C to challenge the medicallsurgical team. Although

Transcription

1 Critical Aortic Stenosis in the First Month of Life: Surgical Results in 26 Infants Tom R. Karl, MD, Shunji Sano, MD, William J. Brawn, FRCS, and Roger B. B. Mee, FRACS Victorian Pediatric Cardiac Surgical Unit, Royal Children's Hospital, Melbourne, Australia Between 1980 and 1989,26 infants aged less than 1 month underwent open aortic valvotomy for critical aortic stenosis. All had congestive heart failure requiring inotropic agents (58%), ventilation (42%), and/or prostaglandin El (35% preoperatively. Nine patients with isolated aortic stenosis had an operative mortality of 0%, whereas 17 patients with other anatomical lesions had a 47% mortality (p < 0.01). Univariate analysis failed to identify additional risk factors other than year of < 0.05). There were four late deaths, three probably related to arrhythmia (actuarial survival at 113 months = 0.53). Two patients have required late reoperation; neither required valve replacement. (Ann Thorac Surg 1990;50:105-9) ritical aortic stenosis (AS) in the newborn continues C to challenge the medicallsurgical team. Although results worldwide have improved, substantial morbidity and mortality are still associated with the various surgical techniques in use [1-8]. We analyzed our own 10-year experience using one surgical technique. The factors affecting long-term and short-term outcome in the current era were examined. Our protocols and their shortcomings for diagnosing and treating critical AS in the newborn are discussed. Material and Methods Patient Population During the period of this study, 152 children underwent aortic valvotomy at the Royal Children's Hospital (3% of all cardiac procedures). We excluded from consideration all patients aged more than 30 days at the time of operation. We believe, as do others, that the ductdependent physiology, usual preoperative condition, and tolerance to cardiopulmonary bypass of the newborn differ from those of older patients [4, 91. There were therefore 26 infants in this series, with a mean age of 8.7 days and a mean weight of 3.2 kg. All patients had serious congestive heart failure requiring inotropic support (58%), ventilation (42%), and/or administration of prostaglandin El (PGE,) (35%) preoperatively. The diagnosis was usually made clinically and confirmed with two-dimensional echocardiogram alone (22/26) or cardiac catheterization plus echocardiogram (426). Mean Doppler-derived or directly measured left ventricle to aortic gradient was 59 mm Hg (standard deviation, 15 mm Hg). Associated Accepted for publication Feb 23, Mr Brawn's current address is Children's Hospital, Ladywood, Middleway, Birmingham, UK B168ET. Address reprint requests to Mr Mee, Victorian Pediatric Cardiac Surgical Unit, Royal Children's Hospital, Flemington Rd, Parkville, Victoria, 3052 Australia. anatomical lesions including persistent ductus arteriosus (PDA) are shown in Table 1. We considered PDA a relatively constant hemodynamic feature of newborns with critical AS, and because our practice is to ligate the duct in all cases, it was not considered an additional lesion. Furthermore, all patients would currently receive PGE, preoperatively to maintain ductal patency. Recognizing the continuum of left heart obstructive lesions, we have excluded from this series patients with aortic atresia, mitral atresia, or both, for whom aortic valvotomy as primary treatment would not have been appropriate. Such patients were treated with various modifications of Norwood's operation, and results will be analyzed in another article. All patients admitted with critical AS survived to undergo open aortic valvotomy, which was the only procedure used during this period for this congenital lesion. Surgical Techniques and Anatomical Finding For patients whom we judged to have adequate-sized left ventricles and in whom aortic valvotomy was deemed to be the method of choice to relieve left ventricular outflow obstruction we used cardiopulmonary bypass, systemic hypothermia (30" to 32"C), and 4 C crystalloid cardioplegic arrest. A single right atrial cannula was used for venous return. The ductus arteriosus was routinely ligated after cardiopulmonary bypass was started. The aortic valve and subaortic area were inspected through an oblique aortotomy. Twenty-two of 26 patients had a bicuspid valve, 3 had a tricuspid valve, and 1 had a unicuspid valve. Only identifiable commissures were incised to the annulus to make the orifice as large as possible. When present, dysplastic nodules or excrescences on the leaflet edges were resected to decrease the bulk obstructing the orifice and to improve mobility and coaptation. The bypass and operative techniques were modified as required to perform the concomitant procedures shown in by The Society of Thoracic Surgeons /90/$3.50

