- Email: [email protected]
Predictors of failure after endocardial resection for sustained ventricular tachycardia
J
THoRAc CARDIOVASC SURG
1988;95:495-500
Predictors of failure after endocardial resection for sustained ventricular tachycardia This study was designed to identify characteristics that might be predictors of failure of surgical
treatment alone (endocardial resection) for sustained ventricular tachycardia. Thirty-three consecutive patients with sustained ventricular tachycardia were studied by standard techniques preoperatively, intraoperatively, and 7 to 36 days postoperatively. Standard endocardial resection was guided by intraoperative mapping in aU patients. Adjuvant cryoablation was used in areas that were not accessible to excision. Patients were divided into two groups on the basis of the results of the postoperative electrophysiologic study. Group I (14) were patients who still had ventricular tachycardia (failure) and Group II (19) were those who did not have ventricular tachycardia (success). On the basis of the postoperative electrophysiologic testing, the time from myocardial infarction to surgical treatment (less than 3 months) was a powerful predictor of failure of operation alone to prevent ventricular tachycardia (p < 0.01). This may indicate a different mechanism of ventricular tachycardia in this group of patients. Another possible predictor of surgical failure was three-vessel disease. The site of origin of ventricular tachycardia, the use of cryoablation, the number of morphologies, and the amount of tissue resected were not significant predictors of success or failure. The result of the postoperative electrophysiologic study was also a strong prognostic predictor of subsequent arrhythmias.
Berkeley Brandt III, MD, James B Martins, MD,b and Michael G. Kienzle, MD,b
Iowa City, Iowa
SUrgiCal treatment of ventricular tachycardia is becoming more routine in surgical practice throughout the world. Surgical techniques include endocardial excision guided by intraoperative mapping,' extended endocardial excision,' and encircling endomyocardial incision.' Results have been good but have varied with the technique, the timing of operation, and the site of tachycardia." The purpose of this study was to identify characteristics predictive of success or failure after endocardial resection for sustained ventricular tachycardia.
Methods The study group included 33 consecutive patients with prior myocardial infarction and drug-resistant sustained ventricular
From the Division of Thoracic and Cardiovascular Surgery. Departments of Surgery'and Internal Medicine," The University of Iowa Hospitals and Clinics Iowa City, Iowa. Received for publication Oct. 31, 1986. Accepted for publication March 23, 1987. Address for reprints: Berkeley Brandt III, MD, 750 East AdamsSt., Syracuse, NY 13210.
tachycardia. All patients underwent preoperative hemodynamic and electrophysiologic catheterization. During electrophysiologic study, ventricular tachycardia was induced and endocardial mapping of as many morphologies as possible was performed.' During the surgical procedure, ventricular tachycardia was initiated and electrophysiologic mapping of as many morphologies as possible was obtained during cardiopulmonary bypass at normothermia. After epicardial mapping, the left ventricle was opened through the area of infarction and endocardial mapping was done with a hand-held probe. Mapping was usually completed in less than I hour of cardiopulmonary bypass time. After the endocardial map was completed, the aorta was cross-clamped, the heart arrested with cold potassium cardioplegic solution, and endocardial excision was performed as described by Harken, Josephson, and Horowitz.' Since 1983, sites that were in locations inaccessible excision (example, posterior papillary muscle) were also treated by cryoablation. This was achieved by applying a nitrous oxicd probe for 1 to 2 minutes in areas selected by mapping. Results were evaluated in terms of the sites of origin of the tachycardias, time between infarction and operation, the number of morphologies, the quality of the mapping, the presence of an aneurysm, the use of cryoablation, the number of stenotic coronary vessels, and the clinical presentation. All patients underwent postoperative electrophysiologic study 7 to 36 days (mean 14 days) after operation to determine
495
496
The Journal of Thoracic and Cardiovascular Surgery
Brandt, Martins, Kienzle
Table I. Clinical characteristics Group II (EPS -)
Group I (EPS +)
Parameter No. of patients Syncope Cardiac arrest No. of drug trials Anterior infarct Inferior infarct Ejection fraction (%) No. with three-vessel disease
14 8 6 2.7 5 9 34.7 11/14
P Value (XC)
19 (57%) (43%)
13 6 2.7 II 8 33.5 7/19
(64%) (79%)
(68%) (32%) (58%) (42%) (37%)
NS NS NS NS NS NS <0.01
Legend: NS, Not significant. EPS, Electrophysiologic study.
Table II. EP characteristics Characteristic No. of patients HV interval> 55 msec VT cycle length VT morphologies Multiple clinical Multiple induced Adequate map VT origin inferior
Group I (EPS+) 14 2 (14%) 270 1/4 0/14 12 (86%) 9/14 (64%)
Group II (EPS-) 19 6 (32%) 245 5/10 16/19 14 (74%) 8/19 (42%)
P Value NS NS NS NS NS NS
Legend: EP, Electrophysiologic. HV, Time from His bundle activation to ventricular activation. VT, Ventricular tachycardia.
