- Email: [email protected]
Surgical treatment of ventricular tachycardia by balloon electric shock ablation
Surgical treatment of ventricular tachycardia by balloon electric shock ablation Potential effects on the mitral valve apparatus We have previously described a new surgical technique for control of arrhythmogenic foci in patients with recurrent ventricular tachycardia that we caD balloon electric shock ablation. With this method sequential shocks are delivered to a grid of electrodes on a baDoon that can be introduced across the mitral valve into the intact ventricle. A series of experiments was undertaken to investigate possible deleterious effects of baDoon electric shock ablation when shocks are delivered directly to the mitral valve apparatus. In six animals shocks totaling 1200 joules were given through a closely spaced electrode grid applied to the area of the mitral valve. Nine to 12 weeks later, left ventricular and mitral valve function were assessed. BaDoon electric shock ablation in the basilar portion of the ventricle was associated with decreased myocardial performance, as evidenced by ejection phase indices. In five of six animals baDoon ablation led to minor thickening of the valve leaflets and chordal attachments plus necrosis of adjacent myocardium, including papillary muscles. In these animals there was no significant dysfunction of the valve observed. In the remaining animal, however, ablation was centered on the posterior papillary muscle and resulted not only in necrosis of the base of the papillary muscle but also in fuD-thickness scarring and thinning of the adjacent left ventricular waD. In this dog, mitral regurgitation was seen on long-term foDow-up. We conclude that when baBoon electric shock ablation is used to destroy a localized area of myocardium in the basilar portion of the intact ventricle, the procedure results in decreased myocardial performance. When shocks were directly applied to the mitral valve apparatus in five of six animals, ablation did not result in significant negative effects on the structure and function of the valve. In the sixth dog, however, shock delivery resulted in transmural necrosis and thinning at the site of papillary muscle insertion and was associated with severe mitral regurgitation with volume loading. Therefore caution should be used when considering clinical application of this technique if the area to be ablated is in the basal portion of the heart. (J THORAC CARDIOVASC SURG 1992;103:629-37)
Lynda L. Mickleborough, MD, Gregory J. Wilson, MD, Akihiko Usui, MD, Tadashi Isomura, MD, Alexandre Varela, MD, Harry Rakowski, MD, and Gordon Gray, BSc, Toronto, Ontario, Canada
From the Division of Cardiovascular Surgery, The Toronto Hospital, the Department of Pathology, The Hospital for Sick Children, and the Departments of Surgery, Pathology, and Medicine, University of Toronto, Toronto, Ontario, Canada. Supported by the Canadian Heart Foundation, the Heart and Stroke Foundation of Ontario, and the Cardiovascular Research Fund, University of Toronto. Received for publication May 29, 1990. Accepted for publication Feb. 18, 1991. Address for reprints: Lynda L. Mickleborough, MD, Division of Cardiovascular Surgery, EN 13-217, The Toronto Hospital, 200 Elizabeth St., Toronto, Ontario, M5G 2C4 Canada.
12/1/28928
Map-directed operations offer a potential cure for patients with life-threatening ventricular arrhythmias.' Most surgical candidates have extensive coronary artery disease and a history of previous infarction. Preoperative ventricular function is often significantly compromised, but only 50% of patients have a clear-cut area of transmural scar or resectable aneurysm. 1-4 Most arrhythmias associated with ischemic heart disease appear to arise from the subendocardium. With standard approaches for endocardial mapping and ablation, a ventriculotomy incision is required to obtain access to the inner layers of the heart. In patients with ventricular tachycardia (VT),
629
The Journal of Thoracic and Cardiovascular
6 30 Mickleborough et al.
Surgery
assessed by echocardiography before and 9 to 12 weeks after balloon ablation. In addition, LV hemodynamic function was evaluated after ablation in these animals and compared with function in a control group not subjected to the procedure. Methods
Fig. 1. Balloon used for electric shock ablation. There are 12 silver bead electrodes through which shocks are delivered.
the ventriculotomy and scarring associated with its healing may contribute to perioperative mortality and postoperative morbidity.v 6 We have recently described a transatrial balloon approach for intraoperative mapping that can be performed in the intact ventricle.' On the basis of previous experience with catheter ablation and as an extension of the transatrial mapping approach, once mapping has been completed we have delivered electric current through selected electrodes on the balloon array to destroy specific "target areas."? We call this procedure balloon electric shock ablation or BESA. Because the same electrode array is used for mapping and ablation, there is no need to precisely orient the balloon array within the intact ventricle. We have previously demonstrated in normal dog hearts that apical delivery of shocks (100 joules each) through a grid of electrodes spaced at 1 em intervals can produce a well-demarcated layer of homogeneous scar.! This series of experiments was designed to demonstrate possible deleterious effects of basilar shock delivery through a 0.5 em grid of electrodes directly applied to the mitral valve apparatus and adjacent papillary muscles. Left ventricular (LV) and mitral valve function were
Description of balloon apparatus used for BESA. A latex double-layered No.7 balloon is covered with expandable mesh. Twelve silver bead electrodes are attached to Teflon-insulated 36-gauge stainless steel wires and sutured to the mesh in a closely spaced three-row array (Fig 1). The wires are attached to a connector that allows electric shocks to be delivered to selected beads. Echocardiographic assessment- BESA group. All dogs in this group underwent serial echocardiographic evaluations. The dogs were mildly sedated with acepromazine (Atrovet), 0.055 mg/kg, By means of a commercially available Hewlett-Packard 77020a scanner (Hewlett-Packard Company, Andover, Mass.), two-dimensional pulsed and color Doppler flow recordings were obtained from the apical and parasternal windows.The presence or absence of mitral regurgitation was established and quantitated by color flow mapping in multiple views."LV diastolic and systolic dimensions, as well as left atrial dimensions, were measured. Percent basal LV shortening was calculated by the formula: Percent basal LV shortening = (LV diastolic - LV systolic dimension) /Diastolic dimension. Mitral valve and LV function were assessed in a similar fashion 1 week after the BESA procedure, and a third study was performed during volume loading 9 to 12 weeks later. This final study was performed in an open chest preparation by direct application of the echocardiographic probe to the heart. During the follow-up studies we looked for the presence of local wall motion abnormalities related to the previous ablation. Operative procedure-BESA group. Six adult mongrel dogs (30 to 36 kg) underwent a left thoracotomy and were supported with cardiopulmonary bypass at 37° C. A series of 200joule shocks were delivered to adjacent 0.5 em electrode pairs with Spacelabs defibrillator (Spacelabs Inc., Redmond, Wash., model 4045, which provides a damped sinusoidal discharge, energy 100 joules per bead). The bead electrodes served as a cathode, and a standard cautery plate (15 x 25 cm-) applied to the right side of the thorax acted as the anode. A Tektronix storage oscilloscope (Tektronix Inc., Beaverton, Ore.) monitored the voltage and current waveforms during delivery of the shocks. If VT or ventricular fibrillation occurred, defibrillation was performed with internal paddles (20 W . sec). After delivery of the shocks, the balloon was deflated and removed. Bypass support was continued for 20 to 30 minutes (total pump time 60 minutes). Air was removed from the left side of the heart, and the left atrial incision was closed. The dogs were weaned from bypass and decannulated. After closure of the thoracotomy incision, they were allowed to recover under careful hemodynamic monitoring. They were kept under observation, and repeat echocardiographic studies to assess LV function, regional wall motion, and mitral regurgitation were performed 1 week postoperatively and 9 to 12 weeks later. Hemodynamic assessment-BESA group and control group. In control animals and in the BESA animals 9 to 12
Volume 103 Number 4 April 1992
Surgical treatment of VT by RESA
Table I. Echocardiographic assessment Preop.
