- Email: [email protected]
The Blalock-Taussig shunt
J
THoRAc CARDIOVASC SURG
85:917-922, 1983
The Blalock-Taussig shunt Low risk, effective palliation, and pulmonary artery growth The Blalock-Taussig shunt has been used at our institution in 64 infants with cyanotic congenital heart disease who required palliation. Thirty-one of these children were less than 2 months of age at the time of operation. There were no early shunt closures. There were two hospital deaths (3%). There was only one hospital death in the 3/ patients less than 60 days old. According to a life-table analysis, 87% ± 9% (± 95% confidence limits) of the shunts were functioning at l year and 78% ± /2% at 2 years. Results are similar in the subgroup of children less than 60 days of age at the time of operation or in the subgroup of children weighing less than 3,999 gm at operation. In infants with small pulmonary arteries, pulmonary arterial growth was excellent (ipsilateral pulmonary artery/descending aorta ratio 0.70 -- 0.95 in 550 days, p < 0.00/; contralateral ratio 0.73 -- 0.99 in 550 days, p < 0.00/). The Blalock-Taussig shunt can be performed with low risk, provides excellent palliation, and is associated with excellent pulmonary artery growth.
Robert A. Guyton, M.D., J. Elwood Owens, M.D., James D. Waumett, M.D., Kenneth J. Dooley, M.D., Charles R. Hatcher, Jr., M.D., and Willis H. Williams, M.D., Atlanta. Ga.
he
Blalock-Taussig shunt has become increasingly popular in small infants as surgical mortality has decreased and the complications of other forms of palliation have been identified.1·7 We have used the Blalock-Taussig shunt in recent years in most children with cyanotic congenital heart disease who required palliation, regardless of age. Over a 5 year period, we performed by Blalock-Taussig shunts in 63 children. Thirty-one of these children were less than a months old at the time of operation. This review was undertaken to evaluate the results of this effort. Particular attention was devoted to the very young children with respect to duration of palliation, distortion of the pulmonary arteries, and pulmonary artery growth. Patients
Between May of 1976 and May of 1981, the BlalockTaussig shunt has been used with increasing frequency
From the Henrietta Egleston Hospital for Children and the Carlyle Fraser Heart Center, Crawford W. Long Memorial Hospital, Emory University School of Medicine, Atlanta, Ga. Received for publication July 30, 1982. Accepted for publication Sept. 10, 1982. Address for reprints: Robert A. Guyton, M.D., Cardiothoracic Research Laboratory, Carlyle Fraser Heart Center, Crawford W. Long Memorial Hospital, Atlanta, Ga. 30365.
at Henrietta Egleston Hospital for Children. Fifty-two operations were performed upon children less than 1 year old, and 31 infants were less than 60 days old (Figs. 1 and 2). Twenty-nine children weighed less than 3,999 gm at operation. The smallest child weighed 1,800 gm. Most of the children in this series had complex cyanotic congenital heart disease with multiple anomalies. Until January of 1981, it was our policy to perform primary total correction in children with "simple" tetralogy of Fallot, regardless of size (the smallest child having successful repair weighed 2,400 gm). Nineteen children in this series had a primary diagnosis of tetralogy of Fallot. Six of these children had anomalous origin of one or more coronary arteries. The other children with tetralogy underwent palliation for a variety of reasons (renal failure, vascular ring, colostomy present, small pulmonary arteries). Nine children had pulmonary atresia, three of whom had ventricular septal defects. Seven children had dextro-transposition of the great arteries (d-TGA) and ventricular septal defect with either pulmonary stenosis or prior pulmonary artery banding. Five children had simple transposition (d-TGA with intact ventricular septum) with inadequate oxygenation after a Blalock-Hanlon procedure in the first few weeks of life. Four patients had a primary diagnosis of tricuspid atresia. Nineteen patients had a variety of other diagnoses including double-outlet right 917
The Journal of Thoracic and Cardiovascular
9 I 8 Guyton et al.
30
Surgery
15
~
-",
0_
~ C Ql::J
.n.s:: EUJ
10
::Jf-
ZeD
20
5
o
o
2
6
12
n
r1
24
> 24
3
4
5
6
7
8
9
Age at Operation (weeks)
-
18
2
Fig. 2. More children underwent the Blalock-Taussig (BT) shunt within the firstweek of lifethan duringany othersimilar time interval.
Age at Operation (months)
Fig. 1. The majority of children included in this series underwent the Blalock-Taussig (BT) shunt within the first 6
months of life.
