Economic outcomes of implantable cardioverter-defibrillators.

BACKGROUND: We reviewed the literature pertaining to the cost-effectiveness of implantable cardioverter-defibrillator (ICD) therapy in the management of ventricular fibrillation and tachycardia. Discussed are the methodology, advantages, and limitations of economic-outcomes analyses as related to ICD therapy; the impact of new technology and physician practice patterns; and methodological recommendations for future studies. METHODS AND RESULTS: Articles published between 1990 and 1997 were screened for cost-effectiveness analyses of ICD versus antiarrhythmic drug therapy. Randomized clinical trials, prospective and retrospective studies, and economic models were included. These studies report incremental cost-effectiveness ratios ranging from cost savings of $13 975 per life-year saved (LYS) to an incremental cost of $114 917 per LYS for ICD therapy. Differences were due to study type, cost-reporting methodology, ICD technology used, and length of follow-up. Assuming current technology and physician practice patterns, we find that ICD total therapy costs may break even in 1 to 3 years. CONCLUSIONS: Recent literature suggests that ICDs are a cost-effective therapy for management of life-threatening ventricular tachyarrhythmias. The advent of new technology and patient management practices should further improve the cost-effectiveness of ICD therapy. Future studies of ICD cost-effectiveness should address the implications of truncated follow-up periods and quality of life.

Atrial therapies reduce atrial arrhythmia burden in defibrillator patients.

BACKGROUND: Approximately 25% of patients who receive an implantable cardioverter-defibrillator (ICD) to treat ventricular tachyarrhythmias have documented atrial tachyarrhythmias before implantation. This study assessed the ability of device-based prevention and termination therapies to reduce the burden of spontaneous atrial tachyarrhythmias. METHODS AND RESULTS: Patients with a standard indication for the implantation of an ICD and 2 episodes of atrial tachyarrhythmias in the preceding year received a dual-chamber ICD (Medtronic 7250 Jewel AF) that uses pacing and shock therapies for prevention and/or termination of atrial tachyarrhythmias. In a multicenter trial, patients were randomized to 3-month periods with atrial therapies "on" or "off" and subsequently crossed over. Analysis was performed on the 52 of 269 patients who had episodes of atrial tachyarrhythmia and had >/=30 days of follow-up with atrial therapies on and off. The atrial therapies resulted in a reduction of atrial tachyarrhythmia burden from a mean of 58.5 to 7.8 h/mo. A paired analysis (Wilcoxon signed-rank test) showed that the median difference in burden (1.1 h/mo) was highly significant (P=0.007). When the subgroup of 41 patients treated only with atrial pacing therapies was analyzed, the reduction in burden persisted (P=0.01). CONCLUSIONS: In this study, patients with a standard ICD indication and atrial tachyarrhythmias had a significant reduction in atrial tachyarrhythmia burden with use of atrial pacing and shock therapies.

Global right atrial mapping of human atrial flutter: the presence of posteromedial (sinus venosa region) functional block and double potentials : a study in biplane fluoroscopy and intracardiac echocardiography.

BACKGROUND: Previous studies of atrial flutter have found linear block at the crista terminalis; this was thought to predispose the patient to the arrhythmia. More recent observations, however, have demonstrated crista conduction. We sought to characterize the posterior boundary of atrial flutter. METHODS AND RESULTS: Patients with counterclockwise flutter (n=20), clockwise flutter (n=3), or both (n=5) were studied using two 20-pole catheters. Biplane fluoroscopy determined catheter positions. During counterclockwise flutter, craniocaudal activation occurred along the entire lateral and posterior right atrial walls. Septal activation proceeded caudocranially. In all patients, a line of block was seen in the posteromedial (sinus venosa) right atrium; this was manifested by the presence of double potentials where the upward and downward activations collided. Anatomic location was confirmed by intracardiac echocardiography in 9 patients. In patients with clockwise flutter, the line of block and double potentials were seen in the same location during counterclockwise flutter, but the activation sequence around the line of block was reversed. Pacing near the site of double potentials during sinus rhythm excluded a fixed line of block, and premature atrial complexes demonstrated functional block with manifest double potentials. In 2 patients, posterior ectopy organized to subsequently initiate isthmus-dependent atrial flutter. CONCLUSIONS: (1) A functional line of block is seen at the posteromedial (sinus venosa region) right atrium during counterclockwise and clockwise atrial flutter. (2) All lateral wall right atrial activation can be uniform during flutter, without linear block or double potentials in the region of the crista terminalis. (3) Activation at the site of posteromedial right atrial functional block can organize to subsequently initiate isthmus-dependent atrial flutter.