the success or failure of the procedure." This study included ventricular pacing following the same protocol used in the preoperative study. This study was considered positive if any preoperative morphology (clinical or nonclinical) was still inducible. For the purposes of this paper, patients were divided into groups on the basis of the results of this postoperative el~ctrophysiologic study, which was done without antiarrhythrmc drugs. Group I comprised those patients who still had the same morphology of ventricular tachycardia inducible postoperatively. They were placed on a tolerable dose of antiarrhythmic drug and discharged without further study. Group II were those patients who did not have ventricular tachycardia inducible postoperatively and they were discharged without antiarrhythmic drugs. Patients were followed up in a cardiac electrophysiology clinic at intervals of 3 to 12 months. When patients died outside the hospital, the circumstances of death were established by contacting the local physician and family. The site of infarction was determined by electrocardiographic criteria and confirmed by regional wall motion analysis of the left ventricular contrast angiogram. Infarction was assumed when hypokinesia, akinesia, or dyskinesia was present in the inferior or anterior wall. Ejection fraction was calculated by a radionuclide ventriculogram, which makes no assumptions about the shape of the ventricle in systole or diastole. The number of coronary vessels with 50% narrowing of luminal diameter was assessed visually on single-plane coronary angiograms. The time from His bundle activation to ventricular act iva-
tion (HV interval) was measured in a standard fashion! Ventricular tachycardia cycle length was averaged over 3 seconds when tachycardia was stable. Every attempt was made to gather all clinical morphologies from referring physicians. Morphologies occurring spontaneously or during electrophysiologic studies were determined to be separate and distinct if bundle branch block configuration was different (based on the configuration of lead VI being left, right, or intermediate) or the axis on the limb leads varied by 90 degrees. The definition of an adequate map was that all clinical morphologies were mapped either during endocardial catheter mapping or during intraoperative mapping and that a site of origin was clearly defined for all morphologies. The origin of ventricular tachycardia usually corresponded to the area of infarction. The origin as described in this paper was based on the activation map according to the criteria published by Miller and associates' and grouped into inferior, anterior, and septal site. An aneurysm was defined as a thin-walled, transmural fibrous scar demonstrated during the operation and confirmed by pathologic evaluation of the tissue. The tissue resected was measured both as an area by measuring the width and length of the resected peel and as a volume by estimating the length, width, and thickness of the entire resected specimen including the transmural left ventricular scar. These measurements were performed by the pathology department, whose staff were unaware of the results of the surgical procedure. Statistical analysis. The data are given as the mean ± one standard deviation. The data were analyzed by both Xl and logistic regression." The latter was performed to determine which of the three variables associated with the surgical failure were operating independently of the other variables. The data were analyzed in two fashions by logistic regression. The first was a categorical analysis based on arbitrary categorization of the data. The second was a continuous variable analysis in which each patient was assigned a number. The latter was independent of arbitrary categorization.
Results Thirty-three patients with sustained ventricular tachycardia proved by electrophysiologic study in the catheterization laboratory underwent surgical treatment. There were 31 men and two women. Ages ranged from 48 to 77 years (mean 61.5) The clinical characteristics of the patients are given in Table I. Sixteen
Volume 95 Number 3 March 1988
49 7
Sustained ventricular tachycardia
Table ill. Surgical characteristics Characteristic No. of patients Aneurysm present Tissue resection Area (cc') Volume (eel) Adjunctive cryotherapy (% of entire series) Adjunctive cryotherapy (inf. wall origin only) Interval between infarct and operation « 3 mo) First half of series
infarctions were anterior and 17 were inferior. Twenty (61 %) patients had an aneurysm demonstrated during the operation. The only clinical characteristic of the patient groups that was a predictor of surgical failure by x' analysis was three-vessel coronary artery disease. Although there were more failures in the group with an inferior infarction (9/17) than in those with an anterior infarction (5/16) this difference was not statistically significant. Electrophysiologic parameters are shown in Table II. No parameter was a predictor, including the site of origin of ventricular tachycardia. The surgical characteristics are shown in Table III. Patients operated on within :fmonths of infarction had a much higher failure rate (6/7) than those operated on after 3 months (8/26). To assess the possibility that patients operated on later benefited from the experience gained early in the series, we compared the results of patients operated on in the first half of the series with those operated on in the second half. Those operated on in the first half had a higher proportion of failure than those in second half by x' analysis. We also analyzed by Xl, in a continuous fashion, variablesthat were associated with failure of the surgical procedure. These variables included the time of operation after myocardia infarction in days, the number of coronary artery vesselsnarrowed, and the sequence from first to last of the patients undergoing operation. The log of the time between the myocardial infarction and the surgical procedure was associated with a higher proportion of success (p = 0.0302). However, this method of analysis of continuous variables did not demonstrate that the number of vessels involved (p = 0.108) and the sequence of operation (p = 0.074) were important predictors. Indeed, regarding the latter, the arbitrary nature of determining in which sequence each patient was operated on made an important contribution to whether the data were significant. For example, analyzing the data comparing the first quarter with the last three
