectively). Patients receiving an ICD were more likely to have been taking thienopyridines (P = 0.003) and beta-blockers (P = 0.01) at the time of initial hospitalization than those not receiving an ICD. Baseline echocardiograms and/or nuclear scans prior to hospitalization for the index MI were available for 48 patients (38%). There were no differences in LVEF and LV chamber size between those who received or not received ICDs (all P > 0.05).
Baseline demographic and clinical characteristics of patients with early VA
| ICD (n = 26) | No ICD (n = 102) | P-value | |
|---|---|---|---|
| Gender, male | 20 (77%) | 73 (72%) | 0.58 |
| Age, mean (SD) | 64.4 (11.5) | 63.4 (13.5) | 0.72 |
| Body mass index, kg/m2 | 31.0 ± 6.5 | 30.2 ± 6.2 | 0.60 |
| Smoker, yes | 17 (65%) | 73 (72%) | 0.54 |
| Pre-existing diabetes, yes | 8 (31%) | 26 (25%) | 0.59 |
| Pre-existing hypertension yes | 21 (81%) | 75 (74%) | 0.45 |
| Pre-existing hyperlipidaemia, yes | 19 (73%) | 76 (75%) | 0.88 |
| Previous MI, yes | 12 (46%) | 17 (17%) | 0.001 |
| Previous PCI, yes | 9 (35%) | 13 (13%) | 0.008 |
| Previous CABG, yes | 9 (35%) | 7 (7%) | <0.001 |
| Echocardiography before MI (n = 48) | |||
| LVEF, % ± SD | 56.1 ± 12.8 | 59.5 ± 10.4 | 0.33 |
| LV end-diastolic diameter, mm ± SD | 52.2 ± 8.0 | 50.2 ± 5.5 | 0.4 |
| LV end-systolic diameter, mm ± SD | 36.3 ± 9.9 | 32.4 ± 5.6 | 0.17 |
| Pre-hospitalization medication use (n = 124) | |||
| Aspirin | 15 (58%) | 42 (43%) | 0.18 |
| Thienopyridine | 5 (19%) | 3 (3%) | 0.003 |
| Statin | 9 (35%) | 42 (43%) | 0.45 |
| ACE-inhibitor/ARB | 13 (50%) | 33 (34%) | 0.13 |
| Beta-blocker | 16 (62%) | 33 (34%) | 0.01 |
| Antiarrhythmic | 0 (0%) | 0 (0%) | 1 |
| Digoxin | 1 (4%) | 2 (2%) | 0.59 |
| ICD (n = 26) | No ICD (n = 102) | P-value | |
|---|---|---|---|
| Gender, male | 20 (77%) | 73 (72%) | 0.58 |
| Age, mean (SD) | 64.4 (11.5) | 63.4 (13.5) | 0.72 |
| Body mass index, kg/m2 | 31.0 ± 6.5 | 30.2 ± 6.2 | 0.60 |
| Smoker, yes | 17 (65%) | 73 (72%) | 0.54 |
| Pre-existing diabetes, yes | 8 (31%) | 26 (25%) | 0.59 |
| Pre-existing hypertension yes | 21 (81%) | 75 (74%) | 0.45 |
| Pre-existing hyperlipidaemia, yes | 19 (73%) | 76 (75%) | 0.88 |
| Previous MI, yes | 12 (46%) | 17 (17%) | 0.001 |
| Previous PCI, yes | 9 (35%) | 13 (13%) | 0.008 |
| Previous CABG, yes | 9 (35%) | 7 (7%) | <0.001 |
| Echocardiography before MI (n = 48) | |||
| LVEF, % ± SD | 56.1 ± 12.8 | 59.5 ± 10.4 | 0.33 |
| LV end-diastolic diameter, mm ± SD | 52.2 ± 8.0 | 50.2 ± 5.5 | 0.4 |
| LV end-systolic diameter, mm ± SD | 36.3 ± 9.9 | 32.4 ± 5.6 | 0.17 |
| Pre-hospitalization medication use (n = 124) | |||
| Aspirin | 15 (58%) | 42 (43%) | 0.18 |
| Thienopyridine | 5 (19%) | 3 (3%) | 0.003 |
| Statin | 9 (35%) | 42 (43%) | 0.45 |
| ACE-inhibitor/ARB | 13 (50%) | 33 (34%) | 0.13 |
| Beta-blocker | 16 (62%) | 33 (34%) | 0.01 |
| Antiarrhythmic | 0 (0%) | 0 (0%) | 1 |
| Digoxin | 1 (4%) | 2 (2%) | 0.59 |
ACE, angiotensinogen-converting enzyme; ARB, angiotensin-receptor blocker; CABG, coronary artery bypass graft; ICD, implantable cardioverter-defibrillator; LV, left ventricular; MI, myocardial infarction; PCI, percutaneous coronary intervention.
Bold P-values are statistically significant (P < 0.05).
