The medical records of 148 patients implanted with an ICD between October 1999 and June 2011 at Samsung Medical Center in Seoul, Korea were analyzed retrospectively. Any additional information was collected by contacting general practitioners, reviewing hospital records, and conducting telephone interviews. The primary outcome was a composite event of all-cause mortality or hospitalization for cardiac causes. The secondary outcome was all-cause mortality during follow-up.

The indication for ICD implantation was secondary prevention in patients who had experienced aborted sudden cardiac death, sustained ventricular tachyarrhythmia, or presumed tachyarrhythmic syncopal attacks. The indication for ICD implantation in all other patients was primary prevention. All defibrillator systems were implanted in the pectoral region. The ICDs were manufactured by St. Jude Medical, Inc. (St. Paul, MN, USA), Medtronic, Inc (Minneapolis, MN, USA) or Guidant Corp (Indianapolis, IN, USA). We often attempted to tailor the therapy based on electrophysiology studies, a patient's arrhythmia history, or both (patient-specific tailored programming). The ICDs can be programmed to provide different therapies for tachyarrhythmias in up to three heart rate zones. The ICD can deliver bursts of antitachycardia pacing, cardioversion, or defibrillation in each therapy zone. Although a variety of algorithms exist, antitachycardia pacing was usually delivered at a slightly faster rate (a cycle length 10% to 12% shorter) than the rate of the detected tachycardia. Devices often delivered synchronized cardioversion for tachyarrhythmias in this range (heart rate below 160 or 180 beats/min), and usually delivered unsynchronized shocks for very rapid ventricular arrhythmias (heart rate > 180 or 200 beats/min). Furthermore, most devices offered 3 algorithms intended to minimize inappropriate shocks: 1) "stability", detecting irregularity in cycle length of the tachyarrhythmia; 2) "sudden onset", monitoring the cycle length for the sudden or abrupt onset of a high ventricular rate rather than a gradually increasing heart rate (10); and 3) "QRS morphology", comparing the electrograms during the tachycardia to the baseline QRS shape, duration, and polarity. The dual-chamber devices provided additional algorithms evaluating the atrial rate (11). The ICD programming, including such discriminator usage, was left to the discretion of the operators.

After ICDs were implanted, patients were followed in our outpatient ICD clinic. The devices were interrogated, and the complete data set (including intracardiac electrograms) was recorded. The delivered therapy was adjudicated by a trained electrophysiologist. Appropriate ICD therapy was defined as either cardioversion, defibrillation of ventricular tachycardia (VT), or defibrillation of ventricular fibrillation (VF) based on the stored electrogram. An inappropriate shock was defined as an episode, starting with a shock not delivered for VT or VF and ending when the ICD detected sinus rhythm. The rhythm triggering therapy was categorized as atrial fibrillation or atrial flutter, supraventricular including sinus tachycardia, or abnormal sensing.

All values are presented as the mean ± standard deviation or median with interquartile range. Comparisons between continuous variables were tested using the t-test or Mann-Whitney U test, as appropriate. Categorical data were tested using the chi-square test or Fisher's exact test, as appropriate. Event-free survival was estimated by the Kaplan-Meier method and compared using the log-rank test. Adjusted hazard rates were compared by multivariable Cox proportional-hazards regression analysis. Covariates that were statistically significant by univariate analysis or that were clinically relevant were considered candidate variables for the multivariate models. The relationship between inappropriate shocks and clinical outcomes was assessed using a Cox proportional-hazards model adjusting for the following covariates: age > 70 yr, previous coronary bypass surgery, history of atrial fibrillation, left ventricular ejection fraction < 40%, serum creatinine ≥ 1.5 (mg/dL), interim appropriate shocks, interim inappropriate shocks, and device type. All tests were two-tailed, and P < 0.05 was considered statistically significant. All analyses were performed with SPSS software version 18.0 (SPSS Inc., Chicago, IL, USA).

The institutional review board at Samsung Seoul Hospital approved this study (IRB No. 2009-09-047). All of subjected patients gave written informed consent for the ICD implantation procedure.

A total of 148 patients with an ICD were identified from the ICD database at Samsung Medical Center. Of these, 34 (23.0%) patients experienced one or more episodes of inappropriate ICD shocks. Baseline clinical characteristics are provided in Table 1. Overall, 34 patients were implanted with an ICD for primary prevention and 114 patients for secondary prevention of sudden cardiac death. The two groups had largely similar clinical characteristics, except significantly fewer inappropriate shock recipients had prior coronary bypass surgery than nonrecipients (0 vs 13.2%, P = 0.023). During the follow-up period, 5 (15.6%) and 58 patients (50.0%) with an ICD for primary and secondary prevention, respectively, had an appropriate shock (P < 0.001). Six (18.8%) and 28 patients (24.1%) with an ICD for primary and secondary prevention, respectively, had an inappropriate shock (P = 0.521). Atrial fibrillation or atrial flutter was the most common cause of inappropriate shock (67.7%), followed by supraventricular tachycardia including sinus tachycardia (23.5%), and then abnormal sensing (8.8%). Of these, two patients suffered inappropriate shocks because of T-wave oversensing and one patient experienced inappropriate shocks because of oversensing due to noise. We analyzed the heart rate at the time of a patient's first inappropriate shock for atrial fibrillation or atrial flutter, and supraventricular tachycardia including sinus tachycardia, but not for shocks due to abnormal sensing. The mean ventricular rate that triggered inappropriate shock for atrial fibrillation or atrial flutter was 179 ± 28 beats/min, and for supraventricular tachycardia including sinus tachycardia it was 172.4 ± 23.2 beats/min.

Cox regression analysis was performed to identify predictors of inappropriate shock (Table 2). Sex, age ≥ 70 yr, secondary prevention, hypertrophic cardiomyopathy, use of beta-blocker, and interim appropriate shock did not influence inappropriate shock in the total patient population. The only significant univariate predictor of inappropriate shock was a history of atrial fibrillation. In the multivariate Cox regression model, the only independent predictor of inappropriate shock was a history of atrial fibrillation. Cox regression analysis was also performed to recognize predictors of mortality (Table 3). Previous coronary bypass surgery, history of atrial fibrillation, serum creatinine ≥ 1.5 (mg/dL), interim appropriate shock, and interim inappropriate shock did not influence mortality in the total patient population. The significant univariate predictors of mortality were age ≥ 70 yr and left ventricular ejection fraction < 40%. In the multivariate Cox regression model, the independent predictors for mortality were age ≥ 70 yr and previous coronary bypass surgery.

The median follow-up duration was 29 months (interquartile range: 8 to 53). Table 4 and Fig. 1 show the cumulative clinical outcomes of the study population during the follow-up period. The composite event of all-cause mortality or hospitalization for any cardiac cause was high in the inappropriate shock group, but the difference between groups was not statistically significant (25.4% for no inappropriate shock vs 35.3% for inappropriate shock, adjusted hazard ratio [HR]: 1.06, 95% confidence interval [CI]: 0.49-2.29, P = 0.877). During the follow-up period, there were no significant differences between groups for all-cause mortality (7.9% vs 11.8%, adjusted HR: 0.59, 95% CI: 0.14-2.45, P = 0.468), death due to cardiac causes (5.3% vs 8.8%, adjusted HR: 0.89, 95% CI: 0.13-6.20, P = 0.889), and hospitalization due to cardiac causes (19.3% vs 32.4%, adjusted HR: 1.51, 95% CI: 0.62-3.67, P = 0.360). In patients implanted for secondary prevention, the incidences of combined all-cause mortality and hospitalization due to a cardiac cause, as well as all-cause mortality only between the two groups did not differ significantly (P = 0.412 for combined all-cause mortality or hospitalization due to a cardiac cause, P = 0.539 for all-cause mortality only).