Since 1954, when Lillehei completed the first successful

Transcription

1 Current Surgical Management of Tetralogy of Fallot Vaughn A. Starnes, MD, Giovanni Battista Luciani, MD, David A. Latter, MD, and Michael 1. Griffin, MD Division of Cardiothoracic Surgery, University of Southern California School of Medicine, Los Angeles; Department of Cardiovascular Surgery and Division of Pediatric Cardiology, Stanford University Medical Center, Stanford, California The surgical management of tetralogy of Fallot has undergone important changes in recent years. Earlier repair of tetralogy of Fallot is now favored by many institutions. At Stanford University Medical Center, we have performed definitive repair of tetralogy of Fallot at the time of presentation, regardless of the child's age, with few exceptions. In this report, we describe our results with early repair, and we believe these support the contention that infants who undergo early repair «1 year of age) have postoperative results similar to those of children who undergo repair at an older age. Complications related to shunts are prevented by the infant repairs, and, in the future, reduced ventricular ectopy may be demonstrated to be a benefit of such repairs. (Ann Thorae Surg ) Since 1954, when Lillehei completed the first successful definitive repair of tetralogy of Fallot (ToF) [1], the medical and surgical management of ToF has continued to evolve. Arterial pulmonary shunting, as first used by Blalock and Taussig and initially the cornerstone of treatment of ToF [2], has become a controversial method in recent years. Dramatic advances in the general field of infant cardiac surgical procedures that have made possible low-risk infant operations have opened the way for early «1 year of age) complete repair of ToF. Castaneda [3] and others [4, 5] have championed the concept of early repair and have shown excellent operative results with a low mortality risk (5% to 7%). This approach has not yet been completely accepted, however, and many authors have presented data suggesting that early repair may carry a greater operative mortality [6, 7]. Kirklin and co-workers [6, 8] maintain that, in the minority of children with ToF who require operative intervention in the first 9 months of life, the best approach is still a two-staged one; a Blalock shunt followed later by complete repair. A previous report from this institution [7] examined the late surgical results in 309 children with ToF and identified younger age as an independent risk factor for operative mortality of complete repair. More recently at this institution, early primary repair has become the standard approach to the surgical management of ToF. Currently, a definitive repair is performed in children with ToF at the time of presentation, regardless of their age or weight. If the child is asymptomatic, then preferably elective repair is done when he or she is younger than 1 year of age. Contraindications to early repair include cases in which there is an anticipated need for a conduit repair, such as an anomalous origin of the left anterior descending from Accepted for publication Nov 27, Address reprint requests to Dr Starnes, Division of CardiothoracicSurgery, Childrens Hospital Los Angeles, University of Southern California Schoolof Medicine, 15 San Pablo St, Ste 415, Los Angeles, CA by The Society of Thoracic Surgeons the right coronary sinus; other major congenital anomalies that would preclude the low risk of the definitive operation; and miscellaneous factors, such as the child's family being Jehovah's Witnesses. The purpose of the present study was to retrospectively analyze the results of the current surgical management of early repair «1 year of age) by assessing the preoperative, perioperative, and intermediate period of follow-up, in order to formulate the optimal strategy and timing of operation. Material and Methods Tetralogy of Fallot was defined as the constellation of findings resultingfrom anterior malalignmentof the conal septum and producing a ventricular septal defect (VSD); infundibular, valvar, and supravalvar pulmonary stenosis; an overriding aortic valve; and right ventricular hypertrophy. Children with a double-outlet right ventricle (as evidenced by aortic-mitral discontinuity), a doublechambered right ventricle, and valvar or truncal pulmonary atresia (as evidenced by right ventricle-pulmonary artery discontinuity) were excluded from the study. From January 1988 to April 1990, the cases of all children operated on by one surgeon (V.A.S.) for the management of ToF were retrospectively assessed. Thirty-seven patients underwent the procedure. One patient received a Blalock shunt because his parents refused transfusions, and this was believed to substantially increase the risk associated with the definitive repair. Thirty-six children were enrolled in the study. The patients' medical records were reviewed to obtain information on the preoperative and perioperative conditions. Intermediate follow-up investigations included telephone interviews with the parents, telephone and letter interviews with the referring pediatric cardiologists, echocardiography (M mode and Doppler), and 24-hour Holter monitoring. Low-flow (0 ml. min- t m- 2 ), hypothermic (20 C), bicaval cardiopulmonary bypass was used in every case /94/$7.00

2 212 STARNES ET AL Ann Thorae Surg Early in the series, 2 patients who did not require transannular patching underwent transventricular repair of the VSD and infundibular obstruction. Although these 2 children did well, the advantage to transatrial repair without ventriculotomy when transannular patching is not necessary became evident. Thereafter, repair of the right ventricular outflow tract obstruction in all patients who did not require a transannular patch was performed through the right atrium and tricuspid valve. A transverse main pulmonary arteriotomy was occasionally added so that the adequacy of the obstruction relief could be directly assessed. Transannular repairs were done whenever intraoperative assessment of the pulmonary valve annulus, as measured with a Hegar dilator passed through the right atrium, showed the annulus to be greater than one standard deviation smaller than that predicted by Rowlatt's standard chart [9]. An incision was made from the main pulmonary artery across the pulmonary valve annulus and down the infundibulum that was just enough to adequately relieve the infundibular obstruction. Little infundibular muscle resection was done; instead, obstructing muscle bundles were divided. The pulmonary valve was excised if it was dysplastic. Native pericardium tanned in glutaraldehyde was used as the patch material and was extended distally onto the branch pulmonary arteries if relief of obstruction was required at this level. A point of great importance in the repair was to not detach the pulmonary artery from the aorta, as it was thought that this could provide support for the reconstructed pulmonary artery and prevent the troublesome problem of recurrent stenosis of the origins of the branch pulmonary arteries. This occurs if the main pulmonary artery is able to "ride up" with each systole. Except for the two transventricular repairs just mentioned, all the VSDs in patients not requiring transannular patching were repaired through the right atrium. When a transannular patch was required, the VSDs were approached through the right ventriculotomy. Dacron was used as the VSD patch material. Simple ligation of the Blalock shunt, when present, was done at the start of cardiopulmonary bypass. Myocardial protection was accomplished with one application of cold potassium blood cardioplegia administered in the aortic root, followed by continuous cold topical pericardial irrigation. Myocardial temperature was measured in all patients and maintained at less than 15 C. Intracardiac monitoring lines were used liberally. The threshold for the use of inotropic agents, pulmonary vasodilators, or both was low, as an adequate perfusion pressure must be maintained to resuscitate the hypertrophied right ventricular myocardium. Statistical analysis was done using a two-tailed unpaired t test for assessing parametric data and the Mann Whitney rank test for evaluating nonparametric data. Results Thirty-six children formed the basis of this study. One child was lost to follow-up. Two children, who lived in Age Groups In Ve.rs Fig 1. Age distribution of the 36 patients at the time of complete repairof tetralogy of Fallot. Fifteen patients were less than 1 year of age at the time of repair (group 1) and 21 were between 1 and years old (group 2). remote areas, were known to be doing well, but were inaccessible to follow-up examination. Therefore, 33 children were available for intermediate follow-up assessment that included echocardiography and Holter monitoring. The mean age of the entire group of 36 children, was 2.17 years (range, 56 days to years). Fifteen children (group 1) underwent repair when they were younger than 1 year old (mean, 0.48 year), and 21 children (group 2) underwent repair when they were older than 1 year old (mean, 3.36 years) (Fig 1). The mode of repair of the right ventricular outflow tract obstruction was determined by the size of the pulmonary valve annulus. There was no substantial difference in the type of repair between the two groups (13 versus 12 transannular and 2 versus 7 transatrial, in group 1 versus group 2, respectively). The indications for operation were also similar for the two groups: 13 children presented electively (7 versus 6, in group 1 versus group 2, respectively); 7 had hypoxic spells (3 versus 4); suffered from cyanosis (4 versus 6), and 6 showed decreased exercise tolerance (1 versus 5). The symptom of decreased exercise tolerance obviously cannot be elicited from an infant, and failure to thrive was therefore used as a similar end point in these patients. Of note are the 7 children in group 1 who were operated on electively. The preoperative hemoglobin levels, oxygen saturation, and weight were recorded. The mean weight was 7.06 kg in the group 1 children and kg in the group 2 children. The hemoglobin (13.6 g/dl versus 13.7 g/dl in group 1 and 2, respectively) and oxygen saturation (88% versus 89%) values were similar for both groups. Associated cardiovascular anomalies consisted mainly of aortic arch abnormalities. Three children had anomalous coronary arteries, but in only 1, who had a left anterior descending coronary artery from the right coronary sinus, was this of any significance to the operative repair. In this particular case, relief of the right ventricular outflow obstruction was achieved by a transatrial and transpulmonary approach. No patients in either group died, and all of the 35 children who could be contacted at follow-up were alive.

3 Ann Thorac Surg STARNES ET AL 213 P. 02 > Q IiII Less Than 1 yr Old Greeter Than 1 yr Old!l e 0.e 0.4 '" l1. IiII Less Than 1 yr Old Greater Than 1 yr Old Days on Respirator Days In leu Days of Hospitalization Fig 2. Perioperative status of 36 patients undergoing complete repair of tetralogy of Fallot. (leu = intensive care unit.) Perioperative management in both groups was characterized by relative ease of management. The durations of respirator dependency, stay in the intensive care unit, and hospitalization for the two groups are depicted in Figure 2. A significant difference was observed for the older age group, who showed more rapid progress. Major postoperative morbidity occurred in only 1 patient, a 5-month-old boy who was given an accidental overdose of dopamine that produced acute tubular necrosis. The child required peritoneal dialysis for 1 week. At follow-up 3 months postoperatively, there was no evidence of renal insufficiency. Minor complications, such as reintubation for the management of respiratory insufficiency, were encountered, but none had led to any significant sequelae. There were no neurologic complications. Five children in group 1 had previously received a systemic-to-pulmonary shunt. Four children had a unilateral Blalock-Taussig shunt, and 1 had bilateral Blalock-Taussig shunts and a previous right ventricular-to-pulmonary artery patch. AIlS children required transannular repair at the time of definitive repair. Two of these children required reoperation. One 4-year-old girl required a valved homograft for the treatment of residual pulmonary artery stenosis thought to be a consequence of the previous Blalock shunt. One lo-year-old girl needed reoperation for a residual VSD, and at the time of follow-up was in New York Heart Association (NYHA) functional class ill (the only child not in class I). Two of the 5 children who had previous shunts now require digoxin. Intermediate follow-up was done at a mean of 1.13 years postoperatively. All the children were doing well, with the exception of the 1 child who was in NYHA class ill (Table 1). Thirty-three children were assessed by echocardiogra- Table 1. Intermediate Follow-up Exercise Age Follow-up Alive NYHAI Digoxin Intolerance <ly 12.0 rna 15/15 15/ >ly 14.8 rna 21/21 20/21 4/ NYHA I = New York Heart Association functional class I mmhg 20 40mmhg Peak RV to PA Gradient 40 60mmhg Fig 3. Right ventricular (RV)-to-pulmonary artery (PA) pressure gradient as estimated by postoperative Doppler echocardiography in 31 patients undergoing complete repair of tetralogy of Fallot. phy at a mean of 1.13 years postoperatively (range, 50 days to 2.28 years). An echocardiographic-doppler assessment for the presence of residual VSD, tricuspid regurgitation, residual right ventricular outflow tract obstruction, and pulmonary insufficiency was done in all 33 children. In 25 echocardiograms, an M-mode assessment of right ventricular dimensions was obtained in an effort to ascertain the status of right ventricular function. No difference in the residual VSD status was found between the two groups: there was one trivial VSD in group 1 versus four trivial VSDs in group 2. As already mentioned, 1 child in group 2 underwent reoperation for the treatment of a residual VSD, and subsequently had no residual VSD. Doppler interrogation was used to obtain the peak systolic right ventricle--to-pulmonary artery pressure gradient. Most children in both groups showed either no or a minimal «20 mm Hg) peak gradient. Only 1 child in each group had peak gradients of 40 to 60 mm Hg. Of the group 1 children, the youngest child in the series, whose repair was done at 56 days of age, underwent balloon angioplasty in an unsuccessful attempt to relieve the residual obstruction. Nonetheless, there was no significant difference between the two groups in terms of the peak systolic gradients (Fig 3). The echocardiographic assessment of right ventricular function included measurements of the degree of tricuspid regurgitation (Fig 4), the degree of pulmonary regurgitation (Fig 5), and the right ventricular end-diastolic dimension, expressed as a ratio to the predicted value (Fig 6). Only tricuspid regurgitation was seen to be significantly different between the two age groups, with less tricuspid regurgitation observed in the younger patients (p < 01). Continuous Holter monitoring was carried out in all 33 children accessible to follow-up. In the age group younger than 1 year (15 children), no ventricular ectopy activity (VEA) was noted in any child, placing them all in Lown's class 0 []. In the older age group (18 children), however, 2 were in Lown's class 1A and 1 was in Lown's class 2.

4 214 STARNES ET AL Ann Thorae Surg Ẹ S! CiI "- 0.6 :! "'" c 0.4 "e ""- 0.9 E " 0.7 :; "- 0.6 " 0.5 :! '" 0.4 c "e 0.3 ""- 0.1!i!I Less Than 1 yr Old Greater Than 1 yr Old Degree of Tricuspid Regurgitation Fig 4. Amount of tricuspid regurgitation (RegurgJ found at postoperative Doppler echocardiography in 33 patients undergoing complete repair of tetralogy of Fallot. e 1.0 >1 - <1.5 >1,5 - <2.0 RVEDD I Predicted RVEDD (For BSA) :>2.0 - <2,5 Fig6. Right ventricular end-diastolic dimension (RVEDD) as calculated by postoperative echocardiography in 24 patients undergoing complete repair of tetralogy of Fallot. The valuesare corrected for the RVEDD estimated for the bodysurface areas (BSA) of the patients. Comment There have been reports from several institutions on series of patients who underwent ToF repair with very low mortality rates [3-6, 7]. In fact, the operative mortality rate appears to be approaching zero in these patients. Much of this advance has stemmed from the improvements in the practice of infant cardiac surgery in general. The issues that remain unresolved in the management of ToF include the role of initial palliative procedures versus primary repair in symptomatic children, the timing of repair in asymptomatic children, and the mode of complete repair with or without transannular repair. Since Castaneda and associates [3] and Barratt-Boyes and Neutz [5] first demonstrated the feasibility of early primary repair, this practice has been widely adopted. Data supporting this approach have been presented by teams from a number of centers [3-5]. Kirklin and coworkers [8], however, have questioned the advisability of widespread adoption of this practice, particularly in younger infants. They suggest that these infants may be best served by an initial Blalock shunt, followed by repair at the age of 2 years. Accordingly, shunt procedures have :l c 1.0.S! CiI "- 0.6 " '" 04 l'l :;; c, El Less Than 1 yr Old Greater Than 1 yr Old "=1 No Putm. Regurg Pulm. Regurg. 3+ Pulm. Aegurg Degree of Pulmonary Regurgitation Fig5. Amount of pulmonary regurgitation (Pulrn. RegurgJ found at postoperative Doppler echocardiography in 33 patients undergoing complete repair of tetralogy of Fallot. a negligible risk and are not associated with an incremental risk at the second stage of treatment, the definitive repair [6]. The operative mortality figures (ranging from 0 to 7%) cited for various published series of early primary repair are very encouraging [3-5]. The results observed at our institution have increased our enthusiasm for this approach. We are pleased with the operative mortality of zero, which, even if corrected for the confidence limits (0 to 5%), compares favorably with previously reported results. Further, we believe that the long-term survival will continue to be good, as has been demonstrated in numerous studies of long-term results in children with ToF repairs [7, 11-13]. Criticism of early primary repair has included the high incidence of transannular patching (85% in the Walsh and colleagues' series [11] and 83% in ours). We believe that the younger the infant with ToF at the time of presentation, the more severe the anatomy involved. Therefore, it is not realistic to compare a 56-day-old child with severe cyanosis who undergoes primary repair with a child who undergoes repair at 2 years of age. Perhaps it would be reasonable to compare infants who undergo repair primarily at a given age with children who had shunts at the same given age and then undergo complete repair at a later time. There is ample evidence that the pulmonary valve annulus does not "grow," even with an arterial pulmonary shunt in place [6, 14]. In other words, once a transannular repair, always a transannular repair. This implies that delaying definitive surgical treatment will not produce fewer transannular repairs. Pulmonary regurgitation is well tolerated in the setting of unobstructed pulmonary arterial systems [6, 7, 11]. Important contributions from Pacifico and associates [15] have clarified when transannular repair, with its obligatory pulmonary insufficiency, is needed. We did not see a significant difference in the amount of pulmonary regurgitation between the two age groups, and therefore have no qualms about early repair with regard to pulmonary regurgitation. We recognize the need for the long-term follow-up assessment of right ventricular function, but do