Mitral regurgitation. Basal fractional shortening (%) Left atrial size (rnm) 'Different from preop. (p
0-1+ 33 ± 7.5
32 ± 3
Early postop. 0-1+ 14.3 ± 5.9*
29 ± 4
63 I
Table II. Hemodynamic response to volume loading Late postop. 0-1+ (n = 5) 2+-4+ (n = I) 25.5 ± 8.7*
33 ± 3
< 0.01).
weeks after the initial operative procedure, LV function was assessed during volume loading. The dogs were premedicated (atropine sulfate 0.6 mg) and anesthetized (thiopental sodium 30 mg/kg). The lungs were ventilated with a volume respirator. An arterial line and a triple-lumen thermodilution Swan-Ganz catheter from Baxter Healthcare Corp., Irvine, California, were inserted. Through a left thoracotomy, a left atrial catheter was inserted. Myocardial function was then evaluated by nuclear ventriculograms. Red cells were labeled with stannous pyrophosphate (2.5 mg) injected directly into the cephalic vein and technetium 99m pertechnetate (600 MBq) injected 25 minutes later. A Technicare Sigma 420 gamma camera (GE Medical Systems, Milwaukee, Wis.) with a high-sensitivity parallel-hole collimator was placed in the best septal view (approximately 45-degree left anterior oblique view) to maximize left and right ventricular separation. A Medical Data Systems A2/A3 computer (Medical Data Systems, Ann Arbor, Mich.) was interfaced with the gamma camera. Scintigraphic data were acquired during 2-minute periods with 16 frames per RR interval. The LV ejection fraction was calculated from commercially available software with a semiautomated edge-detection program. Hemodynamic variables, including heart rate, systolic blood pressure, diastolic blood pressure, mean arterial pressure, left atrial pressure, and cardiac output, were recorded. Cardiac output was indexed to body weight in kilograms. LV stroke volume and stroke work were calculated by standard formulas'P II and indexed to body weight in kilograms. LV end-diastolic volume index (L VEDVI) was calculated from the nuclear-derived ejection fraction (EF) and the thermodilution stroke index (SI) by the formula: LVEDVI = (SI/EF). LV end-systolic volume index was calculated as the difference between the LVEDVI and the stroke index. During volume loading the effects of change in preload (left atrial pressure and EDVI) on cardiac index, stroke index, and LV stroke work index were assessed. Ten to fifteen measurements were made in each animal. Volume loading and unloading resulted in a reproducible change in preload (EDVI) and afterload (mean arterial pressure). In the BESA dogs echocardiographic assessment of mitral regurgitation was repeated with color flow Doppler techniques at peak volume loading. Myocardial performance was evaluated by plotting cardiac index and stroke work index against EDVI. Left atrial pressure was plotted against EDVI, and systolic blood pressure was plotted against end-systolic volume index (ESVI). A linear regression analysis was performed for each of these relationships. To display and analyze the information, we chose a low-volume (EDVI 0.5 to 1.5 ml/kg) and a high-volume (EDVI 2.0 to 3.5
Groups
Before volume loading (LVEDVI 0.5-1.5
After volume loading (LVEDVI 2.()"3.5
mlfkg}
mlfkg)
Pulse (beat/min) Control 120 ± 13 BESA 132 ± 23 Systolic blood pressure (mm Hg) Control 127 ± 29 BESA 119 ± 18 Diastolic blood pressure (mm Hg) Control 92 ± 21 BESA 85 ± 12 Mean arterial pressure (mm Hg) Control 104 ± 23 BESA 94 ± 14 Left atrial pressure (mm Hg) Control 4.4 ± 2.4 BESA 4.4 ± 2.9 Cardiac index (Lyrnin/kg) 0.09 ± 0.Q2 Control BESA 0.08 ± 0.01 Stroke index (nil/kg) Control 0.76 ± 0.20 BESA 0.58 ± 0.13 Stroke workindex (gm . rri/kg) Control \.04 ± 0.37 BESA 0.74 ± 0.25 Ejection fraction (%) Control 67 ± 13 BESA 48 ± 9 EDVI (ml/kg) Control 1.14 ± 0.30 BESA 1.23 ± 0.25 ESVI (ml/kg) Control 0.38 ±0.20 BESA 0.64 ± 0.21 Values are mean ± standard deviation. 'Different from before volume loading (p
119 ± 8 129 ± 16 173 ± 21* 142 ± 14* 120 ± 18* 103 ± 14*,t 138 ± 18* 116 ± 13* 8.8 ± 1.9* I\.6 ± 3.9* 0.14 ± 0.01:1: 0.11 ± 0.01*, § \.20 ± 0.12:1: 0.83 ± 0.14*,§ 2.10 ± 0.37:1: 1.17 ± 0.23*,§ 45 ± 6:1: 32 ± 4:1: 2.72 ± 0.52:1: 2.57 ± 0.40:1: 1.5 ± 0.43:1: 1.7 ± 0.31:1:
< 0.05).
[Different from control group (p < 0.05). :\:Differentfrom before volume loading (p §Different from control group (p
< 0.01).
< 0.01).
ml/kg) measurement from the observations in each dog. These values were used to evaluate myocardial performance and the systolic and diastolic pressure-VOlume relationships for each group. All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication No. 80-23, revised 1978). Pathologic studies-BESA group. After hemodynamic assessment the animals were killed and the hearts excised. The left atrium and left ventricle were opened, and the mitral valve apparatus was examined and photographed. The dimensions of the apparent endocardial scar were recorded. The myocardial lesion was then sectioned. The depth of the lesion at its center
The Journal of Thoracic and Cardiovascular Surgery
6 3 2 Mickleborough et al.
