- Email: [email protected]
Results of total cavopulmonary connection in the treatment of patients with a functional single ventricle
J
THORAC CARDIOVASC SURG
1991;102:280-7
Results of total cavopulmonary connection in the treatment of patients with a functional single ventricle Total cavopulmonary connection was proposed as a modification of the Fontan procedure that might have greater benefits than previous methods. To assess tbis procedure we reviewed case bistories of 38 patients (aged 17 months to 30 years) who underwent Fontan procedures with cavopulmonary anastomoses between January 1987 and December 1989. The group included 32 patients with univentricular heart, 2 with pulmonary atresia and intact ventricular septum, 3 with tricuspid atresia, and 1 with hypoplastic left heart syndrome. One or more previous palliative procedures were performed in 34 patients, including 19 systemic-pulmonary shunts, 16 pulmonary artery bandings, 7 atrial septectomiesjseptostomies, 7 Glenn shunts, and 1 patent ductus arteriosus ligation. Preoperative hemodynamics showed a pulmonary artery pressure of 12 mm Hg (range 6 to 22 mm Hg), pulmonary-systemic flow ratio of 1.6 (range 0.37 to 3.0), left ventricular end-diastolic pressure 9 mm Hg (range 3 to 15 mm Hg), and systemic arterial oxygen saturation of 82 % (range 67 % to 94 % ~ Concomitant with cavopulmonary connection, 13 patients underwent additional procedures, including 9 atrioventricular valve annuloplasties, 4 Damus-Stansel-Kaye procedures, and 2 resections of subaortic membranes. Modifying the Fontan procedure in this fashion was particularly useful in the management of 2 patients with pulmonary atresia and intact ventricular septum who had right ventricular-dependent coronary blood flow. Cavopulmonary anastomosis and atrial septectomy were performed in both patients, with resultant inflow of oxygenated blood to the right ventricle and coronary arteries. ExceUent postoperative results were noted in each. Postextubation hemodynamics for the entire group included a mean right atrial pressure of 13 mm Hg (range 11 to 17 mm Hg), a mean left atrial pressure of 6 mm Hg (range 3 to 12 mm Hg), and a room air oxygen saturation of 96% (range 92% to 98 %). Seven patients had pleural effusions, 3 required postoperative pacemaker placement, and 2 required reoperation for tamponade. A venous assist device was required in one patient on the second postoperative day, but the patient was weaned successfuUy within 24 hours. One early death (2.6%) occurred in a patient who had intractable ventricular fibriUation 2 days after operation. There was one late cardiac death (2.7%) caused by ventricular failure and one late noncardiac death. These results demonstrate that total cavopulmonary connection providesexceUent early definitive treatment, with low morbidity and mortality, for a variety of complex congenital heart lesions.
Darryl G. Stein, MD,a Hillel Laks, MD,a Davis C. Drinkwater, MD,a Lester C. Permut, MD,a Henry W. Louie, MD,a Jeffrey M. Pearl, MD,a Barbra L. George, MD,b and Roberta G. Williams, MD,b Los Angeles, Calif.
From the Division of Cardiothoracic Surgery' and the Division of Pediatric Cardiology," UCLA School of Medicine, Los Angeles, Calif. Read at the Sixteenth Annual Meeting of The Western Thoracic Surgical Association, Coronado, Calif., June 20-23, 1990. Address for reprints: Hillel Laks, MD, Division of Cardiothoracic Surgery, University of California, Los Angeles, UCLA Medical Center, 62-182 CHS, 10833 Le Conte Ave., Los Angeles, CA 90024.
12/6/27993
280
In 1971 Fontan and Bauder' reportedthe firstclinically successful procedure for total bypass of the right sideof the heart fortreatment oftricuspid atresia.The procedure included construction of a Glenn shunt, closure of the atrial septal defect, homograftvalve insertion at the orifice ofthe inferiorvenacava (IVC), placement ofa valved homograftfrom the right atrium (RA) to the left pulmonary artery (PA), and ligation of the main PA. Sincethe
Volume 102 Number 2
Functional single ventricle
August 1991
original description, a number of modifications to Fontan's operation have been described that have simplified the operation and allowed its application to a variety of complex congenital heart lesions. Kreutzer and coworkers.l in 1973, described a method for right ventricular (RV) bypass by construction of an end-to-end anastomosis between the RA appendage and the detached pulmonary anulus without a concomitant Glenn shunt. The success of this technique demonstrated that an RA inlet valve was dispensable and that the entire systemic venous return could be directed to the PA through a single connection. Shemin and colleagues' examined the necessity for an outlet (RA-PA) valve in dogs. They found that blood flow through RA-PA conduits was influenced primarily by RA pressure and was not affected by the presence or absence of a valve. Shortly thereafter Gale and associates" reported a 17% prevalence of valved conduit-related complications and presented favorable results using a direct valveless RA-PA connection. Puga, Chiavarelli, and Hagler-Iater described total cavopulmonary connection as a valveless repair that avoided a direct atriopulmonary connection. In their technique they used a lateral intraatrial baffle to divert Ive blood to the superior vena cava (SVC). This was followed by transection of the SVC at the levelof the right PA and end-to-side anastomoses of each transected end to the P A. To determine the contribution of RA contraction in a valveless atriopulmonary connection, de Leval and coworkers" performed a series of hydrodynamic studies with plastic models of the human RA. These investigations showed that pulsation in a valvelesschamber generated turbulence, with resultant elevation of pressure within the chamber and diminution of forward flow. Additionally, turbulence generated by flow through cavities, around corners, and through stenoses further inhibited forward flow. They concluded that maximal forward output could be achieved by streamlining blood flow through a valvelessRA tunnel of uniform caliber. On the basis of these experimental observations, de Leval and coworkers" described a modified Fontan operation, very similar to the one described by Puga and associates.P The procedure consists of an end-to-side anastomosis of the SVC to the superior aspect of the undivided right PA, construction of a lateral intraatrial tunnel with the use of a patch and the lateral wall of the RA to connect the IVC to the SVC, and anastomosis of the cardiac end of the transected SVC, augmented with pericardium, to the inferior aspect of the main P A. The potential benefits include (1) technical ease of repair, (2) reduced atrial turbulence with resulting improved hemodynamics, and (3) decreased arrhythmias.
281
To assess the possible benefits of such an approach we reviewed the case histories of 38 consecutive patients who underwent the Fontan procedure with total cavopulmonary anastomosis between January 1987 and December 1989. This group made up an increasing percentage of all patients undergoing Fontan procedures at our institution during this period, constituting 71% of all Fontan-type procedures we performed in 1989.
Patient profile Nineteen patients were male and 19 were female. Age at operation ranged from 17 months to 30 years, with a mean of 7 years. The largest group included 32 patients (84%) with univentricular heart. Fifteen of these had left ventricular anatomy, 16 had right ventricular chambers, and one had indeterminate anatomy. Three patients (8%) had tricuspid atresia, 2 (5%) had pulmonary atresia with intact ventricular septum and RV-coronary collaterals, and 1 (3%) had hypoplastic left heart syndrome. Thirty-four patients underwent one or more prior palliative procedures. Nineteen systemic-PA shunts were performed in 17 patients. Sixteen patients underwent PA banding, 7 had atrial septectomies/septostomies, 7 had modified Glenn shunts, 1 underwent ligation of the patent ductus arteriosus, and 1 had undergone pacemaker placement for intermittent complete heart block and bradycardia. Total cavopulmonary connection was performed in patients with cyanotic congenital heart disease with a functionally univentricular heart. The selection criteria were similar to those used for previous Fontan procedures and included acceptable ventricular function and pulmonary vascular resistance. An ejection fraction of greater than 45% and a ventricular end-diastolic pressure of less than 15 mm Hg were considered optimal. A pulmonary vascular resistance of greater than 2 U /m 2 or a PA pressure greater than 25 mm Hg was unacceptable. Ventricular function as assessed by echocardiography was considered good to excellent in all patients. Angiography demonstrated a mean ejection fraction of 58% (range 43% to 72%). Ventricular end-diastolic pressures ranged from 3 mm Hg to 15 mm Hg (mean 9 mm Hg). The mean pulmonary vascular resistance was 1.7 U/m 2 (range 0.3 to 2.6 U/m 2) , with mean PA pressures ranging from 6 mm Hg t022 mm Hg (average 12 mm Hg). Pulmonary-systemic blood flow ratio was available in 22 patients, with a mean value of 1.6 (range 0.37 to 3.0). The mean preoperative hemoglobin concentration was 17.4 mg/dl (range 13.2 to 21.6 mg/dl), with a mean arterial oxygen saturation of 82% (range 67% to 94%).