Electromagnetic interference from welding and motors on implantable cardioverter-defibrillators as tested in the electrically hostile work site.

OBJECTIVES: This study was designed to determine the susceptibility of an implanted cardioverter-defibrillator to electromagnetic interference in an electrically hostile work site environment, with the ultimate goal of allowing the patient to return to work. BACKGROUND: Normal operation of an implanted cardioverter-defibrillator depends on reliable sensing of the heart's electrical activity. Consequently, there is concern that external electromagnetic interference from external sources in the work place, especially welding equipment or motor-generator systems, may be sensed and produce inappropriate shocks or abnormal reed switch operation, temporarily suspending detection of ventricular tachycardia or ventricular fibrillation. METHODS: The effects of electromagnetic interference on the operation of one type of implantable cardioverter-defibrillator (Medtronic models 7217 and 7219) was measured by using internal event counter monitoring in 10 patients operating arc welders at up to 900 A or working near 200-hp motors and 1 patient close to a locomotive starter drawing up to 400 A. RESULTS: The electromagnetic interference produced two sources of potential interference on the sensing circuit or reed switch operation, respectively: 1) electrical fields with measured frequencies up to 50 MHz produced by the high currents during welding electrode activation, and 2) magnetic fields produced by the current in the welding electrode and cable. The defibrillator sensitivity was programmed to the highest (most sensitive) value: 0.15 mV (model 7219) or 0.3 mV (model 7217). The ventricular tachycardia and ventricular fibrillation therapies were temporarily turned off but the detection circuits left on. CONCLUSIONS: None of the implanted defibrillators tested were affected by oversensing of the electric field as verified by telemetry from the detection circuits. The magnetic field from 225-A welding current produced a flux density of 1.2 G; this density was not adequate to close the reed switch, which requires approximately 10 G. Our testing at the work site revealed no electrical interference with this type of defibrillator. Patients were allowed to return to work. The following precautions should be observed by the patient: 1) maintain a minimal distance of 2 ft (61 cm) from the welding arc and cables or large motors, 2) do not exceed tested currents with the welding equipment, 3) wear insulated gloves while operating electrical equipment, 4) verify that electrical equipment is properly grounded, and 5) stop welding and leave the work area immediately if a therapy is delivered or a feeling of lightheadedness is experienced.

High failure rate for an epicardial implantable cardioverter-defibrillator lead: implications for long-term follow-up of patients with an implantable cardioverter-defibrillator.

OBJECTIVES: The purpose of this study was to determine the risk of epicardial lead failure during long-term follow-up and its mode of presentation. BACKGROUND: Despite the high prevalence of epicardial lead-based implantable cardioverter-defibrillators, their long-term performance is unknown, and appropriate follow-up has not been established. METHODS: The study group comprised all patients in whom an epicardial lead system was implanted at the Mayo Clinic between October 31, 1984 and November 3, 1994. The number of lead fractures and leads with fluid within the insulation and the mode of presentation were determined retrospectively by review of patient visits, radiographs of lead systems and data derived from formal lead testing. RESULTS: At 4 years, the survival rate free of lead malfunction, using formal lead testing, for 160 Medtronic epicardial patches (models 6897 and 6921) was 72% compared with 92.5% for the 179 Cardiac Pacemaker, Inc. (CPI) patches (models 0040 and 0041) (p = 0.01). In addition, five Medtronic patches in three patients had fluid within the lead insulation but no obvious fracture. No CPI patches had fluid identified within the leads. Of 330 Medtronic epicardial pace/sense leads (model 6917), the 4-year survival rate free of lead malfunction as assessed by lead testing was 96%. In all, 19 presentations of lead malfunction were found in 17 patients (2 patients had more than one lead fracture at different times). In 11 (58%) of these presentations, the patients were asymptomatic despite the presence of obvious lead fracture. CONCLUSIONS: Epicardial lead malfunction is common on long-term follow-up, and some leads have a failure rate of 28% at 4 years. Many patients with fractured leads remain asymptomatic, despite involvement of multiple leads in some cases. Therefore, consideration should be given to regular periodic lead testing in addition to routine X-ray examination, as asymptomatic lead malfunction can present with normal chest X-ray findings.

Arrhythmogenic effects of antiarrhythmic drugs: a study of 506 patients treated for ventricular tachycardia or fibrillation.

