Comparison of dilated cardiomyopathy and coronary artery disease in patients with life-threatening ventricular arrhythmias: Differences in presentation and outcome in the AVID registry.

BACKGROUND: The etiology of structural heart disease in patients with life-threatening arrhythmias (ventricular tachycardia [VT]/ventricular fibrillation [VF]) may define clinical characteristics at presentation, may require that different therapies be administered, and may cause different mortality outcomes. METHODS: In the Antiarrhythmics Versus Implantable Defibrillators (AVID) registry, baseline clinical characteristics, treatments instituted, and ultimate mortality outcomes from the National Death Index were obtained on 3117 patients seen at participating institutions with VT/VF, irrespective of participation in the randomized trial. By use of these data, 2268 patients with coronary artery disease (CAD) were compared with 334 patients with dilated nonischemic cardiomyopathy (DCM). RESULTS: The CAD group was 7 years older and had a higher percentage of males. DCM patients were more likely to be African American, have severely compromised left ventricular function (52% vs 39%), and have a history of congestive heart failure symptoms (62% vs 44%). Patients with CAD were more likely to be treated with b-blockers and calcium channel blockers and less likely to be treated with angiotensin-converting enzyme inhibitors. Patients with DCM were more likely to be treated with diuretics, warfarin, and an implantable cardioverter defibrillator for VT/VF (54% vs 48% for CAD); the use of other antiarrhythmic therapies did not differ between the 2 groups. Two-year survival was not significantly different between the groups (76.6% [95% CI 74.6%-78.7%] vs 78.2% [95% CI 73.6%-82.9%]). CONCLUSIONS: In AVID registry patients with VT/VF, demographic and clinical characteristics were different between patients with CAD and those with DCM. Despite these differences, overall survival was similar in these 2 groups.

Transvenous low energy internal cardioversion for atrial fibrillation: a review of clinical applications and future developments.

Low energy internal atrial cardioversion can be performed by delivering biphasic shocks between transvenous catheters positioned within the cardiac chambers or great vessels. Delivery of shocks results in effective cardioversion at energies < 6-10 J and the procedure can be effective even when external cardioversion has failed. Shock induced discomfort varies from patient to patient, but the procedure can be usually performed without general anesthesia and eventually under mild sedation. Nevertheless, tolerability has to be improved by obtaining a substantial reduction in defibrillating thresholds. With regard to safety, delivery of shocks for defibrillating the atria implies a potential risk of inducing ventricular fibrillation; to minimize this risk, shock delivery must be synchronous to the QRS and should be avoided during rapid RR cycles (< 300 ms). Presently, transvenous low energy cardioversion is an investigational procedure, but a widening of indications is expected in the near future. The cost of the procedure, which remains invasive and requires a brief hospital stay, must be balanced with the benefit of restoring sinus rhythm and the possibility of maintaining sinus rhythm for the medium- to long-term. Experimental and clinical investigations of low energy internal cardioversion have resulted in the development of devices for atrial defibrillation whose clinical role and cost-benefit ratio is currently under evaluation.

Initial clinical experience with a dual chamber rate responsive implantable cardioverter defibrillator.

The present study reports the first clinical experience with the Photon DR dual chamber rate responsive ICD. Fifty-seven patients (mean age 67.6 +/- 10 years) with a mean LVEF of 0.332 in NYHA Class I (30%), II (51%), or III (19%) met the implant criteria. Tachyarrhythmia diagnoses included VF (21%), VT (49%), or VT/VF (30%). Additional bradycardia pacing diagnoses included AV block (18%), sinus node dysfunction (5%), and sinus bradycardia (11%). All patients were followed on a long-term basis. A device-based direct current fibrillator (DC fibber) successfully induced VF in 96.2% of 243 attempted inductions. Detection times (2.86 +/- 0.47 s) and redetection times (1.29 +/- 0.32 s) compared favorably with historic controls using a morphology detection (MD) clinical data. Defibrillation energy and voltage thresholds were 10.4 +/- 5.5 J (range 3-28) and 421 V +/- 108 (range 230-696), respectively. Ninety-two clinical arrhythmias were recorded. Using a nominal nonindividualized algorithm, the sensitivity for VT detection was 100% and specificity for SVT rejection 81%. There were six patient deaths (2 nonsudden arrhythmic, 2 cardiac, 2 other). In conclusion, early clinical experience with the Photon DR ICD supports its safety, efficacy, and ability to provide advanced dual chamber ICD features.