Surgical procedure Details of the operative technique have been reported previously." The operation was performed through a median sternotomy incision. Direct bicaval cannulation was accomplished with proximal cannulation of the SVC at its confluence with the innominate vein. Previously placed systemic-pulmonary shunts were snared or ligated as cardiopulmonary bypass was initiated. Core cooling to 22° C was followed by aortic crossclamping and antegrade infusion of cold blood cardioplegic solution. The two patients with pulmonary atresia and intact ventricular septum and RV-coronary connections also had delivery of cardioplegic solution via the coronary sinus and the
The Journal of Thoracic and Cardiovascular Surgery
2 8 2 Stein et al.
Fig. 1. Composite lateral tunnel. The tunnel is of uniform diameter. The coronary sinus ostium remains outside the tunnel.
Fig. 2. Completed total cavopulmonary connection.
R V. After cardiac arrest an oblique incision was made in the RA anterior to the crista terminalis. After excision of the medial portion of the atrial septum, a composite tunnel was created by inserting a polytetrafluoroethylene (PTFE)* patch around the orifices of the SVC and the IVC, to the remnant of the atrial septum and to the lateral atrial wall to the levelof the crista terminalis (Fig. 1). The tunnel was fashioned to create a conduit of uniform diameter between the orifices of the SVC and the IVe. The main PA was transected and oversewn proximally or suture ligated. Areas of stenosis in the branch PAs were augmented with pericardium. The SVC was divided proximal to the cavoatrial junction, and an end-to-side anastomosis was performed between the proximal SVC and the superior aspect of the right P A. The distal SVC orifice was anastomosed to the inferior aspect of the right P A. As rewarming was begun the right atriotomy was closed (Fig. 2). Warm blood cardioplegic solution with glutamate and aspartate was used in all patients before release of the aortic crossclamp. Concomitant procedures were required in 13 patients. Nine patients with severe atrioventricular valve regurgitation required atrioventricular valve annuloplasty. Six patients were found to have significant subaortic stenosis (including 2 with both subaortic stenosis and atrioventricular valve regurgitation); 2 had resection of subaortic membranes, and 4 underwent Damus-Stansel-Kaye procedures.
Two patients with elevated PA pressures (means of 22 and 20) preoperatively had adjustable atrial septal defects placed in the PTFE tunnel at the time of operation (Fig. 3). A defect 4 to 6 mm in diameter was left between the posterior edge of the PTFE patch and the interatrial septum. A 1-0 polypropylene suture was placed across the defect, passed through the interatrial septum, and snared with a tube prepared from an 8F endotracheal suction catheter. After cardiopulmonary bypass, the size of the defect was adjusted to maintain RA pressures of 14 to 16 mm Hg and saturations of ± 85%. The end of the snare was fixed in position with medium hemoclips and left beneath the linea alba. In the intensive care unit, opening the subxiphoid portion of the incision under local anesthesia allowed exposure of the snare, which could then be adjusted or progressivelyclosed with additional hemoclips. The mean aortic crossclamp time for all procedures was 88 minutes (range 44 to 165 minutes).
*Gore-Tex patch, registered trade mark of W. L. Gore & Associates, Inc., Elkton, Md.
Results
There were no intraoperative deaths. One early death (2.6%) occurred 2 days after operation in an otherwise hemodynamically stable 3-year-old child in whom recurrent episodes of ventricular tachycardia developed, resistant to pharmacologic management. Death was due to intractable ventricular fibrillation. Autopsy revealed no anatomic cause for the arrhythmias. Cardiomegaly was present, but there were no intracardiac thrombi, and all
Volume 102
Functional single ventricle 2 8 3
Number 2 August 1991
anastomoses were patent. Microscopically there was no evidence of myocardial or endocardialfibrosis. Two late deaths occurred. Viral pneumonia developed 4 months postoperatively in one patient.This patient had a respiratoryarrest withresulting anoxicbrain injury,and shedied subsequently of noncardiaccauses. The other late death occurred in a 2-year-old patient whose initialhospitalization was complicated by serous pericardial tamponade necessitating reoperation. She was discharged in stable condition but required readmission to another hospital 2 weeks later with respiratory distress from a chylous left pleural effusion. Tube thoracostomy was performed and her condition improved, although chest tube drainage remained high. Subsequently fever and symptoms of a viral syndrome developed. Hypotension resulted from reaccumulation of the pericardial effusion, and she died after creation of a pericardial window 40 days after the Fontan procedure. All patients were treated with inotropic agents in the immediate postoperative period. Duration of support ranged from 16to 313hours,with a mean duration of 64 hours. In addition, one patient required a venous assist device on the second postoperative day but was successfully weaned fromthe device within24 hours.No patients required intraaortic balloon or ventricular assist devices. Hemodynamic measurements were obtained 24 hours after operation. The averageRA pressurewas 13mm Hg (range 11 to 17),and the averageleft atrial pressurewas 7 mm Hg (range 3 to 12). The mean room air oxygen saturation was 96% (range 92%to 98%). Three patients required prolonged ventilatory support (>100 hours). The average duration of intubation for the remaining 35 patients was 19 hours (range 6 to 71 hours). Pleural effusions developed in seven patients, one of whom required pleurodesis. Cardiac tamponade caused by serous effusions occurredin two patients;one patient, as mentioned previously, required reoperation 24 hours postoperatively; the other patient had lowcardiac output 2 weeks postoperatively and required open drainage of a loculated pericardial effusion. A third patient whose chest x-ray film showed an enlarging cardiac silhouette was found to have a chylous pericardial effusion 3 months after operation. This necessitated ligationof the thoracic duct and creationof a pericardial window. Three patients required permanent pacemaker insertion. One patient with intermittent preoperative atrioventricular block underwent pacemakerplacementat the time of cavopulmonary connection because he remained in junctional rhythm after cardiopulmonary bypass. Two patients underwent pacemaker placement postoperatively; one being a patient who had sick sinus syndrome with intermittent junctional bradycardia and the other having
Fig. 3. Adjustable atrial septal defect. A pledget-supported 1was placed across the defect, passed through the interatrial septum, and snared. After cardiopulmonary bypass the size of the defect was adjusted by placing hemoclips distal to the snare.
o polypropylene suture
third-degree atrioventricular block. The averagelengthof hospitalization was 12 days (range 7 to 31 days). Follow-up wasavailablein all 35surviving patients.At a mean follow-up of 11 months, 30 patients are in New York Heart Association functional class I and five are in class II. Thirty-three patients were in normal sinus rhythm,one withintermittent supraventriculartachycardia, one with first-degree atrioventricularblock,and one withintermittentjunctionalrhythm. Of the three patients who underwent placement of a permanent pacemaker, one is in normal sinus rhythm and two are pacemaker dependent. Discussion Since Fontan's original description of RV bypass for tricuspid atresia, several modifications have been described. Kreutzer and associates,' soon after Fontan, described an RA-PA connection using the disconnected main PA and pulmonaryvalve. The IVC valvewas eliminated in this repair. Severalmethods have subsequently been used for valveless RA-PA connections.v 9 It is now
2 8 4 Stein et al.