Group I (EP +)
Group II (EP -)
14 8 (57%)
19 15 (79%)
23 ± 26 32 ± 48 4/14 (29%) 3/9 (33%) 6/14 (43%) 8/14 (57%)
33 ± 27 48 ± 39 7/19 (37%) 5/8 (63%) 1/19 (5%) 5/19 (26%)
P Value NS NS NS NS NS < 0.01 <0.01
Table IV. Late follow-up (21 rna)
Operative mortality (30 days) Late mortality Sudden death Myocardial infarct Left ventricular failure Antiarrhythmic success Of total group (operation + drugs)
Group I (EP +)
Group II (EP-)
14
19
o
5 (36%)
3 1 1 7 (50%) 25/33 (76%)
o
4 (21%) 1 1
2 18 (95%)
quarters demonstrated no relationship to success of the operation (p = 0.620). We also analyzed the data to determine if one or more predictors might be dependent on the others. It must be realized that the number of patients does not allow the statistical power to declare that nonsignificant relationships are truly so. With this proviso, however, comparing the time after myocardial infarction to operation and the number of coronary vessels narrowed demonstrated that the time of operation after myocardial infarction was a much better predictor of failure (p = 0.036) than the number of vessels (p = 0.110). In addition, the time to operation from myocardial infarction was also a better predictor of failure (p = 0.054) than when in our sequence of operations a patient was operated on (p = 0.124). Thus, although these observations must be considered somewhat tentative, they suggest the possibility that early operation after myocardial infarction predicts failure of an endocardial excision procedure. In addition, although the number of coronary vessels narrowed and the time of occurrence in our surgical series may have played a role in the success or failure of operation alone, these aspects of our study did not account for the apparently powerful prediction that
The Journal of Thoracic and Cardiovascular Surgery
498 Brandt, Martins, Kienzle
_vr
4/14 50
~ SUDDEN DEATH
40 w u
i3 0::: 0::: :=l
30
~ 20
p
0:::
~
3/14
10
o~---
EP NEG
EP VT
POSTOP
Fig. 1. Recurrence rate of ventricular tachycardia (VT) and sudden death after endocardial excision based on postoperative electrophysiologic (EP) testing. Patients with inducible tachycardia postoperatively are at significantly higher risk for recurrence of ventricular tachycardia or sudden death.
early operation after myocardial infarction was more likely to be ineffective. Late follow-up is shown in Table IV. There were no operative deaths but nine late deaths. Three patients died of congestive heart failure and two of myocardial infarction. There were four sudden deaths that were presumed due to arrhythmia. Fig. 1 shows the relationship between the positive postoperative electrophysiologic study and the risk of recurrent ventricular tachycardia. Patients with a positive postoperative study were at significantly higher risk (p < 0.05) of recurrent ventricular tachycardia or sudden death. Discussion
Our study, based on postoperative electrophysiologic testing, demonstrates that the time from myocardial infarction to surgical treatment was a powerful predictor of the failure of antiarrhythmic operations alone to prevent induction of ventricular tachycardia. The postoperative electrophysiologic study is also a strong prognostic predictor in terms of subsequent arrhythmias. Our data suggest that a postoperative electrophysiologic test demonstrating that ventricular tachycardia is no longer inducible enables prediction of long-term success of operation alone in the therapy of recurrent, inducible, sustained ventricular tachycardia. This conclusion is supported by the work of Kienzle and associates." In disagreement with Kienzle's group, we found that persistent inducibility of ventricular tachycardia off drugs was a significant predictor of subse-
quent clinical occurrence of ventricular tachycardia or sudden death. A major cause for this disagreement is the way patients were evaluated and analyzed. Kienzle and associates" restudied their patients until their tachycardia became noninducible postoperatively with drug therapy. In the present study, if ventricular tachycardia was still inducible postoperatively, patients were not restudied while receiving drug therapy. This discrepancy may account for the higher incidence of subsequent postoperative ventricular tachycardia in our series. Nonetheless, postoperative studies off drugs revealing the same morphology of ventricular tachycardia as preoperatively indicate at least the short-term failure of surgical therapy alone. Comparison of our results with those of other investigators points out the need for a consistent method of postoperative evaluation. Our success rate with surgical therapy alone (58%) is similar to that of Miller and colleagues,' who reported that 64 of 100 patients were cured by surgical treatment alone. When results obtained with surgical plus antiarrhythmic drug therapy were reported, the success rate was 91% in that study, compared to 76% in our study. This addition of antiarrhythmic drugs improves results but makes the comparison of data more difficult. For that reason we have reported our results in terms of success by surgical therapy alone. The data from our study suggesting that occurrence of drug-resistant ventricular tachycardia less than 3 months after a myocardial infarction is an important predictor of failure of endocardial resection has not been
Volume 95 Number 3 March 1988
previously reported. This fact suggests to us the possibility that there may be a different mechanism of ventricular tachycardia in this group of patients which may be similar to experimental work in which early reperfused infarctions may have intramural reentry. I I Thus early ventricular tachycardias may be due to intramural reentry that would not respond to simple endocardial excision. We speculate that, with time and healing, the anatomic and electrophysiologic characteristics may change such that endocardial procedures may become more effective. Dimarco and co-workers" have shown that operations for sustained ventricular tachyarrhythmias can be done within 2 months of acute myocardial infarction and that these patients have acceptable survival rates. It is not clear from these data whether these patients continued to have inducible tachycardia postoperatively. Miller and colleagues" evaluated a large number of patients and failed to show a difference in failure rate in patients at I, 2, and 4 months compared to those operated on more than 1 year after their myocardial infarction. Their control or late group had an operation-alone failure rate of 34% compared to our control group rate of 31%. Patients operated on within 4 months had a 24% failure rate compared to our 85%. The reason for this difference is not apparent from our data, although these various populations may have different anatomic and electrophysiologic components of ventricular tachycardia. These authors did not evaluate the effect of multivessel disease or position of operation in the sequence, which may also be a predictor of failure of operation alone. The meaning of these latter observations is questionable because of the failure of continuous variable analysis to demonstrate predictor information. The questions raised by these data demand more investigative intraoperative mapping to assess these possibilities. Our data also did not confirm the findings of Miller and associates,' who demonstrated that disparate sites of ventricular tachycardia and the presence of multiple, clinical morphologies of ventricular tachycardia had a higher failure rate. These authors also demonstrated that inferior wall site origin was a secondary predictor of failure. Although we tended to have a higher incidence of failure with inferior wall site of origin, this difference was not statistically significant. Other authors 2. 5 have also suggested that an inferior origin of ventricular arrhythmias requires a different operative approach. A radical excision, perhaps combined with resection of the papillary muscle and mitral valve, has been done in this type of patient.' In our experience the origin of tachycardia was not a significant predictor of failure, and we
Sustained ventricular tachycardia
499
have not performed the more radical excision. We have used cryoablation more commonly in patients with inferior myocardial infarction, but statistical analysis did not show that this improved our results. Intraoperative mapping is an important feature in the treatment of ventricular tachycardia. Moran and colleagues' suggested that extensive excision without mapping is as successful as excision guided by mapping. Logic dictates that the larger amount of tissue removed, the more likely that the area responsible for the tachycardia will be included in the resected specimen. In this series, we were unable to correlate the amount of tissue removed with success. We also could not confirm that an adequate map was a predictor of success and therefore did not prove that mapping was necessary. This may be because our experience is too small to prove that either of these parameters is important or that our definition of an adequate map is inappropriate. Conclusion
In conclusion, our data show that ventricular tachycardia can be controlled by these surgical techniques in both anterior and inferior infarcts. The origin of ventricular tachycardia was not a significant factor in determining success. Also insignificant were the use of cryoablation, the number of morphologies, adequacy of the endocardial map, and amount of tissues excised. The time of operation after myocardial infarction that presumably produced the ventricular tachycardia was a significant predictor of failure, which suggests the possibility of a different mechanism in this subgroup of patients. The presence of three-vessel disease and operation early in our series may have been predictive of failure of operation alone to prevent ventricular tachycardia. Positive postoperative electrophysiologic study was a strong predictor of recurrence of ventricular tachycardia and sudden death. We would like to thank Marie P. Klugman, PhD, for help with the statistical analysis.
REFERENCES I. Harken AH, Horowitz LN, Josephson ME. Comparison of standard aneurysmectomy and aneurysmectomy with directed endocardial resection for the treatment of recurrent sustained ventricular tachycardia. J. THORAC CARD10VASC SURG 1980;80:527-34. 2. Moran JM, Kehoe RF, Loeb JM, Lichtenthal PR, Sanders JH Jr, Michaelis LL. Extended endocardial resection for the treatment of ventricular tachycardia and
500
The Journal of Thoracic and Cardiovascular Surgery
Brandt. Martins. Kienzle
ventricular fibrillation. Ann Thorac Surg 1982;34:53852.
3. Guiraudon G, Fontaine G, Frank R, Escande G, Etievent P, Cabrol C. Encircling endocardial ventriculotomy: a new surgical treatment for life-threatening ventricular tachycardias resistant to medical treatment following myocardial infarction. Ann Thorac Surg 1978;26:43844.
4. Miller JM, Kienzle MG, Harken AH, Josephson ME. Subendocardial resection for ventricular tachycardia: predictors of surgical success. Circulation 1984;70:624-31. 5. Ivey TO, Brady GH, Misbach GA, Greene HL. Surgical management of refractory ventricular arrhythmias in patients with prior inferior myocardial infarction. J THORAC CARDIOVASC SURG
1985;89:369-77.
6. Miller JM, Marchlinski FE, Harken AH, Hargrove WC, Josephson ME. Subendocardial resection for sustained ventricular tachycardia in the early period after acute myocardial infarction. Am J Cardiol 1985;55:980-4. 7. Harken AH, Josephson ME, Horowitz IN. Surgical endocardial resection for the treatment of malignant ventricular tachycardia. Ann Surg 1979;190:456-60.
8. Kienzle MG, Doherty JU, Roy 0, Waxman Hl., Harken AH, Josephson ME. Subendocardial resection for refractory ventricular tachycardia: effects on ambulatory electrocardiogram, programmed stimulation and ejection fraction, and relation to outcome. J Am Coll Cardiol 1983;2:853-8.
9. Josephson ME, Horowitz IN, Farshidi A, Kastor JA. Recurrent sustained ventricular tachycardia. I. Mechanisms. Circulation 1978;57:431-40. 10. Schlesselman JJ. Multivariate analysis: case-control studies-design, conduct, analysis. New York: Oxford University Press, 1982:225-90. 11. Kramer JB, Saffitz JE, Witkowski FX, Corr PH. Intramural re-entry as a mechanism of ventricular tachycardia during evolving canine myocardial infarction. Circ Res 1985;56:736-54. 12. Dimarco JP, Lerman BB, Kron Il., Sellers TO. Sustained
ventricular tachyarrhythmias within 2 months of acute myocardial infarction: results of medical and surgical therapy in patients resuscitated from the initial episode. J Am Coll Cardiol 1985;6:759-68.
