Experience with pectoral versus abdominal implantation of a small defibrillator

Transcription

1 European Heart Journal (1998) 19, Article No. hj9816 Experience with pectoral versus abdominal implantation of a small defibrillator A multicenter comparison in 778 patients E. Hoffmann, G. Steinbeck, for the European Jewel Investigators* Aims The aim of the study was to examine the results of implanting small sized cardioverter defibrillators in the pectoral as opposed to the abdominal area. Hitherto, owing to the large size of the early defibrillators, the site of implantation had been confined to the abdomen. Methods Between 3 March 1993 and 1 November 1994, 778 patients from 63 centres in 14 countries underwent their first device implantation. The study was set up to evaluate the safety and the efficacy of Medtronic models 7219 D, a multi-lead abdominal/pectoral implantable cardioverter defibrillator, and 7219 C, a pectoral single-lead Active Can implantable cardioverter defibrillator. There were 155 abdominal and 623 pectoral implants. Survival data were comparable during a mean follow-up period of months, with no difference regarding the pectoral placement of single (n=392) or multi-lead (n=231) devices. The only significant difference was related to severe lead-related events: 5 3% in the pectoral vs 11 6% in the abdominal group (P< 5). These events were mainly related to lead dislodgement. Kaplan Meier estimates showed that both single and multilead systems, in either the pectoral or abdominal position, demonstrated a similar severe adverse event-free survival. Conclusion These findings suggest that an implantable cardioverter defibrillator (18 mm thick, 8 cc volume, 129 g weight) can be implanted in the pectoral position without an increase in clinically relevant adverse events compared to abdominal implantation. implantation was associated with a significantly reduced lead-related severe adverse event rate. (Eur Heart J 1998; 19: ) Key Words: Implantable cardioverter defibrillator, pectoral implantation, abdominal implantation, ventricular tachyarrhythmia Introduction Patients with ventricular tachycardia or ventricular fibrillation have a poor long-term outcome due to the high rate of recurrence of these potentially life threatening arrhythmias, but implantable cardioverter defibrillators are an effective mode of therapy for secondary prevention [5,14,22]. A major innovative step in the development of implantable cardioverter defibrillators was the introduction of a non-thoracotomy lead system, which dramatically reduced procedure-related mortality and morbidity, as compared to previous epicardial electrode systems [1 3,1,16,19]. However, until recently, due to the large size of the implantable cardioverter defibrillator, pulse generators had to be implanted in an abdominal Revision submitted 19 January 1998, and accepted 25 January *Participants are listed in the Appendix. Correspondence: Ellen Hoffmann, MD, Department of Medicine I, Grosshadern Hospital, Ludwig-Maximilians- University, Marchioninistr. 15, Munich, Germany X/98/ $18./ pouch. Now, the availability of smaller devices allows for pectoral implantation, similar to that of pacemakers, which substantially facilitates the implantation procedure and reduces the surgical trauma. The widespread use of this simplified implantation procedure in most European centres, in the absence of definite data on the long-term clinical outcome, suggested the need for a formal evaluation. We present the results of a European multicentre observational clinical trial. The objectives were to compare the efficacy and safety of pectoral as compared to abdominal implantation of a first small pulse generator in a large patient population. Patients and methods The present report comments on 778 patients who received an implantable cardioverter defibrillator between March 1993 and October Sixty-three investigators from 14 countries have collaborated in this prospective observational study (see Appendix). Patients were eligible for implantation if they had clinically significant sustained ventricular tachycardia, had 1998 The European Society of Cardiology

2 186 E. Hoffmann et al. survived cardiac arrest caused by fast ventricular tachycardia or ventricular fibrillation, or if they had suspected arrhythmogenic syncope with inducible ventricular tachycardia/ventricular fibrillation at programmed electrical stimulation of the right ventricle. Patients were excluded from the study if there was a reversible cause for ventricular tachycardia/ventricular fibrillation or if acute myocardial infarction occurred. With the availability of the small pulse generator in March 1993, it was the choice of the investigator to implant in the abdominal or pectoral position, without a consistent pre-selection of patients; only new implants are included in this report. All patients took part in the clinical evaluation of the Medtronic models 7219 D (Jewel) or 7219 C (Jewel Active Can ) implantable cardioverter defibrillators (Medtronic Inc., Minneapolis, U.S.A.), conducted in accordance with the Declaration of Helsinki. Approval was obtained from the local ethics committees, and every patient gave informed consent before participation. In each clinical centre, a principal investigator ensured that the protocol was followed, and that eligible patients were considered for the study. Between the first implant on 3 March 1993 and the final patient inclusion in October 1994, 778 devices were implanted, 623 in the pectoral and 155 in the abdominal position. The lead systems were transvenous in 772 patients. Six patients received epicardial patches. Data obtained from the individual investigators were submitted to the Bakken Research Center, Medtronic, Maastricht, The Netherlands, and centrally documented. All major participating centres had an audit for study protocol compliance. Adverse clinical events and deaths were analysed by an independent adverse event committee of four cardiologists. Electrophysiological evaluation and device implantation Patient evaluation before implantable cardioverter defibrillator implantation included coronary angiography, measurement of left ventricular ejection fraction and programmed ventricular stimulation. The study protocol did not pre-define the methods used for these assessments. The results were recorded on the study forms, as well as details on any concomitant antiarrhythmic treatment. The standard pectoral multi-lead system consisted of a 7 French transvenous silicone superior vena cava or coronary sinus lead (model 6933, Medtronic) and a tripolar right ventricular screw-in lead (model 6936, Medtronic) with a 5 cm coil for defibrillation and a bipolar electrode pair for pacing and sensing. A model 7219 D pulse generator was used, weight 132 g, volume of 83 cc 3. Defibrillation was performed between the defibrillation electrodes in the right ventricular and superior vena cava leads. If necessary, a subcutaneous patch was added to the defibrillation system. The pectoral single-lead system was composed of the same right ventricular lead and a model 7219 C pulse generator of 8 cc 3 and 129 g, featuring an electrically active titanium housing with a surface area of 18 cm 2 used for defibrillation. For abdominal implantation the model 7219 D pulse generator with the non-active shell was used. The leads had the same electrical configuration as the pectoral multi-lead system, but were longer (11 cm). Abdominal implantable cardioverter defibrillator devices were typically implanted by surgeons in the operating theatre, whereas the pectoral devices could be implanted by cardiologists in the cardiac catheterization laboratory. Patients were under general anaesthesia for all abdominal implants; pectoral implantation could