Fig. 4. Preoperative RV angiogram in patient with pulmonary atresia and intact ventricular septum. Extensive R V-coronary connections are demonstrated.
accepted that valves in these locations are not necessary. For patients with tricuspid atresia and ventricular septal defect, an RA-RV connection may be made with'? or without II a valve. We now believe that a homograft valve is necessary only in those unusual patients with a relatively large RV cavity that may be further enlarged by resection of trabeculae. For the remainder, either a nonvalved RA-RV connection may be used, or an RA-PA connection may be performed. In tricuspid atresia, particularly if there is transposition of the great arteries, a direct RA-PA connection using the RA appendage may be performed, particularly if the RA is of normal size. If the RA is dilated, a total cavopulmonary connection is performed to avoid exposing the entire RA to an elevated pressure. In other complex congenital anomalies with functionally univentricular hearts and adequate RA size, we prefer total cavopulmonary connection for Fontan type repair. There are several technical advantages to total cavopulmonary connection when compared with other Fontan modifications. The construction of a lateral tunnel avoids suture placement near the atrioventricular conduction pathways, decreasing the likelihood of postoperative heart block. The posterior location of the PA anastomoses avoids the risk of sternal compression associated with other types of atriopulmonary anastomoses. 12. 13 Pulmonary venous obstruction by the intraatrial baffle, which
The Journal of Thoracic and Cardiovascular Surgery
can occur with mitral atresia, is also avoided by the lateral tunnel. The adjustable atrial septal defect, I 4 which provides a controlled right-to-Ieft shunt to reduce the RA pressure, can also be used with the laterally placed PTFE patch, as was described in the Surgicalprocedure section. This was useful in the management of two patients with elevated PA pressures postoperatively, allowing a controlled rightto-left shunt and improving cardiac output early after operation. When the PA pressures fell during the first few days after operation, the defects in both patients were closed under local anesthesia without the need for thoracotomy. Demonstration of the theoretical hemodynamic advantages of cavopulmonary connection is extremely difficult because there is great variability in preoperative ventricular function and pulmonary vascular resistance between patients. Nevertheless, our patients exhibited a mean RA pressure of only 13 mm Hg compared with pressures ranging between 16 and 18 mm Hg reported after other Fontan modifications. 5. 15-17 This may, however, be related to the excellent preoperative hemodynamic values that this group of patients showed. One of the proposed advantages of the total cavopulmonary connection, demonstrated in de Leval's experimental work, has been the avoidance of turbulence that could cause pressure gradients between the venae cavae and the PA. Pressure measurements taken in our patients at the conclusion of the procedure showed no gradients between the .SVC, the IVC, and the PA. A concern that has also been expressed is that the SVC orifice might not be large enough to conduct the IVC flow. This was not borne out in this experience, although the connection was not augmented by a patch in any of the patients. Additionally, none of our patients has had ascites or evidence of hepatic dysfunction after operation. Significant pleural effusion necessitating intervention occurred in 18.4% of our patients after operation. Although effusion is prevalent, it is markedly less prevalent than the 47% figure we 18 previously reported in 45 patients undergoing modified Fontan operations, again suggesting that the lower venous pressures achieved after total cavopulmonary connection may be advantageous. Atrial arrhythmias are common late after Fontan operations, reported to occur in more than 35% of patients. 19. 20 These arrhythmias may result from the atriotomy and its closure, the suture line of the intraatrial baffle, as well as from the elevated pressures to which the RA is exposed. Creation of a lateral tunnel may reduce the prevalence oflate atrial arrhythmias by decreasing the portion of RA myocardium exposed to relatively high pressures. Our patients had a 3% incidence of late atrial arrhythmias, suggesting the efficacy of cavopulmonary connection in their prevention. Only one patient in our
Volume 102 Number 2 August 1991
series had atrioventricular blockrequiring placement of a permanent pacemaker. Importantly, this patient had concomitant atrioventricular valve annuloplasty, which waslikely to havecausedthiscomplication. There isconcernthat division of the distal SVC closeto the RAjunction could damage the sinoatrial node, resulting in sick sinus syndrome. One of our patients required a pacemaker because of persistent bradycardia with junctional rhythm. Further long-term follow-up of the rhythm in these patients is imperative in establishing the safety of this procedure. The two patients with pulmonary atresia and intact ventricular septumare of particularinterest. Preoperative catheterization in both revealed hypoplastic hypertensive RV chambers withextensive connections to the coronary circulation (Fig. 4). Previous reports have documented the deleterious effects of RV decompression on the myocardial blood supply." The approach taken was to leave the undecompressed RV as a source of coronary blood flow but to substitute oxygenated blood for the desaturated blood previously enteringthe RV. The lateral tunneldiverted IVC blood to the PAs,and the coronarysinus wasunroofed to position the coronaryvenous return at a distance fromthe tricuspid valve. Oxygenated pulmonary venous blood reached the tricuspid valve via the atrial septal defect. Both patients have done extremely well after operation, with noevidence of myocardial ischemia. Although long-term follow-up is necessary to determine the benefits of this approach, it appearsto be a promising technique for the management of this particularly difficult group of patients. Conclusions
Total cavopulmonary connection is an effective modification of the Fontan procedure that can be performed with relative technical ease. It can be used to treat a number of congenital heart defects characterized functionally by a single ventricle, and it is of particular use in patients with complex congenital anomalies, including pulmonary atresia and intact ventricular septum with RV-coronary connections. The operation can be performed withlowmorbidityand mortalityand may result in improved hemodynamics when compared with other Fontan-type procedures. The lateral tunnel may prevent RA distention and decrease late arrhythmias; however, close follow-up is necessary to establish the long-term safetyof this procedure. REFERENCES 1. Fontan F, Baudet E. Surgical repair of tricuspid atresia. Thorax 1971;26:240-8. 2. Kreutzer G, Galindez E, Bono H, De Palma C, Laura JP.
Functional single ventricle 2 8 5
An operation for the correction of tricuspid atresia. J THORAC CARDIOVASC SURG 1973;66:105-11. 3. Shemin RJ, Merrill WH, Pfeifer JS, Conkle DM, Morrow AG. Evaluation of right atrial-pulmonary conduits for tricuspid atresia. J THORAC CARDIOVASC SURG 1979;77:68590. 4. Gale AW, Danielson GK, McGoon DC, Mair DD. Modified Fontan operation for univentricular heart and complicated lesions. J THORAC CARDIOVASC SURG 1979;78: 831-8. 5. Puga FJ, Chiavarelli M, Hagler DJ. Modifications of the Fontan operation applicable to patients with left atrioventricular valve atresia or single atrioventricular valve. Circulation 1987;76(Pt 2):IIl53-60. 6. de Leval MR, Kilner P, Gewillig M, Bull C. Total cavapulmonary connection: a logical alternative to atriopulmonary connection for complex Fontan operations. J THORAC CARDIOVASC SURG 1988;96:682-95. 7. Haas GS, Laks H, Pearl JM. Modified Fontan procedure. Adv Cardiac Surg 1990;1:111-54. 8. Doty DB, Marvin W J Jr, Lauer RM. Modified Fontan procedure: methods to achieve direct anastomosis of the right atrium to the pulmonary artery. J THORAC CARDIO. VASC SURG 1981;81:470-7. 9. Michler RE, Rose EA, Maim JR. Tricuspid atresia. In: Arciniegas E, ed. Pediatric cardiac surgery. St. Louis: Mosby-Year Book, 1985:297. 10. Bowman FO, MaIm JR, Hayes CJ, et al. Physiologic approach to surgery for tricuspid atresia. Circulation 1978; 58(Pt 2):83. II. Bjork VO, Olin CL, Bjarke BB, Thoren CA. Right atrialright ventricular anastomosis for correction of tricuspid atresia. J THORAC CARDIOVASC SURG 1979;77:452-8. 12. Girod DA, Fontan F, Deville C, Ottenkamp J, Choussat A. Long-term results after the Fontan operation for tricuspid atresia. Circulation 1987;75:605-10. 13. Fernandez G, Costa F, Fontan F, Naftel DC, Blackstone EH, Kirklin JW. Prevalence of reoperation for pathway obstruction after Fontan operation. Ann Thorac Surg 1989;48:654-9. 14. Laks H, Haas GS, Pearl JM, et al. The use of an adjustable intraatrial communication in patients undergoing the Fontan and definitive right heart procedures [Abstract]. Circulation I988;78(Pt 2):11357. 15. Nakazawa M, Nakanishi T, Hirohumi 0, et al. Dynamics of right heart flow in patients after Fontan procedure. Circulation 1984;69:306-12. 16. Humes RA, Porter CJ, Mair DD, et al. Intermediate follOW-Up and predicted survival after the modified Fontan procedure for tricuspid atresia and double-inlet ventricle. Circulation 1987;76(Pt 2):III67-71. 17. Abe T, Komatsu S, Sugiki K. Modified Fontan operation for complex cardiac anomalies-postoperative hemodynamics, cardiac function, and clinical status. Jpn Circ J 1988;52:1221-30. 18. Laks H, Milliken JC, Perloff JK, et al. Experience with the Fontan procedure. J THORAC CARDIOVASC SURG 1984;88: 939-51.
