- Email: [email protected]
Trials of the automatic implantable defibrillator in man
J
THoRAc CARDIOVASC SURG
86:381-387, 1983
Trials of the automatic implantable defibrillator
.
In man A three-year program report Since February, 1980, 65 survivors of multiple arrhythmic cardiac arrests unresponsive to therapy were referred for implantation of the automatic defibrillator. In 37 patients (Group I), automatic defibrillator implantation alone was performed by subxiphoid insertion (20 patients) or thoracotomy (17 patients). In another 28 patients (Group 1/), implantation was combined with definitive cardiac procedures-coronary artery bypass grafting in seven patients, bypass grafting and mitral valve replacement in four patients, and left ventricular aneurysmectomy with endocardial resection in 17 patients. There were no surgical deaths in Group I; four operative deaths occurred in Group 1/. The longest follow-up has been 34 months, average 15.6 months. Following hospital discharge, 44 episodes of automatic out-of-hospital resuscitation were observed in II Group I patients. Similarly, four resuscitations were observed in two Group /I patients. Hypothetical survival curves based on the assumption that these out-of-hospital resuscitations were lifesaving indicated expected I year survivals rates of 45% in Group I and 85% in Group /I. Excluding the perioperative deaths that were unrelated to the defibrillator, the actual I year survival rates observed were 75% and 95% in Groups I and 1/, respectively. Although definitive operation markedly reduced the number of out-of-hospital arrhythmic episodes, the automatic defibrillator appears to increase survival both when implanted alone and when used in combination with cardiac procedures.
Levi Watkins, Jr., M.D. (by invitation), Morton M. Mower, M.D. (by invitation), Philip R. Reid, M.D. (by invitation), Edward V. Platia, M.D. (by invitation), Lawrence S. C. Griffith, M.D. (by invitation), and M. Mirowski, M.D. (by invitation), Baltimore, Md.
Sponsored by Vincent L. Gott, M.D., Baltimore, Md.
After over a decade of bench and animal testing;" the first human implantation of the automatic implantable defibrillator was performed in February, 1980, at The Johns Hopkins Hospital. Since that time 65 survivors of sudden death were referred for implantation of
the device at this institution. A great deal has been learned about the defibrillator and its clinical use."? The present report outlines the current status of the trials in patients at The Johns Hopkins Hospital, with particular emphasis on the I and 2 year mortality experience. Methods and materials
From the Departments of Surgery and Medicine, The Johns Hopkins Medical Institutions, Baltimore, Md. and the Department of Medicine, Sinai Hospital of Baltimore, Md. Supported in part by grants from the General Clinical Research Center No. RROOO35-21, Outpatient General Clinical Research Center No. RROO722-9, and the American Heart Association, Maryland Affiliate. Computational assistance was received from CLINFO, sponsored by National Institutes of Health Grant No. 5MOI RR35-20. Read at the Sixty-third Annual Meeting of The American Association for Thoracic Surgery, Atlanta, Ga., April 25-27, 1983. Address for reprints: Levi Watkins, Jr., M.D., The Johns Hopkins Hospital, 600 N. Wolfe St., Baltimore, Md. 21205.
The automatic defibrillator. The original defibrillator* (Fig. 1) was designed to diagnose and terminate ventricular fibrillation. Both sensing and defibrillation were accomplished by electrodes positioned in the superior vena cava and on the left ventricular apex. The determinant for identification of ventricular fibrillation was satisfaction of a probability density function criterion," which required the input signal to spend increased *Developed and manufactured under the name AID defibrillator by Intec Systems, Pittsburgh, Pa.
381
3 8 2 Watkins et al.
Fig. 1. The original AID defibrillator.
time away from the zero potential baseline. In contrast to sinus rhythm, during which the electrical signal spends a relatively great amount of time at the zero potential, ventricular fibrillation signals spend little time at the zero potential. Although these units functioned adequately, several problems related to their use were recognized. These included lead fractures, which resulted in erroneous input signal information causing spurious shocks in four patients. Improvement in lead construction eliminated this problem. Rate miscounting also caused spurious shocks in five patients. In addition, the slow ventricular tachycardias frequently encountered in survivors of sudden death episodes were often rejected by the probability density function sensing circuitry. The spectrum of arrhythmias to which the device reacted was broadened to include slower ventricular tachycardia by the development of a reliable rate detector system which included an additional bipolar ventricular lead that served for R-wave synchronization. The so modified defibrillator, AID-B,9 (Fig. 2) used the superior vena cava-left ventricular apex lead for the probability density function analysis and for delivery of the corrective pulses, while the ventricular bipolar lead permitted accurate determination of the lower heart rates. Several other features of the new model referred to as the internal automatic cardioverter-defibrillator are as
The Journal of Thoracic and Cardiovascular Surgery