Late retesting of system performance in ICD patients without spontaneous shocks.

UNLABELLED: One hundred five implantable cardioverter defibrillator (ICD) patients (71 +/- 9 years of age, 83% men) without spontaneous ICD discharges for > or = 12 months were tested to assess high voltage (HV) circuit integrity and the system's ability to recognize and terminate ventricular fibrillation (VF). Indications for ICD implantation were sustained ventricular tachycardia (VT) (35%), cardiac arrest (27%), and inducible VT (38%). Eighty-two percent of the patients had coronary artery disease (CAD), and the mean left ventricular ejection fraction (LVEF) was 36% +/- 13%. RESULTS: One hundred patients had inducible VF and five did not. Testing led to ICD reprogramming in 50 (49%) patients. Two (1.9%) patients required ICD replacement: (1) a 45-year-old patient with a Ventritex 110 ICD implanted for 13 months interfaced with a CPI 0062 lead implanted for 46 months could not be defibrillated internally (impedance nonmeasurable); (2) an 82-year-old patient with a 23-month-old Medtronic 7219 ICD interfaced with 6936 and 6933 leads whose defibrillation threshold (DFT) had doubled since implantation (24 J from 12 J). Lead fractures were found in both cases (proximal coil of the 0062, and subcutaneously in the 6933). Based on DFT determinations, the first shock output was programmed lower in 37 patients and higher in 10 patients. Shock pulse width was changed in one patient and the ventricular refractory period in another. No programming changes were made in 54 (51%) patients. CONCLUSIONS: (1) Late testing of HV circuit integrity in ICD patients without an ICD shock in > or = 12 months identifies previously unsuspected HV lead fractures; (2) chronic DFT testing resulted in HV output reprogramming in one-half of the patients.

Temporary transvenous cardioversion and defibrillation: a new method for practical tachyarrhythmia management.

The use of transvenous ICD systems and the recent advances in atrial defibrillation techniques have heightened interest in internal defibrillation. However, most shocks for induced or spontaneous arrhythmias in patients without devices are still delivered transthoracically using high energy. We describe the history of temporary internal defibrillation techniques and report the initial clinical results with a custom built disposable catheter for internal cardioversion and defibrillation. This prototype successfully converted more than 95% of 109 episodes of VT or VF in 28 patients, with biphasic energies < or = 20 J. A newer disposable catheter, using 40-wire Matrix technology as the defibrillating electrode, has design features that provide high surface area, low impedance, and low current density when compared to other leads used for similar purposes. Temporary internal cardioversion-defibrillation of induced and spontaneous arrhythmias using such catheter designs is likely to be widely applicable to patients undergoing electrophysiology procedures and to those in critical care units prone to tachyarrhythmias.

Predischarge arrhythmia induction testing of implantable defibrillators may be unnecessary in selected cases.

Complete postoperative evaluation of implantable cardioverter-defibrillators (ICDs) before discharge, including arrhythmia induction, has been the standard since their introduction. Whereas the original ICDs provided little telemetered information and used separate pace-sense and defibrillation leads, modern, third-generation devices provide pace-sense function information in addition to other data and are used in conjunction with integrated transvenous endocardial leads that combine pace-sense and defibrillation function. Changes in lead position, which can potentially result in either an inability to detect fibrillation or to terminate it, should be mirrored by changes in resting pace-sense function. Thus, for newer ICDs implanted with integrated endocardial lead systems, it is possible that in at least some cases predischarge arrhythmia inductions can be avoided. Two hundred patients receiving third-generation ICDs in conjunction with integrated transvenous leads were evaluated before discharge. Defibrillation detection or termination problems were seen in 8. Declines in resting R-wave amplitude and pacing impedance were significantly associated with such complications (-7 +/- 5 vs -0.3 +/- 2.3 mV [p <0.0001] and -158 +/- 138 vs -93 +/- 76 omega [p <0.05], for those with vs without complications, respectively), as were gross right ventricular lead migrations on chest x-ray. No patient with a defibrillation complication had an R-wave change of <3 mV. However, 13% of patients without complications had R-wave changes of >3 mV. It is concluded that a pace-sense evaluation of ICDs may be a satisfactory screen to determine those who need to go on to complete testing with arrhythmia induction in selected cases.

Multicenter comparison of truncated biphasic shocks and standard damped sine wave monophasic shocks for transthoracic ventricular defibrillation. Transthoracic Investigators.