Antiarrhythmic therapy in 506 consecutive patients undergoing 1,268 antiarrhythmic drug trials for ventricular tachycardia or ventricular fibrillation was reviewed for evidence of arrhythmogenic drug effect defined as the occurrence of a new form of ventricular tachyarrhythmia temporally associated with initiation of drug therapy or dosage increase. Arrhythmogenic effects occurred in 6.9% of patients and 3.4% of drug trials. This ranged from a high of 11.8% caused by encainide to none occurring with procainamide, tocainide or beta-adrenergic blocking drugs. The incidence of arrhythmogenesis was significantly greater in patients whose presenting arrhythmia was sustained ventricular tachycardia than it was in those who presented with nonsustained ventricular tachycardia or ventricular fibrillation (p = 0.02). Decreased systolic function measured echocardiographically at the base of the left ventricle was associated with an increased incidence of arrhythmogenic effects (p = 0.006) whereas global left ventricular ejection fraction was not. Age, gender, cardiac diagnosis, location of prior myocardial infarction and New York Heart Association functional class for heart failure were not related to the occurrence of drug-induced arrhythmias. These findings emphasize the need for in-hospital cardiac monitoring during initiation of antiarrhythmic drug therapy for ventricular tachyarrhythmias.

Regional sympathetic denervation after myocardial infarction in humans detected noninvasively using I-123-metaiodobenzylguanidine.

Transmural myocardial infarction in dogs produces denervation of sympathetic nerves in viable myocardium apical to the infarct that may be arrhythmogenic. It is unknown whether sympathetic denervation occurs in humans. The purpose of this study was to use iodine-123-metaiodobenzylguanidine (MIBG), a radiolabeled guanethidine analog that is actively taken up by sympathetic nerve terminals, to image noninvasively the cardiac sympathetic nerves in patients with and without ventricular arrhythmias after myocardial infarction. Results showed that 10 of 12 patients with spontaneous ventricular tachyarrhythmias after myocardial infarction exhibited regions of thallium-201 uptake indicating viable perfused myocardium, with no MIBG uptake. Such a finding is consistent with sympathetic denervation. One patient had frequent episodes of nonsustained ventricular tachycardia induced at exercise testing that was eliminated by beta-adrenoceptor blockade. Eleven of the 12 patients had ventricular tachycardia induced at electrophysiologic study and metoprolol never prevented induction. Sympathetic denervation was also detected in two of seven postinfarction patients without ventricular arrhythmias. Normal control subjects had no regions lacking MIBG uptake. This study provides evidence that regional sympathetic denervation occurs in humans after myocardial infarction and can be detected noninvasively by comparing MIBG and thallium-201 images. Although the presence of sympathetic denervation may be related to the onset of spontaneous ventricular tachyarrhythmias in some patients, it does not appear to be related to sustained ventricular tachycardia induced at electrophysiologic study.

Adenosine: potential modulator for vasovagal syncope.

OBJECTIVES: This study examined the hypothesis that adenosine could provoke a vasovagal response in susceptible patients. Mechanisms of the vasovagal response were further explored by studying the adenosine-mediated reactions. BACKGROUND: Increased sympathetic activity is frequently observed before vasovagal syncope. Recent studies have demonstrated that adenosine, in addition to its direct bradycardiac and vasodilatory effects, can increase sympathetic discharge by activating cardiovascular afferent nerves. METHODS: The effects of adenosine and head-up tilt-table testing with or without isoproterenol were prospectively evaluated in 85 patients examined for syncope after negative results of electrophysiologic testing (51 men and 34 women, mean [+/- SD] age 61 +/- 17 years). Adenosine bolus injections of 6 mg and 12 mg were sequentially administered to patients in the upright position. The same protocol was implemented in 14 normal control subjects (7 men and 7 women, mean [+/- SD] age 38 +/- 10 years). RESULTS: Transient hypertension or tachycardia was observed in 57 (67%) and 20 (24%) patients after administration of 6 mg and 12 mg of adenosine, respectively, during the immediate phase (first 15 s), suggesting direct sympathetic activation. Hypotension and reflex tachycardia were observed in all patients during the delayed phase (15 to 60 s after adenosine injection), suggesting baroreceptor unloading. A vasovagal response was induced in 22 (26%) and 29 (34%) patients after adenosine administration and during tilt-table testing. Inducibility of a vasovagal response by these two methods was comparable (p = 0.12). Of the control subjects, one (7%) had a vasovagal response after adenosine administration and one (7%) had a positive response during tilt-table testing. CONCLUSIONS: These observations support the idea that adenosine is an endogenous modulator of the cardiac excitatory afferent nerves. Sympathetic activation by adenosine can be direct (i.e., cardiac excitatory afferent nerves) and indirect (i.e., vasodilation and reflex sympathetic activation). Adenosine could be an important modulator in triggering a vasovagal response in susceptible patients during examination for syncope.

Effect of failed defibrillation shocks on electrogram amplitude in a nonintegrated transvenous defibrillation lead system.

Concern has been raised regarding the ability of a nonthoracotomy integrated lead system to redetect ventricular fibrillation following failed defibrillation shocks due to diminution in postshock intracardiac electrogram amplitude. Whether such a problem could occur with other lead systems is not known, leading to uncertainty regarding a potential ongoing risk of sudden cardiac death in some patients despite implantable cardioverter-defibrillator therapy. To investigate this problem, we measured the amplitude of 10 consecutive ventricular fibrillation endocardial electrograms immediately before and immediately after failed defibrillation shocks in 15 patients at the time of implantation of a nonintegrated, transvenous, pace/sense/defibrillation lead. Overall, mean electrogram amplitude decreased 21%, from 10.7 +/- 4.6 mV before to 8.5 +/- 4.9 mV immediately after failed defibrillation shocks. The change in electrogram amplitude postshock was directly related to shock energy (r = 0.85, p < 0.0005), but shock waveform had no differential effect. Electrogram amplitude could also increase after failed shocks, particularly following those of low energy. No failures to redetect ventricular fibrillation were found. Thus, intracardiac electrogram amplitude is reduced following failed defibrillation shocks in this nonintegrated lead system, but by an amount less than that previously reported for some integrated lead systems. Our findings reveal that failed low energy defibrillation shocks are likely to result in less diminution in postshock intracardiac electrogram amplitude than high energy shocks, and that the postshock amplitude may even increase after some failed shocks.

A population-based estimate of candidacy rates for the implantable cardioverter-defibrillator.

The implantable cardioverter-defibrillator (ICD) increases survival of patients who receive the device. However, candidacy rates have not been calculated for a defined population, and the potential effect of the device on the survival of all patients with heart disease has not been estimated. To make these calculations, medical records were reviewed for 1976 to 1988 in a population demographically similar to the white population of the United States. Definite and possible candidates were identified on the basis of American Heart Association/American College of Cardiology guidelines. Candidacy rates ranged from 3.3/100,000 (counting only definite candidates for the entire period) to 8.7/100,000 (counting definite and possible candidates after 1980). Extrapolated to the 1990 U.S. population, estimates ranged from 8,207 to 21,637 new candidates each year. During an average follow-up of 5 years, half of all deaths among candidates had the potential to be delayed by an ICD. In a similar population that has a death rate from heart disease of approximately 280/100,000, 0.6 to 1.6% of subjects have the potential to have their deaths delayed to some extent by an ICD.

Comparison of results in two implantable defibrillators. Jewel 7219D Investigators.

The Jewel 7219D was the first non-thoracotomy implantable cardioverter-defibrillator (ICD) with biphasic shock capability small enough to be placed in the prepectoral subcutaneous position. Size reduction of ICDs is desirable, but safety and efficacy of smaller devices must be demonstrated. Outcomes of patients treated with the Jewel 7219D defibrillator (n = 1,781) and with its precursor model PCD 7217B (n = 2,637) were compared. To use PCD patients (n = 2,637) as historical (n = 2,574) and concurrent controls (n = 63), statistical adjustments using the Cox proportional-hazards regression model were made. Jewel recipients (n = 1,781) treated in 106 US and 32 non-US centers exhibited similar characteristics including a mean age of 59 years, 78% men, ejection fraction of 34%, history of aborted sudden cardiac death in 41%, and coronary artery disease in 70%. Implantation was completed in 1,777 of 1,781 (99.9%) attempts and success with the first electrode configuration and polarity was 89.5%. Kaplan-Meier cumulative first-year survivals for cardiac and all-cause mortality were 98.5% and 93.3%. Complication-free first-year survival for Jewel implants in prepectoral subcutaneous (n = 582), subpectoral submuscular (n = 366), and abdominal (n = 449) positions did not differ (p > 0.05). First-year survival free of pocket-related complications exceeded 98% in all locations. Adjusted cardiac and all-cause first-year mortality, and efficacy in terminating spontaneous tachyarrhythmias did not differ between the 2 device groups. In conclusion, the safety and efficacy of Jewel model 7219D in the prepectoral subcutaneous position are at least equal to either those of Jewel models implanted in different positions or to those of the previously extensively characterized PCD 7217B.