ventricle, complete atrioventricular canal with pulmonic stenosis, and variations of single ventricle. Fifty-one patients had no prior operation to alter pulmonary blood flow. Four patients had a previous Blalock-Hanlon procedure, five children had a prior pulmonary artery banding, two patients had a prior central shunt, one patient had a prior Blalock-Taussig shunt, and one patient had a prior Brock procedure. Follow-up was complete through February of 1982 for all patients. Operative technique The subclavian artery arising from the innominate artery was chosen for anastomosis in all but two instances (right, 41 patients; left, 21 patients). In one case, a previous Blalock shunt had been performed on the innominate side. In a second case, a vascular ring was divided at the time of anastomosis, and the subclavian artery arising from the smaller, divided, left aortic arch was used for the anastomosis. The chest was entered through an anterolateral thoracotomy in the fourth intercostal space. The pulmonary artery was isolated, controlled proximally by a vascular clamp, and controlled distally by tension on elastic vessel loops. The subclavian artery was isolated, its branches were divided, and it was brought under the vagus nerve. Care was taken to avoid excessive tension on the pulmonary artery. If necessary, and if the subclavian artery was large enough, the anastomosis was performed distal to the first branches of the subclavian artery. In two cases, a section of prosthetic material (4 mm polytetrafluoroethylene graft) was interposed between the end of the subclavian artery and the pulmonary artery in order to
avoid tension. The carotid and innominate arteries were dissected from surrounding tissue to gain additional length. Next, 6-0 or 7-0 polypropylene sutures were used to construct the anastomosis. Simple interrupted sutures were used on the anterior aspect and either simple interrupted sutures (seven patients) or continuous sutures (54 patients) on the posterior aspect. In three patients (the two patients with prosthetic grafts and one patient with d-TGA with intact septum) a continuous suture was used throughout. Optical magnification was used in every case. Heparin was administered prior to anastomosis (100 units/kg) and additionally was continued for 2 days in 41 patients (26 of the 31 children less than 61 days old) at a rate of 25 units/kg/hr. More recently, postoperative heparin dosage has been reduced to maintain the patient's partial thromboplastin time at I Vl to 2 times the control value. Results Operative mortality and morbidity. There were no intraoperative deaths. Two patients died during the hospitalization for the shunt procedure. The first hospital death occurred in a 57-day-old child with a hypoplastic right ventricle and single ventricle physiology, who had pneumonia after the operation. She died with a patent anastomosis 28 days after creation of the shunt. The second death occurred in a 13-month-old child with single ventricle undergoing his third palliative procedure for cyanosis and congestive heart failure. Cardiac arrest occurred during an episode of tachycardia while he was being treated with dopamine for low cardiac output on the second postoperative day. At postmortem examination he was found to have an unsuspected subaortic chamber with a restrictive bulboventricular foramen. Loss of arm tissue and phrenic nerve paralysis did not occur. In one patient, the left recurrent laryngeal nerve
Volume 85 Number 6 June, 1983
C Q) o
iii
Blalock-Taussig shunt
100
100
75
75
C Q)
50
~
a..
(81
25
Q)
9I9
(8)
78 ± 12% at 2 years (17)
50
a..
25
0 0
500
1000
o
<:..L-
---'-
o
500
Days
---'-
1000
---'
1500
Days
Fig. 3. Depicted is a life-table analysis of the cumulative percent event-free (nooperation or death for any reason) in the 63 patients included in this study.
Fig. 4. Cumulative percent adequate shunt function (no death or second operation related to shunt failure) in the 63 patients included in this study.
was paralyzed when a double aortic arch was divided at the time of creation of the shunt. Six patients underwent reexploration for bleeding, five of 41 (12%) patients receiving heparin postoperatively and one of 23 (4%) patients not heparinized postoperatively (difference significant using chi square test, p < 0.05). Duration of palliation. Evaluation of adequacy of shunt function is complicated by the fact that some patients underwent a second operation electively while the shunt was still functioning and other patients died with a shunt still open and providing adequate pulmonary blood flow. Data analysis was therefore carried out by means of the Kaplan-Meier' modification of the life-table analysis for two different conditions: the event-free interval after shunt (no death or second procedure for any reason) and the shunt failure-free interval after shunt (no death or second procedure related to shunt failure). These results are presented in Table I and illustrated in Fig. 3, 4, and 5. Results are similar for the entire series (n = 64), the patients operated upon within the first 2 months of life (n = 31) and the patients operated upon at a weight less than 3,999 gm (n = 29) (Table I). . Late mortality without a second operation. Four children died without a second operation. Three of these children had multiple cardiac defects with single ventricle physiology complicated by cyanosis and congestive heart failure. These three children died outside the hospital and the cause of death is not known. One child died of renal failure secondary to hydronephrosis 6 months after creation of the shunt. Second procedures. Twenty-nine children have undergone a second operation, with 21 hospital survivors
Table I. Duration of palliation Entire series
Age
<6/ days
Percent event free (± 95% CL) At 1 year 74 ± 11 74 ± 16 53 ± 14 At 2 years Percent shunt functioning (±95% CL) At I year 87 ± 9 85 ± 14 At 2 years 78 ± 12
Weight <3,999 gm 72 ± 17
86 ± 15
Legend: CL, Confidence limits.
(72%). Thirteen children had a second procedure designed to adjust pulmonary blood flow. Four of these patients died after the second operation, all having complex anomalies. Sixteen children had a "corrective" procedure. Three infants (4, II, and 21 months) died after early corrective operation was forced by clinical deterioration and complex anatomy. One older child (12 years) died after a Fontan procedure at another institution. Pulmonary artery distortion and growth. Postoperative angiograrns were reviewed in 42 patients. The pulmonary artery was mildly tented up at the site of anastomosis in four patients, two of whom have since undergone successful correction without repair of the pulmonary artery. Kinking was judged to besevere (that is, probably necessitating surgical revision) in two patients, one of whom has since undergone correction without revision of the pulmonary artery. Thirty-six angiograrns revealed no significant distortion of the pulmonary artery. Pulmonary artery size was estimated relative to the
The Journal of Thoracic and Cardiovascular Surgery
9 2 a Guyton et al.
1.5
100
75
C Q) ~ Q) a,
.......