BACKGROUND: The most important factor for improving out-of-hospital ventricular fibrillation survival rates is early defibrillation. This can be achieved if small, lightweight, inexpensive automatic external defibrillators are widely disseminated. Because automatic external defibrillator size and cost are directly affected by defibrillation waveform shape and because of the favorable experience with truncated biphasic waveforms in implantable cardioverter-defibrillators, we compared the efficacy of a truncated biphasic waveform with that of a standard damped sine monophasic waveform for transthoracic defibrillation. METHODS AND RESULTS: The principal goal of this multicenter, prospective, randomized, blinded study was to compare the first-shock transthoracic defibrillation efficacy of a 130-J truncated biphasic waveform with that of a standard 200-J monophasic damped sine wave pulse using anterior thoracic pads in the course of implantable cardioverter-defibrillator testing. Pad-pad ECGs were also examined after transthoracic defibrillation. After the elimination of data for 24 patients who did not meet all protocol criteria, the results from 294 patients were analyzed. The 130-J truncated biphasic pulse and the 200-J damped sine wave monophasic pulse resulted in first-shock efficacy rates of 86% and 86%, respectively (P = .97). ST-segment levels measured 10 seconds after the shock in 151 patients in sinus rhythm were -0.26 +/- 1.58 and -1.86 +/- 1.93 mm for the 130- and 200-J shocks, respectively (P < .0001). CONCLUSIONS: We found that 130-J biphasic truncated transthoracic shocks defibrillate as well as the 200-J monophasic damped sine wave shocks that are traditionally used in standard transthoracic defibrillators and result in fewer ECG abnormalities after the shock.

Effects of oral propafenone on defibrillation and pacing thresholds in patients receiving implantable cardioverter-defibrillators. Propafenone Defibrillation Threshold Investigators.

OBJECTIVES: The effects of propafenone, a predominantly class IC antiarrhythmic drug, on defibrillation and pacing thresholds were evaluated in patients undergoing cardioverter-defibrillator implantation. BACKGROUND: Previous studies have shown that the class IC agents encainide and flecainide may increase the energy requirements for pacing and defibrillation. Animal studies with propafenone have shown inconsistent results regarding its effect on defibrillation energy requirements. This report investigated the effects of propafenone on defibrillation and pacing thresholds in humans. METHODS: After cardioverter-defibrillator implantation, 47 patients were enrolled in a double-blind, three-way parallel, randomized trial of 450 mg/day (Group 1) or 675 mg/day (Group 2) of oral propafenone or placebo (Group 3) for 3 to 7 days. Predischarge defibrillation and pacing thresholds after treatment were compared with baseline thresholds obtained at implantation. RESULTS: There was no statistically significant difference between implantation and predischarge defibrillation thresholds in the three groups (Group 1: [mean +/- SE] 11.0 +/- 1.3 vs. 12.1 +/- 1.5 J; Group 2: 11.5 +/- 1.1 vs. 13.6 +/- 1.3 J; Group 3: 12.5 +/- 1.2 vs. 13.3 +/- 1.6 J), and no significant difference between treatment groups was found with a 0.86 power to detect a 5-J difference between groups. Paired pulse width pacing thresholds at 2.8 V were compared in 14 patients. A small increase of 0.02 ms was noted at predischarge testing in patients treated with propafenone and placebo. CONCLUSIONS: Short-term oral propafenone (450 and 675 mg/day) does not significantly affect defibrillation or pacing thresholds. Concomitant use of propafenone in patients with implantable cardioverter-defibrillators with recurrent ventricular or atrial tachyarrhythmias should not interfere with proper device function.

Compatibility of a nonthoracotomy lead system with a biphasic implantable cardioverter-defibrillator. Cadence-Endotak 60-Series IDE investigators.