be performed under intravenous analgesia and conscious sedation. The transvenous leads were inserted via a subclavian vein puncture using the Seldinger technique, or via a cephalic vein cut-down. The pectoral pocket for the pulse generator was prepared either subcutaneously or subpectorally; abdominal implantable cardioverter defibrillators were typically implanted behind the rectus sheath. Following the insertion of the transvenous leads, ventricular fibrillation was induced by 5 Hz stimulation, low-energy T-wave shock or programmed electrical stimulation. The criterion for implantation was fulfilled if two ventricular fibrillation episodes were successfully terminated with a 24 J monophasic or biphasic shock. As part of the defibrillation testing protocol, the lowest energy for successful defibrillation was recorded, decreasing the shock energy in 6 J steps from 24 J. A true defibrillation threshold was defined as the lowest shock that successfully terminated ventricular fibrillation with one lower shock failure, using a step-down protocol of 24 J, 18 J, 12 J, 6 J. The defibrillation threshold was determined only if the safety of the patient allowed more than two ventricular fibrillation inductions. The final decision on whether or not to implant a device was left to the physician and depended on the clinical condition of the patient. Automatic ventricular fibrillation detection by the device requires that at least 75% of a consecutive number of detected R-R intervals are shorter than the programmable fibrillation detection interval. Four subsequent defibrillation therapies can be programmed, with energies from 4 to 34 J. Typically, for safety reasons, the maximum available energy of 34 J was programmed for defibrillation. Automatic ventricular tachyarrhythmia detection requires that a consecutive number of programmable intervals is shorter than the programmable ventricular tachyarrhythmia detection interval. Intervals shorter than the fibrillation detection interval are ignored for ventricular tachyarrhythmia detection. Typically, one of two different modes of anti-tachycardia pacing was chosen by the physicians: burst, a sequence of five to eight stimuli with a cycle length shorter than the ventricular tachycardia cycle length; ramp, a sequence of impulses, with a decreasing cycle length within the sequence [5,11,18]. Shock delivery synchronized with the R-wave (cardioversion), with energy programmable from 4 to 34 J, was another

3 Implantation of a small defibrillator 187 option for ventricular tachyarrhythmia termination. Fast ventricular tachycardia was programmed if a patient had more than one ventricular tachycardia with different cycle lengths requiring different therapies. Antibradycardia VVI pacing was activated in all patients at a pacing frequency below the patient s sinus frequency to prevent long post-shock pauses. A sensitivity of 3 mv was recommended for sensing of ventricular activity. Assessment of outcome The primary outcome measure was all-cause mortality. Survival comparisons were conducted between pectoral and abdominal systems. The secondary outcome measures were sudden cardiac, non-sudden cardiac and non-cardiac deaths (according to standardized criteria of implantable cardioverter defibrillator patient outcome) [12], ventricular tachyarrhythmias requiring device therapy and adverse events. All outcome parameters were recorded by the investigators during the initial hospital stay, including implantable cardioverter defibrillator implantation and evaluation prior to hospital discharge (in most cases 1 week post implantation, including a pacing test and if possible ventricular fibrillation and ventricular tachycardia detection and therapy), follow-up visits at 4 and 12 weeks and every 3 months from then on. At follow-up, pacing thresholds were tested and the implantable cardioverter defibrillator memory was interrogated with respect to tachyarrhythmia events. The intervals and electrocardiograms (2 5 1 s) of a maximum of five arrhythmias could be stored in the implantable cardioverter defibrillator. The adverse event committee reviewed and validated evidence for all causes of death and adverse events. Patients were followed from the time of study entry to a common cut-off date, 1 January The mean follow-up time was months. One hundred and seventy patients (22%) had a follow-up time of more than 6 months; 82 patients (11%) had a follow-up time of more than 12 months. For both efficacy and safety analysis, the total pectoral group and the subgroups of single and multi-lead systems were compared to the abdominal implantation group. A detailed comparison of adverse events between pectoral and abdominal implantation was performed. Definitions of adverse events The study was performed in accordance with the European standard for conducting clinical investigations of medical devices for human subjects, EN54 [9]. Thus, all adverse events, defined as any undesirable clinical occurrence in a subject (EN54) were monitored and classified as mild or severe, and as device related or not device related. An adverse event is severe when as a result a subject has to be hospitalized, or their hospitalization is unduly prolonged because of potential disability or danger to life or because an intervention has been necessitated or the event is terminal (EN54); in the present study, any adverse event that is not severe, is mild. Device-related adverse events were further classified according to the device-related cause (shock-related, lead-related, pocket-related, inappropriate therapy or other). Failure to terminate ventricular tachycardia/ ventricular fibrillation by the device and shock induction of arrhythmia were classified as shock related. Right ventricular perforation (with or without tamponade), lead dislodgement, difficulty with lead fixation, inappropriate connection, pulmonary embolism, thrombosis, haemothorax, increased defibrillation threshold and decreased sensing capability were all classified as lead related. Infection or haematoma in the implantable cardioverter defibrillator pocket, failure of the pocket to close, migration of the implantable cardioverter defibrillator and reduced shoulder mobility were classified as pocket related. A review of the adverse events was presented at the annual scientific session of the North American Society of Pacing and Electrophysiology [4]. Since interventional therapy is assessed, the occurrence rate of events requiring surgical intervention (excluding deaths) during follow-up is also determined. Statistical analysis Quantitative variables are reported as mean SD values; for comparisons, the t-test and the Wilcoxon rank-sum test have been used. Qualitative variables are reported as a percentage or number of occurrences; for comparisons, the Pearson chi-square and the Fisher exact test have been used. Time until the first occurrence of a specific event (death, surgical intervention, adverse event, spontaneous episode, etc.) is described by means of Kaplan Meier survival estimates and comparisons are done by means of a log-rank test. Results From March 1993 to October 1994, 778 patients were enrolled in 63 centres in 14 countries. Twenty-five centres performed at least 1 implantations, accounting for 83% of the procedures. After the first implantation in March 1993, the majority of the 385 multi-lead systems were implanted in the pectoral position (n=231, 6%); 154 (4%) devices were implanted in an abdominal pocket. Between March and December 1993, a total of 6 centres implanted multi-lead systems during the study. The 392 single-lead pectoral systems were implanted between November 1993 and October 1994 at 42 centres, with 15 centres performing at least 1 implantations, accounting for 73% of the implantations.