2 8 6 Stein et al.
19. Chen S, Nouri S, Pennington DG. Dysrhythmias after the modified Fontan procedure. Pediatr Cardiol 1988;9:215-9. 20. Weber HS, Hellenbrand WE, Kleinman CS, Perlmutter RA, Rosenfeld LE. Predictors of rhythm disturbances and subsequent morbidity after the Fontan operation. Am J Cardiol 1989;64:762-7. 21. Coles JG, Freedom RM, Lightfoot NE, et al. Long-term results in neonates with pulmonary atresia and intact ventricular septum. Ann Thorac Surg 1989;47:213-7.
Discussion Dr. Winfield J. Wells (Los Angeles. Calif). I notice this is a consecutive series beginning in January 1987. Does this represent all of the patients undergoing total cavopulmonary connection at UCLA? Dr. Stein. It does. We first started doing them in 1987. Dr. Wells. How do you select patients for total cavopulmonary connection as opposed to other modified Fontan connections? In the same time frame as the series, how many other Fontan procedures were done, what other connections were used, and why? Dr. Stein. After the initial description, we first started using the Fontan procedure with total cavopulmonary anastomosis in 1987. In that year 7 patients underwent a lateral tunnel and 26 underwent other modified Fontan procedures. Through the years we have found total cavopulmonary connection to be effective and technically easier to perform. In 1989 22 (71 %) of our patients underwent the Fontan procedure with total cavopulmonary anastomosis. Nine underwent other modifications. Now we prefer this operation to other Fontan applications because of the potential benefits of the improved hemodynamics, possible decreased arrhythmias, late arrhythmias (which time will tell), and technical ease. Some patients with tricuspid atresia and large RVs may have other connections such as RA"RV connections. The other situation in which we do not use it is one in which the RA is small and will not accommodate a baffle. Dr. Wells. You make a point of the technical ease of this operation, yet, even recognizing that a third of the patients had significant concomitant procedures, the average crossclamp time was 88 minutes. This does not seem compatible with "technical ease." The results are outstanding, which implies to me that the UCLA group knows how to take care of the univentricle during the ischemic period. You mention that you use antegrade infusion of cold blood cardioplegic solution. Would you give us more detail on the exact technique? Dr. Hillel Laks (Los Angeles. Calif). The myocardial protection is the standard one that we use for either congenital or adult procedures. We use infusion of cold blood cardioplegic solution every 15 to 20 minutes. We have recently begun to use retrograde infusion in addition to that, and we do use glutamate and aspartate warm blood cardioplegic solution at the end of the procedure before reperfusion. We also use topical hypothermia. We agree with you that myocardial protection is critically important in an operation the success of which depends on low end-diastolic pressure at the end of the procedure. Dr. Wells. Do you measure the infusion pressure of the cardioplegic solution? Dr. Laks, Yes. We monitor it always and vary the pressure depending on the age of the child. In younger children we use
The Journal of Thoracic and Cardiovascular Surgery
lower pressures of 80 mm Hg until cardiac arrest and then 50 to 60 mm Hg for subsequent doses. In the older children, 3 to 4 years of age, we generally use a pressure around 100 mm Hg until cardiac arrest occurs and then drop it down to 70 to 80 mm Hg for subsequent doses. Dr. Wells. You point out several potential technical advantages to this procedure. An additional point that we believemay be important is the placement of the coronary sinus in the lower pressure atrial chamber. We believe this may be a physiologic advantage to these patients. Dr. Laks. We do keep the coronary sinus outside the tunnel, on the left atrial side of the tunnel, so that it is in the low-pressure atrium. Dr. Vaughn A. Starnes (Stanford. Calif). I would like to make one comment and then ask you how you are managing it. When I was at the Hospital for Sick Children and de Leval was doing the Fontan procedures, we had one child with thrombosis of the lateral pulmonary connection; we were using PTFE at that time. I am now using autologous pericardium or even bovine pericardium. Do you use any anticoagulants on these children-aspirin, dipyridamole (Persantine), or even warfarin sodium (Coumadin), in the follow-up period? Dr. Stein. We do not use warfarin sodium. Some patients are placed on a regimen of dipyridamole or aspirin postoperatively, but there is no warfarin sodium used. Dr. D. Craig Miller (Stanford. Calif). I am particularly intrigued by the six cases you had of pulmonary atresia and intact septum associated with a "peach pit" RV and RV-coronary sinusoids. Have you not also told us something equally important here in terms of these particular patients' myocardial metabolism, that is, getting oxygenated blood flow retrograde into the coronary arteries? Do you have any autopsy information for the two deaths concerning myointimal proliferation in the coronary arteries, if the patients who died had pulmonary atresia and intact ventricular septum? Dr. Stein. Regarding pulmonary atresia and intact ventricular septum with the coronary connections, there were two patients with that diagnosis. We know that previous reports of decompression of the R V with these connections have resulted in high mortality postoperatively. This technique allows the pulmonary venous return to enter the RV in an unobstructed manner and then provides well-oxygenated blood to the coronary arteries. We also unroofed the coronary sinus at a distance from the RV to help prevent desaturated blood from entering the RV. We hope that this modification will be beneficial long term to these patients. With two patients and a short series, however, we cannot reach any conclusions. There were two late deaths in patients with other diagnoses. One occurred in a patient who had respiratory arrest from viral pneumonia and an anoxic brain injury. Life support was eventually withdrawn because of brain death. The other death occurred in a patient who had a pleural effusion. Dr. Miller. I know how they died, but do you have any histopathologic information if they were patients with pulmonary atresia and intact ventricular septum? Looking at their coronary arteries histologically would be fascinating. Dr. Stein. The two patients with pulmonary atresia and intact ventricular septum did not die. They are still alive. Dr. Laks. I want to comment on two issues. One is when we would use the lateral tunnel procedure as opposed to some other form of connection. As Dr. Stein mentioned, one situation would be tricuspid atresia in which the RA is not large. In that situation there seems no reason to do an atrial septectomy and
Volume 102 Number 2 August 1991
put in a baffle. We simply unroof the coronary sinus into the left atrium, close the coronary sinus orifice, and then do the two caval connections as in the lateral tunnel procedures, using the small RA as part of the tunnel. The other situation would be total anomalous pulmonary venous return to the SVC. Obviously in that situation one has to use a different approach and cannot use a lateral tunnel procedure. Then there are patients with tricuspid atresia or other conditions with transposition, in whom the RA or atrial appendage is lying on the P A. In these patients we have used different types of baffles. After one patient in whom we partially decompressed the RV, we hoped that keeping RV pressures around two-thirds systemic would help supply the coronary circulation in patients with pulmonary atresia, intact septum, and RV--coronary artery connections. The patient who had normal ventricular function preoperatively died of low output. We thought that if the
Functional single ventricle 2 8 7
patients have normal function before the operation with the RV supplying desaturated blood to the coronary arteries, surely they would be better off if we could supply oxygenated blood and leave the RV as the source of coronary blood supply. We have now actually managed three patients that way; in one we used a nonlateral tunnel approach, and in the other two the lateral tunnel approach. Inall of them we allowed oxygenated blood to reach the tricuspid valve without decompression of the RV. All three patients have been restudied with follow-up to about 2 years. We have done thallium scans, positron-emission tomography scans, left ventricular multigated image acquisition analyses, and ejection fractions. The ventricular function is normal, there is no evidence of ischemia, and Holter monitor revealed no ventricular arrhythmias. This may be, at least for awhile, an alternative to transplantation. We need experience with more patients and we need to monitor them very closely because sudden death is still a concern.