200
15
L 10
L
150
L
OJ
I
E E
n,
« ....J
OJ
I
5
E E 100
1t
o
.:
11
o,
en CfJ
• control eBESA
2
3
50
LVEDVI ml/kg
Fig. 2. Diastolic pressure-volume relationship with volume loading is depicted for control group and BESA group. Mean and standard error of left atrial pressure (LAP) at high and low left ventricular end-diastolic volume indices (LVEDVI) are demonstrated. By analysis of covariance, there was no statistically significantdifference in increase in left atrial pressure with volume loading between the two groups.
was measured, and lesion size was calculated (major axis X minor axis X depth). Transmural LV tissue blockswere taken for histologic studies from the area of intervention, from the border zone (interface of scar with apparently normal myocardium), and from areas remote from the site of ablation. Tissue blocks were fixed with 10% buffered formaldehyde, dehydrated, and embedded in paraffin, and 5 ~m thick sectionswere cut and stained for histologicstudy with hematoxylin-phloxinesaffron, Masson's trichrome, picro-Mallory, and Movat's pentachrome stains. Statistical analysis. Statistical analysis was performed with the Statistical Analysis Program (SAS Institute Inc., Cary, N.e.). Differencesbetween groups were evaluated by x2 statistics for discrete variables and an analysis of covariance for continuous variables. Myocardial performance, diastolic properties, and systolic function were evaluated by a two-way analysis of covariance with the high- and low-volume measurements. Continuous variables are summarized as the mean ± standard deviation in the text and tables and as mean ± standard error in the illustrations. Results Preoperative echocardiographic assessmentBESA group. All dogs had either no mitral regurgitation or 1 + mitral regurgitation (on a scale of 0 to 4+) when studied preoperatively. All ventricles appeared normal with no evidence of localized wall motion abnormalities. The percent basal fractional LV shortening was
• control • BESA
o
2
3
LVESVI ml/kg
Fig. 3. Systolic pressure-volume relationship with volume loading is depicted for the two groups. Mean and standard error of systolicbloodpressure (SBP) at high and low LV end-systolic volume indices (LVESVI) are demonstrated. By analysis of covariance there was a significantlygreater increase in systolic bloodpressure with volume loading in the control group than in the BESA group (p < 0.01). 33.0% ± 7.5%,andtheaverageleftatrialsizewas32mm (Table I). Arrhythmias observed during delivery of electric shocks in BESA group. On delivery of 31 of the 36 shocks, normal sinus rhythm was interrupted. After 24 shocks VT was observed. This arrhythmia either was self-terminated or was modified by the subsequent shock. After 6 of 36 shocks the hearts fibrillated and a countershock had to be applied. In one animal, after delivery of a single shock, a brief episode of asystole occurred that was followed by a slow escape rhythm. In all animals blood pressure and flow were maintained by cardiopulmonary bypass during shock delivery. After the period of reperfusion all dogs had reverted to sinus rhythm and were easily weaned from bypass. As previously reported, during BESA shock delivery we obtained a consistent peak voltage and current within each animal.f In all cases the waveform was smooth without any breaks. Postoperative echocardiographic assessment of mitral regurgitation and regional wall motion-BESA
Volume 103
633
Surgical treatment of VT by BESA
Number 4 April 1992
2.5
L
2.0
.15 Ol
*
15
1-
E E
L
Ol
~ (/J > .....J
.10 OJ
1.0
1-
C
E
:::J
l/ 1
0
.05
.5 • control • BESA
A
o
2
LVEDVI ml/kg
• control • BESA
3
B
0
2
3
LVEDVI ml/kg
Fig. 4. Relationship between cardiac index (eI) and LVEDVI (A)and left ventricular stroke work index (LVSWI) and LVEDVI (B) in response to volume loading is demonstrated. By analysis of covariance there was a statistically significant difference in myocardial performance with poor function after BESA compared with performance in the control group (p < 0.01).
group. In the BESA group in the early study after shock delivery, the area of ablation was indicated by increased echocardiographic density and local wall motion abnormalities (anterior in two dogs, posterior in four dogs). At the time of the late postoperative assessment, wall motion in these areas had returned toward normal in all but one animal. In both the early and late postoperative studies there was no significant mitral regurgitation in fiveof the sixdogs. In the final animal at the late postoperative study there was 2+ mitral regurgitation in the unloaded state and 4+ mitral regurgitation after volume loading (see Table I). The average percent basal LV wall shortening at the early postoperative study was significantly decreased at 14.3% ± 5.9% (p < 0.01 by t test). At 6 weeks postoperatively the percent basal LV shortening had returned toward normal but was still significantly different from the preoperative value at 25.5% ± 8.7% (p < 0.01 by t test). The mean left atrial size was not changed on either the early or late postoperative study (29 ± 4 and 33 ± 3 mm respectively).
Ventricular function-BESA group and control group. The hemodynamic response to volume loading (low- and high-volume measurements) in the control and BESA animals (excluding the one dog with significant mitral regurgitation after BESA) is presented in Table II. Pulse rate did not change significantly, but arterial blood pressure (systolic, diastolic, and mean) increased significantly with volume loading. Left atrial pressure and EDVI also increased significantly. The increase in preload and afterload was associated with a significant increase in cardiac index, stroke index, and stroke work index. Fig. 2 illustrates the diastolic pressure-volume relationship in the two groups. The animals subjected to BESA had a higher left atrial pressure at similar end-diastolic volumes than did the control dogs. However, this trend did not reach statistical significance. Fig. 3 illustrates the systolic pressure-volume relationships in the two groups. The systolic blood pressure was lower in the BESA group at the same end-systolic volume than in control animals. This difference reached statistical significance (p < 0.05). Fig.
The Journal of Thoracic and Cardiovascular Surgery
6 3 4 Mickleborough et al.
Fig. 5. Mitral valve apparatus of BESA dog. Asterisk indicates obvious area of endocardial scar related to previous ablation. Mild thickening of intact chordae and of free edge of valve leaflet.