Baseline demographic and clinical characteristics of patients with early VA
| ICD (n = 26) | No ICD (n = 102) | P-value | |
|---|---|---|---|
| Gender, male | 20 (77%) | 73 (72%) | 0.58 |
| Age, mean (SD) | 64.4 (11.5) | 63.4 (13.5) | 0.72 |
| Body mass index, kg/m2 | 31.0 ± 6.5 | 30.2 ± 6.2 | 0.60 |
| Smoker, yes | 17 (65%) | 73 (72%) | 0.54 |
| Pre-existing diabetes, yes | 8 (31%) | 26 (25%) | 0.59 |
| Pre-existing hypertension yes | 21 (81%) | 75 (74%) | 0.45 |
| Pre-existing hyperlipidaemia, yes | 19 (73%) | 76 (75%) | 0.88 |
| Previous MI, yes | 12 (46%) | 17 (17%) | 0.001 |
| Previous PCI, yes | 9 (35%) | 13 (13%) | 0.008 |
| Previous CABG, yes | 9 (35%) | 7 (7%) | <0.001 |
| Echocardiography before MI (n = 48) | |||
| LVEF, % ± SD | 56.1 ± 12.8 | 59.5 ± 10.4 | 0.33 |
| LV end-diastolic diameter, mm ± SD | 52.2 ± 8.0 | 50.2 ± 5.5 | 0.4 |
| LV end-systolic diameter, mm ± SD | 36.3 ± 9.9 | 32.4 ± 5.6 | 0.17 |
| Pre-hospitalization medication use (n = 124) | |||
| Aspirin | 15 (58%) | 42 (43%) | 0.18 |
| Thienopyridine | 5 (19%) | 3 (3%) | 0.003 |
| Statin | 9 (35%) | 42 (43%) | 0.45 |
| ACE-inhibitor/ARB | 13 (50%) | 33 (34%) | 0.13 |
| Beta-blocker | 16 (62%) | 33 (34%) | 0.01 |
| Antiarrhythmic | 0 (0%) | 0 (0%) | 1 |
| Digoxin | 1 (4%) | 2 (2%) | 0.59 |
| ICD (n = 26) | No ICD (n = 102) | P-value | |
|---|---|---|---|
| Gender, male | 20 (77%) | 73 (72%) | 0.58 |
| Age, mean (SD) | 64.4 (11.5) | 63.4 (13.5) | 0.72 |
| Body mass index, kg/m2 | 31.0 ± 6.5 | 30.2 ± 6.2 | 0.60 |
| Smoker, yes | 17 (65%) | 73 (72%) | 0.54 |
| Pre-existing diabetes, yes | 8 (31%) | 26 (25%) | 0.59 |
| Pre-existing hypertension yes | 21 (81%) | 75 (74%) | 0.45 |
| Pre-existing hyperlipidaemia, yes | 19 (73%) | 76 (75%) | 0.88 |
| Previous MI, yes | 12 (46%) | 17 (17%) | 0.001 |
| Previous PCI, yes | 9 (35%) | 13 (13%) | 0.008 |
| Previous CABG, yes | 9 (35%) | 7 (7%) | <0.001 |
| Echocardiography before MI (n = 48) | |||
| LVEF, % ± SD | 56.1 ± 12.8 | 59.5 ± 10.4 | 0.33 |
| LV end-diastolic diameter, mm ± SD | 52.2 ± 8.0 | 50.2 ± 5.5 | 0.4 |
| LV end-systolic diameter, mm ± SD | 36.3 ± 9.9 | 32.4 ± 5.6 | 0.17 |
| Pre-hospitalization medication use (n = 124) | |||
| Aspirin | 15 (58%) | 42 (43%) | 0.18 |
| Thienopyridine | 5 (19%) | 3 (3%) | 0.003 |
| Statin | 9 (35%) | 42 (43%) | 0.45 |
| ACE-inhibitor/ARB | 13 (50%) | 33 (34%) | 0.13 |
| Beta-blocker | 16 (62%) | 33 (34%) | 0.01 |
| Antiarrhythmic | 0 (0%) | 0 (0%) | 1 |
| Digoxin | 1 (4%) | 2 (2%) | 0.59 |
ACE, angiotensinogen-converting enzyme; ARB, angiotensin-receptor blocker; CABG, coronary artery bypass graft; ICD, implantable cardioverter-defibrillator; LV, left ventricular; MI, myocardial infarction; PCI, percutaneous coronary intervention.
Bold P-values are statistically significant (P < 0.05).
Clinical presentation and use of implantable cardioverter-defibrillator
Table 3 demonstrates presenting clinical characteristics among those treated with or not treated with ICD. Eighty-five (66%) patients initially presented with ST-elevation MI, while the remainder had non-ST-elevation MI. Implantable cardioverter-defibrillator recipients were more likely to have non-ST-elevation MI (65 vs. 24%, P < 0.001). Ventricular arrhythmia type was more likely to be sustained monomorphic VT and less likely to be VF in the ICD group than in the non-ICD group (65 vs. 21%, P < 0.001 and 19 vs. 70%, P < 0.001, respectively). Left ventricular ejection fraction at MI presentation was significantly lower in the ICD group (35 vs. 45%, P = 0.003). Revascularization rates with PCI and CABG were similar in ICD recipients and non-recipients (73 vs. 84%, P = 0.18 and 12 vs. 10%, P = 0.85, respectively). There was a trend towards a decrease in ICD utilization among early VA survivors after January 2005, as 10 of 33 (30.3%) patients admitted between January 2002 and December 2004 received ICD compared with 16 of 95 (16.8%) of those admitted between January 2005 and January 2012 (P = 0.10).
Characteristics at MI presentation of patients with early VA
| ICD (n = 26) | No ICD (n = 102) | P–value | |
|---|---|---|---|
| ST-elevation MI | 9 (35%) | 78 (76%) | <0.001 |
| Cardiogenic shocka | 4 (44%) | 20 (32%) | 0.47 |
| Intraaortic balloon pumpa | 3 (33%) | 6 (10%) | 0.046 |
| Multivessel diseasea | 6 (67%) | 26 (43%) | 0.19 |
| Left main >70% diseasea | 0 (0%) | 1 (2%) | 0.74 |
| LAD >70% diseasea | 5 (56%) | 38 (61%) | 0.6 |
| Circumflex >70% diseasea | 5 (56%) | 24 (39%) | 0.08 |
| RCA >70% diseasea | 2 (22%) | 29 (47%) | 0.14 |
| Early VA type | |||
| Monomorphic VT | 17 (65%) | 21 (21%) | <0.001 |
| Polymorphic VT | 5 (19%) | 12 (12%) | 0.32 |
| VF | 5 (19%) | 71 (70%) | <0.001 |
| Echocardiogram at presentation | |||
| LV EF ± SD | 35.2 ± 15.6 | 45.3 ± 14.9 | 0.003 |
| LV end-diastolic diameter, mm ± SD | 57.6 ± 8.2 | 50.1 ± 6.2 | <0.001 |
