- Email: [email protected]
In vivo study of a new radioisotope-powered cardiac pacer
In vivo study of a new radioisotopepowered cardiac pacer A new radioisotopic pulse generator has been developed. It is 6 em. long, 4.7 em. high, 1.92 em. wide, and weighs 61 Gm. (2 oz.). It is the smallest pulse generator made and has a life expectancy of over 20 years. The circuit is a conventional ventricular-inhibited (V. V.I.) type. In vitro testing has passed all Atomic Energy Commission requirements. The present study is concerned with in vivo testing of the complete pacemaker system, by means of both myocardial and endocardial electrodes, in 20 dogs with and without induced heart block. Extensive testing for electromagnetic compatability was carried out on I animal with induced heart block and a special, fast-rate pulse generator. Based on studies to date, the Atomic Energy Commission has issued a license for limited clinical trial which has already begun at the collaborating institutions.
Nicholas P. D. Smyth, M.D., * George J. Magovern, M.D., ** Ruben G. Ramirez, M.D.,* Mario H. Diaz, M.D.,* Charles M. Dixon, B.S.,** David C. Fecht, B.S., ** and Alvin Johnson, C.V.T., * Pittsburgh, Pa., and Washington, D. C.
A
new radioisotopic pulse generator has been developed, based both on work supported by the Atomic Energy Commission-" and on recent advances in technology by the manufacturer. t The pulse generator is 6 cm. long, 4.7 em. high, and 1.92 em. wide. It weighs 61 Gm. and occupies 33 sq. em. It is the smallest pulse generator made and has a life expectancy of over twenty years (Fig. 1).
A cutaway drawing of the pacer is shown in Fig. 2. A hermetically sealed and welded titanium case surrounds the electronics and nuclear battery. The sealed, welded case protects the electronics and nuclear battery From the Departments of Surgery of the Allegheny General Hospital, Pittsburgh, Pa., and The Washington Hospital Center, Washington, D. C. Supported by Coratomic, Incorporated, Indiana, Pa., and The Potomac Fund for Cardiovascular Research, Washington, D. C. Received for publication Dec. 20, 1974. *From The Washington Hospital Center. **From The Alleghany General Hospital. tCoratomic, Incorporated, Indiana, Pa.
2
from corrosive body fluids, water vapor, or gaseous contaminants, and it also electromagnetically shields the pacer electronics from spurious signals or electromagnetic interference. Within the hermetically sealed titanium enclosure, the electronics and nuclear battery are cushioned to provide for shock and vibration isolation of these two critical subassemblies. The electronic components are the highest reliability components available and are mounted on epoxy fiberglass printed circuit boards. A toroidal transformer in a DC/DC converter provides for increasing the battery output voltage to that required by the transitorized electronic circuit. Redundant ground straps welded to the titanium case provide for the ground connection between the electronics and the anode (or case) of the pacer. The output connectors from the nuclear battery are soldered to the electronic circuit board with high-strength junction interfaces. The thermoelectric radioisotopic battery has been described in detail previously!
Volume 70 Number 1
Radioisotope-powered cardiac pacer
July, 1975
Fig. 1. Photograph of pulse generator.
Pt.-Rh CONNECTOR FEED TfflOUGH
SUTURE HOLE
r ,, ,,
,
NUCLEAR BATTERY
\
ELECTRONICS
--MAGNETIC SWITCH
Fig. 2. Cutaway drawing of pulse generator showing location of major components.
3
4
The Journal of
Smyth et at.
Thoracic and Cardiovascular Surgery
7
7 ACUTE THRESHOLDS (Average values) 10 Dogs
6
ACUTE THRESHOLDS (Average values) 10 Dogs
6
5
5
4
3
3
2
2
0'----':------'----::-'-:------'---::-'-:------'.15
.30
.50
.75
1.0
1.5
.15
CD
PULSE DURATION, Milliseconds
.30
.50
.75
1.0
1.5
PULSE DURATION, Milliseconds
Fig. 3. Acute strength-duration curves. A, Current. B, Voltage. Electrode surface area 0.175 sq. em. in each case.