4 illustrates the relationship between cardiac index and stroke work index and end-diastolic volume in the two groups. There was poorer function after mitral valve BESA with lower cardiac index and stroke work index than in control animals (p < 0.01). Pathologic studies-BESA group. In one animal the mitral valve apparatus after BESA appeared grossly normal. In one animal there was thickening and slight retraction of the edges of the posterior leaflet of the mitral valve. In three animals there was slight thickening of the chordal attachments to the mitral valve leaflets (anterior leaflet in two and posterior leaflet in one) (Fig. 5). In these fivecases the chordae tendineae were intact. In all animals the exact location of the blast could be identified by the presence of a white scar on the endocardial surface of the adjacent myocardium. In two cases the anterior papillary muscle in the scarred area had shrunk and retracted, but the adjacent LV wall was relatively intact (Fig. 6). In three cases in which scar extended onto the septum, ablation had resulted in an almost transmural injury. The volume of scar (major axis X minor axis X depth) produced by BESA in each of the dogs was 3.36, 5.50, 6.00, 6.40, 6.75, and 8.54 em", respectively, with a mean of 6.09 ± 1.69 em", In the single dog with evidence of a persistent wall motion abnormality and significant mitral regurgitation on the late postoperative echocardiographic study, the area of ablation centered over the posterior papillary muscle that had scarred and atrophied. In addition, in this animal there was transmural scarring and thinning of the
surrounding myocardium (Fig. 7). Two chordae originally arising from this papillary muscle had ruptured, but major attachments to the rest of the leaflet appeared to be intact. On microscopic examination, the central region of the area of scar was composed of mature hypocellular dense collagen-rich connective tissue in which intramural coronary arteries were trapped but no viable myocardium was present. At the periphery there was a loosely organized connective tissue matrix containing microvasculature but no myocytes, with a sharp transition to structurally normal myocardium at the edge of the lesions. Microscopic examination of tissue from blocks remote to the area of ablation failed to show any abnormalities. These findings were similar to those previously observed in the apical ablation experiments.f Discussion These long-term animal experiments were designed to demonstrate possible detrimental effects of BESA on LV and mitral valve function. We delivered IOO-joule shocks through electrodes on a 0.5 em grid. This corresponds to two times the current density used in our previous apical ablation experiments. The energy delivered is clearly in excess of that required to achieve limited subendocardial necrosis in the normal dog heart. In these experiments we recorded hemodynamic variables during volume loading to determine the effects of BESA on ventricular function. When compared with a control group, the BESA group showed a trend toward
Volume 103 Number 4 April 1992
Surgical treatment of VT by BESA
635
Fig. 6. Mitral valve apparatus of BESA dog. A, Surface view. Area of ablation centers on tip of anterior papillary muscle, which shows marked scarring and atrophy. B, Transmural section. Adjacent LV wall is intact.
decreased ventricular compliance. However this difference did not reach statistical significance as it had in the previous apical ablation experiments. In the basilar ablation experiments, BESA resulted in poorer myocardial performance as evidenced by ejection phase indices (changes independent in preload). Similarly, the relationship between systolic pressure and systolic volume, independent of preload and incorporating afterload into its formulation.F was also significantly lower than that in control animals. In the previous apical ablation experiments, these indices of LV function were not significantly different from those in control animals.
In the basilar ablation experiments, in five of six animals BESA did not lead to significant dysfunction of the mitral valve. In these dogs, minor degrees of thickening of the free edges of the valve leaflets occurred and were comparable with changes seen in leaflets in the previous apical ablation experiments. We believe that these changes represent a nonspecific reaction to insertion and rubbing of the balloon on leaflet tissue. As expected, BESA caused necrosis of the underlying myocardium in the area of shock delivery. The resulting scar was variable in thickness depending on whether the necrosis is in the region of the septum, the free wall, or the papillary mus-
6 3 6 Mickleborough et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 7. Mitral valveapparatus of BESAdog.A,Surface view. Area of ablationcentersoverthe posterior papillary muscle, which is atrophied. Two chordae arising from this muscle have ruptured. B, Transmural section. There is extensive transmural necrosis and thinning of surrounding myocardium on long-term follow-up.
cle attachment. In one animal, ablation was centered on the posterior papillary muscle attachment and resulted in an area of transmural necrosis, scarring, and thinning. In this dog the early wall motion abnormality persisted in the long-term preparation. At 12 weeks there was significant mitral regurgitation, and two small chordae previously attached to the muscle head were ruptured. Because no ruptured chordae were visualized in the fiveother animals even though there was significant necrosis of the papillary muscle in three other specimens, we believe rupture of the chordae in this dog was due to increased stress as a result
of the persistent localized wall motion abnormality rather than to the direct effects of shock delivery. It is difficult to extrapolate from the results of these animal experiments to the clinical situation. Clearly the long-term effects of BESA depend on location (basal versus apical) and size of the target area, the amount of energy delivered, and the preexisting functional status of this portion of the myocardium. If a large enough area of functional myocardium is ablated, BESA will result not only in decreased compliance but also in deterioration of systolic function. In addition, these experiments show that
Volume 103 Number 4 April 1992
if shocks of high enough energy are delivered, transmural necrosis and persistent wall motion abnormalities will occur. If transmural scarring and thinning occur in the region of papillary muscle insertion, mitral valve dysfunction will result. In conclusion, our findings in this study show that when BESA is used to destroy a localized area of the basilar myocardium in the intact beating normal canine heart, the procedure is associated with decreased ventricular function and potentially harmful effects on mitral valve function. For these reasons BESA must be considered an experimental technique, and caution is advised when considering its clinical use when target areas lie in the basilar portion of the heart. In most patients with VT associated with coronary artery disease, the arrhythmogenic areas identified by mapping appear to be in the subendocardium. In such cases energy delivery that results in transmural injury is clearly excessive. Further experiments are required to determine the appropriate level of energy needed to achieve the desired depth of ablation with minimal injury to the surrounding normal myocardium. We extend our appreciation to Ms. Hilary Vincent for the excellent preparation of the manuscript and to Mr. Richard Adams, Mr. Peter Bertozi, and Ms. Erma Minaker for technical assistance. REFERENCES I. Cox JL. Surgical treatment of ischemic and nonischemic ventricular arrhythmias. In: Modern technics in surgery: Installment II: Cardiac/thoracic surgery. Mount Kisco, N.Y.: Futura, 1985;70:1-18. 2. Hargrove WC, Miller JM, Vasallo JA, Josephson ME. Improved results in the operative management of ventric-
Surgical treatment of VT by BESA
637
ular tachycardia related to inferior wall infarction. J THORAC CARDIOVASC SURG 1986;92:726-32. 3. Mickleborough LL, Harris L, Downar E, Parson I, Gray G. A new intraoperative approach for endocardial mapping of ventricular tachycardia. J THORAC CARDiOVASC SURG 1988;95:271-80. 4. Brodman R, Fisher JD, Johnston DR, et al. Results of electrophysiologically guided operations for drug-resistant recurrent ventricular tachycardia and ventricular fibrillation due to coronary artery disease. J THORAC CARDiOVASC SURG 1984;87:431-8. 5. Cox JL. Laser photoablation for the treatment of refractory ventricular tachycardia and endocardial fibroelastosis. Ann Thorac Surg 1985;39:119-200. 6. Horowitz LN, Harken AH, Kastor JA, Josephson ME. Ventricular resection guided by epicardiaI and endocardial mapping for treatment of recurrent ventricular tachycardia. N Engl J Med 1980;302:589-93. 7. Downar E, Mickleborough L, Harris L, Parson I. Intraoperative electric ablation of ventricular arrhythmias: a closed-heart procedure. J Am Coil Cardiol 1987;10:104856. 8. Mickleborough LL, Wilson J, Harris L, Tashiro T, Parson I, Gray G. Balloon electric shock ablation: effects on ventricular structure, function, and electrophysiology. J THORAC CARDiOVASC SURG 1989;97:135-46. 9. Helmcke F, Nanda NC, Hsiung Me. Color Doppler assessment of mitral regurgitation with orthogonal planes. Circulation 1987;75:175-83. 10. Weisel RD, Lipton IH, Lyau RN, Baird RJ. Cardiac metabolism and performance following cold potassium cardioplegia. Circulation 1978;58(Pt 2):1217-26. II. Weisel RD, Goldman BS, Lipton IH, Teasdale S, Mickle D, Baird RJ. Optimal myocardial protection. Surgery 1978;84:812-21. 12. Sagawa K. The end systolicpressure volume relationship of the ventricle: definition,modifications and clinical use. Circulation 1981;63:1223-7.