This prospective multicenter study was conducted under the Food and Drug Administration Investigational Device Exemption to evaluate the safety and efficacy of the combination of the Cadence implantable defibrillator (Ventritex, Inc.) and 60-series Endotak C leads (Cardiac Pacemakers, Inc.). Implantation was attempted in 148 patients with hemodynamically compromising ventricular tachycardia or fibrillation (VF), or with pace-terminable ventricular tachycardia. The system was successfully implanted in 97% of patients, with 96% of implants in a transvenous-lead-alone configuration. At implantation, the defibrillation threshold was 455 +/- 94 V (14 +/- 6 J) for lead-alone patients and 532 +/- 40 V (19 +/- 3 J) for those requiring a subcutaneous patch. VF conversion efficacy was reconfirmed in patients who underwent a 3-month chronic induction study. The system successfully detected all 763 induced arrhythmias and terminated 99.5% of them; after system modification, successful conversion was demonstrated in the 2 patients who initially had induced episodes requiring external defibrillation (1 lead revision; 1 reprogramming). All spontaneous episodes were terminated with an implantable-cardioverter defibrillator. Postshock VF redetection times were significantly shorter than initial detection times (4.5 +/- 1.8 seconds detection, 2.1 +/- 0.7 seconds redetection; p<0.0001). During an 8-month mean follow-up (range 1 to 31 months), 2 unwitnessed deaths were classified as sudden cardiac deaths, and 11 patients experienced a total of 12 complications, none of which was associated with the Cadence-Endotak combination.

The implantable cardioverter defibrillator: modern implantation techniques and their impact on outcomes.

The implantable cardioverter defibrillator has revolutionized the management of lethal ventricular arrhythmias in susceptible patients. In its second decade of existence, the implantable cardioverter defibrillator has undergone significant technologic enhancements which have resulted in ease of implantation, lower mortality rates, and shorter hospital stays. The newer pectoral size devices have been successfully implanted in a variety of patients, using models from several device manufacturers. Improvements in lead technology have paralleled those of the device itself. These include the unique concept of "unipolar" defibrillation as well as the trend toward dual chamber lead systems. Results of these newer technologies are favorable: comparably low defibrillation thresholds have been reported with the newer lead configurations, with lower operative mortality. However, morbidity attached to earlier lead systems remains as high as 16%. It is anticipated that the results will further improve as shorter transvenous leads and better connector material become routinely available. Finally, the clinical outcomes in the early postoperative phase indicate fewer proarrhythmic effects leading to shorter hospital stays in patient equipped with the latest types of pectoral implants. Continued progress at the level of the patient-device interface is expected to result in every better patient acceptance and proliferation of implantable cardioverter defibrillator therapy.

Initial clinical experience with a dual lead endocardial defibrillation system with atrial pace/sense capability. United States and Canada Enguard Investigations.

Ninety-three patients underwent implants of the Telectronics Model 4211 ICD attached to the Enguard PFX endocardial defibrillation lead system. Eighty-one males and 12 females ranging in age from 25-85 years (mean = 64). Coronary disease was the substrate in the majority (88%); mean left ventricular ejection fraction was 30%. VT was present in 66%, VF in 22%, and both in 7%. Three lead configurations were used in the study: ventriculo-atrial (U1, 86%); bidirectional (B2, 12%); and ventricular to patch (U2, 2%). Mean RV pacing thresholds were 0.46 V pre- and 0.54 V posttesting, with no significant differences between the two. Mean R wave voltage was 11.05 mV pre- and 11.72 mV posttesting, also not significantly different. A subgroup of 13 patients had mean atrial pacing thresholds of 0.59 V at 0.5 msec pulse width, with mean P waves of 4.01 mV. Mean defibrillation threshold for the entire group was 10.6 J using biphasic waveforms. Defibrillation thresholds by configuration were: 399 V (U1); 475 V (U2); and 350 V (B2). All patients but one had thresholds < 550 V in at least one configuration. The 4211/Enguard system was implanted without (86%) or with (14%) a subcutaneous patch electrode. Early postoperative findings related to the ICD system include: one device circuit failure, one early lead dislodgement, and one pacing exist block.(ABSTRACT TRUNCATED AT 250 WORDS)

Cost and length of hospital stay: comparisons between nonthoracotomy and epicardial techniques in patients receiving implantable cardioverter defibrillators.

Twenty-five patients with implantable cardioverter defibrillators (ICDs) implanted intrathoracically (group I) were compared with 25 patients who underwent implant using the nonthoracotomy approach (group II). All systems were implanted by the same medical team, in the same high volume implanting center. Indications for implantation were comparable in both groups. Patient characteristics were not statistically different with the exception of age (66-group I vs 71-group II; P < 0.05). Although left ventricular ejection fractions appeared to differ (32% vs 37%, respectively), this difference was not statistically significant (P = 0.06). ICD models used in group I were: Ventritex Cadence (16), Telectronics Guardian 4211 (2), Medtronic PCD (7); in group II they were: Ventritex Cadence (15), Guardian 4211 (2), and CPI 1600 (1). Total length of hospital stay was 16 +/- 6 days for group I versus 12 +/- 5 for group II (P < 0.05). Number of postoperative days in an intensive care unit was 3.2 +/- 2.8 for group I versus 0.5 +/- 0.6 for group II (P < 0.0001). Postoperative length of stay was 8.2 +/- 3.1 for group I versus 5.7 +/- 4.4 for group II (P < 0.001). Mean total hospital charges for the entire length of stay were $72,918 +/- $26,770 in group I versus $55,031 +/- $42,870 in group II, representing a mean reduction of 21% in global costs for group II patients. These data confirm that nonthoracotomy ICD implantation in an experienced center is associated with significantly shorter hospital stays, a virtual elimination of the need for postoperative intensive care, and globally lower total hospital costs. In addition, the presence of a statistically older population in group II does not negate these beneficial effects.