2 106 KARLETAL Ann Thorac Surg 1990;50: 1059 Table 1. Lesions Associated With Critical Aortic Stenosis in 26 Infants Lesion No. of Patients Patent ductus arteriosus only 9 Coarctation 6 Atrial septal defect 6 Ventricular septal defect 2 Mitral stenosis 6 Aortic annulus less than 5 mm 2 Hypoplastic left ventricle 3 Table 2. Of the last 3 patients shown in Table 2, 2 were initially treated with aortic valvotomy and could not be separated from cardiopulmonary bypass owing to inadequate size of the left ventricle. Atrial septectomy, proximal aortopulmonary window, and distal pulmonary artery banding were performed. The third patient underwent Gore-Tex innominate to right pulmonary artery shunt plus proximal pulmonary artery to aortic endto-side anastomosis with division and patch closure of distal pulmonary artery, plus atrial septectomy. None of these patients survived the immediate postoperative period. Patients with this anatomy currently would be treated primarily with Norwoods operation and would not be considered for aortic valvotomy as a definitive procedure. Of the 6 patients with coarctation, 1 had repair before aortic valvotomy, 2 had repair as a concomitant procedure, and 3 had repair as a second procedure during the same hospitalization (the diagnosis of coarctation was made after aortic valvotomy). Five other patients required rtzoperation in the immediate postoperative period, 1 for ventricular septal defect closure, 1 for modified Konno operation (after valvotomy failed to relieve the annular obstruction), 1 for mitral valve replacement, 1 for revision of a pulmonary artery band, and 1 for valvoplasty to reduce aortic insufficiency. All patients were kept sedated, paralyzed, and ventilated for 12 to 24 hours or longer as required, during which time right and left atrial and pulmonary artery pressures were monitored. All patients operated on after 1985 had peritoneal dialysis catheters inserted at the time of operation for control of postoperative fluid and electrolyte balance and temperature. During the last half of this study, lignocaine infusion was started intraoperatively and maintained until the babies could be converted to oral phenytoin, which was then continued for 6 months postoperatively. Follow-up Follow-up for 1 to 113 months (mean, 37 months) after discharge was available for all hospital survivors in this study. Postoperative data were collected by us or by our referring cardiologists. All patients living in our own state were examined in our institution. For out of state or overseas patients, information was obtained from the referring cardiologist responsible for the child s long-term care. Statistical Analysis Characteristics of the subjects in the study were compared using Fisher s exact, x2, or t tests for discrete or continuous data as indicated. Actuarial data were calculated using the Kaplan-Meier formulas [lo]. Sample proportions are expressed with confidence limits corresponding to one standard deviation with continuity corrections for upper and lower limits [ll]. Results Early Outcome Overall operative mortality was 8/26 (31%; confidence limits, 21% to 45%). The babies with isolated AS with or without PDA had an operative mortality of 0/9 (0%; confidence limits, 0% to 19%). The patients with complex Tzble 2. Operations Performed Concomitantly or During Same Hospitalization as Aortic Valvotomy - No. of Late Operation (in Addition to Aortic Valvotomy) Patients Hospital Deaths Deaths 1 PDA ligation only 11 1 (low cardiac output owing to 3 hypoplastic MV) 2 Coarctation repair Coarctation repair + ASD closure Coarctation repair + VSD closure Coarctation repair + ASD/VSD closure MV repair 1 1 (unable to wean from CPB)... 7 MV replacement ASD closure 3 2 (hypoplastic MV, arrhythmia) 0 9 Konno operation 1 1 (low cardiac output) Hypoplastic left heart operation (low cardiac output)... ASD = atrial septal defect; defect. CPB = cardiopulmonary bypass; MV = mitral valve; PDA = patent ductus arteriosus; VSD = ventricular septal

3 Ann Thorac Surg 1990;50:105-9 KARLETAL (.06 26) Fig 1. Actuarial survival after aortic valvotomy. Standard deviations of probability estimates and number of patients completing each follow-up interval are shown in parentheses > a 3 cn u >- k 1 m a (.08,23) PO STOP E RAT I VE DAYS AS had an operative mortality of 8/17 (47%; confidence limits, 39% to 62%). Details of early postoperative deaths are shown in Table 2. Factors found to influence early mortality were presence of an additional anatomical lesion (p < 0.01) and operation before the last 2 years of the study (p < 0.05). Endocardia1 fibroelastosis (EFE) was identified by preoperative echocardiogram in 2 survivors and 2 nonsurvivors, but was detected postmortem in all 5 nonsurvivors undergoing autopsy. Late Outcome Follow-up of 1,194 patient-months after discharge disclosed four late deaths. Three patients had sudden death at home 2, 3, and 4 months after operation. These deaths were presumed to result from arrhythmias. One patient had initially undergone coarctation repair plus atrial septal defect closure plus aortic valvotomy and was hemodynamically well at evaluation 1 month before