>~
t'Cl
... «0 «>- Cl 0 c:: ~ :c :; c:: c:: a:
181 85 ± 14°0 at 1 year 1201
o
Q)
::::J
Q)
E
50
1.0
~
.5
a.. 0
25
After
Ipsilateral Pulmonary Artery Size
o Cl...a
.l..-
500
----J
1.5
1000
Days
Fig. 5. Cumulative percent adequate shunt function (no death or second operation related to shunt failure) in the 31 patients who underwent operation within the first 60 days of life.
size of the descending aorta at the diaphragm. In patients in whom angiograms were adequate for this analysis, the change in size of the pulmonary arteries was assessed. These results are presented in Table II and depicted in Fig. 6. The patients are divided into two groups: patients in whom the shunt was inserted into a pulmonary artery smaller than the descending aorta (ratio < La) and patients in whom the shunt was inserted into a pulmonary artery larger than the descending aorta (ratio> 1.0). This division is logical and necessary, since one would not expect large pulmonary arteries to grow at the same rate as small pulmonary arteries. The data reveal that large pulmonary arteries grew at the same rate as the descending aorta, but small pulmonary arteries grew faster than the descending aorta after creation of a Blalock-Taussig shunt. Moreover, in this series, the contralateral pulmonary arteries increased in size as much as the ipsilateral pulmonary arteries. There was no difference between right and left Blalock shunts with respect to pulmonary artery growth. Discussion The application of the Blalock-Taussig shunt to small infants has nearly eliminated the use of direct aortopulmonary anastomoses in our institution. This review is an attempt to evaluate the consequences of this change in policy with respect to survival and subsequent complications related to the shunt. Hospital survival rate in this series was 97%, even in patients less than 61 days of age. This result compares
>....... Q)
t'Cl ...
t ~ Cl >- c:: .Q ~ '0 tii c c:: a:
«
o
E::::J
1.0 p< _/ .
Q)
~ Q) a.. 0
00:
.5 I
®
Before
After
Contralateral Pulmonary Artery Size
Fig. 6. Growth of the ipsilateral (Aj and contralateral (Bj pulmonary arteries after creation of the Blalock-Taussig shunt in children in whom the pulmonary artery at the time of anastomosis was smaller than the descending aorta (mean duration 550 days).
favorably with earlier large series of small infants: 91 % survival rate in 33 infants less than 3 months of age from Detroit Michigan, and Birmingham, Alabama, 1 77% survival rate in 31 infants less than 7 days of age from Philadelphia, Pennsylvania,' and 65% survival rate in 34 infant less than 3 months of age from Pittsburgh, Pennsylvania.' Smaller series have emphasized the advantages of the Blalock-Taussig shunt in neonates: 67% survival rate in 18 infants less than 6 months of age from St. Louis University,' 100% survival rate in 11 infants less than 6 months of age from Washington University in St. Louis, Missouri,' 100% survival in seven infants under 1 year of age from Madison, Wisconsin," and 100%survival rate in five infants under 6 months of age from Galveston, Texas.' The improvement in hospital mortality in these children in all institutions reflects improvements in perioperative care as well as surgical and anesthetic techniques. Similar improvement has occurred in survival with the Waterston anastomosis. Ten consecutive survivors of the
Volume 85 Number 6 June, 1983
Blalock-Taussig shunt
921
Table II. Pulmonary artery/descending aorta ratio Duration of shunt (days) Initial ratio <1.0 Ipsilateral PA Contralateral PA Initial ratio> 1.0 Ipsilateral PA Contralateral PA
Before shunt
After shunt
Significant change
550 ± 332
0.70 ± 0.12 0.73 ± 0.18
0.95 ± 0.19 0.99 ± 0.27
p<0.OOI,n=22 p < 0.00I, n = 19
515 ± 580
1.30 ± 0.29 1.30 ± 0.33
1.30 ± 0.39 1.24 ± 0.29
NS, n = 10 NS. n = 10
Legend: PA. Pulmonary artery. NS. Not significant.
Waterston anastomosis performed in the first week of life were reported from Rochester, New York." We conclude that the Blalock-Taussig shunt can be successfully performed in very young or small infants with a hospital survivalrate at least as good as that of any other anastomosis. Hospital morbidity related to the surgical procedure was rare, with only one instance of recurrent laryngeal nerve damage. Phrenic nerve damage and loss of arm tissue did not occur. Early shunt closure was also not observed. In most of the smaller infants (26/31), heparin anticoagulation was employed for the first 2 postoperative days. Postoperative anticoagulation did increase the reexploration rate for postoperative bleeding in the heparinized versus nonheparinized group (12% versus 4%, p < 0.05). We continue to use heparin postoperatively because we feel subjectively that anticoagulation may improve early shunt patency, although we have no data to support this hypothesis. There were no early graft closures in either group, i.e., whether heparinized postoperatively or not. AIl of the patients who were at high risk for early shunt closure were given anticoagulant therapy. Evaluation of the clinical course of these children is complicated by the diverse and complex diagnoses. Within the tetralogy of Fallot group, most of the patients had some additional problem (cardiac or otherwise) which forced palliation rather than correction. During this period, our policy was to correct primarily all tetralogy anomalies in which the anatomy was suitable, regardless of age. The smallest child undergoing successful correction weighed 2,400 gm. The one operative death at the time of tetralogy correction in this series occurred in a child with persistent pulmonary problems after division of a double aortic arch. Four patients died during attempted additional palliative procedures to increase pulmonary blood flow. Three other children died during "corrective" operations for complex situations. Twenty-one patients have survived a second operation. An evaluation of the duration of adequate palliation is
also difficult. One method is to evaluate the event-free interval after the shunt procedure. However, this tends to underestimate the duration of effective palliation. We repair d-TGA at 1 year of age and tetralogy before 2 years of age regardless of the effectiveness of previous palliation. Therefore, an "event" might occur in these children before the end of effective palliation. The second method which we employed, therefore, was an actuarial analysis of the shunt failure-free interval after shunt creation. In this analysis, children who died or underwent a second operation without shunt failure were considered "removed from the series alive" at the time of second operation or death. An example of death without shunt failure would be the child who died of renal failure with a patent shunt. This second method of analysis might tend to overestimate the duration of effective palliation if shunt inadequacy contributed to the death or second operation in an unrecognized manner. The actual duration of effective palliation probably is between these two estimates, that is, between 74% and 87% effective palliation at 1 year (Figs. 3 and 4). It is particularly notable that there is no less effective palliation in the children who underwent operation within the first 60 days of life or at a weight of less than 3,999 gm (Table I and Fig. 5). A critically important part of any evaluation of shunt function is the fate of the pulmonary arteries after creation of the shunt. The development of pulmonary vascular obstruction is uncommon in the first 5 years after creation of a Waterston shunt and very rare after creation of a Blalock-Taussig shunt." Severe pulmonary vascular obstruction, with reversal of flow in the Blalock-Taussig shunt, developed in one child in this series 4 months after shunt creation. Because the pulmonary artery was not entered in pre-shunt catheterizations in this infant with d-TGA, the pulmonary vascular resistance prior to shunt creation is unknown. The primary liability of the Waterston anastomosis is distortion and kinking of the pulmonary artery at the site of anastomosis.I I We have been concerned that the Blalock-Taussig shunt, when performed in these very
922
The Journal of Thoracic and Cardiovascular Surgery
Guyton et al.