Lynda L. Mickleborough, MD, Gregory J. Wilson, MD, Akihiko Usui, MD, Tadashi Isomura, MD, Alexandre Varela, MD, Harry Rakowski, MD, and Gordon Gray, BSc, Toronto, Ontario, Canada
From the Division of Cardiovascular Surgery, The Toronto Hospital, the Department of Pathology, The Hospital for Sick Children, and the Departments of Surgery, Pathology, and Medicine, University of Toronto, Toronto, Ontario, Canada. Supported by the Canadian Heart Foundation, the Heart and Stroke Foundation of Ontario, and the Cardiovascular Research Fund, University of Toronto. Received for publication May 29, 1990. Accepted for publication Feb. 18, 1991. Address for reprints: Lynda L. Mickleborough, MD, Division of Cardiovascular Surgery, EN 13-217, The Toronto Hospital, 200 Elizabeth St., Toronto, Ontario, M5G 2C4 Canada.
12/1/28928
Map-directed operations offer a potential cure for patients with life-threatening ventricular arrhythmias.' Most surgical candidates have extensive coronary artery disease and a history of previous infarction. Preoperative ventricular function is often significantly compromised, but only 50% of patients have a clear-cut area of transmural scar or resectable aneurysm. 1-4 Most arrhythmias associated with ischemic heart disease appear to arise from the subendocardium. With standard approaches for endocardial mapping and ablation, a ventriculotomy incision is required to obtain access to the inner layers of the heart. In patients with ventricular tachycardia (VT),
629
The Journal of Thoracic and Cardiovascular
6 30 Mickleborough et al.
Surgery
assessed by echocardiography before and 9 to 12 weeks after balloon ablation. In addition, LV hemodynamic function was evaluated after ablation in these animals and compared with function in a control group not subjected to the procedure. Methods
Fig. 1. Balloon used for electric shock ablation. There are 12 silver bead electrodes through which shocks are delivered.
the ventriculotomy and scarring associated with its healing may contribute to perioperative mortality and postoperative morbidity.v 6 We have recently described a transatrial balloon approach for intraoperative mapping that can be performed in the intact ventricle.' On the basis of previous experience with catheter ablation and as an extension of the transatrial mapping approach, once mapping has been completed we have delivered electric current through selected electrodes on the balloon array to destroy specific "target areas."? We call this procedure balloon electric shock ablation or BESA. Because the same electrode array is used for mapping and ablation, there is no need to precisely orient the balloon array within the intact ventricle. We have previously demonstrated in normal dog hearts that apical delivery of shocks (100 joules each) through a grid of electrodes spaced at 1 em intervals can produce a well-demarcated layer of homogeneous scar.! This series of experiments was designed to demonstrate possible deleterious effects of basilar shock delivery through a 0.5 em grid of electrodes directly applied to the mitral valve apparatus and adjacent papillary muscles. Left ventricular (LV) and mitral valve function were
Description of balloon apparatus used for BESA. A latex double-layered No.7 balloon is covered with expandable mesh. Twelve silver bead electrodes are attached to Teflon-insulated 36-gauge stainless steel wires and sutured to the mesh in a closely spaced three-row array (Fig 1). The wires are attached to a connector that allows electric shocks to be delivered to selected beads. Echocardiographic assessment- BESA group. All dogs in this group underwent serial echocardiographic evaluations. The dogs were mildly sedated with acepromazine (Atrovet), 0.055 mg/kg, By means of a commercially available Hewlett-Packard 77020a scanner (Hewlett-Packard Company, Andover, Mass.), two-dimensional pulsed and color Doppler flow recordings were obtained from the apical and parasternal windows.The presence or absence of mitral regurgitation was established and quantitated by color flow mapping in multiple views."LV diastolic and systolic dimensions, as well as left atrial dimensions, were measured. Percent basal LV shortening was calculated by the formula: Percent basal LV shortening = (LV diastolic - LV systolic dimension) /Diastolic dimension. Mitral valve and LV function were assessed in a similar fashion 1 week after the BESA procedure, and a third study was performed during volume loading 9 to 12 weeks later. This final study was performed in an open chest preparation by direct application of the echocardiographic probe to the heart. During the follow-up studies we looked for the presence of local wall motion abnormalities related to the previous ablation. Operative procedure-BESA group. Six adult mongrel dogs (30 to 36 kg) underwent a left thoracotomy and were supported with cardiopulmonary bypass at 37° C. A series of 200joule shocks were delivered to adjacent 0.5 em electrode pairs with Spacelabs defibrillator (Spacelabs Inc., Redmond, Wash., model 4045, which provides a damped sinusoidal discharge, energy 100 joules per bead). The bead electrodes served as a cathode, and a standard cautery plate (15 x 25 cm-) applied to the right side of the thorax acted as the anode. A Tektronix storage oscilloscope (Tektronix Inc., Beaverton, Ore.) monitored the voltage and current waveforms during delivery of the shocks. If VT or ventricular fibrillation occurred, defibrillation was performed with internal paddles (20 W . sec). After delivery of the shocks, the balloon was deflated and removed. Bypass support was continued for 20 to 30 minutes (total pump time 60 minutes). Air was removed from the left side of the heart, and the left atrial incision was closed. The dogs were weaned from bypass and decannulated. After closure of the thoracotomy incision, they were allowed to recover under careful hemodynamic monitoring. They were kept under observation, and repeat echocardiographic studies to assess LV function, regional wall motion, and mitral regurgitation were performed 1 week postoperatively and 9 to 12 weeks later. Hemodynamic assessment-BESA group and control group. In control animals and in the BESA animals 9 to 12
Volume 103 Number 4 April 1992
Surgical treatment of VT by RESA
Table I. Echocardiographic assessment Preop.
Mitral regurgitation. Basal fractional shortening (%) Left atrial size (rnm) 'Different from preop. (p
0-1+ 33 ± 7.5
32 ± 3
Early postop. 0-1+ 14.3 ± 5.9*
29 ± 4
63 I
Table II. Hemodynamic response to volume loading Late postop. 0-1+ (n = 5) 2+-4+ (n = I) 25.5 ± 8.7*
33 ± 3
< 0.01).
weeks after the initial operative procedure, LV function was assessed during volume loading. The dogs were premedicated (atropine sulfate 0.6 mg) and anesthetized (thiopental sodium 30 mg/kg). The lungs were ventilated with a volume respirator. An arterial line and a triple-lumen thermodilution Swan-Ganz catheter from Baxter Healthcare Corp., Irvine, California, were inserted. Through a left thoracotomy, a left atrial catheter was inserted. Myocardial function was then evaluated by nuclear ventriculograms. Red cells were labeled with stannous pyrophosphate (2.5 mg) injected directly into the cephalic vein and technetium 99m pertechnetate (600 MBq) injected 25 minutes later. A Technicare Sigma 420 gamma camera (GE Medical Systems, Milwaukee, Wis.) with a high-sensitivity parallel-hole collimator was placed in the best septal view (approximately 45-degree left anterior oblique view) to maximize left and right ventricular separation. A Medical Data Systems A2/A3 computer (Medical Data Systems, Ann Arbor, Mich.) was interfaced with the gamma camera. Scintigraphic data were acquired during 2-minute periods with 16 frames per RR interval. The LV ejection fraction was calculated from commercially available software with a semiautomated edge-detection program. Hemodynamic variables, including heart rate, systolic blood pressure, diastolic blood pressure, mean arterial pressure, left atrial pressure, and cardiac output, were recorded. Cardiac output was indexed to body weight in kilograms. LV stroke volume and stroke work were calculated by standard formulas'P II and indexed to body weight in kilograms. LV end-diastolic volume index (L VEDVI) was calculated from the nuclear-derived ejection fraction (EF) and the thermodilution stroke index (SI) by the formula: LVEDVI = (SI/EF). LV end-systolic volume index was calculated as the difference between the LVEDVI and the stroke index. During volume loading the effects of change in preload (left atrial pressure and EDVI) on cardiac index, stroke index, and LV stroke work index were assessed. Ten to fifteen measurements were made in each animal. Volume loading and unloading resulted in a reproducible change in preload (EDVI) and afterload (mean arterial pressure). In the BESA dogs echocardiographic assessment of mitral regurgitation was repeated with color flow Doppler techniques at peak volume loading. Myocardial performance was evaluated by plotting cardiac index and stroke work index against EDVI. Left atrial pressure was plotted against EDVI, and systolic blood pressure was plotted against end-systolic volume index (ESVI). A linear regression analysis was performed for each of these relationships. To display and analyze the information, we chose a low-volume (EDVI 0.5 to 1.5 ml/kg) and a high-volume (EDVI 2.0 to 3.5
Groups
Before volume loading (LVEDVI 0.5-1.5
After volume loading (LVEDVI 2.()"3.5
mlfkg}
mlfkg)
Pulse (beat/min) Control 120 ± 13 BESA 132 ± 23 Systolic blood pressure (mm Hg) Control 127 ± 29 BESA 119 ± 18 Diastolic blood pressure (mm Hg) Control 92 ± 21 BESA 85 ± 12 Mean arterial pressure (mm Hg) Control 104 ± 23 BESA 94 ± 14 Left atrial pressure (mm Hg) Control 4.4 ± 2.4 BESA 4.4 ± 2.9 Cardiac index (Lyrnin/kg) 0.09 ± 0.Q2 Control BESA 0.08 ± 0.01 Stroke index (nil/kg) Control 0.76 ± 0.20 BESA 0.58 ± 0.13 Stroke workindex (gm . rri/kg) Control \.04 ± 0.37 BESA 0.74 ± 0.25 Ejection fraction (%) Control 67 ± 13 BESA 48 ± 9 EDVI (ml/kg) Control 1.14 ± 0.30 BESA 1.23 ± 0.25 ESVI (ml/kg) Control 0.38 ±0.20 BESA 0.64 ± 0.21 Values are mean ± standard deviation. 'Different from before volume loading (p
119 ± 8 129 ± 16 173 ± 21* 142 ± 14* 120 ± 18* 103 ± 14*,t 138 ± 18* 116 ± 13* 8.8 ± 1.9* I\.6 ± 3.9* 0.14 ± 0.01:1: 0.11 ± 0.01*, § \.20 ± 0.12:1: 0.83 ± 0.14*,§ 2.10 ± 0.37:1: 1.17 ± 0.23*,§ 45 ± 6:1: 32 ± 4:1: 2.72 ± 0.52:1: 2.57 ± 0.40:1: 1.5 ± 0.43:1: 1.7 ± 0.31:1:
< 0.05).
[Different from control group (p < 0.05). :\:Differentfrom before volume loading (p §Different from control group (p
< 0.01).
< 0.01).
ml/kg) measurement from the observations in each dog. These values were used to evaluate myocardial performance and the systolic and diastolic pressure-VOlume relationships for each group. All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication No. 80-23, revised 1978). Pathologic studies-BESA group. After hemodynamic assessment the animals were killed and the hearts excised. The left atrium and left ventricle were opened, and the mitral valve apparatus was examined and photographed. The dimensions of the apparent endocardial scar were recorded. The myocardial lesion was then sectioned. The depth of the lesion at its center
The Journal of Thoracic and Cardiovascular Surgery
6 3 2 Mickleborough et al.
200
15
L 10
L
150
L
OJ
I
E E
n,
« ....J
OJ
I
5
E E 100
1t
o
.:
11
o,
en CfJ
• control eBESA
2
3
50
LVEDVI ml/kg
Fig. 2. Diastolic pressure-volume relationship with volume loading is depicted for control group and BESA group. Mean and standard error of left atrial pressure (LAP) at high and low left ventricular end-diastolic volume indices (LVEDVI) are demonstrated. By analysis of covariance, there was no statistically significantdifference in increase in left atrial pressure with volume loading between the two groups.
was measured, and lesion size was calculated (major axis X minor axis X depth). Transmural LV tissue blockswere taken for histologic studies from the area of intervention, from the border zone (interface of scar with apparently normal myocardium), and from areas remote from the site of ablation. Tissue blocks were fixed with 10% buffered formaldehyde, dehydrated, and embedded in paraffin, and 5 ~m thick sectionswere cut and stained for histologicstudy with hematoxylin-phloxinesaffron, Masson's trichrome, picro-Mallory, and Movat's pentachrome stains. Statistical analysis. Statistical analysis was performed with the Statistical Analysis Program (SAS Institute Inc., Cary, N.e.). Differencesbetween groups were evaluated by x2 statistics for discrete variables and an analysis of covariance for continuous variables. Myocardial performance, diastolic properties, and systolic function were evaluated by a two-way analysis of covariance with the high- and low-volume measurements. Continuous variables are summarized as the mean ± standard deviation in the text and tables and as mean ± standard error in the illustrations. Results Preoperative echocardiographic assessmentBESA group. All dogs had either no mitral regurgitation or 1 + mitral regurgitation (on a scale of 0 to 4+) when studied preoperatively. All ventricles appeared normal with no evidence of localized wall motion abnormalities. The percent basal fractional LV shortening was
• control • BESA
o
2
3
LVESVI ml/kg
Fig. 3. Systolic pressure-volume relationship with volume loading is depicted for the two groups. Mean and standard error of systolicbloodpressure (SBP) at high and low LV end-systolic volume indices (LVESVI) are demonstrated. By analysis of covariance there was a significantlygreater increase in systolic bloodpressure with volume loading in the control group than in the BESA group (p < 0.01). 33.0% ± 7.5%,andtheaverageleftatrialsizewas32mm (Table I). Arrhythmias observed during delivery of electric shocks in BESA group. On delivery of 31 of the 36 shocks, normal sinus rhythm was interrupted. After 24 shocks VT was observed. This arrhythmia either was self-terminated or was modified by the subsequent shock. After 6 of 36 shocks the hearts fibrillated and a countershock had to be applied. In one animal, after delivery of a single shock, a brief episode of asystole occurred that was followed by a slow escape rhythm. In all animals blood pressure and flow were maintained by cardiopulmonary bypass during shock delivery. After the period of reperfusion all dogs had reverted to sinus rhythm and were easily weaned from bypass. As previously reported, during BESA shock delivery we obtained a consistent peak voltage and current within each animal.f In all cases the waveform was smooth without any breaks. Postoperative echocardiographic assessment of mitral regurgitation and regional wall motion-BESA
Volume 103
633
Surgical treatment of VT by BESA
Number 4 April 1992
2.5
L
2.0
.15 Ol
*
15
1-
E E
L
Ol
~ (/J > .....J
.10 OJ
1.0
1-
C
E
:::J
l/ 1
0
.05
.5 • control • BESA
A
o
2
LVEDVI ml/kg
• control • BESA
3
B
0
2
3
LVEDVI ml/kg
Fig. 4. Relationship between cardiac index (eI) and LVEDVI (A)and left ventricular stroke work index (LVSWI) and LVEDVI (B) in response to volume loading is demonstrated. By analysis of covariance there was a statistically significant difference in myocardial performance with poor function after BESA compared with performance in the control group (p < 0.01).
group. In the BESA group in the early study after shock delivery, the area of ablation was indicated by increased echocardiographic density and local wall motion abnormalities (anterior in two dogs, posterior in four dogs). At the time of the late postoperative assessment, wall motion in these areas had returned toward normal in all but one animal. In both the early and late postoperative studies there was no significant mitral regurgitation in fiveof the sixdogs. In the final animal at the late postoperative study there was 2+ mitral regurgitation in the unloaded state and 4+ mitral regurgitation after volume loading (see Table I). The average percent basal LV wall shortening at the early postoperative study was significantly decreased at 14.3% ± 5.9% (p < 0.01 by t test). At 6 weeks postoperatively the percent basal LV shortening had returned toward normal but was still significantly different from the preoperative value at 25.5% ± 8.7% (p < 0.01 by t test). The mean left atrial size was not changed on either the early or late postoperative study (29 ± 4 and 33 ± 3 mm respectively).
Ventricular function-BESA group and control group. The hemodynamic response to volume loading (low- and high-volume measurements) in the control and BESA animals (excluding the one dog with significant mitral regurgitation after BESA) is presented in Table II. Pulse rate did not change significantly, but arterial blood pressure (systolic, diastolic, and mean) increased significantly with volume loading. Left atrial pressure and EDVI also increased significantly. The increase in preload and afterload was associated with a significant increase in cardiac index, stroke index, and stroke work index. Fig. 2 illustrates the diastolic pressure-volume relationship in the two groups. The animals subjected to BESA had a higher left atrial pressure at similar end-diastolic volumes than did the control dogs. However, this trend did not reach statistical significance. Fig. 3 illustrates the systolic pressure-volume relationships in the two groups. The systolic blood pressure was lower in the BESA group at the same end-systolic volume than in control animals. This difference reached statistical significance (p < 0.05). Fig.
The Journal of Thoracic and Cardiovascular Surgery
6 3 4 Mickleborough et al.
Fig. 5. Mitral valve apparatus of BESA dog. Asterisk indicates obvious area of endocardial scar related to previous ablation. Mild thickening of intact chordae and of free edge of valve leaflet.
4 illustrates the relationship between cardiac index and stroke work index and end-diastolic volume in the two groups. There was poorer function after mitral valve BESA with lower cardiac index and stroke work index than in control animals (p < 0.01). Pathologic studies-BESA group. In one animal the mitral valve apparatus after BESA appeared grossly normal. In one animal there was thickening and slight retraction of the edges of the posterior leaflet of the mitral valve. In three animals there was slight thickening of the chordal attachments to the mitral valve leaflets (anterior leaflet in two and posterior leaflet in one) (Fig. 5). In these fivecases the chordae tendineae were intact. In all animals the exact location of the blast could be identified by the presence of a white scar on the endocardial surface of the adjacent myocardium. In two cases the anterior papillary muscle in the scarred area had shrunk and retracted, but the adjacent LV wall was relatively intact (Fig. 6). In three cases in which scar extended onto the septum, ablation had resulted in an almost transmural injury. The volume of scar (major axis X minor axis X depth) produced by BESA in each of the dogs was 3.36, 5.50, 6.00, 6.40, 6.75, and 8.54 em", respectively, with a mean of 6.09 ± 1.69 em", In the single dog with evidence of a persistent wall motion abnormality and significant mitral regurgitation on the late postoperative echocardiographic study, the area of ablation centered over the posterior papillary muscle that had scarred and atrophied. In addition, in this animal there was transmural scarring and thinning of the
surrounding myocardium (Fig. 7). Two chordae originally arising from this papillary muscle had ruptured, but major attachments to the rest of the leaflet appeared to be intact. On microscopic examination, the central region of the area of scar was composed of mature hypocellular dense collagen-rich connective tissue in which intramural coronary arteries were trapped but no viable myocardium was present. At the periphery there was a loosely organized connective tissue matrix containing microvasculature but no myocytes, with a sharp transition to structurally normal myocardium at the edge of the lesions. Microscopic examination of tissue from blocks remote to the area of ablation failed to show any abnormalities. These findings were similar to those previously observed in the apical ablation experiments.f Discussion These long-term animal experiments were designed to demonstrate possible detrimental effects of BESA on LV and mitral valve function. We delivered IOO-joule shocks through electrodes on a 0.5 em grid. This corresponds to two times the current density used in our previous apical ablation experiments. The energy delivered is clearly in excess of that required to achieve limited subendocardial necrosis in the normal dog heart. In these experiments we recorded hemodynamic variables during volume loading to determine the effects of BESA on ventricular function. When compared with a control group, the BESA group showed a trend toward
Volume 103 Number 4 April 1992
Surgical treatment of VT by BESA
635
Fig. 6. Mitral valve apparatus of BESA dog. A, Surface view. Area of ablation centers on tip of anterior papillary muscle, which shows marked scarring and atrophy. B, Transmural section. Adjacent LV wall is intact.