4 188 E. Hoffmann et al. Table 1 Baseline characteristics, including underlying heart disease and indication for implantable cardioverter defibrillator implantation Abdominal Multi-lead n=155 Total n=623 Multi-lead n=231 Single-lead n=392 Total n=778 Mean (SD) Age (years) 54 16* Weight (kg) 73 14* Height (cm) Left ventricular ejection fraction (%) 42 18* No (%) of patients Sex Male 119 (77) 515 (83) 194 (84) 322 (82) 635 (82) Female 36 (23) 17 (17) 37 (16) 7 (18) 143 (18) Underlying heart disease Coronary artery disease 68 (44)* 383 (61) 139 (6) 244 (62) 451 (58) Old myocardial infarction 72 (46)* 35 (56) 134 (58) 216 (55) 422 (54) Cardiomyopathy 57 (37)* 175 (28) 63 (27) 112 (29) 232 (3) NYHA class I 51 (33)* 125 (2) 44 (19) 81 (21) 176 (23) II 67 (43)* 348 (56) 14 (61) 28 (53) 415 (53) III 32 (21) 148 (24) 47 (2) 11 (26) 18 (23) not determined 5 (3) 2 (1) () 2 (1) 7 (1) Indication for implantable cardioverter defibrillator implantation Ventricular tachycardia 63 (41) 237 (38) 19 (47) 154 (39) 35 (39) Sudden cardiac death 68 (44) 275 (44) 83 (36) 166 (42) 338 (43) Ventricular tachycardia and sudden cardiac death 14 (9) 82 (13) 3 (13) 52 (13) 96 (12) Other 1 (6) 28 (5) 8 (3) 2 (5) 38 (5) Differences between the abdominal and the total pectoral group are indicated *P< 5 (t-test for continuous variables, Pearson chi-square for categorical variables). One patient was implanted with a single-lead defibrillation system in the abdominal position. Implanting centres carried out the procedures thus: in 4% of the investigational centres, all implants were done in the pectoral region; in 14% of hospitals, implants were done in the abdominal position; in 46% of centres, both pectoral and abdominal implantations were performed. Patient characteristics The clinical characteristics of the patients are comparable and summarized in Table 1. Eighty-two percent of the study population were men and the mean age was years. Significant differences between the total pectoral group and the total abdominal group were found with respect to age and weight. patients were significantly older (58 12 years pectoral vs years abdominal; P= 3) and heavier (77 12 kg pectoral vs kg abdominal; P= 11). In the study group as a whole, 58% of the patients had coronary artery disease as the underlying disease with a significantly higher incidence in the pectoral as compared to the abdominal group (61% vs 44%; P= 1); 54% of the patients had a history of previous myocardial infarction, again more often in the pectoral group (56%) than in the abdominal group (46%; P= 3); 3% of the patients had cardiomyopathy, but this was consistently less frequent in the pectoral group (28%) than in the abdominal group (37%; P= 34). The mean left ventricular ejection fraction was 39 17% in the total study population, with a significantly lower value for the pectoral group (38 16%) than for the abdominal group (42 18%; P= 1). There was a significant difference between the groups (P= 1) with respect to severity of heart failure where NYHA classes II and III were more frequent in the pectoral group (II: 56% vs 43%; III: 23% vs 19%) and NYHA class I was more frequent in the abdominal group (33% vs 2%). Differences between abdominal and pectoral implants can be, to a large extent, explained by the difference in age. The percentage of young patients (<4 years) was significantly higher (21%) in the abdominal than in the pectoral group (9%; P= 1). The cut-off value of 4 seemed the most appropriate because the average age of patients under 4 and over 4 were almost the same for the pectoral and abdominal groups. An age-related analysis showed that young patients from both groups had a significantly better ejection fraction, a lower NYHA class and suffered less frequently from coronary artery disease and myocardial infarction (P< 1). After correction for age, a difference between the pectoral and the abdominal group with respect to ejection fraction, NYHA and history of an old myocardial infarction was abolished. The differences with respect to coronary artery disease and cardiomyopathy were less significant (P= 4 instead of P= 1 and P= 49 instead of P= 34). The weight difference is also partly explained

5 Implantation of a small defibrillator 189 by the difference in age. For patients under 2 years (16 in total), the mean weight was significantly higher in the pectoral group as compared to the abdominal group (72 6 kg pectoral vs 55 3 kg abdominal; P= 6). For patients over 2 years, there were no significant differences between the two groups (77 5 kg vs 74 9 kg; ns). The primary indication for which patients received an implantable cardioverter defibrillator was ventricular fibrillation and aborted sudden cardiac death in 43% of the patients, spontaneous, recurrent ventricular tachycardia in 39% of the patients and both ventricular tachycardia and aborted sudden cardiac death in 12%, with no significant difference between the pectoral and abdominal groups. Electrophysiological data Prior to 142 abdominal (92%) and 57 pectoral (91%) implantations, invasive electrophysiological studies were performed. In the majority of these patients (588, 83%) ventricular tachyarrrhythmias were inducible and only in 124 patients (17%) was arrhythmia not inducible. Non-sustained ventricular tachycardia was induced in 65 (9%) of all patients, and sustained monomorphic ventricular tachycardia in 44 patients (62%), with a significantly higher incidence in the pectoral (58%) vs the abdominal group (49%; P= 35). This finding is consistent with the higher number of previous myocardial infarctions, the arrhythmogenic substrate for monomorphic ventricular tachycardia. The abdominal group consists of significantly younger and healthier patients with a better ejection fraction, a lower NYHA class, less frequent coronary artery disease and myocardial infarction than the pectoral group. In addition, the abdominal group more often received antiarrhythmic drugs during electrophysiological studies than the pectoral group (39% vs 34%; P= 39), with no difference concerning amiodarone treatment. Polymorphic ventricular tachycardia was induced in 9% of the patients, ventricular flutter in 3% and ventricular fibrillation