| LV end-systolic diameter, mm ± SD | 47.0 ± 12.1 | 35.9 ± 6.7 | <0.001 |
| Revascularization | |||
| Fibrinolytics | 2 (8%) | 26 (25%) | 0.05 |
| CABG | 3 (12%) | 10 (10%) | 0.85 |
| PCI | 19 (73%) | 86 (84%) | 0.18 |
| Successful PCIa,b | 8 (89%) | 52 (84%) | 0.71 |
| Door-to-balloon time, mina | 133 ± 130 | 96 ± 119 | 0.51 |
| PCI native left main)a | 0 (0) | 1 (2) | 0.71 |
| PCI native LADa | 2 (25) | 23 (40) | 0.42 |
| PCI native circumflexa | 3 (38) | 15 (26) | 0.49 |
| PCI native RCAa | 3 (38) | 19 (33) | 0.79 |
| PCI bypass grafta | 1 (13) | 2 (3) | 0.25 |
| Number of stents placed (mean)a | 1.50 ± 0.76 | 1.17 ± 0.73 | 0.24 |
| Medications at hospital discharge | |||
| Aspirin | 25 (96%) | 100 (98%) | 0.57 |
| Thienopyridine | 20 (77%) | 83 (81%) | 0.61 |
| Statin | 22 (85%) | 97 (95%) | 0.06 |
| ACE-inhibitor or ARB | 19 (73%) | 54 (53%) | 0.06 |
| Beta-blocker | 25 (96%) | 101 (99%) | 0.29 |
| Antiarrhythmic | 12 (46%) | 12 (12%) | <0.001 |
| Digoxin | 4 (15%) | 7 (7%) | 0.17 |
| ICD (n = 26) | No ICD (n = 102) | P–value | |
|---|---|---|---|
| ST-elevation MI | 9 (35%) | 78 (76%) | <0.001 |
| Cardiogenic shocka | 4 (44%) | 20 (32%) | 0.47 |
| Intraaortic balloon pumpa | 3 (33%) | 6 (10%) | 0.046 |
| Multivessel diseasea | 6 (67%) | 26 (43%) | 0.19 |
| Left main >70% diseasea | 0 (0%) | 1 (2%) | 0.74 |
| LAD >70% diseasea | 5 (56%) | 38 (61%) | 0.6 |
| Circumflex >70% diseasea | 5 (56%) | 24 (39%) | 0.08 |
| RCA >70% diseasea | 2 (22%) | 29 (47%) | 0.14 |
| Early VA type | |||
| Monomorphic VT | 17 (65%) | 21 (21%) | <0.001 |
| Polymorphic VT | 5 (19%) | 12 (12%) | 0.32 |
| VF | 5 (19%) | 71 (70%) | <0.001 |
| Echocardiogram at presentation | |||
| LV EF ± SD | 35.2 ± 15.6 | 45.3 ± 14.9 | 0.003 |
| LV end-diastolic diameter, mm ± SD | 57.6 ± 8.2 | 50.1 ± 6.2 | <0.001 |
| LV end-systolic diameter, mm ± SD | 47.0 ± 12.1 | 35.9 ± 6.7 | <0.001 |
| Revascularization | |||
| Fibrinolytics | 2 (8%) | 26 (25%) | 0.05 |
| CABG | 3 (12%) | 10 (10%) | 0.85 |
| PCI | 19 (73%) | 86 (84%) | 0.18 |
| Successful PCIa,b | 8 (89%) | 52 (84%) | 0.71 |
| Door-to-balloon time, mina | 133 ± 130 | 96 ± 119 | 0.51 |
| PCI native left main)a | 0 (0) | 1 (2) | 0.71 |
| PCI native LADa | 2 (25) | 23 (40) | 0.42 |
| PCI native circumflexa | 3 (38) | 15 (26) | 0.49 |
| PCI native RCAa | 3 (38) | 19 (33) | 0.79 |
| PCI bypass grafta | 1 (13) | 2 (3) | 0.25 |
| Number of stents placed (mean)a | 1.50 ± 0.76 | 1.17 ± 0.73 | 0.24 |
| Medications at hospital discharge | |||
| Aspirin | 25 (96%) | 100 (98%) | 0.57 |
| Thienopyridine | 20 (77%) | 83 (81%) | 0.61 |
| Statin | 22 (85%) | 97 (95%) | 0.06 |
| ACE-inhibitor or ARB | 19 (73%) | 54 (53%) | 0.06 |
| Beta-blocker | 25 (96%) | 101 (99%) | 0.29 |
| Antiarrhythmic | 12 (46%) | 12 (12%) | <0.001 |
| Digoxin | 4 (15%) | 7 (7%) | 0.17 |
ACE, angiotensinogen-converting enzyme; ARB, angiotensin-receptor blocker; CABG, coronary artery bypass graft; ICD, implantable cardioverter-defibrillator; LAD, left anterior descending; LV, left ventricular; MI, myocardial infarction; PCI, percutaneous coronary intervention; RCA, right coronary artery; VF, ventricular fibrillation; VT ventricular tachycardia.
Bold P-values are statistically significant (P < 0.05).
aData obtained from the 93 patients whose records were available in the Mayo Clinic PCI database (non-ICD cohort: n = 84; ICD cohort: n = 9).
bSuccessful PCI defined as <20% residual stenosis in at least one treated lesion with no in-hospital death, ST-elevation MI, or CABG.
Characteristics at MI presentation of patients with early VA
| ICD (n = 26) | No ICD (n = 102) | P–value | |
|---|---|---|---|
| ST-elevation MI | 9 (35%) | 78 (76%) | <0.001 |
| Cardiogenic shocka | 4 (44%) | 20 (32%) | 0.47 |
| Intraaortic balloon pumpa | 3 (33%) | 6 (10%) | 0.046 |
| Multivessel diseasea | 6 (67%) | 26 (43%) | 0.19 |
| Left main >70% diseasea | 0 (0%) | 1 (2%) | 0.74 |
| LAD >70% diseasea | 5 (56%) | 38 (61%) | 0.6 |
| Circumflex >70% diseasea | 5 (56%) | 24 (39%) | 0.08 |
| RCA >70% diseasea | 2 (22%) | 29 (47%) | 0.14 |
| Early VA type | |||
| Monomorphic VT | 17 (65%) | 21 (21%) | <0.001 |
| Polymorphic VT | 5 (19%) | 12 (12%) | 0.32 |
| VF | 5 (19%) | 71 (70%) | <0.001 |
| Echocardiogram at presentation | |||
| LV EF ± SD | 35.2 ± 15.6 | 45.3 ± 14.9 | 0.003 |
| LV end-diastolic diameter, mm ± SD | 57.6 ± 8.2 | 50.1 ± 6.2 | <0.001 |
| LV end-systolic diameter, mm ± SD | 47.0 ± 12.1 | 35.9 ± 6.7 | <0.001 |
| Revascularization | |||
| Fibrinolytics | 2 (8%) | 26 (25%) | 0.05 |
| CABG | 3 (12%) | 10 (10%) | 0.85 |
| PCI | 19 (73%) | 86 (84%) | 0.18 |
| Successful PCIa,b | 8 (89%) | 52 (84%) | 0.71 |
| Door-to-balloon time, mina | 133 ± 130 | 96 ± 119 | 0.51 |
| PCI native left main)a | 0 (0) | 1 (2) | 0.71 |
| PCI native LADa | 2 (25) | 23 (40) | 0.42 |
| PCI native circumflexa | 3 (38) | 15 (26) | 0.49 |
| PCI native RCAa | 3 (38) | 19 (33) | 0.79 |