A stepwise testing protocol for modern implantable cardioverter-defibrillator systems to prevent pacemaker-implantable cardioverter-defibrillator interactions.

Current use of newer implantable cardioverter-defibrillators (ICDs) has changed the spectrum of pacemaker-ICD interactions and provided new tools for testing and understanding those interactions. Testing for pacemaker-ICD interactions was performed in 31 procedures involving 22 patients. The protocol included: (1) evaluation of pacemaker stimulus artifact amplitude and its ratio to that of the evoked ventricular electrogram, (2) testing for inhibition of ventricular fibrillation (VF) detection by the ICD during asynchronous pacing at maximum output, (3) evaluation by pacemaker event marker recordings of pacemaker sensing behavior while programmed to nonasynchronous mode during ventricular tachycardia (VT) or VF, and (4) evaluation of postshock interactions. Inhibition of detection of VT/VF was found in 6 of 22 patients (27.2%). Large stimulus artifact amplitude (>2 mV) or stimulus artifact:evoked QRS ratio > 1/3 had a positive predictive accuracy of 18% and 14.4%, respectively, and a negative predictive accuracy of 100% and 92.3%, respectively, for clinically significant interaction. Asynchronous pacing occurred in 16 of 31 procedures (51.6%), and was due to underdetection by the pacemaker in 4 of 16 (25%) and noise reversion in 12 of 16 (75%). Postshock phenomena occurred in 6 cases, 3 of which were clinically significant. Overall, 11 of 22 patients (50%) had clinically significant interactions discovered by this protocol, which led to system revisions in 6 and to pacemaker output reprogramming in 5. Thus, pacemaker-ICD interactions are frequently detected using a thorough and systematic protocol. Most cases can be managed by system revision or pacemaker reprogramming.

Narrow QRS complex tachycardias.

Regular narrow QRS complex tachycardias are a common problem encountered by general internists or family practitioners. Although such tachycardias often occur in patients with a normal heart and seldom represent life-threatening conditions, they may cause bothersome symptoms. The key to approaching the diagnosis of these arrhythmias is identifying atrial activity (P waves) on the surface electrocardiogram and classifying the tachycardia as long RP or short RP. On the basis of that information, a differential diagnosis can be generated, logical therapy can be delivered for termination of the tachycardia, and a plan can be developed to prevent recurrence. Because intravenously administered adenosine alleviates 90% of the episodes of supraventricular tachycardias and has minimal side effects, it has become the drug of choice for termination of most types of narrow QRS complex tachycardias.

Propafenone-induced peripheral neuropathy.

Propafenone hydrochloride, a class IC antiarrhythmic drug, is used in the treatment of ventricular and supraventricular arrhythmias. Herein we describe a patient with episodic jabbing and crushing pain in his hands and feet, aching in his forearms, and hyperesthesias of his extremities. He had been taking propafenone for 1 year because of ventricular arrhythmias. Results of a nerve conduction velocity test were abnormal. Electron microscopic findings on a sural nerve biopsy specimen represented distal small fiber neuropathy. Findings on a thermoregulatory sweat test and on autonomic tests were abnormal, compatible with a distal small fiber neuropathy. To our knowledge, peripheral neuropathy has not previously been reported to occur with use of propafenone. In this patient, propafenone seemed to be responsible for the development of peripheral neuropathy, which resolved after use of the drug had been discontinued.

Prolonged QT syndrome: perioperative management.

Intraoperative cardiac arrhythmias related to prolonged QT syndrome are uncommon. We describe a 26-year-old woman in whom ventricular fibrillation occurred during the final minutes of a partial glossectomy and right supraomohyoid selective neck dissection and discuss the role that this specific operation may have had in the development of the intraoperative event. In addition, we review the perioperative management of patients with prolonged QT syndrome.

Randomized prospective pilot study of long-term dual-site atrial pacing for prevention of atrial fibrillation.