Simplified subxiphoid placement of implantable cardioverter defibrillators using a posterior rectus pocket.

A simplified subxiphoid procedure using a single longitudinal epigastric incision and posterior rectus pocket for implantable cardioverter defibrillators was used in 100 patients. Through a single incision, ventricular patches are placed via a transverse pericardiotomy, and a pouch is created behind the rectus abdominis muscle in the left upper quadrant for placement of the implantable cardioverter defibrillator. Patients have minimal discomfort soon after operation, and the implantable cardioverter defibrillator generator is imperceptible to most.

Changing trends in therapy delivery with a third generation noncommitted implantable defibrillator: results of a large single center clinical trial.

Thirty-four patients underwent implantation of a third generation ICD, the 4210 ATP, for sudden cardiac death or ventricular tachycardia. This device incorporates significant telemetry logs as well as a detailed analysis of each arrhythmia episode detected. During the period of clinical follow-up, a mean of 12.2 months, a total of 26,569 VT or VF detections were made. The vast majority of these were either due to atrial fibrillation, nonsustained VT, or "noise" detection, and only 6% led to device therapy. ATP was successful in 86.3% of episodes, with 3.5% accelerations and 2.4% failure of ATP trains. The majority of inappropriate therapy episodes were clustered in seven patients, and all were easily diagnosed with the aid of the extensive telemetry logs and sense histories. Of five late deaths, three were from congestive heart failure, one from cerebrovascular accident, and one unknown. These data reveal that this "tiered" therapy noncommitted ICD performs to expectations; the stored data is of significant value in diagnosing the cause of ICD therapy. In addition, ATP is an effective modality for termination of VT.

Improved patient surveillance and data acquisition with a third generation implantable cardioverter defibrillator.

Thirteen patients were implanted with the Telectronics 4210 ATP implantable cardioverter defibrillator (ICD) for ventricular tachycardia or ventricular fibrillation. This device has multiprogrammable antitachycardia pacing, bradycardia pacing, and shock therapies. In addition, there is extensive data logging and ECG snapshot capability for arrhythmia confirmation and response to therapy. These features permit easy retrieval of all detected and treated events, whatever the eventual outcome. In this study, the data logged at predischarge electrophysiological testing was compared to the data recorded in a standard manner. The bulk of the data, however, was derived from long-term follow-up of spontaneous events over a mean period of 203 days (range 154-257). During this period, a total of 6,193 arrhythmia detections were made: 20 were classified as ventricular fibrillation, and 6,173 as ventricular tachycardia. The vast majority of these (93%) terminated spontaneously without ICD intervention (5,738), underscoring the benefit of a standard second confirmation prior to therapy delivery (noncommitted system). There were 394 arrhythmia episodes treated with antitachycardia pacing; of these a total of 8.3% accelerated to either more rapid ventricular tachycardia or ventricular fibrillation (4.3% and 4.0%, respectively). Events were reported in an "episode log" format, listing all arrhythmia detections with time/date annotation; or in a "sense history" format, detailing each episode from start to conclusion. These data demonstrate that this advanced, "tiered" ICD with data recall contributes to better patient management, and permits a more tailored termination prescription for the individual patient.

Clinical electrophysiological effects of intravenous recainam: an antiarrhythmic drug under investigation for the treatment of ventricular and supraventricular arrhythmias.