THORAC CARDIOVASC SURG
1991;102:280-7
Results of total cavopulmonary connection in the treatment of patients with a functional single ventricle Total cavopulmonary connection was proposed as a modification of the Fontan procedure that might have greater benefits than previous methods. To assess tbis procedure we reviewed case bistories of 38 patients (aged 17 months to 30 years) who underwent Fontan procedures with cavopulmonary anastomoses between January 1987 and December 1989. The group included 32 patients with univentricular heart, 2 with pulmonary atresia and intact ventricular septum, 3 with tricuspid atresia, and 1 with hypoplastic left heart syndrome. One or more previous palliative procedures were performed in 34 patients, including 19 systemic-pulmonary shunts, 16 pulmonary artery bandings, 7 atrial septectomiesjseptostomies, 7 Glenn shunts, and 1 patent ductus arteriosus ligation. Preoperative hemodynamics showed a pulmonary artery pressure of 12 mm Hg (range 6 to 22 mm Hg), pulmonary-systemic flow ratio of 1.6 (range 0.37 to 3.0), left ventricular end-diastolic pressure 9 mm Hg (range 3 to 15 mm Hg), and systemic arterial oxygen saturation of 82 % (range 67 % to 94 % ~ Concomitant with cavopulmonary connection, 13 patients underwent additional procedures, including 9 atrioventricular valve annuloplasties, 4 Damus-Stansel-Kaye procedures, and 2 resections of subaortic membranes. Modifying the Fontan procedure in this fashion was particularly useful in the management of 2 patients with pulmonary atresia and intact ventricular septum who had right ventricular-dependent coronary blood flow. Cavopulmonary anastomosis and atrial septectomy were performed in both patients, with resultant inflow of oxygenated blood to the right ventricle and coronary arteries. ExceUent postoperative results were noted in each. Postextubation hemodynamics for the entire group included a mean right atrial pressure of 13 mm Hg (range 11 to 17 mm Hg), a mean left atrial pressure of 6 mm Hg (range 3 to 12 mm Hg), and a room air oxygen saturation of 96% (range 92% to 98 %). Seven patients had pleural effusions, 3 required postoperative pacemaker placement, and 2 required reoperation for tamponade. A venous assist device was required in one patient on the second postoperative day, but the patient was weaned successfuUy within 24 hours. One early death (2.6%) occurred in a patient who had intractable ventricular fibriUation 2 days after operation. There was one late cardiac death (2.7%) caused by ventricular failure and one late noncardiac death. These results demonstrate that total cavopulmonary connection providesexceUent early definitive treatment, with low morbidity and mortality, for a variety of complex congenital heart lesions.
Darryl G. Stein, MD,a Hillel Laks, MD,a Davis C. Drinkwater, MD,a Lester C. Permut, MD,a Henry W. Louie, MD,a Jeffrey M. Pearl, MD,a Barbra L. George, MD,b and Roberta G. Williams, MD,b Los Angeles, Calif.
From the Division of Cardiothoracic Surgery' and the Division of Pediatric Cardiology," UCLA School of Medicine, Los Angeles, Calif. Read at the Sixteenth Annual Meeting of The Western Thoracic Surgical Association, Coronado, Calif., June 20-23, 1990. Address for reprints: Hillel Laks, MD, Division of Cardiothoracic Surgery, University of California, Los Angeles, UCLA Medical Center, 62-182 CHS, 10833 Le Conte Ave., Los Angeles, CA 90024.
12/6/27993
280
In 1971 Fontan and Bauder' reportedthe firstclinically successful procedure for total bypass of the right sideof the heart fortreatment oftricuspid atresia.The procedure included construction of a Glenn shunt, closure of the atrial septal defect, homograftvalve insertion at the orifice ofthe inferiorvenacava (IVC), placement ofa valved homograftfrom the right atrium (RA) to the left pulmonary artery (PA), and ligation of the main PA. Sincethe
Volume 102 Number 2
Functional single ventricle
August 1991
original description, a number of modifications to Fontan's operation have been described that have simplified the operation and allowed its application to a variety of complex congenital heart lesions. Kreutzer and coworkers.l in 1973, described a method for right ventricular (RV) bypass by construction of an end-to-end anastomosis between the RA appendage and the detached pulmonary anulus without a concomitant Glenn shunt. The success of this technique demonstrated that an RA inlet valve was dispensable and that the entire systemic venous return could be directed to the PA through a single connection. Shemin and colleagues' examined the necessity for an outlet (RA-PA) valve in dogs. They found that blood flow through RA-PA conduits was influenced primarily by RA pressure and was not affected by the presence or absence of a valve. Shortly thereafter Gale and associates" reported a 17% prevalence of valved conduit-related complications and presented favorable results using a direct valveless RA-PA connection. Puga, Chiavarelli, and Hagler-Iater described total cavopulmonary connection as a valveless repair that avoided a direct atriopulmonary connection. In their technique they used a lateral intraatrial baffle to divert Ive blood to the superior vena cava (SVC). This was followed by transection of the SVC at the levelof the right PA and end-to-side anastomoses of each transected end to the P A. To determine the contribution of RA contraction in a valveless atriopulmonary connection, de Leval and coworkers" performed a series of hydrodynamic studies with plastic models of the human RA. These investigations showed that pulsation in a valvelesschamber generated turbulence, with resultant elevation of pressure within the chamber and diminution of forward flow. Additionally, turbulence generated by flow through cavities, around corners, and through stenoses further inhibited forward flow. They concluded that maximal forward output could be achieved by streamlining blood flow through a valvelessRA tunnel of uniform caliber. On the basis of these experimental observations, de Leval and coworkers" described a modified Fontan operation, very similar to the one described by Puga and associates.P The procedure consists of an end-to-side anastomosis of the SVC to the superior aspect of the undivided right PA, construction of a lateral intraatrial tunnel with the use of a patch and the lateral wall of the RA to connect the IVC to the SVC, and anastomosis of the cardiac end of the transected SVC, augmented with pericardium, to the inferior aspect of the main P A. The potential benefits include (1) technical ease of repair, (2) reduced atrial turbulence with resulting improved hemodynamics, and (3) decreased arrhythmias.
281
To assess the possible benefits of such an approach we reviewed the case histories of 38 consecutive patients who underwent the Fontan procedure with total cavopulmonary anastomosis between January 1987 and December 1989. This group made up an increasing percentage of all patients undergoing Fontan procedures at our institution during this period, constituting 71% of all Fontan-type procedures we performed in 1989.
Patient profile Nineteen patients were male and 19 were female. Age at operation ranged from 17 months to 30 years, with a mean of 7 years. The largest group included 32 patients (84%) with univentricular heart. Fifteen of these had left ventricular anatomy, 16 had right ventricular chambers, and one had indeterminate anatomy. Three patients (8%) had tricuspid atresia, 2 (5%) had pulmonary atresia with intact ventricular septum and RV-coronary collaterals, and 1 (3%) had hypoplastic left heart syndrome. Thirty-four patients underwent one or more prior palliative procedures. Nineteen systemic-PA shunts were performed in 17 patients. Sixteen patients underwent PA banding, 7 had atrial septectomies/septostomies, 7 had modified Glenn shunts, 1 underwent ligation of the patent ductus arteriosus, and 1 had undergone pacemaker placement for intermittent complete heart block and bradycardia. Total cavopulmonary connection was performed in patients with cyanotic congenital heart disease with a functionally univentricular heart. The selection criteria were similar to those used for previous Fontan procedures and included acceptable ventricular function and pulmonary vascular resistance. An ejection fraction of greater than 45% and a ventricular end-diastolic pressure of less than 15 mm Hg were considered optimal. A pulmonary vascular resistance of greater than 2 U /m 2 or a PA pressure greater than 25 mm Hg was unacceptable. Ventricular function as assessed by echocardiography was considered good to excellent in all patients. Angiography demonstrated a mean ejection fraction of 58% (range 43% to 72%). Ventricular end-diastolic pressures ranged from 3 mm Hg to 15 mm Hg (mean 9 mm Hg). The mean pulmonary vascular resistance was 1.7 U/m 2 (range 0.3 to 2.6 U/m 2) , with mean PA pressures ranging from 6 mm Hg t022 mm Hg (average 12 mm Hg). Pulmonary-systemic blood flow ratio was available in 22 patients, with a mean value of 1.6 (range 0.37 to 3.0). The mean preoperative hemoglobin concentration was 17.4 mg/dl (range 13.2 to 21.6 mg/dl), with a mean arterial oxygen saturation of 82% (range 67% to 94%).
Surgical procedure Details of the operative technique have been reported previously." The operation was performed through a median sternotomy incision. Direct bicaval cannulation was accomplished with proximal cannulation of the SVC at its confluence with the innominate vein. Previously placed systemic-pulmonary shunts were snared or ligated as cardiopulmonary bypass was initiated. Core cooling to 22° C was followed by aortic crossclamping and antegrade infusion of cold blood cardioplegic solution. The two patients with pulmonary atresia and intact ventricular septum and RV-coronary connections also had delivery of cardioplegic solution via the coronary sinus and the
The Journal of Thoracic and Cardiovascular Surgery
2 8 2 Stein et al.
Fig. 1. Composite lateral tunnel. The tunnel is of uniform diameter. The coronary sinus ostium remains outside the tunnel.
Fig. 2. Completed total cavopulmonary connection.
R V. After cardiac arrest an oblique incision was made in the RA anterior to the crista terminalis. After excision of the medial portion of the atrial septum, a composite tunnel was created by inserting a polytetrafluoroethylene (PTFE)* patch around the orifices of the SVC and the IVC, to the remnant of the atrial septum and to the lateral atrial wall to the levelof the crista terminalis (Fig. 1). The tunnel was fashioned to create a conduit of uniform diameter between the orifices of the SVC and the IVe. The main PA was transected and oversewn proximally or suture ligated. Areas of stenosis in the branch PAs were augmented with pericardium. The SVC was divided proximal to the cavoatrial junction, and an end-to-side anastomosis was performed between the proximal SVC and the superior aspect of the right P A. The distal SVC orifice was anastomosed to the inferior aspect of the right P A. As rewarming was begun the right atriotomy was closed (Fig. 2). Warm blood cardioplegic solution with glutamate and aspartate was used in all patients before release of the aortic crossclamp. Concomitant procedures were required in 13 patients. Nine patients with severe atrioventricular valve regurgitation required atrioventricular valve annuloplasty. Six patients were found to have significant subaortic stenosis (including 2 with both subaortic stenosis and atrioventricular valve regurgitation); 2 had resection of subaortic membranes, and 4 underwent Damus-Stansel-Kaye procedures.
Two patients with elevated PA pressures (means of 22 and 20) preoperatively had adjustable atrial septal defects placed in the PTFE tunnel at the time of operation (Fig. 3). A defect 4 to 6 mm in diameter was left between the posterior edge of the PTFE patch and the interatrial septum. A 1-0 polypropylene suture was placed across the defect, passed through the interatrial septum, and snared with a tube prepared from an 8F endotracheal suction catheter. After cardiopulmonary bypass, the size of the defect was adjusted to maintain RA pressures of 14 to 16 mm Hg and saturations of ± 85%. The end of the snare was fixed in position with medium hemoclips and left beneath the linea alba. In the intensive care unit, opening the subxiphoid portion of the incision under local anesthesia allowed exposure of the snare, which could then be adjusted or progressivelyclosed with additional hemoclips. The mean aortic crossclamp time for all procedures was 88 minutes (range 44 to 165 minutes).
*Gore-Tex patch, registered trade mark of W. L. Gore & Associates, Inc., Elkton, Md.
Results
There were no intraoperative deaths. One early death (2.6%) occurred 2 days after operation in an otherwise hemodynamically stable 3-year-old child in whom recurrent episodes of ventricular tachycardia developed, resistant to pharmacologic management. Death was due to intractable ventricular fibrillation. Autopsy revealed no anatomic cause for the arrhythmias. Cardiomegaly was present, but there were no intracardiac thrombi, and all
Volume 102
Functional single ventricle 2 8 3
Number 2 August 1991
anastomoses were patent. Microscopically there was no evidence of myocardial or endocardialfibrosis. Two late deaths occurred. Viral pneumonia developed 4 months postoperatively in one patient.This patient had a respiratoryarrest withresulting anoxicbrain injury,and shedied subsequently of noncardiaccauses. The other late death occurred in a 2-year-old patient whose initialhospitalization was complicated by serous pericardial tamponade necessitating reoperation. She was discharged in stable condition but required readmission to another hospital 2 weeks later with respiratory distress from a chylous left pleural effusion. Tube thoracostomy was performed and her condition improved, although chest tube drainage remained high. Subsequently fever and symptoms of a viral syndrome developed. Hypotension resulted from reaccumulation of the pericardial effusion, and she died after creation of a pericardial window 40 days after the Fontan procedure. All patients were treated with inotropic agents in the immediate postoperative period. Duration of support ranged from 16to 313hours,with a mean duration of 64 hours. In addition, one patient required a venous assist device on the second postoperative day but was successfully weaned fromthe device within24 hours.No patients required intraaortic balloon or ventricular assist devices. Hemodynamic measurements were obtained 24 hours after operation. The averageRA pressurewas 13mm Hg (range 11 to 17),and the averageleft atrial pressurewas 7 mm Hg (range 3 to 12). The mean room air oxygen saturation was 96% (range 92%to 98%). Three patients required prolonged ventilatory support (>100 hours). The average duration of intubation for the remaining 35 patients was 19 hours (range 6 to 71 hours). Pleural effusions developed in seven patients, one of whom required pleurodesis. Cardiac tamponade caused by serous effusions occurredin two patients;one patient, as mentioned previously, required reoperation 24 hours postoperatively; the other patient had lowcardiac output 2 weeks postoperatively and required open drainage of a loculated pericardial effusion. A third patient whose chest x-ray film showed an enlarging cardiac silhouette was found to have a chylous pericardial effusion 3 months after operation. This necessitated ligationof the thoracic duct and creationof a pericardial window. Three patients required permanent pacemaker insertion. One patient with intermittent preoperative atrioventricular block underwent pacemakerplacementat the time of cavopulmonary connection because he remained in junctional rhythm after cardiopulmonary bypass. Two patients underwent pacemaker placement postoperatively; one being a patient who had sick sinus syndrome with intermittent junctional bradycardia and the other having
Fig. 3. Adjustable atrial septal defect. A pledget-supported 1was placed across the defect, passed through the interatrial septum, and snared. After cardiopulmonary bypass the size of the defect was adjusted by placing hemoclips distal to the snare.
o polypropylene suture
third-degree atrioventricular block. The averagelengthof hospitalization was 12 days (range 7 to 31 days). Follow-up wasavailablein all 35surviving patients.At a mean follow-up of 11 months, 30 patients are in New York Heart Association functional class I and five are in class II. Thirty-three patients were in normal sinus rhythm,one withintermittent supraventriculartachycardia, one with first-degree atrioventricularblock,and one withintermittentjunctionalrhythm. Of the three patients who underwent placement of a permanent pacemaker, one is in normal sinus rhythm and two are pacemaker dependent. Discussion Since Fontan's original description of RV bypass for tricuspid atresia, several modifications have been described. Kreutzer and associates,' soon after Fontan, described an RA-PA connection using the disconnected main PA and pulmonaryvalve. The IVC valvewas eliminated in this repair. Severalmethods have subsequently been used for valveless RA-PA connections.v 9 It is now
2 8 4 Stein et al.