follows: (1) the ability to determine non invasively the number of pulses the device delivered to the patient, (2) the ability to verify R-wave sensing function with the magnet, (3) the ability to easily activate or inactivate the device. Since April, 1982, the AID-B defibrillator has been the model utilized in the study . In addition, high-energy devices with an output of up to 42 joules became available after the clinical trials had demonstrated increased defibrillatory thresholds in patients with increased myocardial cell mass. Patient population. The population of this study consisted of 65 patients who were referred for implantation of the automatic defibrillator between February, 1980, and January, 1983, at The Johns Hopkins Hospital. The patient selection criteria have been described.v' In brief, all recipients had had recurrent ventricular tachyarrhythmias unresponsive to therapy, and all had been resuscitated from at least two episodes of arrhythmic cardiac arrest not associated with acute myocardial infarction. In fact , each patient had received an average of 4.5 antiarrhythmic agents without benefit and continued to experience life-threatening ventricular arrhythmias necessitating an average of 3.9 cardiopulmonary resuscitations. More recently these stringent criteria have been relaxed, and at present only one episode of ventricular fibrillation or ventricular tachycardia is required along with continued evidence of inducibility during electrophysiological testing . There were 51 male and 14 female subjects. Their age range was from 16 to 75 years with a mean of 51. Coronary artery disease was the underlying cause in 51 patients and nonischemic cardiomyopathy was present in 11. One patient had prolonged QT interval syndrome, one had idiopathic hypertrophic subaortic stenosis, and one patient showed no evidence of structural cardiac disease except for the presence of malignant arrhythmias. Seventeen patients had undergone previous cardiac surgical procedures including coronary artery bypass grafting (12), endocardial resection (one), aneurysmectomy (three), and mitral valve replacement (one). Prior to surgical intervention, complete cardiac evaluation including cardiac catheterization and electrophysiological studies were performed. The patients were divided into two groups to facilitate evaluation of the functional performance of the implanted device and to assess the role of concomitant cardiac surgical procedures. Group I consisted of 37 patients who underwent implantation of the defibrillator only. These patients had cardiomyopathy or end-stage cardiac disease believed to be too severe for corrective procedures. The mean ejection fraction of 29% in this group indicated their
Volume 86 Number 3
Automatic implantable defibrillator
September. 1983
383
Fig. 2. The modified AID-B defibrillator.
poor ventricular function. Group II consisted of 28 patients in whom cardiac procedures were performed concomitantly with implantation of the automatic defibrillator. Left ventricular dysfunction was also present in this group, as their mean ejection fraction was 26%. Implantation techniques. The surgical approach for implantation was dictated by the clinical needs of the patient and has been described. 10. I I Briefly, patients with previous cardiac surgical procedures underwent implantation of the defibrillator via a lateral thoracotomy (Fig. 3). In the candidates undergoing combined cardiac procedures and defibrillator implantation, median sternotomy was employed. In the patients who had not had a previous operation and in whom no other cardiac procedure was indicated , the simpler subxiphoid technique was used (Fig. 4). With the AID-B model, two closely spaced (I em) screw-on myocardial electrodes were placed during thoracotomy or median sternotomy (Fig. 5). These provided the signals for the rate channel. When the subxiphoid approach was used, a bipolar endocardial catheter lead was inserted percutaneously and advanced to the right ventricular apex in a manner identical to that used for endocardial pacemaker lead insertion. Postoperative evaluation. Postoperative care included general medical management and adjustment of antiarrhythmic therapy when indicated. Programmed electrical stimulation with an aggressive protocol including the use of alternating current" was performed 7 to 30 days after operation to determine inducibility, to evaluate the efficacy of antiarrhythmic drug therapy, and to document effective function of the implanted
device. After discharge , outpatient visits were required every 2 to 3 months for noninvasive testing of the pulse generator to assess battery life and to reevaluate medical therapy. The determination that an automatic out-of-hospital resuscitation from a malignant arrhythmia had occurred was made on the basis of a characteristic sequence of events described by the patient or bystanders. This sequence included palpitations, weakness, dizziness, and often cardiovascular collapse and evidence of an internal discharge consisting of a sudden brief muscular contractions of the thorax. This was followed by almost immediate recovery and with restoration of hemodynamics. The diagnosis needed unanimous agreement of the investigators with regard to the circumstances surrounding the event and its interpretation. Actuarial analysis of the deaths that occurred in patients who received the implant was done by means of the Kaplan-Meier! ' method. In this analysis, patients who did not receive pulse generators and the operative deaths were excluded. A hypothetical estimate of the late deaths that would have occurred in the same group of patients without implantation of the automatic defibrillator was also performed. In order to derive these hypothetical estimates, a nonsurvival end point was defined as the patient's actual death or the first automatic resuscitation out of the hospital, whichever occurred first. This definition was based on the assumption that patients whose malignant arrhythmias were automatically terminated outside the hospital would have died without the implanted device. In patients with several automatic resuscitations, only the first event was
The Journal of Thoracic and Cardiovascular
3 8 4 Watkins et al.
De fib r ill a t ing e le c !r"o d e c ov ers ~ 1:!opical per lcZirdiu \.