OBJECTIVE: To determine whether dual-site atrial pacing is feasible, safe, and effective. DESIGN: We undertook a randomized prospective single-blind crossover study. MATERIAL AND METHODS: Nine patients with at least two episodes per month of symptomatic paroxysmal atrial fibrillation participated in a randomized crossover study involving three separate 3-month blocks of single-site atrial pacing, dual-site atrial pacing, and control (support-only) pacing. RESULTS: Dual-site atrial pacing resulted in shorter P wave duration (81 +/- 14 ms) than did single-site pacing (111 +/- 12 ms) or control sinus rhythm (123 +/- 9 ms) (P<0.0001) and in fewer premature atrial complexes on Holter monitoring (P = 0.06). The arrhythmia-free interval was longer with dual-site pacing (67 +/- 17 days) than with single-site (62 +/- 30 days) or support-only (49 +/- 34 days) pacing (P = 0.10). This pilot study was not statistically powered to detect a difference between pacing modes. CONCLUSION: (1) Dual-site atrial pacing is feasible and safe; (2) it shortens the P wave duration and tends to decrease premature atrial complexes on Holter monitoring; (3) any atrial pacing tends to prolong the arrhythmia-free interval; and (4) this pilot study enrolled too few patients to determine whether a significant difference in pacing modes exists and supports the need for a larger study.

Torsades de pointes ventricular tachycardia associated with overdose of astemizole.

An overdose of astemizole predisposes the myocardium to ventricular dysrhythmias, including torsades de pointes. Herein we describe a case of astemizole-induced torsades de pointes ventricular tachycardia and also review previous case reports in the literature. All the patients were young, and dysrhythmias developed only in those with corrected QT intervals greater than 500 ms. Although several mechanisms have been postulated, no clear explanation has been provided for why astemizole promotes myocardial dysrhythmias. Treatment of astemizole-induced torsades de pointes includes discontinuing use of astemizole, intravenous administration of magnesium sulfate and isoproterenol, temporary cardiac pacing, and, when necessary, direct current cardioversion. A cardiac cause of syncope or convulsions must not be overlooked, especially in patients taking H1 antagonists because they often have these symptoms before hospitalization or detection of torsades de pointes (or both).

Consistent subcutaneous prepectoral implantation of a new implantable cardioverter defibrillator.

OBJECTIVE: To describe the use of a new implantable cardioverter defibrillator (ICD) that can be placed in the prepectoral region rather than implanted in the abdominal wall. DESIGN: We report the experience of placement of this new ICD in the prepectoral region in 13 patients from Sept. 28, 1993, through Jan. 10, 1994, at the Mayo Clinic. MATERIAL AND METHODS: Thirteen consecutive patients offered this new ICD underwent placement of transvenous defibrillation leads, and the pulse generator was placed in a pocket formed in the subcutaneous, prepectoral space. Testing ensured a defibrillation threshold of 24 J or less. RESULTS: In all 13 patients, the pulse generator could be placed in the subcutaneous, prepectoral space. In all except one patient, acceptable defibrillation thresholds were achieved by using lead systems placed totally transvenously. Only one patient required placement of a subcutaneous patch. All but two patients were dismissed from the hospital within 3 days after the ICD implantation. CONCLUSION: Consistent subcutaneous, prepectoral placement of this new ICD pulse generator is possible. Because the entire procedure can be performed in the pacemaker implantation room, the potential exists for decreasing the duration of the hospitalization and associated costs.

Operation for ventricular tachyarrhythmias: refining current treatment strategies.

For many patients with ventricular tachyarrhythmias, the optimal choice of palliative or curative therapies is not yet well established. To refine optimal current treatment strategies, baseline patient characteristics were studied in relation to likelihood of successful outcome in 240 consecutive patients undergoing operation for treatment of ventricular tachyarrhythmias from 1981 to 1991. Indications for operation were sudden cardiac death or inducible ventricular tachyarrhythmias refractory to medical therapy (or both). Treatment was directed endocardial procedures in 77 patients (32%), other cardiac procedures in 57 patients (24%) (coronary artery bypass grafting in 94% and valve procedure in 14%, either with [35%] or without [65%] concomitant implantable cardioverter-defibrillator), and implantable cardioverter-defibrillator alone in 106 patients (44%). Overall 30-day operative mortality was 5% (70% confidence interval, 4%-7%) and 2-year survival was 74% (70% confidence interval, 71%-77%). Overall 2-year freedom from sudden cardiac death was 97% (70% confidence interval, 96%-98%) and was similar (p = not significant) for all treatment modalities. For each treatment modality, multivariate analysis identified independent risk factors for operative mortality and 2-year tachyarrhythmia recurrence, advanced angina and congestive heart failure New York Heart Association classes, and overall mortality. To characterize better the use and benefit of coronary artery bypass grafting, risk factors related to outcome also were identified for patients stratified according to absence (44 patients) or presence (119 patients) of coronary artery disease excluding patients treated by directed endocardial procedures.(ABSTRACT TRUNCATED AT 250 WORDS)