| PCI bypass grafta | 1 (13) | 2 (3) | 0.25 |
| Number of stents placed (mean)a | 1.50 ± 0.76 | 1.17 ± 0.73 | 0.24 |
| Medications at hospital discharge | |||
| Aspirin | 25 (96%) | 100 (98%) | 0.57 |
| Thienopyridine | 20 (77%) | 83 (81%) | 0.61 |
| Statin | 22 (85%) | 97 (95%) | 0.06 |
| ACE-inhibitor or ARB | 19 (73%) | 54 (53%) | 0.06 |
| Beta-blocker | 25 (96%) | 101 (99%) | 0.29 |
| Antiarrhythmic | 12 (46%) | 12 (12%) | <0.001 |
| Digoxin | 4 (15%) | 7 (7%) | 0.17 |
| ICD (n = 26) | No ICD (n = 102) | P–value | |
|---|---|---|---|
| ST-elevation MI | 9 (35%) | 78 (76%) | <0.001 |
| Cardiogenic shocka | 4 (44%) | 20 (32%) | 0.47 |
| Intraaortic balloon pumpa | 3 (33%) | 6 (10%) | 0.046 |
| Multivessel diseasea | 6 (67%) | 26 (43%) | 0.19 |
| Left main >70% diseasea | 0 (0%) | 1 (2%) | 0.74 |
| LAD >70% diseasea | 5 (56%) | 38 (61%) | 0.6 |
| Circumflex >70% diseasea | 5 (56%) | 24 (39%) | 0.08 |
| RCA >70% diseasea | 2 (22%) | 29 (47%) | 0.14 |
| Early VA type | |||
| Monomorphic VT | 17 (65%) | 21 (21%) | <0.001 |
| Polymorphic VT | 5 (19%) | 12 (12%) | 0.32 |
| VF | 5 (19%) | 71 (70%) | <0.001 |
| Echocardiogram at presentation | |||
| LV EF ± SD | 35.2 ± 15.6 | 45.3 ± 14.9 | 0.003 |
| LV end-diastolic diameter, mm ± SD | 57.6 ± 8.2 | 50.1 ± 6.2 | <0.001 |
| LV end-systolic diameter, mm ± SD | 47.0 ± 12.1 | 35.9 ± 6.7 | <0.001 |
| Revascularization | |||
| Fibrinolytics | 2 (8%) | 26 (25%) | 0.05 |
| CABG | 3 (12%) | 10 (10%) | 0.85 |
| PCI | 19 (73%) | 86 (84%) | 0.18 |
| Successful PCIa,b | 8 (89%) | 52 (84%) | 0.71 |
| Door-to-balloon time, mina | 133 ± 130 | 96 ± 119 | 0.51 |
| PCI native left main)a | 0 (0) | 1 (2) | 0.71 |
| PCI native LADa | 2 (25) | 23 (40) | 0.42 |
| PCI native circumflexa | 3 (38) | 15 (26) | 0.49 |
| PCI native RCAa | 3 (38) | 19 (33) | 0.79 |
| PCI bypass grafta | 1 (13) | 2 (3) | 0.25 |
| Number of stents placed (mean)a | 1.50 ± 0.76 | 1.17 ± 0.73 | 0.24 |
| Medications at hospital discharge | |||
| Aspirin | 25 (96%) | 100 (98%) | 0.57 |
| Thienopyridine | 20 (77%) | 83 (81%) | 0.61 |
| Statin | 22 (85%) | 97 (95%) | 0.06 |
| ACE-inhibitor or ARB | 19 (73%) | 54 (53%) | 0.06 |
| Beta-blocker | 25 (96%) | 101 (99%) | 0.29 |
| Antiarrhythmic | 12 (46%) | 12 (12%) | <0.001 |
| Digoxin | 4 (15%) | 7 (7%) | 0.17 |
ACE, angiotensinogen-converting enzyme; ARB, angiotensin-receptor blocker; CABG, coronary artery bypass graft; ICD, implantable cardioverter-defibrillator; LAD, left anterior descending; LV, left ventricular; MI, myocardial infarction; PCI, percutaneous coronary intervention; RCA, right coronary artery; VF, ventricular fibrillation; VT ventricular tachycardia.
Bold P-values are statistically significant (P < 0.05).
aData obtained from the 93 patients whose records were available in the Mayo Clinic PCI database (non-ICD cohort: n = 84; ICD cohort: n = 9).
bSuccessful PCI defined as <20% residual stenosis in at least one treated lesion with no in-hospital death, ST-elevation MI, or CABG.
According to the AUC, ICD implantation was considered appropriate care for three ICD recipients (no revascularization and LVEF <35%); was maybe appropriate for 17 (16 revascularization success and LVEF <35%; 1 no revascularization and LVEF >35%) and was rarely appropriate for 6 (revascularization success and LVEF >35%) (Table 4). Among those not receiving an ICD, ICD implantation would have been considered appropriate for four based on revascularization status and LVEF (no revascularization and LVEF <35%). Two of these four patients died (after 91 and 122 days), both due to congestive heart failure in hospice care. Thus, both patients likely would not have qualified for ICD since the expected survival with good functional status was less than 1 year. Additionally, ICD implantation may have been appropriate for 30 (25 revascularization success and LVEF <35%; 5 no revascularization and LVEF >35%) and would have been rarely appropriate for 68 (revascularization success and LVEF >35%).
Appropriateness of ICD implantation per 2013 guidelines
| ICD group (n = 26) | Non-ICD group (n = 102) | |
|---|---|---|
| Appropriate, n (%) | 3 (11.5) | 4 (3.9) |
| Experienced appropriate ICD therapy, n (%) | 0 (0) | – |
| Maybe appropriate, n (%) | 17 (65.4) | 30 (29.4) |
| Experienced appropriate ICD therapy, n (%) | 10 (58.8) | – |
| Rarely appropriate, n (%) | 6 (23.1) | 68 (66.7) |
| Experienced appropriate ICD therapy, n (%) | 2 (33.3) | – |
| ICD group (n = 26) | Non-ICD group (n = 102) | |
|---|---|---|
| Appropriate, n (%) | 3 (11.5) | 4 (3.9) |
| Experienced appropriate ICD therapy, n (%) | 0 (0) | – |
| Maybe appropriate, n (%) | 17 (65.4) | 30 (29.4) |
| Experienced appropriate ICD therapy, n (%) | 10 (58.8) | – |
| Rarely appropriate, n (%) | 6 (23.1) | 68 (66.7) |
| Experienced appropriate ICD therapy, n (%) | 2 (33.3) | – |
ICD, implantable defibrillator-cardioverter.
Appropriateness of ICD implantation per 2013 guidelines
| ICD group (n = 26) | Non-ICD group (n = 102) | |
|---|---|---|
| Appropriate, n (%) | 3 (11.5) | 4 (3.9) |
| Experienced appropriate ICD therapy, n (%) | 0 (0) | – |
| Maybe appropriate, n (%) | 17 (65.4) | 30 (29.4) |
| Experienced appropriate ICD therapy, n (%) | 10 (58.8) | – |
| Rarely appropriate, n (%) | 6 (23.1) | 68 (66.7) |
| Experienced appropriate ICD therapy, n (%) | 2 (33.3) | – |
| ICD group (n = 26) | Non-ICD group (n = 102) | |
|---|---|---|
| Appropriate, n (%) | 3 (11.5) | 4 (3.9) |
| Experienced appropriate ICD therapy, n (%) | 0 (0) | – |
| Maybe appropriate, n (%) | 17 (65.4) | 30 (29.4) |
| Experienced appropriate ICD therapy, n (%) | 10 (58.8) | – |
| Rarely appropriate, n (%) | 6 (23.1) | 68 (66.7) |
| Experienced appropriate ICD therapy, n (%) | 2 (33.3) | – |
ICD, implantable defibrillator-cardioverter.
Long-term outcomes
Follow-up data was available for all 128 patients with a median follow-up period of 878 days after discharge. During follow-up, 11 of 26 (42%) ICD recipients died (5 cardiac, 5 non-cardiac, 1 unknown) compared with 17 of 102 (17%) in the non-ICD group (6 cardiac, 7 non-cardiac, 4 unknown). There were four deaths due to recurrent VA in the ICD recipients, although three occurred after ICD deactivation in patients who had transitioned to comfort care measures. The Kaplan–Meier survival estimate showed a lower overall survival rate in ICD recipients (P = 0.02) (Figure 2). As shown in Table 5, older age (P < 0.0001), female gender (P = 0.004), and a prior history of MI (P = 0.01) or CABG (P = 0.02) were independently associated with increased mortality in the univariate analysis. Non-ST-elevation MI (P = 0.01) and sustained monomorphic VT (P = 0.047) were also associated with worse overall survival. Ventricular fibrillation as the presenting VA type (P = 0.001) and thienopyridine prescription at discharge (P = 0.017) were associated with improved survival. The multivariate analysis identified older age at the time of MI to predict higher mortality (HR = 1.09; P < 0.0001), whereas VF as the presenting VA type was associated with lower mortality (HR = 0.37; P = 0.001). Decreased LVEF was not associated with increased mortality (P = 0.86).
Univariate and multivariate analysis of predictors for mortality in survivors of early VA after MI
| Variables | HR | P-value |
|---|---|---|
| Age at MIa | 1.09 [1.05–1.13] | <0.0001 |
| Male gender | 0.33 [−0.69] | 0.004 |
| Body mass index | 1 [0.94–1.06] | 0.95 |
| Smoker | 0.81 [0.38–1.76] | 0.6 |
| Diabetes | 1.43 [0.65–3.14] | 0.38 |
| Hypertension | 1.46 [0.59–3.60] | 0.42 |
| Hyperlipidemia | 0.61 [0.28–1.35] | 0.22 |
| Prior history of MI | 2.72 [1.28–5.82] | 0.01 |
| Prior history of PCI | 1.73 [0.73–4.06] | 0.21 |
| Prior history of CABG | 2.83 [1.19–6.69] | 0.02 |
| MI presentation | ||
| ST-elevation MI | 0.39 [0.19–0.81] | 0.01 |
| Cardiogenic shock | 2.18 [0.55–8.73] | 0.27 |
| Intraaortic balloon pump | 0.83 [0.34–2.05] | 0.69 |
| Multivessel disease | 0.72 [0.16–3.22] | 0.66 |
| Left main >70% disease | 0 | 0.99 |
| LAD >70% disease | 0.21 [0.06–0.76] | 0.02 |
| Circumflex >70% disease | 0.7 [0.22–2.21] | 0.54 |
| RCA >70% disease | 0.54 [0.17–1.72] | 0.3 |
| LVEF | 1 [0.98–1.02] | 0.86 |
| LV end-diastolic diameter | 1.01 [0.96–1.06] | 0.66 |
| LV end-systolic diameter | 1.01 [0.97–1.05] | 0.53 |
| Monomorphic VT | 2.1 [1.01–4.38] | 0.047 |
| Polymorphic VT | 1.65 [0.62–4.39] | 0.32 |
| VFa | 0.37 [0.18–0.79] | 0.001 |
| Revascularization | ||
| Fibrinolytics | 0.4 [0.14–1.15] | 0.09 |
| CABG | 0.49 [0.12–2.07] | 0.33 |
| PCI | 0.57 [0.25–1.29] | 0.18 |
| Successful PCIb | 0.39 [0.05–3.18] | 0.38 |
| Door-to-balloon time | 1 [0.99–1.01] | 0.94 |
| PCI native left main | 7.3 [0.87–61.10] | 0.07 |
| PCI native LAD | 0.58 [0.11–2.98] | 0.51 |
| PCI native circumflex | 2.63 [0.59–11.79] | 0.21 |
| PCI native RCA | 0.82 [0.16–4.24] | 0.81 |
| PCI bypass graft | 3.34 [0.40–27.81] | 0.27 |
| Number of stents placed (mean) | 1.45 [0.51–4.10] | 0.48 |
| Medications at hospital discharge | ||
| Aspirin | 0.44 [0.10–1.85] | 0.26 |
| Thienopyridine | 0.4 [0.19–0.85] | 0.017 |
| Statin | 1.56 [0.36–6.74] | 0.55 |
| ACE-inhibitor or ARB | 1.42 [0.66–3.06] | 0.37 |
| Beta-blocker | 0.43 [0.06–3.19] | 0.41 |
| Antiarrhythmic | 2.13 [0.97–4.69] | 0.06 |
| Digoxin | 1.5 [0.52–4.32] | 0.45 |
| Variables | HR | P-value |
|---|---|---|
| Age at MIa | 1.09 [1.05–1.13] | <0.0001 |
| Male gender | 0.33 [−0.69] | 0.004 |
| Body mass index | 1 [0.94–1.06] | 0.95 |
| Smoker | 0.81 [0.38–1.76] | 0.6 |
| Diabetes | 1.43 [0.65–3.14] | 0.38 |
| Hypertension | 1.46 [0.59–3.60] | 0.42 |
| Hyperlipidemia | 0.61 [0.28–1.35] | 0.22 |
| Prior history of MI | 2.72 [1.28–5.82] | 0.01 |
| Prior history of PCI | 1.73 [0.73–4.06] | 0.21 |
| Prior history of CABG | 2.83 [1.19–6.69] | 0.02 |
| MI presentation | ||
| ST-elevation MI | 0.39 [0.19–0.81] | 0.01 |
| Cardiogenic shock | 2.18 [0.55–8.73] | 0.27 |
| Intraaortic balloon pump | 0.83 [0.34–2.05] | 0.69 |
| Multivessel disease | 0.72 [0.16–3.22] | 0.66 |
| Left main >70% disease | 0 | 0.99 |
| LAD >70% disease | 0.21 [0.06–0.76] | 0.02 |
| Circumflex >70% disease | 0.7 [0.22–2.21] | 0.54 |
| RCA >70% disease | 0.54 [0.17–1.72] | 0.3 |
| LVEF | 1 [0.98–1.02] | 0.86 |
| LV end-diastolic diameter | 1.01 [0.96–1.06] | 0.66 |
| LV end-systolic diameter | 1.01 [0.97–1.05] | 0.53 |
| Monomorphic VT | 2.1 [1.01–4.38] | 0.047 |
| Polymorphic VT | 1.65 [0.62–4.39] | 0.32 |
| VFa | 0.37 [0.18–0.79] | 0.001 |
| Revascularization | ||
| Fibrinolytics | 0.4 [0.14–1.15] | 0.09 |
| CABG | 0.49 [0.12–2.07] | 0.33 |
| PCI | 0.57 [0.25–1.29] | 0.18 |
| Successful PCIb | 0.39 [0.05–3.18] | 0.38 |
| Door-to-balloon time | 1 [0.99–1.01] | 0.94 |
| PCI native left main | 7.3 [0.87–61.10] | 0.07 |
| PCI native LAD | 0.58 [0.11–2.98] | 0.51 |
| PCI native circumflex | 2.63 [0.59–11.79] | 0.21 |
| PCI native RCA | 0.82 [0.16–4.24] | 0.81 |
| PCI bypass graft | 3.34 [0.40–27.81] | 0.27 |
| Number of stents placed (mean) | 1.45 [0.51–4.10] | 0.48 |
| Medications at hospital discharge | ||
| Aspirin | 0.44 [0.10–1.85] | 0.26 |
| Thienopyridine | 0.4 [0.19–0.85] | 0.017 |
| Statin | 1.56 [0.36–6.74] | 0.55 |
| ACE-inhibitor or ARB | 1.42 [0.66–3.06] | 0.37 |
| Beta-blocker | 0.43 [0.06–3.19] | 0.41 |
| Antiarrhythmic | 2.13 [0.97–4.69] | 0.06 |
| Digoxin | 1.5 [0.52–4.32] | 0.45 |
ACE, angiotensinogen-converting enzyme; ARB, angiotensin receptor blocker; CABG, coronary artery bypass graft surgery; LAD, left anterior descending coronary artery; LV, left ventricular; MI, myocardial infarction; PCI, percutaneous coronary intervention; RCA, right coronary artery; VT, ventricular tachycardia.
Bold P-values are statistically significant (P < 0.05).
aStepwise multivariate analysis run using covariates with P-value <0.02 identified age to be an independent mortality predictor, while VF was associated with lower mortality.
bSuccessful PCI defined as <20% residual stenosis in at least one treated lesion with no in-hospital death, ST-elevation MI, or CABG.
Univariate and multivariate analysis of predictors for mortality in survivors of early VA after MI
| Variables | HR | P-value |
|---|---|---|
| Age at MIa | 1.09 [1.05–1.13] | <0.0001 |
| Male gender | 0.33 [−0.69] | 0.004 |
| Body mass index | 1 [0.94–1.06] | 0.95 |
| Smoker | 0.81 [0.38–1.76] | 0.6 |
| Diabetes | 1.43 [0.65–3.14] | 0.38 |
| Hypertension | 1.46 [0.59–3.60] | 0.42 |
| Hyperlipidemia | 0.61 [0.28–1.35] | 0.22 |
| Prior history of MI | 2.72 [1.28–5.82] | 0.01 |
| Prior history of PCI | 1.73 [0.73–4.06] | 0.21 |
| Prior history of CABG | 2.83 [1.19–6.69] | 0.02 |
| MI presentation | ||
| ST-elevation MI | 0.39 [0.19–0.81] | 0.01 |
| Cardiogenic shock | 2.18 [0.55–8.73] | 0.27 |
| Intraaortic balloon pump | 0.83 [0.34–2.05] | 0.69 |
| Multivessel disease | 0.72 [0.16–3.22] | 0.66 |
| Left main >70% disease | 0 | 0.99 |
| LAD >70% disease | 0.21 [0.06–0.76] | 0.02 |
| Circumflex >70% disease | 0.7 [0.22–2.21] | 0.54 |
| RCA >70% disease | 0.54 [0.17–1.72] | 0.3 |
| LVEF | 1 [0.98–1.02] | 0.86 |
| LV end-diastolic diameter | 1.01 [0.96–1.06] | 0.66 |
| LV end-systolic diameter | 1.01 [0.97–1.05] | 0.53 |
| Monomorphic VT | 2.1 [1.01–4.38] | 0.047 |
| Polymorphic VT | 1.65 [0.62–4.39] | 0.32 |
| VFa | 0.37 [0.18–0.79] | 0.001 |
| Revascularization | ||
| Fibrinolytics | 0.4 [0.14–1.15] | 0.09 |
| CABG | 0.49 [0.12–2.07] | 0.33 |
| PCI | 0.57 [0.25–1.29] | 0.18 |
| Successful PCIb | 0.39 [0.05–3.18] | 0.38 |
| Door-to-balloon time | 1 [0.99–1.01] | 0.94 |
| PCI native left main | 7.3 [0.87–61.10] | 0.07 |
| PCI native LAD | 0.58 [0.11–2.98] | 0.51 |
| PCI native circumflex | 2.63 [0.59–11.79] | 0.21 |
| PCI native RCA | 0.82 [0.16–4.24] | 0.81 |
| PCI bypass graft | 3.34 [0.40–27.81] | 0.27 |
| Number of stents placed (mean) | 1.45 [0.51–4.10] | 0.48 |
| Medications at hospital discharge | ||
| Aspirin | 0.44 [0.10–1.85] | 0.26 |
| Thienopyridine | 0.4 [0.19–0.85] | 0.017 |
| Statin | 1.56 [0.36–6.74] | 0.55 |
| ACE-inhibitor or ARB | 1.42 [0.66–3.06] | 0.37 |
| Beta-blocker | 0.43 [0.06–3.19] | 0.41 |
| Antiarrhythmic | 2.13 [0.97–4.69] | 0.06 |
| Digoxin | 1.5 [0.52–4.32] | 0.45 |
| Variables | HR | P-value |
|---|---|---|
| Age at MIa | 1.09 [1.05–1.13] | <0.0001 |
| Male gender | 0.33 [−0.69] | 0.004 |
| Body mass index | 1 [0.94–1.06] | 0.95 |
| Smoker | 0.81 [0.38–1.76] | 0.6 |
| Diabetes | 1.43 [0.65–3.14] | 0.38 |
| Hypertension | 1.46 [0.59–3.60] | 0.42 |
| Hyperlipidemia | 0.61 [0.28–1.35] | 0.22 |
| Prior history of MI | 2.72 [1.28–5.82] | 0.01 |
| Prior history of PCI | 1.73 [0.73–4.06] | 0.21 |
| Prior history of CABG | 2.83 [1.19–6.69] | 0.02 |
| MI presentation | ||
| ST-elevation MI | 0.39 [0.19–0.81] | 0.01 |
| Cardiogenic shock | 2.18 [0.55–8.73] | 0.27 |
| Intraaortic balloon pump | 0.83 [0.34–2.05] | 0.69 |
| Multivessel disease | 0.72 [0.16–3.22] | 0.66 |
| Left main >70% disease | 0 | 0.99 |
| LAD >70% disease | 0.21 [0.06–0.76] | 0.02 |
| Circumflex >70% disease | 0.7 [0.22–2.21] | 0.54 |
| RCA >70% disease | 0.54 [0.17–1.72] | 0.3 |