,
Surgery
Table I. Automatic resuscitations of sudden
death episodes No. of patients
Group I In hospital Out of hospital Group II In hospital Out of hospital
Fig. 3. implantation of the defibrillator via a lateral thoracotomy.
counted as an end point. For actual survival to be compared with the predicted mortality, the 95% confidence limits for the actual survival experience have been calculated. The longest follow-up has been 34 months; the average was 15.6 months. Results Operative results. In Group I, 37 patients underwent implantation of the automatic defibrillator via thoracotomy or the subxiphoid technique. In two of these patients only the leads were implanted. There were no surgical deaths. One patient required reexploration for postoperative bleeding. Twenty-eight patients in Group II underwent implantation of the defibrillator. This was combined with endocardial resection and aneurysmectomy in 17, coronary artery bypass grafting in seven, and coronary artery bypass grafting and mitral valve replacement in four. In five of these patients only the leads were implanted. In Group II there were four operative deaths, two of which occurred during emergency procedures. One patient required reexploration for mediastinal bleeding. Arrhythmia control. In the hospital, under monitored conditions, the implanted device correctly identified and terminated 30 episodes of spontaneous ventricular tachyarrhythmias in six Group I patients and 11 episodes in one Group II patient. Following discharge,
No. of resuscitations
6
30
II
44
I 2
II 4
44 out-of-hospital resuscitations were observed in 11 Group I patients and four were observed in two Group II patients (Table I). Late mortality. There have been a total of 14 late deaths in this series; 13 in Group I and one in Group II. Late deaths were due to the progression of end-stage myocardial disease in three, myocardial infarction in three, and cerebral vascular accident in one. One patient had electromechanical dissociation. Four presumably died of unwitnessed sudden death episodes (Fig. 6). For Group I, actuarial display of mortality with the defibrillator implanted is shown in Fig. 7. The I and 2 year actual survival rates are 75% and 65%, respectively. The bottom curve is a hypothetical estimate of survival had the out-of-hospital resuscitations not been successful. The mortality for Group II , excluding the operative deaths, is displayed in Fig. 8 and demonstrates 1 and 2 year survival rates of 95%. The bottom curve, again , shows the hypothetical estimate of survival had the out-of-hospital resuscitations failed . Discussion After more than 3 years of clinical evaluation, considerable evidence now indicates that the automatic defibrillator is an effective and reliable device capable of accurate diagnosis and prompt termination of malignant ventricular arrhythmias in man.4-7.9-1 4The surgical techniques for implantation have been found safe, and serious hazards from long-term implantation of the device have not been observed. The few problems which appeared during the early part of the study seem to have been overcome by appropriate design modifications . These changes eventually led to the development of a more advanced model with a broader diagnostic and therapeutic spectrum. The analysis of Group I patients provides a unique opportunity for a close evaluation of the functional performance of the implanted defibrillator. All of these patients had multiple arrhythmic arrests which remained inducible during programmed electrical stim-
Volume 86 Number 3
Automatic implantable defibrillator
September, 1983
385
Pe r ic a r d i a l /! fl a ps •...•.. ..
/
I
A p ex of He ad
i i
•.••....-:
I ncis io n in
pe r t c a t- d tum
al att achme n t t o di aph) -a ~ ~
A:'
~ :.
;l o.> _ .
Fig. 4. Implantation of the defibr illator by the subxiphoid technique. RV, Right ventricle. LV. Left ventricle.
D eftbr
i llahn
~
e leclro de
c overs ap ical myocar d iu m
Fig. 5. Placement of the defibrillator via thoracotomy or median sternotomy.
ulation, despite multiple antiarrhythmic drug trials. Because of drug resistance and the inability to offer additional surgical interventions, the underlying myocardial process responsible for the arrhythmias remained essentially unchanged. This undoubtedly accounted for the large number (74) of malignant arrhythmias observed in this group. These potentially lethal episodes were correctly terminated by the implanted device. The 44 successful automatic resuscitations outside the hospital setting are particularly important, because they appear to be largely responsible
for the 75% 1 year survival rate. Deaths in this group were due mainly to the progression of native disease and were possibly arrhythmic in nature in only four instances . Group II patients were found suitable for more definitive cardiac operations. More than half underwent endocardial resection, which controlled ventricular tachycardia in over 90% of those so treated. As a whole, Group II patients fared much better. In this group only four out-of-hospital sudden death episodes were observed and these were automatically terminated. Of
The Journal of Thoracic and Cardiovascular Surgery
3 8 6 Watkins et al.
"!:
95% ACTIlAL SURVIVAL- CROUP II HYPOTHETICAL SUllVIVAL- CROUP II
"
z
.8
>
;: a:
~
'"
.6
z
... 0
De 0
•4
"0
De
Do
.2 B
•
.5
1.5
2
2.5
3
YEARS ELAPSED
Fig. 8. The I year and 2 year survival rates. Fig. 6. Late mortality. new surgical techniques, and defibrillator implantation should all be viewed as complementary. We express our gratitude to Mr. Paul Freund, P.A., Ms. Sue Johnson, P.A., and Ms. Jennifer Becker for technical support and to Ms. Irma Silkworth for secretarial assistance.
> .8 > 0:
ACTIlAL SURVIVAL- CROUP I
~
'" 5
.6
to:
o
l5
0:
.4
"-
.2
2 B
B
.5
1.5
2
2.5
:3
YEARS ELAPSED
Fig. 7. Survival rates.
the patients surviving operation, 95% are alive at 2 years. This experience suggests that whenever possible an aggressive definitive cardiac surgical approach is warranted and will decrease the likelihood of malignant arrhythmias occurring in the years after the operation. Implantation of the defibrillator is also desirable because it provides additional protection against sudden death secondary to late arrhythmic events. The favorable results of the clinical trials support continued clinical use and further evaluation of the automatic defibrillator. It appears from these data and previously published information 14 that effective function of the device may prolong survival in selected high-risk patients," However, the implanted defibrillator should not be viewed as a substitute for antiarrhythmic or surgical therapy. The new antiarrhythmic drugs, the
3
4
5
6
REFERENCES Mirowski M, Mower MM, Staewen WS, Tabatznik, B, Mendeloff AL: Standby automatic defibrillator. An approach to prevention of sudden coronary death. Arch Intern Med 126:158-161, 1970 Mirowski M, Mower MM, Langer A, Heilman MS, Schreibman J: A chronically implanted system for automatic defibrillation in active conscious dogs. Experimental model for treatment of sudden death from ventricular fibrillation. Circulation 58:90-94, 1978 Mirowski M, Mower MM, Bhagavan BS, Langer A, Kolenik SA, Fischell RE, Heilman MS: Chronic animal and bench testing of the implantable automatic defibrillator, Chap 27-2, Pacesymp: Proceedings of the Sixth World Symposium on Cardiac Pacing, C Meere, ed., Montreal, 1979. Mirowski M, Mower MM, Reid PR, Watkins L Jr, Langer A: The automatic implantable defibrillator. New modality for treatment of life-threatening ventricular arrhythmias. PACE 5:384-401,1982 Mirowski M, Reid PR, Mower MM, Watkins L, Gott VL, Schauble JF, Langer A, Heilman MS, Kolenik SA, Fischell RF, Weisfeldt ML: Termination of malignant ventricular arrhythmias with an implanted automatic defibrillator in human beings. N Engl J Med 303:322-324, 1980 Mirowski M, Reid PR, Watkins L Jr, Weisfeldt ML, Mower MM: Clinical treatment of life-threatening ventricular tachyarrhythmias with the automatic implantable defibrillator. Am Heart J 102:265-270, 1981
Volume 86 Number 3 September, 1983
7 Mirowski M, Reid PR, Mower MM, Watkins L Jr: Successful conversion of out-of-hospital life-threatening arrhythmias with the implanted automatic defibrillator. Am Heart J 103:147-148, 1982 8 Langer A, Heilman MS, Mower MM, Mirowski M: Considerations in development of the automatic implantable defibrillator. Med Instrum 10:163-167, 1976 9 Reid PR, Mirowski M, Mower MM, Platia EV, Griffith LSC, Watkins L Jr, Bach SM, Imran M, Thomas A: Clinical evaluation of the internal automatic cardioverterdefibrillator in survivors of sudden cardiac death. Presented in part at the American Heart Scientific Sessions, Dallas, Texas, 1982. 10 Watkins L Jr, Mirowski M, Mower MM, Reid PR, Griffith LSC, Vlay SC, Weisfeldt ML, Gott VL: Automatic defibrillation in man. The initial surgical experience. J THoRAc CARDIOVASC SURG 82:492-500, 1981 11 Watkins L Jr, Mirowski M, Mower MM, Reid PR, Freund PA, Thomas A, Weisfeldt ML, Gott VL: Implantation of the automatic defibrillator. The sub-xiphoid approach. Ann Thorac Surg 34:515-520, 1982 12 Mower MM, Reid PR, Watkins L Jr, Mirowski M: Use of alternating current during diagnostic electrophysiologic studies. Circulation 67:69-72, 1983 13 Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 14 Mirowski M, Reid PR, Winkle RA, Mower MM, Watkins L Jr, Stinson EB, Griffith LSC, Kallman CH, Weisfeldt ML: Mortality in patients with implanted automatic defibrillators. Ann Intern Med 98:585-588, 1983
Discussion DR. ALDEN H. HARKEN Phi/adelphia, Pa.
Dr. Watkins and the group at Johns Hopkins have made substantial and innovative progress toward the surgical therapy of ventricular fibrillation and associated "sudden death." They are to be congratulated. We know that a substantial number of persons in the United States die suddenly, perhaps as many as 400,000 a year. One might guess how many of that
Automatic implantable defibrillator
387
group could be helped by this kind of procedure. In addition, when these patients get to the hospital, following successful resuscitation, what kind of electrophysiological testing should they undergo? DR. WATKINS (Closing) Thank you very much for the interesting and stimulating discussion. To address the question regarding the numbers of patients involved, it is estimated that about 400,000 deaths annually result from malignant ventricular arrhythmias. With the advent of modern electrophysiological testing and programmed electrical stimulation, it has become apparent that approximately 75% of that number could potentially be controlled with antiarrhythmic agents if programmed electrical stimulation were used to determine the choice of drugs utilized. The remaining 25% would be potential candidates for operation. These persons should undergo a complete work-up, including cardiac catheterization, angiography, hemodynamic and electrophysiological studies, and particularly programmed electrical stimulation. Analysis of this work-up will reveal that some patients will not be candidates for cardiac operations. Others will be candidates for coronary artery bypass or mitral valve replacement, and a good number will be candidates for endocardial resection. At present we implant the device in all persons who meet the criteria for implantation, that is, survival after one or two cardiac arrests and overall drug refractoriness. We implant the device in all patients undergoing endocardial resection, because we believe that late recurrence rates of 10% or more justify additional long-term protection for these patients. In closing, arrhythmias in the vast majority of patients can be controlled by medications; only the 25% whose arrhythmias are resistant to medications will be candidates for further work-up and operation. The type of operation indicated is determined by the results of the cardiac work-up. Because of the inability to predict accurately which patients will have late recurrences of malignant ventricular arrhythmias, we believe implantation of the device is indicated at present. We believe the defibrillator will become a valuable addition to our therapeutic armamentarium for the treatment of malignant ventricular arrhythmias.
THoRAc CARDIOVASC SURG
86:381-387, 1983
Trials of the automatic implantable defibrillator
.
In man A three-year program report Since February, 1980, 65 survivors of multiple arrhythmic cardiac arrests unresponsive to therapy were referred for implantation of the automatic defibrillator. In 37 patients (Group I), automatic defibrillator implantation alone was performed by subxiphoid insertion (20 patients) or thoracotomy (17 patients). In another 28 patients (Group 1/), implantation was combined with definitive cardiac procedures-coronary artery bypass grafting in seven patients, bypass grafting and mitral valve replacement in four patients, and left ventricular aneurysmectomy with endocardial resection in 17 patients. There were no surgical deaths in Group I; four operative deaths occurred in Group 1/. The longest follow-up has been 34 months, average 15.6 months. Following hospital discharge, 44 episodes of automatic out-of-hospital resuscitation were observed in II Group I patients. Similarly, four resuscitations were observed in two Group /I patients. Hypothetical survival curves based on the assumption that these out-of-hospital resuscitations were lifesaving indicated expected I year survivals rates of 45% in Group I and 85% in Group /I. Excluding the perioperative deaths that were unrelated to the defibrillator, the actual I year survival rates observed were 75% and 95% in Groups I and 1/, respectively. Although definitive operation markedly reduced the number of out-of-hospital arrhythmic episodes, the automatic defibrillator appears to increase survival both when implanted alone and when used in combination with cardiac procedures.
Levi Watkins, Jr., M.D. (by invitation), Morton M. Mower, M.D. (by invitation), Philip R. Reid, M.D. (by invitation), Edward V. Platia, M.D. (by invitation), Lawrence S. C. Griffith, M.D. (by invitation), and M. Mirowski, M.D. (by invitation), Baltimore, Md.
Sponsored by Vincent L. Gott, M.D., Baltimore, Md.
After over a decade of bench and animal testing;" the first human implantation of the automatic implantable defibrillator was performed in February, 1980, at The Johns Hopkins Hospital. Since that time 65 survivors of sudden death were referred for implantation of
the device at this institution. A great deal has been learned about the defibrillator and its clinical use."? The present report outlines the current status of the trials in patients at The Johns Hopkins Hospital, with particular emphasis on the I and 2 year mortality experience. Methods and materials
From the Departments of Surgery and Medicine, The Johns Hopkins Medical Institutions, Baltimore, Md. and the Department of Medicine, Sinai Hospital of Baltimore, Md. Supported in part by grants from the General Clinical Research Center No. RROOO35-21, Outpatient General Clinical Research Center No. RROO722-9, and the American Heart Association, Maryland Affiliate. Computational assistance was received from CLINFO, sponsored by National Institutes of Health Grant No. 5MOI RR35-20. Read at the Sixty-third Annual Meeting of The American Association for Thoracic Surgery, Atlanta, Ga., April 25-27, 1983. Address for reprints: Levi Watkins, Jr., M.D., The Johns Hopkins Hospital, 600 N. Wolfe St., Baltimore, Md. 21205.
The automatic defibrillator. The original defibrillator* (Fig. 1) was designed to diagnose and terminate ventricular fibrillation. Both sensing and defibrillation were accomplished by electrodes positioned in the superior vena cava and on the left ventricular apex. The determinant for identification of ventricular fibrillation was satisfaction of a probability density function criterion," which required the input signal to spend increased *Developed and manufactured under the name AID defibrillator by Intec Systems, Pittsburgh, Pa.
381
3 8 2 Watkins et al.
Fig. 1. The original AID defibrillator.
time away from the zero potential baseline. In contrast to sinus rhythm, during which the electrical signal spends a relatively great amount of time at the zero potential, ventricular fibrillation signals spend little time at the zero potential. Although these units functioned adequately, several problems related to their use were recognized. These included lead fractures, which resulted in erroneous input signal information causing spurious shocks in four patients. Improvement in lead construction eliminated this problem. Rate miscounting also caused spurious shocks in five patients. In addition, the slow ventricular tachycardias frequently encountered in survivors of sudden death episodes were often rejected by the probability density function sensing circuitry. The spectrum of arrhythmias to which the device reacted was broadened to include slower ventricular tachycardia by the development of a reliable rate detector system which included an additional bipolar ventricular lead that served for R-wave synchronization. The so modified defibrillator, AID-B,9 (Fig. 2) used the superior vena cava-left ventricular apex lead for the probability density function analysis and for delivery of the corrective pulses, while the ventricular bipolar lead permitted accurate determination of the lower heart rates. Several other features of the new model referred to as the internal automatic cardioverter-defibrillator are as
The Journal of Thoracic and Cardiovascular Surgery
follows: (1) the ability to determine non invasively the number of pulses the device delivered to the patient, (2) the ability to verify R-wave sensing function with the magnet, (3) the ability to easily activate or inactivate the device. Since April, 1982, the AID-B defibrillator has been the model utilized in the study . In addition, high-energy devices with an output of up to 42 joules became available after the clinical trials had demonstrated increased defibrillatory thresholds in patients with increased myocardial cell mass. Patient population. The population of this study consisted of 65 patients who were referred for implantation of the automatic defibrillator between February, 1980, and January, 1983, at The Johns Hopkins Hospital. The patient selection criteria have been described.v' In brief, all recipients had had recurrent ventricular tachyarrhythmias unresponsive to therapy, and all had been resuscitated from at least two episodes of arrhythmic cardiac arrest not associated with acute myocardial infarction. In fact , each patient had received an average of 4.5 antiarrhythmic agents without benefit and continued to experience life-threatening ventricular arrhythmias necessitating an average of 3.9 cardiopulmonary resuscitations. More recently these stringent criteria have been relaxed, and at present only one episode of ventricular fibrillation or ventricular tachycardia is required along with continued evidence of inducibility during electrophysiological testing . There were 51 male and 14 female subjects. Their age range was from 16 to 75 years with a mean of 51. Coronary artery disease was the underlying cause in 51 patients and nonischemic cardiomyopathy was present in 11. One patient had prolonged QT interval syndrome, one had idiopathic hypertrophic subaortic stenosis, and one patient showed no evidence of structural cardiac disease except for the presence of malignant arrhythmias. Seventeen patients had undergone previous cardiac surgical procedures including coronary artery bypass grafting (12), endocardial resection (one), aneurysmectomy (three), and mitral valve replacement (one). Prior to surgical intervention, complete cardiac evaluation including cardiac catheterization and electrophysiological studies were performed. The patients were divided into two groups to facilitate evaluation of the functional performance of the implanted device and to assess the role of concomitant cardiac surgical procedures. Group I consisted of 37 patients who underwent implantation of the defibrillator only. These patients had cardiomyopathy or end-stage cardiac disease believed to be too severe for corrective procedures. The mean ejection fraction of 29% in this group indicated their
Volume 86 Number 3
Automatic implantable defibrillator
September. 1983
383
Fig. 2. The modified AID-B defibrillator.
poor ventricular function. Group II consisted of 28 patients in whom cardiac procedures were performed concomitantly with implantation of the automatic defibrillator. Left ventricular dysfunction was also present in this group, as their mean ejection fraction was 26%. Implantation techniques. The surgical approach for implantation was dictated by the clinical needs of the patient and has been described. 10. I I Briefly, patients with previous cardiac surgical procedures underwent implantation of the defibrillator via a lateral thoracotomy (Fig. 3). In the candidates undergoing combined cardiac procedures and defibrillator implantation, median sternotomy was employed. In the patients who had not had a previous operation and in whom no other cardiac procedure was indicated , the simpler subxiphoid technique was used (Fig. 4). With the AID-B model, two closely spaced (I em) screw-on myocardial electrodes were placed during thoracotomy or median sternotomy (Fig. 5). These provided the signals for the rate channel. When the subxiphoid approach was used, a bipolar endocardial catheter lead was inserted percutaneously and advanced to the right ventricular apex in a manner identical to that used for endocardial pacemaker lead insertion. Postoperative evaluation. Postoperative care included general medical management and adjustment of antiarrhythmic therapy when indicated. Programmed electrical stimulation with an aggressive protocol including the use of alternating current" was performed 7 to 30 days after operation to determine inducibility, to evaluate the efficacy of antiarrhythmic drug therapy, and to document effective function of the implanted
device. After discharge , outpatient visits were required every 2 to 3 months for noninvasive testing of the pulse generator to assess battery life and to reevaluate medical therapy. The determination that an automatic out-of-hospital resuscitation from a malignant arrhythmia had occurred was made on the basis of a characteristic sequence of events described by the patient or bystanders. This sequence included palpitations, weakness, dizziness, and often cardiovascular collapse and evidence of an internal discharge consisting of a sudden brief muscular contractions of the thorax. This was followed by almost immediate recovery and with restoration of hemodynamics. The diagnosis needed unanimous agreement of the investigators with regard to the circumstances surrounding the event and its interpretation. Actuarial analysis of the deaths that occurred in patients who received the implant was done by means of the Kaplan-Meier! ' method. In this analysis, patients who did not receive pulse generators and the operative deaths were excluded. A hypothetical estimate of the late deaths that would have occurred in the same group of patients without implantation of the automatic defibrillator was also performed. In order to derive these hypothetical estimates, a nonsurvival end point was defined as the patient's actual death or the first automatic resuscitation out of the hospital, whichever occurred first. This definition was based on the assumption that patients whose malignant arrhythmias were automatically terminated outside the hospital would have died without the implanted device. In patients with several automatic resuscitations, only the first event was
The Journal of Thoracic and Cardiovascular
3 8 4 Watkins et al.
De fib r ill a t ing e le c !r"o d e c ov ers ~ 1:!opical per lcZirdiu \.
,
Surgery
Table I. Automatic resuscitations of sudden
death episodes No. of patients
Group I In hospital Out of hospital Group II In hospital Out of hospital
Fig. 3. implantation of the defibrillator via a lateral thoracotomy.