Fig. 4. Preoperative RV angiogram in patient with pulmonary atresia and intact ventricular septum. Extensive R V-coronary connections are demonstrated.
accepted that valves in these locations are not necessary. For patients with tricuspid atresia and ventricular septal defect, an RA-RV connection may be made with'? or without II a valve. We now believe that a homograft valve is necessary only in those unusual patients with a relatively large RV cavity that may be further enlarged by resection of trabeculae. For the remainder, either a nonvalved RA-RV connection may be used, or an RA-PA connection may be performed. In tricuspid atresia, particularly if there is transposition of the great arteries, a direct RA-PA connection using the RA appendage may be performed, particularly if the RA is of normal size. If the RA is dilated, a total cavopulmonary connection is performed to avoid exposing the entire RA to an elevated pressure. In other complex congenital anomalies with functionally univentricular hearts and adequate RA size, we prefer total cavopulmonary connection for Fontan type repair. There are several technical advantages to total cavopulmonary connection when compared with other Fontan modifications. The construction of a lateral tunnel avoids suture placement near the atrioventricular conduction pathways, decreasing the likelihood of postoperative heart block. The posterior location of the PA anastomoses avoids the risk of sternal compression associated with other types of atriopulmonary anastomoses. 12. 13 Pulmonary venous obstruction by the intraatrial baffle, which
The Journal of Thoracic and Cardiovascular Surgery
can occur with mitral atresia, is also avoided by the lateral tunnel. The adjustable atrial septal defect, I 4 which provides a controlled right-to-Ieft shunt to reduce the RA pressure, can also be used with the laterally placed PTFE patch, as was described in the Surgicalprocedure section. This was useful in the management of two patients with elevated PA pressures postoperatively, allowing a controlled rightto-left shunt and improving cardiac output early after operation. When the PA pressures fell during the first few days after operation, the defects in both patients were closed under local anesthesia without the need for thoracotomy. Demonstration of the theoretical hemodynamic advantages of cavopulmonary connection is extremely difficult because there is great variability in preoperative ventricular function and pulmonary vascular resistance between patients. Nevertheless, our patients exhibited a mean RA pressure of only 13 mm Hg compared with pressures ranging between 16 and 18 mm Hg reported after other Fontan modifications. 5. 15-17 This may, however, be related to the excellent preoperative hemodynamic values that this group of patients showed. One of the proposed advantages of the total cavopulmonary connection, demonstrated in de Leval's experimental work, has been the avoidance of turbulence that could cause pressure gradients between the venae cavae and the PA. Pressure measurements taken in our patients at the conclusion of the procedure showed no gradients between the .SVC, the IVC, and the PA. A concern that has also been expressed is that the SVC orifice might not be large enough to conduct the IVC flow. This was not borne out in this experience, although the connection was not augmented by a patch in any of the patients. Additionally, none of our patients has had ascites or evidence of hepatic dysfunction after operation. Significant pleural effusion necessitating intervention occurred in 18.4% of our patients after operation. Although effusion is prevalent, it is markedly less prevalent than the 47% figure we 18 previously reported in 45 patients undergoing modified Fontan operations, again suggesting that the lower venous pressures achieved after total cavopulmonary connection may be advantageous. Atrial arrhythmias are common late after Fontan operations, reported to occur in more than 35% of patients. 19. 20 These arrhythmias may result from the atriotomy and its closure, the suture line of the intraatrial baffle, as well as from the elevated pressures to which the RA is exposed. Creation of a lateral tunnel may reduce the prevalence oflate atrial arrhythmias by decreasing the portion of RA myocardium exposed to relatively high pressures. Our patients had a 3% incidence of late atrial arrhythmias, suggesting the efficacy of cavopulmonary connection in their prevention. Only one patient in our
Volume 102 Number 2 August 1991
series had atrioventricular blockrequiring placement of a permanent pacemaker. Importantly, this patient had concomitant atrioventricular valve annuloplasty, which waslikely to havecausedthiscomplication. There isconcernthat division of the distal SVC closeto the RAjunction could damage the sinoatrial node, resulting in sick sinus syndrome. One of our patients required a pacemaker because of persistent bradycardia with junctional rhythm. Further long-term follow-up of the rhythm in these patients is imperative in establishing the safety of this procedure. The two patients with pulmonary atresia and intact ventricular septumare of particularinterest. Preoperative catheterization in both revealed hypoplastic hypertensive RV chambers withextensive connections to the coronary circulation (Fig. 4). Previous reports have documented the deleterious effects of RV decompression on the myocardial blood supply." The approach taken was to leave the undecompressed RV as a source of coronary blood flow but to substitute oxygenated blood for the desaturated blood previously enteringthe RV. The lateral tunneldiverted IVC blood to the PAs,and the coronarysinus wasunroofed to position the coronaryvenous return at a distance fromthe tricuspid valve. Oxygenated pulmonary venous blood reached the tricuspid valve via the atrial septal defect. Both patients have done extremely well after operation, with noevidence of myocardial ischemia. Although long-term follow-up is necessary to determine the benefits of this approach, it appearsto be a promising technique for the management of this particularly difficult group of patients. Conclusions
Total cavopulmonary connection is an effective modification of the Fontan procedure that can be performed with relative technical ease. It can be used to treat a number of congenital heart defects characterized functionally by a single ventricle, and it is of particular use in patients with complex congenital anomalies, including pulmonary atresia and intact ventricular septum with RV-coronary connections. The operation can be performed withlowmorbidityand mortalityand may result in improved hemodynamics when compared with other Fontan-type procedures. The lateral tunnel may prevent RA distention and decrease late arrhythmias; however, close follow-up is necessary to establish the long-term safetyof this procedure. REFERENCES 1. Fontan F, Baudet E. Surgical repair of tricuspid atresia. Thorax 1971;26:240-8. 2. Kreutzer G, Galindez E, Bono H, De Palma C, Laura JP.
Functional single ventricle 2 8 5
An operation for the correction of tricuspid atresia. J THORAC CARDIOVASC SURG 1973;66:105-11. 3. Shemin RJ, Merrill WH, Pfeifer JS, Conkle DM, Morrow AG. Evaluation of right atrial-pulmonary conduits for tricuspid atresia. J THORAC CARDIOVASC SURG 1979;77:68590. 4. Gale AW, Danielson GK, McGoon DC, Mair DD. Modified Fontan operation for univentricular heart and complicated lesions. J THORAC CARDIOVASC SURG 1979;78: 831-8. 5. Puga FJ, Chiavarelli M, Hagler DJ. Modifications of the Fontan operation applicable to patients with left atrioventricular valve atresia or single atrioventricular valve. Circulation 1987;76(Pt 2):IIl53-60. 6. de Leval MR, Kilner P, Gewillig M, Bull C. Total cavapulmonary connection: a logical alternative to atriopulmonary connection for complex Fontan operations. J THORAC CARDIOVASC SURG 1988;96:682-95. 7. Haas GS, Laks H, Pearl JM. Modified Fontan procedure. Adv Cardiac Surg 1990;1:111-54. 8. Doty DB, Marvin W J Jr, Lauer RM. Modified Fontan procedure: methods to achieve direct anastomosis of the right atrium to the pulmonary artery. J THORAC CARDIO. VASC SURG 1981;81:470-7. 9. Michler RE, Rose EA, Maim JR. Tricuspid atresia. In: Arciniegas E, ed. Pediatric cardiac surgery. St. Louis: Mosby-Year Book, 1985:297. 10. Bowman FO, MaIm JR, Hayes CJ, et al. Physiologic approach to surgery for tricuspid atresia. Circulation 1978; 58(Pt 2):83. II. Bjork VO, Olin CL, Bjarke BB, Thoren CA. Right atrialright ventricular anastomosis for correction of tricuspid atresia. J THORAC CARDIOVASC SURG 1979;77:452-8. 12. Girod DA, Fontan F, Deville C, Ottenkamp J, Choussat A. Long-term results after the Fontan operation for tricuspid atresia. Circulation 1987;75:605-10. 13. Fernandez G, Costa F, Fontan F, Naftel DC, Blackstone EH, Kirklin JW. Prevalence of reoperation for pathway obstruction after Fontan operation. Ann Thorac Surg 1989;48:654-9. 14. Laks H, Haas GS, Pearl JM, et al. The use of an adjustable intraatrial communication in patients undergoing the Fontan and definitive right heart procedures [Abstract]. Circulation I988;78(Pt 2):11357. 15. Nakazawa M, Nakanishi T, Hirohumi 0, et al. Dynamics of right heart flow in patients after Fontan procedure. Circulation 1984;69:306-12. 16. Humes RA, Porter CJ, Mair DD, et al. Intermediate follOW-Up and predicted survival after the modified Fontan procedure for tricuspid atresia and double-inlet ventricle. Circulation 1987;76(Pt 2):III67-71. 17. Abe T, Komatsu S, Sugiki K. Modified Fontan operation for complex cardiac anomalies-postoperative hemodynamics, cardiac function, and clinical status. Jpn Circ J 1988;52:1221-30. 18. Laks H, Milliken JC, Perloff JK, et al. Experience with the Fontan procedure. J THORAC CARDIOVASC SURG 1984;88: 939-51.
