Survivors of ventricular fibrillation and of sustained monomorphic ventricular tachycardia (SMVT) have a high risk of arrhythmia recurrence, especially in the first year after the presenting arrhythmic event. The results from the Antiarrhythmic Versus Implantable Defibrillator (AVID) trial (1) and the Multicenter Unsustained Tachycardia Trial (MUSTT) (2) demonstrated that the implantable cardioverter defibrillator (ICD) reduced mortality compared with drug treatment. Few data, however, are available on patient symptoms associated with arrhythmia recurrence with ICD therapy (3- 4), the results of which could influence the ability to resume higher risk activity such as driving (5). Some studies have shown a lower incidence of first ICD therapy after six to 12 months (5- 6), especially in patients with left ventricular ejection fraction (LVEF) above 25% (7). Importantly, little is known about the probability of a second arrhythmic event and symptoms. Thus, the purpose of our study was to investigate which clinical or electrophysiologic characteristics could predict if and when first and second ICD therapy would occur and whether initial symptoms correlate with subsequent symptoms associated with ICD therapy.

To exclude inappropriate shocks and to verify that therapy delivered to an asymptomatic patient was appropriate, only patients with an ICD with interval (Ventak PRx Model 1700, 1705, 1710 and 1715, CPI, Minneapolis, Minnesota) or electrogram recording capability (Cadence Model V100, V100B, V100C, Ventritex, Sylmar, California, and Ventak PRx Model 1715, Guidant, Indianapolis, Indiana) were included. The records of 125 consecutive patients who had undergone implantation using one of these ICDs for symptomatic ventricular tachycardia (VT) or cardiac arrest and who were followed by one of the authors were examined. Patients in whom an ICD was implanted prophylactically were excluded from the study. Baseline data for each patient included gender, age, type of heart disease, history of myocardial infarction, history of congestive heart failure, past history of coronary artery bypass grafting, LVEF, electrogram QRS duration and signal-averaged electrogram. Type of presenting arrhythmia, cardiac arrest (CA) or SMVT was determined. For patients with SMVT, the arrhythmia cycle lengths and symptoms were noted.

The electrophysiologic study (EPS) procedure has been reported in detail (8). In patients with hemodynamically stable SMVT, serial electrophysiologic-pharmacologic testing was done. An ICD was implanted in patients who did not have adequate response to the drugs tested. Drug therapy at the time of the presenting arrhythmia and at discharge was noted.

In a predischarge EPS, patients with hemodynamically stable SMVT had antitachycardia pacing tested and enabled only if the SMVT was reproducibly terminated.

After hospital discharge, patients were seen by one of the authors (ENP, RIF) every two to six months and also after a clinically apparent arrhythmic event. At each visit a history and physical examination were done, and the ICD was interrogated for new arrhythmic events. When such an event was noted, the patient activity and symptoms at that time were recorded. Appropriate ICD therapy was identified on the basis of device interrogation and clinical information.

In patients not treated with an antiarrhythmic drug (AAD) before the first ICD therapy, initiation of an AAD was deferred until after the next ICD therapy. Similarly, patients receiving an AAD before the first ICD therapy continued with the same AAD.

The chi-square test or the Fisher exact test was used for analysis of categorical variables. The Student t test or analysis of variance was used for analysis of continuous variables as appropriate. Actuarial survival and freedom from arrhythmia were computed by the Kaplan-Meier product limit method (9). Survival and freedom from arrhythmia between two groups were compared using the log-rank test (10). Multivariate analysis was done by the Cox proportional hazard model (11). Statistical computations were done with Statistica for Macintosh software.

One hundred twenty-five patients with an ICD with interval or electrogram recording capability were followed at the Care Group arrhythmia clinic for 408 ± 321 (range, 1 to 1,277) days. Five patients died one to 188 days after implantation (one- and three-year survival of 95%). All deaths were due to cardiac causes, and no death was sudden.

Thirty-nine patients (31%) were treated with AAD: 14 (36%) for atrial fibrillation (AF), 12 (31%) for VT slowing at EPS and 13 (33%) for suppression of frequent nonsustained VT. Twenty-seven patients (70%) were treated with amiodarone, four with mexiletine, three with propafenone, two each with quinidine and sotalol, and one with procainamide.

Fifty-eight patients (46%) had an initial ICD therapy after 152 ± 193 (range, 1 to 896) days. Actuarial freedom from initial ICD therapy is shown in (Figure 1). Clinical characteristics of patients who had ICD therapy are shown in (Table le2). Patients with LVEF ≤25% had one- and two-year freedom from ICD therapy of 39% and 31% versus 57% and 52%, respectively, for patients with LVEF >25% (p = 0.024, log-rank test) (Figure 2). In patients presenting with SMVT versus CA, one- and two-year freedom from ICD therapy were 43% and 34% versus 62% and 52%, respectively (p = 0.0045, log-rank test) (Figure 2). Patients treated with AAD also had more ICD therapies, with one- and two-year freedom from ICD therapy of 44% and 24% versus 53% and 50%, respectively (p = 0.039, log-rank test) (Figure 2). In contrast, treatment with a beta-blocker did not predict ICD therapy.

In age- and gender-adjusted multivariate analysis, presenting arrhythmia and LVEF ≤25% (but not AAD therapy) were independent predictors for initial ICD therapy, with odds ratios (95% confidence interval) for presenting arrhythmia of SMVT 2.573 (1.321 to 5.013) and for LVEF ≤25% 1.953 (1.107 to 3.447) (p < 0.05). Age, gender or type of heart disease did not predict the time from implant to first ICD therapy.

In a follow-up period of 329 ± 295 (range, 7 to 1,179) days after the initial ICD therapy, only 12 patients (21%) remained free of further ICD therapy. No patient had ADD therapy changed between initial ICD therapy and second ICD therapy.

There were no significant differences in clinical characteristics between patients who had one or more than one ICD therapy.

(Figure 1) shows Kaplan-Meier analysis of the period from the first to second ICD therapy. The median actuarial freedom from ICD therapy for the second shock was 22 days, and all second shocks occurred within one year after the initial ICD therapy. There were no significant differences in the clinical characteristics of the nine patients with recurrence within 48 h compared with the 37 patients who had subsequent ICD therapy >48 h later. In patients in whom subsequent ICD therapy occurred after >48 h, mean time to second ICD therapy was 66 ± 93 days compared with 138 ± 168 days for first ICD therapy (p < 0.05). None of the clinical predictors that predict first ICD therapy was useful for the prediction of the next ICD therapy. Similarly, the actuarial freedom from subsequent ICD therapy could not be predicted by any clinical characteristics of the patients who had the first ICD therapy. No correlation was found between time to the first and second ICD therapy.

The symptoms during first ICD therapy are summarized in (Table le3). Only 9% of the patients who presented with SMVT had syncope during the first ICD therapy, versus 36% of the patients who presented with CA (p < 0.05). However, 22% of patients who presented with SMVT had near syncope during first ICD therapy, versus none of the CA patients (p = 0.09).

Ventricular tachycardia cycle length induced at baseline EPS was shorter than 250 ms in five of six patients in whom first ICD therapy was associated with syncope (83%), compared with 10 of 41 patients who had no syncope (24%) (p < 0.005). Ventricular tachycardia cycle length ≤250 ms at baseline EPS was noted in five of 12 patients (58%) who had impaired consciousness (near syncope or syncope), versus eight of 35 (23%) patients who did not (p < 0.05). Presenting symptoms in patients with SMVT did not correlate with symptoms during first ICD therapy (Figure 3), nor with any clinical variable including LVEF or the presence of AF.

In contrast, patients tended to have similar symptoms with the first and second ICD therapy, respectively (p = 0.0001): syncope (6 vs. 3 syncope with the first ICD, 2 near syncope, 1 asymptomatic with the second), near syncope (8 vs. 6 near syncope, 2 asymptomatic), other symptoms (2 vs. 2), asymptomatic (30 vs. 28 asymptomatic, 2 syncope) (Figure 4). In patients whose initial arrhythmia was SMVT, only one of 25 patients who were asymptomatic with the first ICD therapy had syncope with the second ICD therapy (p < 0.0001). One of five patients with CA who were asymptomatic on the first ICD therapy had syncope on the second ICD therapy (p = 0.07). The cycle length of VT preceding first and second ICD therapy was similar (r = 0.75; p = 0.0001) (Figure 5).

This study evaluated whether clinical or electrophysiologic characteristics associated with patients who had first ICD therapy added prognostic information regarding recurrence and symptoms of subsequent ICD events. The significant new observations noted in this study are as follows: 1) subsequent ICD therapy occurs sooner than first ICD therapy, and neither time nor recurrence can be predicted by any clinical variable; 2) symptoms during first ICD therapy predict subsequent symptoms; and 3) patients presenting with SMVT and asymptomatic first ICD therapy are at very low risk for future syncopal ICD therapy.

Our policy has been to delay prescribing AAD after the first ICD therapy, because the frequency of recurrence of sustained VT is not known for a particular patient. Importantly, the time to the second ICD therapy was therefore not influenced by change in drug therapy in this study. Additional ICD therapy was observed in 79% of the patients after the initial event. The second ICD therapy, however, occurred much sooner than the first, and the time to first therapy did not predict the time to second therapy. The median actuarial ICD therapy-free survival to the second ICD therapy was only 22 days (46 days excluding patients who had second ICD therapy within 48 h), versus 348 days for the first ICD therapy. In contrast to the initial ICD therapy, actuarial freedom from subsequent ICD therapy of patients presenting with CA was not significantly different from patients presenting with SMVT. Similarly, patients treated with antiarrhythmic drugs before initial hospital discharge did not have higher incidence of second ICD therapy compared with first ICD therapy. Finally, symptoms during the first ICD therapy did not predict incidence or actuarial freedom from second ICD therapy. Thus, an initial ICD therapy per se selects patients who will have an approximately 80% chance of another ICD therapy within one year. That incidence is much higher than seen with any of the clinical variables that predict the first ICD therapy. No clinical predictor was found for very early (<48 h) second ICD therapy.

Although a recent paper suggested that treatment with sotalol may reduce the frequency of ICD therapy (12), the influence of adding or changing AAD therapy at VT recurrence on symptoms is unknown. Our data, however, showed that ICD therapy without alteration of AAD treatment occurs early and is not predictable. Thus, one should consider initiation of a new AAD regimen soon after the first ICD therapy, especially when therapy is a direct current shock and accompanied by syncope or near syncope (3,12).

During first ICD therapy, 14% of the patients had syncope and 17% near syncope. Impaired consciousness (syncope or near syncope) occurred in 30% of patients who presented with SMVT and in 36% of CA patients. However, patients who presented with SMVT tended to have near syncope whereas patients who presented with CA had syncope (p < 0.05). No other clinical baseline variable or presenting symptom predicted symptoms of first ICD therapy. Inducible VT cycle length ≤250 ms at EPS did, however, predict syncope and impaired consciousness at subsequent arrhythmic event.

Importantly, symptoms during the first ICD therapy correlated highly with symptoms of the next ICD therapy (p = 0.0001). Only two of 30 patients who were asymptomatic at the first ICD therapy had syncope with the next ICD therapy. In patients presenting with SMVT and initial asymptomatic ICD therapy, the probability of syncope on the second ICD therapy was only one in 25 (4%). The fact that we did not change AAD therapy or device setting after the first ICD therapy, and the observation that tachycardia cycle length during first and second ICD therapy was similar, may explain the consistency of symptoms between ICD therapies.

Our data differ from those of Bansch et al. (3), who found risk factors for syncope to be LVEF ≤40%, chronic AF and inducible VT with a cycle length <300 ms. However, our patient population differs significantly from their study (age 66 ± 8 years vs. 58 ± 13, LVEF 29 ± 15 vs. 43 ± 17, coronary artery disease in 84% vs. 62%, inducible VT cycle length >300 ms in 32% of VT induced vs. 46%, respectively) Our study population is similar to several other major ICD studies (1- 2,13- 14). Additionally, more than half of Bansch’s patients had a change in therapy after the first event, either by addition of drugs or by alternation of ICD settings (3), which may have prevented subsequent syncope.

Using strict criteria for appropriate ICD therapy, we found that 58 patients (46%) had an initial ICD therapy after 152 ± 193 (range, 1 to 896) days. This is in agreement with most other studies (7,15- 16). Patients who had ICD therapy tended to have LVEF ≤25% (63% vs. 42%, p = 0.025), as noted in some (7,17- 22) but not other studies (5,15,23); patients more commonly had SMVT (81% vs. 58%, p = 0.006), as previously reported (15). Type of heart disease, age and gender did not predict ICD therapy or time from implant to first ICD therapy. Most of the ICD therapy in our study occurred in asymptomatic patients.

To exclude inappropriate shocks and to verify that the therapy delivered to asymptomatic patients was appropriate, only patients who were initially implanted with an ICD having electrogram or RR interval recording capabilities were included in our study. However, even with intracardiac electrograms, occasional supraventricular tachycardia masquerading as VT cannot be ruled out (24- 25). Patients can also have a change in cardiac function that may alter symptoms with recurrent VT or ventricular fibrillation. In this situation, a previously hemodynamically stable VT may result in syncope.

One cannot predict the effect of adding an AAD on subsequent symptoms. Although a slower VT may occur during AAD treatment, symptoms may be unchanged or even worsen. Thus, after a change in AAD therapy close observation and restriction from driving appear warranted.

After an ICD is implanted, arrhythmia recurrence tends to occur early in patients treated with AAD who have a history of SMVT and LVEF ≤25%. Because a second ICD therapy occurs soon and is unpredictable, we recommend early consideration for AAD therapy for patients who have significant symptoms with the initial ICD event and restriction of driving (12). A recent policy statement suggests that driving be restricted for six months after initial ICD implant (26). However, if the first ICD therapy is asymptomatic in a patient with SMVT with a stable disease state, chances of syncope or presyncope with subsequent ICD therapy is minimal, providing no change in AAD therapy is made. In these patients it may be reasonable to consider resumption of driving with appropriate precautions.