The electronics are those of a conventional ventricular-inhibited (V.V.I.) pulse generator." The pulse width is 1 msec. and the pulse amplitude 8 Ma. The basic rate is 70 beats per minute. The noise and magnet rates are the same: 85 beats per minute. The R-wave sensitivity is 1.75 mv. and the refractory period is 275 msec. The unit has successfully passed all required Atomic Energy Commission tests for fire, corrosion, crushing, and vibration. The results of these complex tests have been reported in detail.' With the completion of this bench testing program, the first phase in the development of the pacer was completed. The next phase involved testings for in vivo behavior in animals. Combined with this program was an extensive in vitro and in vivo electromagnetic compatibility test, carried out at the Georgia Institute of Technology." Materials and methods
A joint testing program in animals was initiated early in 1973 in the surgical research laboratories of The Allegheny General Hospital, Pittsburgh, Pennsylvania, and The Washington Hospital Center, Washington, D. C, following a protocol approved
by
the Atomic Energy Commission. Mongrel dogs were used. All animals were anesthetized by intravenous pentobarbital (II mg per kilogram). If thoracotomy was performed, the dogs were intubated and ventilated by a Harvard respirator. Initially, either Cordis* 2 mm. tip or Coratomic L-lO I transvenous unipolar leads were used. They were inserted through the right jugular vein, which was exposed through a neck incision. The lead was then tunneled subcutaneously to be connected to the pulse generator, which was placed in a subcutaneous pocket near the animal's right shoulder to minimize injury to the unit. This implantation technique was used to simulate the technique most commonly used in clinical practice. Instability of the endocardial lead was encountered in several of the animals early in the study-a common problem with endocardial leads in dogs-and the technique was therefore modified. The Allegheny General Hospital group performed right thoracotomies and inserted the leads into the right ventricular myocardium. At the Washington Hospital Center endocardial leads were used in all animals, but with a two'Cordis Corp., Miami, Fla.
Volume 70
Radioisotope-powered cardiac pacer
Number 1
5
July, 1975
7
7 CHRONIC THRESHOLDS (Average values) 10 Dogs
6
6
CHRONIC THRESHOLDS (Average values) 10 Dogs
5
5
4
~
o
>
3
3
2
2
O'--..l---::'::----:o':--~----:'-=__-.J.:_
.15
.30
.50
.75
1.0
1.5
PULSE DURATION, Milliseconds
OL.-,.~15-----::'::---=-':------=':---c-'-::--.......L=.30 .50
o
.75
1.0
1.5
PULSE DURATION, Milliseconds
Fig. 4. Chronic strength-duration curves. A, Current. B, Voltage. Changes from acute values are within acceptable range.
stage technique. The leads were inserted in the conventional way through the right jugular vein and carefully positioned at the apex of the right ventricle under fluoroscopic control. A complete series of stimulating thresholds was then recorded in milliamperage and voltage at pulse widths of 0.15, 0.3, 0.5, 0.75, 1.0 and 1.5 msec. The R wave was also recorded. The lead was then capped and coiled in a subcutaneous pocket lateral to the incision. The pulse generator was not implanted at this time. One month later the animal studied by fluoroscopy or roentgenography, and if the lead appeared to be in satisfactory position the second stage was carried out. The lead was exteriorized and uncapped. The R wave and threshold measurements were repeated. If satisfactory values were obtained, fixation of the lead was assumed and the lead was tunneled subcutaneously to a pocket near the animal's right shoulder, where it was connected to the pulse generator and buried subcutaneously. Conventional electrocardiograms were taken at monthly intervals on all animals. Since the dog's heart rate was much faster than the pulse generator's basic rate, the pulse generator was always inhibited. Magnet rates were therefore determined with
each electrocardiogram to check for capture and rate consistency. In 1 dog intermittent heart block developed spontaneously; in 3 dogs complete heart block was induced, in 1 by a cautery technique" and in 2 by transseptal injection of 40 per cent formalin into the atrioventricular node." Pacing was established at the same operation by means of a Cordis myocardial electrode implanted in the left ventricle and a special fast-rate pulse generator (basic rate 115 pulses per minute). These animals were prepared so that we would have pacer-dependent dogs available for the electromagnetic compatibility studies. Results A number of complications were encountered, not related to the pulse generator. Five animals were reoperated upon for insertion of myocardial leads to replace displaced endocardial leads. Six dogs died, 3 from distemper and empyema, 1 from intussusception and intestinal gangrene, 1 from hemorrhagic pneumonia and gastroenteritis, and 1 from pneumonia. A colony of 20 animals was finally assembled in a kennel licensed by the Atomic Energy Commission for long-term follow-up of the pulse generators.