This open-label, multicenter study was designed to assess the electrophysiological properties of intravenous recainam, an investigational Class I antiarrhythmic agent. In 25 patients undergoing electrophysiological studies for the evaluation of arrhythmias, recainam was administered intravenously in a loading infusion (0.1 mg/kg/min) for 40 minutes, followed by a maintenance infusion (0.02 mg/kg/min) until the completion of the study. Electrophysiological measurements were obtained at baseline, 30 minutes after initiation of the loading infusion, and 30 minutes after termination of the infusion during washout. Conduction intervals, refractory periods, and sinus node recovery times were measured during sinus rhythm and during atrial or ventricular pacing. Vital signs were obtained and recorded before, during, and after recainam infusion. The results showed no change in mean arterial pressure, but heart rate increased slightly by 4 beats/min following recainam infusion. Recainam produced a generalized slowing of intracardiac conduction. The mean intraatrial conduction time, measured at an atrial paced cycle length of 600 msec, increased during recainam loading infusion by 44%, from 38.8% +/- 2.8 to 53.0 +/- 5.4 msec; intranodal conduction time increased by 10%, from 102.0 +/- 5.5 to 112.1 +/- 5.2 msec; and infranodal conduction time increased by 31% from 53.1 +/- 3.0 to 70.7 +/- 3.8 msec. Slowed conduction persisted during washout. The mean right atrial effective refractory period was significantly prolonged (+7% at 600 msec cycle length and +8% at 450 msec cycle length, P less than 0.05 and P less than 0.01, respectively) during recainam loading and remained so during washout.(ABSTRACT TRUNCATED AT 250 WORDS)

PR interval behavior during exercise: implications for physiological pacemakers.

The relationship between heart rate response and the dynamic changes in the PR interval was assessed in 631 patients undergoing routine cardiac exercise tests for a variety of clinical indications. Patients were stratified into four subsets: nonmedicated normals (n = 437), patients on beta-antagonist agents (n = 118), those on antiarrhythmic agents alone (n = 61) and those with a clinical diagnosis of advanced (New York Heart Association [NYHA] Class III or IV) congestive heart failure. All patients were in stable sinus rhythm throughout the test. PR intervals were measured at rest, at mid-exercise and at peak exercise. Mean PR intervals shortened to a statistically significant degree in most subgroups. This effect was predominantly observed in the earlier stages of exercise. In patients with advanced heart failure, there was no statistically significant shortening of exercise PR intervals later in exercise, demonstrating a parallel with their relatively blunted heart rate response. These changes in exercise PR intervals suggest that implanted pacemaker algorithms may be constructed to maximize hemodynamic benefit in patients requiring physiological pacemakers.

Effects of supernormal capture on directly measured nonmodulated parasystolic cycles.

Pacemaker capture during the supernormal period was seen in a case of modulated parasystole where ectopic cycles occurred without any interposed nonparasystolic beat. This contrasts with previous clinical reports, since, in the latter, the parasystolic cycle length could not be measured directly. As in experimental studies, supernormal modulation resulted in a triphasic phase response curve. Although less likely, similar electrocardiographic changes could be produced by a conceptually different phenomenon, namely full parasystolic resetting due to loss of protection occurring exclusively early in the cycle, combined with modulation in the other (late) parts of the cycle.

Time to first shock and clinical outcome in patients receiving an automatic implantable cardioverter-defibrillator.

The relation between time to first shock and clinical outcome was studied in 60 patients who received an automatic implantable cardioverter-defibrillator (AICD) from August 1983 through May 1988. The mean (+/- SD) patient age was 64 +/- 10 years, 82% were men and the mean ejection fraction was 33 +/- 13%. During follow-up, 38 patients (63%) had one or more shocks; there were no differences in age, gender distribution or ejection fraction at entry between the shock and no shock groups. Among 51 patients with coronary artery disease, 31 (61%) had one or more shocks, whereas all seven patients with cardiomyopathy had one or more shocks (p less than 0.05). Neither of the two patients with idiopathic ventricular fibrillation had shocks. Of the 13 deaths, 12 occurred during post-hospital follow-up and 1 during the index hospitalization. Of the four sudden post-hospital deaths, only one was due to tachyarrhythmia in the absence of acute myocardial infarction. All four sudden deaths and five of eight post-hospital nonsudden deaths occurred in patients who had had one or more appropriate shocks during follow-up. Eight of the nine first appropriate shocks among patients who subsequently died occurred within the first 3 months of follow-up, but the actual deaths were delayed to a mean of 14.1 +/- 13.9 months (p less than 0.05). The mean time to all deaths was 14.8 +/- 13.1 months. The ejection fraction was significantly lower among patients who died than among patients who survived (25 +/- 7% versus 35 +/- 14%, p less than 0.02), but it did not distinguish risk of first shocks.(ABSTRACT TRUNCATED AT 250 WORDS)