counted as an end point. For actual survival to be compared with the predicted mortality, the 95% confidence limits for the actual survival experience have been calculated. The longest follow-up has been 34 months; the average was 15.6 months. Results Operative results. In Group I, 37 patients underwent implantation of the automatic defibrillator via thoracotomy or the subxiphoid technique. In two of these patients only the leads were implanted. There were no surgical deaths. One patient required reexploration for postoperative bleeding. Twenty-eight patients in Group II underwent implantation of the defibrillator. This was combined with endocardial resection and aneurysmectomy in 17, coronary artery bypass grafting in seven, and coronary artery bypass grafting and mitral valve replacement in four. In five of these patients only the leads were implanted. In Group II there were four operative deaths, two of which occurred during emergency procedures. One patient required reexploration for mediastinal bleeding. Arrhythmia control. In the hospital, under monitored conditions, the implanted device correctly identified and terminated 30 episodes of spontaneous ventricular tachyarrhythmias in six Group I patients and 11 episodes in one Group II patient. Following discharge,
No. of resuscitations
6
30
II
44
I 2
II 4
44 out-of-hospital resuscitations were observed in 11 Group I patients and four were observed in two Group II patients (Table I). Late mortality. There have been a total of 14 late deaths in this series; 13 in Group I and one in Group II. Late deaths were due to the progression of end-stage myocardial disease in three, myocardial infarction in three, and cerebral vascular accident in one. One patient had electromechanical dissociation. Four presumably died of unwitnessed sudden death episodes (Fig. 6). For Group I, actuarial display of mortality with the defibrillator implanted is shown in Fig. 7. The I and 2 year actual survival rates are 75% and 65%, respectively. The bottom curve is a hypothetical estimate of survival had the out-of-hospital resuscitations not been successful. The mortality for Group II , excluding the operative deaths, is displayed in Fig. 8 and demonstrates 1 and 2 year survival rates of 95%. The bottom curve, again , shows the hypothetical estimate of survival had the out-of-hospital resuscitations failed . Discussion After more than 3 years of clinical evaluation, considerable evidence now indicates that the automatic defibrillator is an effective and reliable device capable of accurate diagnosis and prompt termination of malignant ventricular arrhythmias in man.4-7.9-1 4The surgical techniques for implantation have been found safe, and serious hazards from long-term implantation of the device have not been observed. The few problems which appeared during the early part of the study seem to have been overcome by appropriate design modifications . These changes eventually led to the development of a more advanced model with a broader diagnostic and therapeutic spectrum. The analysis of Group I patients provides a unique opportunity for a close evaluation of the functional performance of the implanted defibrillator. All of these patients had multiple arrhythmic arrests which remained inducible during programmed electrical stim-
Volume 86 Number 3
Automatic implantable defibrillator
September, 1983
385
Pe r ic a r d i a l /! fl a ps •...•.. ..
/
I
A p ex of He ad
i i
•.••....-:
I ncis io n in
pe r t c a t- d tum
al att achme n t t o di aph) -a ~ ~
A:'
~ :.
;l o.> _ .
Fig. 4. Implantation of the defibr illator by the subxiphoid technique. RV, Right ventricle. LV. Left ventricle.
D eftbr
i llahn
~
e leclro de
c overs ap ical myocar d iu m
Fig. 5. Placement of the defibrillator via thoracotomy or median sternotomy.
ulation, despite multiple antiarrhythmic drug trials. Because of drug resistance and the inability to offer additional surgical interventions, the underlying myocardial process responsible for the arrhythmias remained essentially unchanged. This undoubtedly accounted for the large number (74) of malignant arrhythmias observed in this group. These potentially lethal episodes were correctly terminated by the implanted device. The 44 successful automatic resuscitations outside the hospital setting are particularly important, because they appear to be largely responsible
for the 75% 1 year survival rate. Deaths in this group were due mainly to the progression of native disease and were possibly arrhythmic in nature in only four instances . Group II patients were found suitable for more definitive cardiac operations. More than half underwent endocardial resection, which controlled ventricular tachycardia in over 90% of those so treated. As a whole, Group II patients fared much better. In this group only four out-of-hospital sudden death episodes were observed and these were automatically terminated. Of
The Journal of Thoracic and Cardiovascular Surgery
3 8 6 Watkins et al.
"!:
95% ACTIlAL SURVIVAL- CROUP II HYPOTHETICAL SUllVIVAL- CROUP II
"
z
.8
>
;: a:
~
'"
.6
z
... 0
De 0
•4
"0
De
Do
.2 B
•
.5
1.5
2
2.5
3
YEARS ELAPSED
Fig. 8. The I year and 2 year survival rates. Fig. 6. Late mortality. new surgical techniques, and defibrillator implantation should all be viewed as complementary. We express our gratitude to Mr. Paul Freund, P.A., Ms. Sue Johnson, P.A., and Ms. Jennifer Becker for technical support and to Ms. Irma Silkworth for secretarial assistance.
> .8 > 0:
ACTIlAL SURVIVAL- CROUP I
~
'" 5
.6
to:
o
l5
0:
.4
"-
.2
2 B
B
.5
1.5
2
2.5
:3
YEARS ELAPSED
Fig. 7. Survival rates.
the patients surviving operation, 95% are alive at 2 years. This experience suggests that whenever possible an aggressive definitive cardiac surgical approach is warranted and will decrease the likelihood of malignant arrhythmias occurring in the years after the operation. Implantation of the defibrillator is also desirable because it provides additional protection against sudden death secondary to late arrhythmic events. The favorable results of the clinical trials support continued clinical use and further evaluation of the automatic defibrillator. It appears from these data and previously published information 14 that effective function of the device may prolong survival in selected high-risk patients," However, the implanted defibrillator should not be viewed as a substitute for antiarrhythmic or surgical therapy. The new antiarrhythmic drugs, the
3
4
5
6
REFERENCES Mirowski M, Mower MM, Staewen WS, Tabatznik, B, Mendeloff AL: Standby automatic defibrillator. An approach to prevention of sudden coronary death. Arch Intern Med 126:158-161, 1970 Mirowski M, Mower MM, Langer A, Heilman MS, Schreibman J: A chronically implanted system for automatic defibrillation in active conscious dogs. Experimental model for treatment of sudden death from ventricular fibrillation. Circulation 58:90-94, 1978 Mirowski M, Mower MM, Bhagavan BS, Langer A, Kolenik SA, Fischell RE, Heilman MS: Chronic animal and bench testing of the implantable automatic defibrillator, Chap 27-2, Pacesymp: Proceedings of the Sixth World Symposium on Cardiac Pacing, C Meere, ed., Montreal, 1979. Mirowski M, Mower MM, Reid PR, Watkins L Jr, Langer A: The automatic implantable defibrillator. New modality for treatment of life-threatening ventricular arrhythmias. PACE 5:384-401,1982 Mirowski M, Reid PR, Mower MM, Watkins L, Gott VL, Schauble JF, Langer A, Heilman MS, Kolenik SA, Fischell RF, Weisfeldt ML: Termination of malignant ventricular arrhythmias with an implanted automatic defibrillator in human beings. N Engl J Med 303:322-324, 1980 Mirowski M, Reid PR, Watkins L Jr, Weisfeldt ML, Mower MM: Clinical treatment of life-threatening ventricular tachyarrhythmias with the automatic implantable defibrillator. Am Heart J 102:265-270, 1981
Volume 86 Number 3 September, 1983
7 Mirowski M, Reid PR, Mower MM, Watkins L Jr: Successful conversion of out-of-hospital life-threatening arrhythmias with the implanted automatic defibrillator. Am Heart J 103:147-148, 1982 8 Langer A, Heilman MS, Mower MM, Mirowski M: Considerations in development of the automatic implantable defibrillator. Med Instrum 10:163-167, 1976 9 Reid PR, Mirowski M, Mower MM, Platia EV, Griffith LSC, Watkins L Jr, Bach SM, Imran M, Thomas A: Clinical evaluation of the internal automatic cardioverterdefibrillator in survivors of sudden cardiac death. Presented in part at the American Heart Scientific Sessions, Dallas, Texas, 1982. 10 Watkins L Jr, Mirowski M, Mower MM, Reid PR, Griffith LSC, Vlay SC, Weisfeldt ML, Gott VL: Automatic defibrillation in man. The initial surgical experience. J THoRAc CARDIOVASC SURG 82:492-500, 1981 11 Watkins L Jr, Mirowski M, Mower MM, Reid PR, Freund PA, Thomas A, Weisfeldt ML, Gott VL: Implantation of the automatic defibrillator. The sub-xiphoid approach. Ann Thorac Surg 34:515-520, 1982 12 Mower MM, Reid PR, Watkins L Jr, Mirowski M: Use of alternating current during diagnostic electrophysiologic studies. Circulation 67:69-72, 1983 13 Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 14 Mirowski M, Reid PR, Winkle RA, Mower MM, Watkins L Jr, Stinson EB, Griffith LSC, Kallman CH, Weisfeldt ML: Mortality in patients with implanted automatic defibrillators. Ann Intern Med 98:585-588, 1983
Discussion DR. ALDEN H. HARKEN Phi/adelphia, Pa.
Dr. Watkins and the group at Johns Hopkins have made substantial and innovative progress toward the surgical therapy of ventricular fibrillation and associated "sudden death." They are to be congratulated. We know that a substantial number of persons in the United States die suddenly, perhaps as many as 400,000 a year. One might guess how many of that
Automatic implantable defibrillator
387
group could be helped by this kind of procedure. In addition, when these patients get to the hospital, following successful resuscitation, what kind of electrophysiological testing should they undergo? DR. WATKINS (Closing) Thank you very much for the interesting and stimulating discussion. To address the question regarding the numbers of patients involved, it is estimated that about 400,000 deaths annually result from malignant ventricular arrhythmias. With the advent of modern electrophysiological testing and programmed electrical stimulation, it has become apparent that approximately 75% of that number could potentially be controlled with antiarrhythmic agents if programmed electrical stimulation were used to determine the choice of drugs utilized. The remaining 25% would be potential candidates for operation. These persons should undergo a complete work-up, including cardiac catheterization, angiography, hemodynamic and electrophysiological studies, and particularly programmed electrical stimulation. Analysis of this work-up will reveal that some patients will not be candidates for cardiac operations. Others will be candidates for coronary artery bypass or mitral valve replacement, and a good number will be candidates for endocardial resection. At present we implant the device in all persons who meet the criteria for implantation, that is, survival after one or two cardiac arrests and overall drug refractoriness. We implant the device in all patients undergoing endocardial resection, because we believe that late recurrence rates of 10% or more justify additional long-term protection for these patients. In closing, arrhythmias in the vast majority of patients can be controlled by medications; only the 25% whose arrhythmias are resistant to medications will be candidates for further work-up and operation. The type of operation indicated is determined by the results of the cardiac work-up. Because of the inability to predict accurately which patients will have late recurrences of malignant ventricular arrhythmias, we believe implantation of the device is indicated at present. We believe the defibrillator will become a valuable addition to our therapeutic armamentarium for the treatment of malignant ventricular arrhythmias.