| LVEF | 1 [0.98–1.02] | 0.86 |
| LV end-diastolic diameter | 1.01 [0.96–1.06] | 0.66 |
| LV end-systolic diameter | 1.01 [0.97–1.05] | 0.53 |
| Monomorphic VT | 2.1 [1.01–4.38] | 0.047 |
| Polymorphic VT | 1.65 [0.62–4.39] | 0.32 |
| VFa | 0.37 [0.18–0.79] | 0.001 |
| Revascularization | ||
| Fibrinolytics | 0.4 [0.14–1.15] | 0.09 |
| CABG | 0.49 [0.12–2.07] | 0.33 |
| PCI | 0.57 [0.25–1.29] | 0.18 |
| Successful PCIb | 0.39 [0.05–3.18] | 0.38 |
| Door-to-balloon time | 1 [0.99–1.01] | 0.94 |
| PCI native left main | 7.3 [0.87–61.10] | 0.07 |
| PCI native LAD | 0.58 [0.11–2.98] | 0.51 |
| PCI native circumflex | 2.63 [0.59–11.79] | 0.21 |
| PCI native RCA | 0.82 [0.16–4.24] | 0.81 |
| PCI bypass graft | 3.34 [0.40–27.81] | 0.27 |
| Number of stents placed (mean) | 1.45 [0.51–4.10] | 0.48 |
| Medications at hospital discharge | ||
| Aspirin | 0.44 [0.10–1.85] | 0.26 |
| Thienopyridine | 0.4 [0.19–0.85] | 0.017 |
| Statin | 1.56 [0.36–6.74] | 0.55 |
| ACE-inhibitor or ARB | 1.42 [0.66–3.06] | 0.37 |
| Beta-blocker | 0.43 [0.06–3.19] | 0.41 |
| Antiarrhythmic | 2.13 [0.97–4.69] | 0.06 |
| Digoxin | 1.5 [0.52–4.32] | 0.45 |
ACE, angiotensinogen-converting enzyme; ARB, angiotensin receptor blocker; CABG, coronary artery bypass graft surgery; LAD, left anterior descending coronary artery; LV, left ventricular; MI, myocardial infarction; PCI, percutaneous coronary intervention; RCA, right coronary artery; VT, ventricular tachycardia.
Bold P-values are statistically significant (P < 0.05).
aStepwise multivariate analysis run using covariates with P-value <0.02 identified age to be an independent mortality predictor, while VF was associated with lower mortality.
bSuccessful PCI defined as <20% residual stenosis in at least one treated lesion with no in-hospital death, ST-elevation MI, or CABG.
Kaplan–Meier survival curve depicting overall survival in ICD recipients vs. non-recipients.
During a mean follow-up of 3.1 ± 2.7 years, 12 of 26 ICD recipients (46%) experienced appropriate ICD therapy resulting in the termination of recurrent VA (10 of whom received ICD initially for sustained monomorphic VT). Among the 10 patients who presented initially with sustained monomorphic VT who experienced appropriate ICD therapies during follow-up, VA type was recurrent monomorphic VT in 8 and VF in 2. Six patients (23%) experienced inappropriate ICD therapies (1 antitachycardia pacing, 5 ICD shocks). Five-year incidence of ICD therapies is displayed in Figure 3. Of the 12 patients who received appropriate ICD therapy during follow-up, 6 were still alive after a median follow-up period of 5.4 years, while the 6 who died lived an average of 1.3 years after first appropriate ICD therapy (including 1 patient who died of cardiogenic shock the same day, shortly after appropriate ICD therapy terminated recurrent VA). As demonstrated in Table 4, in the 26 early VA survivors in whom ICD was implanted, 0 of 3 (0%) appropriate candidates, 10 of 17 (58.8%) maybe appropriate candidates and 2 of 6 (33%) rarely appropriate candidates had at least one episode of recurrent VA during follow-up which was effectively terminated by appropriate ICD shock or antitachycardia pacing therapy. Meanwhile, 3 of 6 (50%) ICD recipients who would have been considered rarely appropriate per AUC experienced at least one inappropriate ICD therapy.
Time from ICD implantation to initial appropriate (A) and inappropriate (B) therapy in the first 5 years following ICD implantation in early VA survivors.
Discussion
Risk stratification after early ventricular arrhythmia
To our knowledge, this is the first study examining the characteristics and outcomes of patients receiving ICDs for early VA after MI. Prior studies have suggested that early VA may portend worse prognosis in the first few weeks after discharge, but long-term survival after 60 days does not appear to be affected.12–15 In our study, the subset of early VA survivors treated with ICDs prior to hospital discharge had significantly higher rates of death at long-term follow-up despite similar revascularization with PCI and CABG. The trend towards decreased long-term mortality in patients who were treated with fibrinolytics (P = 0.09) and PCI (P = 0.18) likely did not reach statistical significance due to our small sample size. The lower rate of fibrinolytic use in ICD recipients reflects the predominance of non-ST-elevation MI in those patients, for which fibrinolytics are contraindicated. Patients receiving ICDs tended to be higher risk, with more extensive cardiac history (prior MI, PCI, or CABG) and worrisome presenting arrhythmia (sustained monomorphic VT as opposed to VF) and MI (non-ST-elevation vs. ST-elevation MI) characteristics, prompting ICD implantation per clinician discretion. The high rate of appropriate ICD shocks and antitachycardia pacing therapies in ICD recipients suggests that clinicians identified these patients as being at higher risk for recurrent VA, although their outcomes remained worse despite ICD implantation.
Although decreased LVEF at the time of MI is considered a risk factor for death by the 2013 AUC, decreased LVEF did not predict mortality in our small single-center cohort. Instead, we found that VF as early VA type was associated with a lower mortality. While VF is more likely to be a transient phenomenon during the peri-infarct period, early sustained monomorphic VT may imply the presence of an underlying irreversible substrate for recurrent VA and has been associated with lower in-hospital and 1-year survivals.16,17 Particularly in patients with prior MI, sustained monomorphic VT likely originates from fixed myocardial scar which may be more vulnerable in the setting of acute ischaemia. Of the 17 ICD recipients in our cohort whose initial VA type was sustained monomorphic VT, 10 had appropriate ICD therapy (8 for VT and 2 for VF) during follow-up, none of which occurred in the setting of recurrent ischaemia. The high rate of recurrent VA of the same type in survivors of early sustained monomorphic VT after MI suggests that these patients in particular may have irreversible substrate putting them at increased risk for recurrent VT and cardiac death, irregardless of LVEF or VA timing at MI presentation.
Implantable cardioverter-defibrillator implantation for early ventricular arrhythmia
Implantable cardioverter-defibrillators are intended to improve patient health outcomes by reducing sudden cardiac death. However, the decision to implant an ICD following early VA after acute MI remains highly subjective. Implantable cardioverter-defibrillator implantation is costly and device implant complications and inappropriate shocks were previously reported to occur in over 30% of patients,18 although optimization of device programming has recently been shown to minimize inappropriate therapies and improve survival.19 The Defibrillators in Acute Myocardial Infarction (DINAMIT)20 and Immediate Risk Stratification Improves Survival (IRIS)21 trials showed that prophylactic ICD after acute MI in high-risk patients resulted in no overall survival benefit despite a significant reduction in arrhythmic mortality. While both of these studies excluded patients with late VA (48 h after symptom onset), neither study specified whether the study patients developed early VA after acute MI. Thus, the impact of ICD therapy on long-term outcomes in early VA survivors remains to be examined. In our study, there was only one arrhythmic death among the ICD recipients which was not prevented by the ICD (since ICD had been deactivated in the other 3 arrhythmic deaths) while the majority of deaths were non-cardiac.
While the supporting evidence is weak, the 2013 AUC is the first consensus statement to provide guidance for the management of post-MI early VA, classifying patients into appropriate, maybe appropriate, or rarely appropriate categories mainly based on revascularization status and presenting LVEF (Table 1).11 Although the AUC advisory committee notes that the decision to implant an ICD ultimately remains subjective, the AUC recommendations do not specifically mention the impact of presenting MI characteristics, VA type, or past medical history on the utility of ICDs for patients experiencing early VA after MI. Clinicians for our study cohort were more likely to implant an ICD in early VA survivors with a low LVEF, as recommended by the 2013 AUC, and other than VF, more likely to implant an ICD in those presenting with monomorphic VT, concerning the existing ventricular substrate. Three quarters of the 26 ICD recipients in our study were considered to be appropriate or maybe appropriate ICD implant candidates per AUC guidelines. Of these, half had at least one episode of recurrent VA during follow-up which was effectively terminated by appropriate ICD therapy, supporting the AUC recommendations. There was a trend suggesting that ICDs were more freely implanted in the early years of patient enrollment (2002–05), coinciding with the release of The Multicenter Automated Defibrillator Implantation Trial II (MADIT-II) results in March 2002, which showed improved survival with prophylactic ICD implantation in patients with LV dysfunction and prior MI. It is likely that the decline in ICD utilization in subsequent years at our institution was due to more stringent National Coverage Determination reimbursement indications released January 2005 requiring a 40-day waiting period after acute MI (or 3 months after CABG or PCI) before reevaluation of LVEF for consideration of ICD for primary prevention of sudden death.22,23
Our findings suggest that when considering ICD implantation in patients with early VA, clinicians should take into account not only VA timing (early or late), revascularization status and LVEF, but also prior cardiac history, VA and/or MI type. Survivors of early sustained monomorphic VT are at high risk for VA recurrence and consideration for ICD therapy for secondary prevention of sudden death may be warranted. Meanwhile, those patients with lower risk features (younger age, no cardiac history, ST-elevation MI, and VF as the presenting VA) have a better long-term prognosis and ICD implantation prior to discharge may not be necessary.
Limitations
This study is a retrospective analysis and thus should be considered exploratory. As ICDs are uncommonly implanted prior to discharge in patients with early VA, our sample size (n = 26) of ICD recipients was small. Additionally, there are several limitations to this study inherent to the fact that our institution is a tertiary referral center. Many of the patients in our early VA cohort were transferred to our institution for acute care and some may have followed up elsewhere. Since some patients were revascularized elsewhere prior to transfer to our institution, we only had access to angiographic and periprocedural PCI data for 9/26 (34.6%) of the ICD recipients and 84/102 (82.4%) of the ICD non-recipients for analysis. Therefore, the group of ICD recipients may represent a higher-risk cohort since patients with low-risk clinical characteristics may have been managed locally, while those thought to require higher level of care were preferentially transferred. Although overall death and ICD therapies were accurately confirmed using the National Social Security Registry and National and Mayo Clinic ICD registries, cause of death may not have been known in those patients who died elsewhere.
Conclusions
While the 2013 AUC attempt to provide much-needed guidance on ICD implantation for early VA in the setting of acute MI, the evidence on which they are based is lacking. We identified additional risk factors for poor patient outcomes that are not included in the AUC. Patients with a history of MI who develop early sustained monomorphic VT after subsequent MI, especially non-ST-elevation MI, are at particularly high risk for death. While ICD therapy may decrease the risk for arrhythmic mortality in these patients, medical comorbidities and expected lifespan must be considered. Given the paucity of data on which the current AUC are based, further studies are necessary to better understand the prognostic and therapeutic implications of early VA after MI.
Acknowledgements
The authors thank Melissa A. Elafros for her assistance with reviewing and revising this manuscript.
Conflicts of interest: none declared.