small children with tiny, delicate pulmonary arteries, might lead to similar distortion. The results of this study, with severe kinking in only two of 64 patients, indicate that distortion of the pulmonary artery is not a major problem in the subsequent management of these children. A final important consideration is the growth (or failure to grow) of the pulmonary arteries after shunt creation. Since one would not expect large pulmonary arteries to grow after shunt creation, these children were separated from the children with small pulmonary arteries. A significant increase in the size of small pulmonary arteries did occur, bringing the size of these arteries very close to the size of the descending aorta. This occurred in both left and right Blalock shunts, and growth was equivalent in the ipsilateral and in the contralateral pulmonary arteries (Table II, Fig. 6). Such growth was not observed in large pulmonary arteries, which tended to stay the same size relative to the descending aorta. This study supports the contention of others that the Blalock-Taussig shunt can be performed with a high hospital survival rate (97% in this series) even in very young and small infants. This success reflects the improvements in perioperative care and anesthesia and surgical techniques that have occurred in recent years. The large number of patients included in this series allows a statistical evaluation of the effectiveness of the Blalock-Taussig shunt, an analysis heretofore unavailable. Effective palliation can be expected in about 85% of these infants 1 year after shunt creation, regardless of age or size at the time of operation. Two thirds of the children in this series have undergone postoperative cineangiograms, providing substantial information about the fate of the pulmonary artery after the Blalock-Taussig procedure. Pulmonary artery distortion or kinking (severe in only two cases in this series) is not a major problem in the subsequent management of these children. The growth of small pulmonary arteries, both the ipsilateral and the contralateral pulmonary arteries, is excellent in these infants. The Blalock-Taussig shunt is a superb technique in the palliation of cyanotic congenital heart disease, even in the first few days of life.
We wish to acknowledge the superb clinical assistance of Mr. Lemuel Yerby, Ms. Ann Yerby, Ms. Beth Bain, Ms. Martha Wilkins, Ms. Susan Davis, and our intensive care nurses in the care of these patients. We also appreciate the statistical assistance of Dr. Elmer Hall in performing the analyses described herein. We would like to acknowledge the excellent secretarial assistance of Ms. Margaret Rose in the preparation of this manuscript. REFERENCES Arciniegas E, Blackstone EH, Pacifico AD, Kirklin JW: Classic shunting operations as part of two-stage repair for tetralogy of Fallot. Ann Thorac Surg 27:514-518, 1979 2 Edmunds LH Jr, Stephenson LW, Gadzik JP: The Blalock-Taussig anastomosis in infants younger than I week of age. Circulation 62:597-603, 1980 3 Neches WH, Naifeh JG, Park SC, Lenox CC, Zuberbuhler JR, Siewers RD, Pontius RG, Bahnson HT: Systemicpulmonary artery anastomoses in infancy. J THORAC CARDIOVASC SURG 70:921-927, 1975 4 Laks H, Marco JD, Willman VL: The Blalock-Taussig shunt in the first six months of life. J THORAC CARDIOVASC SURG 70:687-691, 1975 5 Marbarger JP Jr, Sandza JG Jr, Hartmann AF Jr, Weldon CS: Blalock-Taussig anastomosis. The preferred shunt in infants and newborns. Circulation 58:5uppl 1:73-77,1978 6 Chopra PS, Levy JM, Dacumos GC Jr, Berkoff HA, Loring LL, Kahn DR: The Blalock-Taussig operation. The procedure of choice in the hypoxic infant with tetralogy of Fallot. Ann Thorac Surg 21:235-238, 1976 7 Tyson KRT, Larrieu AJ, Kirchmer JT Jr: The BlalockTaussig shunt in the first two years of life. A safe and effective procedure. Ann Thorac Surg 26:38-41, 1978 8 Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 9 Stewart S, Mahoney EB, Manning J: The Waterston anastomosis with no deaths in the neonate. J THORAC CARDIOVASC SURG 72:588-592, 1976 10 Tay OJ, Engle MA, Ehlers KH, Levin AR: Early results and late developments of the Waterston anastomosis. Circulation 50:220-229, 1974 11 Newfeld EA, Waldman JD, Paul MH, Muster AJ, Cole RB, Idriss F, Rikep W: Pulmonary vascular disease after systemic-pulmonary arterial shunt operations. Am J Cardiol 39:715-720,1977
THoRAc CARDIOVASC SURG
85:917-922, 1983
The Blalock-Taussig shunt Low risk, effective palliation, and pulmonary artery growth The Blalock-Taussig shunt has been used at our institution in 64 infants with cyanotic congenital heart disease who required palliation. Thirty-one of these children were less than 2 months of age at the time of operation. There were no early shunt closures. There were two hospital deaths (3%). There was only one hospital death in the 3/ patients less than 60 days old. According to a life-table analysis, 87% ± 9% (± 95% confidence limits) of the shunts were functioning at l year and 78% ± /2% at 2 years. Results are similar in the subgroup of children less than 60 days of age at the time of operation or in the subgroup of children weighing less than 3,999 gm at operation. In infants with small pulmonary arteries, pulmonary arterial growth was excellent (ipsilateral pulmonary artery/descending aorta ratio 0.70 -- 0.95 in 550 days, p < 0.00/; contralateral ratio 0.73 -- 0.99 in 550 days, p < 0.00/). The Blalock-Taussig shunt can be performed with low risk, provides excellent palliation, and is associated with excellent pulmonary artery growth.