death on no medications. The second patient had isolated AS and was well just before death, receiving digoxin, furosemide, and phenytoin. Postmortem examination showed advanced EFE, although the AS was well relieved. The third late death occurred in a patient known to have severe EFE preoperatively. He was clinically well just before death, but autopsy showed a small dysplastic mitral valve in addition to the EFE. The fourth late death occurred during a bout of severe gastroenteritis 38 months after the original operation. Postmortem examination showed severe EFE. In the interim, this patient had required two reoperations for valvar and subvalvar AS although she had been asymptomatic. Postmortem examination showed severe EFE. One other patient required reoperation 9 years after valvotomy. At the original operation, she had a dysplastic valve and underwent unicommisurotomy. At the second operation, four true cusps were noted, with minimal valvar stenosis but severe membranous and muscular subaortic stenosis. Her obstruction was successfully relieved by resection. Of the remaining survivors, 3 have mild aortic insufficiency and none has more than a 40-mm Hg transaortic Doppler-derived gradient. Actuarial survival for all newborns with critical AS was 0.53 at 113 months (strictly speaking, at 38 months, with no deaths occurring after that time), as shown in Figure 1. The small number of late deaths does not lend itself to further meaningful statistical analysis except to indicate that hospital survivors with both complex and isolated AS are at risk. Comment Various surgical techniques for treatment of AS in critically ill newborns have been proposed. These include valvotomy under inflow occlusion [5], transventricular or transaortic closed valvotomy [l-3, 6, 71, and open valvotomy with cardiopulmonary bypass and aortic crossclamping with or without cardioplegia [8, 121. Hypothermic circulatory arrest with initial surface cooling has also been used [13]. Although comparisons of techniques in different institutions and heterogenous patient populations may be misleading, our technique compares favorably with other techniques for the group of patients with isolated AS. Results with the more complex AS lesions have been less encouraging. Analysis of our own patients in this group suggests that at least six of the eight operative deaths occurred in patients with degrees of left heart hypoplasia more severe than valvar AS (small left ventricle, small aortic annulus, small mitral valve, subvalvar

4 108 KARLETAL Ann Thorac Surg lyyo;50:1059 stenosis). The prognostic significance of preoperative left ventricular volume determination is controversial. Gun- (dry and Behrendt [4] reported similar left ventricular end-systolic indexes for survivors and nonsurvivors in infants undergoing valvotomy for critical AS. Other investigators [14], however, have suggested that survival is unlikely with indexes less than 20 cm3/m2 [14]. Although we have not rigorously measured left ventricular volume in all cases, we currently consider size of the aortic annulus, ascending aorta, mitral valve, and left ventricular cavity when deciding which patients are suitable candidates for aortic valvotomy as initial treatment [8]. If the cardiac apex is not formed by the left ventricle, or if dimensions are less than 60% mean for body weight, we currently do not regard aortic valvotomy alone as appropriate treatment. This concept was not uniformly applied, however, in the earlier years of this series. Future results might be improved if all patients at this end of the spectrum could be identified with more certainty preoperatively and treated primarily with Norwood-type procedures or cardiac transplantation. The exact incidence and degree of EFE in survivors is currently unknown. The high incidence and severity of this lesion at autopsy, however, suggests that some babies with complex AS and EFE might also be better candidates for cardiac transplantation than for aortic valvotomy (we have not yet performed transplantations for this indication). Our current treatment protocol emphasizes noninvasive diagnosis and adequate resuscitation before operation [5, 141. Although PGE, has been available since 1983 ir our unit, it did not figure statistically in reducing operative mortality in the complex group. We do find PGE, to be an extremely useful resuscitation drug, however, and currently use it in all patients, as has been advocated by other investigators [15]. We also consider inotropic agents and ventilation important aspects of preoperative treatment in most patients, deferring operation until the baby is metabolically and hemodynamically more stable. Regarding our surgical technique, we favor use of hypothermic (32 C) cardiopulmonary bypass with cardioplegic arrest. This technique allows a detailed inspection of the valve and a more precise repair than indirect techniques. Incising true commissures to the annulus provides a substantial increase in orifice size as compared