decreased ventricular compliance. However this difference did not reach statistical significance as it had in the previous apical ablation experiments. In the basilar ablation experiments, BESA resulted in poorer myocardial performance as evidenced by ejection phase indices (changes independent in preload). Similarly, the relationship between systolic pressure and systolic volume, independent of preload and incorporating afterload into its formulation.F was also significantly lower than that in control animals. In the previous apical ablation experiments, these indices of LV function were not significantly different from those in control animals.
In the basilar ablation experiments, in five of six animals BESA did not lead to significant dysfunction of the mitral valve. In these dogs, minor degrees of thickening of the free edges of the valve leaflets occurred and were comparable with changes seen in leaflets in the previous apical ablation experiments. We believe that these changes represent a nonspecific reaction to insertion and rubbing of the balloon on leaflet tissue. As expected, BESA caused necrosis of the underlying myocardium in the area of shock delivery. The resulting scar was variable in thickness depending on whether the necrosis is in the region of the septum, the free wall, or the papillary mus-
6 3 6 Mickleborough et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 7. Mitral valveapparatus of BESAdog.A,Surface view. Area of ablationcentersoverthe posterior papillary muscle, which is atrophied. Two chordae arising from this muscle have ruptured. B, Transmural section. There is extensive transmural necrosis and thinning of surrounding myocardium on long-term follow-up.
cle attachment. In one animal, ablation was centered on the posterior papillary muscle attachment and resulted in an area of transmural necrosis, scarring, and thinning. In this dog the early wall motion abnormality persisted in the long-term preparation. At 12 weeks there was significant mitral regurgitation, and two small chordae previously attached to the muscle head were ruptured. Because no ruptured chordae were visualized in the fiveother animals even though there was significant necrosis of the papillary muscle in three other specimens, we believe rupture of the chordae in this dog was due to increased stress as a result
of the persistent localized wall motion abnormality rather than to the direct effects of shock delivery. It is difficult to extrapolate from the results of these animal experiments to the clinical situation. Clearly the long-term effects of BESA depend on location (basal versus apical) and size of the target area, the amount of energy delivered, and the preexisting functional status of this portion of the myocardium. If a large enough area of functional myocardium is ablated, BESA will result not only in decreased compliance but also in deterioration of systolic function. In addition, these experiments show that
Volume 103 Number 4 April 1992
if shocks of high enough energy are delivered, transmural necrosis and persistent wall motion abnormalities will occur. If transmural scarring and thinning occur in the region of papillary muscle insertion, mitral valve dysfunction will result. In conclusion, our findings in this study show that when BESA is used to destroy a localized area of the basilar myocardium in the intact beating normal canine heart, the procedure is associated with decreased ventricular function and potentially harmful effects on mitral valve function. For these reasons BESA must be considered an experimental technique, and caution is advised when considering its clinical use when target areas lie in the basilar portion of the heart. In most patients with VT associated with coronary artery disease, the arrhythmogenic areas identified by mapping appear to be in the subendocardium. In such cases energy delivery that results in transmural injury is clearly excessive. Further experiments are required to determine the appropriate level of energy needed to achieve the desired depth of ablation with minimal injury to the surrounding normal myocardium. We extend our appreciation to Ms. Hilary Vincent for the excellent preparation of the manuscript and to Mr. Richard Adams, Mr. Peter Bertozi, and Ms. Erma Minaker for technical assistance. REFERENCES I. Cox JL. Surgical treatment of ischemic and nonischemic ventricular arrhythmias. In: Modern technics in surgery: Installment II: Cardiac/thoracic surgery. Mount Kisco, N.Y.: Futura, 1985;70:1-18. 2. Hargrove WC, Miller JM, Vasallo JA, Josephson ME. Improved results in the operative management of ventric-
Surgical treatment of VT by BESA
637
ular tachycardia related to inferior wall infarction. J THORAC CARDIOVASC SURG 1986;92:726-32. 3. Mickleborough LL, Harris L, Downar E, Parson I, Gray G. A new intraoperative approach for endocardial mapping of ventricular tachycardia. J THORAC CARDiOVASC SURG 1988;95:271-80. 4. Brodman R, Fisher JD, Johnston DR, et al. Results of electrophysiologically guided operations for drug-resistant recurrent ventricular tachycardia and ventricular fibrillation due to coronary artery disease. J THORAC CARDiOVASC SURG 1984;87:431-8. 5. Cox JL. Laser photoablation for the treatment of refractory ventricular tachycardia and endocardial fibroelastosis. Ann Thorac Surg 1985;39:119-200. 6. Horowitz LN, Harken AH, Kastor JA, Josephson ME. Ventricular resection guided by epicardiaI and endocardial mapping for treatment of recurrent ventricular tachycardia. N Engl J Med 1980;302:589-93. 7. Downar E, Mickleborough L, Harris L, Parson I. Intraoperative electric ablation of ventricular arrhythmias: a closed-heart procedure. J Am Coil Cardiol 1987;10:104856. 8. Mickleborough LL, Wilson J, Harris L, Tashiro T, Parson I, Gray G. Balloon electric shock ablation: effects on ventricular structure, function, and electrophysiology. J THORAC CARDiOVASC SURG 1989;97:135-46. 9. Helmcke F, Nanda NC, Hsiung Me. Color Doppler assessment of mitral regurgitation with orthogonal planes. Circulation 1987;75:175-83. 10. Weisel RD, Lipton IH, Lyau RN, Baird RJ. Cardiac metabolism and performance following cold potassium cardioplegia. Circulation 1978;58(Pt 2):1217-26. II. Weisel RD, Goldman BS, Lipton IH, Teasdale S, Mickle D, Baird RJ. Optimal myocardial protection. Surgery 1978;84:812-21. 12. Sagawa K. The end systolicpressure volume relationship of the ventricle: definition,modifications and clinical use. Circulation 1981;63:1223-7.