THoRAc CARDIOVASC SURG
1988;95:495-500
Predictors of failure after endocardial resection for sustained ventricular tachycardia This study was designed to identify characteristics that might be predictors of failure of surgical
treatment alone (endocardial resection) for sustained ventricular tachycardia. Thirty-three consecutive patients with sustained ventricular tachycardia were studied by standard techniques preoperatively, intraoperatively, and 7 to 36 days postoperatively. Standard endocardial resection was guided by intraoperative mapping in aU patients. Adjuvant cryoablation was used in areas that were not accessible to excision. Patients were divided into two groups on the basis of the results of the postoperative electrophysiologic study. Group I (14) were patients who still had ventricular tachycardia (failure) and Group II (19) were those who did not have ventricular tachycardia (success). On the basis of the postoperative electrophysiologic testing, the time from myocardial infarction to surgical treatment (less than 3 months) was a powerful predictor of failure of operation alone to prevent ventricular tachycardia (p < 0.01). This may indicate a different mechanism of ventricular tachycardia in this group of patients. Another possible predictor of surgical failure was three-vessel disease. The site of origin of ventricular tachycardia, the use of cryoablation, the number of morphologies, and the amount of tissue resected were not significant predictors of success or failure. The result of the postoperative electrophysiologic study was also a strong prognostic predictor of subsequent arrhythmias.
Berkeley Brandt III, MD, James B Martins, MD,b and Michael G. Kienzle, MD,b
Iowa City, Iowa
SUrgiCal treatment of ventricular tachycardia is becoming more routine in surgical practice throughout the world. Surgical techniques include endocardial excision guided by intraoperative mapping,' extended endocardial excision,' and encircling endomyocardial incision.' Results have been good but have varied with the technique, the timing of operation, and the site of tachycardia." The purpose of this study was to identify characteristics predictive of success or failure after endocardial resection for sustained ventricular tachycardia.
Methods The study group included 33 consecutive patients with prior myocardial infarction and drug-resistant sustained ventricular
From the Division of Thoracic and Cardiovascular Surgery. Departments of Surgery'and Internal Medicine," The University of Iowa Hospitals and Clinics Iowa City, Iowa. Received for publication Oct. 31, 1986. Accepted for publication March 23, 1987. Address for reprints: Berkeley Brandt III, MD, 750 East AdamsSt., Syracuse, NY 13210.
tachycardia. All patients underwent preoperative hemodynamic and electrophysiologic catheterization. During electrophysiologic study, ventricular tachycardia was induced and endocardial mapping of as many morphologies as possible was performed.' During the surgical procedure, ventricular tachycardia was initiated and electrophysiologic mapping of as many morphologies as possible was obtained during cardiopulmonary bypass at normothermia. After epicardial mapping, the left ventricle was opened through the area of infarction and endocardial mapping was done with a hand-held probe. Mapping was usually completed in less than I hour of cardiopulmonary bypass time. After the endocardial map was completed, the aorta was cross-clamped, the heart arrested with cold potassium cardioplegic solution, and endocardial excision was performed as described by Harken, Josephson, and Horowitz.' Since 1983, sites that were in locations inaccessible excision (example, posterior papillary muscle) were also treated by cryoablation. This was achieved by applying a nitrous oxicd probe for 1 to 2 minutes in areas selected by mapping. Results were evaluated in terms of the sites of origin of the tachycardias, time between infarction and operation, the number of morphologies, the quality of the mapping, the presence of an aneurysm, the use of cryoablation, the number of stenotic coronary vessels, and the clinical presentation. All patients underwent postoperative electrophysiologic study 7 to 36 days (mean 14 days) after operation to determine
495
496
The Journal of Thoracic and Cardiovascular Surgery
Brandt, Martins, Kienzle
Table I. Clinical characteristics Group II (EPS -)
Group I (EPS +)
Parameter No. of patients Syncope Cardiac arrest No. of drug trials Anterior infarct Inferior infarct Ejection fraction (%) No. with three-vessel disease
14 8 6 2.7 5 9 34.7 11/14
P Value (XC)
19 (57%) (43%)
13 6 2.7 II 8 33.5 7/19
(64%) (79%)
(68%) (32%) (58%) (42%) (37%)
NS NS NS NS NS NS <0.01
Legend: NS, Not significant. EPS, Electrophysiologic study.
Table II. EP characteristics Characteristic No. of patients HV interval> 55 msec VT cycle length VT morphologies Multiple clinical Multiple induced Adequate map VT origin inferior
Group I (EPS+) 14 2 (14%) 270 1/4 0/14 12 (86%) 9/14 (64%)
Group II (EPS-) 19 6 (32%) 245 5/10 16/19 14 (74%) 8/19 (42%)
P Value NS NS NS NS NS NS
Legend: EP, Electrophysiologic. HV, Time from His bundle activation to ventricular activation. VT, Ventricular tachycardia.