Robert A. Guyton, M.D., J. Elwood Owens, M.D., James D. Waumett, M.D., Kenneth J. Dooley, M.D., Charles R. Hatcher, Jr., M.D., and Willis H. Williams, M.D., Atlanta. Ga.
he
Blalock-Taussig shunt has become increasingly popular in small infants as surgical mortality has decreased and the complications of other forms of palliation have been identified.1·7 We have used the Blalock-Taussig shunt in recent years in most children with cyanotic congenital heart disease who required palliation, regardless of age. Over a 5 year period, we performed by Blalock-Taussig shunts in 63 children. Thirty-one of these children were less than a months old at the time of operation. This review was undertaken to evaluate the results of this effort. Particular attention was devoted to the very young children with respect to duration of palliation, distortion of the pulmonary arteries, and pulmonary artery growth. Patients
Between May of 1976 and May of 1981, the BlalockTaussig shunt has been used with increasing frequency
From the Henrietta Egleston Hospital for Children and the Carlyle Fraser Heart Center, Crawford W. Long Memorial Hospital, Emory University School of Medicine, Atlanta, Ga. Received for publication July 30, 1982. Accepted for publication Sept. 10, 1982. Address for reprints: Robert A. Guyton, M.D., Cardiothoracic Research Laboratory, Carlyle Fraser Heart Center, Crawford W. Long Memorial Hospital, Atlanta, Ga. 30365.
at Henrietta Egleston Hospital for Children. Fifty-two operations were performed upon children less than 1 year old, and 31 infants were less than 60 days old (Figs. 1 and 2). Twenty-nine children weighed less than 3,999 gm at operation. The smallest child weighed 1,800 gm. Most of the children in this series had complex cyanotic congenital heart disease with multiple anomalies. Until January of 1981, it was our policy to perform primary total correction in children with "simple" tetralogy of Fallot, regardless of size (the smallest child having successful repair weighed 2,400 gm). Nineteen children in this series had a primary diagnosis of tetralogy of Fallot. Six of these children had anomalous origin of one or more coronary arteries. The other children with tetralogy underwent palliation for a variety of reasons (renal failure, vascular ring, colostomy present, small pulmonary arteries). Nine children had pulmonary atresia, three of whom had ventricular septal defects. Seven children had dextro-transposition of the great arteries (d-TGA) and ventricular septal defect with either pulmonary stenosis or prior pulmonary artery banding. Five children had simple transposition (d-TGA with intact ventricular septum) with inadequate oxygenation after a Blalock-Hanlon procedure in the first few weeks of life. Four patients had a primary diagnosis of tricuspid atresia. Nineteen patients had a variety of other diagnoses including double-outlet right 917
The Journal of Thoracic and Cardiovascular
9 I 8 Guyton et al.
30
Surgery
15
~
-",
0_
~ C Ql::J
.n.s:: EUJ
10
::Jf-
ZeD
20
5
o
o
2
6
12
n
r1
24
> 24
3
4
5
6
7
8
9
Age at Operation (weeks)
-
18
2
Fig. 2. More children underwent the Blalock-Taussig (BT) shunt within the firstweek of lifethan duringany othersimilar time interval.
Age at Operation (months)
Fig. 1. The majority of children included in this series underwent the Blalock-Taussig (BT) shunt within the first 6
months of life.