in 17%, with no differences between the groups. In 59 patients (8%) both sustained ventricular tachycardias and ventricular fibrillation were induced at electrophysiological studies. Implantation data The study design was such that it was difficult to compare implant success for the multi-lead and the single-lead systems. If, for an intended multi-lead implantation, a system consisting of two leads had an insufficient defibrillation safety margin (a non-active Can failure), a subcutaneous or epicardial patch could be added to the system and the non-active Can implantable cardioverter defibrillator could be implanted anyway. However, if for an intended single-lead Active Can implantation a normal system with one lead had an insufficient defibrillation safety margin (an Active Can failure), a switch to a multi-lead non-active Can system had to be made. Information about the number of right-sided pectoral implants of 7219 C devices are not available; however, in six patients the lead-introduction side was the right subclavian vein and therefore the devices may have possibly been implanted on the right side. Of the 385 primary multi-lead implantations, 71 patients (18%) received a system other than the standard 6933/6936 combination. Of these, 58 patients required a subcutaneous and six patients an epicardial patch. In three of the six patients the epicardial patches were implanted after failure to obtain satisfactory defibrillation thresholds with transvenous leads, in the other three patients the patches were part of the first system that was tested. During the study, nine Active Can failures were reported. One patient received the Active Can in the abdominal position, with a right ventricular lead for pacing and sensing and a subcutaneous array lead for defibrillating. The other 392 Active Can patients all received a single-lead system. Figure 1 illustrates that both the lowest energy tested and the defibrillation threshold were significantly lower in the pectoral vs the abdominal group and in the pectoral single-lead vs the multi-lead group (P< 3). The number of tested lead systems was also significantly higher in the abdominal than in the pectoral single-lead group. In the Active Can group, there were more successes in the first lead configuration tested (93%) as compared to the multi-lead group (54%; P= 1). Defibrillation threshold criteria for implantation were met for the Active Can system in 95%, as compared to 88%, for the multi-lead systems (P= 1). The mean shock impedance at implantation was significantly lower in the pectoral group with Ohms (single-lead 52 1, multi-lead Ohms, P= 1) as compared to Ohms for the abdominal group (P= 1). Implantation duration ( skin-to-skin time) was significantly shorter in the single-lead group as compared to the abdominal and the multi-lead pectoral implantations (P< 3; Fig. 1). The mean number of hospital days from implantation to discharge was days for pectoral (single-lead , multilead days) and significantly shorter than the days for abdominal implantations. The mean number of hospitalization days was for the total study population. Concomitant antiarrhythmic medication At the time of implantation 398 (51%) of the patients were receiving antiarrhythmic drug therapy, with equal distribution between the abdominally and pectorally implanted devices for any of the drug classes according to the Vaughan Williams classification (Fig. 2). Class III

6 19 E. Hoffmann et al. Defibrillation threshold (Joule) Lowest energy tested (Joule) Implantation duration (min) Abdominal Abdominal 5 Abdominal * * * total total total multi-lead multi-lead multi-lead antiarrhythmic drugs accounted for the majority of these treatments with about half of the patients receiving sotalol (155, 19 9%) or amiodarone (192, 24 7%), respectively. At the pre-hospital discharge test, usually one week after implantation, antiarrhythmic drugs were used by 251 patients (32%). The decreased use of sotalol (used by 122 patients, 16%) and amiodarone (used by 58 patients, 7%) is most marked. However, a slight but non-significant increase in sotalol therapy (17 4%) could be observed during follow-up. Following an initial reduction of antiarrhythmic drug treatment postimplantation, during follow-up 284 (36 5%) of the patients continued to receive a single antiarrhythmic drug. Only 5% of the patients in the total study population received beta-blockers, with a slight increase before hospital discharge to 5 7% and during follow-up to 6 3%. * * * single-lead single-lead single-lead Figure 1 Implantation data: defibrillation threshold, lowest energy tested and duration of implantation. *P< 3 between abdominal/pectoral total group and between pectoral multi-lead/pectoral single-lead group. Implantable cardioverter defibrillator therapy efficacy A pre-hospital discharge test, including non-invasive induction of ventricular tachyarrhythmia and termination by the device, was documented in 593 (95%) of the patients in the pectoral group and 142 (92%) in the abdominal group. Five hundred and eighty two (75%) of all patients completed 4 weeks of follow-up, 332 (43%) 3 months and 17 (22%) more than 6 months from the date of implant. The mean follow-up for the whole study population was months (range 1 day to 21 months), with documented data of last follow-up available at evaluation for more than 95% of the patients (pectoral 64, 97% and abdominal 147, 95%). At the last follow-up, in 58 of 751 patients (68%) ventricular tachycardia detection and in 422 patients (56%) ventricular tachycardia therapies were programmed. Antitachycardia pacing was the first ventricular tachycardia therapy to be activated in 376 of 422 patients; the remaining 46 patients had cardioversion as first ventricular tachycardia therapy. In the remaining patients, only defibrillation therapy was programmed. A total of 374 episodes was detected: 3354 ventricular tachyarrhythmia episodes were appropriately treated and 226 spontaneously aborted. The percentage of successfully terminated episodes was 96 7% for all episodes; 95 9% for episodes of ventricular tachycardia, 96 5% for fast ventricular tachycardia and 99 % for ventricular fibrillation. Comparison of