2 8 6 Stein et al.
19. Chen S, Nouri S, Pennington DG. Dysrhythmias after the modified Fontan procedure. Pediatr Cardiol 1988;9:215-9. 20. Weber HS, Hellenbrand WE, Kleinman CS, Perlmutter RA, Rosenfeld LE. Predictors of rhythm disturbances and subsequent morbidity after the Fontan operation. Am J Cardiol 1989;64:762-7. 21. Coles JG, Freedom RM, Lightfoot NE, et al. Long-term results in neonates with pulmonary atresia and intact ventricular septum. Ann Thorac Surg 1989;47:213-7.
Discussion Dr. Winfield J. Wells (Los Angeles. Calif). I notice this is a consecutive series beginning in January 1987. Does this represent all of the patients undergoing total cavopulmonary connection at UCLA? Dr. Stein. It does. We first started doing them in 1987. Dr. Wells. How do you select patients for total cavopulmonary connection as opposed to other modified Fontan connections? In the same time frame as the series, how many other Fontan procedures were done, what other connections were used, and why? Dr. Stein. After the initial description, we first started using the Fontan procedure with total cavopulmonary anastomosis in 1987. In that year 7 patients underwent a lateral tunnel and 26 underwent other modified Fontan procedures. Through the years we have found total cavopulmonary connection to be effective and technically easier to perform. In 1989 22 (71 %) of our patients underwent the Fontan procedure with total cavopulmonary anastomosis. Nine underwent other modifications. Now we prefer this operation to other Fontan applications because of the potential benefits of the improved hemodynamics, possible decreased arrhythmias, late arrhythmias (which time will tell), and technical ease. Some patients with tricuspid atresia and large RVs may have other connections such as RA"RV connections. The other situation in which we do not use it is one in which the RA is small and will not accommodate a baffle. Dr. Wells. You make a point of the technical ease of this operation, yet, even recognizing that a third of the patients had significant concomitant procedures, the average crossclamp time was 88 minutes. This does not seem compatible with "technical ease." The results are outstanding, which implies to me that the UCLA group knows how to take care of the univentricle during the ischemic period. You mention that you use antegrade infusion of cold blood cardioplegic solution. Would you give us more detail on the exact technique? Dr. Hillel Laks (Los Angeles. Calif). The myocardial protection is the standard one that we use for either congenital or adult procedures. We use infusion of cold blood cardioplegic solution every 15 to 20 minutes. We have recently begun to use retrograde infusion in addition to that, and we do use glutamate and aspartate warm blood cardioplegic solution at the end of the procedure before reperfusion. We also use topical hypothermia. We agree with you that myocardial protection is critically important in an operation the success of which depends on low end-diastolic pressure at the end of the procedure. Dr. Wells. Do you measure the infusion pressure of the cardioplegic solution? Dr. Laks, Yes. We monitor it always and vary the pressure depending on the age of the child. In younger children we use
The Journal of Thoracic and Cardiovascular Surgery
lower pressures of 80 mm Hg until cardiac arrest and then 50 to 60 mm Hg for subsequent doses. In the older children, 3 to 4 years of age, we generally use a pressure around 100 mm Hg until cardiac arrest occurs and then drop it down to 70 to 80 mm Hg for subsequent doses. Dr. Wells. You point out several potential technical advantages to this procedure. An additional point that we believemay be important is the placement of the coronary sinus in the lower pressure atrial chamber. We believe this may be a physiologic advantage to these patients. Dr. Laks. We do keep the coronary sinus outside the tunnel, on the left atrial side of the tunnel, so that it is in the low-pressure atrium. Dr. Vaughn A. Starnes (Stanford. Calif). I would like to make one comment and then ask you how you are managing it. When I was at the Hospital for Sick Children and de Leval was doing the Fontan procedures, we had one child with thrombosis of the lateral pulmonary connection; we were using PTFE at that time. I am now using autologous pericardium or even bovine pericardium. Do you use any anticoagulants on these children-aspirin, dipyridamole (Persantine), or even warfarin sodium (Coumadin), in the follow-up period? Dr. Stein. We do not use warfarin sodium. Some patients are placed on a regimen of dipyridamole or aspirin postoperatively, but there is no warfarin sodium used. Dr. D. Craig Miller (Stanford. Calif). I am particularly intrigued by the six cases you had of pulmonary atresia and intact septum associated with a "peach pit" RV and RV-coronary sinusoids. Have you not also told us something equally important here in terms of these particular patients' myocardial metabolism, that is, getting oxygenated blood flow retrograde into the coronary arteries? Do you have any autopsy information for the two deaths concerning myointimal proliferation in the coronary arteries, if the patients who died had pulmonary atresia and intact ventricular septum? Dr. Stein. Regarding pulmonary atresia and intact ventricular septum with the coronary connections, there were two patients with that diagnosis. We know that previous reports of decompression of the R V with these connections have resulted in high mortality postoperatively. This technique allows the pulmonary venous return to enter the RV in an unobstructed manner and then provides well-oxygenated blood to the coronary arteries. We also unroofed the coronary sinus at a distance from the RV to help prevent desaturated blood from entering the RV. We hope that this modification will be beneficial long term to these patients. With two patients and a short series, however, we cannot reach any conclusions. There were two late deaths in patients with other diagnoses. One occurred in a patient who had respiratory arrest from viral pneumonia and an anoxic brain injury. Life support was eventually withdrawn because of brain death. The other death occurred in a patient who had a pleural effusion. Dr. Miller. I know how they died, but do you have any histopathologic information if they were patients with pulmonary atresia and intact ventricular septum? Looking at their coronary arteries histologically would be fascinating. Dr. Stein. The two patients with pulmonary atresia and intact ventricular septum did not die. They are still alive. Dr. Laks. I want to comment on two issues. One is when we would use the lateral tunnel procedure as opposed to some other form of connection. As Dr. Stein mentioned, one situation would be tricuspid atresia in which the RA is not large. In that situation there seems no reason to do an atrial septectomy and
Volume 102 Number 2 August 1991
put in a baffle. We simply unroof the coronary sinus into the left atrium, close the coronary sinus orifice, and then do the two caval connections as in the lateral tunnel procedures, using the small RA as part of the tunnel. The other situation would be total anomalous pulmonary venous return to the SVC. Obviously in that situation one has to use a different approach and cannot use a lateral tunnel procedure. Then there are patients with tricuspid atresia or other conditions with transposition, in whom the RA or atrial appendage is lying on the P A. In these patients we have used different types of baffles. After one patient in whom we partially decompressed the RV, we hoped that keeping RV pressures around two-thirds systemic would help supply the coronary circulation in patients with pulmonary atresia, intact septum, and RV--coronary artery connections. The patient who had normal ventricular function preoperatively died of low output. We thought that if the
Functional single ventricle 2 8 7
patients have normal function before the operation with the RV supplying desaturated blood to the coronary arteries, surely they would be better off if we could supply oxygenated blood and leave the RV as the source of coronary blood supply. We have now actually managed three patients that way; in one we used a nonlateral tunnel approach, and in the other two the lateral tunnel approach. Inall of them we allowed oxygenated blood to reach the tricuspid valve without decompression of the RV. All three patients have been restudied with follow-up to about 2 years. We have done thallium scans, positron-emission tomography scans, left ventricular multigated image acquisition analyses, and ejection fractions. The ventricular function is normal, there is no evidence of ischemia, and Holter monitor revealed no ventricular arrhythmias. This may be, at least for awhile, an alternative to transplantation. We need experience with more patients and we need to monitor them very closely because sudden death is still a concern.