5 Ann ThoraeSurg STARNES ET AL 215 not share the belief of others [16] that a significant percentage of patients will go on to require a valved pulmonary circuit. Disturbing instances of sudden death have been noted during the long-term follow-up of patients with ToF who have undergone repair. Walsh and colleagues [12] reviewed the literature and found many series in which the incidence of VEA graded as greater than Lown's class 2 ranged from 17% to 44% in ToF patients who had undergone repair. In their own data, they noted a much lower incidence of significant VEA and attributed this to performing repair at an early age, thereby effecting early relief of right ventricular hypertension, preventing left ventricular volume overload from occurring as the result of arterial pulmonary shunts, and correcting systemic hypoxemia. Sullivan and co-workers [17] presented data that likewise implicate the duration of uncorrected ToF physiology as the cause of VEA, seen both preoperatively and postoperatively. One can extrapolate from published data regarding the incidence of sudden death in association with premortern VEA [] and conclude from this that the children undergoing repair at an older age will have more problems with VEA, and therefore may be at increased risk for suffering late sudden death. Given all these arguments, we believe that an early primary repair of symptomatic infants, regardless of their age or weight, is the optimal way to manage these patients. Our data encourage us to continue with this treatment protocol. The technique of repair has been well described. We use the transatrial approach whenever possible, but we do not hesitate to perform a transannular repair when indicated. The results presented in this paper are from a relatively small number of patients (36 children) and represent an intermediate duration of followup, but they are in agreement with those from other reported series. We plan to continue to follow the progress of these children and look forward to determining the incidence of VEA and status of cardiac function at later follow-up. References 1. Gott VL. Walton Lillehei and total correction of tetralogy of FaIJot. Ann Thorac Surg 1990;49: Blalock A, Taussig HB. Treatment of malformations in which there is pulmonary stenosis or atresia. JAMA 1945;29: Castaneda AR, Freed MD, Williams RG, Norwood WI. Repair of tetralogy of FaIJot in infancy: early and late results. J Thorae Cardiovasc Surg 1977;74: Gustafson RA, Murray GF, Warden HE, et al. Early primary repair of tetralogy of FaIJot. Ann Thorac Surg 1988;45: Barratt-Boyes BG, Neutz JM. Primary repair of tetralogy of FaIJot in infancy using profound hypothermia with circulatory arrest and limited cardiopulmonary bypass: a comparison with conventional two-stage management. Ann Surg 1973;178: Kirklin JW, Blackstone EH, Kirklin JK, et al. Surgical results and protocols in the spectrum of tetralogy of FaIJot. Ann Surg 1983;198: Zhao HX, Miller DC, Reitz BA, et al. Surgical repair of tetralogy of FaIJot. J Thorac Cardiovasc Surg 1985;89: Kirklin JW, Blackstone EH, Colvin EV, et al. Early primary correction of tetralogy of FaIJot. Ann Thorac Surg 1988;45: Rowlatt JF, Rimoldi HJA, Lev M. The quantitative anatomy of the normal child's heart. Pediatr Clin North Am 1963;: Lown B, Graboys TB. Management of patients with malignant ventricular arrhythmias. Am J Cardiol 1977;39: Walsh EP, Rockenmacher S, Keane JF, et al. Late results in patients with tetralogy of FaIJot repaired during infancy. Circulation 1988;77: Chen 0, MolJer JH. Comparison of late clinical status between patients with different hemodynamic findings after repair of tetralogy of FaIJot. Am Heart J 1987;113: Horneffer PI, Zahka KG, Rowe SA. Long-term results of total repair of tetralogy of FaIJot in childhood. Ann Thorac Surg 1990;50: Sade R, Sloss I, Treves S. Repair of tetralogy of FaIJot after aorto-pulmonary anastomosis. Ann Thorac Surg 1977;23: Pacifico AD, Kirklin JW, Blackstone EH. Surgical management of pulmonary stenosis in tetralogy of FaIJot. J Thorac Cardiovasc Surg 1977;74: Ilbawi MN, Idriss FS, Serafin Y. Factors that exaggerate the deleterious effects of pulmonary insufficiency on the right ventricle after tetralogy repair. J Thorac Cardiovasc Surg 1987;93: Sullivan, Prebiterio P, Gooch VM. Is ventricular arrhythmia in repaired tetralogy of Fallot an effect of operation or a consequence of the course of the disease? Br Heart J 1987;58: 40-4.