pectoral (single/multi-lead systems) and abdominal implantable cardioverter defibrillators revealed no differences in the distribution of the 2649 ventricular tachycardia episodes, the 828 ventricular fibrillation episodes, the 227 fast ventricular tachycardia episodes, or the total number of ventricular tachyarrhythmia episodes (Fig. 3). Two hundred and one patients (28%) experienced at least one appropriately treated ventricular tachyarrhythmia, 11 patients (14%) experienced at least one episode of ventricular tachycardia successfully terminated by anti-tachycardia pacing, 33 patients (4 2%) had at least one shock-terminated episode of fast ventricular tachycardia and 133 patients (17%) of ventricular fibrillation (Fig. 4). In the abdominal group fewer patients had ventricular tachycardia episodes (15% vs 18%) and more patients had ventricular fibrillation episodes (19% vs 17%). This can be explained by differences between underlying disease in these two groups. There were significantly more pectoral patients with programmed ventricular tachycardia detection (71% pectoral vs 55% abdominal; P= 1), with activated ventricular tachycardia therapies (62% pectoral vs 36% abdominal; P= 1) and with programmed ATP as first ventricular tachycardia therapy (55% pectoral vs 31% abdominal; P= 1). The number of episodes per patient was higher on average in the abdominal group as compared to the pectoral group (ventricular tachycardia 25/pt vs 19/pt; ns; ventricular fibrillation 7/pt vs 6/pt; ns; all episodes 19/pt vs 16/pt; ns), due to longer follow-up. However, this trend showed no significant impact of the

7 Implantation of a small defibrillator Percent patients on antiarrhythmic drugs I A-C II III Sot III Amio IV Total Figure 2 Concomitant antiarrhythmic drug treatment at the time of pectoral and abdominal implantable cardioverter defibrillator implantation (left-hand columns), at pre-hospital discharge (middle) and last follow-up (right) according to Vaughan Williams classification of antiarrhythmic drugs. =pectoral; =abdominal. IA C=Class IA C antiarrhythmic drugs; II, III, IV=classes II, III and IV antiarrhythmic drugs; Sot=sotalol; Amio=amiodarone. implanted device on the occurrence of spontaneous episodes. Analysis of the demographic data of patients with episodes compared to those without episodes revealed only a significant difference with respect to the implantable cardioverter defibrillator indications: ventricular tachycardia (65% of the patients with episodes vs 53% of the patients without episodes; P= 1) and sudden cardiac death (45% of the patients with episodes vs 54% of the patients without episodes; P= 2). No significant differences were found with respect to age, ejection fraction, coronary artery disease, previous myocardial infarction and cardiomyopathy as underlying heart disease. The location of the implantable cardioverter defibrillator (abdominal vs pectoral) revealed no more information. Figure 4 presents Kaplan Meier curves of the percentage of patients who did not experience a successfully treated arrhythmic event as a function of the time since implantation. These curves were terminated if fewer than 1 patients were included in the follow-up. The occurrence of device treatment for a first ventricular tachycardia/ventricular fibrillation episode was 53% at one year for all patients, with no significant difference between the abdominal and pectoral groups. The occurrence of anti-tachycardia pacing was 28% in the group as a whole after one year, with a significant difference between the abdominal group (occurrence 17%) and the total pectoral group (occurrence 33%; P= 39), which might be explained by the differences in underlying disease. The occurrence of shock therapy for ventricular tachycardia/ventricular fibrillation showed no significant difference between pectoral and abdominal device implantation. Adverse events and complications Table 2 gives a detailed overview of all adverse events and complications. Overall, 12 3% of the patients experienced at least one severe device-related adverse event during follow-up. A significant difference was found between the pectoral and the abdominal group, with respect to the frequency of severe lead-related adverse events (5 3% vs 11 6%, P< 5). A total of 15 right ventricular lead dislodgements occurred. The subclavian vein was used for insertion of one or more leads in 427 patients (54 9%). In the 385 non-active Can patients, the subclavian vein was used in 24 patients (62 3%), in the 393 Active Can patients the subclavian vein was used in 187 patients (47 6%). We identified 3 adverse events in 29 patients that might have been influenced by the introduction site of the lead, or by the track of the lead in the body. These events were selected from all lead-related events, including pneumothorax, right ventricular lead dislodgements, superior vena cava

8 192 E. Hoffmann et al Number of episodes VT FVT VF VT + VF + FVT No. of patients: % patients 24/11/41/69 16/18/18/18 4/3/11/19 3 / 5 / 5 / 5 29/14/34/7 19/17/15/18 43/179/62/117 28/29/27/3 Figure 3 Occurrence of spontaneous ventricular tachycardia and/or ventricular fibrillation episodes. Comparison between the groups revealed no significant differences. =abdominal; =total pectoral; =pectoral multi-lead; =pectoral single lead. lead dislodgements, right ventricular lead repositioning during implantation, haemothorax and thrombosis. In patients in whom the left or right subclavian vein was used, an adverse event, influenced by the site of introduction, was experienced by 21 patients (4 9%). In patients in whom the subclavian vein was not used, such an event was experienced by only eight patients (2 3%); P= 4 (Fisher s exact test). There was no difference among the three treatment groups in terms of any other individual severe adverse event. Inappropriate device therapy was the most common mild adverse event, almost twice as frequent as any other adverse event. There were few patients with other individual mild adverse events and none of these events had a considerable clinical impact on the patient. Survival and cause of death Figure 5 gives an overview of cumulative survival of all patients and freedom from device-related adverse events or events requiring surgical intervention. At 6 months, total patient survival was 97 5%, and at one year it was 97%. There were no differences with respect to the site of implantation or the lead system used. Table 3 details the