6
Smyth et al.
The Journal 01 Tho rac ic and Ca rd iovascular Surgery
Fig. S. Electrocardiographic rhythm strip. A, Normal sinus rhythm. Time lines 1 second. Rate 95 beats per minute. B, Magnet applied. Rate at this time is 108 beats per minute. Magnet rate is 92 beats per minute, with capture occurring when the paced beat falls outside the heart's refractory period.
The threshold values obtained acutely are averaged and expressed as strengthduration curves in Fig. 3. Those obtained 1 month later are shown in F ig. 4. The low values in each case confirm good endocardial contact. The records obtained from the animals without atrioventricular block resembled those shown in Fig. 5. Fig. 6 shows the preoperative electrocardiogram and the changes following production of complete atrioventricular block in 1 of the 3 dogs in which heart block was surgically induced. In Fig. 7, the postoperative basic paced rate and magnet rate are shown for the special fast-rate unit implanted in 1 of the animals with surgically induced complete heart block. Electromagnetic interference testing One of the dogs with stable induced heart block and complete pacer dependence
was used in an extensive test program at the Engineering Experiment Station of the Georgia Institute of Technology. Six different electromagnetic environments were chosen : ( 1) 60 Hz magnetic field, (2) 450 Hz magnetic field, (3) 3.1 GHz pulsed field, (4) 915 MHz microwave oven, (5) 2450 MHz microwave oven, and (6) broadband "noise" from automobile ignition systems. The 60 Hz magnetic field test was first done with the test dog in a uniform 60 Hz field generated by a Helmholtz coil pair, with a maximum field intensity in excess of 20 gau ss. No effect was produced even at maximum intensity. Results of all the other tests were negative, at maximum field strengths generated by the equipment. The results of these in vivo tests and the in vitro air and saline immersion tests (also negative) will be reported elsewhere. G To date, 144 dog months of pacer im-
Volume 70 Number 1
Radioisotope-powered cardiac pacer
7
July, 1975
,::!::mi.·
::::
Fig. 7. Electrocardiographic tracing in dog with surgically induced complete heart block and implanted fast-rate pulse generator. Initial basic rate is 115 beats per minute. Line shows application of magnet with immediate change to faster magnet rate of 136 beats per minute. On removal of magnet, pulse generator immediately reverts to basic rate.
The Journal of
8
Smyth et al.
Thoracic and Cardiovascular Surgery
plant have been accumulated. In 1 animal that died, the pulse generator was found to be in its interference mode (noise rate) while the animal was in the refrigerator. When the pulse generator was removed and warmed to room temperature it returned to normal function. The unit was dismantled and a random failure in one of the capacitors was found. Except for this, all of the pulse generators have functioned flawlessly. A license for human implants has been granted by the Atomic Energy Commission, and clinical trial of the pacer began in October, 1974. REFERENCES Shapiro, Z. M., and Purdy, D. L.: Design of Isotopic Generators, Eighth 1apan Conference on Radioisotopes, Paper No. CIE, November 16, 1967. 2 Prosser, M.: Final Term Report, Phase I, lan. 31, 1969, to lan. 31, 1970, Radioisotope Powered Cardiac Pacemaker Program, Contract AT (30-1 )-3731, lune 24, 1970, submitted to the Atomic Energy Commission.
3 Cole, D.: Quarterly Progress Report, Feb. 1, 1973, to April 30, 1974, Radioisotope Powered Cardiac Pacemaker Program, ARCO-3057-l7, prepared for the Atomic Energy Commission Contract AT (11-1 )-3057. 4 Purdy, D. L., Magovern, G. 1., and Smyth, N. P. D.: A New Radioisotope-Powered Cardiac Pacer, 1. THoRAc. CARDIOVASC. SURG. 69: 82, 1975. 5 Parsonnet, Y., Furman, S., and Smyth. N. P. D.: Implantable Cardiac Pacemaker: Status Report and Resource Guideline, Pacemaker Study Group, Report of Inter-Society Commission for Heart Disease Resources, Circulation 50: A-2l, 1974. 6 Toler, 1., Johnson, W., Dixon, C., and Smyth, N. P. D.: Environmental Electromagnetic Compatability of a New Radioisotopic Powered Cardiac Pacemaker. To be published. 7 Smyth, N. P. D., and Magassy, C. L.: Experimental Heart Block in the Dog: An Improved Method, 1. THoRAc. CARDIOVASC. SURG. 59: 201, 1970. 8 Steiner, C., and Kovalik, T. W.: Simple Technique for Production of Chronic Complete Heart Block in Dogs, 1. Appl. Physiol. 25: 631, 1968.
Nicholas P. D. Smyth, M.D., * George J. Magovern, M.D., ** Ruben G. Ramirez, M.D.,* Mario H. Diaz, M.D.,* Charles M. Dixon, B.S.,** David C. Fecht, B.S., ** and Alvin Johnson, C.V.T., * Pittsburgh, Pa., and Washington, D. C.
A
new radioisotopic pulse generator has been developed, based both on work supported by the Atomic Energy Commission-" and on recent advances in technology by the manufacturer. t The pulse generator is 6 cm. long, 4.7 em. high, and 1.92 em. wide. It weighs 61 Gm. and occupies 33 sq. em. It is the smallest pulse generator made and has a life expectancy of over twenty years (Fig. 1).
A cutaway drawing of the pacer is shown in Fig. 2. A hermetically sealed and welded titanium case surrounds the electronics and nuclear battery. The sealed, welded case protects the electronics and nuclear battery From the Departments of Surgery of the Allegheny General Hospital, Pittsburgh, Pa., and The Washington Hospital Center, Washington, D. C. Supported by Coratomic, Incorporated, Indiana, Pa., and The Potomac Fund for Cardiovascular Research, Washington, D. C. Received for publication Dec. 20, 1974. *From The Washington Hospital Center. **From The Alleghany General Hospital. tCoratomic, Incorporated, Indiana, Pa.
2
from corrosive body fluids, water vapor, or gaseous contaminants, and it also electromagnetically shields the pacer electronics from spurious signals or electromagnetic interference. Within the hermetically sealed titanium enclosure, the electronics and nuclear battery are cushioned to provide for shock and vibration isolation of these two critical subassemblies. The electronic components are the highest reliability components available and are mounted on epoxy fiberglass printed circuit boards. A toroidal transformer in a DC/DC converter provides for increasing the battery output voltage to that required by the transitorized electronic circuit. Redundant ground straps welded to the titanium case provide for the ground connection between the electronics and the anode (or case) of the pacer. The output connectors from the nuclear battery are soldered to the electronic circuit board with high-strength junction interfaces. The thermoelectric radioisotopic battery has been described in detail previously!
Volume 70 Number 1
Radioisotope-powered cardiac pacer
July, 1975
Fig. 1. Photograph of pulse generator.
Pt.-Rh CONNECTOR FEED TfflOUGH
SUTURE HOLE
r ,, ,,
,
NUCLEAR BATTERY
\
ELECTRONICS
--MAGNETIC SWITCH
Fig. 2. Cutaway drawing of pulse generator showing location of major components.
3
4
The Journal of
Smyth et at.
Thoracic and Cardiovascular Surgery
7
7 ACUTE THRESHOLDS (Average values) 10 Dogs
6
ACUTE THRESHOLDS (Average values) 10 Dogs
6
5
5
4
3
3
2
2
0'----':------'----::-'-:------'---::-'-:------'.15
.30
.50
.75
1.0
1.5
.15
CD
PULSE DURATION, Milliseconds
.30
.50
.75
1.0
1.5
PULSE DURATION, Milliseconds
Fig. 3. Acute strength-duration curves. A, Current. B, Voltage. Electrode surface area 0.175 sq. em. in each case.
The electronics are those of a conventional ventricular-inhibited (V.V.I.) pulse generator." The pulse width is 1 msec. and the pulse amplitude 8 Ma. The basic rate is 70 beats per minute. The noise and magnet rates are the same: 85 beats per minute. The R-wave sensitivity is 1.75 mv. and the refractory period is 275 msec. The unit has successfully passed all required Atomic Energy Commission tests for fire, corrosion, crushing, and vibration. The results of these complex tests have been reported in detail.' With the completion of this bench testing program, the first phase in the development of the pacer was completed. The next phase involved testings for in vivo behavior in animals. Combined with this program was an extensive in vitro and in vivo electromagnetic compatibility test, carried out at the Georgia Institute of Technology." Materials and methods
A joint testing program in animals was initiated early in 1973 in the surgical research laboratories of The Allegheny General Hospital, Pittsburgh, Pennsylvania, and The Washington Hospital Center, Washington, D. C, following a protocol approved
by
the Atomic Energy Commission. Mongrel dogs were used. All animals were anesthetized by intravenous pentobarbital (II mg per kilogram). If thoracotomy was performed, the dogs were intubated and ventilated by a Harvard respirator. Initially, either Cordis* 2 mm. tip or Coratomic L-lO I transvenous unipolar leads were used. They were inserted through the right jugular vein, which was exposed through a neck incision. The lead was then tunneled subcutaneously to be connected to the pulse generator, which was placed in a subcutaneous pocket near the animal's right shoulder to minimize injury to the unit. This implantation technique was used to simulate the technique most commonly used in clinical practice. Instability of the endocardial lead was encountered in several of the animals early in the study-a common problem with endocardial leads in dogs-and the technique was therefore modified. The Allegheny General Hospital group performed right thoracotomies and inserted the leads into the right ventricular myocardium. At the Washington Hospital Center endocardial leads were used in all animals, but with a two'Cordis Corp., Miami, Fla.
Volume 70
Radioisotope-powered cardiac pacer
Number 1
5
July, 1975
7
7 CHRONIC THRESHOLDS (Average values) 10 Dogs
6
6
CHRONIC THRESHOLDS (Average values) 10 Dogs
5
5
4
~
o
>
3
3
2
2
O'--..l---::'::----:o':--~----:'-=__-.J.:_
.15
.30
.50
.75
1.0
1.5
PULSE DURATION, Milliseconds
OL.-,.~15-----::'::---=-':------=':---c-'-::--.......L=.30 .50
o
.75
1.0
1.5
PULSE DURATION, Milliseconds
Fig. 4. Chronic strength-duration curves. A, Current. B, Voltage. Changes from acute values are within acceptable range.
stage technique. The leads were inserted in the conventional way through the right jugular vein and carefully positioned at the apex of the right ventricle under fluoroscopic control. A complete series of stimulating thresholds was then recorded in milliamperage and voltage at pulse widths of 0.15, 0.3, 0.5, 0.75, 1.0 and 1.5 msec. The R wave was also recorded. The lead was then capped and coiled in a subcutaneous pocket lateral to the incision. The pulse generator was not implanted at this time. One month later the animal studied by fluoroscopy or roentgenography, and if the lead appeared to be in satisfactory position the second stage was carried out. The lead was exteriorized and uncapped. The R wave and threshold measurements were repeated. If satisfactory values were obtained, fixation of the lead was assumed and the lead was tunneled subcutaneously to a pocket near the animal's right shoulder, where it was connected to the pulse generator and buried subcutaneously. Conventional electrocardiograms were taken at monthly intervals on all animals. Since the dog's heart rate was much faster than the pulse generator's basic rate, the pulse generator was always inhibited. Magnet rates were therefore determined with
each electrocardiogram to check for capture and rate consistency. In 1 dog intermittent heart block developed spontaneously; in 3 dogs complete heart block was induced, in 1 by a cautery technique" and in 2 by transseptal injection of 40 per cent formalin into the atrioventricular node." Pacing was established at the same operation by means of a Cordis myocardial electrode implanted in the left ventricle and a special fast-rate pulse generator (basic rate 115 pulses per minute). These animals were prepared so that we would have pacer-dependent dogs available for the electromagnetic compatibility studies. Results A number of complications were encountered, not related to the pulse generator. Five animals were reoperated upon for insertion of myocardial leads to replace displaced endocardial leads. Six dogs died, 3 from distemper and empyema, 1 from intussusception and intestinal gangrene, 1 from hemorrhagic pneumonia and gastroenteritis, and 1 from pneumonia. A colony of 20 animals was finally assembled in a kennel licensed by the Atomic Energy Commission for long-term follow-up of the pulse generators.
6
Smyth et al.
The Journal 01 Tho rac ic and Ca rd iovascular Surgery
Fig. S. Electrocardiographic rhythm strip. A, Normal sinus rhythm. Time lines 1 second. Rate 95 beats per minute. B, Magnet applied. Rate at this time is 108 beats per minute. Magnet rate is 92 beats per minute, with capture occurring when the paced beat falls outside the heart's refractory period.
The threshold values obtained acutely are averaged and expressed as strengthduration curves in Fig. 3. Those obtained 1 month later are shown in F ig. 4. The low values in each case confirm good endocardial contact. The records obtained from the animals without atrioventricular block resembled those shown in Fig. 5. Fig. 6 shows the preoperative electrocardiogram and the changes following production of complete atrioventricular block in 1 of the 3 dogs in which heart block was surgically induced. In Fig. 7, the postoperative basic paced rate and magnet rate are shown for the special fast-rate unit implanted in 1 of the animals with surgically induced complete heart block. Electromagnetic interference testing One of the dogs with stable induced heart block and complete pacer dependence
was used in an extensive test program at the Engineering Experiment Station of the Georgia Institute of Technology. Six different electromagnetic environments were chosen : ( 1) 60 Hz magnetic field, (2) 450 Hz magnetic field, (3) 3.1 GHz pulsed field, (4) 915 MHz microwave oven, (5) 2450 MHz microwave oven, and (6) broadband "noise" from automobile ignition systems. The 60 Hz magnetic field test was first done with the test dog in a uniform 60 Hz field generated by a Helmholtz coil pair, with a maximum field intensity in excess of 20 gau ss. No effect was produced even at maximum intensity. Results of all the other tests were negative, at maximum field strengths generated by the equipment. The results of these in vivo tests and the in vitro air and saline immersion tests (also negative) will be reported elsewhere. G To date, 144 dog months of pacer im-
Volume 70 Number 1
Radioisotope-powered cardiac pacer
7
July, 1975
,::!::mi.·
::::
Fig. 7. Electrocardiographic tracing in dog with surgically induced complete heart block and implanted fast-rate pulse generator. Initial basic rate is 115 beats per minute. Line shows application of magnet with immediate change to faster magnet rate of 136 beats per minute. On removal of magnet, pulse generator immediately reverts to basic rate.
The Journal of
8
Smyth et al.
Thoracic and Cardiovascular Surgery
plant have been accumulated. In 1 animal that died, the pulse generator was found to be in its interference mode (noise rate) while the animal was in the refrigerator. When the pulse generator was removed and warmed to room temperature it returned to normal function. The unit was dismantled and a random failure in one of the capacitors was found. Except for this, all of the pulse generators have functioned flawlessly. A license for human implants has been granted by the Atomic Energy Commission, and clinical trial of the pacer began in October, 1974. REFERENCES Shapiro, Z. M., and Purdy, D. L.: Design of Isotopic Generators, Eighth 1apan Conference on Radioisotopes, Paper No. CIE, November 16, 1967. 2 Prosser, M.: Final Term Report, Phase I, lan. 31, 1969, to lan. 31, 1970, Radioisotope Powered Cardiac Pacemaker Program, Contract AT (30-1 )-3731, lune 24, 1970, submitted to the Atomic Energy Commission.
3 Cole, D.: Quarterly Progress Report, Feb. 1, 1973, to April 30, 1974, Radioisotope Powered Cardiac Pacemaker Program, ARCO-3057-l7, prepared for the Atomic Energy Commission Contract AT (11-1 )-3057. 4 Purdy, D. L., Magovern, G. 1., and Smyth, N. P. D.: A New Radioisotope-Powered Cardiac Pacer, 1. THoRAc. CARDIOVASC. SURG. 69: 82, 1975. 5 Parsonnet, Y., Furman, S., and Smyth. N. P. D.: Implantable Cardiac Pacemaker: Status Report and Resource Guideline, Pacemaker Study Group, Report of Inter-Society Commission for Heart Disease Resources, Circulation 50: A-2l, 1974. 6 Toler, 1., Johnson, W., Dixon, C., and Smyth, N. P. D.: Environmental Electromagnetic Compatability of a New Radioisotopic Powered Cardiac Pacemaker. To be published. 7 Smyth, N. P. D., and Magassy, C. L.: Experimental Heart Block in the Dog: An Improved Method, 1. THoRAc. CARDIOVASC. SURG. 59: 201, 1970. 8 Steiner, C., and Kovalik, T. W.: Simple Technique for Production of Chronic Complete Heart Block in Dogs, 1. Appl. Physiol. 25: 631, 1968.