with leaving 1 mm of fusion and does not lead to an appreciable increase in insufficiency. We also believe that debridement of the valve is important, especially if the configuration is trileaflet. Open repair makes this possible. In a bileaflet valve, a commissurotomy will usually lead to a substantial increase in orifice size, whereas a small trileaflet valve often has fairly well-developed commissures before valvotomy. In the latter case, debridement also allows further increase in orifice size without contributing to insufficiency. Currently, we routinely close the foramen ovale through a small right atrial incision, using a very brief period of total circulatory arrest at 32 C. Associated mitral insufficiency of moderate degree is very likely to resolve after successful left ventricular decompression; in our experience, this does not require inspection or specific treatment at the time of valvotomy. Postoperatively, we have found rapid atrial pacing to be useful in increasing cardiac output in infants with very hypertrophied left ventricles and relatively fixed stroke volume. We also have a low threshold for initiation of low-volume (10 ml/kg) peritoneal dialysis as adjunctive treatment for acidosis, renal insufficiency, electrolyte disorders, and hypothermia or hyperthermia. The 3 sudden deaths at 2, 3, and 4 months, as well as some hospital deaths, are presumed to be related to arrhythmias. Because we have not observed complications attributable to long-term antiarrhythmic therapy, we currently use it for 6 months, recognizing that efficacy would be difficult to establish statistically. We are aware that catheter balloon valvoplasty has been used successfully for treatment of newborn AS, although not in our own institution. As equipment and techniques improve, this procedure may be more widely and successfully applied. In our opinion it has not yet been proved to be superior to surgical treatment in either simple or complex newborn AS. The procedure is not free of complications, and long-term data are not available [16]. With regard to long-term follow-up, our data suggest that for survivors of the newborn period hemodynamically important residual AS and insufficiency are uncommon and not rapidly progressive during childhood. We presume that some of these patients will require repeat valvotomy or valve replacement later in life, because our longest follow-up is only 9 years. We conclude that isolated AS in the newborn can be treated safely with open aortic valvotomy on cardiopulmonary bypass with the expectation of good mediumterm results and low incidence of reoperation during childhood. Aortic stenosis complicated by other clinically significant hemodynamic lesions has a less favorable prognosis immediately after operation, but hospital survivors are likely to have medium-term results similar to those of patients with isolated AS. References 1. Midgley FM, Perry LW, Shapiro SR, et al. Closed transventricular valvotomy for critical aortic stenosis in the neonate. In: Doyle EF, Engle MA, Gersony WM, et al, eds. Pediatric cardiology. Proceedings of the 2nd World Congress. New York: Springer Verlag, 1986:54&1. 2. Brown JW, Robison RG, Waller BF. Transventricular balloon catheter aortic valvotomy in neonates. Ann Thorac Surg 1985;39: Trinkle JK, Grover FL, Arom KV. Closed aortic valvotomy in infants. J Thorac Cardiovasc Surg 1978;76:19% Gundry SR, Behrendt DM. Prognostic factors in valvotomy for critical aortic stenosis in infancy. J Thorac Cardiovasc Surg 1986;92: Sink JD, Smallhorn JF, Macartney FJ, et al. Management of critical aortic stenosis in infancy. J Thorac Cardiovasc Surg 1984; Duncan K, Sullivan 1, Robinson P, et al. Transventricular aortic valvotomy for critical aortic stenosis in infants. J Thorac Cardiovasc Surg 1987;93:

5 Ann Thorac Surg 1990;50: 1059 KARLETAL Pelech AN, Dyck JD, Trusler GA, et al. Critical aortic stenosis: survival and management. J Thorac Cardiovasc Surg 1987;94:51&7. 8. Messina LM, Turley K, Stanger P, et al. Successful aortic valvotomy for severe congenital aortic valvular stenosis in the newborn infant. J Thorac Cardiovasc Surg 1984;88: Turley K. Commentary on [4]. 10. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53: Brownlee KA. Statistical theory and methodology in science and engineering. 2nd ed. New York: Wiley, 1965:14& Wheller JJ, Hosier DM, Teske DW, et al. Results of operation for aortic valve stenosis in infants, children and adolescents. J Thorac Cardiovasc Surg 1988;96: Kirklin J, Barratt-Boyes 8. Cardiac surgery. New York Wiley, 1986: Keane JF, Norwood WI, Bernhard WF. Surgery for aortic stenosis in infancy. Circulation 1983;68(Suppl 3): Jonas RA, Lang P, Mayer JE, et al. The importance of PGE, in resuscitation of the neonate with critical aortic stenosis. J Thorac Cardiovasc Surg 1985;89: Sholler GF, Keane JF, Perry SB, et al. Balloon dilation of congenital aortic valve stenosis. Results and influence of technical and morphologic features on outcome. Circulation 1988;78: