Implantable cardioverter-defibrillator (ICD) implantation outside practice guidelines remains contentious, particularly during the mandated waiting periods in patients with recent cardiac events. We assessed the prevalence and outcomes of non–guideline-based (NGB) ICD implantations in a tertiary academic medical center, with a specific focus on adjudication of arrhythmia events. All patients who underwent initial primary prevention ICD implantation at our institution from 2004 to 2012 were categorized as having received guideline-based (GB) or NGB implants and were retrospectively assessed for first episode of appropriate ICD therapy and total mortality. Of 807 patients, 137 (17.0%) received NGB implants. During a median follow-up of 2.9 years, patients with NGB implants had similar times to first appropriate ICD therapy (median time to event 1.94 vs 2.17 years in patients with GB implants, p = 0.20). After multivariable analysis, patients with NGB implants remained at higher risk for death (hazard ratio 1.54, 95% confidence interval 1.1 to 2.2, p = 0.03) but not appropriate ICD therapy (hazard ratio 0.83, 95% confidence interval 0.5 to 1.3, p = 0.51). Furthermore, only 1 of 125 patients who underwent implant within the 40-day waiting period after myocardial infarction or 3-month waiting period after revascularization or cardiomyopathy diagnosis received an appropriate therapy within this period. In conclusion, few patients received NGB ICD implants in our academic medical center. Although these patients have similar long-term risk of receiving appropriate ICD therapy compared with patients with GB implants, this risk is very low during the waiting periods mandated by clinical practice guidelines. These results suggest that there is little need to rush into implanting ICDs during these waiting periods.
Current practice guidelines for primary prevention implantable cardioverter-defibrillator (ICD) implantation mandate a 40-day period after a myocardial infarction (MI) before ICD implantation, recommend against ICD implantation for patients with New York Heart Association (NYHA) class IV symptoms, and emphasize that ICDs have not been shown to improve survival when implanted within 3 months of myocardial revascularization. The guidelines also state that ICD therapy should be considered in patients with a recent cardiomyopathy diagnosis only after their response to optimal medical therapy for ≥3 months has been assessed. For a variety of reasons, however, ICDs may be implanted in patients who do not meet guideline-based (GB) criteria. A retrospective cohort study of cases submitted to the National Cardiovascular Data Registry-ICD Registry found that 22.5% of patients did not meet evidence-based criteria. This study was controversial in part because of the possibility that some variables in the ICD Registry may not have been recorded accurately. Furthermore, no information was available regarding long-term rates of appropriate ICD therapy or overall mortality. We undertook the present study to assess the prevalence of non–guideline-based (NGB) ICD implants at an academic medical center using prespecified criteria, leveraging full access to patients’ medical records and associated clinical decision making. In addition, we compared rates of appropriate ICD therapy and survival between patients with NGB and GB implants, with particular emphasis on the rates of appropriate ICD therapy during the waiting periods mandated by the clinical guidelines.
Methods
All patients who underwent ICD implantation for primary prevention of sudden cardiac death at Beth Israel Deaconess Medical Center from January 1, 2004, to December 31, 2012, were included in this retrospective analysis. Patients were considered to have undergone primary prevention ICD implants if they had no history of cardiac arrest, spontaneous sustained ventricular tachycardia (VT), or syncope compatible with a ventricular arrhythmia with a subsequent electrophysiology study demonstrating inducible sustained VT. Patients with hypertrophic cardiomyopathy, cardiac sarcoidosis, arrhythmogenic right ventricular cardiomyopathy, long QT and Brugada syndromes, and other channelopathies were excluded from this analysis. Patients were also excluded if it was impossible or difficult to determine whether the implant was GB using our prespecified criteria or if there was insufficient information to estimate >3 of the clinical variables discussed subsequently. Patients were followed through December 31, 2013. Outcome data were extracted from the medical records by investigators blinded to the indication for the device.
Clinical variables were identified from medical records, including review of primary data (e.g., electrocardiograms, echocardiograms, and device electrograms) when available. The primary independent variable was NGB versus GB ICD implantation. Patients were classified as having an NGB ICD if their initial ICD was implanted within 40 days of the MI that resulted in their reduced left ventricular ejection fraction (LVEF), within 3 months of either percutaneous coronary intervention or coronary artery bypass graft surgery (revascularization) or within 3 months of cardiomyopathy diagnosis. Patients were also classified as having an NGB ICD if they had NYHA class IV symptoms and did not undergo concurrent cardiac (CRT) at the time of initial implant. For the multivariable analysis, patients who met >1 NGB criteria were included in only 1 NGB category using the hierarchy specified earlier (e.g., a patient with NYHA class IV heart failure with no CRT whose ICD was implanted within 40 days of MI would be categorized as a patient whose ICD was implanted within 40 days of MI). Cases were adjudicated through chart review involving an internal medicine resident (MKT or RG) and an electrophysiology fellow (YCL). An attending electrophysiologist (AEB) was consulted in approximately 150 cases when further review was required.
The primary outcome variable was time to appropriate ICD therapy, which was defined as antitachycardia pacing or shock therapy for VT or ventricular fibrillation. We defined inappropriate therapy as intervention delivered for atrial tachyarrhythmias (including atrial fibrillation, atrial flutter, atrial tachycardia, and sinus tachycardia) or therapy delivered in response to extraneous potentials generated by other cardiac or extracardiac electromagnetic signals. For all ICDs, we recorded programming parameters, especially to ensure that there was no significant difference in detection times. Analysis of stored ICD electrograms was performed using previously published methods by an electrophysiology fellow (YCL) and an attending electrophysiologist (AEB) blinded to the identity of the patient being studied. The secondary outcome variable was death from any cause, determined by the medical record and social security death index. A censor date corresponding to the most recent ICD interrogation or last known follow-up was recorded for all patients who did not receive appropriate ICD therapy in the study period.
Continuous variables were expressed as mean and SD, and categorical variables were expressed as absolute and relative frequencies. Discrepancies in baseline characteristics were evaluated for statistical significance using analysis of variance or the chi-square test. To determine the risk of appropriate ICD therapy and mortality in the NGB categories, time-to-event analysis was performed using the Kaplan-Meier method. Multivariate analysis included the prognostic variables described earlier in a Cox proportional hazards regression model. The Fine and Gray approach was used for competing risk analysis. We performed separate survival analyses for appropriate ICD therapy and for total mortality using STATA IC/13 software (STATA Corporation, College Station, Texas). The study protocol was approved by the Institutional Review Board of the Beth Israel Deaconess Medical Center.
Results
The 806 patients in the study were selected and classified as shown in Figure 1 . Patients with NGB ICDs underwent ICD implantation an average of 12, 45, and 42 days after MI, revascularization, or cardiomyopathy diagnosis, respectively. Of the NGB implants, 10 of 137 patients (7.3%) needed a permanent pacemaker, and because of this, it was decided to implant an ICD, rather than just a pacemaker.
Compared with patients whose implants were GB, those with NGB implants were more likely to have coronary artery disease (CAD), higher NYHA class, lower LVEF, shorter QRS duration, and tobacco use and were less likely to be taking antiarrhythmic medications or to have undergone concurrent CRT ( Table 1 ). Other clinical variables were not significantly different between the 2 cohorts.
| Guideline-based | Non-guideline-based | p | |
|---|---|---|---|
| Total | 669 (82.9%) | 137 (17.0%) | |
| Men | 523 (78.2%) | 108 (78.8%) | 0.87 |
| Age (mean±SD) (years) | 65.8±12.1 | 65.5±13.1 | 0.84 |
| Ischemic cardiomyopathy | 350 (52.2%) | 86 (62.8%) | 0.04 |
| Non-ischemic cardiomyopathy | 260 (38.8%) | 37 (27.0%) | 0.03 |
| Coronary artery disease | 436 (65.2%) | 104 (76.5%) | 0.01 |
| History of myocardial infarction | 369 (55.2%) | 86 (62.8%) | 0.10 |
| History of PCI | 236 (35.3%) | 84 (62.8%) | 0.001 |
| History of CABG | 217 (32.4%) | 51 (37.5%) | 0.25 |
| History of congestive heart failure | 497 (74.3%) | 113 (82.5%) | 0.04 |
| Cardiac resynchronization therapy | 201 (30.0%) | 22 (16.1%) | 0.01 |
| NYHA class I | 177 (26.5%) | 30 (21.9%) | 0.04 |
| NYHA class II | 191 (28.6%) | 39 (28.5%) | 0.70 |
| NYHA class III | 280 (41.9%) | 41 (29.9%) | 0.03 |
| LVEF (mean±SD) | 26.2±9.2 | 23.8±8.5 | 0.005 |
| Body mass index (mean±SD) | 28.8±6.3 | 28.9±6.8 | 0.88 |
| Diabetes | 236 (36.2%) | 55 (40.7%) | 0.32 |
| Hypertension | 515 (77.3%) | 106 (78.5%) | 0.76 |
| History of atrial fibrillation or flutter | 235 (35.3%) | 44 (32.4%) | 0.51 |
| Tobacco use (ever) | 344 (52.7%) | 82 (62.1%) | 0.04 |
| Tobacco use (current) | 54 (8.8%) | 25 (10.2%) | 0.02 |
| Peripheral vascular disease | 87 (13.1%) | 20 (14.6%) | 0.63 |
| Glomerular filtration rate (mean±SD) | 82.0±39.7 | 84.6±37.5 | 0.50 |
| Blood urea nitrogen (mean±SD) | 26.6±15.2 | 25.8±12.7 | 0.62 |
| Beta-blocker use | 556 (83.1%) | 113 (83.1%) | 1.00 |
| ACEi or ARB use | 526 (78.6%) | 103 (75.2%) | 0.38 |
| Aldosterone antagonist | 149 (22.3%) | 24 (17.5%) | 0.22 |
| Antiarrhythmic drug use | 81 (12.1%) | 7 (5.2%) | 0.02 |
| QRS duration (ms, mean±SD) | 132.7±31.9 | 125.8±30.7 | 0.02 |
| Echocardiogram LVDD (cm, mean±SD) | 6.0±0.9 | 6.1±0.8 | 0.33 |
| Echocardiogram LVH | 138 (25.0%) | 19 (16.7%) | 0.10 |
| Electrophysiology study | 207 (30.9%) | 48 (35.0%) | 0.17 |
| Inducible monomorphic VT | 91 (44.0% ∗ ) | 21 (43.8% ∗ ) | 0.70 |
∗ Denotes percentage of all patients who underwent an EP study.
We compared ICD programming parameters in patients with GB and NGB implants. The number of tachycardia zones, cut-off values for VT detection, and number of intervals to detect (Medtronic, Minneapolis, MN and St. Jude Medical Devices, St. Paul, MN) or detection times (Boston Scientific Devices, Marlborough, MA) were similar in both groups.
During a median follow-up of 2.9 years, 16.6% of all patients (134 of 807) received appropriate ICD therapy and 22.0% (177 of 807) died. Compared with patients with GB implants, patients with NGB implants had similar times to first appropriate ICD therapy (median time to event 1.94 vs 2.17 years, p = 0.20) but significantly shorter survival (median time to death 2.57 vs 3.03 years, adjusted p = 0.015; Table 2 ). After adjustment for initial LVEF, CAD, atrial fibrillation/flutter, cardiomyopathy type, NYHA class, antiarrhythmic drug use, and QRS duration, patients with NGB implants had similar risk of appropriate ICD therapy (hazard ratio [HR] 0.83, 95% confidence interval [CI] 0.5 to 1.3, p = 0.51, Figure 2 ) but significantly higher mortality (HR 1.54, 95% CI 1.1 to 2.2, p = 0.03, Figure 2 ). We also performed a competing risk analysis and found no significant differences in the risk of receiving appropriate ICD therapy between the 2 groups, when accounting for competing risk of death (subdistribution HR for NGB implants 0.75; 95% CI 0.5 to 1.2; p = NS). Appropriate ICD therapy was not associated with total mortality risk.
| Variable | All implants (N=806) | Guideline-based implants (N=669) | Non-guideline-based implants (N=137) |
|---|---|---|---|
| Appropriate ICD therapy | 134 (16.6%) | 115 (17.2%) | 19 (13.9%) |
| Median time-to-first appropriate ICD therapy (years) | 2.12 | 2.17 | 1.94 |
| Inappropriate ICD therapy | 50 (6.2%) | 47 (7.0%) | 3 (2.2%) |
| Death (# of patients) | 177 (22.0%) | 135 (20.2%*) | 42 (30.7%*) |
| Median time-to-death (years) | 2.92 | 3.03* | 2.57* |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