the success or failure of the procedure." This study included ventricular pacing following the same protocol used in the preoperative study. This study was considered positive if any preoperative morphology (clinical or nonclinical) was still inducible. For the purposes of this paper, patients were divided into groups on the basis of the results of this postoperative el~ctrophysiologic study, which was done without antiarrhythrmc drugs. Group I comprised those patients who still had the same morphology of ventricular tachycardia inducible postoperatively. They were placed on a tolerable dose of antiarrhythmic drug and discharged without further study. Group II were those patients who did not have ventricular tachycardia inducible postoperatively and they were discharged without antiarrhythmic drugs. Patients were followed up in a cardiac electrophysiology clinic at intervals of 3 to 12 months. When patients died outside the hospital, the circumstances of death were established by contacting the local physician and family. The site of infarction was determined by electrocardiographic criteria and confirmed by regional wall motion analysis of the left ventricular contrast angiogram. Infarction was assumed when hypokinesia, akinesia, or dyskinesia was present in the inferior or anterior wall. Ejection fraction was calculated by a radionuclide ventriculogram, which makes no assumptions about the shape of the ventricle in systole or diastole. The number of coronary vessels with 50% narrowing of luminal diameter was assessed visually on single-plane coronary angiograms. The time from His bundle activation to ventricular act iva-
tion (HV interval) was measured in a standard fashion! Ventricular tachycardia cycle length was averaged over 3 seconds when tachycardia was stable. Every attempt was made to gather all clinical morphologies from referring physicians. Morphologies occurring spontaneously or during electrophysiologic studies were determined to be separate and distinct if bundle branch block configuration was different (based on the configuration of lead VI being left, right, or intermediate) or the axis on the limb leads varied by 90 degrees. The definition of an adequate map was that all clinical morphologies were mapped either during endocardial catheter mapping or during intraoperative mapping and that a site of origin was clearly defined for all morphologies. The origin of ventricular tachycardia usually corresponded to the area of infarction. The origin as described in this paper was based on the activation map according to the criteria published by Miller and associates' and grouped into inferior, anterior, and septal site. An aneurysm was defined as a thin-walled, transmural fibrous scar demonstrated during the operation and confirmed by pathologic evaluation of the tissue. The tissue resected was measured both as an area by measuring the width and length of the resected peel and as a volume by estimating the length, width, and thickness of the entire resected specimen including the transmural left ventricular scar. These measurements were performed by the pathology department, whose staff were unaware of the results of the surgical procedure. Statistical analysis. The data are given as the mean ± one standard deviation. The data were analyzed by both Xl and logistic regression." The latter was performed to determine which of the three variables associated with the surgical failure were operating independently of the other variables. The data were analyzed in two fashions by logistic regression. The first was a categorical analysis based on arbitrary categorization of the data. The second was a continuous variable analysis in which each patient was assigned a number. The latter was independent of arbitrary categorization.
Results Thirty-three patients with sustained ventricular tachycardia proved by electrophysiologic study in the catheterization laboratory underwent surgical treatment. There were 31 men and two women. Ages ranged from 48 to 77 years (mean 61.5) The clinical characteristics of the patients are given in Table I. Sixteen
Volume 95 Number 3 March 1988
49 7
Sustained ventricular tachycardia
Table ill. Surgical characteristics Characteristic No. of patients Aneurysm present Tissue resection Area (cc') Volume (eel) Adjunctive cryotherapy (% of entire series) Adjunctive cryotherapy (inf. wall origin only) Interval between infarct and operation « 3 mo) First half of series
infarctions were anterior and 17 were inferior. Twenty (61 %) patients had an aneurysm demonstrated during the operation. The only clinical characteristic of the patient groups that was a predictor of surgical failure by x' analysis was three-vessel coronary artery disease. Although there were more failures in the group with an inferior infarction (9/17) than in those with an anterior infarction (5/16) this difference was not statistically significant. Electrophysiologic parameters are shown in Table II. No parameter was a predictor, including the site of origin of ventricular tachycardia. The surgical characteristics are shown in Table III. Patients operated on within :fmonths of infarction had a much higher failure rate (6/7) than those operated on after 3 months (8/26). To assess the possibility that patients operated on later benefited from the experience gained early in the series, we compared the results of patients operated on in the first half of the series with those operated on in the second half. Those operated on in the first half had a higher proportion of failure than those in second half by x' analysis. We also analyzed by Xl, in a continuous fashion, variablesthat were associated with failure of the surgical procedure. These variables included the time of operation after myocardia infarction in days, the number of coronary artery vesselsnarrowed, and the sequence from first to last of the patients undergoing operation. The log of the time between the myocardial infarction and the surgical procedure was associated with a higher proportion of success (p = 0.0302). However, this method of analysis of continuous variables did not demonstrate that the number of vessels involved (p = 0.108) and the sequence of operation (p = 0.074) were important predictors. Indeed, regarding the latter, the arbitrary nature of determining in which sequence each patient was operated on made an important contribution to whether the data were significant. For example, analyzing the data comparing the first quarter with the last three
Group I (EP +)
Group II (EP -)
14 8 (57%)
19 15 (79%)
23 ± 26 32 ± 48 4/14 (29%) 3/9 (33%) 6/14 (43%) 8/14 (57%)
33 ± 27 48 ± 39 7/19 (37%) 5/8 (63%) 1/19 (5%) 5/19 (26%)
P Value NS NS NS NS NS < 0.01 <0.01
Table IV. Late follow-up (21 rna)
Operative mortality (30 days) Late mortality Sudden death Myocardial infarct Left ventricular failure Antiarrhythmic success Of total group (operation + drugs)
Group I (EP +)
Group II (EP-)
14
19
o
5 (36%)
3 1 1 7 (50%) 25/33 (76%)
o
4 (21%) 1 1
2 18 (95%)
quarters demonstrated no relationship to success of the operation (p = 0.620). We also analyzed the data to determine if one or more predictors might be dependent on the others. It must be realized that the number of patients does not allow the statistical power to declare that nonsignificant relationships are truly so. With this proviso, however, comparing the time after myocardial infarction to operation and the number of coronary vessels narrowed demonstrated that the time of operation after myocardial infarction was a much better predictor of failure (p = 0.036) than the number of vessels (p = 0.110). In addition, the time to operation from myocardial infarction was also a better predictor of failure (p = 0.054) than when in our sequence of operations a patient was operated on (p = 0.124). Thus, although these observations must be considered somewhat tentative, they suggest the possibility that early operation after myocardial infarction predicts failure of an endocardial excision procedure. In addition, although the number of coronary vessels narrowed and the time of occurrence in our surgical series may have played a role in the success or failure of operation alone, these aspects of our study did not account for the apparently powerful prediction that
The Journal of Thoracic and Cardiovascular Surgery
498 Brandt, Martins, Kienzle
_vr
4/14 50
~ SUDDEN DEATH
40 w u
i3 0::: 0::: :=l
30
~ 20
p
0:::
~
3/14
10
o~---
EP NEG
EP VT
POSTOP
Fig. 1. Recurrence rate of ventricular tachycardia (VT) and sudden death after endocardial excision based on postoperative electrophysiologic (EP) testing. Patients with inducible tachycardia postoperatively are at significantly higher risk for recurrence of ventricular tachycardia or sudden death.
early operation after myocardial infarction was more likely to be ineffective. Late follow-up is shown in Table IV. There were no operative deaths but nine late deaths. Three patients died of congestive heart failure and two of myocardial infarction. There were four sudden deaths that were presumed due to arrhythmia. Fig. 1 shows the relationship between the positive postoperative electrophysiologic study and the risk of recurrent ventricular tachycardia. Patients with a positive postoperative study were at significantly higher risk (p < 0.05) of recurrent ventricular tachycardia or sudden death. Discussion
Our study, based on postoperative electrophysiologic testing, demonstrates that the time from myocardial infarction to surgical treatment was a powerful predictor of the failure of antiarrhythmic operations alone to prevent induction of ventricular tachycardia. The postoperative electrophysiologic study is also a strong prognostic predictor in terms of subsequent arrhythmias. Our data suggest that a postoperative electrophysiologic test demonstrating that ventricular tachycardia is no longer inducible enables prediction of long-term success of operation alone in the therapy of recurrent, inducible, sustained ventricular tachycardia. This conclusion is supported by the work of Kienzle and associates." In disagreement with Kienzle's group, we found that persistent inducibility of ventricular tachycardia off drugs was a significant predictor of subse-
quent clinical occurrence of ventricular tachycardia or sudden death. A major cause for this disagreement is the way patients were evaluated and analyzed. Kienzle and associates" restudied their patients until their tachycardia became noninducible postoperatively with drug therapy. In the present study, if ventricular tachycardia was still inducible postoperatively, patients were not restudied while receiving drug therapy. This discrepancy may account for the higher incidence of subsequent postoperative ventricular tachycardia in our series. Nonetheless, postoperative studies off drugs revealing the same morphology of ventricular tachycardia as preoperatively indicate at least the short-term failure of surgical therapy alone. Comparison of our results with those of other investigators points out the need for a consistent method of postoperative evaluation. Our success rate with surgical therapy alone (58%) is similar to that of Miller and colleagues,' who reported that 64 of 100 patients were cured by surgical treatment alone. When results obtained with surgical plus antiarrhythmic drug therapy were reported, the success rate was 91% in that study, compared to 76% in our study. This addition of antiarrhythmic drugs improves results but makes the comparison of data more difficult. For that reason we have reported our results in terms of success by surgical therapy alone. The data from our study suggesting that occurrence of drug-resistant ventricular tachycardia less than 3 months after a myocardial infarction is an important predictor of failure of endocardial resection has not been
Volume 95 Number 3 March 1988
previously reported. This fact suggests to us the possibility that there may be a different mechanism of ventricular tachycardia in this group of patients which may be similar to experimental work in which early reperfused infarctions may have intramural reentry. I I Thus early ventricular tachycardias may be due to intramural reentry that would not respond to simple endocardial excision. We speculate that, with time and healing, the anatomic and electrophysiologic characteristics may change such that endocardial procedures may become more effective. Dimarco and co-workers" have shown that operations for sustained ventricular tachyarrhythmias can be done within 2 months of acute myocardial infarction and that these patients have acceptable survival rates. It is not clear from these data whether these patients continued to have inducible tachycardia postoperatively. Miller and colleagues" evaluated a large number of patients and failed to show a difference in failure rate in patients at I, 2, and 4 months compared to those operated on more than 1 year after their myocardial infarction. Their control or late group had an operation-alone failure rate of 34% compared to our control group rate of 31%. Patients operated on within 4 months had a 24% failure rate compared to our 85%. The reason for this difference is not apparent from our data, although these various populations may have different anatomic and electrophysiologic components of ventricular tachycardia. These authors did not evaluate the effect of multivessel disease or position of operation in the sequence, which may also be a predictor of failure of operation alone. The meaning of these latter observations is questionable because of the failure of continuous variable analysis to demonstrate predictor information. The questions raised by these data demand more investigative intraoperative mapping to assess these possibilities. Our data also did not confirm the findings of Miller and associates,' who demonstrated that disparate sites of ventricular tachycardia and the presence of multiple, clinical morphologies of ventricular tachycardia had a higher failure rate. These authors also demonstrated that inferior wall site origin was a secondary predictor of failure. Although we tended to have a higher incidence of failure with inferior wall site of origin, this difference was not statistically significant. Other authors 2. 5 have also suggested that an inferior origin of ventricular arrhythmias requires a different operative approach. A radical excision, perhaps combined with resection of the papillary muscle and mitral valve, has been done in this type of patient.' In our experience the origin of tachycardia was not a significant predictor of failure, and we
Sustained ventricular tachycardia
499
have not performed the more radical excision. We have used cryoablation more commonly in patients with inferior myocardial infarction, but statistical analysis did not show that this improved our results. Intraoperative mapping is an important feature in the treatment of ventricular tachycardia. Moran and colleagues' suggested that extensive excision without mapping is as successful as excision guided by mapping. Logic dictates that the larger amount of tissue removed, the more likely that the area responsible for the tachycardia will be included in the resected specimen. In this series, we were unable to correlate the amount of tissue removed with success. We also could not confirm that an adequate map was a predictor of success and therefore did not prove that mapping was necessary. This may be because our experience is too small to prove that either of these parameters is important or that our definition of an adequate map is inappropriate. Conclusion
In conclusion, our data show that ventricular tachycardia can be controlled by these surgical techniques in both anterior and inferior infarcts. The origin of ventricular tachycardia was not a significant factor in determining success. Also insignificant were the use of cryoablation, the number of morphologies, adequacy of the endocardial map, and amount of tissues excised. The time of operation after myocardial infarction that presumably produced the ventricular tachycardia was a significant predictor of failure, which suggests the possibility of a different mechanism in this subgroup of patients. The presence of three-vessel disease and operation early in our series may have been predictive of failure of operation alone to prevent ventricular tachycardia. Positive postoperative electrophysiologic study was a strong predictor of recurrence of ventricular tachycardia and sudden death. We would like to thank Marie P. Klugman, PhD, for help with the statistical analysis.
REFERENCES I. Harken AH, Horowitz LN, Josephson ME. Comparison of standard aneurysmectomy and aneurysmectomy with directed endocardial resection for the treatment of recurrent sustained ventricular tachycardia. J. THORAC CARD10VASC SURG 1980;80:527-34. 2. Moran JM, Kehoe RF, Loeb JM, Lichtenthal PR, Sanders JH Jr, Michaelis LL. Extended endocardial resection for the treatment of ventricular tachycardia and
500
The Journal of Thoracic and Cardiovascular Surgery
Brandt. Martins. Kienzle
ventricular fibrillation. Ann Thorac Surg 1982;34:53852.
3. Guiraudon G, Fontaine G, Frank R, Escande G, Etievent P, Cabrol C. Encircling endocardial ventriculotomy: a new surgical treatment for life-threatening ventricular tachycardias resistant to medical treatment following myocardial infarction. Ann Thorac Surg 1978;26:43844.
4. Miller JM, Kienzle MG, Harken AH, Josephson ME. Subendocardial resection for ventricular tachycardia: predictors of surgical success. Circulation 1984;70:624-31. 5. Ivey TO, Brady GH, Misbach GA, Greene HL. Surgical management of refractory ventricular arrhythmias in patients with prior inferior myocardial infarction. J THORAC CARDIOVASC SURG
1985;89:369-77.
6. Miller JM, Marchlinski FE, Harken AH, Hargrove WC, Josephson ME. Subendocardial resection for sustained ventricular tachycardia in the early period after acute myocardial infarction. Am J Cardiol 1985;55:980-4. 7. Harken AH, Josephson ME, Horowitz IN. Surgical endocardial resection for the treatment of malignant ventricular tachycardia. Ann Surg 1979;190:456-60.
8. Kienzle MG, Doherty JU, Roy 0, Waxman Hl., Harken AH, Josephson ME. Subendocardial resection for refractory ventricular tachycardia: effects on ambulatory electrocardiogram, programmed stimulation and ejection fraction, and relation to outcome. J Am Coll Cardiol 1983;2:853-8.
9. Josephson ME, Horowitz IN, Farshidi A, Kastor JA. Recurrent sustained ventricular tachycardia. I. Mechanisms. Circulation 1978;57:431-40. 10. Schlesselman JJ. Multivariate analysis: case-control studies-design, conduct, analysis. New York: Oxford University Press, 1982:225-90. 11. Kramer JB, Saffitz JE, Witkowski FX, Corr PH. Intramural re-entry as a mechanism of ventricular tachycardia during evolving canine myocardial infarction. Circ Res 1985;56:736-54. 12. Dimarco JP, Lerman BB, Kron Il., Sellers TO. Sustained
ventricular tachyarrhythmias within 2 months of acute myocardial infarction: results of medical and surgical therapy in patients resuscitated from the initial episode. J Am Coll Cardiol 1985;6:759-68.