9 Implantation of a small defibrillator 193 VT/VF episode free survival p = Survival ATP therapy free survival Shock therapy free survival Time (days) p =.39 5 p = Figure 4 Episode and device treatment free survival (Kaplan Meier analysis) (a logrank test gives the P-values). =pectoral total; =abdominal. cause of death in 2 patients: six patients died perioperatively during the first 3 days after cardioverter defibrillator implantation. No sudden cardiac death occurred during follow-up. Frequencies of cardiac and non-cardiac deaths were similar for each treatment group. The most common cause of death was heart failure (in 11 patients) within days (range days) after implantation. There were no significant differences between the devices Time (days) Figure 5 Cumulative rates of survival (Kaplan Meier analysis) for all patients, severe and mild adverse event free survival and surgical procedure free survival. =survival; =mild device-related adverse events; =severe device-related adverse events; =surgical procedures. generation implantable cardioverter defibrillator in both pectoral and abdominal implantation sites. Differences in patient characteristics In this multicentre prospective clinical evaluation, the implantable cardioverter defibrillator treatment groups revealed differences at baseline, with respect to age, weight, underlying heart disease, degree of depression of left ventricular function and severity of heart failure. The pectoral group seemed to be significantly older and heavier than the abdominal group, more patients had coronary artery disease, more patients had a history of previous myocardial infarction, and the mean left ventricular ejection fraction was lower. The differences at baseline can be, to a large extent, explained by the differences in age. There were significantly younger patients in the abdominal group. Young patients had a significantly better ejection fraction and a lower NYHA class, less frequent coronary artery disease or a history of previous myocardial infarction. After correction for the large number of young patients in the abdominal group, there was no difference between the abdominal and the pectoral groups, with respect to ejection fraction, NYHA or previous myocardial infarction. Thus, after age correction, the baseline characteristics of the pectoral and abdominal implantation group were well matched. Discussion This study provides initial feedback from wide experience on implantation of a small transvenous third Implantation practice and procedure Between March and November 1993, when only multilead devices were available, 38% of the investigators

10 194 E. Hoffmann et al. Table 2 Severe and mild adverse events Abdominal multi-lead n=155 Total n=623 Multi-lead n=231 Single-lead n=392 Total n=778 Severe device related Shock 2 (1 3) 5 ( 8) 2 ( 9) 3 ( 8) 7 ( 9) Lead 21/18 (11 6)* 33 (5 3) 15 (6 5) 18 (4 6) 54/51 (6 6) Pocket 3 (1 9) 19 (3 ) 7 (3 ) 12 (3 1) 22 (2 8) Inappropriate therapy 9 (1 4) 5 (2 2) 4 (1 ) 9 (1 2) Other programming 2 (1 3) 9/8 (1 3) 6/5 (2 2) 3 ( 8) 11/1 (1 3) protocol 3 ( 5) 1 ( 4) 2 ( 5) 3 ( 4) psychological 4/3 ( 5) 4/3 (1 3) 4/3 ( 4) drug induced inefficacy 1 ( 6)* 1 ( 1) At least one adverse event 29/24 (15 5) 82/72 (11 6) 4/33 (14 4) 42/39 (9 9) 111/96 (12 3) Mild device related Shock 3 (1 9) 6 (1 ) 3 (1 3) 3 ( 8) 9 (1 2) Lead 3 (1 9)* 16/15 (2 4) 7 (3 ) 9/8 (2 ) 19/18 (2 3) Pocket 6/5 (3 2) 17 (2 7) 6 (2 6) 11 (2 8) 23/22 (2 8) Inappropriate therapy 31/27 (17 4) 88/75 (12 ) 38/32 (13 9) 5/43 (11 ) 119/12 (13 1) Other programming 1 ( 2) 1 ( 3) 1 ( 1) protocol 5 (3 2)* 3 ( 5) 3 (1 3) 8 (1 ) external device 5 (3 2) 15 (2 4) 9 (3 9) 6 (1 5) 2 (2 6) At least one mild adverse event 53/41 (26 5) 146/124 (19 9) 66/54 (23 4) 8/7 (17 8) 199/165 (21 2) General totals Shock 5/4 (2 6) 11/1 (1 6) 5 (2 2) 6 (1 5) 16/14 (1 8) Lead 24/21 (13 5)* 49/47 (7 5) 22 (9 5) 27/25 (6 4) 73/68 (8 7) Pocket 9/8 (5 1) 36/34 (5 5) 13 (5 6) 23/21 (5 4) 45/42 (5 4) Inappropriate therapy 31/27 (17 4) 97/82 (13 2) 43/36 (15 6) 54/46 (11 7) 128/19 (14 ) Severe non-device related 1 ( 6)* 25 (4 ) 6 (2 6) 19 (4 8) 26 (3 3) All other events 13 (8 4) 35/33 (5 3) 23/21 (9 1) 12 (3 1) 48/46 (5 9) Deaths 3 (1 9) 18 (2 9) 5 (2 2) 12 (3 1) 2 (2 6) At least one severe adverse event 86/57 (36 8) 27/22 (32 4) 117/88 (38 1) 153/114 (29 1) 365/259 (33 3) Results as numbers of episodes and patients if different (and % patients). Any patient experiencing more than one adverse event may appear more than once. *P< 5 between abdominal multi-lead and pectoral total; P< 5 between pectoral multi-lead and pectoral single-lead. Table 3 Analysis of death during follow-up Abdominal multi-lead n=155 Total n=623 Multi-lead n=231 Single-lead n=392 Total n=778 Deaths Total 3 (1 9) 17 (2 7) 5 (2 2) 12 (3 1) 2 (2 6) Sudden cardiac Cardiac 1 ( 6) 1 (1 6) 2 ( 9) 8 (2 ) 11 (1 4) Non-cardiac 1 ( 2) 1 ( 4) 1 ( 1) Procedure (peri-operative) 2 (1 3) 4 ( 6) 1 ( 4) 3 ( 8) 6 ( 8) Unknown 2 ( 3) 1 ( 4) 1 ( 3) 2 ( 3) implanted 385 implantable cardioverter defibrillators in equal numbers in both positions, pectoral and abdominal, 37% only in the pectoral and 25% only in the abdominal region. With the availability of the single-lead system, 393 devices out of 42 attempts (98%) were subsequently implanted via a single skin incision in the pectoral region by all investigators (one patient received the device abdominally). In the total study group, less than 1% of all implantations needed an additional subcutaneous or epicardial patch for successful defibrillation, with no difference between pectoral and abdominal implantations, reflecting the very high efficacy of transvenous implantable cardioverter defibrillator systems. The ability to implant a small multi-lead device in the pectoral position alone did not significantly reduce the operation time, as compared to the abdominal position ( min vs min), although the

11 Implantation of a small defibrillator 195 surgical trauma was already reduced from two to one skin incision and a subcutaneous thoracic preparation for the leads was no longer necessary. The mean operation time for transvenous multi-lead implantable cardioverter defibrillator implantations performed by both electrophysiologists [13] and cardiac surgeons [19] in recently published studies was comparable to the mean multi-lead implantation time in this study and markedly longer than the single-lead implantation duration. The introduction of pectoral single-lead implantation with the housing as an active electrode reduced the mean operation time to min. The mean operation time for single-lead devices reported in the recently published Active Can study was comparable at min [21].In the Active Can group, the first lead system tested was often successful, whereas, in the multi-lead group, several lead systems had to be tested to meet the defibrillation safety margin. Both the lowest energy tested and the true defibrillation threshold were significantly lower in the pectoral than in the abdominal group and in the pectoral single-lead than in the multi-lead group (P< 3). The mean defibrillation thresholds for transvenous implantable cardioverter defibrillator systems reported in the literature [3,7,23] range between 11 and 15 J and are comparable to the results in our study. In addition, there was no difference in the defibrillation threshold in the pectoral single-lead group as compared with the recently published Active Can study by Bardy and co-authors [21]. However defibrillation threshold criteria for implantation were met in 95% of single-lead as compared to 89% of multi-lead pectoral systems, which may, in addition to technical implantation aspects, have also contributed to the shorter implantation time and may also reflect a slightly increased safety margin for single-lead systems. Another major procedure-related advantage of pectoral systems may be the possibility of implantation without general anaesthesia, which may simplify the whole procedure and shorten the hospital stay [8,15,2]. These data confirm other evidence supporting the efficacious implantable cardioverter defibrillator design of single-lead active-housing systems [6,15,21]. The potential necessity of additional electrodes to achieve effective defibrillation thresholds should also be diminished, but this need remains in about 2% of patients. Patient outcome The rate of successful termination of any ventricular tachyarrhythmia reached 97%. The efficacy rates seen in this study are similar to those previously reported, which vary between 96% and 98% for transvenous implantable cardioverter defibrillators at one year for all episodes of ventricular tachycardia/ventricular fibrillation [3,23,24]. Kaplan Meier analysis revealed a 53% occurrence of device treatment for a first ventricular tachycardia/ ventricular fibrillation episode at one year for all patients (Kaplan Meier estimate). The occurrence of anti-tachycardia pacing was 28% in the total study group after one year, with a significant difference between the abdominal group and the total pectoral group (occurrence 17% vs 33%; P= 39). Experience with a small device in the abdominal position compared with pectoral cardioverter defibrillator implantation has not been reported previously. Comparisons of pectoral (single/multi-lead systems) and abdominal cardioverter defibrillators revealed no differences in the distribution of the ventricular tachycardia, ventricular fibrillation or fast ventricular tachycardia episodes or of the total number of episodes during follow-up. Ventricular tachyarrhythmia detection and ventricular tachyarrhythmia therapies activated were comparable in both groups and were not related to the implantation site and the use of multi- or single-lead systems. In other words, abdominal vs pectoral device implantation did not influence the efficacy of implantable cardioverter defibrillator treatment for ventricular tachyarrhythmias. Interventional device therapy carries the risk of a wide range of complications. This study provides the only comparative data on the performance of a small device in the pectoral and abdominal implantation positions. There was an excess of severe lead-related adverse effects in the abdominal group (abdominal 11 6% vs pectoral 5 3%; P< 5), which can be interpreted as being a consequence of the leads travelling subcutaneously along the exterior wall of the thorax and thereby being exposed to traction of shoulder movements on the abdominally fixed leads as compared to the pectorally fixed leads. Additionally, a smaller number of leads is probably associated with fewer lead-related problems; the pectoral single-lead group showed this complication in only 4 6% of the implantations. In the non-active Can patients, the subclavian vein was used in 62 3% of the patients, in the Active Can group, only in 47 6% of the patients. Some lead-related complications were significantly higher using the subclavian vein compared with the patients where the subclavian vein was not used (4 9% vs 2 3%); P= 4. We found no significant difference between the three treatment groups in the frequency of any other individual severe adverse event. Inappropriate therapies (observed in 82 patients, 13 1%) accounted for the majority of mild events, which compared favourably to previously published data, with a slightly higher frequency in the abdominal group. The Kaplan Meier estimate of adverse event free survival rate was surprisingly low, 51% after one year of followup, but it should be considered that the definition of adverse events used in this study, including any undesirable clinical event, will by itself cause higher rates than usually reported. With respect to the severe devicerelated adverse events, the rate was 85% and the surgical intervention-free rate was 93% without differences in any of the device groups. Fifty-one percent of the patients received antiarrhythmic drug therapy at the time of implantation. After an initial reduction of the antiarrhythmic therapy to 32%, during long-term follow-up, 36 5% of patients still needed antiarrhythmic drug treatment, with equal distribution in the abdominal and pectoral implantation groups.