ventricle, complete atrioventricular canal with pulmonic stenosis, and variations of single ventricle. Fifty-one patients had no prior operation to alter pulmonary blood flow. Four patients had a previous Blalock-Hanlon procedure, five children had a prior pulmonary artery banding, two patients had a prior central shunt, one patient had a prior Blalock-Taussig shunt, and one patient had a prior Brock procedure. Follow-up was complete through February of 1982 for all patients. Operative technique The subclavian artery arising from the innominate artery was chosen for anastomosis in all but two instances (right, 41 patients; left, 21 patients). In one case, a previous Blalock shunt had been performed on the innominate side. In a second case, a vascular ring was divided at the time of anastomosis, and the subclavian artery arising from the smaller, divided, left aortic arch was used for the anastomosis. The chest was entered through an anterolateral thoracotomy in the fourth intercostal space. The pulmonary artery was isolated, controlled proximally by a vascular clamp, and controlled distally by tension on elastic vessel loops. The subclavian artery was isolated, its branches were divided, and it was brought under the vagus nerve. Care was taken to avoid excessive tension on the pulmonary artery. If necessary, and if the subclavian artery was large enough, the anastomosis was performed distal to the first branches of the subclavian artery. In two cases, a section of prosthetic material (4 mm polytetrafluoroethylene graft) was interposed between the end of the subclavian artery and the pulmonary artery in order to
avoid tension. The carotid and innominate arteries were dissected from surrounding tissue to gain additional length. Next, 6-0 or 7-0 polypropylene sutures were used to construct the anastomosis. Simple interrupted sutures were used on the anterior aspect and either simple interrupted sutures (seven patients) or continuous sutures (54 patients) on the posterior aspect. In three patients (the two patients with prosthetic grafts and one patient with d-TGA with intact septum) a continuous suture was used throughout. Optical magnification was used in every case. Heparin was administered prior to anastomosis (100 units/kg) and additionally was continued for 2 days in 41 patients (26 of the 31 children less than 61 days old) at a rate of 25 units/kg/hr. More recently, postoperative heparin dosage has been reduced to maintain the patient's partial thromboplastin time at I Vl to 2 times the control value. Results Operative mortality and morbidity. There were no intraoperative deaths. Two patients died during the hospitalization for the shunt procedure. The first hospital death occurred in a 57-day-old child with a hypoplastic right ventricle and single ventricle physiology, who had pneumonia after the operation. She died with a patent anastomosis 28 days after creation of the shunt. The second death occurred in a 13-month-old child with single ventricle undergoing his third palliative procedure for cyanosis and congestive heart failure. Cardiac arrest occurred during an episode of tachycardia while he was being treated with dopamine for low cardiac output on the second postoperative day. At postmortem examination he was found to have an unsuspected subaortic chamber with a restrictive bulboventricular foramen. Loss of arm tissue and phrenic nerve paralysis did not occur. In one patient, the left recurrent laryngeal nerve
Volume 85 Number 6 June, 1983
C Q) o
iii
Blalock-Taussig shunt
100
100
75
75
C Q)
50
~
a..
(81
25
Q)
9I9
(8)
78 ± 12% at 2 years (17)
50
a..
25
0 0
500
1000
o
<:..L-
---'-
o
500
Days
---'-
1000
---'
1500
Days
Fig. 3. Depicted is a life-table analysis of the cumulative percent event-free (nooperation or death for any reason) in the 63 patients included in this study.
Fig. 4. Cumulative percent adequate shunt function (no death or second operation related to shunt failure) in the 63 patients included in this study.
was paralyzed when a double aortic arch was divided at the time of creation of the shunt. Six patients underwent reexploration for bleeding, five of 41 (12%) patients receiving heparin postoperatively and one of 23 (4%) patients not heparinized postoperatively (difference significant using chi square test, p < 0.05). Duration of palliation. Evaluation of adequacy of shunt function is complicated by the fact that some patients underwent a second operation electively while the shunt was still functioning and other patients died with a shunt still open and providing adequate pulmonary blood flow. Data analysis was therefore carried out by means of the Kaplan-Meier' modification of the life-table analysis for two different conditions: the event-free interval after shunt (no death or second procedure for any reason) and the shunt failure-free interval after shunt (no death or second procedure related to shunt failure). These results are presented in Table I and illustrated in Fig. 3, 4, and 5. Results are similar for the entire series (n = 64), the patients operated upon within the first 2 months of life (n = 31) and the patients operated upon at a weight less than 3,999 gm (n = 29) (Table I). . Late mortality without a second operation. Four children died without a second operation. Three of these children had multiple cardiac defects with single ventricle physiology complicated by cyanosis and congestive heart failure. These three children died outside the hospital and the cause of death is not known. One child died of renal failure secondary to hydronephrosis 6 months after creation of the shunt. Second procedures. Twenty-nine children have undergone a second operation, with 21 hospital survivors
Table I. Duration of palliation Entire series
Age
<6/ days
Percent event free (± 95% CL) At 1 year 74 ± 11 74 ± 16 53 ± 14 At 2 years Percent shunt functioning (±95% CL) At I year 87 ± 9 85 ± 14 At 2 years 78 ± 12
Weight <3,999 gm 72 ± 17
86 ± 15
Legend: CL, Confidence limits.
(72%). Thirteen children had a second procedure designed to adjust pulmonary blood flow. Four of these patients died after the second operation, all having complex anomalies. Sixteen children had a "corrective" procedure. Three infants (4, II, and 21 months) died after early corrective operation was forced by clinical deterioration and complex anatomy. One older child (12 years) died after a Fontan procedure at another institution. Pulmonary artery distortion and growth. Postoperative angiograrns were reviewed in 42 patients. The pulmonary artery was mildly tented up at the site of anastomosis in four patients, two of whom have since undergone successful correction without repair of the pulmonary artery. Kinking was judged to besevere (that is, probably necessitating surgical revision) in two patients, one of whom has since undergone correction without revision of the pulmonary artery. Thirty-six angiograrns revealed no significant distortion of the pulmonary artery. Pulmonary artery size was estimated relative to the
The Journal of Thoracic and Cardiovascular Surgery
9 2 a Guyton et al.
1.5
100
75
C Q) ~ Q) a,
.......
>~
t'Cl
... «0 «>- Cl 0 c:: ~ :c :; c:: c:: a:
181 85 ± 14°0 at 1 year 1201
o
Q)
::::J
Q)
E
50
1.0
~
.5
a.. 0
25
After
Ipsilateral Pulmonary Artery Size
o Cl...a
.l..-
500
----J
1.5
1000
Days
Fig. 5. Cumulative percent adequate shunt function (no death or second operation related to shunt failure) in the 31 patients who underwent operation within the first 60 days of life.
size of the descending aorta at the diaphragm. In patients in whom angiograms were adequate for this analysis, the change in size of the pulmonary arteries was assessed. These results are presented in Table II and depicted in Fig. 6. The patients are divided into two groups: patients in whom the shunt was inserted into a pulmonary artery smaller than the descending aorta (ratio < La) and patients in whom the shunt was inserted into a pulmonary artery larger than the descending aorta (ratio> 1.0). This division is logical and necessary, since one would not expect large pulmonary arteries to grow at the same rate as small pulmonary arteries. The data reveal that large pulmonary arteries grew at the same rate as the descending aorta, but small pulmonary arteries grew faster than the descending aorta after creation of a Blalock-Taussig shunt. Moreover, in this series, the contralateral pulmonary arteries increased in size as much as the ipsilateral pulmonary arteries. There was no difference between right and left Blalock shunts with respect to pulmonary artery growth. Discussion The application of the Blalock-Taussig shunt to small infants has nearly eliminated the use of direct aortopulmonary anastomoses in our institution. This review is an attempt to evaluate the consequences of this change in policy with respect to survival and subsequent complications related to the shunt. Hospital survival rate in this series was 97%, even in patients less than 61 days of age. This result compares
>....... Q)
t'Cl ...
t ~ Cl >- c:: .Q ~ '0 tii c c:: a:
«
o
E::::J
1.0 p< _/ .
Q)
~ Q) a.. 0
00:
.5 I
®
Before
After
Contralateral Pulmonary Artery Size
Fig. 6. Growth of the ipsilateral (Aj and contralateral (Bj pulmonary arteries after creation of the Blalock-Taussig shunt in children in whom the pulmonary artery at the time of anastomosis was smaller than the descending aorta (mean duration 550 days).
favorably with earlier large series of small infants: 91 % survival rate in 33 infants less than 3 months of age from Detroit Michigan, and Birmingham, Alabama, 1 77% survival rate in 31 infants less than 7 days of age from Philadelphia, Pennsylvania,' and 65% survival rate in 34 infant less than 3 months of age from Pittsburgh, Pennsylvania.' Smaller series have emphasized the advantages of the Blalock-Taussig shunt in neonates: 67% survival rate in 18 infants less than 6 months of age from St. Louis University,' 100% survival rate in 11 infants less than 6 months of age from Washington University in St. Louis, Missouri,' 100% survival in seven infants under 1 year of age from Madison, Wisconsin," and 100%survival rate in five infants under 6 months of age from Galveston, Texas.' The improvement in hospital mortality in these children in all institutions reflects improvements in perioperative care as well as surgical and anesthetic techniques. Similar improvement has occurred in survival with the Waterston anastomosis. Ten consecutive survivors of the
Volume 85 Number 6 June, 1983
Blalock-Taussig shunt
921
Table II. Pulmonary artery/descending aorta ratio Duration of shunt (days) Initial ratio <1.0 Ipsilateral PA Contralateral PA Initial ratio> 1.0 Ipsilateral PA Contralateral PA
Before shunt
After shunt
Significant change
550 ± 332
0.70 ± 0.12 0.73 ± 0.18
0.95 ± 0.19 0.99 ± 0.27
p<0.OOI,n=22 p < 0.00I, n = 19
515 ± 580
1.30 ± 0.29 1.30 ± 0.33
1.30 ± 0.39 1.24 ± 0.29
NS, n = 10 NS. n = 10
Legend: PA. Pulmonary artery. NS. Not significant.
Waterston anastomosis performed in the first week of life were reported from Rochester, New York." We conclude that the Blalock-Taussig shunt can be successfully performed in very young or small infants with a hospital survivalrate at least as good as that of any other anastomosis. Hospital morbidity related to the surgical procedure was rare, with only one instance of recurrent laryngeal nerve damage. Phrenic nerve damage and loss of arm tissue did not occur. Early shunt closure was also not observed. In most of the smaller infants (26/31), heparin anticoagulation was employed for the first 2 postoperative days. Postoperative anticoagulation did increase the reexploration rate for postoperative bleeding in the heparinized versus nonheparinized group (12% versus 4%, p < 0.05). We continue to use heparin postoperatively because we feel subjectively that anticoagulation may improve early shunt patency, although we have no data to support this hypothesis. There were no early graft closures in either group, i.e., whether heparinized postoperatively or not. AIl of the patients who were at high risk for early shunt closure were given anticoagulant therapy. Evaluation of the clinical course of these children is complicated by the diverse and complex diagnoses. Within the tetralogy of Fallot group, most of the patients had some additional problem (cardiac or otherwise) which forced palliation rather than correction. During this period, our policy was to correct primarily all tetralogy anomalies in which the anatomy was suitable, regardless of age. The smallest child undergoing successful correction weighed 2,400 gm. The one operative death at the time of tetralogy correction in this series occurred in a child with persistent pulmonary problems after division of a double aortic arch. Four patients died during attempted additional palliative procedures to increase pulmonary blood flow. Three other children died during "corrective" operations for complex situations. Twenty-one patients have survived a second operation. An evaluation of the duration of adequate palliation is
also difficult. One method is to evaluate the event-free interval after the shunt procedure. However, this tends to underestimate the duration of effective palliation. We repair d-TGA at 1 year of age and tetralogy before 2 years of age regardless of the effectiveness of previous palliation. Therefore, an "event" might occur in these children before the end of effective palliation. The second method which we employed, therefore, was an actuarial analysis of the shunt failure-free interval after shunt creation. In this analysis, children who died or underwent a second operation without shunt failure were considered "removed from the series alive" at the time of second operation or death. An example of death without shunt failure would be the child who died of renal failure with a patent shunt. This second method of analysis might tend to overestimate the duration of effective palliation if shunt inadequacy contributed to the death or second operation in an unrecognized manner. The actual duration of effective palliation probably is between these two estimates, that is, between 74% and 87% effective palliation at 1 year (Figs. 3 and 4). It is particularly notable that there is no less effective palliation in the children who underwent operation within the first 60 days of life or at a weight of less than 3,999 gm (Table I and Fig. 5). A critically important part of any evaluation of shunt function is the fate of the pulmonary arteries after creation of the shunt. The development of pulmonary vascular obstruction is uncommon in the first 5 years after creation of a Waterston shunt and very rare after creation of a Blalock-Taussig shunt." Severe pulmonary vascular obstruction, with reversal of flow in the Blalock-Taussig shunt, developed in one child in this series 4 months after shunt creation. Because the pulmonary artery was not entered in pre-shunt catheterizations in this infant with d-TGA, the pulmonary vascular resistance prior to shunt creation is unknown. The primary liability of the Waterston anastomosis is distortion and kinking of the pulmonary artery at the site of anastomosis.I I We have been concerned that the Blalock-Taussig shunt, when performed in these very
922
The Journal of Thoracic and Cardiovascular Surgery
Guyton et al.
small children with tiny, delicate pulmonary arteries, might lead to similar distortion. The results of this study, with severe kinking in only two of 64 patients, indicate that distortion of the pulmonary artery is not a major problem in the subsequent management of these children. A final important consideration is the growth (or failure to grow) of the pulmonary arteries after shunt creation. Since one would not expect large pulmonary arteries to grow after shunt creation, these children were separated from the children with small pulmonary arteries. A significant increase in the size of small pulmonary arteries did occur, bringing the size of these arteries very close to the size of the descending aorta. This occurred in both left and right Blalock shunts, and growth was equivalent in the ipsilateral and in the contralateral pulmonary arteries (Table II, Fig. 6). Such growth was not observed in large pulmonary arteries, which tended to stay the same size relative to the descending aorta. This study supports the contention of others that the Blalock-Taussig shunt can be performed with a high hospital survival rate (97% in this series) even in very young and small infants. This success reflects the improvements in perioperative care and anesthesia and surgical techniques that have occurred in recent years. The large number of patients included in this series allows a statistical evaluation of the effectiveness of the Blalock-Taussig shunt, an analysis heretofore unavailable. Effective palliation can be expected in about 85% of these infants 1 year after shunt creation, regardless of age or size at the time of operation. Two thirds of the children in this series have undergone postoperative cineangiograms, providing substantial information about the fate of the pulmonary artery after the Blalock-Taussig procedure. Pulmonary artery distortion or kinking (severe in only two cases in this series) is not a major problem in the subsequent management of these children. The growth of small pulmonary arteries, both the ipsilateral and the contralateral pulmonary arteries, is excellent in these infants. The Blalock-Taussig shunt is a superb technique in the palliation of cyanotic congenital heart disease, even in the first few days of life.
We wish to acknowledge the superb clinical assistance of Mr. Lemuel Yerby, Ms. Ann Yerby, Ms. Beth Bain, Ms. Martha Wilkins, Ms. Susan Davis, and our intensive care nurses in the care of these patients. We also appreciate the statistical assistance of Dr. Elmer Hall in performing the analyses described herein. We would like to acknowledge the excellent secretarial assistance of Ms. Margaret Rose in the preparation of this manuscript. REFERENCES Arciniegas E, Blackstone EH, Pacifico AD, Kirklin JW: Classic shunting operations as part of two-stage repair for tetralogy of Fallot. Ann Thorac Surg 27:514-518, 1979 2 Edmunds LH Jr, Stephenson LW, Gadzik JP: The Blalock-Taussig anastomosis in infants younger than I week of age. Circulation 62:597-603, 1980 3 Neches WH, Naifeh JG, Park SC, Lenox CC, Zuberbuhler JR, Siewers RD, Pontius RG, Bahnson HT: Systemicpulmonary artery anastomoses in infancy. J THORAC CARDIOVASC SURG 70:921-927, 1975 4 Laks H, Marco JD, Willman VL: The Blalock-Taussig shunt in the first six months of life. J THORAC CARDIOVASC SURG 70:687-691, 1975 5 Marbarger JP Jr, Sandza JG Jr, Hartmann AF Jr, Weldon CS: Blalock-Taussig anastomosis. The preferred shunt in infants and newborns. Circulation 58:5uppl 1:73-77,1978 6 Chopra PS, Levy JM, Dacumos GC Jr, Berkoff HA, Loring LL, Kahn DR: The Blalock-Taussig operation. The procedure of choice in the hypoxic infant with tetralogy of Fallot. Ann Thorac Surg 21:235-238, 1976 7 Tyson KRT, Larrieu AJ, Kirchmer JT Jr: The BlalockTaussig shunt in the first two years of life. A safe and effective procedure. Ann Thorac Surg 26:38-41, 1978 8 Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 9 Stewart S, Mahoney EB, Manning J: The Waterston anastomosis with no deaths in the neonate. J THORAC CARDIOVASC SURG 72:588-592, 1976 10 Tay OJ, Engle MA, Ehlers KH, Levin AR: Early results and late developments of the Waterston anastomosis. Circulation 50:220-229, 1974 11 Newfeld EA, Waldman JD, Paul MH, Muster AJ, Cole RB, Idriss F, Rikep W: Pulmonary vascular disease after systemic-pulmonary arterial shunt operations. Am J Cardiol 39:715-720,1977