12 196 E. Hoffmann et al. Patient survival, the primary outcome measure, was as high as 97% after one year, and was independent of the site of device implantation or the lead system used. In particular, peri-operative morbidity ( 8%) was lower, not only as compared to epicardial electrode systems, but also as compared to second- and thirdgeneration large implantable cardioverter defibrillator devices [1,3,16,19,2,24]. The main cause of death was heart failure unrelated to ventricular tachyarrhythmias. The 12-month survival rates of transvenous electrode systems and abdominal generator implantation seen in our study are better than those reported previously [3,21]. Data for pectoral comparisons are not available. Survival rates reflect the combined losses resulting from cardiac death, sudden cardiac death and death from other causes; however, differentiation of each of these other important end-points revealed no differences. Figures for all three implantable cardioverter defibrillator groups compare favourably with published incidence rates. Clinical implications This prospective observational clinical evaluation study revealed surprisingly few differences related to the site of device implantation. The results suggest that pectoral implantation as compared to abdominal cardioverter defibrillator implantation has advantages only in respect of simplifying the implant procedure and reducing leadrelated adverse events significantly. Single-lead systems in the pectoral position were associated with a shorter operation time than multi-lead pectoral systems, reflecting an easier implantation procedure. The lower defibrillation threshold results in a higher safety margin for the Active Can group; however, the efficacy of treatment of ventricular tachyarrhythmias was the same. The considerable overall incidence of 33% of at least one adverse event, despite the use of a small device and transvenous electrodes, has been taken into account if implantable cardioverter defibrillator therapy is chosen. Independent of implantation site, there was an excellent 97% survival rate of patients with ventricular tachyarrhythmias treated with the third-generation of small implantable cardioverter defibrillators. We thank Dr Petra Nimmermann and Dr Uwe Dorwarth for their help in preparing the publication. This study was supported by Bakken Research Center, Medtronic Inc., Maastricht, Netherlands. References [1] Saksena S. The PCD Investigator Group. Clinical outcome of patients with malignant ventricular tachyarrhythmias and a multiprogrammable implantable cardioverter defibrillator implanted with or without thoracotomy: an international multicenter study. JACC 1994; 23: [2] Hoffmann E for the European Jewel Investigators. Chronic experience with pectoral defibrillator implantation. PACE 1995; 18: 372. [3] Zipes D, Roberts D; for the Pacemaker-Cardioverter- Defibrillator Investigators. Results of the international study of the implantable pacemaker cardioverter-defibrillator: a comparison of epicardial and endocardial lead systems. Circulation 1995; 92: [4] Rosenqvist M, Beyer T, Block M, Dulk Kd, Minten J, for the European Jewel Investigators. Adverse events in patients with implantable cardioverter defibrillators a prospective study. PACE 1996; 19: 598. [5] Fromer M, Brachmann J, Block M et al.efficacy of automatic multimodal device therapy for ventricular tachyarrhythmias as delivered by a new implantable pacing cardioverterdefibrillator: results of a European multicenter study of 12 implants. Circulation 1992; 86: [6] Markewitz A, Kaulbach H, Mattke S et al. One-incision approach for insertion of implantable cardioverter defibrillators. Ann Thorac Surg 1994; 58: [7] Thakur R, Ip J, Mehta D et al. Subpectoral implantation of ICD generators: long-term follow-up. PACE 1995; 18: [8] Bardy GH, Johnson G, Poole JE et al. A simplified single-lead unipolar transvenous cardioversion-defibrillation system. Circulation 1993; 88: 543. [9] EN54, European Standard Clinical Investigation of medical devices for human subjects. European Committee for Standardization. Brussels, Belgium, [1] Bardy GH, Hofer B, Johnson G et al. Implantable transvenous cardioverter-defibrillators. Circulation 1993; 87: [11] Leitch JW, Gillis AM, Wyse DG et al. Reduction in defibrillator shocks with an implantable device combining antitachycardia pacing and shock therapy. J Am Coll Cardiol 1991; 18: [12] Kim SG, Fogoros RN, Furman S, Connolly SJ, Kuck KH, Moss AJ. Standardized reporting of ICD patient outcome. PACE 1993; 16: [13] Strickberger SA, Hummel JD, Daoud E et al. Implantation by electrophysiologists of 1 consecutive cardioverter defibrillators with nonthoracotomy lead systems. Circulation 1994; 9: [14] Akhtar M, Avitall B, Jazayeri M et al. Role of implantable cardioverter defibrillator therapy in the management of high-risk patients. Circulation 1992; 85: [15] Jones GK, Bardy GH, Kudenchuk PJ et al. Mechanical complications after implantation of multiple-lead nonthoracotomy defibrillator systems: implications for management and future system design. Am Heart J 1995; 13: [16] Raviele A, Gasparini G. Italian multicenter clinical experience with endocardial defibrillation: Acute and long-term results in 37 patients. PACE 1995; 18: [17] Jones GK, Bardy GH. Considerations for ventricular fibrillation detection by implantable cardioverter-defibrillators. Am Heart J 1994; 127: [18] Bardy GH, Troutman C, Poole JE et al. Clinical experience with a tiered therapy multiprogrammable antiarrhythmic device. Circulation 1992; 85: [19] Hammel D, Scheld HH, Block M, Breithardt G. Nonthoracotomy defibrillator implantation: a single-center experience with 2 patients. Ann Thorac Surg 1994; 58: [2] Tung RT, Baja AK. Safety of implantation of a cardioverterdefibrillator without general anesthesia in an electrophysiology laboratory. Am J Cardiol 1995; [21] Bardy GH, Yee R, Jung W for the Active Can Investigators. Multicenter experience with a pectoral unipolar implantable cardioverter defibrillator. JACC 1996; 28: 4 1. [22] Nisam S, Kaye SA, Mower MM, Hull M. AICD automatic cardioverter defibrillator clinical update: 14 years experience in over 34 patients. PACE 1995; 18: [23] Bardy GH, Hofer B, Johnson G et al. Implantable transvenous cardioverter-defibrillators. Circulation 1993; 87: